ACNP 57th Annual Meeting: Poster Session I

    Sponsorship Statement: Publication of this supplement is sponsored by the ACNP.

    Individual contributor disclosures may be found within the abstracts. Asterisks in the author lists indicate presenter of the abstract at the annual meeting.

    M1. Lifespan Effects of Early Life Stress on Aging-Related Trajectory of Memory Decline

    Benedetta Bigio*, Danielle Zelli, Timothy Lau, Paolo de Angelis, Daniella Miller, Jonathan Lai, Anisha Kalidindi, Susan Harvey, Anjali Ferris, Aleksander Mathe, Francis Lee, Natalie Rasgon, Bruce McEwen, Carla Nasca

    Rockefeller University, New York, New York, United States

    Background: Early life adversities (ELA), such as variations in maternal care of offspring, are critical factors underlying the individual likelihood to development of multiple psychiatric and medical disorders. For example, our new translation findings suggest a role of ELA in the form of childhood trauma on development of metabolic dysfunction, such as a deficiency in acetyl-L-carnitine (LAC) and insulin resistance (IR), in patients with major depression (MDD). Arising from the postulate of insulin resistance (IR) as a pathophysiological link in diabetes, depression and dementia (“the 3 D’s”), we implemented a mating rodent model with high and low nurturing dams to study the lifespan effects of ELA on peripheral metabolic physiology, central glucocorticoid function and behavioral states.

    Methods: We devised a computerized setup to record and assess the maternal care provided by wild-type or heterozygous BDNF val66met (hets) dams on offspring which were screened for individual anxiety-like behavior at the light dark, peripheral markers of IR and central glucocorticoid function in adulthood, and subsequently assessed for memory function in later age. Maternal care was assessed on the basis of previously employed ethological parameters by analyzing in and out of the nest caring and self-maintenance behaviors. Behavioral and molecular analyses in adulthood and midlife mice were performed as previously described by four observers who were blind to the sample conditions. Two-tailed t-tests and multiple regression were used as appropriate to specific analyses.

    Results: We find that offspring receiving low quality maternal care in the early life show peripheral IR in adulthood as suggested by increased levels of insulin, glucose and tryglicerides in comparison with offspring that received high quality maternal care. This state of IR in adulthood is concomitant with increased expression of the mineralocorticoid receptors (MR) in hippocampus and corresponding anxiety-like behavior at the light-dark test in low nurtured offspring. Furthermore, low nurtured mice showed cognitive impairments in contextual memory at the Y maze in midlife, suggesting dementia-like symptoms, which were not apparent before.

    Conclusions: These findings suggest a role for ELA in the form of poor maternal care in increasing the likelihood to development of peripheral IR, altered central glucocorticoid function and corresponding anxiety states in adulthood, and that these factors may encode lifelong susceptibility to pathophysiological aging. Given our earlier reported association between IR and a LAC deficiency, a candidate biomarker of major depression that is a risk factor for aging-associated memory decline, we are currently assessing LAC levels in this mechanistic framework. This model may provide endpoints for identification of early windows of opportunities for preemptive tailored interventions.

    Keywords: Early Life Adversity, Glucocorticoids, Insulin Resistance, Glutamate, Memory Function

    Disclosure: Nothing to disclose.

    M2. Genetic Signatures of Response to Combination Escitalopram-Memantine Treatment for Geriatric Depression: A Randomized, Double-Blinded, Placebo-Controlled Pilot Study

    Adrienne Grzenda*, Helen Lavretsky, Prabha Siddarth, Natalie St. Cyr

    UCLA, Los Angeles, California, United States

    Background: Geriatric depression is frequently associated with mild cognitive impairment (MCI), which responds poorly to antidepressant monotherapy. Furthermore, comorbid depression in patients with cognitive impairment appears to accelerate the neurodegenerative process, accelerating conversation to dementia. Our objective was to evaluate the efficacy of memantine augmentation of escitalopram treatment in elderly patients in improving clinical and cognitive outcome. Determine if unique gene signatures underlie escitalopram-memantine treatment response compared to escitalopram alone.

    Methods: We conducted a 12-month double-blinded, placebo-controlled trial in older adults to assess the efficacy of combination memantine and escitalopram therapy compared to escitalopram and placebo for the treatment of depression ( NCT01902004). For the current study, we analyzed the first 37 participants to completed 6-month follow up. The primary outcome was remission of depressive symptoms, defined as Hamilton Depression Rating Scale (HAM-D) scores of 6 or less for 3 consecutive weeks following initiation of treatment. Secondary measures included apathy, quality of life, and cognition. Genome-wide transcriptional profiles were generated from peripheral blood leukocytes sampled at baseline and 3 months following initiation of treatment.

    Results: Escitalopram daily doses ranged between 10-20 mg; memantine daily doses ranged between 10-20 mg. Both treatment groups demonstrated similar remission rates at 6-month follow-up with significant improvement in depression, anxiety, and overall clinical illness severity. Escitalopram with memantine treatment resulted in additional improvements in apathy, quality of life, and intermediate/delayed memory. There exists unique baseline and post-treatment gene signatures that predicted response to combination escitalopram-memantine compared to escitalopram-placebo treatment. Furthermore, bioinformatics analysis of differentially expressed gene targets indicates unique and synergistic mechanisms of action mediating antidepressant response in monotherapy and memantine augmentation.

    Conclusions: Pilot data indicate that memantine augmentation of escitalopram treatment in geriatric depression results in an enhanced clinical response profile, including improvements in apathy, resilience, quality of life, and cognition. There was no difference in depression remission rates between the two groups. Remission of depression, however, was associated with unique gene signatures in each treatment group, which may direct selection of patients to receive combination therapy in future precision medicine initiatives. The current findings need to be replicated in a larger sample with additional time-course data to determine if additional differences exist between the treatments remission rate and overall degree of response.

    Keywords: Geriatric Depression, Antidepressant Response, Neuroprotection, Escitalopram, Memantine

    Disclosure: Nothing to disclose.

    M3. Online Repetitive Transcranial Magnetic Stimulation Enhances Working Memory Performance in Younger and Older Adults

    Lysianne Beynel, Simon W. Davis, Susan Hilbig, Wesley Lim, Angel V. Peterchev, Roberto Cabeza, Greg Appelbaum, Sarah Lisanby, Bruce Luber*

    National Institute of Mental Health, Bethesda, Maryland, United States

    Background: Working memory (WM) is the ability to perform mental operations on information that is stored in a flexible, limited capacity buffer. Through the interface between long-term memories and moment-to-moment information available in the environment, WM allows humans to carry out successful goal-directed behaviors. As such, WM is widely involved in the achievement of complex tasks such as learning, decision making, and reasoning. Critically, decline in WM is a major factor in cognitive impairment that accompanies healthy aging. Therefore, many different approaches are being explored as possible interventions to enhance these abilities, including non-invasive brain stimulation methods such as repetitive transcranial magnetic stimulation (rTMS). Previously, we have shown 5 Hz rTMS to enhance WM performance in different groups of young and older healthy adults. However, these studies only examined only the maintenance component of WM, the process that sustains information in an activated state, rather than the ability to manipulate information in WM (the “working” part of WM), which is central to many aspects of human cognition. Further, compared to maintenance, manipulation is more dependent on dorsolateral prefrontal cortex (DLPFC) which is one of the brain regions shown to be most impaired by aging, although as of yet, no rTMS studies have tried to enhance WM manipulation by stimulating DLPFC. The current study tested the capacity to enhance working memory manipulation with online 5 Hz rTMS to DLPFC in healthy young and older adults.

    Methods: Online high frequency rTMS was applied over the left DLPFC to test the hypothesis that active rTMS would lead to significant improvements in memory recall accuracy compared to sham stimulation, and that these effects would be most pronounced in WM manipulation conditions with the highest cognitive demand in both young and older healthy adults. rTMS was applied while participants performed a delayed response alphabetization task with three individually-titrated levels of difficulty. The TMS coil placement was optimized by integrating electric field modeling with individualized functional magnetic resonance imaging activation maps in left DLPFC based on the WM task and was implemented during the experiment using stereotactic neuronavigation with real-time robotic guidance, allowing accurate targeting during the stimulation. Participants were scheduled for 6 sessions. The first visit consisted of consenting, exclusionary screening, resting motor threshold (rMT) assessment and 6 blocks of practice with the behavioral task. Participants then returned for an MRI visit and four more rTMS sessions. On each rTMS visit, 25 pulses of 5 Hz rTMS were applied at rMT while subjects performed the WM task with the titrated 3 difficulty levels (Easy, Medium, and Hard), with accuracy feedback given at the end of each block. Ten blocks of the task were performed: a first block without stimulation, four blocks of active or sham stimulation, one block without stimulation, and four more blocks with the sham or active stimulation. The order of stimulation type (Active or Sham) was presented according to an ABBA schedule

    Results: Twenty-nine young adults (17 females and 12 males; 22.9 ± 4.8 years old) and eighteen older adults (13 females and 5 males; age: 69.7 ± 4.8 years old) completed the full protocol. A significant interaction between Difficulty Level and Stimulation Type (Active or Sham rTMS) was found (F(2, 88) = 8.49, p < .001). The decomposition of this interaction with Bonferroni correction revealed a significant difference in the Hard condition, with active rTMS producing increased accuracy compared to sham rTMS (F(1,44) = 17.15, p = .006). Investigating this effect in both young and older groups separately, the one-way interaction between Difficulty and Stimulation Type, was found to be significant for both the young (F(2, 54) = 4.11, p = .022) and the older cohort (F(2, 32) = 4.79, p = .015). Thus, the results yielded two main findings. First, active 5 Hz rTMS delivered to individualized left DLPFC targets, identified with fMRI and electric field modeling, significantly enhanced WM manipulation abilities over the course of the intervention, but only in the most difficult condition of the task, namely the hardest difficulty level. Second, active versus sham rTMS produced similar effects in young and older adults, suggesting that rTMS enhancement was not modulated by age.

    Conclusions: The current study showed that 5 Hz rTMS applied over the left dorsolateral prefrontal cortex can enhance WM manipulation abilities, but only in the most difficult condition, with no differences between young and older adults. This result points towards further investigation of rTMS for memory enhancement, focusing on high difficulty conditions as those most likely to exhibit benefits.

    Keywords: TMS, Working Memory, Aging, Memory Enhancement, TMS Targeting

    Disclosure: Nothing to disclose.

    M4. Individualized Structural Connectome for Diagnostic and Prognostic Prediction for Alzheimer’s Disease

    Yun Wang*, Chenxiao Xu, Ji-Hwan Park, Seonjoo Lee, Yaakov Stern, Shinjae Yoo, Hyoung Seop Kim, Jiook Cha

    Columbia University/New York State Psychiatric Institute, New York City, New York, United States

    Background: There is an urgent, unmet need for clinically useful biomarkers for Alzheimer’s disease (AD) based on non-invasive and affordable measures suited for screening for individuals with subthreshold symptoms. More Studies have been focusing on brain MRI-derived markers. Cortical thinning and reduced hippocampal volumes based on structural MRI are known for markers for Alzheimer’s disease, but these structural estimates alone are insufficient for implementation at clinical settings because of insufficient accuracy and generalizability. Here we tested the utility of the white matter structural connectomes in predicting diagnosis, and disease trajectory.

    Methods: We used data from 208 elderly patients who visited the dementia clinic at National Health Insurance Service Ilsan Hospital (NHIS-IH), Goyang, South Korea from 2010 to 2015. Participants included 110 with diagnosis of Alzheimer’s disease (AD; median age = 82; interquartile intervals (Q3-Q1) = 85-77), 62 with mild cognitive impairment (MCI; median age = 73; Q3-Q1 = 77-66), and 37 subjective mild cognitive impairment (SMC; median age = 74; Q3-Q1 = 78-72). To test the generalizability of our approach, we also used ADNI-2 (Alzheimer’s Disease Neuroimaging Initiative), of which structural and diffusion MRI was collected.

    For both data, using multi-modal brain MRI (structural and diffusion MRI), we performed high-throughput brain phenotyping, including automated morphometry and white matter structural connectomics (probabilistic tractography) to generate large-scale multi-modal, multi-parameteric imaging-derived phenotypes used as features in machine learning. For structural MRI analysis, firstly, we estimated morphometrics measures using Freesurfer image analysis pipeline from T1 and T2-FLAIR images. Morphometric measures (N = 948 per subject) include volumes of the hippocampal subdivisions, and thickness, surface area, and volume of cortical/subcortical regions using two different atlases available in freesurfer. For diffusion MRI analysis, we used MRtrix 3. The connectome measures (N = 33,698 per subject) include counts of streamlines, a surrogate measure of structural connectivity, and mean length of streamlines given any two brain regions based on multiple atlases.

    We built several machine learning models using the large-scale brain MRI-derived phenotypes to predict diagnosis of AD and MCI, respectively. We benchmarked three classifiers available at a python library for machine learning, scikit-learn (4): random forest, logistic regression (LR) with L2 regularization, and support vector machine (SVM) with linear kernel. Model training and validation was done using nested cross validation to avoid overfitting due to bias to training data (5, 6). For hyper-parameter optimization, we used the grid search method, varying C parameter for SVM and LR classifier, and varying the number of estimators and the minimum samples per leaf for random forest classifier. We used 3-fold cross validation with 10 iterations with stratification by target labels (repeated, stratified, nested 3-fold cross validation). We estimated accuracy, sensitivity, specificity, F1 score, and receiver operating characteristic (ROC) curve from repeated 3-fold cross-validation to evaluate the performance of our classifiers.

    Results: In dataset1, for classification of AD vs. SMC, “morphometry + connectome” model showed optimal classification accuracy (accuracy = 0.99; RF. While “connectome” model showed similar accuracy (accuracy = 0.99; RF), “morphometry” model showed least significant accuracy (accuracy = 0.88; RF). In classification of MCI vs. SMC, “morphometry + connectome” models also showed best classification accuracy (accuracy = 0.90; RF), followed by “connectome” model (accuracy = 0.88; RF) and, “morphometry” model showed poor performance (accuracy = 0.86; RF). In classification of AD vs. MCI, “morphometry + connectome” models showed a best accuracy (accuracy = 0.99; RF), followed by “connectome” model (accuracy = 0.99; RF), and “morphometry” model (accuracy = 0.87; RF). Throughput all classifications, connectomes and morphometry showed greater diagnostic accuracies compared with the WMH benchmark. In dataset2, to compare the performance of our classifiers, we used models based on CSF biomarkers as a benchmark. In classifying diagnosis of AD and MCI, the combined connectome and morphometric model shows similar accuracy compared with the CSF benchmark model (0.63 vs 0.69). Of these samples, data from 60 elders with initial diagnosis of MCI who were followed at least two years were used to test prognosis. The same CSF benchmark was used. In predicting MCI-to-AD progression, all MRI-based models outperformed the CSF benchmark. However, in this case, the combined connectome and morphometric model underperformed either connectome or morphometric only model.

    Conclusions: In this study, we showed that large-scale MRI-derived brain phenotypes could be used to accurately classify Alzheimer’s Disease (AD) and Mild Cognitive Impairment (MCI) and to predict disease progression to AD from MCI in multiple datasets with moderate sample sizes. These promising results show potential utility of white matter structural connectomes in addition to widely use morphometry as a proxy measure of neurodegeneration in AD pathology.

    Keywords: Diffusion Tensor Imaging (DTI), machine learning classification, Alzheimer’s Disease

    Disclosure: Nothing to disclose.

    M5. Sensitivity to Maternal Buffering of Fear in Infancy is Sex-Dependent

    Amanda White, Jacek Debiec*

    University of Michigan, Ann Arbor, Michigan, United States

    Background: Beginning in infancy, social cues play a crucial role in the regulation of emotional state and behavior. In addition to providing their child with nutrition, warmth, and safety, mothers can suppress amygdala activity, cortisol release, emotional distress, and fear learning in their child, an effect termed “maternal buffering.” When infant rats undergo fear conditioning in the presence of a calm mother, they do not acquire the association between the neutral cue and mild foot shock. There is an emerging literature on sex differences in expression of fear and in the social regulation of fear in adults; however, sex differences in maternal buffering of fear have not received much attention. Here, we examined whether maternal presence differentially modulates fear in female and male rat pups.

    Methods: All experiments were conducted using Sprague-Dawley rats. Rat mothers were anesthetized with 50 mg/kg ketamine and 5 mg/kg xylazine. Once dams were fully anesthetized, they were placed in a clean cage lined with an absorbent blue pad. Postnatal day 13 (P13) pups were then placed in the cage with the mother and given a 10-min adaptation period. Following the adaptation period, pups received 11 presentations of a 30-s peppermint odor (CS) and a 1-s, 0.5-mA tail shock unconditioned stimulus (US) (Lafayette scrambled shock generator), with an intertrial interval (ITI) of 4 min. Peppermint odor was delivered by a flow dilution olfactometer (2 L/min flow rate) at concentration of 1:10 peppermint vapor. Paired odor–shock pups received a shock overlapping with the last second of the 30-s odor presentation. On P18, pups were placed in clear plastic boxes and given a 3-min adaptation period. A 30-second CS odor was presented to the pups 3 times with an ITI of 2 minutes. Pup behavior was recorded via video camera. Neuronal activation and endocrine response will be examined in future experiments.

    Results: In females, we found a significant main effect of maternal presence (p = 0.008); females that were conditioned in the presence of an anesthetized mother (n = 6) froze significantly less at test than females that were conditioned in the absence of an anesthetized mother (n = 7). However, we did not observe significant differences between males that were conditioned in the presence of an anesthetized mother (n = 7) and males that were conditioned in the absence of an anesthetized mother (n = 5) (p = 0.75). In subsequent experiments we assessed corticosterone levels and neural activity in pups during maternal buffering of fear.

    Conclusions: Our data demonstrate that female pups are more susceptible to maternal buffering of fear, and suggest that sex differences in social regulation of emotion emerge very early in life. Caregivers buffer their child’s emotional response to stressful events and protect their child’s developing brain from the deleterious effects of stress. A better understanding of the mechanisms underlying maternal buffering of fear may help clinicians improve treatment outcomes for children that have experienced trauma

    Keywords: Fear Conditioning, Maternal Behavior, Infancy, Early Life Stress, Social Stress Buffering

    Disclosure: Nothing to disclose.

    M6. Tracing and Intersectional Optogenetic Approaches to Identify a Novel Serotonergic System That Projects to the Hypothalamus to Increase Arousal and Place Preference and Inhibit Panic

    Cristian Bernabe, Izabela Caliman, Zachary Rodd, Anantha Shekhar, Philip Johnson*

    Indiana University School of Medicine, Indianapolis, Indiana, United States

    Background: Although selective serotonin transporter reuptake inhibitors are the first line pharmacological treatment for severe anxiety disorders such as panic disorder (associated with recurrent panic attacks) and highly comorbid phobias, our understanding of how serotonin regulates panic and fear networks is poorly understood. Our preclinical work has identified that the perifornical hypothalamus (PeF) and local orexin (OX)/glutamate neurons play a critical role in animal models of adaptive and pathological panic, but little is known about how serotonin regulates this system in the context of panic and phobias. Based on previous immunohistochemistry (IHC) and electrophysiological studies showing that 5-HT1a and 5-HT2a receptors are respectively expressed on and inhibit and excite OX/glutamatergic and melanin concentrating hormone (MCH)/GABA neurons in the PeF and lateral hypothalamus (LH), we hypothesized that serotonin is inhibiting panic locally.

    Methods: In experiment 1 we wanted to use optogenetics to confirm that the PeF/LH is pro panic/phobia. We bilaterally injected an adeno-associated viral construct (AAV) with a glutamatergic promoter into the PeF/LH to induce local expression of channelrhodopsin (ChR2) and a fluorescent reporter. Wireless LED implants were later implanted into the PeF/LH to excite ChR2 expressing glutamate neurons. In experiment 2, we injected the retrograde tracer cholera toxin B (CtB) conjugated to a flourophore into the PeF/LH of rats and immunostained for tryptophan hydroxylase to identify serotonergic neurons that project to the PeF/LH. In experiment 3, we next bilaterally injected a canine adenovirus (CAV) with a nonspecific neuronal promoter into the PeF/LH. The CAV is retrogradely transported and induces expression of Cre-recombinase (Cre-R) in all neurons that project to the PeF/LH. We then injected an AAV into the lwDRN and MRN with a nonspecific neuronal promoter to induce a double floxed inverted (DIO) sequence of ChR2 which in the presence of Cre-R is expressed. Wireless LED were later implanted into the lwDRN/MRN.

    Results: In experiment 1 - LED excitation of glutamate neurons in the PeF/LH induced: 1) panic-associated flight/escape behaviors and robust cardioexcitation; 2) and also phobia associated unconditioned and conditioned place avoidance. Ex vivo IHC confirmed that the reporter (evidence of ChR2 expression) was expressed on all local OX neurons (n = 4-5/group). In experiment 2 – Retrograde tracing revealed that CtB expressing serotonergic neurons that project to the PeF/LH were concentrated in the lateral wings of the dorsal (lwDRN) raphe and also medial (MRN) raphe nucleus (n = ~7) and did not overlap with pro-fear amygdala projecting serotonergic neurons in the midline DRN (n = ~15). In experiment 3, LED excitation of ChR2 + lwDRN/MRN terminals in the PeF/LH: 1) increased non-agitated locomotion; 2) unconditioned and conditioned place preference and, 3) inhibited CO2 (a suffocation stimuli) induced panic (e.g., flight/escape behaviors and cardioexcitation). Ex vivo analyses revealed robust expression of the fluorescent reporter (evidence of ChR2 expression) in the lwDRN/MRN and the PeF/LH, but also in collateral fields associated with reward such as the ventral tegmental area and the nucleus accumbens (n = 9-10/group).

    Conclusions: The data here support the hypothesis that serotonergic neurons in the lwDRN and MRN that project to the PeF/LH represents a panic-off circuit that also mobilizes arousal and motivated behaviors. Future studies will focus on elucidating this circuit’s role in reward, addiction, and depression and if inhibition of this circuit induces pathological panic and or depression.

    Keywords: Anxiety, Panic, Agoraphobia, Hypothalamus, Serotonin

    Disclosure: Nothing to disclose.

    M7. Adolescent Intermittent Alcohol Exposure Enhances Sensitivity to Future Stress Events That Promote Abnormal Fear-Related Behavior in Adulthood

    Lawrence Chandler*, Justin Gass

    Medical University of South Carolina, Charleston, South Carolina, United States

    Background: Post-traumatic stress disorder (PTSD) and alcohol use disorder (AUD) are two of the most common mental disorders and are highly comorbid. Accumulating evidence indicates that the neuropathology of both of these conditions arise from engagement of aberrant learning and memory processes, and recent evidence further suggests that the neurocircuitry of alcohol addiction and the neurocircuitry of fear and anxiety disorders overlap and have many common characteristics. It is becoming increasing apparent that abuse of alcohol during adolescence has long lasting effects of brain and behavior in adulthood. Although the important relationship between PTSD and AUD is an increasing focus of research, the question of whether the abuse of alcohol during adolescence adversely impacts how an individual reacts to a traumatic stress event, as well as their resilience to subsequent stressors, has received little attention.

    Methods: Adolescent Long Evans rats were subjected to repeated episodes of alcohol exposure by vapor inhalation followed by a single episode of acute restraint stress to simulate a traumatic event. Rats were subsequently tested for stress resilience using a fear conditioning and extinction procedure.

    Results: We observed that adult rats that had been subjected to adolescent intermittent ethanol (AIE) acquired fear conditioning at a faster rate but exhibited a slower rate of extinction of the fear memory. AIE-exposed rats also exhibited increased freezing behavior during testing for both extinction recall and spontaneous recover. However, treatment with an mGluR5 positive allosteric modulator during fear extinction training attenuated these deficits in extinction learning and retention. An especially interesting observation was that while extinction recall was not altered by stress alone, it had a synergist effect with AIE on extinction recall (e.g., deficits in the ability to retain the extinction memory).

    Conclusions: These findings in rats provide support for the hypothesis that adults with a history of abuse of alcohol during adolescent who experience a traumatic stress event later in life have reduced stress resiliency and are at a significantly greater risk of developing PTSD and AUD and identify potential therapeutic targets for the treatment of these comorbid disorders.

    Keywords: PTSD, Fear Conditioning, Adolescent Alcohol Exposure, Fear Extinction, Alcohol Use Disorder

    Disclosure: Nothing to disclose.

    M8. Prefrontal Regulation of the Extended Amygdala and Anxious Temperament in Young Nonhuman Primates

    Margaux Kenwood*, Jonathan A. Oler, Andrew S. Fox, Do P. Tromp, Rasmus M. Birn, Elisabeth A. Murray, Ned H. Kalin

    University of Wisconsin Madison, Madison, Wisconsin, United States

    Background: Anxious temperament (AT) is an early-life risk factor for the later development of stress-related psychopathology. Using a well characterized nonhuman primate model of AT with FDG-PET in a large sample (n = 592), we identified the AT neural circuit. This circuit includes the bed nucleus of the stria terminalis (BST) and central nucleus of the amygdala (Ce), components of the central extended amygdala (EAc). Studies from our laboratory suggest that the EAc is regulated by the prefrontal cortex, as orbitofrontal cortex (OFC) lesions reduce BST metabolism. The connectivity between these structures is via fibers that are carried in the uncinate fasciculus (UF), a major white matter tract between the OFC and subcortical structures. Interestingly, decreased UF fractional anisotropy (FA) has been demonstrated in adults and children with anxiety disorders. Here, we used a selective lesioning strategy to assess the role of UF fibers in the expression of AT, and multimodal imaging to understand the neural circuit mechanisms underlying this effect.

    Methods: With the intent of disrupting fibers of passage, we lesioned a narrow strip of the posterior OFC in 10 young female rhesus monkeys. 10 age- and sex-matched animals served as cage-mate controls. Before and after the lesions, animals were phenotyped for AT and assessed with multimodal brain imaging (DTI, fMRI, and FDG-PET). To assess lesion-induced alterations of white matter microstructure, deterministic tractography was used to create regions of interest (ROIs) of the UF and other major white matter tracts. Average FA was extracted from these ROIs and examined for lesion-induced changes. To assess changes in BST function resulting from the lesions, threat-related metabolic activity was extracted from the FDG-PET images using a bilateral BST seed. To assess changes in BST functional connectivity resulting from the lesions, fMRI signal was averaged across all voxels in the same BST seed, and voxelwise temporal correlations were computed using AFNI. The behavioral and imaging data were analyzed using a (LESION[post-pre] - CONTROL[post-pre]) design. Finally, to determine whether lesion-induced changes in BST fMRI connectivity occurred in regions that are functionally connected to the BST, intrinsic functional connectivity of the BST was assessed using the same seed in separate animals (n = 378).

    Results: Lesioned animals demonstrated the predicted decrease in AT (Student’s t-test, t = 1.589, p = .06, one-tailed). Extraction of average FA values demonstrated substantial lesion-induced alterations in white matter microstructure of the UF (Student’s t-test, t = 4.16, p = 0.0005, two-tailed). Additionally, in the lesioned animals, the decrease in UF FA was associated with individual differences in the decrease in AT (Pearson’s r = 0.72, p = 0.029). Notably, no other fiber tracts showed significant decreases in FA. Replicating previous findings, lesioned animals showed a significant decrease in BST metabolism (Student’s t-test, t = 3.98, p = 0.0009). Finally, the lesioned animals demonstrated decreased functional coupling between the BST and regions of the left ventrolateral PFC (vlPFC, area 12/47), the upper bank of the rostral portion of the cingulate sulcus (ACC) and the right superior temporal sulcus (p < 0.005, uncorrected). Importantly, all three clusters demonstrating lesion-induced changes in BST functional connectivity overlapped with regions intrinsically coupled to the BST in the large sample (n = 378).

    Conclusions: This study, combining causal manipulations with multimodal imaging in young nonhuman primates, provides insight into the importance of white matter fibers that connect prefrontal with subcortical structures in mediating the early risk to develop stress-related psychopathology. The findings from this study, along with those in children demonstrating altered UF FA in relation to anxiety, highlight the importance of the fibers passing nearby or through the OFC, particularly those of the UF, in mediating anxiety regulation. The results point to BST interactions with frontal cortex in mediating the adaptive and maladaptive regulation of anxiety. Furthermore, the functional connectivity data from the operated monkeys implicates not only OFC, but also the vlPFC and ACC in the regulation of BST function. Taken together, findings from this study demonstrate the value of a strategy that combines causal manipulations with multimodal imaging methods in NHPs to discover mechanisms underlying the early life risk to develop psychopathology.

    Keywords: Extended Amygdala, Orbitofrontal Cortex (OFC), Rhesus

    Disclosure: Nothing to disclose.

    M9. Effect of Exogenous Hormones on Rodent Models of Fear Memory

    Teniel Ramikie*, Kerry Ressler

    McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

    Background: Posttraumatic Stress disorder (PTSD) affects approximately 7% of the US population in their lifetime (Kessler et al., 2005). Alarmingly, PTSD rates are 2-fold higher in women as compared to men and women experience greater PTSD symptom severity and longer symptom duration. Although women and men experience differences in trauma type exposure (e.g. women experience higher rates of sexual assault), PTSD-related sex differences persist even when controlling for trauma type and severity (Breslau et al., 1998). Collectively, these data suggest that increased PTSD vulnerability in women is, in part, shaped by sex differences in non-environmental factors. Notably, epidemiological studies have shown that PTSD-related sex differences rise following puberty onset and fall following menopause, highlighting the impact of fluctuating gonadal hormone levels on shaping PTSD vulnerability in women of reproductive age. In support of this, many studies now show that gonadal hormone status does influence PTSD symptom development in women (Milad et al., 2006; Glover et al., 2012, 2013). While this remains a critical area of study, most of these investigations have focused on endogenous gonadal hormones while only a few studies have examined the role of exogenous gonadal hormones in shaping PTSD outcomes in women of reproductive age. This is an important research area as hormone-based contraceptives (HC), have been used by women for over 50 years and are one of the 1st-lines of treatment for women after sexual assault. In addition to suppressing endogenous gonadal hormone levels and decreasing pregnancy risk, clinical evidence shows that peri-traumatic administration of exogenous hormones, via HC intake, impact memory related PTSD symptoms (Ferree et al., 2012)—suggesting that exogenous hormones also shape fear memory. The biological mechanisms through which exogenous gonadal hormones shape emotional memory remains unknown. Our study will begin to address this question by examining the impact of exogenous gonadal hormones on fear memory in mice.

    Methods: All experiments were carried out using female C57/BL6J mice, 8-12 weeks of age. Each experimental group had 8-12 mice. Estrous cycle was examined using vaginal cytology and subsequent H&E staining. HC used for these experiments was administered orally, via drinking water. Either HC-containing or control drinking water was the only source of water available to experimental groups. An estradiol (E2) ELISA was used to confirm HC biological efficacy. Auditory fear conditioning and fear extinction behavioral experiments were performed to assess fear memory-related behaviors (% freezing), startle chambers were used to measure shock reactivity; open field (OF) and elevated plus maze (EPM) behavioral protocols were used to assess anxiety-like behaviors. Two-way ANOVA and Student’s t-test were used to analyze group differences. Data analysis was performed using GraphPad Prism 7.

    Results: Vaginal cytology and ELISAs were used to characterize the efficacy of sub-chronic and chronic HC intake at different concentrations. Results revealed that chronic HC intake effectively suppressed E2 levels (measured as a % of proestrus levels) in HC treated groups as compared to controls (sub-chronic HC intake F(2,9) = 0.0505, p = 0.951; chronic HC intake F(2,7) = 65.13 p < 0.0001). We assessed the effects of chronic HC use on basal measures of anxiety-like behaviors using OF and EPM. These results revealed that chronic HC did not affect measures of anxiety-like behaviors or locomotor measures (control vs. HC-treated groups, p > 0.05).

    Next, we assessed the effects of peri-traumatic HC administration on fear memory. To do this, we assessed the following forms of peri-traumatic HC intake: 1) acute HC intake immediately following trauma, 2) chronic HC intake following trauma and 3) chronic HC intake prior to and after trauma. These HC intake protocols best reflected the HC administration methods used in clinical settings (Ferree et al., 2012). These experiments revealed that acute HC intake immediately following fear conditioning had no effect on fear learning or fear expression but significantly improved fear extinction memory in HC treated group as compared to controls (Drug effect: F(1,22) = 4.611, p < 0.05). In addition, chronic HC intake immediately following fear conditioning led to both decreased fear expression in HC-treated groups as compared to controls (p < 0.05) and enhanced fear extinction learning and retention in HC-treated groups as compared to controls (Drug effect: F(1,22) = 7.811, p < 0.05). Lastly, our subsequent experiments revealed that chronic HC intake prior to and after fear conditioning had no effect on fear learning (p > 0.05) but resulted in a relatively greater decrease in fear expression (p < 0.01) and also enhanced fear extinction memory in HC treated groups as compared to controls (Drug effect: F(1,22) = 7.811, p < 0.01). Given evidence of gonadal hormone effects on pain sensitivity, we also measured the pain sensitivity of our HC-treated animals and found no difference between our control and HC-treated groups (p > 0.05).

    Conclusions: These results indicate that though HC-intake does not affect basal levels of anxiety-like behaviors or pain sensitivity, acute and chronic exogenous gonadal hormone intake, via HC, decrease fear conditioning recall and enhance fear extinction memory. Collectively, these experimental outcomes indicate that peri-traumatic exogenous hormone intake can positively modulate fear-related memory.

    Keywords: Gonadal Hormones, Fear Memory, Sex Differences

    Disclosure: Nothing to disclose.

    M10. A Subpopulation of Extended Amygdala Neurons Drives Physiological Arousal and Anxiety States

    Jose Rodriguez-Romaguera*, Randall Ung, Hiroshi Nomura, James Otis, Marcus Basiri, Pengcheng Zhou, Hanna van den Munkhof, Vijay Namboodiri, Xueqi Zhu, J. Elliott Robinson, Jenna McHenry, Shanna Resendez, Louisa Eckman, Oksana Kosyk, Liam Paninski, Thomas Jhou, Thomas Kash, Michael Bruchas, Garret Stuber

    University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States

    Background: Anxiety is a complex state consisting of cognitive, emotional and physiological arousal components. The bed nucleus of the stria terminalis (BNST, part of the extended amygdala) is a key region implicated in the regulation of anxiety. Due to the large genetic diversity in cell types within BNST, it is unclear whether specific components of anxiety, such as physiological arousal, are driven by precisely defined BNST subpopulations. Neuropeptide systems contribute to define BNST cell-types. BNST is enriched with the neuropeptide nociceptin/orphanin FQ (N/OFQ), and besides its role in pain sensitivity, recent studies suggest its involvement in positive and negative emotional states across many brain regions. Given that painful and emotional stimuli alter arousal, we hypothesized that neurons within BNST that selectively express the genetic precursor to N/OFQ-prepronociceptin (Pnoc-BNST) may contribute to physiological arousal and modulate anxiety-mediated behavior.

    Methods: We characterized the anatomical and genetic identity of Pnoc-BNST neurons using histological (n = 11), electrophysiological (n = 4), and single-cell sequencing (n = 24) approaches. We then determined whether Pnoc-BNST neurons encode increases in physiological arousal, by visualizing the activity dynamics of individual Pnoc-BNST neurons in response to aversive (2,3,5-Trimethyl-3-thiazoline) and rewarding (peanut oil) odors in head-fixed mice. Furthermore, to assess anxiety-mediated behaviors, freely-moving mice were exposed to the elevated plus maze (EPM). We did so by using either head-fixed two-photon (n = 1,180 neurons, n = 7 mice) or freely-moving one-photon (n = 258 neurons, n = 10 mice) calcium imaging approaches. Lastly, we tested whether cell-specific optogenetic photoactivation of Pnoc-BNST neurons drive physiological arousal as measured by pupillary responses (n = 8) and promote anxiety-mediated behavior in the EPM (n = 13).

    Results: Histological, electrophysiological, and single-cell sequencing data revealed that Pnoc-BNST neurons consists of a previously unstudied subset of GABAergic neurons that are interconnected and dispersed across BNST subnuclei. In vivo single-cell calcium imaging revealed that activity of Pnoc-BNST neurons correlates with increased pupillary responses in mice presented with aversive (p < 0.0001) or rewarding (p < 0.0001) odors. Similarly, exposure to open arms of the EPM increased Pnoc-BNST activity (p < 0.001). In addition, in vivo optogenetic photoactivation revealed these neurons increase physiological arousal (p = 0.0398) and anxiety-mediate avoidance (p = 0.0037).

    Conclusions: Pnoc-BNST neurons show high correlations with increases in physiological arousal to salient stimuli and also signal anxiety-mediated states, suggesting a role of these neurons in driving the physiological arousal component of an anxiety state. Our data demonstrate that driving neurons that mediate physiological arousal in combination with exposure to an anxiogenic environment induces a hyperarousal-like state that produces excessive and aberrant anxiety-mediated behavior. Moreover, our data strongly suggest that targeting these neurons may be effective for patients suffering from anxiety disorders.

    Keywords: Anxiety Circuitry, Hyperarousal, In Vivo Calcium Imaging

    Disclosure: Nothing to disclose.

    M11. Corticostriatal Circuit Defects in Hoxb8 Mouse Model of Repetitive Behaviors

    Abstract not included.

    M12. Transcriptional Profiling of Primate Central Nucleus of the Amygdala Neurons to Understand the Molecular Underpinnings of Early Life Anxious Temperament

    Rothem Kovner*, Tade Souaiaia, Andrew Fox, Patrick Roseboom, Delores A. French, Jonathan A. Oler, James A. Knowles, Julie Fudge, Ned Kalin

    University of Wisconsin, Madison, Madison, Wisconsin, United States

    Background: Children exhibiting stable and extreme anxious temperament (AT) are at an increased risk to develop anxiety and depression later in life. Numerous studies point to the importance of the amygdala in mediating adaptive responses to threat as well as stress-related psychopathology. Using a well validated non-human primate model with neuroimaging and lesions strategies, our laboratory identified the central nucleus of the amygdala (Ce) as a key component in the neural circuitry mediating AT. Our previous work assessing punch biopsies of the dorsal amygdala with microarray technology revealed alterations in neuroplasticity genes that were related to AT. To refine our understanding of the molecular alterations related to AT, we now focus on neurons within the lateral Ce (CeL) using laser capture microscopy (LCM) to selectively capture neurons combined with deep RNA sequencing (RNA-Seq).

    Methods: Using LCM, approximately 600 CeL neurons were collected from each of 47 AT phenotyped young male and female rhesus monkeys and RNA-Seq was performed. Multiple regression analysis investigated the relationship between exonic transcripts and individual differences in AT. We focused on transcripts from genes known to be implicated in anxiety and also used a discovery approach to identify alterations in the function of new transcripts. Gene ontology was performed using Panther to determine biological processes related to AT. Transcripts predictive of AT were selected for further anatomical validation using triple-labeling immunofluorescence combined with confocal microscopy and stereological cell counting.

    Results: RNA-Seq analysis demonstrated 600 transcripts to be associated with AT (p < 0.05, two-tailed uncorrected). Permutation testing was done for each of the 600 transcripts in relation to AT and 576 transcripts passed permutation testing (p < 0.05 out of 10,000 tests). Gene ontology analysis demonstrated a variety of biological processes that are associated with AT such as neuron development (GO:0048666, GO:0031175) and differentiation (GO:0030182) as well as enrichment in transcripts related to protein phosphorylation (GO:0006468). More specifically, we identified a number of transcripts related to epigenetic mechanisms (e.g. SS18, SMYD2, and DNMTA3) as well as transcripts predicted to be related to anxiety and psychopathology such as NPY2R and VMAT-2. Of particular interest was the discovery that protein kinase C type-delta (gene name: PRKCD, protein name: PKC-delta) was positively associated with AT (R2, 0.15, t = 3.280, p = 0.017). PKC-delta is an interesting candidate due to its demonstrated role in rodent CeL neurons involved in threat processing and conditioned fear learning. To understand the extent to which these rodent microcircuitry studies are translatable to primates, we characterized the distribution of PKC-delta and other neuropeptide expressing neurons such as somatostatin (SST) in the monkey CeL and compared this to the mouse CeL. The estimated percent of neurons that expressed PKC-delta was 59% in the monkey and 43% in the mouse. Surprisingly, there was a significant difference in the percent of SST expressing neurons with only 7% of neurons expressing SST in the monkey compared to 20% in the mouse (p = 0.02, t = 3.6).

    Conclusions: Here, we demonstrate a method in non-human primates to identify alterations in gene function within neurons of the CeL that are relevant to early life pathological anxiety. The findings implicating PRKCD in CeL neurons are particularly exciting as they build on findings in rodents demonstrating an important functional role for PKC-delta CeL neurons in threat processing. In addition, the data point to other new molecular targets that could be used for circuit-based interventions aimed at modulating CeL microcircuit function. The use of viral vector-based methods to alter gene expression and manipulate circuits in the CeL of primates will allow for further translation of these findings. These studies in nonhuman primates have the potential to inform the development of new treatments for patients suffering from anxiety disorders.

    Keywords: Central Nucleus of the Amygdala, Anxiety Circuitry, Gene Expression

    Disclosure: Nothing to disclose.

    M13. Mitochondrial Pathways Associated With Anxiety-Related Behavior in Mice and Panic Disorder Patients Through a Multi-Omics Analysis

    Zuzanna Misiewicz, Stella Iurato, Natalia Kulesskaya, Laura Salminen, Luis Rodrigues, Giuseppina Maccarrone, Jade Martins, Darina Czamara, Mikaela Laine, Ewa Sokolowska, Kalevi Trontti, Ingrid Balcells, Christiane Rewerts, Suvi Saarnio, Bozo Novak, Dong Ik Park, Angelika Erhardt, Christoph W. Turck, Iiris Hovatta*

    University of Helsinki, Helsinki, Finland

    Background: Anxiety disorders are influenced by genetic and environmental factors and an interaction between them, which is largely undefined. The chronic social defeat stress (CSDS) mouse model allows identification of factors underlying resilience and susceptibility to chronic psychosocial stress-induced anxiety-like behavior, in a controlled manner not possible in human settings. We have taken a cross-species approach to identify biological pathways associated with anxiety through integrative analyses of omics data from the bed nucleus of the stria terminalis (BNST), a part of the extended amygdala involved in stress response, and blood cells from mice and from panic disorder patients who experienced exposure-induced panic attacks.

    Methods: To induce an anxiety-like response in mice, we subjected two inbred strains, innately non-anxious C57BL/6NCrl (B6) and anxious DBA/2NCrl (D2), to CSDS. Following CSDS, we carried out the social avoidance test, to divide the mice into those susceptible and resilient to the behavioral effects of stress. We analyzed the BNST for differences in gene (RNA-seq) and protein (liquid chromatography-tandem mass spectrometry) expression between the stress-resilient, susceptible and control mice. In addition, we carried out RNA-seq from blood cells of stressed mice and performed microarray gene expression profiling of blood samples from panic disorder patients, collected directly and 24 h after exposure-induced panic attacks.

    Results: The two mouse strains had distinct response to CSDS as 87% of D2, but only 30% of B6 mice showed social avoidance behavior. To assess the active and passive stress-coping response following CSDS, we performed the forced swim test and measured the latency to immobility. We observed a high correlation between active coping strategy and resilience to psychosocial stress in the CSDS-subjected D2 mice (r = 0.920, P = 0.009), but not the B6 mice (r = 0.026, P = 0.852). In the BNST, mitochondria-related pathways and gene sets were down-regulated in the D2 (both on mRNA and protein levels) and up-regulated in the B6 strain (on mRNA level). We also found up-regulation of cytochrome c (CYCS), a key molecule of the mitochondrial electron transport chain involved in energy production and initiation of apoptosis, in the B6 resilient and susceptible mice versus controls, which we validated by western blot analysis (P = 0.015 and P = 0.027, respectively). We next conducted gene set enrichment analyses of the mouse and panic disorder patient blood cell transcriptomic data. In the blood cells, mitochondria-related gene sets were significantly up-regulated (FDR < 0.05) in both the stressed B6 mice and panic disorder patients after exposure-induced panic attack. Mitochondria-related gene sets were not differentially expressed in the blood cells of D2 mice. To investigate if CSDS leads to changes in mitochondrial number or morphology, we carried out electron microscopy in the BNST. We observed that susceptible B6 mice had a larger number of pre-synaptic mitochondrial cross-sections than B6 resilient mice (P < 0.05), but that the mitochondria were smaller than those of B6 control mice by 8.4% (P < 0.01). We did not observe differences in the mitochondria size or number between stressed and control D2 mice, but D2 susceptible mice had more total mitochondrial cross-sections in comparison to the B6 susceptible mice (P = 0.013).

    Conclusions: We found a global evolutionarily conserved response of mitochondrial pathways in anxiety-related behaviors. Identification of the underlying mechanisms will provide much needed insight into the molecular basis of anxiety, a critical step in developing targeted therapy.

    Keywords: Anxiety, Panic Attacks, Gene Expression, Proteomics, Mouse Model

    Disclosure: Nothing to disclose.

    M14. In Vivo Evidence for mGluR5 Dysregulation as a Biomarker of Suicidality in PTSD, but Not MDD

    Margaret Davis*, Ansel Hillmer, Sophie Holmes, Nicole DellaGioia, Robert Pietrzak, Nabeel Nabulsi, David Matuskey, Gustavo Angarita-Africano, Richard Carson, Steven Southwick, John Krystal, Irina Esterlis

    Yale University School of Medicine, New Haven, Connecticut, United States

    Background: Posttraumatic stress disorder (PTSD) is associated with high risk for both suicide attempt and mortality. Unfortunately, few available pharmacological interventions have demonstrated efficacy in reducing suicide risk in this population. Exploration of the molecular underpinnings of suicidal behavior in PTSD is essential to targeted treatment development. Recent in vivo, genetic, and postmortem evidence suggest dysregulation of the metabotropic glutamate receptor type 5 (mGluR5) may play a role in the pathophysiology of PTSD and suicidal behavior. Using positron emission tomography (PET) and [18 F]FPEB, we aimed to quantify mGluR5 availability in frontal and limbic areas implicated in the pathophysiology of both PTSD and suicidal ideation.

    Methods: Participants (n = 29 PTSD, 14 with reported scan-day suicidal ideation (SI; based on scan-day self-report); n = 29 with major depressive disorder (MDD), 15 with SI; n = 29 healthy control (HC)) were recruited and matched for age, gender, and smoking status. All participants participated in 1 MRI scan and 1 PET scan with 18F-FPEB and as well as psychiatric and cognitive assessments. Volume of distribution (VT: the ratio of activity in tissue relative to that in blood) was computed using a venous input function. The radiotracer was injected as bolus plus constant infusion and subjects were scanned during steady state (90-120 mins post-injection). Primary analyses focused on a corticolimbic circuit comprised of 5 brain regions: dorsolateral prefrontal cortex (dlPFC); orbitofrontal cortex (OFC); ventromedial prefrontal cortex (vmPFC); amygdala; hippocampus.

    Results: We observed higher mGluR5 availability in PTSD compared to HC individuals in all regions of interest (15%-20% difference, p’s = .001-.01), and compared to MDD individuals in the OFC (15%, p = .0007), dlPFC (17%, p = .007), and hippocampus (15%, p = .007). mGluR5 availability was not significantly different between MDD and HC individuals (p = 0.17). When examining the relationship between suicidality and diagnosis, we observed higher mGluR5 availability in the PTSD-SI group compared to PTSD-non (F7,16 = 4.03, p = .01), but not between the MDD groups. Post hoc analyses indicated PTSD individuals with reported SI had significantly higher mGluR5 availability in all regions of interest (average 24% difference, p’s = .001-.007). Exploratory analyses revealed that mGluR5 availability was positively correlated with avoidance (vmPFC r = .57, p = .007), and measures of tension (vmPFC, r = .53, p = .007) and anxiety (vmPFC r = .47, p = .008) in individuals with PTSD. Interestingly, the opposite was observed among individuals with MDD; mGluR5 availability was inversely correlated with measures of tension (e.g., dlPFC = r = -.52, p = .01) and anxiety (e.g. OFC, r = -.38, p = .01).

    Conclusions: Individuals with PTSD had higher mGluR5 availability in frontal and limbic brain regions relative to both MDD and HC. Further, highest mGluR5 availability was observed in those reporting scan-day SI relative to those without SI in the PTSD group only, implicating mGluR5 as a possible marker of SI specific to individuals with PTSD. The variability in the directionality between mGluR5 and psychological measures further proposes differential role for mGluR5 in PTSD versus MDD. Findings underscore the importance of continued investigation of the glutamatergic system, and mGluR5 specifically, as a target for intervention and risk management in PTSD.

    Keywords: MDD, PTSD, Suicidal Ideation, PET, mGluR5

    Disclosure: Nothing to disclose.

    M15. Applying Computational Modeling for Quantifying Fear Conditioning and Extinction Learning Processes Across Anxiety and Age

    Rany Abend*, Andrea Gold, Jennifer Britton, Bruno Averbeck, Ellen Leibenluft, Daniel Pine

    National Institute of Mental Health, Bethesda, Maryland, United States

    Background: Fear conditioning and extinction are complementary learning processes by which one flexibly acquires, and extinguishes, associations between neutral and harmful stimuli in the environment. Translational research proposes that these processes may play key roles in the etiology and maintenance of anxiety disorders, but several important gaps remain. First, while prior research provides indirect evidence for stronger fear conditioning in anxious individuals, no study formally tests the acquired associative strength between neutral and aversive stimuli as a function of anxiety. Second, some research suggests that fear conditioning and extinction learning may be influenced by age, with developmental differences potentially contributing to the emergence of anxiety disorders. However, extant developmental findings are mixed, limiting insights into the etiology and maintenance of anxiety. Third, findings on the neural correlates of fear conditioning and extinction learning in humans are also limited and mixed, limiting our understanding of underlying mechanisms. Such gaps in knowledge may arise from several reasons, including insufficient sample sizes, variability in study paradigms and parameters, and challenges in quantifying associative learning indices.

    Here, in a novel approach, we leverage computational modeling to directly quantify associative learning strength between neutral and aversive stimuli during a fear conditioning and extinction paradigm, in a large sample (N = 349) of clinically-anxious and healthy youths and adults. Learning indices were then used to investigate three questions: 1) Does associative learning strength during conditioning or extinction differ as a function of anxiety? 2) Does learning strength differ as a function of age? 3) Does learning strength relate to brain morphometry? Insights may link aberrations in fear conditioning and extinction process to anxiety and brain structure across development and may bear implications for treatment of anxiety disorders.

    Methods: Clinically anxious (n = 155) and non-anxious (n = 194) youths and adults (ages 8-50 years; 196 females) completed an established fear conditioning and extinction paradigm. During conditioning, a conditioned stimulus (CS + ) is repeatedly followed by an aversive unconditioned stimulus (UCS). A non-conditioned stimulus (CS-) is never reinforced. Next, during extinction, each CS repeatedly appears, never reinforced. Skin conductance responses (SCR), an index of physiological fear response, were continuously recorded. Reinforcement learning models were fitted to individual trial-by-trial SCR to the CS + during the conditioning and extinction phases. The models yield one associative learning parameter per phase, αCOND and αEXT, indicating how strongly one learns to associate the CS + with the UCS or its absence, respectively. Linear regression models predicted these parameters based on anxiety group and age.

    High-resolution structural brain imaging (T1-weighted) data were available for a subsample of 92 youths. Standard FreeSurfer procedures were used for processing, including cortical thickness estimation and subcortical volume segmentation. We tested for associations between learning indices and cortical thickness and subcortical gray matter volume (GMV) in regions previously implicated in fear conditioning and extinction.

    Results: Across the sample, fear conditioning and extinction learning parameters were significantly greater than 0 (αCOND: M = 0.26, αEXT: M = 0.17, ps < 0.001). A linear regression analysis predicting αCOND indicated a significant effect of anxiety group, α = 0.12, t = 2.23, p = 0.027, with stronger associative learning during conditioning among anxious participants. Additionally, diminished learning during conditioning was noted with age, α = -0.20, t = 3.89, p < 0.001. Anxiety and age did not significantly interact. αEXT was not significantly predicted by anxiety group and age, p = 0.11.

    Associations emerged between αCOND and GMV in right caudate (r = 0.23, p = 0.03), left putamen (r = -0.21, p = 0.05), and right putamen (r = -0.22, p = 0.04), but these did not survive Bonferroni corrections. No other significant associations emerged. Of note, structural data for 149 additional youth and adult participants are currently analyzed and will be ready for the meeting and are expected to increase statistical power to detect effects.

    Conclusions: The current study applied computational modeling to quantify associative learning during fear conditioning and extinction. Anxiety was associated with stronger learning during fear conditioning, whereas age was related to diminished learning. Preliminary imaging data suggest potential associations between conditioning learning strength and subcortical regions previously implicated in learning. These findings reveal novel evidence for associations between fear conditioning processes in anxiety and across age.

    Keywords: Adolescent Anxiety, computational modeling, brain structure

    Disclosure: Nothing to disclose.

    M16. Neural Correlates of Provoked Aggression and PTSD: Preliminary Findings

    Jennifer Fanning*, Lauren Pasetes, Victoria Okuneye, Jamie Wren-Jarvis, Sarah Keedy

    University of Chicago, Chicago, Illinois, United States

    Background: Exposure to psychological trauma contributes significantly to the public health burden of aggressive behavior, and this association is mediated in part by the development of posttraumatic stress disorder (PTSD). A rich literature associates PTSD symptoms with increased anger, hostility, and aggressive behavior. Neurobiological models of PTSD and pathological aggression implicate hyper-activity in limbic regions that mediate threat detection and response (e.g., amygdala) and hypo-activity in regions that support emotion regulation (e.g., orbitofrontal cortex). However, we know little about how patterns of neural activity may account for the relationship between PTSD and aggressive behavior. The purpose of this study is to address this question using fMRI neuroimaging in a sample of adults with and without exposure to trauma who vary in PTSD symptom severity.

    Methods: Adult women and men recruited from the community were studied as part of an ongoing investigation of the effects of interpersonal trauma on brain function and behavior. The overall study uses an RDoC approach to study the neural correlates of interpersonal provocation and retaliation in healthy non-trauma exposed subjects (HC) and trauma-exposed subjects (TE) who vary in PTSD symptom severity, as assessed in preliminary analyses using the PTSD Checklist for DSM-5 (PCL-5). Participants completed a social interaction in task in the scanner against a confederate who was actually fictitious. The task was adapted for the scanner from a well-validated social psychology task with extensive data to support its validity as a measure of provoked aggressive behavior. While in the scanner, the participant and opponent engaged in a competitive cover task, during which the participant was provoked by the opponent with threatened electric shocks to the fingertips. On losing trials the participant received the selected shock. The level of interpersonal threat (i.e., shock intensity) was manipulated on a trial-by-trial basis, and the participant had the opportunity to retaliate on each trial by setting a shock for the opponent, of the intensity of their choice. The task included sensorimotor control trials in which no shocks were delivered. Scans were acquired on a Philips Achieva Quasar 3 T scanner. Structural MRI images were acquired using a T1-weighted 3D Turbo Field Echo (MP-RAGE) anatomical scan. To reduce signal dropout in the orbitofrontal cortex (OFC), a volume-selective z-shim technique was implemented in four slices covering the OFC. Data were pre-processed and analyzed in SPM12 software. First-level statistical models were specified and estimated for the time series per voxel per subject with motion parameters included as regressors. Trial events were modeled using box-car functions convolved with the canonical hemodynamic response function. A first-level general linear model (GLM) with regressors for the opponent’s provocation and the subject’s retaliation was used to model the BOLD signal for the contrast of decision > control. The decision regressor was associated temporally with the subject’s button press response during retaliation. The data were analyzed using a priori region of interest analysis (combined ROI, including prefrontal, limbic, striatal, and parietal regions and motor cortex) and whole-brain analysis. PTSD symptom severity was analyzed as a covariate of the contrast.

    Results: The current sample includes 26 women and 9 men with a mean age of 30 (SD = 7), including 9 HC subjects and 26 TE subjects. The current ratio of women to men participants precludes analysis of sex differences at the present time. Subjects varied in PTSD symptom severity (mean PCL-5 = 16.5, SD = 18.7). Over the course of 48 trials, participants most frequently matched the level of provocation by their opponent (40%) or set a lower intensity (38%). Participants escalated on 22% of trials on average. During the decision phase, whole brain analysis revealed significant activity in the supplementary motor area (SMA) and nucleus accumbens (p-uncorrected < .001, FWE < .05). PTSD symptom severity did not correlate with aggression on the task (r = .18, p = .32) but did correlate with self-reported aggression (Buss Perry physical aggression: r = .40, p = .023; anger: r = .57, p < .001) as expected. PTSD symptom severity was associated with significant activity in the periaqueductal gray (PAG) region (p-uncorrected < .005, FWE < .05).

    Conclusions: Preliminary data from this ongoing study suggests that PTSD symptom severity is associated with heightened activity in the PAG region during retaliation against a provocative confederate. The PAG is implicated in defensive responding to imminent threat and animal models of defensive behavior suggest that subregions of PAG differentially relate to the topography of defensive behavior with respect to freezing, escaping, and attacking. Human neuroimaging studies are suggestive of similar functional specificity within the PAG. The current study points to a potential important role of the PAG in PTSD-related aggressive behavior.

    Keywords: Posttraumatic Stress Disorder, Aggression, Functional MRI (fMRI)

    Disclosure: Nothing to disclose.

    M17. Striatal Activation to Social Reward Anticipation Predicts Symptoms of Social Anxiety Disorder in Adolescent Girls Varying in Level of Risk for Anxiety Disorders

    Cecile Ladouceur*, Neil Jones, Jennifer Silk

    University of Pittsburgh, Pittsburgh, Pennsylvania, United States

    Background: Social anxiety disorder (SAD) is characterized by impairing fear of social situations, such as being negatively evaluated by others (Kessler et al., 2005). SAD typically onsets during early adolescence and there is evidence suggesting that shyness represents a risk factor for SAD, particularly among girls (Tsui et al., 2017). Recent studies have reported striatal hyperactivation to reward anticipation in youth at-risk for (Guyer et al., 2006) and diagnosed with SAD (Guyer et al., 2012). It remains unclear how variability in the functioning of cortico-striatal regions during anticipation of social reward could representation a marker of risk for the development of SAD symptoms. Here, we examined neural activation to social reward cues in girls at risk for anxiety disorders and tested whether such activation levels predicted levels of SAD symptoms at follow-up.

    Methods: Consistent with the NIMH RDoC framework, adolescent girls (n = 59; mean age = 12.6 (SD = .95) were recruited varying in level of risk for anxiety disorders (41 high; 18 low) based on scores being 0.75 SD above or below the mean on the Fearfulness and Shyness subscales of the Early Adolescent Temperament Questionnaire-Revised. Participants completed a modified version of the Social Incentive Delay (SID) task (Spreckelmeyer et al., 2009) to be used with adolescents during fMRI acquisition. This task was adapted from the original MID task to examine neural activation to anticipation and receipt of social feedback from a virtual peer (i.e., virtual peer facial expression) based on the participants’ task performance. An ROI mask was constructed by combining meta-analytic maps for the terms social (z > 3.30, FDR p < .05), reward (z > 8.84, FDR p < .05), and punishment (z > 4.45, FDR p < .05) from Four separate voxelwise analyses were conducted for the: reward anticipation vs. neutral anticipation; punishment anticipation vs. neutral anticipation, reward outcome vs. neutral outcome, punishment outcome vs. neutral outcome contrasts masking activation using the ROI mask defined above. Analyses were conducted using AFNI’s 3dttest with the -Clustsim option, which uses a nonparametic approach to cluster-size thresholding with a cluster forming threshold of p < .001, extent threshold ~ 23 voxels. Linear regression analyses focused on neural activation to social reward anticipation and SAD symptoms, measured using the SAD subscale of the SCARED at 12 and 18 months following the scan and controlling for age and baseline SAD symptoms. Bonferonni correction was used to correct for type-I error as needed.

    Results: Analyses focusing on task-related activation indicated greater activation in the caudate and medial frontal gyrus (MFG) to reward (vs neutral) anticipation; no significant clusters were found for punishment anticipation. There was greater activation in a number of subcortical (amygdala, striatum) and prefrontal cortical (e.g., VMPFC, VLPFC) regions as well as the superior temporal gyrus (STG) to reward (vs. neutral) outcome. Similar regions were activated to punishment (vs. neutral) outcome in addition to activation in the posterior cingulate gyrus (PCG). Regression analyses revealed that greater activation in the caudate to social reward anticipation predicted higher levels of SAD symptoms at 18 months, (R2 = .48, Beta = .63, t = 3.18, p = .008).

    Conclusions: Findings suggest that the adapted version of the SID task recruited expected regions implicated in processing social (e.g., STG, PCG), reward (e.g., caudate, putamen, VMPFC), and punishment (e.g., amygdala) information. Our findings of greater caudate activation to anticipation of social reward predicting later symptoms of SAD in these at-risk girls is consistent with previous findings in youth with early-life behavioral inhibition (Guyer et al., 2014). They also support a neurodevelopmental model of SAD suggesting that aberrant information processing of social reward cues may escalate the risk for SAD at a time when increased sensitivity to rewarding experiences such as peer acceptance typically occurs (Caouette and Guyer, 2014). Future analyses, including a follow-up scan with a larger sample, will deepen our understanding of neurodevelopmental pathways of SAD in at-risk girls.

    Keywords: Risk for Social Anxiety Disorder, Early Adolescence, Social Reward, Functional Neuroimaging

    Disclosure: Nothing to disclose.

    M18. Dissecting the Genetically Regulated Transcriptomic, and Phenotypic Complexity of PTSD Across 9400 Trauma-Exposed Individuals and 30 Million Observations With Bayesian Machine Learning

    Yue Li, Vasiliki Michopoulos, Adriana Lori, Abigail Lott, Bekh Bradley, Tanja Jovanovic, Kerry Ressler, Manolis Kellis, Nikolaos Daskalakis*

    McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

    Background: Post traumatic stress disorder (PTSD) is a severe mental disorder occurring in some trauma-exposed individuals, but its molecular basis remains uncharacterized. Transcriptomic imputation models (e.g., PrediXcan) leverage large expression quantitative trait loci reference panels to translate genome-wide genotype data into more biologically meaningful measures. Previously, we have associated genetically regulated gene expression in the brain and peripheral tissues with PTSD which were validated by translational findings in mice and humans (Huckins…Daskalakis, In Preparation Manuscript by Psychiatric Genomic Consortium for PTSD). Here, we study the relationship between genetically-regulated transcriptional variation, and phenotypic variation in PTSD, in a cohort of 9400 inner-city, low-socioeconomic-status, primarily-African-American patients of the Grady Memorial Hospital.

    Methods: Each individual was ascertained phenotypically using interview-based assessments and self-report questionnaires including childhood trauma history and PTSD diagnosis, and 12 lab tests including endocrine, metabolic and immune markers (e.g., cortisol, cholesterol, IL-6). In addition, we imputed expression for ~5000 genes across 48 tissues using Genotype-Tissue Expression (GTEx) project’s expression quantitative trait loci (v7 data release) and selecting the top 200 most variable and tissue-specific genes.

    To study the comorbidity patterns of tissue-specific gene expression and phenotypic information across individuals, we used MixEHR, a Bayesian unsupervised learning method for phenotype correlation and imputation analysis (Li & Kellis, 2018), which explicitly deals with the non-missing-at-random nature of phenotypic information and lab tests. We applied MixEHR systematically and found several noteworthy disease topics of high probability of PTSD diagnosis that were concordant with PTSD symptom severity, depression symptom severity and childhood trauma levels, thereby recapitulating the PTSD clinical presentation.

    Results: We found enrichments for well-studied genes and tissues in specific disease topics across multiple tissues. For example, a human-lineage-specific gene, NBPF3, which is implicated in neurodevelopmental disorders, showed high probability for PTSD across multiple brain tissues. Similarly, the top tissues with the highest marginal PTSD associations are basal ganglia, amygdala, substantia nigra, and pituitary, which have been associated with emotion regulation and PTSD. We also constructed a ranked list of PTSD genes by calculating the marginal probabilities of each gene with respect to PTSD across multiple disease topics. The top genes were: MXRA8, the top gene, is important in the maturation and maintenance of blood-brain barrier; ATP6AP1L, ranked #2, is implicated in the glucocorticoid receptor pathway and neural responses to stress; and AP3S2, ranked #3, is essential for vesicles delivery into neurites and nerve terminals.

    Conclusions: By combining transcription imputation across the brain and the rest of the body with Bayesian machine learning in a trauma-exposed sample, we discovered genes that yield new biological insights into the genetic and phenotypic architecture of PTSD.

    Keywords: Bayesian Modeling, PTSD, Childhood trauma, transcription imputation, Genetic variability

    Disclosure: Nothing to disclose.

    M19. Amygdala Subregions Volumes are Associated With PTSD

    Rajendra Morey*, Emily Clarke, Rachel Phillips, Courtney Haswell, Kevin LaBar

    Duke University, Durham, North Carolina, United States

    Background: The amygdala is perhaps the most strongly implicated brain structure in the pathophysiology of posttraumatic stress disorder (PTSD). The amygdala is a subcortical structure involved in emotion processing and associative fear learning. Prevalent models of anxiety have focused on an amygdalocentric neurocircuitry that is critical in fear response, conditioning, generalization, and facilitates the response to stressful experiences. Functional magnetic resonance imaging (MRI) studies have shown that individuals with PTSD have an exaggerated amygdala response to emotional stimuli when compared with control subjects. Animal studies have demonstrated changes in amygdala morphology with chronic stress, evident primarily in the growth of dendritic spines. Experimental studies of amygdala volume in mice and humans have shown an association among smaller amygdala volumes, increased levels of fear conditioning, and an exaggerated glucocorticoid response to stress. However, efforts to find evidence of an association between amygdala volume and PTSD in humans have produced equivocal results. These studies point to specialized roles of basolateral (BLA) and centromedial amygdala (CMA) complexes in associative fear learning. However, this level of investigation has been largely absent in human studies, despite the critical role of individual amygdala complexes predicted in PTSD.

    Methods: The study assessed US military veterans (N = 346) with lifetime PTSD (N = 141) and trauma-exposed non-PTSD controls (N = 206), who served since 9/11 (2001). Participants completed clinical and neuroimaging assessments that included the Structured Clinical Interview for DSM-IV Disorders (PTSD), Alcohol Use Disorders Identification Test, Combat Exposure Scale, Trauma Life Events Questionnaire to assess childhood trauma exposure, Beck Depression Inventory, and Drug Abuse Screening Test. High resolution T1-weighted anatomical scans optimized for tissue contrast were acquired on a 3 T GE MR750 scanner. Automated segmentation of amygdala subregions was completed using FreeSurfer (version 6.0) on nine nuclei or subregions per hemisphere including basal, lateral, accessory basal, anterior amygdaloid, central, medial, cortical, corticoamygdaloid transition area, and paralaminar. The FreeSurfer 6.0 tool is based on a probabilistic atlas of hippocampal anatomy created upon manually segmented ultra-high-resolution (0.13-mm isotropic) ex vivo MRI scans from 15 post-mortem samples and manual segmentations of nearby structures performed on 1-mm resolution isotropic in vivo T1 images. The subregion volumes obtained from FreeSurfer were the dependent variable in an ordinary least square (OLS) regression model run separately for each subfield. Bonferroni correction for multiple testing was applied to the volumetric analyses given that nine regions from two hemispheres were assessed (p-corrected = 0.05/18 = 0.003). Covariates for age, gender, ipsilateral whole amygdala volume, alcohol use, depressive symptom score, childhood trauma exposure, combat exposure, and use of psychotropic medications were modeled.

    Results: We found that lifetime PTSD was associated with larger volume of left medial nucleus (p = 0.005; Cohen’s d = 0.35; Bonferroni corrected), left cortical nucleus (p = 0.007; Cohen’s d = 0.26; Bonferroni corrected), right accessory basal nucleus (p = 0.01; Cohen’s d = 0.07; Bonferroni corrected), and right central nucleus (p = 0.02; Cohen’s d = 0.18; Bonferroni corrected). Right accessory basal nucleus volume was affected by alcohol use (p = 0.04) and marginally affected by age (p = 0.06) and childhood trauma (p = 0.06). Right central nucleus volume was affected by gender (p = 0.02). Left medial nucleus volume was marginally affected by alcohol use (p = 0.09).

    Conclusions: The present results highlight that PTSD is associated with larger volumes in amygdala subregions strongly implicated in fear processing that include the medial nucleus, central nucleus, and the accessory basal nucleus. Our results are consistent with rodent models that demonstrate chronic threat and stress lead to corticosterone-mediated spinogenesis and dendritic arborization.

    Keywords: Combat PTSD, Amygdala, Central Nucleus of the Amygdala, Basolateral Amygdala, Structural MRI

    Disclosure: Nothing to disclose.

    M20. Brain Activation During Interference Inhibition and Error Processing Predicts Symptom Improvement Following Cognitive Behavioral Therapy in Obsessive Compulsive Disorder

    Kate Fitzgerald*, Luke Norman, Mike Angstadt, Kristin Manella, James Abelson, Joseph Himle, Stephan Taylor

    University of Michigan, Ann Arbor, Michigan, United States

    Background: Cognitive behavioral therapy with exposure and response prevention (CBT-ERP) is the gold standard psychotherapeutic treatment method for obsessive compulsive disorder (OCD). However, the extent to which patients experience symptom reduction from CBT-ERP varies, and up to half of those treated do not fully respond. Moreover, the mechanisms that underlie differences in symptom reduction following treatment are poorly understood. Given that effective engagement with CBT-ERP is assumed to rely on executive functions, it might be hypothesized that individual differences in brain activation during executive function tasks would be associated with treatment response to CBT-ERP. In OCD, recent meta-analytic evidence points to abnormal engagement of cingulo-opercular network in patients relative to unaffected controls, with underactivation during interference inhibition and hyperactivation during error processing, but whether these neural abnormalities predict response to CBT-ERP is unknown.

    Methods: 65 patients with OCD (age range 12-45, 40 males, 31 medicated) were randomly allocated to receive either 12 weeks of CBT-ERP or an active control treatment (stress management training; SMT). Prior to treatment, patients completed a functional magnetic resonance imaging (fMRI) version of the incentive flanker task (IFT), which probes neural activation to interference inhibition and error processing. Symptoms were assessed at baseline and post-treatment using the (Children’s) Yale-Brown Obsessive-Compulsive Scale ((C)Y-BOCS). Whole-brain regression analyses examined whether brain activation during interference inhibition and error processing was associated with symptom change following treatment in CBT-ERP and SMT groups (height threshold p < .001 uncorrected, cluster threshold p < .05 FWE corrected).

    Results: As predicted, the CBT-ERP group (mean (C)Y-BOCS change = -12.31, SD = 6.61, range -24.5 to 1.5) showed greater symptom reduction following treatment compared with the SMT group (mean (C)Y-BOCS change = -4.49, SD = 7.98, range -26 to 11.5; F(1,163) = 18.18, p < .001). Nonetheless, there was a wide range of response in both CBT-ERP and SMT groups, and even some patients receiving SMT exhibited reduction in OCD symptoms. In the CBT-ERP group, symptom reduction was predicted by greater baseline activation of left inferior frontal gyrus (IFG), right IFG/anterior insula and dorsomedial prefrontal cortex during interference inhibition, but lesser left anterior insula activation during error-processing (all p < .05, corrected). In the SMT group, there were no areas of activation that significantly predicted treatment response.

    Conclusions: Patients with OCD typically show reduced activation of cingulo-opercular network regions during interference inhibition relative to unaffected controls, as well as increased brain activation in the same network following erroneous responses. We have previously argued that this pattern of neural abnormalities explains why patients become stuck in ‘compulsive loops’, as detected erroneous OCD behaviors remain uncorrected by an underactivated inhibitory control network. The current findings suggest that intact cingulo-opercular functioning may allow patients to engage more effectively with CBT-ERP treatment, perhaps due to a greater capacity for inhibiting maladaptive responses to symptom triggers and for learning and implementing newer, more adaptive behaviors and coping strategies. The presence of significant predictors of symptom reduction for CBT, but not SMT, suggests that cingulo-opercular circuit function may have mechanistic significance specific to CBT-ERP. Additional work examining subjects before and after treatment will be needed to determine whether modulation of cingulo-opercular network functioning also mediates OCD symptom change with CBT, and whether it could serve as a target for novel interventions designed to augment CBT effects.

    Keywords: Obsessive-Compulsive Dlerisorder (OCD), Cingulo-Opercular Network, Cognitive Behavioral Therapy, Predictors of Response

    Disclosure: Nothing to disclose.

    M21. Regionally Specific Alterations and Clinical Correlates of Cerebral GABA in Posttraumatic Stress Spectrum Adults

    Isabelle Rosso*, Elizabeth Olson, J. Eric Jensen, Scott Rauch

    McLean Hospital, Belmont, Massachusetts, United States

    Background: Posttraumatic stress disorder (PTSD) has been associated with hyper-responsivity of the insular cortex and hypo-responsivity of the medial prefrontal cortex (mPFC) during symptom provocation and exposure to fear-inducing stimuli, consistent with altered excitability of brain regions mediating fear conditioning and extinction. As the brain’s principal inhibitory neurotransmitter, γ-aminobutyric acid (GABA) exerts a prominent role in modulating neuronal excitability. Although brain GABA can be reliably assessed with advances in magnetic resonance spectroscopy (MRS), there have been few MRS GABA studies in PTSD. In a pilot study from our laboratory, we previously reported lower right anterior insula (AI) GABA in adults with PTSD compared with healthy comparison (HC) participants (Rosso et al 2013). In the current study, we compared GABA levels in two brain areas implicated in PTSD and fear (mPFC, right AI) and in one control brain area (right posterior temporal cortex) between adults with DSM-IV PTSD, trauma-exposed non-PTSD comparison (TENC) subjects, and HC subjects without any reported Criterion A trauma. We hypothesized that there would be group differences in mPFC and AI GABA but not in temporal cortex GABA. We also predicted that GABA alterations would be associated with dimensional measures of PTSD symptom severity and trauma exposure.

    Methods: All participants (ages 20-50) underwent a psychiatric interview with the Structured Clinical Interview for DSM-IV (SCID-IV I/P), and PTSD and TENC participants also completed the Clinician Administered PTSD Scale (CAPS). The CAPS was used to establish PTSD diagnosis and dimensional symptom cluster scores (re-experiencing, hyperarousal, and avoidance symptoms). The Dissociative Experiences Scale (DES), completed by all participants, provided a dimensional measure of dissociation. Trauma exposure also was assessed quantitatively in all participants using the Life Events Checklist (LEC) which indexes lifetime exposure (number of traumas directly experienced and witnessed). Participants underwent MRS scanning on a 3 T Siemens Trio with a 32-channel head coil. MEGAPRESS data were collected from voxels in the mPFC (2.5 × 2.5 × 3 mL; 39 PTSD, 32 TENC, 37 HC with good quality spectra), right AI (2 × 2 × 3 mL; 38 PTSD, 31 TENC, 38 HC), and temporal cortex (2 × 2 × 3 mL; 39 PTSD, 31 TENC, 36 HC). GABA was normalized to total creatine (Cr).

    Results: The three groups were comparable on age and sex distributions, and PTSD and TENC participants did not differ significantly on number of lifetime traumas (LEC score). Compared with HC, PTSD and TENC participants had significantly lower GABA/Cr in the mPFC voxel (F (2,104) = 5.56, p = .005) and in the AI voxel (F (2,104) = 4.01, p = .02), but not in the temporal cortex voxel (F (2,103) = 0.25, p = 0.78). Neither mPFC GABA/Cr nor AI GABA/Cr was associated with CAPS re-experiencing, hyperarousal, or avoidance scores. However, there was a significant association of DES scores with mPFC GABA/Cr, both in the combined PTSD + TENC group (r (70) = -0.24, p = .04) and in the overall sample (r (107) = -0.24, p = .02). LEC scores were significantly negatively correlated with mPFC GABA/Cr (r (107) = -0.21, p = .03) and AI GABA/Cr (r (107) = -0.29, p = .003). Finally, in a multiple regression entering both DES and LEC scores as predictors of mPFC GABA/Cr, DES score remained a statistically significant effect (t = -1.99, p = .05) but not LEC score (t = -1.36, p = 0.18).

    Conclusions: We found regionally specific lower GABA in the mPFC and right AI but not right posterior temporal cortex, in patients with PTSD and trauma-exposed adults without PTSD, compared with non-traumatized healthy adults. These findings suggest that GABA is lower in brain areas implicated in fear and arousal regulation, and that this is related to aspects of trauma-related psychopathology but not a biomarker of diagnosis per se. We replicate our initial report that right AI GABA is lower in PTSD and extend that finding by showing that lower GABA also is seen in adults with histories of significant trauma who never developed PTSD. This relationship with trauma exposure is further corroborated by the negative correlation between right AI GABA and number of traumatic life events. Although mPFC GABA reductions also scaled with lifetime trauma, they were more strongly associated with a dimensional measure of dissociation. This finding may be consistent with growing evidence that dissociative states have distinct neural correlates within PTSD symptomatology, including differentially high excitability of the medial prefrontal cortex.

    Keywords: Post Traumatic Stress Disorder, MR spectroscopy, Dissociation, insular cortex, Medial Prefrontal Cortex

    Disclosure: Aptinyx Inc, Consultant

    M22. Thyroid Hormone Modulation of Anxiety and Trauma-Related Plasticity in the Amygdala

    Stephanie A. Maddox*, Michelle X. Chen, Brianpaul J. Robert, Kerry J. Ressler

    McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

    Background: The thyroid hormone (TH) system has been associated with anxiety, depression and well implicated in early nervous system development. Despite this early progress, almost nothing is known about the potential role that the TH system may play in mediating synaptic plasticity that is relevant for anxiety and trauma-related behaviors.

    Methods: Non-biased RNA-sequencing and targeted qPCR was used to examine the regulation of genes in the amygdala in fear conditioned mice. Local amygdala infusions of the TH triiodothyronine (T3) or the dual TH receptor (TR) TR-alpha (TRa) and TR-beta (TRb) antagonist 1-850 were used to examine the role of TRs in fear memory consolidation and associated gene transcription. Additional experiments examined the effect of these pharmacological agents on anxiety-like behaviors using the open field test. Additional studies employing a diet-induced model of hypothyroidism were conducted to examine the behavioral and molecular consequences that occur as a result of diminished TH function. Hypothyroid animals were administered local amygdala infusions of T3 hormone to examine the role of amygdala TRs on traumatic memory formation.

    Results: Using nonbiased RNA-sequencing from amygdala punches taken 2 h after fear conditioning, we revealed dynamic regulation of seven TH related genes in the amygdala in response to trauma, Ttr (lg2fc = -2.52, q < 0.05), Trhr (lg2fc = -0.44, q,0.05), Trh (lg2fc = -1.21, q < 0.05), Trip11 (lg2fc = 1.05, q < 0.05), Rxrg (lg2fc = 0.31, q < 0.05), Dio2 (lg2fc = 0.41, q < 0.05), and Med12 (lg2fc = -0.28, q < 0.05. Additionally, using targeted qPCR, we demonstrated a significant fear conditioning-related increase in TRa (T(17) = -2.48, p < 0.05) and TRb (T(17) = -1.47, p = 0.05) mRNA in the amygdala compared to control animals and TRa and TRb protein expression levels in the amygdala using immunohistochemistry. Next, using directed intra-amygdala infusions of either T3 or 1-850, we examined the effect of TR activation and inhibition at the time of fear conditioning on fear memory consolidation 24 h later. We found that mice receiving infusions of T3 30 min prior to conditioning had enhanced levels of freezing during the conditioning session (T(21) = 4.101, p < 0.05) and enhanced levels of freezing 24 h later during the long-term memory test (T(21) = 1.65, p = 0.05) compared to vehicle controls. Conversely, we found that infusions of 1-850 resulted in no difference in freezing level during fear conditioning (T(22) = -0.28, p > 0.05) but dramatically reduced long-term memory 24 h later (T(22) = 2.96, p < 0.05) suggesting impaired memory consolidation. We also examined the effect of intra-amygdala T3 on anxiety behavior using the open field test and T3 on fear memory with infusions occurring immediately after conditioning. Infusions of T3 resulted in increased time spent in the periphery of the open field arena (T(13) = -2.11, p = 0.027). Infusions immediately after fear conditioning failed to alter fear memory expression 24 h later (T(13) = 0.41, p > 0.05). Interestingly, hypothyroid animals also displayed heightened anxiety, as measured using the open field test (T(22) = -4.34, p < 0.01), and impaired traumatic memory consolidation (T(22) = 2.82 p < 0.01). Western blots examining protein level molecular differences with hypothyroidism revealed reduced expression of EGR1, Arc, and GAD67 (all p < 0.05). Examination of transcriptional alterations in the amygdala that accompany hypothyroidism revealed increased Reln, Trh, and Nr3c1 mRNA (all p < 0.05) and decrease Dio3 mRNA (p < 0.05). Infusions of T3 into the amygdala of hypothyroid animals was found to reverse the traumatic memory consolidation impairment (T(10) = 2.04, p = 0.03).

    Conclusions: These data suggest that T3 is anxiogenic and that activation of TRs at the time of trauma exposure or fear conditioning is rapid and must occur at the initiation of trauma to impact subsequent fear expression. Additional experiments have revealed that infusions of 1-850 at the time of fear conditioning alter the expression of many fear and memory related genes suggesting that TRs may mediate trauma-related gene expression in the amygdala. Examination of molecular and behavioral alterations that accompany hypothyroidism revealed alterations in classic plasticity related genes and proteins in the amygdala, as well as increased anxiety-like behaviors. While the mechanisms for heightened anxiety and yet impaired trauma-relevant memory consolidation in hypothyroid animals remain under active investigation, our rescue experiment which replaced T3 directly into the amygdala suggests that THs play a critical role in mediating trauma-relevant memory formation. In sum, these data strongly implicate for the first time a dynamic role for the thyroid hormone axis in the amygdala in modulating anxiety and trauma-relevant neuroplasticity.

    Keywords: Thyroid Hormone, Traumatic Memories, Anxiety

    Disclosure: Nothing to disclose.

    M23. Exploring Floatation-REST as a Novel Intervention for Anxiety Sensitivity

    Justin Feinstein*, Sahib Khalsa, Hung-Wen Yeh, Obada Al Zoubi, Murray Stein, Martin Paulus

    Laureate Institute for Brain Research, Tulsa, Oklahoma, United States

    Background: Floatation-REST (Reduced Environmental Stimulation Therapy), an intervention which attenuates exteroceptive sensory input to the nervous system, has recently been found to reduce state anxiety across a diverse clinical sample with high levels of anxiety sensitivity (Feinstein et al., 2018). To further examine this anxiolytic effect, the present study investigated the physiological changes induced by Floatation-REST and assessed whether individuals with high anxiety sensitivity experienced any alterations in their awareness for interoceptive sensation while immersed in an environment lacking exteroceptive sensation.

    Methods: Thirty-seven participants (23 females; 14 males) with high anxiety sensitivity (Anxiety Sensitivity Index-3 total score ≥ 30) were recruited across a spectrum of anxiety and stress-related disorders (posttraumatic stress, generalized anxiety, panic, agoraphobia, and social anxiety), most (n = 35) with comorbid unipolar depression. Each participant was randomly assigned to undergo a 90-minute session of Floatation-REST or an exteroceptive comparison condition, and then crossed over to the other condition following a 1-week washout period. Measures of self-reported affect and interoceptive awareness were collected before and after each session, and electrocardiogram and blood pressure were collected during each session using wireless and waterproof sensors. All measures were analyzed by linear mixed-effects models using fixed-effects of Time and Session, and a subject-specific random intercept.

    Results: Relative to the comparison condition, Floatation-REST significantly enhanced awareness and attention for cardiorespiratory sensations (p < .0001). Physiological measures collected during the float session showed indications of a rapid relaxation response, including a significant reduction in diastolic blood pressure (on the order of 10 mm Hg) and a significant increase in normalized high-frequency heart rate variability (p < .0001 for both measures). The degree to which floating increased heart rate variability was associated with the magnitude of the anxiolytic effect.

    Conclusions: Floatation-REST induced a strong relaxation response consistent with increased parasympathetic outflow. At the same time, the float environment enhanced awareness for cardiorespiratory sensations in a clinical sample with high anxiety sensitivity. The unique juxtaposition between heightened interoceptive awareness in the context of a relaxed physiological background state may confer unique opportunities for the treatment of anxiety sensitivity.

    Keywords: Anxiety, Interoception, Novel Therapeutics, Mindfulness

    Disclosure: Nothing to disclose.

    M24. Computational Modeling of Biases Influencing Threat Detection

    Tamara Sussman*, Jingwen Jin, Akos Szekely, Christian Luhmann, Aprajita Mohanty

    Columbia University Medical Center, New York, New York, United States

    Background: We frequently use prior knowledge in a “top-down” manner to anticipate and detect threats in our environment in day-to-day life. However, most research has attributed fast and accurate detection of threatening stimuli to automatic or “bottom-up” processing. This relationship between anticipation and threat detection is particularly important in anxiety disorders, which are characterized by inaccurate, negative predictions about the future. In two experiments, we examined whether prior knowledge regarding threat and the likelihood of its occurrence facilitates perceptual decision-making. We examined the role of bias in facilitating threat-related perceptual decision-making by modeling this process using the drift–diffusion model (DDM), in which bias is computed as a shift in starting point or rate of evidence accumulation.

    Methods: In experiment 1, participants (n = 31; 21 women; mean age 20.26 ± 1.90 years) viewed fear cues (the letter F), or neutral cues (the letter N) followed perceptually degraded fearful face (FF) or neutral face (NF) images, presented for 100 ms at each participant’s individually determined perceptual threshold for FF and NF. The fear cue indicated that participants would make a “fearful or not” decision whereas the neutral cue indicated a “neutral or not” decision regarding the subsequently presented faces. We will refer to this cue-related information as “prior salience.” Based on their response, we determined whether participants perceived the face as fearful (FR) or neutral (NR).

    In experiment 2, a separate sample of participants (n = 29; 22 women; mean age 20.41 ± 2.11 years) performed a modified version of the same task such that cues additionally provided information regarding the probability of the targets to be presented (referred to as “prior probability”; 25%, 50%, or 75%) in addition to information regarding the salience of the subsequent face.

    To examine the impact of prior salience (experiment 1) or prior salience and prior probability (experiment 2) on decision parameters, the DDM was fitted to the RT and response data. In the DDM, stochastically sampled evidence is accumulated in a single decision-variable from a starting point between the two decision boundaries. This starting point forms the baseline from which evidence accumulation starts. Evidence accumulation begins upon stimulus presentation and the rate at which evidence is accumulated towards a decision is called the drift rate. The drift rate reflects the effectiveness of the signal extraction from the surrounding noise.

    Results: Results from the both experiments demonstrate that threat cues improve perceptual sensitivity (experiment 1: t(30) = 4.36, p < .001; experiment 2: F(1, 28) = 18.24, p < .001) and speed of decision-making (experiment 1: F (1, 30) = 11.92, p = .002; experiment 2: F(1, 28) = 38.15, p < .001). Additionally, in experiment 2, prior probability also increased the sensitivity (F(2, 56) = 5.26, p = .008) and speed (F(2, 56) = 3.46, p = .038) of decisions, with higher probability cues leading to faster, more accurate choices.

    Results from DDM model comparison indicated that the best-fitting model for experiment 1 allowed starting point to vary by prior salience, and drift rate to vary by both prior salience and the type of stimulus. The best fitting model for experiment 2 was the same, except that it additionally allowed starting point to vary by prior probability.

    DDM analyses of experiment 1 data revealed that the starting point was shifted towards the fear decision boundary (posterior probability that zFC > .5 = .993), for trails starting with FC. NC did not shift the starting point towards an NR decision boundary, the posterior probability that zNC < .5 = .305. In both experiments, compared to NC, FC also resulted in faster evidence accumulation for both FF, with posterior probability of vFC/FF > vNC/FF = .967 (experiment 1), vFC/FF > vNC/FF = .986 (experiment 2) and for NF, with a posterior probability of vFC/NF > vNC/NF = .999 (experiments 1 & 2).

    In experiment 2, for FC trials, change in prior probability shifted the starting point closer to FR decision boundary, with the posterior probability that zFC75% > zFC50% = .802 and that zFC50% > zFC25% = .958, with the absolute estimates of all three shifted closer to FR decision boundary (the probability of z > .5 was .996 for FC25%, .999 for FC50%, and .999 for FC75%). In contrast, increases in prior probability of NC did not shift the starting point.

    Conclusions: These findings establish the importance of top-down factors in understanding how the perceptual system prioritizes threatening stimuli. Results indicate that threat cues enhanced the sensitivity and speed of perceptual decision-making. DDM modeling results indicate this threat-related enhancement occurs via 1) threat- and probability-related shifting of the starting point of the evidence accumulation process closer to the threat decision boundary and 2) faster evidence accumulation or drift rate after encounter with the stimulus following fear cues.

    This study quantifies bias and clarifies overlapping and dissociable biasing mechanisms by which prior knowledge impacts perceptual decision-making. Elucidation of these mechanisms particularly import in anxiety, which is distinguished from fear due to its anticipatory nature. Extending the present research could clarify how anticipatory factors and biased threat-perception contribute to the development and maintenance of anxiety disorders.

    Keywords: Computational Modeling, Threat Faces, Top-Down Control, Visual Perception

    Disclosure: Nothing to disclose.

    M25. Early Life Adverse Experiences and Obsessive-Compulsive Disorder: A Study With Patients, Siblings and Controls

    Fabiana Costa, Carolina Cappi, Marcelo Batistuzzo, Roseli Shavitt, Guaraci Requena, Euripedes Miguel, Marcelo Hoexter*

    University of Sao Paulo, Sao Paulo, Brazil

    Background: Early life adverse experiences are risk factors for obsessive-compulsive disorder (OCD), but studies investigating the relationship between early life adverse experiences, obsessive-compulsive disorder and psychiatric comorbidities are scarce. Moreover, the whole of experiential avoidance, resilience, and dissociation experiences in mediating these relationships is less clear. The present study examines the relationship between early life adverse experiences, experiential avoidance, resilience, dissociation and general psychopathology in a sample of patients with OCD, siblings and controls.

    Methods: The Maltreatment and Abuse Chronology of Exposure (MACE) scale was used to assess early-life adversity in 24 participants with OCD, 24 siblings and 24 matched healthy controls. Experiential avoidance was measured with the Multidimensional Experiential Avoidance Questionnaire (MEAQ) and with the Acceptance and Action Questionnaire (AAQ II); resilience was measured with the Resilience Scale for Adults (RSA); and the Dissociative Experiences Scale (DES) assessed dissociation. The Kruskal–Wallis nonparametric analysis of variance was used to test between-group differences. Then, post hoc pair wise comparisons (Wilcoxon test) were performed considering a p < 0.017.

    Results: MACE emotional abuse scores were different between groups. Post hoc pair wise comparisons showed that patients with OCD presented higher emotional abuse scores compared to healthy controls (p < 0.017). There were no differences between patients and siblings and between siblings and controls in terms of MACE scores. Patients presented higher scores in experiential avoidance compared to controls as measures by the MEAQ (p < 0.0001), but there were no difference between patients and siblings and between siblings and controls. In terms of AAQ II, avoidance was higher in OCD patients compared both to siblings (p = 0.002) as well as to controls (p < 0.0001). Siblings and controls did not differ. Resilience was lower in patients with OCD compared both to siblings (p < 0.003) and to controls (p < 0.014), though there were no differences between siblings and controls. Dissociation scores were not different between-groups.

    Conclusions: Results of this study suggests that emotional abuse during childhood and adolescence may play a role as a risk factor for OCD. On the other hand, experimental avoidance and resilience should be considered as potential factors mediating the relationship between early life adverse experiences and OCD.

    Keywords: Adverse Childhood Experiences (ACE), Obsessive-Compulsive Disorder (OCD), Resilience, Avoidance, Dissociation

    Disclosure: Nothing to disclose.

    M26. The Lipidome in PTSD

    Sabra Inslicht*, Ritu Roy, Adam Olshen, Callan Lujan, Aditi Bhargava, Thomas Neylan

    University of California, San Francisco, San Francisco, California, United States

    Background: Post-traumatic stress disorder (PTSD) affects approximately 8% of the US population and 22% of US veterans returning from Iraq and Afghanistan. Not only does PTSD impact psychological well-being, it also has been associated with physical health concerns, including greater inflammation, metabolic syndrome, gastrointestinal (GI) illness, and even early mortality. Impaired glucocorticoid regulation associated with PTSD may increase risk for chronic health conditions, since glucocorticoids regulate metabolism of proteins, carbohydrates, and lipids including processes such as gluconeogenesis (protein to carbohydrate conversion) and redistribution of fat. In addition, specific fatty acid/amino acid metabolites can render PTSD patients more susceptible to neuronal excitability, signal transduction, and inflammation. Elevated concentrations of serum lipids were associated with combat-related PTSD. Veterans with combat-related PTSD are at a higher risk for arteriosclerosis, thus elevated or dysregulated serum lipids may contribute to increased immune and cardiovascular responses in PTSD. The aims of this study were two-fold. First, to ascertain if plasma lipid metabolites were associated with PTSD relative to healthy controls. Second, to examine if the plasma lipid profile was associated with sex. We predicted that both PTSD and male sex would be associated with higher levels of peripheral lipid metabolites that are known to traverse the blood-brain barrier and negatively affect health.

    Methods: A metabolomics analysis has the potential to explain the regulation of metabolic pathways and networks of physiologically relevant interactions that lead to increased health risks in PTSD. As a global and unbiased approach, metabolomics identifies changes in circulating small molecules that affect cell and physiological function and can provide a more comprehensive examination of the broad range of physiological pathways that may be missed with more traditional, targeted approaches. Metabolomic analyses were performed on plasma samples obtained from fasting male and pre-menopausal follicular phase female subjects with chronic PTSD (N = 44) and trauma-exposed, age-matched controls (N = 44). Participants were between the ages of 20 and 50, primarily civilians (89%), healthy, free of medications, alcohol and drugs, and limited to one cup of caffeine daily. Plasma samples were assayed for lipids, fatty acids, sphingolipids, and short-chain fatty acid metabolites using Time-of-Flight Mass Spectrometer (Agilent Technologies 6220 TOF) coupled with an Ultra HPLC. Statistical analyses were performed using the R/Bioconductor statistical framework. Metabolomic data were transformed to log base 2 scale. Undetected metabolomic values were replaced by 1/10 of minimum value where necessary. Wilcoxon’s rank-sum tests were performed for univariate and moderated t-tests from Bioconductor limma package for multivariate analyses. P-values were adjusted for multiple testing.

    Results: Metabolic analyses indicated that 100/2,390 lipid metabolites were significantly associated with PTSD (97 lipid metabolites were upregulated and 3 were downregulated; adjusted p-value < .05). Lipid metabolites that were altered in PTSD included sphingomyelins, ceramides, and phosphatidylcholines. Two lipids were upregulated (one of which was acylcarnitine) and 1 (sphingomyelin) downregulated in males compared to females. Multivariate analyses revealed no significant interactions between PTSD and sex for any of the lipid metabolites.

    Conclusions: Complementing previous findings of an association between hyperlipidemia in PTSD, the metabolomics approach allowed for the identification of lipid metabolites that were associated significantly with PTSD. Our findings are consistent with a previous report of alterations in sphingomyelin metabolism as measured peripherally in combat Veterans with PTSD and in several studies of individuals with major depressive disorders. Sphingolipids constitute a physical barrier in the brain and provide key functions including cell signaling. Sphingomyelin hydrolyses into ceramide and phosphorylcholine. Acid sphingomyelinase, a lipid metabolizing enzyme responsible for this effect is increased in major depression and is reduced with tricyclic antidepressant intake. Acid sphingomyelinase has a role in interleukin-1 release from brain astrocytes, hypothalamic-pituitary-adrenal axis function and has been associated with cardiovascular disease. As such, sphingolipid metabolism may have a role in neuroendocrine and immune alterations in PTSD and associated cardiometabolic health conditions. While there is some evidence that ceramides in the plasma and brain are correlated and cross blood–brain barrier in rats, whether peripheral lipids are an indirect marker of central lipid metabolism has yet to be determined. Identification of key lipid metabolites in conjunction with steroid hormone biomarkers of PTSD may lead to a more precise and reliable phenotype and a clearer understanding of associations between the brain and glucocorticoid and inflammatory pathways to disease in individuals with PTSD.

    Keywords: PTSD, Lipids, Metabolomics

    Disclosure: Nothing to disclose.

    M27. A Computational Approach to Predict Clinical Features of PTSD: An fMRI and Machine Learning Study

    Amin Zand Vakili*, Jennifer Barredo, Emily Aiken, Benjamin Greenberg, Linda Carpenter, Noah Philip

    Alpert Medical School, Brown University, Providence, Rhode Island, United States

    Background: Post Traumatic Stress Disorder (PTSD) is a prevalent and debilitating condition with complex and variable presentation. Core PTSD symptoms (re-experiencing, avoidance, mood/cognition, and arousal) are highly correlated, yet patients often manifest different symptom subsets despite sharing the same categorical diagnosis. Prior research suggests that different PTSD symptom domains might arise from at least partially distinct brain networks. Here, we employed data-driven machine learning to evaluate whether cortical and subcortical functional connectivity might predict a given patient’s PTSD symptom profile. This could provide dimensional approaches to PTSD diagnosis, a greater understanding of the underlying neurobiology, and lead to individualized and rationally targeted treatments.

    Methods: Resting state fMRI was acquired in 47 patients with PTSD (some with comorbid depression) in a mixed veteran and general population sample. We extracted time course data from 100 brain regions of interest (ROIs) based on multi-modal parcellation of cortical regions (Glasser et al. 2016). These regions were within neural networks implicated in PTSD, named in the default mode, salience, and executive networks. The resulting ROI time series data were cross-correlated and then converted to z-scores to improve normality using Fisher’s transformation.

    The resulting high-dimensional matrix of 4,950 potential independent variables was reduced to 46 components with principal component analysis. Sparse (LASSO) regression with a wrapper feature selection algorithm was then applied to this dimensionality-reduced dataset to predict each of the four PCL-5 subscales (i.e., re-experiencing, avoidance, mood/cognition, and arousal). We then utilized full iterative leave-one-out cross-validation to evaluate regression performance.

    Results: LASSO regression applied to fMRI connectivity data significantly predicted PTSD symptom profiles in this sample. The goodness of fit of the model, defined by the coefficient of determination (R2) for re-experiencing, avoidance, mood symptoms, and arousal, was 0.45, 0.35, 0.33, and 0.37, respectively. All were statistically significant (t(45) = 6.01, t(45) = 4.953, t(45) = 4.706, and t(45) = 5.12 respectively, p < 0.0001 in all cases). The results did not depend on the number of items left out in cross-validation, reducing the likelihood of overfitting. Subscales were well-predicted by 4 to 7 of the 46 available principal components.

    The principal components contributing to the prediction of cognitive/mood symptoms, arousal, and re-experiencing were distinct. Functional connectivity amongst anterior lateral orbitofrontal cortex (associated with affective processing) was stronger in individuals with greater mood/cognitive symptoms. In contrast, strong connectivity amongst dorsolateral prefrontal subregions implicated in executive control was seen in those with fewer of these symptoms. Functional connectivity between pars triangularis, a region implicated in both executive and default mode networks, was associated with greater arousal symptoms, whereas connectivity between dorsolateral prefrontal executive regions was observed in those with fewer arousal symptoms. Moreover, connectivity of pars opercularis, a control region contributing to action planning, was more strongly anticorrelated with anterior cingulate subregions involved in salience processing in individuals with more severe mood/cognitive symptoms.

    Conclusions: This, to our knowledge, is the first attempt to identify biological correlates of PTSD symptom subclasses using a purely computational approach. A machine learning algorithm trained on functional connectivity data was able to successfully predict the detailed symptom profile and subscale scores in patients with PTSD. The resulting predictions were made using shared yet distinct connectivity components. Expansion of this work could lead to a better understanding of PTSD psychopathology and help us target treatments to specific symptom domains. More directly targeting the biological nexus of a particular class of symptoms could potentially lead to more rapid and complete clinical improvements in patients with PTSD.

    Keywords: Post Traumatic Stress Disorder, Resting State Functional Connectivity, Machine Learning, fMRI Biomarkers

    Disclosure: Nothing to disclose.

    M28. Extracellular Regulated Kinase 2 Within the Lateral Habenula Promotes Resilience to Emotional and Physical Stress During Adolescence

    Lyonna Parise*, Omar Sial, Astrid Cardona, Eric Parise, Eric Nestler, Scott Russo, Carlos Bolanos-Guzman

    Icahn School of Medicine at Mount Sinai, New York, New York, United States

    Background: Most stress research to date has focused on the effects of physical stress, yet evidence demonstrates that emotional stress is just as deleterious as physical offenses, thus highlighting the necessity for models of stress that focus on non-physically invasive insults. To this end, the chronic social defeat stress (CSDS) paradigm has been adapted to include both a physical (PS) and emotional stress (ES) component. Known as the vicarious social defeat stress model (VSDS), mice exposed to ES exhibit similar behavioral and biological outcomes as those exposed to PS such as increased anxiety-like responses and social avoidance, as well as deficits in the forced swim test, and increased serum corticosterone. One advantage of the CSDS paradigm is the ability to separate between susceptible and resilient phenotypes. Interestingly, the parameters which identify these phenotypes have not yet been applied to behavioral outcomes of VSDS experienced during adolescence. This is surprising given the high incidence of mood-related disorders diagnosed during adolescence and the lack of effective treatment options for this particular population. Indeed, there is much interest in targeting novel brain substrates that may be involved in stress-related dysfunction and develop more efficient therapeutics. In particular, the lateral habenula (LHb) has recently been implicated in mediating depression-related phenotype, however, the underlying mechanisms are poorly understood. Extracellular regulated kinase ½ (ERK2) activity within the mesolimbic reward pathway has been shown to regulate stress and antidepressant-like behavioral reactivity and given the prominent regulatory function LHb serves in this circuit, it is likely that ERK2 plays a role within the LHb in mediating stress-related behaviors.

    Methods: Adolescent male mice were exposed to 10 days of VSDS and their behavioral reactivity to anxiety- and stress-eliciting situations was assessed 24 h and 1 month after the last stress exposure. A separate group of mice were used for tissue collection 48 h after the last defeat bout in order to measure mRNA expression of ERK2 within the LHb. To determine the functional role of ERK2 signaling within the LHb, we delivered herpes simplex virus (HSV) vectors overexpressing GFP (HSV-GFP) or ERK2 (HSV-wtERK2) and then assessed reactivity to stress. To help elucidate ERK2’s role in antidepressant treatment, a separate group of mice were exposed to CSDS and then treated with a single intraperitoneal injection of 20 mg/kg ketamine. Twenty-four hours later, KET-exposed mice were given an intra-LHb infusion of the ERK inhibitor U0126, or its vehicle and their behavior assessed.

    Results: Adolescent mice exposed to VSDS demonstrate a similar split in susceptibility versus resilience as observed in adult mice exposed to CSDS. Surprisingly, the phenotype identified (i.e., susceptible or resilient) during adolescence is maintained into adulthood. When tested 24 h after VSDS exposure, the susceptible mice spent significantly less time in the center of the open field arena, suggesting increased anxiety. In addition, when tested one month later in the elevated-plus maze (EPM), susceptible mice spent significantly less time in the open arms of the EPM as compared to the resilient mice, also indicative of an anxiogenic response. Additionally, when given the choice between water and a sucrose solution, susceptible mice do not show the expected preference for sucrose, suggesting heightened levels of anhedonia. Interestingly, after 10 days of stress exposure, ERK2 mRNA was decreased in the LHb of both ES- and PS-exposed susceptible mice. Resilient mice did not show this reduction in ERK2 mRNA expression within the LHb. When subjected to the accelerated CSDS paradigm, HSV-GFP-infused ES- and PS-exposed mice exhibit avoidance to the novel social target, however, the mice microinfused with HSV-wtERK2 showed interaction scores similar to non-stressed controls. Given that overexpression of ERK2 in the LHb promotes resilience, we assessed how ERK2 inhibition could influence the antidepressant efficacy of KET. After 10d of CSDS, mice given KET exhibited an attenuation of social avoidance and, surprisingly, inhibition of ERK2 by U0126 blocked the antidepressant effects of KET.

    Conclusions: These findings suggest that ERK2 in the LHb is capable of modulating responsiveness to stress and supports the hypothesis that the antidepressant action of KET are, at least in part, mediated by ERK2 activity within the LHb.

    Keywords: ERK, Lateral Habenula, Adolescence, Social Defeat Stress, Emotional Stress

    Disclosure: Nothing to disclose.

    M29. Maternal Care Affects the Development of Fear Learning in Adolescent Nonhuman Primates: Relationship With Prefrontal 5HT1A Receptor Binding & Attention Bias to Threat in Nonhuman Primates

    Elyse Morin*, Alison Wakeford, Brittany Howell, Dora Guzman, Jerrold Meyer, Erin Siebert, Andrew Kazama, Jonathon Nye, Mar Sanchez

    Emory University Yerkes National Primate Research Center, Atlanta, Georgia, United States

    Background: Childhood maltreatment is associated with increased risk for psychopathology, and social and cognitive deficits. Consistent with human evidence linking early life stress (ELS) with poor emotional/stress regulation, as seen in anxiety and mood disorders, our lab has reported that maternal maltreatment (MALT) leads to increased emotional reactivity, anxiety, and impulsive aggression in nonhuman primates. It is unclear if this is due to enhanced fear learning or impaired modulation of fear responses. These questions can be probed using fear-potentiated startle (FPS) paradigms where the inability to modulate fear-conditioned responses by safety cues have been found to be translational biomarkers for disorders such as PTSD. This study used a well-established rhesus model of MALT, which consists of spontaneous and comorbid physical abuse and rejection, leading to infant distress.

    Methods: The long-term outcome of ELS on fear learning was assessed in 25 adolescent macaques (4.5-5.5 yrs), half experienced MALT during infancy (n = 14; 8 M, 6 F), and the other half competent maternal care (Control: n = 11; 5 M, 6 F). It was hypothesized that MALT animals would have higher baseline and FPS, take longer to discriminate fear/safety to attenuate startle, and show impaired extinction. An AX + /BX- fear/safety signal paradigm measured baseline acoustic startle response as an indicator of anxiety, FPS, attenuation of startle by safety signals, and extinction. Attention bias to emotionally valent stimuli was also determined during adolescence using the dot probe task. Due to its role in emotional/stress regulation, prefrontal serotonin (5HT) 1 A receptor binding potential (BP) was also examined using PET imaging during adolescence.

    Results: Baseline startle was higher in MALT than Control, particularly in females, and remained high throughout several rounds of testing, suggesting impaired desensitization to the startle cue. During discrimination training, % FPS was significantly lower in females than males; however, MALT females showed a higher transfer of fear to the safety cue during early training than Control females, suggesting fear generalization. A significant negative correlation was found between PFC 5HT1A receptor BP and startle measures (r = -0.481, p = 0.0173), such that reduced BP was predictive of increased baseline startle amplitude, consistent with human studies on anxiety and depression. In the dot-probe task, MALT animals showed higher reaction times to social threatening images than Controls, suggesting interference of stimuli with negative valence on attentional control and cognitive processes (F(1,22) = 4.5, p = 0.0463). Higher emotional reactivity during infancy in MALT animals predicted attention bias towards threat (β = -0.14, t = -6.1, p = 0.002), whereas higher levels of prenatal cortisol exposure was associated with attention bias away (avoidance of) threat (β = 0.27, t = 2.8, p = 0.037) in both MALT and Control groups.

    Conclusions: This suggests developmental alterations in fear learning related to MALT, especially in females, leading to difficulties in discrimination learning and generalized fear. This interpretation is supported by higher measures of trait anxiety in baseline startle, as well as reduced PFC 5HT1A receptor BP in maltreated animals. Altogether, these findings suggest that different postnatal experiences and early biobehavioral mechanisms regulate the development of fear/stress regulation and emotional attention biases during adolescence.

    Keywords: Early Life Stress, Adolescence, Nonhuman Primate Models

    Disclosure: Nothing to disclose.

    M30. Maternal Immune Activation Impairs Behavioral Flexibility and Alters Transcription in Frontal Cortex

    Susan Powell*, Dionisio Amodeo, ChiYu Lai, Omron Hassan, Eran Mukamel, Margarita Behrens

    University of California, San Diego, La Jolla, California, United States

    Background: Epidemiological studies suggest that the risk of neurodevelopmental disorders such as autism spectrum disorder (ASD) and schizophrenia is increased by prenatal exposure to viral or bacterial infection during pregnancy. It is still unclear how activation of the maternal immune response interacts with underlying genetic factors to influence observed ASD phenotypes.

    Methods: The current study investigated how maternal immune activation (MIA) in mice impacts gene expression in the frontal cortex in adulthood, and how these molecular changes relate to deficits in cognitive flexibility and increases in repetitive behavior that are prevalent in ASD. Poly(I:C) (20 mg/kg) was administered to dams on E12.5 and offspring were tested for social approach behavior in a 3-chambered social approach task, repetitive grooming, and probabilistic spatial reversal learning in adulthood (n = 8 vehicle; n = 9 Poly(I:C)). The frontal cortex of MIA offspring were harvested at 27 weeks (Poly(I:C) offspring = 9 and vehicle offspring = 8). We employed next-generation high-throughput mRNA sequencing (RNA-seq) to comprehensively investigate the transcriptome profile in frontal cortex of adult offspring of Poly(I:C)-exposed dams.

    Results: Both Poly(I:C)-MIA and control mice demonstrated comparable learning of the initial spatial discrimination, with no significant main effect of treatment on the number of trials to criterion (t(15) = -1.47, NS). On the following day mice were first required to demonstrate retention of the previous spatial acquisition. We found no main effect of treatment on the number of trials to retention criterion (t(15) = -0.34, NS), showing that both groups had comparable retention and did not differ on recall of the previous day’s learning. Once performance reached retention criterion, mice immediately began reversal trials. We found that MIA mice required more trials to reach a performance criterion for reversal learning (significant main effect of treatment, t(15) = -5.18, p < 0.001). To determine the nature of the reversal learning deficits in Poly(I:C)-MIA mice, we therefore analyzed perseverative errors. There was no significant difference in perseverative errors between Vehicle and Poly(I:C)-MIA mice (t(15) = 1.56, NS). Poly(I:C)-MIA mice committed significantly more regressive errors during reversal learning (t(15) = 7.92, p < 0.05). Impaired social behavior is a core ASD phenotype and can be altered by MIA. We found that Poly(I:C)-MIA mice spent less time sniffing an unfamiliar mouse compared to vehicle-treated mice (t(15) = 3.90, p < 0.01), indicating reduced social approach behavior. We found a trend toward increased grooming in MIA mice (145 ± 26 s for Poly(I:C)-MIA vs. 95 ± 11 s for controls, onnt(15) = -1.65, p = 0.12). Out of 12,626 genes expressed in frontal cortex, we identified 24 differentially expressed (DE) genes in Poly(I:C)-MIA mice relative to controls (FDR < 0.1, fold-change > 1.1). The disruption of genes involved in glutamatergic neurotransmission (Grm7), K + ion channel activity (Kcnk1) as well as mTOR signaling (Rictor) can be long-lasting after an environmental insult during embryonic development. We hypothesized that Poly(I:C) treatment affect developmentally regulated genes. Consistent with this, 87.5% of DE genes in MIA offspring are genes that are differentially expressed during fetal to adult brain development. Furthermore, Grm7, Rictor, Rpl29rt and Hist1h2bc were associated with altered behavioral phenotypes. The causal mediation and interaction analyses indicated that expression of Kcnk1 was altered by MIA and was associated with reversal learning. Finally, the GSEA analysis indicated that prenatal Poly(I:C) may disrupt ion channel activity, particularly K + ion channels, through a subtle increase in expression of a set of genes involved in voltage-gated and inward rectifying K + channels.

    Conclusions: Our study characterized frontal cortex transcriptome profile of a viral-mimic of MIA effects and its correlation with behavioral phenotypes relevant to ASD. Poly(I:C)-MIA led to an impairment in reversal of spatial discrimination and in social approach behavior, similar to two core features of ASD. Long-term effects of MIA involved dysregulated expression of genes involved in glutamatergic pathway, mTOR signaling and potassium ion channel activity. These gene expression-phenotype correlations provide insight into genes that may underlie the ASD-like behavioral phenotype in the MIA mouse model.

    Keywords: Neurodevelopmental Disorders, RNA-Sequencing, Maternal Immune Activation, Reversal Learning, Social Behavior

    Disclosure: Nothing to disclose.

    M31. Social Isolation During Adolescence Modulates Cognition and Motivation in Adulthood

    Gregory Carr*, Gongliang Zhang, Michael Noback, Noelle White, James Barrow

    Lieber Institute for Brain Development, Baltimore, Maryland, United States

    Background: Social isolation (SI) during adolescence affects neurodevelopment and is a significant risk factor for psychiatric disorders in later life. In mammals, adolescence is a critical period in which social interaction with peers is essential for normal brain development. SI during adolescence is associated with increased neuroinflammation and altered brain structure and neurochemistry. These SI-mediated effects may underlie SI-associated risk, but the pathologically relevant effects of SI are not clear. Previous research has shown that SI from postnatal day 21 to 35 in mice impairs myelination, alters neuronal excitability in the medial prefrontal cortex (mPFC), and modulates behavior in adulthood. SI in humans has also been shown to alter sleep quality, but sleep has not been studied in the mouse model. Here we describe a new set of studies investigating the effects of SI during adolescence on sleep quality and cognitive function during adulthood.

    Methods: Male and female C57BL/6 J mice were weaned at postnatal day 21 and randomly assigned to group housing with littermates or isolated housing until postnatal day 35. After postnatal day 35, isolated mice were paired and lived in these pairs for the remainder of the experiments. All behavioral experiments were started after postnatal day 63, at least four weeks after the end of the isolation period. Separate cohorts of mice were tested in behavioral assays including locomotor activity, 48-hour home cage monitoring, social interaction, touchscreen-based progressive ratio breakpoint tasks and continuous performance tasks (CPT). The behavioral results presented include data from independent cohorts of mice from different litters and breeding pairs (n = 8-10/group).

    Results: Male SI mice are hyperactive (p < 0.05) and have impaired sleep quality compared to group-housed male mice. Male SI mice have more and shorter bouts of sleep during the dark phase of the light cycle compared to group-housed mice (p < 0.01). There was no difference in the total amount of sleep. Female SI mice were no different from group-housed female mice on locomotor activity or sleep behavior. In the progressive ratio breakpoint task, male SI mice have a significantly higher breakpoint compared to group-housed mice, indicating greater motivation to work for a food reward. SI increased the breakpoint without modulating the total number of rewards taken on an FR1 reinforcement schedule. Additionally, preliminary results suggest that male SI mice also show better performance on the CPT sustained attention task utilizing the same food reward as the progressive ratio task (strawberry milk).

    Conclusions: These results provide further evidence that SI during adolescence modulates behavior in adulthood. To our knowledge, this is the first evidence that SI in mice impacts sleep behavior, a well-established relationship in the clinical literature. Additionally, we show that SI in males increases motivation to work for a food reward without significantly altering the hedonic value of the specific reward. This suggests that SI may increase the salience of specific environmental stimuli. Our preliminary data showing improved performance in the CPT supports this conclusion. Future experiments will aim to determine whether these findings extend to other types of appetitive and aversive stimuli.

    Keywords: Attention, Sleep Disturbance, Reward Processing

    Disclosure: Nothing to disclose.

    M32. A Randomized Controlled Trial of Ketamine for Adolescent Treatment-Resistant Depression

    Jennifer Dwyer*, Sarah Yip, Gerard Sanacora, Michael Bloch

    Yale Child Study Center, New Haven, Connecticut, United States

    Background: Nearly one in four adolescents will experience major depressive disorder (MDD), and suicide is the 2nd leading cause of death in this age group. 40% of adolescents with MDD fail to respond to initial treatment with selective serotonin reuptake inhibitors (SSRIs). Better treatments for adolescent depression are urgently needed. Ketamine has rapid antidepressant and anti-suicidal effects in adults with treatment resistant-depression (TRD), but there have been no prospective controlled trials in adolescents. The adolescent brain is a unique pharmacologic substrate and the proposed sites of ketamine’s action (e.g. prefrontal cortex and hippocampus) are actively maturing during this time. Thus, it is important to carefully and directly test ketamine’s antidepressant effects in adolescents with TRD.

    Methods: We have conducted a midazolam-controlled crossover trial to evaluate the effects of ketamine in treatment-refractory adolescent MDD over four weeks. Adolescents (13-17 years old) must have failed at least one adequate trial of a standard antidepressant to enroll. On day 1 and day 14 adolescents receive either ketamine (0.5 mg/kg over 40 minutes) or midazolam (0.045 mg/kg over 40 minutes). Subjects stayed on their psychiatric mediations, with stable dosing for the four weeks prior to the trial and the duration of the trial. For the primary outcome, paired t-tests compare MADRS score at 1 day following infusion between midazolam and ketamine. Scores of the other rating scales and timecourse are assessed as secondary measures. A subset of subjects (n = 5) also underwent MRI neuroimaging at baseline, one day following ketamine, and one day following midazolam.

    Results: Recruitment for this trial will complete as of October 1, 2018 at which point the we will have enrolled at least 18 pediatric subjects with significant, refractory depressive symptoms. Five of these underwent neuroimaging. Preliminary analysis will be presented in this poster.

    Conclusions: Adolescent TRD is a significant public health problem that is associated with significant morbidity and mortality. The brain undergoes substantial maturation during childhood and adolescence, and novel therapeutics must be carefully tested with attention to developmental context. Here we report the results of the first randomized controlled clinical trial of ketamine in adolescents with treatment-resistant depression.

    Keywords: Adolescent Depression, Ketamine, Clinical trial

    Disclosure: Nothing to disclose.

    M33. Decision-Making Predicts Escalating Cannabis Use Among Adolescent Girls and Boys

    Raul Gonzalez*, Samuel Hawes, Ileana Pacheco-Colón

    Florida International University, Miami, Florida, United States

    Background: Individuals with substance use disorders often exhibit impairments in decision-making (DM): the ability to make optimal choices among competing options with uncertain outcomes (reviewed in Bechara, 2005). Deficits in DM have been linked to dysfunction in several cortico-limbic structures including the OFC, PFC, ACC, and insula, which are implicated in addiction neuropathophysiology. However, the evidence for DM impairments among cannabis users is mixed (reviewed in Crean, Crane, & Mason, 2011). Importantly, questions remain regarding whether DM deficits are evident in adolescent cannabis users and whether DM performance may precede and influence cannabis use trajectories. In this study, we examine whether DM performance predicts prospective growth in cannabis use among an adolescent sample and whether sex-differences are evident.

    Methods: Participants were 401 predominantly Hispanic (90%) adolescents recruited from middle- and high-schools in the Miami-Dade area. Most were at risk for cannabis use escalation based on their self-reported drug use history and ranged in age from 14 to 17 (M = 15.40, SD = .72). Data were collected at five bi-annual assessments across 2.5 years. The sample exclusion criteria comprised history of neurological, learning, or severe psychiatric disorders, head injury with loss of consciousness greater than 10 minutes, and history of significant drug and alcohol use. The 4-hour assessment spanned demographic, mental health, patterns and prevalence of substance use and substance use disorders, as well as neurocognitive functioning. DM was assessed with the Iowa Gambling Task (IGT; Bechara et al. 1994), the Game of Dice Task (GDT; Brand et al., 2005), and the Cups Task (CT; Levin & Hart, 2003), which assess DM under conditions of either ambiguity or specified risk, and under conditions of possible loss or gain. All study participants (n = 401) provided baseline data and retention rates across follow-up assessments remained high (94%-98%).

    Results: Full information maximum likelihood (FIML) estimates were used to handle missing data. No differences were found for any study variable when comparing between participants having complete data and those missing data for any assessment. Four indices of performance across three DM tasks were used to construct a latent DM variable. These included the reverse-scored IGT Net Total, the number of risky choices in the GDT, and the total number of risky choices in the gain and loss domains from the CT. Confirmatory factor analysis (CFA) revealed that all four indices loaded significantly on the latent DM variable and the data provided a good fit to the model (CFI = .98; RMSEA = .03). Latent growth curve modeling (LGCM) was used to delineate the longitudinal course of CU. First, analyses examined the shape and model fit of the CU trajectories. Subsequently, analyses were conducted to determine whether baseline DM scores predicted the developmental course of CU when controlling for participant age, estimated IQ, and race/ethnicity. Finally, we examined whether there was an interaction effect between participant sex and scores on the latent DM variable in the prediction of the LGCM of CU. This interaction effect was estimated using the latent moderated structural equation (LMS) method (Klein & Moosbrugger, 2000). All models used maximum likelihood estimation with standard errors and a chi-square statistic that are robust to nonnormality (MLR) in Mplus 7.0. Initial examination of the unconditional CU growth model revealed acceptable model fit (χ2 (32) = 53.76, p = .009; CFI = .93, RMSEA = .08). The mean estimated intercept (M = .82, SE = .18) and linear slope (M = .41, SE = .04) were both significant and positive (p-values < .001), indicating a systematic increase in the average rate of CU across the adolescent study period. Statistically significant variance estimates were also found for the intercept (δ2 = 7.12, p < .001) and linear slope (δ2 = .46, p < .001) of the CU trajectory, indicating significant within-individual variability in initial levels and change over time in CU. In addition, results from subsequent analyses showed that DM predicted increased rates of CU across the adolescent follow-up period (B = .15, SE = .05; p = .01), even after controlling for participant age, estimated IQ, and race/ethnicity. However, there was no evidence of a significant interaction between participant sex and the latent DM variable (B = .11, SE = .19; p = .55) on CU growth.

    Conclusions: Our results indicate that poorer DM predicts increasing use of cannabis during adolescence in both girls and boys. This lends further support to the contention that prevention strategies aimed at improving DM (e.g., Alfonso et al., 2011) may be of some benefit in reducing adolescent cannabis use. These results need to be replicated in more racially/ethnically diverse samples. Future analyses will explore whether specific aspects of DM are more relevant than others and whether mental health variables moderate or mediate this effect. Funded by R01DA031176 to RG.

    Keywords: Cannabis Use, Risky Decision-Making, Adolescent

    Disclosure: Nothing to disclose.

    M34. Mistimed Network Connectivity Associated With Cognitive Control Deficits in Adolescents and Young Adults With ASD

    Marjorie Solomon*, Marie Krug, Cory Coleman, Matthew Elliott, Jeremy Hogeveen, Tara Niendam, Ragland Dan, Cameron Carter

    University of California, Davis, Sacramento, California, United States

    Background: Most individuals with Autism Spectrum Disorder (ASD) exhibit cognitive control (CC) deficits that persist into adolescence and young adulthood. CC deficits may be associated with the difficulties in social functioning, adaptive functioning and restricted and repetitive behaviors that are characteristic of many affected individuals, making it important to understand the precise nature of these deficits. Previous fMRI studies have found that typically developing (TYP) individuals develop a mature LPFC/Parietal network in adolescence, which aids them in successfully inhibiting responses on incongruent trials. Individuals with ASD, however, may continue to rely more heavily on a less mature dACC/LPFC network when required to implement CC. We use data from the completed first wave of a five-year cohort-sequential study of adolescents and young adults with ASD to investigate potential group differences in behavioral performance and neural recruitment during a rapid event-related version of the Preparing to Overcome Prepotency (POP) task in participants ages 12-22 years old.

    Methods: Participants included 60 individuals with ASD (mean age = 17.6 years; mean IQ = 103) diagnosed using gold standard measures and 69 individuals with TYP (mean age = 17.3 years; mean IQ = 109). They completed 4 28 trial runs of the rapid Preparing to Overcome Prepotency (rPOP) task in the scanner. In the rPOP, they were first presented with a fixation cross. This was followed by a cue signaling whether they should push a button on the same side (cued by a green square) as indicated by the probe, which was an arrow, or to the opposite side relative to where the arrow pointed (cued by a red square). After an inter-stimulus interval, the probe arrow was presented. Both inter-stimulus and inter-trial intervals were jittered (mean time = 4,500 milliseconds (ms)). Data was acquired using a 3 Tesla Siemens Tim Trio with a 32-channel head coil. Data were preprocessed and analyzed using SPM12. Cue and probe phases of red and green trials were modeled separately in the GLM. Error and post-error trials were modeled separately from correct trials and excluded from contrasts. The GLM included translational and rotational movement and ART outlier regressors. Seed to voxel whole brain functional connectivity (fc) analyses were implemented using PPI in the CONN functional connectivity toolbox ( Dorsolateral prefrontal cortex (DLPFC) and dorsal anterior cingulate (dACC) seeds were derived from the Schaefer et al. (2017) Atlas.

    Results: Behavioral performance on the task as indexed by the inverse efficiency score (IES; RT/accuracy) showed that there was a main effect of diagnosis (F = 3.95, p < .05) a trend level main effect of cue type and a significant cue type X diagnosis interaction (F = 3.99, p < .05), indicating that the ASD group was less efficient at the task. Whole brain analyses showed no group differences in recruitment for the red-green contrasts in either the cue or the probe. FC analyses showed that during the cue phase, the TYP versus the ASD group exhibited greater fc between the DLPFC seed and parietal cortex, left hippocampus, and pre and postcentral gyri; and greater fc between the dACC seed and bilateral parahippocampal gyrus, and the medial precuneus. During the cue phase of the task, the ASD versus the TYP group demonstrated greater fc between the DLPFC seed and the parietal cortex, right insula, and right middle and inferior frontal gyri; and greater fc between the dACC seed and the insula and the right inferior frontal gyrus. During the probe phase of the task, the TYP versus the ASD group showed greater fc between the DLPFC seed and the ACC, anterior PFC, and the insula; and greater fc between the dACC seed and middle and inferior frontal gyri, while the ASD versus the TYP group showed greater fc between the dACC and other PFC regions. At cue, for the ASD group, greater dACC-right inferior frontal gyrus and DLPFC-middle frontal gyrus fc was associated with poorer task performance, while greater dACC to fc connectivity (characteristic of the TYP group) was correlated with better task performance. At the probe, greater DLPFC-insula fc (characteristic of the TYP group) was associated with better task performance in the ASD group.

    Conclusions: Seed to voxel fc analyses using PPI illustrated that during the presentation of red cues compared to green cues, both the ASD and TYP groups showed significant fc between frontal seeds and parietal regions. However, at cue the TYP group showed greater fc between seed regions and the hippocampal regions perhaps suggesting the retrieval of learned rules, while the ASD group showed greater fc between dACC and DLPFC seeds and insula and middle and inferior frontal gyri. At the probe, greater connectivity in these regions was seen for TYP compared to ASD, providing evidence that cognitive control processes are mistimed in ASD and involve greater insula and less hippocampal recruitment in the preparatory phases of the task. Associations with task performance provide evidence that the mistiming found in ASD is detrimental to task performance and that the networks recruited by the TYP group are more effective.

    Keywords: Cognitive Control Network, Adolescent Brain and Cognitive Development Study, Functional Magnetic Resonance Imaging

    Disclosure: Nothing to disclose.

    M35. Maturational Changes in Event-Related Potential Components, Cortical Thickness, and Substance Use Disorder Outcome in Young Adulthood

    Shirley Hill*, Stuart Steinhauer, Brian Holmes, Jeannette Locke-Wellman

    Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania, United States

    Background: Familial risk differences in amplitude of specific ERP components have been demonstrated in multiple replications across labs. Some of these differences appear to change with development most likely in association with maturation of brain structures. Cortical thickness is one measure of brain morphology that changes with development. With the capability of assessing multiple brain regions using FreeSurfer, we tested whether regional differences in cortical thickness map to specific ERP measures.

    Methods: Adolescent and young adult male and female participants (N = 217) with a mean age of 25.2 + 4.9 years; from high (N = 118) and low risk (N = 99) for alcohol dependence families were scanned using magnetic resonance imaging (MRI) at 3 T. Cortical thickness was analyzed using FreeSurfer. The participants were part of longitudinal study in which multiple event-related potential recordings were obtained during childhood, adolescence, and young adulthood. Using the visual ERP recording nearest in time to the MRI scan, preliminary analyses were completed to determine the patterns of associations between brain regions and ERP components. Those regions showing significant association were tested for familial risk group differences with those showing significance entered into a survival model to determine their relationship to substance use disorder outcome.

    Results: Among the significant associations with familial risk were cortical thickness of the right hemisphere (rh) parsopercularis and the temporal pole in the left hemisphere (lh). These regions were significantly related to risk group status: lh temporal pole (F = 4.97, df 1, 215, p = 0.027) and the rh parsopercularis (F = 6.98, df = 215, p = 0.009). These variables, along with ERP characteristics at baseline (median age = 11 years) and follow up (median age = 24 years), Iowa Gambling Test scores, and impulsivity scores when entered into a survival model showed several significant predictors of SUD outcome, some spanning a 13 year follow up period.

    Conclusions: Neural activity reflected in P300 amplitude is positively related to cortical thickness in regions found to differ by familial risk status and involved in impulsivity and decision-making.

    Keywords: Cortical Thickness, Event-Related Potentials, Alcohol and Substance Use Disorders

    Disclosure: Nothing to disclose.

    M36. ADHD Symptom Severity in Children is Associated With Reduced Sleep Slow Wave Activity and Default-Mode Network Connectivity After Acute Sleep Restriction

    Jared Saletin*, Jessica Haddad, Gabriela De Queiroz Campos, Mary Carskadon, Daniel Dickstein

    Alpert Medical School, Brown University, Providence, Rhode Island, United States

    Background: Attention-deficit-hyperactivity disorder (ADHD) is commonly associated with disrupted sleep. Unclear, however, is whether ADHD symptoms convey vulnerability to neural consequences of sleep loss. Thus, we conducted a within-subject extended wakefulness protocol combining quantitative analysis of the sleep EEG with next-day functional neuroimaging of resting-state networks in a sample of children across symptom-severity.

    Methods: Eight medication-naïve children (2 F; age: 11.3 ± 1.3 years) without mood or anxiety comorbidity were assessed for ADHD symptoms (inattention, hyperactivity) using average parent and teacher Conners-3 T-scores. Participants slept on a 1-week 9.5-hour schedule (2100h-0630h) before entering the lab for an adaptation session followed by two consecutive nights of sleep: a night of 9.5-hours time-in-bed, and a subsequent night of sleep, restricted to 4-hours time-in-bed (0230h-0630h). Sleep was monitored by a 25-channel topographical EEG array. Each morning (~0900 h) participants completed a resting-state functional magnetic resonance imaging scan, allowing within-subject comparisons of rested wakefulness (RW) with sleep restriction (SR). We examined (1) whether cortical dynamics of sleep homeostasis (indexed by NREM slow wave activity [SWA]; EEG power in the 0.6 – 4.8 Hz band) following sleep restriction are modulated by ADHD symptoms, and (2) how connectivity within the default-mode-network shifts as a function sleep loss and symptom severity.

    Results: When examining polysomnographic sleep, our sleep restriction protocol was successful in increasing the proportion of time spent in slow wave sleep (p < .001) as well as increasing EEG SWA across the entire scalp (p < .05). Turning next to inter-individual differences, we identified a local difference in recovery SWA as a function of symptoms. Heightened ADHD symptoms were associated with an attenuated recovery increase in SWA over frontocentral cortex (EEG derivation FCz; p < .05). Turning next to brain connectivity the morning after SR, we identified a group-wise reduction in DMN connectivity in SR, compared to RW (t(7) = -3.03, p = .019). However, more severe ADHD symptoms were associated with greater reductions in DMN connectivity; both for hyperactivity (b = -.0031; t = -4.01; p = .007; r2 = .45) and inattention (b = -.0029; t = -2.56; p = .042; r2 = .35).

    Conclusions: These initial results suggest young children struggling with attention deficits may be more severely impacted by sleep loss. First, their ability to illicit a robust recovery response in EEG slow wave activity is diminished, perhaps implying insufficient restoration of brain function the next morning. Second, symptoms were equally associated with greater sleep-restriction-induced reductions in neural connectivity following extended wakefulness. We are currently expanding the sample while exploring interactions between sleep physiology and next-day network integrity in both accounting for symptoms, and ultimately, cognitive and behavioral impairments. More generally, however, these data underscore the critical importance of considering sleep as a contributory factor to the neurobiology of ADHD and indicates sleep as a potential modifiable factor in these youth with profound impact for daytime function.

    Keywords: Sleep, Sleep Deprivation, ADHD, Resting State Functional Connectivity, Quantitative EEG

    Disclosure: Nothing to disclose.

    M37. Computational Approaches for Early Emerging Heterogeneity and Disorder Risk

    Suma Jacob*, Jason Wolff, Colleen Doyle, Michael Steinbach, Vipin Kumar, Jed Elison

    University of Minnesota, Minneapolis, Minnesota, United States

    Background: Variability across the typical-to-atypical continuum and heterogeneity within and across early emerging psychiatric disorders remains a challenge for identifying biological mechanisms and targeted treatments. In this project, we are using two unsupervised data-driven computational approaches to identify profiles of risk for early atypicality in neurodevelopment. Using 1) Factor Mixture Modeling (FMM) and 2) Anomaly Detection to identify single cases or outlier clusters of children, we combine these complementary procedures in order to stratify clusters of cases based on a behavioral risk profile. Our goal is to identify high-risk children at rates consistent with epidemiological estimates of ASD and associated developmental disabilities (e.g., language or global developmental delay).

    Methods: We collected parent-report data from 1570 children between 17- 25 months of age in a community sample. Caregivers filled out Video-Referenced Rating of Reciprocal Social Behavior, the Repetitive Behavior Scale for Early Childhood (RBS-EC), and the MacArthur-Bates Communicative Development Inventories (MCDI), and demographic information online. A subset of participants was invited for follow-up assessments before thirty-six months of age.

    Results: With the FMM analyses, we identified five subgroups. Follow-up with a subsample of 107 children confirmed the predictive validity of the risk profiles, showing significant differences between high-, moderate-, and low-risk groups on internalizing, externalizing, and dysfunctional behavior. Comparison of high- and low-risk groups revealed large effect sizes for internalizing (d = 1.39), externalizing (d = 0.83), and dysregulation (d = 1.87). With the Anomaly Detection analyses, we determined that 80 children had local outlier factor (LOF) anomaly scores > 1.32 out of the overall sample of 1570 toddlers. This represents about 5% of the sample and is consistent with prevalence data for developmental delay and ASD. The longitudinal follow up subsample (N = 107) allowed us to determine that the subset of children with higher LOF scores maintained higher risk symptoms when they were reassessed six months later (e.g. continued to show social impairment).

    Conclusions: Combining computational approaches that identify subgroups and outliers may be required to capture the heterogenous profiles of children at risk for neurodevelopmental disabilities. We anticipate that clusters and profiles identified through computational analyses (FMM, Anomaly Detection) will be differentiated by dimensional cognitive and behavioral features, including cases of ASD or related childhood disorders that are clustered into distinct groupings. Next steps include larger samples and longitudinal clinical assessments to determine co-existing risk for emerging anxiety, affective, and/or behavioral disorders. These computational approaches will allow us to model heterogeneity in early emerging child psychiatry disorders.

    Keywords: Computational Psychiatry, Neurodevelopmental Disorders, Early Identification of Risk

    Disclosure: Roche/Genentech, Advisory Board

    M38. Microbiome-Bile Acid Cross-Talk as a Mechanism Underlying Antipsychotic-Induced Weight Gain

    Erika Nurmi*, Jude McElroy, Reimond Costa, Laughlin Christopher, James McCracken, Gerhard Hellemann, Kym Faull, Chadi Calarge, Lauren Seaman

    University of California, Los Angeles, Los Angeles, California, United States

    Background: Weight gain and metabolic syndrome are common and serious treatment-limiting adverse effects of medications commonly used to treat mental illness, especially antipsychotic drugs. Children and adolescents are at increased risk for these side effects, resulting in substantial morbidity and dysfunction. While the mechanism of antipsychotic-induced weight gain (AIWG) is poorly understood, emerging preclinical and clinical data suggest a role for the gut microbiome. AIWG is absent in germ-free mice but can be induced by microbiome transplant. Likewise, AIWG in rodents can be ameliorated by an oral antibiotic cocktail. In humans, treatment with the antipsychotic risperidone (RSP) produces “obesogenic” microbiome shifts in children, an effect that was magnified in those with significant weight gain.

    Given a growing appreciation for the role of bile acids (BA) in energy balance and microbiome-BA interaction in the gut, we tested whether BAs might represent a downstream effector of microbiome changes. BAs can serve as signaling molecules regulating energy balance through their activity at the hepatic nuclear receptor farnesoid X (FXR). Primary BAs (cholic acid [CA] and chenodeoxycholic acid [CDCA]) synthesized by the host are FXR agonists. Endogenous primary BAs are converted to secondary BAs by epimerase + gut bacteria and reabsorbed into the portal circulation. In contrast to the primary BAs, the secondary BA ursodeoxycholic acid (UDCA) acts as an FXR antagonist. Therefore, shifts in the balance of host-derived primary BAs and microbiota-derived secondary BAs could result in weight gain and metabolic dysfunction through downstream competitive effects on FXR signaling.

    Methods: In 30 youth from the NIMH-funded RUPP-RSP clinical trial of children with autism spectrum disorder, we measured plasma levels of the five most prevalent human BAs at baseline and after 8 weeks of RSP treatment. Using mass spectrometry, we analyzed and compared changes in primary versus secondary BAs longitudinally.

    Results: We observed marked changes in the total BA pool and diversity, with predominant increases in primary BAs. Primary BAs were significantly (p = 0.01) more pronounced in those with high (increased 6-fold) versus low (doubled) AIWG. This pattern was replicated in an independent sample of 4 subjects started on RSP with moderate weight gain. Both CDCA and UDCA measures at baseline and 8 weeks were available for 28/30 RUPP participants. 12 of these 28 displayed a significant change (from baseline to 8 weeks) in either CDCA or UDCA ( > .5 SD). In this subset showing BA change with RSP, the ratio of the change in CDCA to the change in UDCA differentiated those with AIWG (∆CDCA/∆UDCA > 1) from those without (∆CDCA/∆UDCA < 1). This algorithm correctly predicts AIWG status in 11/12 (92%) participants whose ratios change (p = 0.001), with a sensitivity of 100% and specificity of 86%. Sex, age, race/ethnicity, and RSP dose/plasma level were tested but did not contribute to BA differences. Test-Retest reliability of CDCA, UDCA, and primary BA repeated measurements in the same individual (N = 2) were precise within 4 nmol/L.

    We hypothesized that a ∆CDCA/∆UDCA < 1 could occur when epimerase + bacteria are unavailable to convert CDCA to UDCA, resulting in excess CDCA and AIWG. Therefore, we examined whether relative abundance of the most common species of epimerase + bacteria (B. fragilis) was affected by RSP exposure. As predicted, in a previously published longitudinal sample of 5 subjects, the relative abundance of B. fragilis decreased over 200-fold after 1-3 months of RSP exposure.

    In order to explore whether these effects are specific to AIWG, we tested a small pilot sample of children taking SSRIs. 10 youth were selected, 5 with moderate ( > 1 Z-score) SSRI-induced weight gain (SIWG) and 5 without. While no effects were statistically significant in this very small sample and the effect sizes of differences were smaller than those observed in AIWG, the trends were in the same direction, with individuals with greater weight gain showing greater elevation in primary BA and ∆CDCA/∆UDCA > 1. Using this ratio, SIWG status can be correctly predicted in 6/8 (75%) of individuals with BA change on an SSRI.

    Conclusions: Our preliminary data suggests that antipsychotics impact BA pathways and that alterations to the BA balance may contribute to AIWG, a link that to our knowledge has not been previously investigated. We have shown that RSP-exposed children exhibit substantial changes in their BA pool, with opposite patterns of primary vs. secondary BAs seen in youth at risk versus protected from AIWG. If proven, clear targeted interventions are currently available that may mitigate adverse effects. UDCA is available orally as a gallstone treatment and epimerase + bacteria are a chief component of probiotics. Drugs directly impacting FXR signaling are under development and prebiotic strategies could also prove useful. A better understanding of this complex mechanism will provide the foundation for targeted treatment strategies for drug-induced and, potentially, other-cause weight gain and metabolic dysfunction.

    Keywords: Antipsychotic-Associated Obesity, Gut Microbiome, Bile Acids

    Disclosure: Nothing to disclose.

    M39. Association of Childhood Adversity With Differential Susceptibility of Transdiagnostic Psychopathology in Adulthood

    Cristina Albott*, Miriam Forbes, Justin Anker

    University of Minnesota, Minneapolis, Minnesota, United States

    Background: Multivariable comorbidity research indicates that childhood adversity increases the risk for the development of common mental disorders. This risk is explained by two underlying transdiagnostic constructs, internalizing and externalizing, which are amplified by environmental stressors. The differential susceptibility model suggests that this interaction of risk and environment is bidirectional: at risk individuals respond to stressful environments “for worse” but also respond “for better” in low-stress environments. The present study tested the differential susceptibility model by examining how a history of childhood adversity influences the impact of environmental stress on transdiagnostic psychopathology factors.

    Methods: Data were drawn from the National Epidemiological Survey on Alcohol and Related Conditions (NESARC), an observational longitudinal survey. Data from adults (aged 18 years and older) that completed the survey at two time points: Wave 1 (2001 to 2002) and Wave 2 (2004 to 2005) (N = 34,458) were used for the analyses. We created latent variables of internalizing, externalizing, and general psychopathology factors based on the number of psychiatric diagnoses. By comparing data collected at Wave 1 to Wave 2, intra-individual change scores were calculated for adult stressful life events and transdiagnostic psychopathology factors (i.e., internalizing-fear, internalizing-distress, externalizing, and general psychopathology). Analyses examined how childhood adversity exposure moderated the relationship between change in adult stressful life events and change in adult transdiagnostic psychopathology factors. Data were analyzed using repeated measures linear mixed-effects models.

    Results: Childhood adversity significantly moderated the relationship between changes in adult stressful life events and changes in all transdiagnostic psychopathology factors. Specifically, higher levels of childhood adversity were related to a stronger relationship between adult stress and adult transdiagnostic psychopathology factors. Further, there were significant differences between childhood adversity groups in the mean scores of all transdiagnostic psychopathology factors for both increases in stressful life events and decreases in stressful life events, providing preliminary evidence for differential susceptibility.

    Conclusions: The findings of this study contribute to our understanding of the enduring effects of childhood adversity. The results provide empirical support for the hypothesis that childhood adversity may represent a plasticity factor—rather than only increasing individuals’ vulnerability to stressful life events. Further research is needed to replicate this finding, and subsequently to understand whether the goal of psychiatric interventions should be to focus on reducing environmental stressors, or on attempting to identify and modify the endogenous processes that underlie heightened reactivity to the environment.

    Keywords: Childhood Adversity, Internalizing Disorders, Externalizing Disorders, Early Life Adversity, Stress Reactivity

    Disclosure: Nothing to disclose.

    M40. Genetic Risk for Alzheimer’s Disease and Functional Brain Connectivity in Children and Adolescents

    Luiza Axelrud, João Sato, Marcos Santoro, Fernanda Talarico, Daniel Pine, Luis Rohde, Andre Zugman, Edson Amaro, Rodrigo Bressan, Rodrigo Grassi-Oliveira, Pedro Pan, Andre Simioni, Hakon Hakonarson, Elisa Brietzke, Ary Gadelha, Renata Pelegrino, Marcelo Hoexter, Euripedes Miguel, Sintia Belangero, Giovanni Salum*

    Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil

    Background: Previous research suggested regions prone to accumulate tau aggregates in late-life form ten tau pathology networks. Nevertheless, studies investigating the impact of genetic risk for Alzheimer’s disease on early brain connections are lacking. Here we aim to investigate whether the polygenic risk score for Alzheimer’s disease influences the connectivity of brain regions susceptible to tau pathology during neurodevelopment.

    Methods: Participants were youth aged 6 to 14 years, recruited in Porto Alegre (discovery sample, n = 349) and São Paulo (replication sample, n = 315), Brazil. The polygenic risk score for Alzheimer’s disease was calculated using summary statistics from the International Genomics of Alzheimer’s Project. 6-min resting state fMRI was obtained from all participants. The connections between the local maxima of tau pathology networks were used as dependent variables.

    Results: The polygenic risk score for Alzheimer’s disease was associated with the connectivity between right precuneus and right superior temporal gyrus (β = 0.2, padjusted = 0.032 for discovery sample; β = 0.2, p = 0.018 for replication sample), regions associated with the ventral Default Mode Network and language network, respectively.

    Conclusions: Therefore, brain connectivity between areas susceptible to tau pathology might be affected in youth with genetic susceptibility to Alzheimer’s disease. This suggests a neurodevelopmental contribution to Alzheimer’s disease pathogenesis.

    Keywords: Alzheimer’s Disease, Brain Connectivity, Children and Adolescents

    Disclosure: Nothing to disclose.

    M41. Differences in Brain Myelination Patterns Associated With 7q11.23 Copy Number Variations

    Enock Teefe*, Michael Gregory, J. Shane Kippenhan, Tiffany Nash, Shannon Grogans, Samuel Elliott, Franchesca Kuhney, Daniel Eisenberg, Carolyn Mervis, Karen Berman

    National Institutes of Health, Bethesda, Maryland, United States

    Background: Genetic variation is translated to cognitive and behavioral features through molecular, cellular, and systems-level neurocircuitry. Copy number variations (CNVs), in particular, have been strongly associated with neuropsychiatric disorders, but the underlying neurogenetic mechanisms are still largely unknown and likely complex. Disorders with well-circumscribed CNVs and highly consistent behavioral and cognitive features provide a privileged setting to understand how these neurogenetic mechanisms translate into brain phenotypes and ultimately into complex behaviors. In this study, we investigated two such disorders, Williams syndrome (WS) and 7q11.23 duplication syndrome (Dup7). WS and Dup7 are related, rare neurodevelopmental disorders for which affected individuals possess either hemideletions (one copy, as in WS) or duplications (three copies, as in Dup7) of the same ~1.5 megabases of the 7q11.23 chromosomal locus. These two CNV-related disorders also contrast in their cognitive and behavioral presentations. While WS is typified by a hypersocial personality, deficits in visuospatial construction, and relative strength in concrete language abilities, individuals with Dup7 typically exhibit social anxiety and have visuospatial skills at the level expected for their language abilities. Though prior neuroimaging studies have identified gray- and white-matter alterations in WS, structural brain changes in Dup7 have yet to be examined. No studies have directly measured brain myelination in either disorder. Here, we tested for 7q11.23 CNV-related changes in brain myelination using an MRI technique called Multicomponent Driven Equilibrium Single Pulse Observation of T1 and T2 (mcDESPOT), which provides regional quantitative measures of Myelin Water Fraction (MWF).

    Methods: Fifteen children with WS (mean age = 13.3 ± 3.9 years; 11 girls), 11 children with Dup7 (age = 13.1 ± 3.4 years; five girls), and 32 unrelated typically developing children ([TD]; age = 13.3 ± 3.7 years; 15 girls) underwent neuroimaging using a GE 3T-MRI scanner. mcDESPOT data acquisition includes 8 flip angles of a Spoiled Gradient recalled echo (SPGR) MR sequence, 8 flip angles of Steady State Free Precession (SSFP) MR sequences at phase 0 and phase 180, and an inversion-recover SPGR sequence. Voxel-wise MWF maps for each participant were calculated using a three-pool model, which determined MWF, intra/extracellular water fraction, and the free water fraction. Multivariate modeling was carried out to determine voxels where MWF was associated with 7q11.23 CNV (p < 0.05, FWE-corrected for multiple comparisons), covarying for age and sex.

    Results: Nine clusters showed a positive association between MWF and 7q11.23 copy number (Dup7 > TD > WS). These clusters included white-matter underlying bilateral occipito-temporal junction, inferior parietal lobules, orbitofrontal cortex, left superior temporal gyrus and frontal operculum. No associations in the opposite direction (WS > TD > Dup7) were observed.

    Conclusions: These results indicate that 7q11.23 CNVs alter myelination of brain regions implicated in visuospatial and social processing and suggest a central role for white matter alterations in the neuropathology of these behavioral domains. Ongoing studies will combine multimodal neuroimaging methods and longitudinal modeling in these participants to identify neural pathways where there is a convergence of evidence for effects of 7q11.23 CNV on developmental trajectories.

    Keywords: Williams Syndrome, Myelin Imaging, Copy Number Variants, 7q11.23

    Disclosure: Nothing to disclose.

    M42. Measuring Regional Cerebral Perfusion in Adults and Children With Williams Syndrome Using Oxygen-15 Water PET and Arterial Spin Labelling MRI

    Michael Gregory*, J. Shane Kippenhan, Carolyn Mervis, Daniel Eisenberg, Mbemba Jabbi, Philip Kohn, Tiffany Nash, Lalith Talagala, Karen Berman

    National Institute of Mental Health, Bethesda, Maryland, United States

    Background: Williams syndrome (WS) is caused by hemizygous deletion of approximately 1.6 megabases at chromosomal locus 7q11.23. Individuals with WS are characterized by marked visuospatial construction deficits, hypersociability, and mild to moderate intellectual disability with relative sparing of concrete language. Because the genes within the hemideleted region of chromosome 7 are known, WS has been used as a model for understanding how genetic changes translate into alterations in complex behaviors through brain function. Previous work has shown that adults with WS have altered regional cerebral blood flow (rCBF), as measured by oxygen-15 water positron emission tomography (PET) scanning, particularly of the hippocampus and insula (Meyer-Lindenberg, et al., 2005; Jabbi, et al., 2012). However, it is unclear when during development these differences arise. Because PET scanning requires administration of radiotracers, use of this method with children is precluded. Instead, arterial spin labelling (ASL) with MRI scanning has allowed for non-invasive quantification of rCBF without exposure to radiation or the need for an IV contrast agent. Here, we measured rCBF in adults with WS using the oxygen-15 water PET method, and in children with WS using ASL MRI.

    Methods: Fourteen adults with WS (mean age = 27.8 ± 9.3 years, seven males) and 14 healthy adults (32.6 ± 7.9 years, eight males) completed resting-state oxygen-15 water PET scanning on a GE PET scanner (four 60-second scans/participant using 12 mCi/scan). Ten children with WS (15.2 ± 4.6 years, one male) and 19 TD children (15.2 ± 4.7 years, three males) underwent 3D pseudo-continuous ASL scanning using a 3 T GE MRI scanner. PET images were background subtracted, aligned across scans, registered to each participant’s structural MRI image, normalized to a group-specific template in MNI space, and smoothed using a 10 mm FWHM kernel. Data were compared across groups using SPM5. ASL rCBF images were aligned to each individual’s structural MRI scan, normalized to a group specific template in MNI space, and smoothed using a 6 mm FWHM kernel. Data were compared across groups using AFNI tools. Resulting statistical maps were thresholded at p < 0.005, uncorrected and compared across modalities for convergence of results.

    Results: PET and ASL methods both revealed differential rCBF in the cerebellum such that individuals with WS had greater rCBF than unaffected individuals, particularly in the vermis. Individuals with WS also showed decreased rCBF with both methods in the bilateral insulae, bilateral intraparietal sulci (IPS), left fusiform gyrus, left amygdala, and bilateral motor cortex. PET imaging, but not ASL, showed significantly lower rCBF in the hippocampi of participants with WS, while ASL showed significantly lower rCBF of the left orbitofrontal cortex and dorsal anterior cingulate in participants with WS.

    Conclusions: We used two independent methods to show that regional cerebral blood flow is significantly altered in individuals with WS, in two independent cohorts at different stages of development. The results were largely convergent both between the two modalities and with previously published work showing structural and functional brain alterations in WS. Previous work has shown that the cerebellum is physically larger in both children and adults with WS, and this size has been related to variations in cognitive abilities (Menghini, et al., 2013). Prior work also has identified altered rCBF, gray matter volume, and function of the insula in WS (Jabbi, et al., 2012), consistent with the current findings. Further, significant evidence implicates the IPS as a hub of dysfunction in WS, particularly as related to visuospatial deficits (Sarpal, et al., 2008). Other work has implicated the amygdala in the hypersocial personality (Schwartz, et al., 2017) and has found altered organization of the corticospinal tracts leading to the motor cortex (Marenco, et al., 2007). At the same time, the methods were not convergent with regard to the hippocampi, anterior cingulate, and orbitofrontal cortex. However, these regions all have been previously identified as having significance for the WS neurobehavioral phenotype. Further work will be required to determine whether this lack of convergence is due to the relatively small sample size, to the participants being at different stages of development, or to methodological differences between the PET and ASL rCBF methods.

    Keywords: Perfusion, Cerebral Blood Flow, 7q11.23, Copy Number Variation, Neurodevelopmental Disorders

    Disclosure: Nothing to disclose.

    M43. Sex Differences in Brain Network Connectivity Prior to Puberty

    Katherine Reding*, Shau-Ming Wei, Michael Gregory, Pedro Martinez, Elizabeth Robinson, Jasmin Czarapata, Jordan Barone, Austin Boroshok, J. Shane Kippenhan, Philip Kohn, Lynnette Nieman, Jack Yanovski, Peter Schmidt, Karen Berman

    National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland, United States

    Background: Sex differences in the prevalence of several behavioral disorders emerge during adolescence and are thought to be related, at least in part, to the (re)activation of the hypothalamic pituitary gonadal axis (i.e., puberty). One hypothesis is that the rise of gonadal steroids during this time affects neurodevelopment such that boys and girls differentially become more (or less) vulnerable to specific disorders. Neuroimaging data point to both sex and developmental differences in the functional connectivity (FC) of several resting state networks, including the Default Mode Network (DMN) and Executive Control Network (ECN), and alterations in FC have been associated with psychopathology during adolescence. First, to establish a normative baseline, we tested for sex differences in resting-state FC in healthy, typically developing children prior to puberty. Then, to probe the stability of these sex differences, we examined FC across pubertal development and into adulthood.

    Methods: One hundred and eighteen children, adolescents, and adults were enrolled. Pubertal stage (PS) for child and adolescent participants was defined by clinical exam based on a five- stage scale assessing breast development in girls and testicular volume in boys. Prepuberty was defined by the absence of secondary sex characteristics (PS1), while the pubertal group ranged from PS2-5. All participants completed two six-minute resting state fMRI scans on a 3 T GE MRI scanner, and data were concatenated (for a total of 12 minutes), motion-corrected (framewise displacement < 0.5 mm), spatially smoothed (6 mm FWHM), and bandpass filtered (0.008 < f < 0.1) using AFNI and normalized to MNI space using ANTs. Signals associated with nuisance variables, as determined by anatomical CompCor, and motion parameters were removed. Resting-state data were first analyzed in the prepubertal cohort using a connectome-wide association study (CWAS) to identify clusters of voxels showing sex differences in connectivity across the brain. These CWAS-derived clusters, delineating overall, across-the-brain prepubertal sex differences, were then used as seed regions in a whole-brain voxel-wise analysis to delineate the direction, magnitude, and anatomical locale of the underlying differential FC. To examine the development of prepubertal sex differences, the significant regions identified in the CWAS analysis of the prepubertal children were also used to probe FC in both the pubertal and adult groups. Specifically, Pearson’s r values were extracted from each region of interest to identify the main and interaction effects of developmental group (prepubertal, pubertal, adult) by sex (male, female) on FC.

    Results: There were 41 prepubertal children (8.7 ± 0.3 yrs, 43.9% girls), 31 pubertal adolescents (13 ± 0.6 yrs, 45.2% girls), and 46 adults (30.5 ± 3.8ys, 58.7% women). CWAS within the prepubertal group identified a cluster in the medial prefrontal cortex (mPFC) that showed the most robust sex difference (p < 0.001, uncorrected). The mPFC was then used as a seed-region to identify the directionality and magnitude of this sex difference in whole-brain FC. There were seven resulting clusters with more robust mPFC FC in girls than boys, including areas within the Default Mode and Executive Control Networks ([DMN, ECN], p’s < 0.05, FDR-corrected). There were no regions where FC was more robust in boys than girls. To extend these prepubertal DMN and ECN findings to the pubertal and adult groups, we examined functional connectivity between the mPFC seed and one canonical cluster from each network (as defined by the prepubertal group findings). For the DMN, the mPFC seed showed a significant sex-by-developmental group interaction with the canonical posterior cingulate cortex (p < 0.001, corrected): the prepubertal sex difference (girls > boys) in FC was not observed in puberty but re-emerged in adulthood (p’s < 0.005, corrected). For the ECN, the mPFC seed showed a significant sex-by-developmental group interaction with the canonical dorsolateral prefrontal cortex (dlPFC) seed (p < .001, corrected): the observed prepubertal sex difference (girls > boys) in FC was not found in the pubertal or adult groups (p’s < 0.005, corrected).

    Conclusions: Prior to the onset of puberty, girls show higher mPFC FC to regions within the DMN and ECN compared to boys. These sex differences prior to puberty may reflect earlier sex-specific exposures to adrenal sex steroids (or their metabolites) secondary to an earlier onset of adrenarche in girls and/or greater ovarian-derived estrogen exposure, or they may occur independently of direct hormonal action. Across developmental groups, prepubertal sex differences in FC between mPFC and DMN were not observed in puberty but reemerged in adulthood, whereas prepubertal sex differences in FC between mPFC and ECN were not observed in either puberty or adulthood. These data are consistent with previous neuroimaging studies across development suggesting that the DMN and ECN are still maturing through early adolescence and, thus, may undergo significant restructuring during puberty. Moreover, these results emphasize the importance of establishing sex differences prior to the onset of puberty – a time of relative absence of sex steroid hormone in comparison to mid to late puberty – for elucidating the roles of sex steroid hormones on brain function throughout pubertal development.

    Keywords: Puberty, Default mode network (DMN), Resting State Functional Connectivity, Connectome-wide association study, Sex Differences

    Disclosure: Nothing to disclose.

    M44. Functional MRI Markers of Transdiagnostic Risk for Psychopathology: A Meta-Analysis

    Jacqueline Clauss*, Daphne Holt

    Harvard Medical School, McLean Hospital, Massachusetts General Hospital, Belmont, Massachusetts, United States

    Background: In recent decades, increasing emphasis has been placed on identifying high-risk individuals and understanding the neurobiology of risk for psychiatric disease. Reproducible, brain-based biomarkers for psychiatric disease may provide the basis for using quantitative targets in early intervention. While individual illnesses, such as depression or psychosis, have been associated with specific risk profiles, increasing evidence suggests that many of the genetic and environmental risk factors initially linked to distinct psychiatric illnesses are shared across conditions. This is consistent with the well-established clinical overlap observed across psychiatric disorders, i.e., comorbidity is the rule rather than the exception, as well as data showing that early, subsyndromal symptoms, such as mild depression and psychosis-like symptoms, increase risk for a range of illnesses. However, despite this convergent evidence for some shared neurobiological mechanisms across illness and risk categories, there have been few studies that have tested for common patterns of brain structure or function across at-risk individuals. Given this gap in the literature, we conducted a voxel-wise meta-analysis of functional magnetic resonance imaging (fMRI) studies of individuals who were considered at-risk for the following disorders: affective, anxiety, psychotic, and substance use.

    Methods: A systematic review of published task-based fMRI studies of subjects at risk for psychopathology was performed, using search terms such as “risk for psychopathology”, “longitudinal neuroimaging”, “at risk mental state”, “depression risk”, “behavioral inhibition”, and “inhibited temperament”. References in identified publications were also examined for additional studies to include. Included studies used a task during fMRI data collection, conducted whole brain analyses, and examined a “high-risk” group, as defined by clinical high risk (i.e., the presence of subsyndromal symptoms), temperamental high risk (i.e., behavioral inhibition), or familial high risk. Studies which did not list peak voxel coordinates in either Montreal Neurological Institute (MNI) or Talairach space were excluded. For each study, if a longitudinal analysis was performed, the coordinates for the comparison of the high-risk group who later developed symptoms versus high risk group without symptoms were included. Otherwise, the coordinates for the comparison of the high-risk group versus the control group were included. Thirty-one studies were identified as examining risk for psychopathology using functional MRI, totaling 1,236 participants. Of these studies, 13 examined risk for anxiety (513 subjects), six examined risk for depression (194 subjects), 10 examined risk for psychosis (331 subjects), and two examined risk for substance use (200 subjects). Study samples included adolescents and young adults (age range (years): 12-34; mean: 20). A wide range of tasks, i.e., cognitive and emotion paradigms, were employed. The meta-analysis was conducted using GingerALE software (version 2.3.6; This software uses an activation likelihood estimation (ALE) to test for overlap among foci across studies. A random-effects model was used. All data were transformed to MNI space using the tal2icbm function within the GingerALE software and peak coordinates were modeled with a 3D Gaussian kernel with a FWHM of 8.5–9.5 (i.e., a larger FWHM was used for studies with fewer subjects). Studies were combined by calculating the union of each individual study, with a voxel p-value of .01 and a cluster size of 25 voxels (200 mm3).

    Results: Compared to the controls, the transdiagnostic high-risk subjects showed lower activation across paradigms of the left caudate and left thalamus (both p < .01, k > 25), and showed increased activation of the right globus pallidus.

    Conclusions: The findings of this meta-analysis of 31 fMRI studies suggest that dysregulation of key nodes of the basal ganglia, the caudate nucleus, thalamus, and globus pallidus, is associated with signs of increased risk (i.e., subsyndromal symptoms or familial risk) for a number of serious mental illnesses. These results are consistent with prior evidence for abnormalities of the basal ganglia in affective, anxiety, psychotic and substance use disorders. Abnormalities in basal ganglia function may be related to alterations in hedonic processing, as hedonic function has been shown to be affected in each of these disorders. We can interpret these results as suggesting that one perturbation in basal ganglia function increases risk for many or all of these illnesses; alternatively, distinct (but overlapping) forms of basal ganglia dysfunction may increase risk for separate conditions or illness categories. These findings must be interpreted with caution, however, given the heterogeneity of the sample and analysis methods used, and the wide range of tasks employed in these studies. Replication of these results in an independent, prospectively collected sample is needed. Overall, this study indicates that such transdiagnostic approaches may produce new clues regarding the potential links between broadly shared genetic and environmental risk factors and the biological mechanisms of neuropsychiatric disease.

    Keywords: Early identification of risk, Functional MRI (fMRI), meta-analysis

    Disclosure: Nothing to disclose.

    M45. Adolescent Changes in Task fMRI Bold Signal Accounting for Longitudinal Reductions in Global Grey Matter Perfusion

    Marisa Silveri*, Huanjie LI, Julia Cohen-Gilbert, Jennifer Sneider, Derek Hamilton, Emily Oot, Anna Seraikas, Maya Rieselbach, Carolyn Caine, Sion Harris, Lisa Nickerson

    McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

    Background: Adolescence is a period of development characterized by rapid changes in brain structure, function, and neurochemistry, including changes in prefrontal cortex (PFC) and related neurocircuitry that are paralleled by improvements in frontally-mediated cognitive skills such as inhibitory control. Developmental changes in adolescence also have been observed in spatial memory skills, subserved by integration of prefrontal and hippocampal circuitry. Incongruous maturation rates across these brain systems can manifest in maladaptive behaviors, helping to explain why rates of risk-taking are elevated during adolescence. Longitudinal studies identifying developmental changes in functional activation during blood oxygen level dependent (BOLD) task fMRI typically do not account for developmental changes in cerebral blood flow (CBF), which through neurovascular coupling can lead to non-specific changes in BOLD signals. Thus, the objective of the current study was to examine the impact CBF has on developmental changes in BOLD signal acquired during inhibitory control on an emotional Go-NoGo task (eGNG) and during memory retrieval on a virtual Morris Water task (MWT), while controlling for developmental changes in CBF.

    Methods: Multimodal neuroimaging data were acquired using a Siemens 3 T TIM Trio, using multiband BOLD fMRI during the eGNG and virtual MWT, and using pseudo-continuous arterial spin labeling (PCASL) to quantify CBF at rest. Alcohol- and substance-naïve and psychiatrically healthy adolescents (n = 74) were recruited into a three-year longitudinal study, with task-related BOLD activation and CBF currently being examined in a subset of participants who completed baseline and one-year follow-up scans. Longitudinal changes in CBF averaged over gray matter were assessed via paired-t test (p < 0.05). Longitudinal changes in brain activation during tasks were first assessed via higher-level general linear modeling (GLM) for selected contrasts of interest from statistical modeling at the subject level via paired-t test with FLAME mixed effects analysis (cluster-based thresholding, z = 3.1, p < .05). To assess the impact of longitudinal changes in global gray matter CBF on BOLD signal changes between year one and year two, two higher-level models were run that assessed differences in activation first without and then with average grey matter CBF included as a regressor (non-parametric permutation testing, p < 0.05 corrected).

    Results: Across imaging visits, there was a significant decrease in global gray matter perfusion, with 13.7% lower CBF at one-year follow-up, p < .001. Significant BOLD activation differences were observed between baseline and one-year follow-up on the MWT in left PFC, left lateral occipital cortex and precuneus cortex (N = 17, retrieval > motor contrast), all nodes in a left-lateralized frontoparietal network implicated in explicit memory retrieval. BOLD activation changes were observed in the absence of performance differences, and results remained unchanged when controlling for CBF. For the emotional Go-NoGo task (N = 16, negative > neutral contrast), no activation differences were observed between baseline and one-year follow-up (with no effect of controlling for CBF), although faster reaction times at follow-up were evident without a change in error rates over time. Results also remained unchanged when controlling for CBF.

    Conclusions: Neurobiological interpretations of developmental changes in BOLD fMRI, or lack thereof, remain difficult to elucidate due to potential confounds associated with longitudinal changes in other factors that contribute to BOLD signals such as CBF. Indeed, longitudinal quantification of CBF is necessary to reconcile functional brain activation patterns, observed during rest or task performance, that change during development. To this end, in this preliminary investigation and using the current data processing procedures, the significant 13.7% decrease in CBF did not appear to influence longitudinal profiles of fMRI task BOLD signal on memory retrieval or emotional response inhibition. Additional analyses of these two distinct cognitive tasks, including assessing the contribution of regional CBF values, are underway to more deeply probe relationships between brain function, task performance and CBF. Taken together, dynamic functional changes suggestive of increased brain efficiency are clearly underway during this short one-year time span. Characterization of such developmental patterns are crucial for informing the search for biomarkers of risk for hazardous behaviors during adolescence, including initiation of alcohol and substance use and mood and anxiety symptoms that increasingly emerge during the second decade of life.

    Keywords: Adolescence, Functional MRI (fMRI), Perfusion, Emotional Response Inhibition, Spatial Memory

    Disclosure: Nothing to disclose.

    M46. Maternal Diet, Placenta Leptin Methylation, and Infant Growth

    Teresa Daniels, Alexander Sadovnikoff, Corina Lesseur, Carmen Marsit, Audrey Tyrka*

    Medical School of Brown University / Butler Hospital, Providence, Rhode Island, United States

    Background: High and low birthweight, as well as early gestational age, increase risk for a variety of medical and psychiatric conditions. Several factors influence birthweight and gestational age; preclinical and clinical studies show that maternal diet is an important factor. Leptin is a hormone that regulates satiety, body weight, and reproductive and developmental functions, and may be involved in the association between maternal nutrition, maternal obesity and infant outcomes. DNA methylation of placenta genes, such as leptin, may occur in response to exposures and may program subsequent infant development. This study examined maternal diet, placenta leptin gene methylation, and neonatal growth in a sample of healthy neonates and their mothers.

    Methods: Healthy neonates and their mothers (N = 135) were recruited within 1-2 days following delivery at Women and Infants Hospital in Providence, RI.

    A structured interview was conducted to assess the weekly servings of various types of food consumed by the mothers. Maternal weight, pregnancy weight gain, infant weight, and maternal health, medications, and vitamin use were obtained from the medical record. The leptin single nucleotide polymorphism (SNP) rs2167270 was genotyped, and bisulfate pyrosequencing was used to measure methylation of CpG sites in the promoter region of the leptin gene in placenta samples.

    Results: Genotype was a significant predictor of leptin methylation (p < .05), and after controlling for this and demographic and weight-related covariates, lower levels of leptin methylation were significantly associated with higher intake of sugared drinks (p < .05), white carbohydrates (p < .05), and whole grain carbohydrates (p < .05). Lower levels of leptin methylation were associated with smaller infant head circumference (p < .05), but not significantly associated with birthweight (p = .11) or gestational age at birth.

    Conclusions: These findings show the importance of intake of sugar and carbohydrate consumption for methylation of the placental leptin gene. Because methylation reduces gene transcription, lower methylation may indicate a placental response to high carbohydrate food that would result in higher levels of this satiety/metabolic hormone during fetal development. Lower levels of methylation (which are expected to lead to higher levels of the hormone) were also linked to smaller head circumference, suggesting the possibility that higher placenta leptin concentration may have important effects on growth and brain development.

    Keywords: Leptin, Developmental Psychopathology, Diet Induced Obesity

    Disclosure: Nothing to disclose.

    M47. Probing Supplementary Motor Area in Tic Disorders Using Repetitive Transcranial Magnetic Stimulation: A Preliminary Study

    Christine Conelea*, Brianna Wellen, Sarah Coleman, Lizzy Egbert, Kathryn Cullen, Benjamin Greenberg, Kelvin Lim

    University of Minnesota, Minneapolis, Minnesota, United States

    Background: Chronic tic disorders, including Tourette Syndrome, are neurodevelopmental conditions that impact 1-3% of children and are associated with multiple child-onset psychiatric disorders. Tics are rapid, repetitive movements and vocalizations frequently cued by aversive somatosensory experiences called “premonitory urges.” Tic suppression is the primary target of treatment, but between 30-70% of those with tics do not benefit from available interventions, which include medication and behavioral therapy. The mechanisms underlying tic suppression remain poorly understood but likely involve complex interactions between the brain and one’s immediate environmental context (e.g., tic-specific consequences, concurrent behavior, emotional state). Supplementary motor area (SMA) is a cortical region that plays a key role in facilitating context-dependent motor output, is hyperactive in those with tics, and has begun to be explored as a neural target for treatment using repetitive transcranial magnetic stimulation (rTMS). The current study aimed to probe the role of SMA in voluntary tic suppression using a novel methodology integrating rTMS with the Tic Suppression Task (TST), an established direct-observation behavioral paradigm capable of assessing context-specific tic fluctuations.

    Methods: Participants in this preliminary study are youth ages 12-18 years with chronic tics (N = 9 to date; n = 6 males) who completed a 1) clinical assessment, 2) brain MRI to functionally localize SMA, and 3) single session of 1 Hz or sham rTMS over SMA paired with pre-post administration of the TST. Contexts assessed in the TST included 3 min conditions of free to tic, instructions to suppress tics, and instructions to suppress tics paired with contingent reinforcement for tic-free periods. Conditions were randomly repeated twice each using a multi-element withdrawal design, and dependent variables included tic frequency (tics per min) and participant self-rated premonitory urge intensity (on a scale of 0-8). Tic frequency was determined by a video coder blind to stimulation and TST conditions, and all videos were double coded for inter-rater reliability purposes. Preliminary analyses utilizing single subject analysis visual inspection procedures and Cohen’s d effect size calculations (Olive & Smith, 2005) were used to identify data trends.

    Results: Results indicate that sham rTMS was generally associated with little change in tic suppression, while active 1 Hz rTMS participants showed a general enhancement in tic suppression. For example, tic frequency during suppression showed greater pre-post reduction for those receiving active (mean change = 3.56 tics per min, d = .81) vs. sham rTMS (mean change = 0.38 tics per min; d = .37). Further, premonitory urge intensity during suppression showed greater pre-post reduction in active rTMS (mean change = 2.88, d = 1.77) compared to sham (mean change = 0.38, d = 0.19). Emerging patterns of heterogeneous response were also identified within the active rTMS group. Youth who were already able to suppress (i.e., near-zero tic frequency in TST suppression contexts pre-rTMS) had no noticeable change. Most notably, those who had difficulty suppressing before rTMS showed greater tic suppressibility following rTMS. Post-rTMS, several youth showed equal tic frequencies across suppression contexts, suggesting enhanced suppression without the need for a supporting reward contingency.

    Conclusions: Overall, these preliminary data support methodology feasibility, sensitivity to detect clinically meaningful change across different contexts following acute rTMS administration, and ability to detect patterns of individual response. The results support the potential promise of this method for youth with the least ability to suppress tics, the group in greatest need of new clinical interventions.

    Keywords: Tourette Syndrome, Tic Disorders, Repetitive Transcranial Magnetic Stimulation (rTMS), Neuromodulation, Adolescent

    Disclosure: Nothing to disclose.

    M48. Biotypes Defined by Topological Data Analysis Distinguish Adolescent Suicidality More Precisely Than Depression Diagnosis

    Jessica Nielson*, Kathryn Cullen

    University of Minnesota Medical School, Minneapolis, Minnesota, United States

    Background: Depression is a complex disorder that commonly emerges during adolescence. Traditional approaches investigating the biology of depression using diagnostic criteria have been hampered by inconclusive results, potentially due to significant biological heterogeneity associated with this syndrome. Novel strategies are needed to parse the heterogeneity and identify biologically meaningful subgroups in order to tailor treatment around that biology. Machine learning approaches are taking center stage in the age of computational psychiatry as promising tools to aid in our understanding of complex neuropsychiatric disorders. The present study leverages existing data from a completed, multi-modal study of adolescents with and without major depressive disorder (MDD) and applies a novel computational strategy with the goal of generating new hypotheses about precision treatment targets for adolescents to aid in future clinical trials and treatment decision support for adolescent MDD. We hypothesized that this approach would reveal a biologically-based stratification of adolescents that would be capable of guiding the design of personalized interventions.

    Methods: Data from 122 adolescent males and females aged 12-20 years (82 patients with MDD and 40 age-matched healthy controls) were examined. Data included basic demographics and diagnostic information, depression severity, multiple measures of other psychopathology, cognitive performance for decision making and attention, heart rate variability, assessment of stress through cortisol values, and neuroimaging data. The neuroimaging data included structural information (cortical thickness and brain volume derived from FreeSurfer processing of high-resolution T1 images), functional MRI results (amygdala responses to negative stimuli) and resting-state functional connectivity metrics within key networks of interest. Data were analyzed using an emerging big-data tool known as topological data analysis (TDA) that combines methods of algebraic topology and geometry to derive a multidimensional network of data that can be used in combination with standard machine learning approaches to cluster patients based on health metrics. Although the original study was designed to examine biological differences between adolescents with MDD versus controls, based on the emerging understanding that MDD diagnosis may not be the most accurate strategy for parsing biological data, we did not include MDD diagnosis or symptoms to define the network in the TDA analysis. Rather, the network was defined by brain structure and function, autonomic function, cortisol responses, cognitive performance, and other (non-MDD) psychopathology measures. Patients were clustered using a Norm Correlation metric combined with Multidimensional Scaling (MDS) set at a resolution of 30 and a gain of 3.0 to render a network topology with approximately 2-3 patients per node to maximize the distribution of heterogeneity across the resulting network. To better understand the clusters of patients that were identified and to generate new hypotheses for treatment targets, we examined how biological and behavioral variables differed between clusters of patients using the Kolmogorov-Smirnoff (KS) test.

    Results: Several clusters emerged from the TDA, with a large cross-shaped network topology that clustered the majority of the adolescents into 108 connected nodes (N = 97). Smaller separate clusters emerged, one with 9 nodes (N = 11), one with 3 nodes (N = 3), and multiple individuals that were not connected to any other groups (N = 11). Subsequent analyses focused on the large network that contained 97 adolescents. Within this larger network, 4 subgroups were identified that showed distinct differences between domains. Between the top and the bottom clusters of the larger network, adolescents differed significantly in cortical gray (KS = .97, p < .0001) and white matter volume (KS = .88, p < .0001) and suicidality (KS = .81, p < .0001), with less of a significant difference in total MDD symptoms (KS = .58, p = .009). Between the left and right clusters of the network, adolescents were significantly different on measures of cognitive function related to decision making (KS = .94, p < .0001) and attention (KS = .94, p < .0001), and heart rate variability (KS = .79, p = .001). In contrast, the diagnoses of the adolescents were distributed throughout the network, suggesting that while these clusters were driven by important clinical and biological information, they were not driven by diagnosis.

    Conclusions: Here we provide proof of concept for a novel machine learning tool to cluster adolescents based on cognitive function, psychopathology, neuroimaging, and heart rate variability, as opposed to standard DSM diagnoses for MDD. Our findings suggest relationships between suicidality and brain structure may be independent of relationships between cognitive function and heart rate variability in adolescents. Further, biologically-driven clusters separated adolescents with suicidality from those without, even though suicidal symptoms weren’t included in the algorithm. With this approach, we are beginning to hone in on precise hypotheses about subgroups of adolescents characterized by features such as suicidality, impaired heart rate variability, abnormal brain structure, and cognitive impairment that may guide development of future biologically-motivated treatments that are personalized to the adolescents’ clinical and biological profile.

    Keywords: Machine Learning Clustering, Adolescent Depression, Suicidality, Computational Psychiatry

    Disclosure: Nothing to disclose.

    M49. Prenatal Fatty Acid Levels Longitudinally Predict Childhood Processing Speed in a Large Population Sample

    Katherine Karlsgodt*, Tanja Vrijkotte, Arne Popma, Bart Peters

    University of California, Los Angeles, Los Angeles, California, United States

    Background: The pre and perinatal environment, including availability of critical nutrients, can have a profound impact on offspring across development. In particular, fatty acids have been associated with a range of developmental outcomes including cognitive variability and neuropsychiatric symptoms. An early deficiency in long-chain polyunsaturated fatty acids (LC-PUFAs), which play an essential role in early neurodevelopment, may have long term impacts on cognitive function. Here, we investigated the association of prenatal LC-PUFA biostatus, specifically, the ratio of omega-6 to omega-3 fatty acids, to childhood cognitive function at age 5-6 years. We specifically examined processing speed, as it may be a core cognitive process that can have substantial effects on other downstream higher-level cognitive functions and it is impaired in schizophrenia.

    Methods: Our participants were children from the Amsterdam Born Children and their Development Study (Amsterdam-ABCD) birth cohort. We selected participants who had both prenatal fatty acid data along with childhood cognitive function data (n = 1793) as measured with the Amsterdam Neuropsychological Tasks (ANT) battery, specifically the Baseline Speed (BS) task. Regression analyses were employed, predicting processing speed and response stability with omega-6 and omega-3 fatty acid levels, adjusted for potential confounders.

    Results: We found that both the prenatal omega-6/omega-3 ratio as well as omega-6 and omega-3 fatty acid levels separately significantly predicted performance on the cognitive measures. Specifically, we found significant relationships between response speed and omega-6/omega-3 ratio (p = .002), such that a higher omega-6/omega-3 ratio predicted higher (slower) processing speed. When we looked at the individual components of this ratio, we found that total omega-3 levels showed a significant prediction of processing speed, such that a higher percentage of omega-3 predicted lower (faster) processing speed (p = .005). Further, total omega-6 percentage significantly predicted processing speed such that higher levels of omega-6 were associated with slower processing speed (p < .0001). We also looked at individual PUFAs, finding significant relationships between the omega-3 fatty acids EPA (p = .008), ALA (p = .004), and DHA (.018), as well as the omega-6 fatty acid AA (p = .027) but not LA.

    Conclusions: These findings indicate that prenatal fatty acid levels may be one potential factor that can contribute to later variability in cognitive function. Furthermore, fatty acids are of interest, as they have been linked to a number of neuropsychiatric disorders, including schizophrenia. Taken together with findings of early childhood cognitive deficits in schizophrenia, this indicates that early exposure to fatty acids may be an important developmental factor to consider.

    Keywords: Cognition, Fatty Acids, Prenatal, Brain Development, Nutrition

    Disclosure: Nothing to disclose.

    M50. Sub-Clinical Brain Failure: Identification of High Risk Mortality Patients and Prediction of Patient Outcomes Associated With Delirium With a Novel Bispectral EEG Device

    Gen Shinozaki*, Nicholas Bormann, Aubrey Chan, Kasra Zarei, Nicholas Sparr, Mason Kilsares, Lindsey Gaul, Gabrielle Duncan, Sydney Jellison, Kumi Yuki, Sayeh Sabbagh, Matthew Karam, Nicolas Noiseux, Eri Shinozaki, Sangil Lee, John Cromwell

    University of Iowa, Iowa City, Iowa, United States

    Background: Delirium is common and dangerous yet is underdiagnosed and undertreated due to lack of effective screening methods. Undetected delirium in hospitalized elderly patients substantially increases mortality, length of stay (LOS), and post-discharge institutionalization rates. In addition to delirium being a very dangerous condition for patients, the cost associated with delirium are immense. Annual financial losses due to delirium are estimated to be over $150 billion in the U.S. alone. To address this issue, we aimed to develop an innovative, noninvasive, user-friendly, objective EEG device with bispectral electroencephalography (BSEEG) for mass screening of delirium. Also, we aimed to test if such objective EEG device can predict patient outcomes related to delirium, such as hospital LOS, discharge disposition, and mortality.

    Methods: Study subjects were recruited from the University of Iowa Hospitals and Clinics upon admission and assessed for the presence of delirium with the Confusion Assessment Method for ICU (CAM-ICU), and Delirium Rating Scale (DRS). For enrolled subjects, we obtained 10 minutes of EEG recording with two channels from a handheld EEG device twice a day during their hospital stay. We analyzed those raw EEG data by converting it to digital signals with spectral density analysis and an additional algorithm to differentiate the two groups with and without delirium; we also compared their outcomes including hospital LOS, discharge to nursing home, and mortality.

    Results: From January 30, 2016, to October 30, 2017, 820 patients were approached and a total of 428 patients were enrolled in the study. Three hundred thirty-seven out of 428 patients were 55 yo or older, and 274 out of 337 had BSEEG scores available for analysis. In some cases, BSEEG scores could not be calculated due to poor signal quality. In the group 55 yo or older, 37.2% of patients were categorized as delirious. The study population was also independently divided into two groups—BSEEG-positive, indicative of more low-frequency components in their brain waves; and BSEEG-negative, indicative of less low-frequency components in their brain waves—based on a threshold to differentiate patient outcomes as described in the following section.

    Data from 274 subjects was analyzed to establish an association between BSEEG score and delirium status. Logistic regression showed significant association between delirium category and BSEEG score (P = 6.39 × 10–6, unadjusted; P = 1.22 × 10–5, adjusted for age, gender, and CCI).

    BSEEG Score and Patient Outcomes: To test the usefulness of the BSEEG score to predict patient outcomes, we used outcome data available from 274 subjects who were 55 yo or older to investigate the association of BSEEG scores obtained at the time of enrollment with patient outcomes commonly affected by delirium. Specifically, we assessed hospital LOS, discharge disposition, and mortality. First, LOS and BSEEG scores were significantly correlated (P = 0.00099, unadjusted; P = 0.0014, adjusted for age, gender and CCI). This indicates that a higher BSEEG score coincides with an increase in a patient’s LOS.

    Second, we compared the discharge outcome and BSEEG score. When BSEEG was compared between those who were discharged to their home and those discharged not to home including death during hospitalization, a higher BSEEG score was significantly associated with discharge not to home (P = 0.0038, unadjusted; P = 0.0090, adjusted for age, gender, and CCI). Third, when we analyzed the hazard ration (HR) for mortality controlling for age, gender, and CCI, the HR based on 1 SD change of BSEEG score was 1.44 (1.12 to 1.84, P = 0.004). Even after controlling for clinical delirium status in addition to age, gender, and CCI, the HR based on BSEEG score remained significant at 1.35 (95% confidence interval = 1.04 to 1.76, P = 0.025). Further, we assessed if there was also a correlation between groups based on BSEEG scores and all-cause mortality at the end of our study period in patients in our dataset, because association between delirium and mortality has been shown previously in the literature. We first assessed overall survival rates among our study participants, which showed differences between those with and without delirium (P = 0.0038). Second, we tested a group difference based on a BSEEG cut-off score and confirmed that the BSEEG-positive group showed worse survival compared to the BSEEG-negative group (P = 0.0032).

    We further analyzed the survival data to detect differences among delirious cases and non-delirious controls as a function of group differences in BSEEG scores. Delirious cases with a positive BSEEG score showed the highest mortality. In contrast, those patients categorized as delirious but with a negative BSEEG score had lower mortality, similar to that of non-delirious controls with a negative BSEEG score. Moreover, those thought to be non-delirious controls based on results of clinical assessment but with a positive BSEEG score had a higher mortality, even compared to those patients with clinical delirium but with a negative BSEEG score.

    Conclusions: The BSEEG algorithm was used to differentiate a patient population with prolonged hospital LOS, discharge not to home, and also poor survival. This result indicates that among those who we cannot clinically identify as delirious, there are certain populations showing EEG changes who are at high risk of poor survival. If we can identify this population appropriately with our BSEEG method and intervene properly, those lives can be saved.

    Keywords: Mortality, Delirium, Electroencephalography

    Disclosure: Predelix Medical LLC, Stock / Equity

    M51. Signaling Pathway Screening to Identify Neuroinflammatory Mediators of Ion Channel Dysfunction

    Paul Wadsworth*, Oluwarotimi Folorunso, Nghi Nguyen, Aditya Singh, David Brunell, Clifford Stephan, John Allen, Fernanda Laezza

    University of Texas Medical Branch of Galveston, Galveston, Texas, United States

    Background: Neuronal firing is a highly regulated process that depends on the stability of ion channel macromolecular complexes, which in turn rely on the integrity of cellular signaling networks. In neuropsychiatric disease states, the cellular milieu of neurons is skewed toward pro-inflammatory mediators that may have drastic physiological consequences on neuronal firing, especially in subpopulations of neurons vulnerable to disease such striatal medium spiny neurons (MSNs). In these cells, intrinsic firing depends upon the voltage-gated Na + (Nav) channel macromolecular complex. This complex is made up of multiple Nav channel regulators including fibroblast growth factor 14 (FGF14), which we have previously demonstrated affects the biophysical properties and cellular targeting of the Nav channel. However, the specific upstream signaling events that control this complex have not been well characterized. Here, we posited that the cytokine tumor necrosis factor-α (TNF-α), which has been previously shown to modulate Nav1.6 firing, could play a role in regulating the FGF14 interaction with the regulatory C-tail of Nav1.6 through Ser/Thr and/or Tyr kinase phosphorylation events downstream of the TNF receptor. This pathway might provide a mechanistic basis of selective neuronal vulnerability in MSNs and open the door for rapid screening and discovery of signaling pathways that may be implicated in reward-related disorders.

    Methods: LCA constructs were subcloned into vectors under the control of G418 and puromycin and linearized prior to insertion in HEK293 cells for the development of a double stable cell line. We employed LCA for the HTS of six compound libraries available in 384-well formats: the Custom Clinical Collection, National Cancer Institute, Prestwick Chemical Library, Selleck Bioactive Collection, Sigma-Aldrich LOPAC Collection, and the UT Austin Combined Kinase Collection.

    Results: To characterize the specific cellular signaling mechanisms that regulate the Nav channel complex, we conducted an in-cell HTS against the FGF14:Nav1.6 complex with LCA. We developed a double-stable HEK293 cell line expressing both CLuc-FGF14 and CD4-Nav1.6-NLuc and validated this line using control treatments and comparison with transient transfection. We subsequently fine-tuned the assay for 384-well plates by optimizing multiple parameters including cell density, media volume, incubation times, and enhancer/inhibitor controls to achieve a Z-score of ≥ 0.5, signifying a viable in-cell assay. Using homogeneously suspended cells in 40 µL of serum-free media and TNF-α and MNS as the enhancing and inhibitory controls, respectively, a Z-score of 0.6-0.7 was obtained for each plate. Following assay development, we screened > 4,500 small molecules targeting known mediators of cellular signaling. Compound libraries were screened in duplicate, and an overall Z-score was calculated for each plate to ensure that assay rigor was maintained across the screening. Compound toxicity was assessed in parallel using the CellTiter-Blue Cell Viability Assay, and effects on luciferase alone were assessed via counter-screening against full-length luciferase reconstituted in HEK293 cells. Following removal of these potential false-positives, compounds were ranked by a combination of %maximal luminescence and individual Z-scores, which were calculated based on the mean and SD of its respective plate controls (0.3% DMSO). Preliminary hits were identified using cut-offs of Z ≤ -3.5 and ≥ 50% reduction in complex formation for inhibitors and Z ≥ 2 and ≥ 130% increase in complex formation for enhancers. Hits were then selected based on potency as assessed by an 8-point dose response. Rationally guided hit assessment revealed an over-representation of structurally diverse compounds against two known mediators of TNF-α signaling, JAK2 and Src tyrosine kinases, including TG101348 (IC50 = 12.3 µM), WP1066 (IC50 = 4.1 µM), Bosutinib (IC50 = 9 µM), and Quercetin (IC50 = 5.7 µM). Bioinformatic analysis revealed a probable JAK2 phosphorylation motif in FGF14 at Y158, which is a previously identified hot-spot at the FGF14:Nav1.6 channel interface. We are currently validating these findings using a variety of biochemical, biophysical and electrophysiological assays to evaluate the impact of TNF-α signaling on intrinsic firing of MSN.

    Conclusions: We have developed a rigorous assay capable of rapidly screening compounds against PPI in the cellular milieu, and this tool could be adapted to study other protein pairs implicated in neuropsychiatric disorders. Furthermore, our screening of kinase inhibitors has identified JAK2 as a key regulator of the Nav complex, which may contribute to vulnerability to depression and mood disorders in MSNs in states of neuroinflammation. Altogether, our results for the FGF14:Nav complex demonstrate that this assay is a powerful new tool capable of identifying potentially dysfunctional signaling pathways that could be targeted for therapeutic development in neuropsychiatric disorders.

    Keywords: Neuroinflammation, High-Throughput, Signaling networks, Ion Channels, Protein:Protein Interactions

    Disclosure: Nothing to disclose.

    M52. Blood-Brain Barrier Dysfunction and Autoantibodies Promote Amyloid Internalization in Early Alzheimer’s Disease Pathogenesis With Diagnostic and Therapeutic Implications

    Eric Goldwaser*, Nimish Acharya, Robert Nagele

    University of Maryland Medical Center / Sheppard-Pratt, Baltimore, Maryland, United States

    Background: Aging related neurocognitive disorders such as Alzheimer’s disease (AD) are poorly understood from a pathophysiologic standpoint. Moreover, treatment efforts have remained limited to symptomatic management, and proven minimally effective at delaying the progression of the disease when it is clinically diagnosed. Recently, antibody subsets, namely natural autoantibodies (aAb), have been shown to be effective blood-borne biomarkers able to accurately detect mild-moderate AD. Their mechanistic role in early AD, including blood-brain barrier (BBB) dysfunction and amyloid plaque development, as well as utility at diagnosing pre-symptomatic stages are of extreme importance in helping to define new therapeutic targets.

    Methods: To explore features of early AD pathophysiology, immunohistochemistry (IHC) slides were prepared from various human cortical sections, including entorhinal and hippocampal tissue, in AD, disease- and age-matched controls. For BBB breakdown and antibody leakage experiments, a traditional secondary antibody was foregone, and biotinylated anti-human IgG was used as a primary antibody, detecting endogenous immunoglobulin already present in the native tissue. To study the dynamic sequences observed on IHC, a cell culture model was adapted using SH-SY5Y cells following established protocol for their maintenance, growth, and differentiation. Immunofluorescence microscopy assessed the internalization and subcellular trafficking of Aβ42 and various surface proteins co-administered with human IgG aAb species. Cultures were treated with 100 nM Aβ42, reproducing physiologic concentrations, for up to 72 h. Other dishes were treated with Aβ42 and human serum obtained from young-aged non-demented control (YC), old-aged non-demented control (OC), and AD patient at 1:50 dilutions. Lastly, human protein microarrays were employed to assay these aAbs as diagnostic biomarkers for early-stage AD detection and disease staging, as well as identifying individual proteins that may be targets of amyloid burdening sequences. Sera from 50 early AD subjects (with confirmed low CSF Aβ42), along with 186 other samples from mild-moderate AD, Parkinson’s disease, multiple sclerosis, breast cancer, and healthy controls were compared, and candidate biomarkers and test utility were evaluated with Random Forest and receiver operating characteristic (ROC) curve analysis.

    Results: IHC experiments showed that blood-borne, brain-reactive aAbs bind selectively to pyramidal neurons and trigger pathological changes, including intraneuronal accumulation of A□42 and expansion of the lysosomal compartment. Concurrent findings included BBB breakdown and leakage of serum emanating radially in a cloud-like morphology. Neuronal surface protein α7nAChR, known Aβ42 high-affinity target, co-localized in consecutive IHC slides with CathepsinD, a lysosomal marker, as well as with IgG-positive intracytoplasmic granules throughout the main axonal process and dendritic tree.

    We then investigated aAb-mediated receptor endocytosis following BBB breakdown in a cell culture model better able to capture the neuronal dynamics at play. Co-localization of IgG, α7nAChR, and Aβ42 were temporally related to the early endosomal marker, Rab11, and at later time points to the lysosomal marker, LAMP-1. Lastly, results using monovalent F(ab) antibody fragments generated from purified IgG in AD patient serum suggest that endocytosis of Aβ42 is triggered by the cross-linking capacity of neuron-binding aAbs. When tracked over time, these coupled events showed aAb-mediated endocytosis and accumulation of exogenous surface protein, IgG, and Aβ42 within the lysosomal compartment, consistent with the AD brain IHC correlates.

    The cell culture experiments showed that this co-localization represented the retrograde transport of an endocytic process attempting to digest Aβ42. The extent of Aβ42 internalization needed to cause cellular damage occurred when coalesced Aβ42 granules were rendered indigestible, under conditions of co-administered purified IgG fractions from AD sera. By cleaving the purified aAbs into monovalent F(ab) fragments, basal internalization rates of Aβ42 were not able to induce the same degree of neuronal damage. However, when the fragments were treated with a secondary F(ab) antibody and cross-linking capabilities restored, internalization rates were rescued, and cellular death again ensued.

    Lastly, the microarray data generated a panel of the top 50 autoantibody biomarkers necessary to detect early AD and diagnosed with 100 percent overall accuracy (AUC = 1; sensitivity = 100%; specificity = 100%; PPV = 100%; NPV = 100%) from later stages of AD, Parkinson’s disease, multiple sclerosis, breast cancer, and age-, disease-, and gender-matched controls.

    Conclusions: These results suggest that BBB breakdown and subsequent leakage of Aβ42 and brain-reactive aAbs act together to drive amyloid plaque development in early AD. Furthermore, these aAbs can be identified in the sera and stratified against phenotypic states of disease, thus allowing for a novel diagnostic capable of differentiating and staging AD, even at the preclinical phases with perfect accuracy. Overall, the data submit aAbs and BBB compromise may be an important risk factor for the initiation and progression of AD as well as other neurodegenerative diseases.

    Keywords: Translational Biomarker Development, Preclinical Alzheimer’s Disease, Blood-Brain-Barrier, Antibody

    Disclosure: Nothing to disclose.

    M53. Depressive Symptoms, Cortical Amyloid, and Cognitive Decline in Older Adults

    Jennifer Gatchel*, Jennifer Rabin, Rachel Buckley, Joseph Locascio, Yakeel Quiroz, Patrizia Vannini, Rebecca Amariglio, Dorene Rentz, Nancy Donovan, Deborah Blacker, Keith Johnson, Reisa Sperling, Gad Marshall

    Harvard Medical School, Massachusetts General Hospital, McLean Hospital, Boston, Massachusetts, United States

    Background: Previous work has shown a cross-sectional association between depressive symptoms and in vivo cerebral tau (FTP positron emission tomography (PET) imaging) in cognitively normal older adults. While this and other evidence support a relationship between depressive symptoms and in vivo measures of Alzheimer’s Disease (AD) pathology, it remains unclear how depressive symptoms together with cortical amyloid (Pittsburgh Compound B (PiB) PET) relate to cognitive decline in older adults.

    Methods: To address this question, 279 cognitively normal older individuals (males and females) from the Harvard Aging Brain Study underwent annual assessment with the Geriatric Depression Scale (GDS) and the preclinical Alzheimer’s Disease Cognitive Composite (PACC), as well as baseline amyloid (Pittsburgh Compound B (PiB)) PET imaging; average follow up time was 4.4 years. A mixed model was run with dependent variable PACC, a random intercept and slope for each subject, and fixed predictors: baseline PACC, baseline GDS and amyloid, baseline GDS X amyloid interaction, time, the interaction of predictors with time, and covariates sex, baseline age, and education. To investigate longitudinal depressive symptoms, time-varying GDS was entered into the model in place of baseline GDS.

    Results: In the model with baseline GDS, greater age (p = 0.0001), greater amyloid (0.002), and lower baseline PACC (p < 0.0001) were associated with PACC decline, but baseline GDS did not predict PACC decline. In the time-varying GDS model, higher GDS was associated with worse PACC (p = 0.0003), but only at amyloid levels above 1.1.

    Conclusions: Greater depressive symptoms over time in the setting of elevated amyloid (PiB PET) are associated with worse cognition in older adults who may be at risk for AD. Together, findings support the potential prognostic utility of depressive symptoms together with AD pathology in identifying individuals at risk for cognitive decline.

    Keywords: Depression, Alzheimer’s Disease, Cognitive Decline, Amyloid, PET Imaging

    Disclosure: Nothing to disclose.

    M54. Multiplexed, Quantitative Proteomic Analysis of Cynomolgus Monkey Tissue as Well as Tau + and TDP43 + Behavioral Variant Frontotemporal Dementia Human Tissue to Enable Biomarker Discovery and Development

    Abstract not included.

    M55. Efficacy and Safety of Dasotraline in Adults With Binge-Eating Disorder: A Randomized, Double-Blind, Fixed-Dose Trial

    Robert Goldman, James Hudson, Susan McElroy, Carlos Grilo, Joyce Tsai, Ling Deng, Justine Kent, Antony Loebel*

    Sunovion Pharmaceuticals, Inc., Fort Lee, New Jersey, United States

    Background: Dasotraline is a potent inhibitor of human dopamine and norepinephrine transporters (uptake IC50 of 3 nM and 4 nM, respectively) with a PK profile characterized by slow absorption (tmax, 10-12 h), and a long elimination half-life (t½, 47-77 h) resulting in stable plasma concentrations over 24 h with once-daily dosing. In a previous placebo-controlled, flexible-dose trial, dasotraline (4-8 mg/d) demonstrated significant efficacy in the treatment of binge eating disorder (BED). The aim of this replication study was to evaluate efficacy and safety of fixed doses of dasotraline (4 mg/d and 6 mg/d) in the treatment of patients with moderate to severe BED.

    Methods: Patients meeting DSM-5 criteria for BED were randomized to 12 weeks of double-blind treatment with fixed doses of dasotraline (4 mg/d and 6 mg/d), or placebo. All patients randomized to dasotraline received 4 mg/d during the first 2 weeks, after which those assigned to the higher dose group received 6 mg/d. The primary efficacy endpoint was change in the number of binge-eating days per week at week 12. Secondary efficacy endpoints included change on the Binge Eating Clinical Global Impression of Severity (BE-CGI-S), the Yale-Brown Obsessive-Compulsive Scale Modified for Binge Eating (Y-BOCS-BE) and the proportion of patients who achieved 4-week cessation of binge eating episodes. Efficacy was assessed using a mixed model for repeated measures analysis estimating the least squares (LS) mean difference in change from baseline between groups on the intent-to-treat (ITT) population. To control the overall type I error rate, a pre-specified sequential testing procedure was used.

    Results: A total of 491 patients were randomized, of whom a total of 485 were in the ITT population (dasotraline 6 mg/d (N = 162), 4 mg/d (N = 161), or placebo (N = 162). Baseline characteristics for the combined patient sample: female, 84%; mean age, 37.6 years, mean number of binge eating days per week, 4.22; mean CGI-S score, 4.4; mean BMI, 34.5 kg/m2. At week 12, treatment with dasotraline was associated with significant reduction in number of binge eating days per week in the 6 mg/d group vs. placebo (LS mean -3.47 vs. -2.92; difference [SE], -0.55 [0.19], P < 0.01), and non-significant improvement in the 4 mg/d group vs. placebo (LS mean: -3.21; difference, -0.29 [0.19], P = 0.12). Outcomes on secondary measures and p-values (not adjusted for multiplicity) generally also favored dasotraline. Changes on the BE-CGI-S for the 6 mg/d and 4 mg/d groups vs. placebo were -2.27 vs. 1.77 (P < 0.01), and -2.13 vs. 1.77 (P < 0.05), respectively. On the YBOCS-BE scores for the 6 mg/d and 4 mg/d groups the changes were -15.2 vs. -11.8 (P < 0.01) and -14.1 vs. -11.8 (P < 0.05), respectively. The proportion of patients who achieved 4-week cessation of binge eating episodes was 34.0%, 33.5% and 30.2% for the dasotraline 6 mg/d (p = 0.64), dasotraline 4 mg/d (p = 0.80), and placebo groups, respectively. The most common adverse events (incidence ≥ 5% and ≥ 2-times placebo) on dasotraline 4 mg/d and 6 mg/d vs. placebo were insomnia (40.1% and 29.8% vs. 13.5%), dry mouth (26.5% and 21.1% vs. 6.7%), decreased appetite (16.0% and 9.3% vs. 6.7%), nausea (13.0% and 11.8% vs. 5.5%), anxiety (13.6% and 8.7% vs. 2.5%), decreased weight (7.4% and 8.7% vs. 1.2%), constipation (6.8% and 3.7% vs. 1.8%), and dizziness (5.6% and 5.0% vs. 1.8%). Discontinuation due to an adverse event occurred in 14.2% of patients on dasotraline 6 mg/d, 8.7% on dasotraline 4 mg/d, and 1.2% on placebo. Changes in systolic and diastolic blood pressure were minimal. Mean baseline to endpoint change in supine heart rate on dasotraline 6 mg/d and 4 mg/d vs. placebo was + 6.2 bpm and + 4.8 vs. + 0.2 bpm.

    Conclusions: In this double-blind, 12-week, placebo-controlled, fixed-dose study, treatment with dasotraline 6 mg/d (but not 4 mg/d) was associated with a significant reduction in the frequency of binge eating days per week. Both dasotraline 6 mg/d and 4 mg/d treatment groups were associated with reduced obsessional thoughts and compulsive behaviors and greater overall improvement in symptom severity. Dasotraline was safe and generally well-tolerated at both doses; most common adverse events were insomnia, dry mouth and headache.

    Keywords: Binge Eating Disorder, Dasotraline, Efficacy, Safety, Randomized Double-Blind

    Disclosure: Sunovion, Employee

    M56. Increased Functional Connectivity Between Ventral Attention and Default Mode Networks in Adolescents With Bulimia Nervosa

    Mirjana Domakonda*, He Xiaofu, Lee Seonjoo, Marilyn Cyr, Rachel Marsh

    Hartford Hospital / Institute of Living, Hartford, Connecticut, United States

    Background: Bulimia nervosa (BN) is characterized, in part, by excessive concern with and selective attention to body shape and weight (shape/weight). Such negative self-referential thoughts contribute to the maladaptive pattern of food restriction, binge eating, and purging that defines BN. Behavioral data suggest that BN pathology is associated with excessive attention to self- and shape/weight-related stimuli, but the ventral attention (VAN) and default mode (DMN) networks that support attentional and self-referential processes are understudied in BN. We assessed whether altered functional connectivity within and between the VAN and DMN contributes to such excessive concerns in adolescents with BN, early the course of the disorder.

    Methods: Resting-state functional magnetic resonance imaging scans (GE Signa 3 T) were acquired from 33 BN and 37 healthy comparison (HC) adolescents (aged 12 to 21 years), group-matched by age, race, and body mass index. Seed-based region-of-interest (ROI) analyses were performed via the CONN toolbox to examine group differences in functional connectivity within and between the VAN and DMN. ROIs for the VAN, including the ventrolateral prefrontal cortex (VLPFC), right posterior superior temporal gyrus (pSTG) and ventral supramarginal gyrus (rvSMG) were selected based on a prior study of this network. ROIs for the DMN were defined a priori in the CONN toolbox. Results were age-adjusted and thresholded at a false discovery rate (FDR) of p-FDR < .05. Spearman’s correlations were performed to explore associations of VAN-DMN connectivity with BN symptoms, body shape/weight concerns, and sustained attention on the Continuous Performance Test (CPT).

    Results: Compared to HC adolescents, those with BN showed significantly increased positive connectivity between rvSMG and all DMN regions (p-FDRs < .021), as well as between VLPFC and left lateral parietal cortex (p-FDR < .036). Positive mean VAN-DMN connectivity was also detected in BN compared to healthy adolescents (p ≤ .003). Within-network connectivity did not differ between groups. VAN-DMN connectivity was associated with the frequency of BN symptoms (ps < .028, uncorrected), and body shape/weight concerns in the BN group (ps < .047, uncorrected). No significant group-by-CPT interactions on VAN-DMN connectivity were detected.

    Conclusions: Increased positive VAN-DMN connectivity in adolescents with BN may reflect abnormal engagement of VAN-mediated attentional processes at rest, perhaps related to their excessive attention to self-referential thoughts about body shape/weight. Future studies should investigate whether altered VAN-DMN connectivity precedes the onset of BN or arises as a result of the disorder, in hopes that these neuroimaging findings could guide the development of circuit-based targeted interventions aimed at reducing excessive body shape/weight concerns that tend to perpetuate maladaptive BN behaviors.

    Keywords: Eating Disorders, Functional MRI (fMRI), Adolescence

    Disclosure: Nothing to disclose.

    M57. Investigating Neural Taste Discrimination in the Insula Across Adolescence and Adulthood

    Marisa DeGuzman*, Megan Shott, Guido Frank

    University of Colorado Anschutz Medical Campus, Aurora, Colorado, United States

    Background: Investigating how taste stimuli are perceived and encoded in the brain could improve our understanding of eating disorder neurobiology. For example, previous evidence from our lab suggests that adults with either anorexia nervosa or obesity exhibit reduced taste discrimination compared to adults who are healthy or recovered from anorexia nervosa. However, there remains a gap in our understanding of what role age plays in taste discrimination. Adolescence is an important period in both normal neurodevelopment and eating disorder onset, therefore, we compared neural taste discrimination between healthy adolescents and adults within the insula, the primary taste cortex.

    Methods: Study participants included a total of seventy-one healthy females with no medication use and no history of major medical or psychiatric illness. Thirty of these were in the adolescent group (mean age = 15.2 + /- 1.9 y, age range = 12-17 y) and forty-one in the adult group (mean age = 26.5 + /- 5.4 y, age range = 19-43 y). Participants rated 2 ml samples of a 1 M sucrose solution and a neutral artificial saliva solution for both pleasantness and sweetness on a 1-9 scale. During functional magnetic resonance imaging, participants received 1 M sucrose solution, neutral artificial saliva solution, or no solution (nothing). First-level contrast images were analyzed using general linear models and three contrasts of interest were computed for each subject: Sucrose-Neutral, Sucrose-Nothing, and Neutral-Nothing. Whole brain activation was compared between groups at FWE-corrected and 10 voxel threshold and parameter estimates were extracted from insula regions of interest for post-hoc group comparisons of each contrast. Multivariate Bayesian pattern analysis and cross-validation tested taste classification accuracy (i.e. neural taste discrimination) in bilateral insula. Rank transformation was applied to data that were determined non-normally distributed using the Shapiro-Wilk test.

    Results: No differences in subjective taste pleasantness ratings were found between the age groups. While the groups rated the sweetness of the sucrose solutions similarly, adults rated the neutral solution significantly sweeter (M = 1.2 + /- 0.6) than the adolescents (M = 1.0 + /- 0.0); t(69) = -2.08, p < 0.05. However, no group differences were found in insula classification accuracy. Further, no significant correlations were found between age and classification accuracy. The whole brain analysis (FWE-corrected, 0 voxel threshold) revealed greater overall bilateral insula and striatum activation in adolescents compared to adults across all three taste contrasts. Post-hoc analyses showed statistically significantly greater adolescent activation in bilateral insula activation for two of the taste contrasts: Sucrose-Nothing F(6, 62) = 2.99, p < 0.05; Wilk’s Λ = 0.775, partial eta squared = .225); Neutral-Nothing F(6, 62) = 5.02, p < 0.0005; Wilk’s Λ = 0.673, partial eta squared = .327). There was also a trend toward significantly greater activation in adolescents for the Sucrose-Neutral contrast F(6, 62) = 2.22, p < 0.053; Wilk’s Λ = 0.823, partial eta squared = .177).

    Conclusions: Results of our neural taste discrimination analysis suggest that healthy adolescents and adults discriminate the sucrose and neutral taste stimuli similarly. However, the group difference in subjective sweet taste ratings of the neutral solution, point toward potential age effects on higher order flavor perception, but this requires further investigation. The greater insula BOLD response found in the adolescents is a novel finding that suggests adolescence is associated with hyper-responsiveness to both sweet and neutral taste stimuli. Perhaps the insula desensitizes from adolescence to adulthood or it “learns” with age to discriminate the tastes, resulting in less activation later in life. While the insula contains the primary taste cortex, it is also central for interoceptive awareness and is a part of the brain’s reward circuitry. Interestingly, the adolescents here also showed greater activation in the striatum in response to the taste stimuli, supporting evidence of higher reward circuit activation in adolescence. This apparent discrepancy, where taste discrimination is similar, yet overall insula BOLD response differs, raises further questions about how sensory integration may interact with the greater sensitivity that has previously been shown in adolescent reward circuitry. Future studies should aim to understand how these neural responses relate to risk for eating disorder development in healthy individuals and to current eating behaviors in individuals with eating disorders. In addition, neural taste discrimination should later be explored in adolescents with anorexia nervosa and other eating disorders.

    Keywords: Insula, fMRI, Taste

    Disclosure: Nothing to disclose.

    M58. Subcortical Shape Deformations in Bulimia Nervosa

    Laura Berner*, Zhishun Wang, Mihaela Stefan, Seonjoo Lee, Zhiyong Huo, Marilyn Cyr, Rachel Marsh

    University of California, San Diego, San Diego, California, United States

    Background: Bulimia nervosa (BN) is associated with significant medical complications and psychosocial impairment, and more than half of individuals with BN do not respond to first-line treatments. Improved understanding of the neurobiological mechanisms that may promote and maintain symptoms is critical to the development of novel, targeted interventions. FMRI data suggest BN is associated with functional abnormalities in frontostriatal and frontolimbic circuits, but subcortical alterations remain relatively uninvestigated. Moreover, findings from existing studies have been mixed, likely because of small sample sizes and the inherent inability of voxel-based morphometry to differentiate effects of shape, location, and size. The current study used vertex-wise shape analyses to overcome these limitations and estimate precisely localized deformations on the surface of subcortical structures in a large sample of adolescent and adult females with BN compared with group-matched healthy controls.

    Methods: High-resolution anatomical MRI scans were acquired from 62 females with full and subthreshold BN (18.73 ± 3.98 years) and 65 matched healthy controls (19.31 ± 5.71 years). A Bayesian model-based segmentation toolbox in FSL (FIRST) was used to segment each anatomical image and create vertex meshes for 15 subcortical structures (brain stem and bilateral nucleus accumbens, putamen, caudate, palladium, thalamus, hippocampus, and amygdala). Vertex indices were calculated based on the signed perpendicular distance from the surface mesh of the corresponding structure in the MNI template. Positive indices represented outward deformations on the surface of a given structure and negative indices represented inward deformations or inversions. Additionally, we computed the volume of each subcortical structure by generating a mask file in 3D volume space for each structure and multiplying the number of voxels in the mask by voxel size (mm). General linear models compared groups and assessed the significance of group-by-age interactions on the shape and volume of subcortical structures. Exploratory analyses in the BN group examined associations of the shape of each subcortical structure with illness severity in the past 28 days and illness duration.

    Results: Subcortical volumes did not differ across groups, but vertex-wise analyses revealed inward shape deformations in the BN compared to HC group on the surface of anterior aspects of the lateral and medial right pallidum (pFWE = 0.038). Inward deformations on the right pallidum were associated with more frequent binge eating episodes (pFWE = 0.013) and longer illness duration (pFWE = 0.005), whereas inversions on the surface of right caudate and left putamen were specifically associated with self-induced vomiting. Inward deformations on the thalamus (pFWE = 0.028) and amygdala (pFWE = 0.0002) were more pronounced with advancing age in the BN group.

    Conclusions: This is the first study to examine subcortical shape abnormalities in a large sample of adolescents and adults with BN. As has been observed in studies of other disease states, vertex-wise shape analyses were more sensitive than volumetric analyses, permitting detection of localized deformations on the surface of subcortical structures that comprise both reward and cognitive control circuits. Inversions detected on the ventromedial pallidal surface may contribute to altered reward sensitivity, excessive eating, and difficulties learning new goal-directed behaviors, whereas dorsal anterior lateral inversions may contribute to generalized and eating-specific disinhibition. Inward deformations on the surface of left thalamus were more pronounced in older individuals with BN but unassociated with illness duration, perhaps suggesting abnormal development of this structure. Inward deformations on the surface of right amygdala were associated with both age and illness duration, suggesting that these surface abnormalities may reflect an altered developmental trajectory, a consequence of prolonged bulimic symptoms, or both.

    Keywords: Bulimia Nervosa, Subcortical Shape Analysis, Binge Eating, Basal Ganglia, Mesolimbic Circuitry

    Disclosure: Nothing to disclose.

    M59. Examining Interoceptive Awareness as a Differential Predictor of Eating Disorder Treatment Outcomes Across Diagnoses in a Partial Hospital Setting

    Tiffany Brown*, Angeline Krueger, Erin Reilly, Christina Wierenga, Walter Kaye

    UCSD, San Diego, California, United States

    Background: Interoceptive awareness (IA) has received attention in the literature as a factor that may be relevant to the onset and maintenance of eating disorders (EDs). Both self-report and behavioral data suggest that individuals with EDs demonstrate alterations in IA, and that IA may relate to worse outcomes. However, IA is a multi-faceted construct, and different facets of IA may be more relevant to certain subgroups. Research testing links between IA and ED symptoms are challenging to interpret, due to the fact that popular measures of IA assess awareness of both interoceptive states and emotions. The Multidimensional Assessment of Interoceptive Awareness (MAIA) provides a nuanced assessment of various factors contributing to IA. To date, no studies have examined which MAIA subscales predict ED outcomes longitudinally, and no studies have evaluated whether these associations vary across diagnostic categories. Thus, the present study examined whether MAIA subscales accounted for treatment response, and whether effects varied across diagnoses.

    Methods: Adult and adolescent patients with anorexia nervosa – restricting subtype (AN-R, n = 171), anorexia nervosa binge-purge subtype (AN-BP, n = 78), and bulimia nervosa (BN, n = 141), enrolled in a partial hospitalization program (PHP), completed the MAIA and the Eating Disorder Examination-Questionnaire (EDE-Q) at treatment admission and discharge. The average length of stay in treatment was M(SD) = 83.72(60.85) days. The MAIA is a validated self-report instrument that assesses eight facets of interoceptive body awareness: Noticing (the awareness of one’s body sensations); Not-Distracting (the tendency to not ignore sensations of pain/discomfort), Not-Worrying (the tendency to not react with distress/worry to sensations of pain/discomfort), Attention Regulation (the ability to sustain attention to body sensations), Emotional Awareness (the awareness of connection between body sensations and emotional states), Self-Regulation (the ability to regulate distress by attending to body sensations), Body Listening (listening to the body for insight), and Trusting (experiencing one’s body as safe and trustworthy). The EDE-Q is a well-validated self-report measure that assesses Global eating pathology over the last 28 days, along with four subscales: Restraint, Eating Concern, Shape Concern, and Weight Concern. Regression analyses examined whether changes in the subscales of the MAIA at admission predicted EDE-Q scores at discharge, controlling for EDE-Q scores at admission. All analyses were run separately in AN-R, AN-BP, and BN patients.

    Results: Regression analyses predicting EDE-Q Global and EDE-Q Restraint scores demonstrated that for AN-BP patients, higher Noticing (p-values < .05), lower Not Worrying (p-values < .04), and lower Emotional Awareness scores at admission (p-values < .02) predicted worse outcome at discharge, while no subscales of the MAIA predicted outcomes for AN-R or BN patients. Similarly, for EDE-Q Eating Concern, in AN-BP patients, higher Noticing (p = .04), lower Not Worrying (p < .001), lower Emotional Awareness scores (p = .02), and lower Self-Regulation (p = .03) predicted worse eating concerns at discharge. For BN patients, higher Body Listening (p = .046) at admission predicted higher eating concerns at discharge. No subscales predicted outcome in AN-R patients. Regarding EDE-Q Shape Concern, lower Not Worrying scores (p = .02) predicted higher shape concerns for AN-BP patients, while lower Trusting (p = .03) predicted higher shape concerns for BN patients. No subscales predicted outcome in AN-R patients. Similarly, for EDE-Q Weight Concern, lower Not Worrying (p = .001) and lower Emotional Awareness scores (p = .03) predicted higher shape concerns for AN-BP patients, while subscales predicted outcome in AN-R or BN patients.

    Conclusions: Findings provide support for the importance of IA in accounting for ED treatment response but suggest that the relevance of the construct to outcome may vary across diagnoses. In particular, IA as measured by the MAIA appears to have greater prognostic value for AN-BP patients as compared to AN-R or BN patients. Specifically, greater distress/worry in response to uncomfortable body sensations and lower awareness of the connection between body sensations and emotional states were the most potent predictors of poor eating disorder outcomes in AN-BP patients. These results align with previous research supporting the negative prognostic value of anxious and alexithymic symptoms in AN-BP samples and suggest that targeting these constructs may be particularly useful for improving outcome in this group. For BN patients, greater listening to the body for insight predicted higher eating concerns at discharge, which may reflect faulty interoceptive signaling in BN patients around sensations of hunger/fullness. The lack of prognostic validity for the MAIA in the AN-R sample is surprising and may reflect poor insight into interoceptive difficulties using self-report measures. Overall, findings support recent research demonstrating altered interoceptive processing in eating disorders and call for continued research examining how this construct plays a role in the etiology and maintenance of eating disorders.

    Keywords: Eating Disorders, Interoceptive Awareness, Clinical Predictors, Anorexia Nervosa, Interoception

    Disclosure: Nothing to disclose.

    M60. Sex-Dependent Neurobiological Features of Prenatal Immune Activation via TLR7

    Galen Missig*, James Robbins, Emery Mokler, Christopher McDougle, William Carlezon

    Harvard Medical School, Belmont, Massachusetts, United States

    Background: Epidemiological evidence indicates that immune activation during pregnancy via infection or autoimmune disease is a risk factor for neuropsychiatric illness. Work from our lab and others has demonstrated that prenatal immune can cause long-lasting behavioral and neurophysiological alterations in offspring. Previous prenatal immune activation protocols have primarily involved administration of agents to mimic infection that target subtypes of the toll-like receptor (TLR) family, a class of receptor proteins that regulate innate immune responses. As examples, TLR3 recognizes Poly I:C and TLR4 recognizees lipopolysaccharide (LPS). In this study we examined the role of TLR7 in prenatal immune activation, considering evidence that this receptor subtype is implicated in the etiology of autoimmune diseases.

    Methods: We administered subcutaneous injections of the selective TLR7 agonist imiquimod (IMQ, 5.0 mg/kg) or vehicle to timed-pregnant dams (C57BL/6 J mice) on embryonic days (E) 12, 14, and 16. Over the first 13 weeks of postnatal development, we assessed the offspring on a battery of behavioral tests that include ultrasonic vocalizations (USVs), open field, social approach, reciprocal social interaction, as well as on measures of circadian activity and temperature. In a parallel set of experiments, we collected whole brain sections for microglia histology and tissue punches of the dorsal striatum for RNA-sequencing. All data were analyzed using ANOVAs and post hoc tests.

    Results: Mice exposed to prenatal IMQ exhibit a behavioral phenotype characterized by decreases in anxiety-like behavior, a fragmentation of social behavior, and an alteration in USV production (p’s < 0.05). This phenotype is readily distinguishable from those seen following prenatal activation of TLR3 and/or TLR4. On many of these measures there are significant sex differences. Additionally, mice exposed to prenatal IMQ have normal baseline locomotor activity but are hyperactive in response to various types of stimuli including the presence of a social partner, circadian cues, or gonadal hormone fluctuations (p’s < 0.05). Prenatal IMQ exposure causes a decrease in microglia density and an increase in the number of microglia ramifications in numerous brain areas, with particularly strong effects in striatum (p’s < 0.05). RNA-sequencing of the dorsal striatum revealed that prenatal IMQ exposure induces differential expression of hundreds of genes (adjusted p’s < 0.05), especially those encoding synaptic components, cell adhesion molecules, and glial markers. However, there are dramatic sex differences, with virtually no overlap in differentially expressed genes between males and females.

    Conclusions: Prenatal immune activation with a TLR7 agonist produces a behavioral phenotype and changes in microglia that are distinct from previous models using other TLR ligands. Underlying this phenotype is a propensity for “conditional hyperactivity”, reflected by an exaggerated response to some types of internal and external stimuli. Further, genome-wide analysis of mRNA identified numerous molecular pathways affected by prenatal IMQ exposure but demonstrated profound sex differences in the directions and patterns of expression. Considered with the existing literature, our findings suggest that early immune system activation can promote various—and sometimes even opposite—developmental trajectories, depending on the type and/or pattern of TLRs activated.

    Keywords: Maternal Immune Activation, Neuroimmunology, Toll-Like Receptors (TLRs), Prenatal Exposure, Prenatal Infection

    Disclosure: Nothing to disclose.

    M61. Genome-Wide Expression Profiling of Suicide

    Giovanna Punzi*, Gianluca Ursini, Leonardo Collado-Torres, Joo Heon Shin, Roberto Catanesi, Andrew Jaffe, Thomas Hyde, Joel Kleinman, Daniel Weinberger

    Lieber Institute for Brain Development, Baltimore, Maryland, United States

    Background: Suicide is the 10th leading cause of death for all age group combined and is on the rise across the U.S., according to recent reports by the Centers for Disease Control and Prevention (CDC).

    As suicidal ideation is a limited predictor of outcome, recent research has turned to the detection of biological markers for the development of prevention approaches. This may be especially relevant for suicides by violent means, since features of impulsive aggression represent a better predictor of self-destructive acting-out, a strong behavioral endophenotype, as well as a valuable research target, being associated with higher prevalence and lethality. In this regard, in completed suicides, the biology underling the choice of a violent method should represents a more precise feature to target in order to detect genetic signatures for the behavior at large.

    Previous discovery findings and further replication suggest that differences in dorso-later prefrontal cortex (DLPFC) expression of a human-specific non-coding RNA (lincRNA) may influence emotional regulation, aggressive behavior and suicide by violent means.

    Methods: In the present study, RNA sequencing (RNA-seq) data from post mortem human brain (228 Caucasian patients; adults) were examined to validate, at genome-wide level of significance, association of the lincRNA specifically with suicide by violent means, and to detect further candidates potentially related to the same signal. The sample size had the appropriate power ( > 80%) to detect genes differentially expressed with a log2 fold change of 0.3. Attribution of suicidal method was determined blind to the post mortem RNA-seq data. Cause and manner of death and contributory causes or medical conditions related to death were obtained from medical examiner documents. Cases where manner of death was pending or not determined at the time of the curation, and suicidal samples with ambiguous, or indefinable means of suicide, in regard to the level of violence employed, were excluded. Among the remaining suicides, most deaths distinctly fell within the violent or non-violent category. When this was not obvious, an in-depth behavioral assessment was obtained using detailed narrative summaries based on all available sources of historical information, including interviews with next of kin. The differential expression analysis was conducted on all features, including genes, exons, junctions and expressed regions data, correcting for diagnosis, sex, age, and qSVs, a measure of RNA integrity. A gene-set enrichment analysis was also performed on the top-list differentially expressed features.

    Results: At PFDR-corr. ≤ 0.05, minimal signal (i.e. ~10 expressed regions) arose when comparing non-suicide with suicide (all kinds of method); differentially expressed features further decreased when looking only at suicide by non-violent means. However, comparison between non-suicides and suicides specifically by violent means produced a remarkably greater list of features (i.e. over 1400 expressed regions). These results confirm our previous findings and suggest the engagement of specific signaling. Finally, gene-set analysis testing a catalog of immune genes on the genes differentially expressed in suicide by violent means points to the microglia, and complement, as the most enriched features (p = 6.72e-05 and p = 2.67e-05, respectively).

    Conclusions: These results confirm that classifying suicide by method is key in revealing the underlying different biology.

    Keywords: Brain, Suicide, RNA-seq

    Disclosure: Nothing to disclose.

    M62. Resting State Connectivity of the VTA Predicts Trait-Based Impulsivity in Healthy Adults

    Pamela DeRosse*, Toshikazu Ikuta, Ashley Moyett, Anil Malhotra

    The Donald & Barbar Zucker School of Medicine at Hofstra/Northwell, Glen Oaks, New York, United States

    Background: The Ventral Tegmental Area (VTA), a group of neurons located close to the midline on the floor of the midbrain, is the origin of the dopaminergic cell bodies of the mesocorticolimbic dopamine system. Not surprisingly, the VTA has been widely implicated in disorders characterized by impulsive behavior including addiction. To date, studies examining the relationship between VTA activity and variations in impulsive behavior have primarily focused on state-dependent activation to reward using task-based fMRI. Few studies have examined whether task-free resting state activity of the VTA is associated with variation in trait-based measures of impulsivity. Thus, the present study examined whether connectivity of the VTA during rest was associated with variation in a trait-based measure of impulsivity.

    Methods: Healthy volunteers (N = 89; 55% female; Mage = 36.60 ± 14.53) were administered the Temperament and Character Inventory (TCI), which provides scores for four domains of temperament that are thought to be largely heritable including: novelty seeking (NS), harm avoidance (HA), reward dependence (RD) and persistence (P). Resting-state fMRI (rs-fMRI) data were collected on a GE Signa HDx 3 T scanner. The primary analyses examined the association between NS and whole brain connectivity of the VTA, defined by the Harvard Ascending Arousal Network Atlas (Edlow et al. 2012). To examine the specificity of our results to Novelty Seeking, the remaining 3 subscales from the TCI (HA, RD and P) were also examined.

    Results: Novelty Seeking (NS) showed significant inverse correlations with connectivity between the VTA and the 1) right inferior frontal gyrus (IFG) (MNI: + 14 + 22 + 80; 819 voxels), 2) right superior frontal gyrus (SFG) (MNI: + 48 + 22 + 12; 448 voxels), and 3) left superior parietal lobule (SPL) (MNI:-34 -42 + 64; 404 voxels). No significant correlations were found for any other TCI domains and VTA connectivity. Post hoc regression analyses examining the contributions of the 4 traits comprising the Novelty Seeking score, including Exploratory Excitability (NS1), Impulsiveness (NS2), Extravagance (NS3) and Disorderliness (NS4), revealed that the association between VTA connectivity and Novelty Seeking was driven entirely by Impulsiveness (NS2).

    Conclusions: The present results suggest that spontaneous fluctuations in BOLD signal that are synchronized between the VTA and frontal/parietal regions may be related to trait-based measures of impulsivity. These findings add to prior studies implicating task-dependent activation in the VTA to impulsivity and may provide further insight into the role of the VTA in disorders characterized by impulsive behavior.

    Keywords: Impulsivity, Ventral Tegmental Area (VTA), Healthy Individuals

    Disclosure: Nothing to disclose.

    M63. Dexmedetomidine (dex) and Agitation – Part 2: Randomized, Placebo-Controlled, Adaptive Design Study to Determine Optimal the Dose in Healthy Elderly Volunteers and Volunteers With Mild Probable Dementia of the Alzheimer’s Type

    Sheldon Preskorn*, Neil O’Neill, Martin Kankam, Bradley Vince, Frank Yocca

    University of Kansas School of Medicine, Wichita, Kansas, United States

    Background: Dexmedetomidine (DM) is a selective alpha-2 adrenergic agonist currently marketed for intravenous (IV) administration to sedate non-intubated patients prior to and/or during surgical and other procedures and to sedate intubated and mechanically ventilated patients during treatment in an intensive care setting. IV DM has also been used for the treatment of agitation in individuals with a range of disorders, most notably delirium. For these reasons, DM may be useful for the treatment of agitation in patients with various neuropsychiatric disorders including senile dementia of the Alzheimer’s Type (SDAT) and acute psychotic disorders including schizophrenia. For these uses, an oral formulation would offer advantages over the currently approved product which is limited to IV administration. The goal of this adaptive design study was to establish the optimal dose of DM capable of producing sedation as a surrogate for an anti-agitation effect without producing any clinically meaningful effects on blood pressure (BP)/heart rate (HR).

    Methods: This Phase 1 b study employed a randomized, placebo-controlled, double-blind, adaptive design. Sequential cohorts (n = 8, 6 on DM & 2 on placebo) were infused intravenously with DM. Subjects in the initial cohorts were initially healthy, elderly (55 – 75 years of age) volunteers. Subjects in later cohorts were elderly volunteers healthy except for having mild probable SDA. Both sexes were included in all cohorts. In the first cohort, the drug was administered IV over a dose range from 0.1 to 0.6 mcg/kg/hour. Each infusion rate was administered for 30 minutes. After 30 minutes, the dose was increased by an additional 0.1 mcg/kg/hour until the subject achieved a predetermined level of sedation, mild drowsiness as assessed by the Richmond Agitation-Sedation Scale (RASS) or a pre-specified reduction in BP and/or HR. The dosing rate was adjusted for subsequent cohorts based on the results of preceding cohorts. Continuous assessment was made of level of arousal, BP and HR. Plasma samples were collected before dosing and then every 15 minutes for the determination of DM concentration. Statistics were descriptive and correlational in terms of pharmacodynamic and pharmacokinetic relationships.

    Results: The results of cohorts 1 and 2 established 0.3 mcg/kg/hour as the optimal starting dose. DM produced a RASS score of -1 (drowsiness) in subjects at concentration which did not produce clinically meaningful effects on BP and/or HR. The effect was evident in 11/12 subjects (92%) on DM. A RASS – 1 was seen in 1 out of 4 Individuals (25%) on placebo. The effect occurred within 30 minutes of starting the dose which produced the desired effect. For subjects on DM, the decrease in arousal persisted for 1.5-2 h, a clinically relevant duration. Plasma drug concentration correlated with dosing rate and with drug effect both within and between subjects. Gender affected drug responsiveness with males requiring twice the dose of DM compared to females. Interindividual variability in drug concentration and effect was established and will be addressed by the development of dose titration strategies. The cohort with SDAT is being conducted at the time this poster is being submitted but should be completed in time for inclusion in the poster presented at the meeting.

    Conclusions: This study also demonstrated the efficiency of an adaptive design approach to early phase CNS drug development. Specific to DM, this study found:

    1. (1)

      Arousal sedation can be achieved without producing clinically meaningful effects on BP/HR.

    2. (2)

      Plasma concentrations of the drug were correlated with both onset and offset of effect within a subject.

    3. (3)

      There was interindividual variability is response with approximately 1/3 being sensitive, 1/3 being intermediate, and 1/3 being insensitive.

    4. (4)

      There was a gender difference in drug effect (i.e., females more sensitive)

    5. (5)

      With completion of the SDAT cohort, the study will determine whether there is in meaningful shift in the dose-response curve because of having this condition.

    Using this information, an oral formulation is being developed for further trials in individuals with clinically meaningful agitation.

    Keywords: Dexmedetomidine, Agitation, Locus Coerulues

    Disclosure: Bioxcel Therapeutics Inc, Consultant

    M64. Dynamic Coding of Reward and Impulsivity in Dopaminergic VTA Neurons

    Tzvia Pinkhasov*, Sarah Starosta, Adam Kepecs

    Stony Brook University School of Medicine, Stony Brook, New York, United States

    Background: Impulsivity is a debilitating trait that characterizes numerous psychiatric conditions including attention deficit/hyperactivity disorder, bipolar disorder, and substance abuse disorders. Imaging and pharmacological studies have implicated the dopamine (DA) system in impulsivity across species, a neuromodulator also central to reward processing, motivation and learning. Because impulsive actions often occur during the process of obtaining a reward, we sought to parse out the dopaminergic contribution to impulsive actions, distinct from reward expectancy. To address this question, we recorded dopaminergic activity in ventral tegmental area (VTA) of mice, a major dopaminergic center that projects strongly to the frontal cortex and striatum.

    Methods: We developed a new behavioral task to probe the relationship between reward expectation and impulsive action on a trial-to-trial basis and tease apart dopamine’s roles in reward expectancy, impulsivity and attentional lapses from the ability to withhold pre-potent actions. We designed a cued-reward lick withholding task for head-fixed mice in which subjects were trained to withhold from licking during a reward-predicting auditory cue (2 s) in order to earn water reward. We used three different auditory cues that predicted three different reward sizes (big, small, none). The cue was restarted as many times as the animal licked prematurely (before the 2 second cue period was over). Thus, the duration of the tone and licking patterns provided a quantitative, trial-by-trial measure of an animal’s impulsivity level. We employed fiber photometry to study the temporal profile of DA activity and record the activity of DA cells in the VTA during the cued-reward lick withholding task. We used a transgenic Cre-driver line to specifically target dopaminergic cells and have them express GCaMP6f, an activity dependent fluorophore. Fluorescence collected through an implanted optic fiber in the VTA was then used as a proxy for the average activity of the DA population. Data from 3 male and 2 female mice were collected in this study.

    Results: Our behavioral results show that mice acted more impulsively in anticipation of the larger reward and were least impulsive when the cue predicted no reward (F(4) = 25.9327, P < 0.01, η2 = 0.5583). Our photometry results demonstrate that VTA DA neurons were activated by both the reward and reward-predicting cues and their activity was proportional to reward size and value, consistent with previous observations. We found that DA activity during the reward cue was predictive of the impulsivity level (no, low, or high impulsivity) during each trial, independent of predicted reward size. Moreover, DA activity prior to the reward cue also predicted impulsivity, even before a mouse could predict the value of upcoming reward. Interestingly, we found that DA activity in the VTA during reward consumption was greater when the mouse was impulsive compared to when it was not impulsive, for the exact same reward amounts.

    Conclusions: Using a novel cued-reward lick withholding task we confirmed that DA encodes reward value, in line with the previous literature. We were also able to identify for the first time that VTA DA activity predicts impulsive vs. non-impulsive states independent of reward, suggesting a direct relationship between the activity of these neurons and the ability to suppress impulsive actions. Based on these results we will be able to determine DA’s causal role in impulsivity, using optogenetic manipulation of VTA DA neurons during distinct phases of the trial based on our photometry recordings. By building on the results from classical neuropharmacology studies, we were able to tease apart how DA predicts reward and impulsivity in distinct moments during lick withholding performance.

    Keywords: Dopamine, Impulsivity, Photometry, Reward-Based Decision-Making, Ventral Tegmental Area (VTA)

    Disclosure: Nothing to disclose.

    M65. Serotonin 1B Receptors Modulate Impulsive Action but Not Impulsive Choice via GABAergic Signaling

    Valerie Magalong, Maliha Sultana, Peter Balsam, Rene Hen, Kate Nautiyal*

    Dartmouth College, Hanover, New Hampshire, United States

    Background: Dysregulated impulsive behavior is found in many psychiatric disorders including substance use disorder, pathological gambling, and attention deficit hyperactivity disorder. Many studies have addressed the role of dopamine in the neural mechanisms underlying impulsivity, however, serotonin also modulates impulsive behavior. Adding to the complexity, impulsivity can be fractionated into multiple dimensions (e.g., action and choice), and is also associated with a number of related behavioral phenotypes such as reward sensitivity, sensation seeking, risky decision making, and habit formation. Our studies aim to investigate the neural circuits which modulate impulsive behavior, as well as the behavioral constructs that contribute to impulsivity.

    Methods: Our previous work shows that an absence of serotonin 1B (5-HT1B) receptors in mice results in increased impulsive behavior that is due to a lack of heteroreceptor (but not autoreceptor) expression during adulthood (rather than development). Recently, our work explored the mechanisms through which 5-HT1B receptor signaling influences the neural and behavioral mechanisms of impulsivity. Impulsive action, impulsive choice, reward sensitivity, habit formation and motivation were all measured with operant paradigms following whole brain and tissue specific knockout of 5-HT1B in male and female mice. Using miniature microscopes to image calcium activity of single cells in vivo during operant behavior, we also assessed the impact of the absence of 5-HT1B receptors on circuits that are known to influence impulsivity.

    Results: Tissue-specific knockouts revealed that an absence of 5-HT1B receptors on GABAergic neurons resulted in increased impulsive behavior. Behavioral screens revealed that an absence of 5-HT1B receptor expression results in increased impulsive action, and no change in impulsive choice. Additionally, the knockouts showed increased motivation for reward and increased reward sensitivity, but no effects on habit formation, compulsive behavior, or extinction. Furthermore, changes in impulsive action correlated with increases in motivation but not reward sensitivity or other behavioral measures. Current in vivo calcium imaging studies are aimed at investigating the mechanisms underlying 5-HT1B receptor-dependent inhibitory control. Our preliminary experiments have focused on imaging dopamine cell activity in a 5-HT1B-sensitive GABA-GABA-DA circuit, as well as 5-HT1B receptor positive dopamine-responsive cells.

    Conclusions: Overall, our results point to a role for serotonin modulation of impulsivity via influence on GABAergic signaling. Current work is aimed at identifying the neural circuit mechanisms through which 5-HT1B receptors mediate this behavioral regulation and identifying the extent to which 5-HT1B receptors influence related behavioral phenotypes. Importantly, these effects are dependent on adult expression suggesting a potential avenue for pharmacological treatment for psychiatric disorders which include dysregulated impulsive behavior.

    Keywords: Impulsivity, 5-ht1b, Neural Circuits, Reward Sensitivity, GABA

    Disclosure: Nothing to disclose.

    M66. Susceptibility to Stress-Induced Anhedonia is Associated With Neurotransmitter Plasticity in the Dorsal Raphe Nucleus

    Nandkishore Prakash, Davide Dulcis, Andre Der-Avakian*

    University of California, San Diego, La Jolla, California, United States

    Background: Anhedonia, or diminished interest or pleasure, is a core symptom of major depressive disorder and is common in other mood disorders. Stress is known to produce anhedonia in some individuals (susceptible), but not others (resilient). We hypothesized that neurotransmitter plasticity, an alteration of neurotransmitter expression in response to activity leading to changes in behavior, may play a role in susceptibility to stress.

    Methods: We used the intracranial self-stimulation (ICSS) procedure to measure anhedonia in rats. Rats were trained to respond to stimulation of the posterior lateral hypothalamus, part of the brain’s reward circuitry. The minimum current needed to elicit a response was defined as the reward threshold. Rats underwent 21 days of social defeat and their reward thresholds were measured daily. An increased threshold from baseline was interpreted as a sign of anhedonia. Their brains were subsequently processed for immunohistochemical detection of neurotransmitter markers. To examine effects of activity manipulation on neurotransmitter plasticity, Cre-dependent viral expression of excitatory DREADDs in a CRH-Cre transgenic rat strain was used to drive changes in neuronal activity.

    Results: Susceptible rats developed anhedonia in response to chronic social defeat while resilient rats did not. Susceptible rats showed a significant 2-fold increase in number of serotonergic (TPH2 + ) neurons in the ventral sub-nucleus of the dorsal raphe nucleus (DRv), compared to resilient rats or non-stressed controls. Other sub-nuclei of the dorsal raphe nucleus did not show significant differences in TPH2 + neuron number across groups. The total number of neurons in the DRv did not change across groups, indicating that a reserve pool of mature neurons in the DRv acquired serotonergic identity only in susceptible rats following chronic stress. Preliminary immunohistochemistry data suggest that the reserve pool may be a glutamatergic population in the DRv, known to co-express serotonin and project to reward regions of the brain. Preliminary results suggest that activating a GABA/CRH input to the dorsal raphe nucleus from the central amygdala reduces the number of serotonergic neurons in controls and ameliorates the effect of stress on the number of serotonergic neurons in the dorsal raphe nucleus.

    Conclusions: Our results suggest that neurotransmitter plasticity occurs in the dorsal raphe nucleus in response to chronic social stress, is bidirectional, and is subject to modulation by the central amygdala. Current experiments aim to manipulate inputs to the dorsal raphe nucleus to understand how circuit activity drives transmitter plasticity and behavior. Collectively, these data reveal a potentially novel mechanism underlying susceptibility and resilience to stress-induced anhedonia.

    Keywords: Social Defeat Stress, Anhedonia, Reward, Intracranial Self-Stimulation, Dorsal Raphe Serotonin Neurons

    Disclosure: Nothing to disclose.

    M67. Social Adversity During Adolescence Weakens Goal-Directed Decision Making in Adulthood

    Elizabeth Hinton, Shannon Gourley*

    Emory University School of Medicine, Atlanta, Georgia, United States

    Background: Depression is often precipitated by social adversity, particularly during adolescence, and is more prevalent in women. Identifying the durable consequences of social adversity experienced during adolescence is of great importance.

    Methods: Female mice were isolated or group-housed from postnatal day (P) 31­-60. At P60, all mice were re-housed into social groups to identify and correct the durable consequences of social poverty – specifically, those that present despite normalization of the social environment.

    Results: A history of isolation reduced 2’,3’-Cyclic-nucleotide 3’-phosphodiesterase (CNPase) expression, consistent with prior investigations, and validating our method. Isolation during adolescence also induced hypocortisolemia early in the dark cycle, when corticosterone (CORT) should otherwise be elevated. Given that CORT tone is a key regulator of dendritic spine turnover, we investigated dendritic spine densities on deep-layer ventromedial prefrontal cortical (vmPFC) excitatory neurons, revealing spine excess. Mice with a history of social isolation were simultaneously less adept at selecting actions based on their consequences, a vmPFC-dependent function, and they developed anhedonic-like behavior. Chemogenetic silencing of the vmPFC impaired the ability of typical mice to select actions based on their outcomes but corrected decision-making abnormalities and anhedonic-like behavior in mice with a history of social isolation.

    Conclusions: The social milieu sculpts prefrontal cortical development. Our findings reveal previously unappreciated neuroanatomical and functional consequences of social poverty experienced during adolescence. This information may inform strategies for improving mental health in humans.

    Keywords: Juvenile, Decision Making, Habit, Depression, Cortisol

    Disclosure: Nothing to disclose.

    M68. Acetylcholinergic Mechanisms of Depressive-Like Behaviors Induced by Seasonally Relevant Reductions in Active Photoperiod

    Zackary Cope*, Maria Lavadia, Aniek Joosen, Davide Dulcis, Marina Picciotto, Yann Mineur, Jared Young

    University of California, San Diego, La Jolla, California, United States

    Background: Environmental factors, such as seasonal variations in day length, can influence switching between mood states in bipolar disorder (BD) in a direction-specific manner. Consistent with sufferers of Seasonal Affective Disorder, BD patients exhibit depressive episodes as light availability decreases in Fall and Winter. Treatments seeking to maintain stability of photoperiod (timed light exposure, social rhythm therapy) can support a euthymic state (reviewed in Young and Dulcis, 2015). Pharmacological-induced increase of acetylcholine using physostigmine induces depressive-relevant behavior in mice, such as increased immobility of mice in the Forced Swim Test (FST, Mineur et al., 2013) and can induce depression in sufferers of BD (Janowsky et al. 1972). The present studies were designed to test the hypothesis that acetylcholine (ACh) signaling is necessary for the expression of depressive-like behaviors following exposure to short-active photoperiod (SAP).

    Methods: Expt. 1: C57BL/6 mice were pretreated 0.03 mg/kg with the AChesterase (AChE) inhibitor physostigmine (0.03 mg/kg) or saline 30 minutes before FST followed by the muscarinic ACh receptor (mAChR) antagonist scopolamine (SCOP, 0.03, 0.05 mg/kg) or the nicotinic ACh receptor (nAChR) antagonist mecamylamine (MEC, 0.56, 0.75 mg/kg). Expt. 2: Mice were housed in SAP (19 H light: 5 H dark) for 2 weeks before FST testing, on which they received 0.03 mg/kg SCOP or 0.56 mg/kg MEC. Expt. 3: Mice were administered low doses of SCOP and MEC to assess possible additive contributions of each ACh target. Expt. 4: AAV was used to deliver human (h)AChE to increase ACh metabolism specifically in the hippocampus to determine its sufficiency to block SAP effects.

    Results: Expt. 1: SCOP (F(2,49) = 5.3, p < 0.01) and MEC (F(2,52) = 4.6, p < 0.05) selectively decreased FST immobility in mice pretreated with the physostigmine. Immobility was not significantly decreased in mice pretreated with saline (F < 1, ns). Expt. 2: SAP significantly increased immobility in two separate cohorts (F(1,70) = 6.5, p < 0.05, F(1,143) = 4.3, p < 0.05). In cohort 1, treatment with 0.03 mg/kg SCOP reduced immobility irrespective of photoperiod condition (F(1,70) = 6.9, p < 0.05). As per our a priori planned hypotheses, it was discovered that this main effect was driven by significantly reduced immobility in the SAP only, not in the NAP mice. In cohort 2, 0.56 mg/kg MEC slightly reduced immobility irrespective of photoperiod condition, but not significantly (F(1,143) = 2.0, p = 0.15). Expt. 3: Two ineffective doses of antagonist (0.3 mg/kg MEC, 0.01 mg/kg SCOP), additively decreased FST immobility (F(3,142) = 3.0, p < 0.05), [although SAP did not significantly increase immobility in this cohort (F(1,142) < 1, p > 0.05)]. Expt. 4: Preliminary results indicate a significant virus*photoperiod interaction (F(1,43) = 16.98, p < 0.01) with hAChe decreasing immobility selectively in SAP mice. Full data and confirmatory immunohistochemistry for experiment 4 will be shown.

    Conclusions: These results indicate a necessity for increased hippocampal ACh neurotransmission in the expression of SAP-exposure-induced depressive-like behaviors. Further, although nAChR and mAChRs both influence depression-relevant behaviors, antagonism of mAChRs may be more effective in reducing SAP exposure-induced depression-relevant behaviors. In sum, these results provide evidence for AChergic mechanisms in driving seasonally depressed affective states in response to shortened day lengths.

    Keywords: Depression, Seasonal Affective Disorder, Bipolar Disorder, Hippocampus, Acetylcholine

    Disclosure: Nothing to disclose.

    M69. Microglia and Complement Activation in Chronic Stress-Induced Depressive-Like Behavior

    Talisha Davis, Babak Baban, Anilkumar Pillai*

    Medical College of Georgia, Augusta, Georgia, United States

    Background: Accumulating evidence suggests that the immune system plays an important role in the pathophysiology of depression. As a fundamental factor in the provocation of depression, chronic stress is associated with dysregulated immunity and subsequent inflammation. However, there is a critical need for studies that are designed to determine the role of specific components of the immune system in MDD in order to identify novel therapeutic targets. We recently reported a critical role of complement system in depression. In the present study, we examined the mechanisms involved in chronic stress-induced complement activation.

    Methods: We used real time PCR to analyse gene expression whereas immunoblotting and immunofluorescence analyses for protein levels of complement components. The C1q and C3 expression were determined in mice following chronic unpredictable stress (CUS).

    Results: C3 and C1qa (a key component of classical pathway) are highly expressed in the brain samples from mice exposed to CUS. CUS increased the expression of tumor necrosis factor α (TNFα, a key inflammatory molecule implicated in depression and known to induce complement activation) in microglia. Moreover, CUS-induced complement activation and depressive-like behavior were attenuated by TNF inhibition using TNFα KO mice or inhibition of microglia activation using minocycline.

    Conclusions: Our findings identify the role of microglia in chronic stress-induced complement activation and depressive-like behavior in mice.

    Keywords: Complement, Stress, Microglia

    Disclosure: Nothing to disclose.

    M70. The Involvement of SKA2, a Novel GR Interaction Partner, in Stress-Related Psychiatric Disorders

    Jakob Hartmann*, Nils Gassen, Harry Pantazopoulos, Claudia Klengel, Nina Dedic, Katelyn Li, Kathrin Hafner, Kenneth McCullough, Rhomberg Thomas, Torsten Klengel, Sabina Berretta, Kerry Ressler

    Harvard Medical School, Belmont, Massachusetts, United States

    Background: Mood and anxiety disorders represent a major disease and social burden worldwide, but the underlying molecular mechanisms are still poorly understood. In recent years, evidence has emerged for the crucial role of genes involved in the regulation of the hypothalamic-pituitary-adrenal (HPA) axis, especially in the context of stress-related psychopathologies such as anxiety and depression. The glucocorticoid receptor (GR) is the main mediator of the negative feedback loop of the HPA axis in response to stress. The Ska2 gene, encoding the spindle and kinetochore associated complex subunit 2, has previously been identified as GR interaction partner. Interestingly, single nucleotide polymorphisms and epigenetic status within the Ska2 gene, as well as gene expression alterations, have been associated with posttraumatic stress disorder and suicide risk in several studies in the past. Yet, little is known about the underlying molecular mechanisms and the role of Ska2 in the brain. Therefore, we set out to further investigate the role of Ska2 in the CNS and validate it as a potential candidate gene in the context of stress-associated psychiatric disorders.

    Methods: We conducted Immunohistochemistry (IHC) to study the expression pattern of Ska2 in human postmortem hippocampus and amygdala samples and in the mouse brain. Additional co-immunoprecipitation (co-IP) assays addressed whether SKA2 physically interacts with the GR in the mouse brain and in human postmortem brain samples. Moreover, we performed GR reporter gene activity (MMTV-Luc) assays and qPCR following Ska2 perturbation to investigate the receptors activity and downstream effects in neuronal cell culture. Western blot analyses were performed to examine SKA2 protein expression in basolateral amygdala (BLA) samples of individuals with bipolar depression and matched controls. Stress-induced changes in Ska2 mRNA expression in mice were investigated via qPCR following fear conditioning.

    Results: IHC analyses in human postmortem hippocampus and amygdala samples revealed a prominent expression of SKA2 in Vglut1-positive, glutamatergic neurons. Moreover, co-IPs revealed a physical interaction between SKA2 and GR in the mouse brain and human postmortem samples. In addition, Ska2 knockdown in mouse neuroblastoma (N2a) cells led to significantly reduced GR reporter gene assay activity, while Ska2 overexpression resulted in the opposite effect (2-Way ANOVA, interaction: F(14,42) = 16.38, p < 0.0001). Along these lines, Ska2 knockdown led to significantly altered mRNA expression of GR target genes Fkbp5 (t-test, T(21) = 2.683, p < 0.05) and Id3 (t-test,, T(22) = 5.762, p < 0.001). Interestingly, we detected significantly increased SKA2 protein levels in the BLA of individuals with bipolar depression compared to matched controls (ANCOVA, F = 5.83, p < 0.025, n = 14 per group). Detailed mapping and co-labeling studies in mice also revealed a distinct pattern of Ska2 expression in neurons of the BLA, as well as in the hippocampus (HC), medio-dorsal thalamus, the paraventricular nucleus of the hypothalamus and throughout the cortex. Most of the Ska2-positive neurons also expressed the GR. Consequently, we assessed whether acute stress is able to modulate Ska2 gene expression. Using qPCR, we found dynamic changes of Ska2 mRNA expression four hours after stress (fear conditioning, 5 tone/foot shock pairings). Stressed mice showed significantly decreased Ska2 mRNA levels in the HC (T(18) = 2.446, p < 0.05, n = 8 (ctrl), n = 12 (stress)) and increased levels in whole amygdala punches (T(15) = 2.693, p < 0.05, n = 5 (ctrl), n = 12 (stress)) compared to baseline controls (home cage group).

    Conclusions: Our findings reveal Ska2 as a novel GR interaction partner in brain regions associated with emotion processing and cognition. Further experiments showed that Ska2 gene expression is dynamically regulated by stress exposure and that SKA2 is able to positively modulate GR signaling and its downstream targets, providing a mechanistic link to its association with stress-related psychopathologies. Collectively, our data point to an important, and thus far unappreciated, role of Ska2 in stress-related psychiatric disorders, which is relevant to our understanding of the molecular mechanisms underlying such diseases.

    Keywords: Bipolar Disorder, PTSD, Suicide, HPA axis, Glucocorticoid Receptor

    Disclosure: Nothing to disclose.

    M71. Omega-3 Polyunsaturated Fatty Acid Biostatus Modulates Behavioral and Neurochemical Responses to Ketamine in Rats

    Ruth Asch, Diana Lindquist, Jennifer Schurdak, Robert McNamara*

    University of Cincinnati College of Medicine, Cincinnati, Ohio, United States

    Background: Although the pathophysiology of mood and psychotic disorders is associated with a dysregulation in prefrontal cortex (PFC) glutamate homeostasis, little is known about pathogenic mechanisms and associated risk factors. Mood and psychotic disorders are also associated with deficits in long-chain omega-3 fatty acids including docosahexaenoic acid (DHA), and we recently reported that glutamate and DHA levels are inversely correlated in the rat PFC using in vivo proton magnetic resonance spectroscopy (1 H MRS). Ketamine has acute antidepressant effects as well as psychotogenic effects, and has been shown to stimulate glutamate release and promote synaptic connectivity in the PFC. The present study investigated the effects of altered PFC DHA levels on dynamic glutamate synaptic responses elicited by ketamine using behavioral, gene expression, and 1 H MRS assays.

    Methods: From P21 to P90, male rats were maintained on a diet with no n-3 fatty acids (Deficient, DEF), a diet fortified with preformed DHA (fish oil, FO), or a control diet fortified with alpha-linolenic acid (CON). In the first cohort, locomotor activity in response to acute ketamine (30 mg/kg, n = 8/diet group) or vehicle (saline, 1 ml/kg, n = 8/diet group) was assessed in automated activity chambers. At 90 minutes post-injection, rats were sacrificed, and the PFC isolated for qRT-PCR analysis. Genes associated with glutamatergic synaptic activity (cFos, ARC), neurotransmission (NMDA, AMPA, mGluR), and plasticity (GAP-43, PKCgamma, PKCzeta) were investigated. In the second cohort, metabolite concentrations were acquired from the mPFC of isoflurane-anesthetized CON (n = 11), DEF (n = 10), and FO (n = 10) rats prior to and following a ketamine challenge (30 mg/kg) using 1 H MRS. Postmortem PFC DHA composition was determined by gas chromatography.

    Results: Compared with CON rats, PFC DHA levels were significantly lower in DEF rats (-30%, p < 0.0001) and significantly higher in FO rats ( + 13%, p = 0.003). There were no diet group differences in novelty- or saline-induced locomotor activity. Ketamine significantly increased locomotor activity and this response was significantly attenuated in FO rats compared with CON and DEF rats. Ketamine robustly increased cFos expression in the PFC, and this response did not differ significantly between diet groups. Ketamine did not significantly alter the expression of other genes of interest. In the 1 H MRS experiment, the ketamine challenge significantly decreased glutamate concentrations, and increased the glutamine/glutamate ratio, and these responses did not differ between diet groups. Ketamine significantly increased the GABA/glutamate ratio in DEF rats but not CON or FO rats.

    Conclusions: Ketamine-induced increases in PFC neuronal activation and reductions in glutamate concentrations are not significantly altered by PFC DHA levels. However, higher PFC DHA levels are associated with a diminished locomotor response to acute ketamine and are protective against ketamine-induced alterations in the inhibitory/excitatory balance in the PFC. Together these preclinical findings suggest that DHA biostatus modulates behavioral and neurochemical responses to acute ketamine and may therefore impact the psychotropic effects of ketamine in human subjects.

    Keywords: Ketamine, Omega-3 Fatty Acids, Glutamate

    Disclosure: Nothing to disclose.

    M72. Hippocampal Proteome Biosignatures of Major Depressive Disorder

    Maura Boldrini*, Huber Gonzalez, Shujuan Tao, Yan Liu, Gorazd B. Rosoklija, Victoria Arango, Andrew J. Dwork, Lewis Brown, J. John Mann

    Columbia University, New York, New York, United States

    Background: Major depressive disorder (MDD) will be the leading cause of global disease burden by 2030 (WHO, 2012), is associated with cognitive dysfunctions involving the hippocampus (HC): negative feedback and set shifting, attention control, memory and executive function (Keilp et al, 2012). The prefrontal cortex (PFC)-amygdala-HC circuit (Thierry et al, 2000) is involved in the pathogenesis of MDD and neuronal, glial and dendrite spine changes were detected in these regions (Licznerski and Duman, 2013). We found more smaller DG volume in unmedicated MDD (uMDD) compared with non-psychiatric controls (NC) and MDD treated with selective serotonin reuptake inhibitors (MDD*SSRI) (Boldrini et al, 2013). Smaller HC volume in MDD (McKinnon et al, 2009) may be due to decreased number and length of dendrites, and cell size and/or density (Boldrini et al, 2013; Rajkowska et al, 2005; Rajkowska et al, 1999; Stockmeier et al, 2004), apoptosis (Lucassen et al, 2001), and/or blunted neurogenesis (Boldrini et al, 2014; Boldrini et al, 2012), although there is an ongoing debate on levels of adult neurogenesis in human (Boldrini et al, 2018; Kempermann et al, 2018; Sorrells et al, 2018). Data-driven and hypothesis-driven approaches to neuroscience are complementary (Bouchard et al, 2016). The complexity of the human brain and the polygenic nature of neuropsychiatric diseases require data-driven approaches to assess numerous pathways at the same time (Shin et al, 2014). Gene expression and proteomics studies are used to identify biosignatures of psychiatric diseases and treatment response (McKenzie et al, 2017; Rangaraju et al, 2017; Rha et al, 2017; Wingo et al, 2017).

    Methods: We studied the proteome of the postmortem anterior DG/CA4 from fully characterized (Kelly and Mann, 1996) uMDD, MDD*SSRI and NC (n = 12/group). Subjects were 14-84 yrs of age, 10 MDD subjects died by suicide; females were four per group. The DG/CA4 region was cut with a bistoury from fresh frozen post-fixed HC, digested and used for proteomics assays. Shotgun proteomics liquid chromatography-tandem mass spectrometry (LC-MS/MS) was run for 120 minutes; two chromatograms were recorded for each biological replicate, in traveling-wave ion mobility spectrometry (TWIMS) MSE resolution mode (Brown, 2014). Results were analyzed with MSE/IdentityE algorithm (PLGS software v2.5 RC9), and annotations with identifications from MS data were processed using ProteinLynx Global Server (PLGS 2.2.5, Waters Corp.), and analyzed with Elucidator Protein Expression Analysis System (v4., Ceiba Solutions). We used Rosetta Elucidator post-processing ( (Levin et al, 2011), created Association Networks for Differentially Expressed Proteins, and used Database for Annotation, Visualization and Integrated Discovery ( Our quality control indices, for considering a protein a good candidate, were: p < 0.001, fold change of at least 20-25%, peptide count above two, good maximum protein score and peptide score. To validate differentially expressed proteins in the DG/CA4, we performed immunohistochemistry (IHC) following our protocols (Boldrini et al, 2018), and Western Blots (WB) and quantified proteins using Image Studio Lite (LI-COR, Lincoln, NE), correcting optical density by that of the housekeeping gene GAPDH.

    Results: Functions of the proteins differentially expressed between uMDD, MDD*SSRI and NC included one or more of the following: acetylation, adhesion, axon, differentiation, Golgi apparatus, maturation, methylation, microtubule, migration, neurogenesis, neuron, plasticity, proliferation, secretory granule, synaptic. This list included Secretogranin-2 (Scg2), VGF (non-acronymic), ES1 protein homolog, mitochondrial (ES1), and Heat Shock Protein 70 (HSP70). Using IHC cells immunoreactive for each protein were found in the human DG and proteins were detectable with WB in DG/CA4 from the same subjects included in the study.

    Conclusions: Scg2 is a marker for cell differentiation (Courel et al, 2014) that has been proposed as tractable target for neurological and psychiatric disorders (Bartolomucci et al, 2010). Scg2 is lower in cerebrospinal fluid in Bipolar Disorder (Jakobsson et al, 2013), clusters with the neuropeptide VGF, which is induced after fear conditioning, increasing BDNF/TrkB signaling and CREB phosphorylation (Lin et al, 2015). ES1 currently has no known functions but was found increased in fetal down syndrome brains (Shin et al, 2004). Hsp70 is involved in cellular repair (Mashaghi et al, 2016) and its expression is an indicator of neuronal stress (Sasara et al, 2004). There have been mixed findings regarding the association of mutated Hsp70 and MDD (Takimoto et al, 2003), and Hsp70 is higher in women with MDD (Pasquali et al, 2017). Different variants of the Hsp70 gene may affect antidepressant response (Pae et al, 2007).

    Proteins identified as differentially expressed in MDD and with SSRI treatment are involved in neuroplasticity, cell repair, stress responses. Proteins are the ultimate functional product of genetic and epigenetic changes, and proteomics analyses can reveal protein networks of biological significance. We can identify proteins that co-regulate biological processes in MDD and with SSRI treatment, which could hardly be discovered studying one molecule at the time. Proteins identified can be tested as potential biomarkers for in vivo imaging, screening tests, and drug targets.

    Keywords: Proteomics, Hippocampus, Mitochondria, Neuroplasticity, Antidepressants

    Disclosure: Nothing to disclose.

    M73. Essential Role for NPAS4 in Chronic Social Defeat Stress-Induced Anhedonia- and Anxiety-Like Behavior

    Makoto Taniguchi*, Jordan Carter, Jaswinder Kumar, Christopher Cowan

    Medical University of South Carolina, Charleston, South Carolina, United States

    Background: In susceptible individuals, the experience of stressful life events can suppress the future ability to experience pleasure, reduce motivation for reward pursuit and increase general anxiety. Deficits in the reward-related behavior, or anhedonia, is a key feature of the diagnosis of Major Depressive Disorder (MDD) and is strongly associated with the predicted severity and persistence of the disorder. In patients suffering from stress-related mental illnesses and MDD, and in several preclinical animal models, there are enduring changes observed in the prefrontal cortex (PFC) activity and E/I synapse balance, and restoration of PFC E/I balance has anti-depressant effects. Several preclinical depression models have linked chronic pathological stress experiences with changes in PFC gene expression, epigenetic modifications to DNA and histones, dysfunction of neuronal circuits, and persistent deficits in reward-related behaviors. Physical or emotional stress alters the expression of several transcription factors, including activity-dependent immediate early genes (IEGs). We recently found that the activity-regulated transcription factor, Neuronal PAS domain protein 4 (NPAS4), is transiently induced in the PFC by stress. NPAS4 is reported to modulate excitatory and inhibitory synapses and synaptic transmission, and as such, is well-positioned to mediate stress-induced changes in neural circuitry and stress-regulated behavior. In this study, we investigated the role and regulation of PFC NPAS4 in the chronic stress-induced changes in social, natural reward preference and motivation, and anxiety-like behaviors.

    Methods: The chronic social defeat stress (CSDS) assay was performed using standard conditions, including 5-10-min daily defeat bouts with a novel CD-1 strain aggressor followed by 24-hrs of co-housing in a divided cage. After 10 days of CSDS, experimental mice were subjected to a battery of test including social interaction, sucrose preference, Y-maze, open field exploration, elevated plus maze, and sucrose self-administration. Npas4 mRNA (qRT-PCR) and protein (IHC) expression were examined in reward-related brain regions following acute or chronic SDS. NPAS4 IHC was performed to examine co-localization with cell type-specific markers, including CaMKIIalpha and Parvalbumin. mPFC-specific knockdown of NPAS4 levels was accomplished in adult C57Bl/6 J mice using adeno-associated viruses (AAVs) expressing Npas4 shRNA or a scrambled control.

    Results: We found that acute and chronic SDS induces transient Npas4 mRNA expression in the mPFC and NAc, and the induction of Npas4 mRNA in the mPFC was significantly lower in CSDS compared to the first defeat experience. We find that NPAS4 protein in the mPFC is expressed predominantly in CaMKII-alpha-( + ), layer 2/3 excitatory pyramidal neurons with no NPAS4 detected in PV( + ) GABAergic interneurons. We found that reduction of NPAS4 levels using bilateral infusions of AAV-shNpas4 or AAV-shControl into the PFC of adult C57Bl6J mice altered behavioral changes induced by CSDS. Unlike control mice, the AAV-shNpas4PFC mice after CSDS failed to develop a reduction in sucrose preference (anhedonia-like behavior). AAV-shNpas4PFC mice also failed to develop an increase in anxiety-like behavior in the OFT. In contrast, AAV-shNpas4PFC mice do develop normal social interaction deficits in the 24-hr novel animal interaction test. Interestingly, AAV- shNpas4PFC mice alter motivation to work for a naturalistic reward in the progressive ratio test following a stable acquisition of sucrose self-administration – an effect observed in the susceptible animals as defined during by post-CSDS social interaction test.

    Conclusions: Our findings indicate that acute and chronic SDS induces strong and transient NPAS4 expression predominantly in the pyramidal neurons of layer 2/3 in the mPFC, and this induction is attenuated following chronic SDS. Viral-mediated reduction of Npas4 in PFC blocks the development of CSDS-induced anhedonia-like and anxiogenic phenotypes, without altering social avoidance. These findings reveal a novel role for PFC NPAS4 in the development of these CSDS-induced depression-related behaviors and indicate that social avoidance and anhedonia/anxiety phenotypes are dissociable and are possibly produced by distinct molecular and circuit mechanisms. Finally, since NPAS4 is reported to regulate GABAergic synapses formed onto excitatory cortical neurons, ongoing studies are investigating how NPAS4 might regulate circuit and synaptic changes underlying chronic stress-induced hypofrontality states.

    Keywords: Social Defeat Stress, Transcription Factor, Immediately Early Gene, Anhedonia, Anxiety

    Disclosure: Nothing to disclose.

    M74. Impact of CYP2C19 and CYP2D6 Genotypes on Clinical Outcomes and Side Effects in Patients Receiving Escitalopram and Aripiprazole for Major Depression: Results From the Can-Bind Cohort

    Kazunari Yoshida*, Farhana Islam, Victoria Marshe, Malgorzata Maciukiewicz, Raymond Lam, Glenda MacQueen, Benicio Frey, Roumen Milev, Sagar Parikh, Claudio Soares, Rudolf Uher, Gustavo Turecki, Susan Rotzinger, Jane Foster, Sidney Kennedy, Daniel Mueller

    Pharmacogenetic Research Clinic, Centre for Addiction and Mental Health, Toronto, Canada

    Background: Existing pharmacotherapy for major depressive disorder (MDD) is modestly effective, with vastly variable outcomes. The activity of cytochrome P450 (CYP) drug‐metabolizing enzymes has been shown to have major contributions to the variability in antidepressant plasma levels and thereby affecting treatment outcome. Specifically, increased, decreased, or absent CYP enzyme function due to genetic variability influences plasma drug and metabolite concentrations. The objective of the investigation was to examine the contribution of genetic variations in CYP2C19 and CYP2D6 genes in individual’s treatment with escitalopram (ESC) and aripiprazole (ARI) augmentation.

    Methods: Our sample included 178 individuals between 18-60 years of age and diagnosed with MDD (MADRS ≥ 24) from the Canadian Biomarker Integration Network in Depression (CAN-BIND) cohort. Participants receive protocolized treatment with open-label ESC (10-20 mg/d) for 8 weeks, where responders (i.e. ≥ 50% reduction in baseline MADRS score) continued ESC treatment while non-responders were augmented with an add-on medication, ARI (2-10 mg/d); each group for up to 16 weeks. Plasma levels of drug and metabolite concentrations were collected at Weeks 2, 10, and 16. Side effects were assessed using the Toronto Side Effects Scale (TSES) and Sexual Side Effects Questionnaire (SexFX).

    Given that CYP2C19 and CYP2D6 are the major metabolizing enzymes of ESC and ARI, respectively, metabolizer types (i.e., extensive, intermediate, poor and rapid metabolizers) were determined by assessing functionally relevant polymorphism by use of an established open-array platform. We investigated associations between metabolizer status and plasma drug/metabolite concentrations, as well as response status and side effects.

    Results: ESC plasma levels and metabolite/ESC ratios were significantly different across CYP2C19 metabolizer status groups (Kruskal-Wallis, FDR corrected p < 0.001). In comparing the largest groups, extensive (EM, n = 123) and intermediate (IM, n = 42) metabolizers, IMs showed consistently higher ESC levels across week 2 (10 mg, MIM = 26 ± 9 ng/mL, MEM = 18 ± 8 ng/mL), week 10 (20 mg, MIM = 43 ± 18 ng/mL, MEM = 32 ± 15 ng/mL), and week 16 (20 mg, MIM = 46 ± 20 ng/mL, MEM = 30 ± 14 ng/mL). CYP2D6 EMs showed the highest metabolite/ESC ratio at Weeks 2 and 16 after adjusting for time-since last dose (F(2, 176) = 3.6, p < 0.01). CYP2D6 metabolizer types showed significant differences in ARI/metabolite ratio at Weeks 10 and 16 in individuals in the ESC + ARI treatment arm (p < 0.001), as well as, nominal difference in ARI blood concentrations at Week 16 (p < 0.05). However, there were no association between CYP2C19 and CYP2D6 metabolizer status and clinical response at Week 8 or 16.

    For side effects, the frequency and severity of drowsiness were associated with ESC levels nominally at Week 2 (per 100 ng/mL, OR = 0.8 [0.7, 1.0], uncorrected p = 0.08) and significantly at Week 10 for both the ESC monotherapy (OR = 1.5 [1.1, 2.0], p = 0.01; p = 0.02) and ESC + ARI treatment groups (OR = 0.7 [0.6, 1.0], p = 0.08; p = 0.03). ESC metabolite levels showed a strong association with frequency (OR = 1.6 [1.1, 2.1]; p = 0.005) and severity (OR = 1.6 [1.1, 2.2]; p = 0.008) of dizziness at Week 10 and a trend for frequency (OR = 1.3 [1.0, 1.8]; p = 0.06) at Week 16 for only the ESC + ARI treatment group. Likewise, frequency and severity of blurred vision (OR = 0.7 [0.5, 1.0]; OR = 0.6 [0.4, 1.0]; p = 0.03, p = 0.02) and headache (OR = 1.5 [1.0, 2.2]; p = 0.05) was influenced by ESC levels in this group at Week 16. Finally, ESC levels were associated nominally with frequency of weight gain (OR = 1.4 [1.0, 2.0], p = 0.08) and significantly with severity of weight gain (OR = 1.7 [1.1, 2.7], p = 0.03) at Week 10 for only the ESC + ARI treatment group. In terms of sexual side effects, after adjusting for baseline sexual function, age and sex, metabolite/ESC ratio at Week 16 was significantly associated with total sexual functioning and global impression scores in the ESC + ARI treatment group (β = 2.6, β = 1.5, p < 0.05) and nominally in ESC monotherapy (β = -1.4, β = -0.8, p < 0.1).

    Conclusions: CYP2C19 and CYP2D6 genotypes are associated with steady state ESC and ARI levels, respectively, metabolite levels and metabolite/drug ratio, but does not show an association with response status. For side effects, frequency and severity of drowsiness show an association with ESC levels in both ESC monotherapy and ESC + ARI adjunctive therapy groups. However, only the ESC + ARI adjunctive therapy showed an association between frequency and severity of dizziness, blurred vision, headache and weight gain with ESC or its metabolite levels. Cases where opposite effects were observed between ESC monotherapy and ESC + ARI treatment groups suggest that ARI augmentation moderates the side effects associated with ESC. Further analyses will be presented how metabolizer status and plasma levels is associated with cognitive performance scores.

    Keywords: Pharmacogenetics, Escitalopram, Aripiprazole, Depression, CYP2D6

    Disclosure: Nothing to disclose.

    M75. Synapses and Cognition in Aging – Does Depression Accelerate the Decline?

    Sophie Holmes, Robert Pietrzak, Sjoerd Finnema, Maggie Davis, Nicole DellaGioia, Mika Naganawa, Nabulsi Nabeel, Matuskey David, Ronald Duman, Gerard Sanacora, Carson Richard, Krystal John, Irina Esterlis*

    Yale Translational Brain Imaging Program, New Haven, Connecticut, United States

    Background: Aging has widespread effects on the brain, affecting molecules, cells, vasculature, gross morphology, and cognitive function. Major depressive disorder (MDD), a highly prevalent and disabling psychiatric disorder, may accelerate brain aging, with some data suggesting it adds up to 10 years to the chronological age of chronically depressed individuals. To date, however, no known in vivo study has evaluated how MDD may accelerate brain aging at the molecular level. We conducted a novel positron emission tomography (PET) imaging study to evaluate synaptic density, emotional processing and cognitive functioning in MDD individuals across a wide age range. The radioligand [11 C]UCB-J was used to quantify synaptic vesicle glycoprotein 2 A (SV2A) in the brain, which we previously showed to be a robust marker of synaptic density.

    Methods: Ten unmedicated individuals with MDD (mean age = 40.1 ± 14.6) and ten age-, sex-, and smoking-matched healthy controls (HC; mean age = 36.4 ± 13.8) participated in a [11 C]UCB-J PET scan. Binding potential (BPND) was computed in the dorsolateral prefrontal cortex (dlPFC) and hippocampus (HIP) with the white matter centrum semiovale as reference region. Magnetic resonance imaging (MRI) was performed prior to PET to guide the placement of regions of interest. Control and MDD individuals participated in a comprehensive psychiatric assessment battery to assess mood (HAM-D, MADRS), anxiety (STAI, HAM-A), and other depression related symptoms. Cognitive function was assessed using the Cogstate cognitive battery.

    Results: Results revealed a significant interaction of MDD x age in predicting synaptic density, with individuals with MDD having significantly more pronounced age-related decline in synaptic density in the HIP (F = 13.55, p = 0.002) and dlPFC (F = 7.47, p = 0.015). The magnitudes of differences in slopes between the MDD and HC groups were very large in both the HIP and dlPFC (Cohen d = 1.8 in HIP and 1.4 in dlPFC). Results also revealed that lower synaptic density in the dlPFC was associated with greater severity of anhedonia symptoms (β = -0.55). Furthermore, we observed that lower synaptic density in the dlPFC in the MDD group was associated with worse performance on measures of attention (r = -0.54) and executive function (r = -0.43); and that lower synaptic density in the HIP was associated with worse performance on measures of visual learning (r = -0.30) and verbal memory (r = -0.43). Individuals with MDD also showed greater age-related reductions in verbal memory, attention, and working memory (all F’s > 3.41, all p’s < 0.09), with lower synaptic density in HIP linked to greater reductions in verbal learning (r = -0.54), and lower synaptic density in dlPFC linked to greater reductions in attention and working memory (r’s = -0.51 and -0.54).

    Conclusions: To our knowledge, these data provide the first in vivo evidence for a role of MDD in accelerating age-related changes in synaptic density in the HIP and dlPFC. They further suggest that greater changes in synaptic density are associated with more pronounced age-related changes in cognition, thus supporting the hypothesis that MDD may be a prodrome to dementia.

    Keywords: Synaptic Density, Depression, Aging, Human Imaging

    Disclosure: Nothing to disclose.

    M76. Relationship Between the Antidepressant Effects of Esketamine Nasal Spray and Perceptual Disturbances

    Li Chen, Xiang Li, Rosannne Lane, Maura Furey, Jaskaran Singh, Wayne Drevets, Guang Chen*

    Janssen Research & Development LLC of Johnson & Johnson, San Diego, California, United States

    Background: Clinical studies report that ketamine and esketamine produce antidepressant effects at subanesthetic doses. Both compounds also produce transient perceptual side effects within the same dose ranges. Previous reports showed that the increases in perceptual measures are associated with greater improvement on depression severity ratings after single iv ketamine infusion (Luckenbaugh et al., 2014; Niciu et al., 2018). Niciu et al also reported that depersonalization component scores also correlated with better improvements on depression ratings after single iv ketamine infusion. Other studies have been unable to replicate these findings (Valentine et al., 2011; Murrough et al., 2013). Notably, Nugent et al (2018) reported a statistical trend toward a correlation between the increases in perceptual effects and the increases in depression measures in healthy subjects receiving single ketamine iv infusion. In the current study, we conducted post-hoc analyses of data from a Janssen R&D phase 3 study to assess the relationship between the effects of esketamine nasal spray on ratings of depression severity and ratings of perceptual disturbance in participants with TRD.

    Methods: The clinical sample was drawn from a randomized, double-blind, active-controlled, multicenter study in male and female adult participants with TRD (NCT02418585). Subjects received flexible dosing intranasal esketamine (56 mg or 84 mg) plus an oral antidepressant or an oral antidepressant plus intranasal placebo twice per week plus a newly initiated oral antidepressant (AD) for 28 days. The Montgomery–Åsberg Depression Rating Scale (MADRS) and the Clinician Administered Dissociative States Scale (CADSS) were used as measures of depression severity and perceptual effects, respectively. The primary outcome measure was assessed by the difference in the change in MADRS between study arms at Day 28.

    Spearman coefficients were computed to assess the relationships between MADRS changes (i.e. change from baseline) at Day-2 (the day after the first IN treatment) and Day 28 (the day of last assessment of the induction phase of the study), and CADSS increases following the first administration of esketamine (i.e. post-treatment peak values) at Day-1.

    In the post-hoc analyses on the temporal profile study of the CADSS and MADRS changes, a mixed model for repeated measures (MMRM) was used. The model included treatment (IN esketamine + oral antidepressant, oral antidepressant + IN placebo), visit, country, class of oral antidepressant (SNRI or SSRI) and treatment-by- visit as fixed effects, and baseline CADSS (before treatment at Day 1) as covariates.

    Results: Of the randomized participants, 197 completed the double-blind period. Change in MADRS total score with esketamine nasal spray and oral antidepressant was superior to oral antidepressant and placebo nasal spray at Day 28 (LS mean [SE] difference vs. antidepressant plus placebo -4.0 [1.69], 95% CI: -7.31, -0.64; 1-sided p = 0.010); likewise, clinically meaningful improvement was observed with esketamine nasal spray plus oral antidepressant at earlier time points, including Day 2 (1-day post-treatment) (Popova V., et al. reported at the American Psychiatric Association Annual Meeting, 2018).

    The CADSS total, depersonalization, derealization and amnesia scores after first treatment did not correlate significantly with changes on the MADRS either at Day 2 in the esketamine + AD (n = 109) or the placebo + AD groups (n = 99) (lowest p-value > 0.48) or at Day 28 in the esketamine + AD (n = 101) or placebo + AD (n = 96) groups (lowest p-value > 0.14).

    The temporal profile of MADRS response over repeated esketamine dosing indicates that MADRS continues to improve over time (day 2-28). The least square mean of the reductions from baseline scores are -8.2 on day 2 and -19.8 on Day 28. The effects of Day and of Treatment are significant (Day: p < 0.001, Treatment: p = 0.0049), while the interaction of Day and treatment is not significant (p = 0.5341).

    The temporal profile of CADSS response over repeated esketamine dosing indicates that the peak CADSS scores decrease over time. The least square mean of CADSS increases from baseline scores are 8.1 and 3.3 after first and last treatments in esketamine arm. Meanwhile, the mean CADSS scores in the group receiving placebo + AD maintains between 0.3 to 0.7 over time. The interaction of Day and treatment is significant (p < 0.001).

    Conclusions: During the administration of esketamine nasal spray plus oral AD, the increases in the CADSS total and component scores did not correlate significantly with the improvement in MADRS scores at either of the two timepoints examined. Furthermore, the MADRS total scores continued to decrease, indicating improvement of depression with repeated dosing of either esketamine + AD or placebo + AD, while the CADSS scores showed attenuation with repeated dosing in the esketamine arm. The observation that with repeated dosing of esketamine, the perceptual effects diminish while the antidepressant effects increase, is consistent with the lack of correlation between the change in MADRS scores and the change in CADSS scores. A variety of evidence suggests that NMDAR antagonism is the primary direct pharmacological action of ketamine and esketamine at antidepressant doses. The current data nevertheless suggest that distinct mechanisms may underlie the antidepressant and perceptual actions of esketamine, although both actions might be triggered by the same initial target(s) or through different subtypes of NMDARs.

    Keywords: Ketamine, Esketamine, Depression, Antidepressant, Dissociation

    Disclosure: Johnson and Johnson, Employee

    M77. Testing a Synergistic, Neuroplasticity-Based Intervention for Depression: Study Design and Preliminary Findings From an RCT of Intravenous Ketamine Plus Neurocognitive Training

    Rebecca Price*, Benjamin Panny, Michelle Degutis, Crystal Spotts, M. Beth Minnigh, Sam Poloyac, Cassandra Gibson, Sanjay Mathew, Robert Howland

    University of Pittsburgh, Pittsburgh, Pennsylvania, United States

    Background: Depression has been described as a problem of impaired neuroplasticity (e.g., prefrontal synaptic depression) at the molecular level, and decreased cognitive flexibility and prefrontal cortex (PFC) control at the neurocognitive level. Intravenous ketamine, which displays rapid antidepressant properties, is posited to reverse depression by rapidly enhancing molecular neuroplasticity; but surprisingly little is known regarding its effects on depressed patients’ neurocognitive processing. We posit that ketamine will rapidly increase cognitive flexibility and the PFC’s influence on affective regions, allowing for rigid, negative biases in cognition to be rapidly reversed. We hypothesize these neurocognitive changes may be linked to individual differences in blood levels of a specific ketamine metabolite, (2 R,6 R)- hydroxynorketamine (HNK), which has been linked to antidepressant-like effects in animal models. We further hypothesize these neurocognitive changes will provide a clinical window of opportunity in which to introduce automated neurocognitive training techniques, which will consolidate adaptive forms of cognitive processing (specifically, positive implicit representations of self) while neuroplasticity remains high. Training adaptive forms of processing after first ‘priming’ the brain with ketamine represents a potentially synergistic treatment approach that could extend the acute effects of a single ketamine infusion beyond its typical 3-7-day window, efficiently fostering antidepressant effects that are both rapid and enduring.

    Methods: In this randomized controlled trial (NCT03237286), 150 unipolar depressed adults who have failed at least one FDA-approved antidepressant medication are randomized to one of three intervention arms: (1) a single infusion of ketamine (0.5 mg/kg over 40 min) followed by 4 days of active cognitive training (n = 50); (2) ketamine infusion followed by four days of sham cognitive training (n = 50); or (3) saline infusion followed by four days of cognitive training (n = 50). Concurrent antidepressant treatments (e.g., medications, therapy) are allowed as long as treatments are stable for 4 weeks prior to enrollment. Outcome measures span multiple levels of analysis. Patients complete repeated in-person assessments over a 30-day period (including clinical, neuroimaging, cognitive, implicit information processing, and serum ketamine metabolite measures) and are followed with remote assessments (e.g., self-report symptoms, implicit information processing) over the following year.

    Results: This poster will present the study design and preliminary analyses from this ongoing NIMH-funded R01. Key findings from the first cohort of 21 patients include the following:

    1. 1.

      (2 R,6 R) HNK blood levels are sustained at detectable levels in some patients for several days following a single ketamine infusion. The timeframe observed roughly corresponds to the 3-7-day clinical response window observed in prior ketamine trials.

    2. 2.

      The overall clinical response rate post-infusion (collapsing across ketamine/saline arms for blinded analysis) is 52.4% (11/21 responders). Assuming a null response to saline and based on the 2:1 allocation to ketamine:saline, this 52.4% response would be consistent with a 70-80% response rate in the ketamine arm, consistent with previous trials.

    3. 3.

      In blinded analyses comparing infusion acute responders (n = 11) to non-responders (n=10), decreased self-reported depression (QIDS-SR) scores are observed in responders and are sustained well into the follow-up period, which now extends to 4-months in some patients. Decreased self-reported depression (relative to baseline) is observed at every assessment out to 4-months post-infusion among responders (p<.05), while non-responders’ self-reported depression does not differ from baseline.

    Conclusions: In this ongoing study, we hypothesize that after priming brain plasticity with ketamine, training positive self-representations could provide an exceedingly efficient, low-cost, portable, non-invasive, safe, and highly dissemination-ready strategy for exploiting and extending ketamine’s rapid antidepressant effects. In preliminary blinded analyses, acute infusion responders show improvement in depression symptoms over a more sustained interval (i.e., 4 months following a single infusion) than has been observed in previous ketamine trials which (a) recruited more treatment-resistant patients and (b) did not include a synergistic neurocognitive/behavioral intervention. This study also aims to provide novel, integrative information on neurocognitive intermediaries bridging ketamine’s molecular and mood effects. Preliminary analyses show that (2 R,6 R)-HNK levels are detectable in patients’ serum for several days following a single infusion. Individual differences in these blood levels will be examined in relation to other outcomes across clinical and neurocognitive levels of analyses.

    Keywords: Ketamine, Cognitive Neuroscience, Depression

    Disclosure: Nothing to disclose.

    M78. Is Sleep Disturbance Linked to Short- And Long-Term Outcomes Following Acute- and Continuation- Phase Treatments for Recurrent Depression?

    Elaine Boland*, Jeffrey R. Vittengl, Lee Anna Clark, Michael E. Thase, Robin B. Jarrett

    Veterans Administration Medical Center, Philadelphia, Pennsylvania, United States

    Background: The field continues to search for treatments that eliminate depression. Sleep disturbance has been shown to predict increased rates of non-remission, with mixed evidence that persistent sleep disturbance is associated with greater likelihood of relapse and recurrence after antidepressant therapies. To date, studies have been limited by single, baseline assessments of sleep disturbance that do not reflect how changes in sleep disturbance over the course of acute-phase treatment may affect outcomes. Similarly, the literature is lacking studies that evaluate repeated measures of sleep disturbance across continuation-phase treatment, assess how changes in sleep disturbance during acute-phase treatment response may be related to subsequent depression relapse and recurrence, and examine how continuation-phase treatment may additionally improve sleep disturbance. The current study addresses these limitations, assessing how changes in sleep disturbance measured repeatedly across both acute- and continuation-phase treatment affected post-treatment outcomes. We hypothesized that greater overall levels of sleep disturbance and less improvement in sleep disturbance during acute-phase cognitive therapy (A-CT) would predict lower remission and recovery rates, and shorter time to relapse or recurrence of major depressive disorder (MDD). We also examined the potential impact of continuation-phase treatment on sleep among higher risk A-CT responders.

    Methods: Analyses were conducted on 523 adults with MDD who consented to 12 weeks of A CT followed by a randomized controlled trial of continuation-phase cognitive therapy (C-CT) or fluoxetine (FLX) versus pill placebo with clinical management to prevent depressive relapse for A-CT responders considered at higher risk of relapse/recurrence (defined by any Hamilton Rating Scale for Depression (HRSD-17) score ≥ 7 at any of the final seven consecutive acute-phase treatment visits). Continuation-phase treatment across all modalities was delivered over 10 sessions; 4 biweekly and 6 monthly visits for a total of 8 months. Patients were monitored for an additional 24-month, treatment-free follow-up period totaling 32 months of follow-up. Sleep disturbance was assessed using the global severity scale of the Pittsburgh Sleep Quality Index (PSQI) and a sleep problems scale derived from 8 sleep items from depression rating scales (i.e., Hamilton Rating Scale for Depression (HRSD-17), Beck Depression Inventory, and the Inventory of Depressive Symptomatology Self-Report). Sleep was assessed at weeks 1, 4, 5 and 12 of A CT, and approximately every 4 months throughout the continuation and follow-up phases. We used multilevel models to evaluate trajectories of sleep disturbance and depression improvement across A-CT. Individual intercepts and slopes were retained and entered into Cox regression models to predict remission (defined as the last 7 consecutive HRSD-17 scores < 7 or “return to usual self” according to the Longitudinal Interview Follow-up Evaluation), recovery (remission lasting 8 or more months), relapse (exacerbation of the presenting episode after a response but before recovery), and recurrence (meeting DSM-IV criteria for MDD following recovery) of MDD. We used multilevel models to test the effects of continuation treatments on sleep disturbance.

    Results: Although patients with greater average sleep disturbance during A-CT were less likely to achieve response, defined as an absence of DSM-IV major depressive disorder and an HRSD-17 score ≤12, and remission, overall there were significant improvements in depression symptoms (d = 2.6) and sleep quality (PSQI d = 1.0, sleep problems d = 1.2) over the course of A-CT. Response and remission were more likely for patients who exhibited greater reduction in sleep disturbance during A-CT. Patients with greater reduction in sleep disturbance during A-CT also achieved post-acute remission and recovery sooner. Continuation-phase treatment did not further improve sleep disturbance, but improvements made during the A-CT were maintained for 32 months among the higher risk responders. Neither average levels of A-CT sleep disturbance nor change in sleep disturbance predicted relapse or recurrence.

    Conclusions: Sleep disturbance among patients with recurrent MDD improves substantially during A-CT. Because greater reductions in sleep disturbance during A-CT were associated with more rapid recovery from depression, our findings support careful targeting of sleep disturbance during A-CT. Further research is needed to determine the extent to which full normalization of sleep is related to complete recovery from a depressive episode. Continuation-phase treatment does not appear to improve sleep quality further in this sample, but the extent to which maintenance of sleep quality during continuation-phase treatments may confer protection against depressive relapse/recurrence should be explored in future studies.

    Keywords: Sleep, Depression, Treatment

    Disclosure: Nothing to disclose.

    M79. The Use of a Mobile Cognitive Intervention to Target Cognitive Control Dysfunction in Middle Aged and Older Adults With Depression: Evidence for Target Engagement

    Faith Gunning*, Joaquin Anguera, Pat Arean

    Weill Cornell Medical College, White Plains, New York, United States

    Background: Poor cognitive control is common in older adults with major depressive disorder (MDD) and is linked to poor antidepressant response. Thus, alternative treatments for depression are needed, particularly for those individuals who exhibit cognitive control dysfunction. Further, interventions for depression that can be easily disseminated could greatly enhance the clinical impact of a neuroscience-informed interventions. Project EVO (or “EVO”) is a mobile therapeutic video game that has been shown to reduce older adults’ susceptibility to interference by augmenting cognitive control abilities (i.e., sustained attention and working memory abilities) through individualized adaptive algorithms (Anguera et al., Nature, 2013).

    Methods: In a sample of 36 middle-aged and older adults (both sexes), we examined the mechanism of action of EVO at the circuitry level (resting state and task-based functional MRI), performance level (computerized measures of working memory and sustained attention), and at the self-report level of analysis. These measures of cognitive control network functions were completed at baseline during a major depressive episode and following 4-weeks of at home use of EVO. Participants ranged in age from 45 to 75 years of age and had evidence of cognitive control dysfunction at the time of enrollment into the study. Change was defined a priori as a z score of > .5 from pre to post use of EVO. Z scores were calculated for each individual from each measure (task-based fMRI, resting state fMRI, task performance, self-report).

    Results: Results indicate that greater than 66% of patients showed a significant improvement, as measured by z score > .5, in cognitive control network functions following EVO. These changes were observed as increased bilateral activation of the middle frontal gryri and the dorsal anterior cingulate cortex in response to incongruent stimuli on a Stroop Flanker Task. In addition, greater patients showed a significant increase in functional connectivity in the left middle frontal gyrus and bilateral dorsal anterior cortex from pre to post use of EVO. Significant improvements in cognitive control functions were also apparent at the performance (TOVA) and self-report (FrsBe) levels of analyses. Mood symptoms, while not a primary outcome for this phase of the work, showed significant improvement from pre to post-EVO (p < .05). In addition, we explored changes in other brain networks as measured by resting state functional connectivity.

    Conclusions: These results indicate that a therapeutic video game approach may target cognitive control dysfunction and offer an alternative treatment for middle-aged and older adults who suffer from MDD with concurrent cognitive control dysfunction.

    Keywords: Experimental Therapeutics, Cognitive Control Network, Major Depressive Disorder, Resting-State Functional MRI, Functional MRI (fMRI)

    Disclosure: Nothing to disclose.

    M80. Transcranial Magnetic Stimulation Causally Influences the Subcallosal Cingulate Cortex Indexed by Interleaved TMS/fMRI

    Desmond Oathes*, Romain Duprat, Morgan Scully, Hannah Long, Kristin Linn, Joseph Kable, Joshua Gold, Michael Platt, Yvette Sheline

    Perelman School of Medicine University of Pennsylvania, Philadelphia, Pennsylvania, United States

    Background: There is burgeoning evidence that targeting regions of the lateral prefrontal cortex with transcranial magnetic stimulation (TMS) according to high resting correlation with the subcallosal cingulate cortex results in effective treatment of refractory major depressive disorder (MDD). However, an outcome as complex as clinical improvement is likely mediated by changing many brain areas/networks. Therefore, the question of whether targeted TMS actually has causal influence over the region of subcallosal cingulate related to MDD remained open prior to this study.

    Methods: We used individualized brain stimulation targets from an initial resting fMRI scan generated by seeding an MDD abnormality meta-analysis defined region of the subcallosal cingulate and choosing a surface accessible region with the strongest time series correlation for targeting TMS. In a mixed gender cohort (approximately equal) of 32 healthy participants, we subsequently interleaved 70 single TMS pulses with functional MRI acquisitions (2.4 s TR including 400 ms gap for TMS) at 120% of motor threshold defined for each participant. As a control site and as typically done in TMS experiments, we stimulated the primary motor cortex with the same sequence for comparison with the subcallosal cingulate target. Following standard event-related fMRI processing, we extracted average percent signal BOLD responses for the subcallosal cingulate region of interest then used a within-subjects paired-t test to establish the differential effect of TMS site on target response.

    Results: TMS to the primary target induced a negative deflection in the BOLD signal (>3% signal change) that was significantly larger than the deflection in response to the control site (motor cortex) stimulation (~1% change), t(31)=2.273, p=0.30, partial eta2=0.143. There was no interaction between the TMS evoked response and gender, age or years of education.

    Conclusions: We for the first time provide causal evidence that a region of the lateral prefrontal cortex with highly correlated BOLD signal to the subcallosal cingulate cortex has the ability to causally drive this target region in healthy individuals. We show that individualized fMRI guided brain stimulation can be an effective tool in guiding non-invasive circuit engagement that is assumed to underlie clinical response across a variety of conditions to this treatment modality. Our ongoing work will continue this investigation in a patient population.

    Keywords: Interleaved TMS/fMRI, Anterior Cingulate Cortex (ACC), Subcallosal Cingulate, Major Depressive Disorder (MDD), Resting-state fMRI

    Disclosure: Nothing to disclose.

    M81. Is There an Association Between Subjective and Objective Measures of Cognition and Brain Structure in Bipolar Disorder?

    Isabelle Bauer*, Thomas Meyer, MonJu Wu, Benson Mwangi, Giovana Zunta-Soares, Jair Soares

    University of Texas Medical School, Health Science Center at Houston, Houston, Texas, United States

    Background: Cognitive deficits are strong predictors of social functioning and treatment response. Further, experimental studies have demonstrated a strong link between subjective rating of cognitive deficits, poor intrinsic motivation, and reduced cognitive performance. To date, the neurocognitive literature of bipolar disorder (BD) has primarily investigated objectively measured cognitive function deficits in relation to structural abnormalities in fronto-limbic, cingulate, and striatal regions. There is sparse but clinically relevant evidence showing that BD patients are more likely than healthy controls to report cognitive deficits prior to performing cognitive tasks. No published study has, however, explored the link between subjective and objective measures of cognitive performance and brain structures in BD.

    Methods: The study included 24 healthy controls (36.18±11.18 years, 15 F) and 45 remitted patients with BD (35.92±12.33 years, 31 F). Computerized measures of complex processing speed, sustained attention, and reward-related decision making were used for the objective assessment of cognitive functioning. Participants completed mood questionnaires and measures of cognitive self-concept and intrinsic motivation. Brain measures of interest included Freesurfer-generated cortical thickness and volumes of the frontal pole, hippocampus, precuneus, inferior temporal, thalamus, caudate, and striatum.

    Results: BD patients self-reported poorer attentional abilities and intrinsic motivation compared to healthy controls. They displayed increased risk-taking and had fronto-temporal structural abnormalities. Self-reported cognitive difficulties correlated weakly with risk-taking but were strongly associated with current residual depressive symptoms. There was no significant correlation between fronto-temporal measures, risk taking and subjective cognitive measures.

    Conclusions: BD patients’ perceptions of their abilities may not be an indicator of their cognitive performance. Patients’ current mood state influences their perceptions of cognitive competence. Functional brain imaging measures may be better suited for answering questions on the link between brain performance, and subjective/ objective cognitive abilities.

    Keywords: self-Efficacy, Risk-Taking, Neuroanatomy, Bipolar Disorder

    Disclosure: Nothing to disclose.

    M82. Assessment of Cognitive Function in Bipolar Disorder With Passive Smartphone Keystroke Metadata: A BiAffect Digital Phenotyping Study

    Olusola Ajilore*, John Bark, Alexander Demos, John Zulueta, Jonathan Stange, Jennifer Duffecy, Faraz Hussain, Scott Langenecker, Peter Nelson, Kelly Ryan, Melvin McInnis, Alex Leow

    University of Illinois at Chicago, Chicago, Illinois, United States

    Background: Cognitive dysfunction in bipolar disorder has been shown to be present even in the euthymic state. Cognitive dysfunction in bipolar disorder has been associated with poor treatment response and increased risk of relapse. There has been recent attention on digital phenotyping and passive sensing through smart, connected devices to probe cognition in real-world settings. Our passive sensing, BiAffect, is a custom-built smartphone keyboard that captures keystroke metadata (‘how you type, not what you type’). In previous studies, our group has demonstrated that BiAffect-derived keystroke metadata is associated with mood symptom severity and can be used prospectively to predict changes in mood in patients with bipolar disorder. For the present study, we hypothesized that typing metadata would be significantly associated with cognitive domains measured with traditional neuropsychological testing.

    Methods: 18 participants with bipolar disorder and 12 healthy comparison subjects from the Prechter Longitudinal Study of Bipolar Disorder at the University of Michigan were provided a mobile phone with a customized keyboard that passively collected keystroke metadata. Participants also completed a neuropsychological battery including the Tower of London task and the Trail Making Test. Select BiAffect-derived time-based metrics (interkey delay, typing speed) were associated with processing speed and set-shifting on the Trail Making Test. A measure of disorder in typing and time to make a move on the Tower of London were compared using Shannon entropy.

    Results: Processing speed, as measured by Trail Making Test (part A), was significantly correlated with average interkey delay (i.e., time since last key, r = .5, p < .001) and keys/second (r = -.54, p < .001). Set shifting, as measured by Trail Making Test-Part B, was highly associated with average time since last key (r = .68, p < .00001) and keys/second (r = -.62, p < .00001). Participants with bipolar disorder had significant increases in entropy in interkey delay times (p = .048, d = -.83) and entropy of Tower of London move times (p = .029, d = -.84). Furthermore, Entropy in interkey delay was significantly associated with entropy in Tower of London moves in participants with bipolar disorder only (r=.78, p=.001), with a trend level group x association interaction (p = .05).

    Conclusions: This pilot study demonstrates that passive, unobstrusive smartphone keystroke metadata can be used to probe cognitive function and dysfunction in bipolar disorder, revealing multi-scalar behavioral features accessible through digital assays.

    Keywords: Bipolar Disorder, Cognitive Functioning, Digital Assessment, Smartphone-Based App

    Disclosure: Nothing to disclose.

    M83. High Sensitivity C-Reactive Protein is Associated With Poorer Cognitive Performance in Bipolar Disorder

    Caitlin Millett*, M. Mercedes Perez-Rodriguez, Emmett Larsen, Megan Shanahan, Katherine Burdick

    Brigham & Women’s Hospital, Boston, Massachusetts, United States

    Background: Data support the notion that 40-60% of patients with bipolar disorder (BD) have neurocognitive deficits. It is increasingly accepted that functioning in BD is negatively impacted by these deficits, yet they have so far not been a target for treatment. This is, in part, due to a lack of understanding of the biological mechanisms underlying cognitive impairment in BD. The biomarkers that predict cognitive deficits in BD are largely unknown, however, recent evidence suggests that inflammation may be associated with poorer cognitive outcomes in BD. To address gaps in knowledge regarding inflammatory biomarkers in cognition in BD, we measured peripheral high sensitivity C-reactive protein (hsCRP) in a large sample of euthymic men and women with BD (n=223). We hypothesized that higher hsCRP concentrations would predict poorer performance on neurocognitive assessments.

    Methods: We have completed an analysis of hsCRP in a large euthymic BD sample (n=223, 48% female, 18-65 y) and a healthy control (HC) sample (n=52, 58% female, 20-70 y). Blood was drawn within one week of neurocognitive testing by a research nurse and tested for hsCRP concentration on an electrochemiluminescence (ECL) multiplex-based assay platform. A one-way multivariate analysis of covariance (MANOVA) was used to assess the performance of the highest hsCRP quartile (n=54) against the remaining lower quartiles (n=158) on cognitive tests. Results were covaried with clinical correlates where appropriate.

    Results: We found a statistically significant effect of hsCRP quartile on cognitive performance in the BD group, with significant effects on several individual cognitive tests: those will high hsCRP had worse performance on Controlled Oral Word Association Task (p=0.003), Wisconsin Card Sorting Task (p=0.003), WAIS block design (p=0.005), vocabulary (p=0.004) and similarities (p=0.002), also reading mind in the eyes (p=0.001) and MCCB reasoning and problem solving (p=0.037). We also found evidence of cognitive decline in the highest hsCRP quartile compared to the lowest hsCRP quartile (p=0.047). Lastly, we performed a regression analysis of a composite cognitive Z score. A significant model (n=180, F=34.8, p<0.001, Adjusted R2=0.57) revealed significant predictors of overall cognition in the BD group, including CRPlog (β=-0.123, p=0.016) and other predictors.

    Conclusions: We have examined hsCRP, a peripheral inflammatory marker, in a large euthymic BD group. Our results indicate that higher hsCRP concentrations are associated with broad cognitive dysfunction in BD. Elucidating the underlying pathophysiology of cognitive decline in BD will help to develop new treatments in an area that has thus far been understudied.

    Keywords: Inflammation, Cognition, Mood Disorder

    Disclosure: Nothing to disclose.

    M84. Combinatorial Pharmacogenetic Testing-Informed Medication Switching From Genetically Incongruent to Congruent Medications Leads to Decreased Side Effect-Burden for Patients With Depression

    John Greden*, Sagar Parikh, Anthony Rothschild, Michael Thase, Boadie Dunlop, Charles DeBattista, Charles Conway, Brent Forester, Francis Mondimore, Richard Shelton, Matt Macaluso, James Li, Krystal Brown, Alexa Gilbert, Lindsay Burns, Michael Jablonski, Bryan Dechairo

    University of Michigan Comprehensive Depression Center, Ann Arbor, Michigan, United States

    Background: Major depressive disorder (MDD) is a leading cause in world burden of disease (World Health Organization). Standard antidepressant treatments result in partial, delayed or absent treatment responses and increasing side effect (SE) burden with each antidepressant trial. Medication safety and SEs play a large role in the path to treatment resistance in MDD (TRD). Precision medicine, such as genetic biomarkers, would hypothetically improve rates of clinical response, remission, and maintenance of overall patient wellness.

    A field of precision medicine, pharmacogenetics (PGx), has been studied for several decades with the ultimate goal to optimize medication treatment decisions. Previous results were disappointing and have become controversial, partially because of the small number of genes and variants studied, small sample sizes, and short-term studies. Recently, clinical advances have led to combinatorial PGx algorithms that integrate multiple pharmacokinetic (PK) and pharmacodynamic (PD) genes specific to each medication rather than the historical approach of large panels including single gene-single medication pairs alone. Here, we present SE data from the largest, longest, randomized, and patient/rater-blinded psychiatric combinatorial PGx study in patients with MDD and inadequate response and/or SEs to medication.

    Methods: This study included 60 academic and community sites and was funded by Assurex Health. Patients had a diagnosis of MDD, QIDS-C16 score ≥11, and had failed ≥1 antidepressant trial in the current episode. Patients were randomized 1:1 to a combinatorial PGx-guided or unguided treatment group. PGx test results were withheld from clinicians in the unguided arm until un-blinding after the 8 wk visit whereas the guided group clinicians had access to the testing to inform medication decisions throughout the study. Raters were blinded to treatment group and timeline. Patients were blinded to treatment group at baseline, 4 wk, and 8 wk; un-blinding could occur after week 8. The study continued with time points at 12 and 24 wk. The PGx test stratified medications based on severity of gene-drug interactions (GDI) based on a combinatorial algorithm incorporating PK and PD GDIs. Congruency is defined as medications with fewer GDIs residing in the “use as directed” and “use with caution” bins whereas incongruence is defined as medications with severe GDIs residing in the “use with increased caution and more frequent monitoring” bin. Patient-reported side effects were assessed as the mean number of SEs and proportion of patients reporting SEs according to study arm and congruence of baseline medication. SEs were only included with a probable causal link to medication (e.g. categorized as likely, probably, possibly, or definitely relating to medication).

    Results: 1,671 patients with MDD were evaluated for 24 wks. At 8 wks, there were no significant differences in patients experiencing SEs and SE burden between the PGx-guided and unguided groups [15.6% (88/560) versus 15.3% (93/607), p=0.881 and 0.243 versus 0.237, p=0.855, respectively]. In the subgroup of patients taking “incongruent” medications at baseline (n=213), significantly fewer patients who switched to congruent medications by 8 wk experienced SEs compared to those who remained on incongruent medications [6.5% (5/77) versus 16.5% (22/136), p=0.045]. Patients in this baseline incongruent subgroup who became congruent by 8 wk also had significantly fewer SEs compared to those who remained incongruent (0.065 versus 0.242, p=0.002).

    Conclusions: Medication treatment decisions guided by combinatorial PGx testing for patients with MDD who have failed at least one antidepressant medication trial result in reduced side effect frequency and burden. Patients switched from genetically incongruent to genetically congruent medications have the greatest benefit from combinatorial PGx testing.

    Keywords: Pharmacogenomic-Guided Treatment Recommendations, Treatment Resistant Depression, Major Depressive Disorder (MDD), Drug Side Effects, Pharmacogenetic Response

    Disclosure: Janssen Pharmaceutical, Advisory Board, Naurex (Allergan), Advisory Board, Cerecor Pharmaceutical, Advisory Board, Assurex Health, Consultant, NeuralStem, Advisory Board, Sage Therapeutics, Advisory Board, National Network of Depression Centers, Board Member, American Foundation for Suicide Prevention, Board Member, Canadian Depression Research Intervention Network, Advisory Board, Le Royal, Advisory Board, Depression and Bipolar Support Alliance, Board Member, Psychiatric Research Society Board of Directors, Board Member, Genomind, Inc., Advisory Board

    M85. Pharmacological Considerations in Pregnant and Postpartum Women With Bipolar Disorder

    Wendy Marsh*, Kailai Duan, John Straus, Nancy Byatt

    University of Massachusetts, Worcester, Massachusetts, United States

    Background: With a rate of up to 1 in 8 women becoming depressed in the pregnancy and postpartum (perinatal) period, mood episodes during the perinatal period are receiving increased and due attention. While evidence-based treatments for depression (and anxiety) include non-pharmacological approaches such as therapy, the best data for mood stabilization in bipolar disorder is medication based. While the pharmacological treatment of unipolar depression is gaining traction in obstetrics, the status of pharmacological considerations of bipolar disorder in perinatal women is less well documented. We hypothesize that in perinatal care providers of pts with mental illness both bipolar disorder and medications used to treat bipolar are increasingly being discussed in phone call consultations with perinatal psychiatrists.

    Methods: Data was collected from, the Massachusetts Child Psychiatry Access Program for Moms (MCPAP for Moms), a population-based program to help health care providers address mental health and substance use disorders among pregnant and postpartum women. Clinical encounter data from community health care providers with MCPAP for Moms perinatal psychiatrists were collected during telephone consultations. Descriptive statistical methods were used.

    Results: From Jul 2014 to June 2018, MCPAP for Moms served 4,544 women, of whom 579 or 12.7% in which bipolar spectrum was considered in the differential. Bipolar disorder came in third of mental health issues requested to be addressed. First anxiety/stress disorders spectrum 39.8% (n=1810) followed closely by depressive disorder spectrum at 36.2% (n=1650). Medication treatments (antipsychotics, anti-epileptics, lithium) associated with bipolar disorder treatment were discussed in 16.0% (n=729) of encounters. More than one medication type or diagnosis can be associated with one encounter. Frequency of bipolar diagnosis and bipolar medication discussion will be presented over time.

    Conclusions: Bipolar disorder is being considered in perinatal women in whom their health care provider is concerned about their mental health. Despite relatively high risk of teratogenicity with many traditional mood stabilizers (example valproate) and relative paucity on information on safety in pregnancy on many newer mood stabilizers (example atypical antipsychotics) compared to fist line treatments in depression and anxiety such as selective serotonin reuptake inhibitors, discussion of mood stabilizing medication in pregnant and postpartum women is taking place.

    Keywords: Bipolar Disorder, Pregnancy, Pharmacology

    Disclosure: Nothing to disclose.

    M86. Cellular Aging Associated With Severe Stress in Suicidal Patients

    Ricardo Caceda*, Sue Griffin, Pedro Delgado, Jorge Gamboa

    Stony Brook University, Stony Brook, New York, United States

    Background: Telomere length, a marker of cellular aging, is associated with psychological stress and several psychiatric disorders including major depression. Suicide is usually associated with mental illness. Suicidal crises are usually triggered by overwhelming psychological pain and overtly psychosocial stressful situations. We previously described shorter telomere length in individuals with current suicidal ideation than in recent suicide attempters. Here we aim to test the temporal dynamics of psychological pain and stress on cellular aging in suicidal individuals.

    Methods: We measured buccal cellular and mitochondrial telomere length, salivary cortisol, psychological pain, depression, hopelessness and demographic variables in depressed adult patients hospitalized for acute suicide risk in an academic center: current suicidal ideation (n = 131) and recent suicide attempt (n= 84) within the previous 5 days.

    Results: The suicidal ideation group had more severe depression, psychological pain, longer mitochondrial telomere length, and higher salivary cortisol than the suicide attempt group. Telomere length was negatively associated with depression, suicidal ideation, salivary cortisol and psychological pain severity. In the suicide attempter group telomere length was associated with time since the attempt.

    Conclusions: Our findings suggest that the decrease telomere length in suicidal patients may be more influenced by the overtly stressful nature of a suicidal crisis, rather than suicidal behavior itself.

    Keywords: Suicide, Acute Stress, Telomere

    Disclosure: Nothing to disclose.

    M87. Real-Time fMRI as a CBT Adjunct: Predicting the Behavioral Impact of Neurofeedback

    Kathryn Dickerson*, Katherine MacDuffie, Jeff MacInnes, Kari Eddington, Timothy Strauman, R. Alison Adcock

    Duke University, Durham, North Carolina, United States

    Background: Cognitive behavioral therapy (CBT) focuses on changing inaccurate, negative thoughts and behaviors. Though a handful of studies have begun to examine the neural correlates of CBT, no study to date has combined rtfMRI neurofeedback with CBT with the goal of evaluating therapeutic efficacy. An experiment from our group aimed to do so by showing individuals how brain activation within the anterior cingulate cortex (ACC) changes in response to 1) personalized, negative autobiographical memories or worries (e.g., death of parent) and 2) use of a therapeutic strategy (e.g., regulate). Prior analyses of our data showed that ACC neurofeedback “strength” (cingulate activation for STRATEGY – MEMORY trials) correlated with self-reported immediate post-scan strategy efficacy ratings and predicted strategy efficacy and frequency ratings one-month post-scan. The aim of the present report is to identify whole brain predictors of behavioral change following neurofeedback, both to understand the neural mechanisms of CBT and for continued intervention development.

    Methods: We recruited individuals (male and female) who had completed a standardized course of CBT as part of a clinical trial (N=24). Participants in our study (N=13) provided a list of personal episodic negative memories or worries as well as strategies used to cope with their mood. During neurofeedback runs, participants viewed a feedback summary after each MEMORY and STRATEGY trial as an index of how brain activity from their cingulate cortex changed in response to negative memories/worries and therapeutic strategies. During post-scan analyses, we constructed a whole brain general linear model with the following predictors: baseline BDI score, medication status, time since CBT completion, feedback values for the MEMORY trials, feedback values for the STRATEGY trials, emotionality of the MEMORY trials, follow-up strategy frequency ratings, and follow-up strategy efficacy ratings. We were specifically interested in examining brain regions engaged during MEMORY trials, STRATEGY trials, comparison between the two trial types, and given our prior findings – what correlated with self-reported strategy efficacy and frequency of use.

    Results: We first examined whether brain activation correlated with clinical symptoms. We observed that during Memory trials (relative to Strategy trials), baseline BDI scores positively correlated with activation within the inferior frontal gyrus and supramarginal gyrus. Given our prior findings that ACC neurofeedback “strength” correlated with self-reported strategy efficacy and frequency, we next examined if any additional brain regions tracked perceived strategy efficacy or frequency. Indeed, during Strategy trials (relative to Memory trials), activation within the periaqueductal gray nucleus (PAG), insula, and temporal pole predicted self-reported frequency of strategy use one-month post-scan (i.e., more PAG activation was associated with greater frequency of strategy use). All results are whole brain, cluster corrected with FSL Flame 1 to p<0.05 (thus correcting for multiple comparisons).

    Conclusions: Activation in regions associated with physical and emotional pain (e.g., PAG and insula) were engaged when using strategies to regulate sad memories/worries. This suggests that though using strategies to regulate negative memories/worries may be an emotionally painful process, it may still be fruitful as PAG and insula activation also predicted strategy use one month later. These results contribute novel findings regarding the neural mechanisms underlying CBT and suggest novel brain regions as potential future targets.

    Keywords: Major Depressive Disorder (MDD), Real-time fMRI Neurofeedback, Anterior Cingulate Cortex (ACC), Cognitive Behavioral Therapy

    Disclosure: NeuroLeadership Institute, Consultant

    M88. Reward-Circuit Biomarkers of Risk and Resilience in Adolescent Depression

    Adina Fischer*, Monica Ellwood-Lowe, Natalie Colich, Anna Cichocki, Tiffany Ho, Ian Gotlib

    Stanford University, San Carlos, California, United States

    Background: There is an estimated 36% cumulative incidence of Major Depressive Disorder (MDD) in adolescent females in the United States. Dysfunctional reward processing underlies core features of major depressive disorder, including anhedonia, depressed mood, and decreased motivation. While there is growing knowledge of reward processing dysfunction in adolescent depression, researchers have ignored underlying neural mechanisms of resilience. Here, we examine neural correlates of reward processing that characterize resilience and risk in adolescents at risk for depression, an essential step toward advancing our understanding of how to strengthen resilience to psychopathology in at-risk youth.

    Methods: 50 adolescent females were followed longitudinally through age 18: 32 at-risk adolescents who either did not experience depression (resilient; n=17) or had previously experienced a depressive episode (remitted-depressed; n=15), and 18 low-risk healthy controls. Participants completed clinical assessments at 18-month intervals and completed an fMRI reward- processing task in late adolescence. We conducted predictive modeling using logistic regression with a priori reward regions of interest (ROIs).

    Results: At-risk resilient and remitted-depressed adolescents exhibited less striatal and supramarginal gyrus activation than did controls during anticipation of reward versus loss (Z>2.3 cluster-level p<0.05). Resilient adolescents also exhibited greater activation than did remitted-depressed adolescents in the middle frontal gyrus during anticipation of reward versus loss, and less activation in the superior frontal gyrus and cuneus during outcome of reward versus loss (Z>2.3 cluster-level p<0.05). Using predictive modeling, we found that ventral anterior cingulate cortex and bilateral putamen activation during reward processing distinguished resilient from remitted-depressed at-risk adolescents with 83% accuracy.

    Conclusions: Distinct patterns of neural activation in the processing of reward appear to be markers of risk and resilience that may be targets for prevention and treatment approaches aimed at strengthening adaptive reward processing in at-risk adolescents.

    Keywords: Risk and Resilience, Major Depressive Disorder (MDD), Adolescence, Reward Circuitry, fMRI Biomarkers

    Disclosure: Nothing to disclose.

    M89. Interleukin-8 (but Not Interleukin-6) Shares Genetic Overlap With Risk for Suicide Attempt: The Confounding Roles of Sex and BMI

    Emma Knowles*, Joanne Curran, Ana Leandro, Harald Göring, Rene Olvera, Ravi Duggirala, Laura Almasy, John Blangero, David Glahn

    Yale University, Hartford, Connecticut, United States

    Background: Suicide is major public health concern. In the United States alone the rate of suicide attempts has risen by 30% since 1999 and research shows that those attempts are not always preceded by a known mental health condition. Suicide is preventable if risk can be identified and appropriate interventions provided promptly. To this end it is imperative to identify robust and reliable indicators, or biomarkers, of suicide risk that transcend traditional diagnostic boundaries. Previous work has revealed a relationship between increased inflammation and risk for suicidal behavior, but it is not clear the extent to which these alterations are genetically linked or if they are driven by environmental confounds (e.g. sex or BMI). We present the first study to directly address whether the shared etiology is due to genetic factors while considering the role of potential confounds.

    Methods: Genetic analyses were conducted in SOLAR in a sample of 1884 Mexican-American individuals from extended pedigrees (83 families, mean size = 21.67 individuals, range 2-178 individuals). The sample was 60.43% female and had a mean age of 42.04 years (sd = 15.74 years, range = 18-97 years). Using bivariate polygenic models, genetic correlations were calculated between IL-6 and IL-8 and lifetime risk for suicide attempt. The effects of sex, BMI, depression, medication, and smoking status were considered on the genetic overlap between phenotypes.

    Results: 159 individuals endorsed having attempted suicide in their lifetime (h2 = 0.47, p = 2.68×10-05). IL-6 (h2 = 0.17, p = 5.46×10-05) and IL-8 (h2 = 0.30, p = 4.54×10-14) were significantly heritable. Genetic correlation analysis (including significant covariates) revealed a significant correlation between risk for suicide attempt and IL-8 (ρg = 0.49, se = 0.17, p = 3.16×10-03). BMI was a significant covariate of IL-6 (p=9.85×10-29) and interestingly the genetic correlation between suicide attempt and IL-6 was significant only if BMI was removed as a covariate (ρg = 0.53, se = 0.19, p = 6.93×10-03). Suicide attempts were significantly more common in females (p=0.01) and the genetic relationship between IL-8 and suicide attempt appeared to be driven by females (ρg = 0.57, se = 0.17, p = 6.49×10-03), it was not significant in males alone (ρg = 0.26, se = 0.49, p = 0.61). The inclusion of major depression, and medication or smoking status in the model did not alter the results.

    Conclusions: Previous literature on inflammation and suicide has placed greater focus on IL-6 than IL-8 as a marker of risk; but the present results demonstrate that IL-8 (but not IL-6) have shared genetic influences with risk for suicide attempt and that this is a sex-specific effect driven by females, which may explain why it has previously been overlooked. It appears that the effects of BMI on inflammation levels might have confounded results from previous studies that have placed greater emphasis on IL-6 for suicide risk.

    Keywords: Suicide, Inflammation, Peripheral Biomarker, Mood Disorder, Sex

    Disclosure: Nothing to disclose.

    M90. A Magnetic Resonance Spectroscopy Study of Glutamate and Other Metabolites in Fronto-Limbic Regions in Adolescents and Young Adults at Increased Genetic Risk for Bipolar Disorder

    Rhoshel Lenroot*, Bronwyn Overs, Brenda Ta, Gloria Roberts, Caroline Rae, Janice Fullerton, Philip Mitchell

    University of New South Wales, Albuquerque, New Mexico, United States

    Background: Individuals at high genetic risk for bipolar disorder (BD) have shown differences in structure and function of fronto-limbic brain regions that are associated with emotional regulation. Genetic risk for BD is also associated with risk of depressive and anxiety disorders, which may indicate progression towards BP. Glutamate has been implicated in the pathophysiology of BD. There have been few previous magnetic resonance spectroscopy (MRS) studies in at-risk individuals and results have been mixed. We therefore used proton magnetic resonance spectroscopy (MRS) to measure glutamate (Glu) N-acetylaspartate (NAA), creatine (Cre), choline (Cho), and myo-inositol (MI) in a large cross-sectional sample of at-risk adolescents and young adults and to test for effects of age and lifetime history of a depressive or anxiety disorder. We hypothesized that glutamate levels would be increased in high-risk youth and NAA would be decreased.

    Methods: Population: 188 individuals aged 14-33, including 98 (43 male) at-risk (AR) first degree relatives of individuals with BP and 90 (41 male) community controls (CON) recruited as part of the Bipolar Kids & Sibs study (for details see (1)). CON excluded history of BP or psychosis but could have other anxiety or depressive disorders. AR or CON with history of either a DSM-IV anxiety or depressive disorder were designated Dx+, those without as DX-. The DX+ group included 47 AR and 20 CON. MRS acquisition: MRS data was acquired on a 3 T Philips Achieva scanner using PRESS (TR/TE 2000/40 ms, 90° flip angle, spectral bandwidth 2000hz, 2 acquisitions per ROI) anterior cingulate (ACC, 20 ×20 ×20 mm), left and right hippocampi (30 ×20 ×15 mm). Structural MRI was acquired to quantify voxel tissue composition. MRS analysis: Spectra were quantified using jMRUI(2). Residual water signals were removed prior to fitting and water peaks measured from unsuppressed spectra. Final sample size after removal of data due to poor-quality spectra, missing covariate data, diagnostic discrepancies, and outliers was 159 ACC (78 control), 120 L hippocampus (49 control), and 98 R hippocampus (43 control). The Dx+ subset included 39 Dx+ AR and 13 Dx+ CON. Gln was analyzed within subset of data with adequate GLN estimates (Cramer Rao lower bound < 50% for ACC, Gln estimate < 3SDs for Hippocampus): this included 78 for the ACC (33 control), 104 and 89 for left and right hippocampi respectively. Metabolites are reported as ratio to water. Statistical analysis: Linear mixed effects models were used including fixed effects of group (AR or CON), gender, age, gray matter percentage (gm%) and presence of lifetime diagnosis of depression or anxiety disorder. Bilateral hippocampal metabolite ratios were included in a single model, with hemisphere as a repeated and fixed factor. In instances where hemisphere interactions were observed, additional unilateral models were run to explore the effects within each hemisphere separately. A backwards selection method was used to define the final model for each metabolite ratio.

    Results: Cr levels were higher in hippocampus in AR than CON (t=2.20, p=.03). There were no other significant group differences between AR and CON. For the ACC model containing NAA, an interaction of age with group was seen (all reported results significant at p<.05): NAA levels decreased with age in CON but increased in AR. History of a depressive or anxiety disorder affected metabolite levels: in R and L hippocampus Gln and Cho levels were higher in Dx+ than Dx-. Several metabolites showed interactions between risk status and history of anxiety or depression. In the ACC, Cr was higher in AR Dx+ but not different in CON. NAA was lower in CON Dx- than CON Dx+, but the pattern as reversed in the AR subjects. In R and L hippocampus, Cho was lower in AR Dx+ than AR Dx-, but higher in CON Dx+ than CON Dx-. Age also interacted with Dx: in the ACC, Gln increased with age in Dx+, but did not show age effects in Dx-.

    Conclusions: In the largest study reported to date of metabolite levels in fronto-limbic areas in adolescents and young adults at increased genetic risk for BD, we found increased hippocampal Cr in the at-risk group and evidence of complex interactions of genetic risk, history of anxiety or depressive disorder, and age. Previous reports in smaller studies have had mixed results, but overall not found a strong effect of genetic risk on metabolite levels (3). Our findings are consistent with this, but the presence of interaction effects suggests that history of mood and anxiety symptoms and developmental effects may be contributing to variability in MRS results. Limitations of the study include cross-sectional design, smaller number of CON subjects with history of anxiety and mood disorders, and challenges measuring Gln at 3 T. Future directions include analysis of metabolite levels against symptom scores rather than categorical diagnoses, and comparison of metabolite levels in the subset of AR individuals who develop bipolar disorder.

    Keywords: Proton Magnetic Resonance Spectroscopy, Bipolar Disorder, High Risk, Glutamate

    Disclosure: Nothing to disclose.


    1) Perich et al., J Psych Res 2015 62:71.

    2) Stefan et al., Measurement Sci and Tech 2009 20:104036

    3) Hajek et al., J Psych Neurosci, 2008 33:531

    M91. Escitalopram Response is Associated With Pre-Treatment Neuraland Clinical Correlates of Anhedonia

    Katharine Dunlop*, Sakina Rizvi, Xiamen Leng, Stefanie Hassel, Stephen Strother, Raymond Lam, Roumen Milev, Susan Rotzinger, Jane Foster, Benicio Frey, Sagar Parikh, Claudio Soares, Rudolf Uher, Gustavo Turecki, Glenda MacQueen, Jonathan Downar, Sidney Kennedy

    Weill Cornell Medical College, New York, New York, United States

    Background: Anhedonia is a core feature of major depression, and is associated with dysfunction

    in mesocorticolimbic regions for reward. High anhedonia a negative predictor of antidepressant

    response. However, it is unknown whether anhedonia and brain dysfunction also negatively

    predicts escitalopram response. Therefore, the aim of this study is to identify pre-treatment

    predictors of escitalopram response using clinical and fMRI correlates of anhedonia.

    Methods: 134 men and women were recruited as part of the Canadian Biomarker Integration

    Network in Depression. Patients underwent 8 weeks of open-label escitalopram treatment. K-means clustering using MADRS scores was used to identify MDD groups with similar trajectories of antidepressant response. Using two valid instruments, we assessed anhedonia severity at baseline and at 8-weeks. All participants completed a monetary incentive delay task (MIDT) in the fMRI at baseline, 2-weeks and 8-weeks. All analyses were executed in FSL using default preprocessing and first-level statistics.

    Results: Patients clustered into ‘fast’ responders, ‘slow’ responders and non-responders. Slow

    responders and non-responders reported significantly higher pre-treatment anhedonia (p < 0.05)

    relative to fast responders and controls. Ventral striatal functional connectivity throughout the task correlated with anhedonia severity in healthy controls; this relationship was absent in the MDD

    group. Task-evoked mesocorticolimbic and medial parietal activity differentiated escitalopram

    non-responders from escitalopram responder groups.

    Conclusions: The results indicate that escitalopram non-responders display significantly higher

    clinical and neural correlates of anhedonia. Future work will incorporate genetic and molecular

    factors of reward processing that may likely impact escitalopram response, as well as machine

    learning processes to assess the individual-level prediction accuracy of these features.

    Keywords: Brain Based Markers for Depression, Anhedonia, Depression, Functional MRI (fMRI), Escitalopram

    Disclosure: Nothing to disclose.

    M92. Association Between Depression Severity and Hippocampal Volumes in Vietnam War Veterans With PTSD, TBI, Both, or Neither

    An Li, Sonja Stojanovski, Arielle Levy, Gabriel Devenyi, Mallar Chakravarty, Anne Wheeler*

    SickKids Research Institute, Toronto, Canada

    Background: High comorbidity is found between post traumatic stress disorder (PTSD), traumatic brain injury (TBI) and depression and reduced hippocampal volume has been associated with all three conditions suggesting that the etiology and neural basis of these disorders may overlap. The objective of this study was to examine whether the association between depression severity and hippocampal volume is moderated by the presence of TBI or PTSD.

    Methods: The Alzheimer’s Disease Neuroimaging Initiative – Department of Defence (ADNI-DOD) dataset of aging veterans was utilized for the study. Hippocampal volumes including subfields and extrahippocampal white matter projections were automatically extracted from T1-weighted MRIs in veterans with PTSD (n=48), TBI (n=13), PTSD&TBI (n=19), and healthy aging controls (n=44) using the MAGeT brain segmentation pipeline. Depression severity was assessed with the geriatric depression scale. Linear models were used to examine relationships between hippocampal volumes and depression severity, as well as interactions with group membership.

    Results: A negative association between hippocampal volumes and depression severity in veterans was moderated by the presence of PTSD such that in the two PTSD groups there was no association between hippocampal volume and depression severity. This effect was observed in both the hippocampus and extrahippocampal white matter projections and was most pronounced in the left hemisphere.

    Conclusions: These results suggest that the relationship between hippocampal volume and depression severity is moderated by the presence of PTSD. Distinct neural bases of depression severity in veterans with TBI and PTSD potentially indicate distinct etiology of depression symptoms that may benefit from different types of treatments.

    Keywords: Depression, Traumatic Brain Injury, Post Traumatic Stress Disorder, Hippocampal Volume, Veterans

    Disclosure: Nothing to disclose.

    M93. Aberrant White Matter Microstructure Relates to Elevated Irritability in Pediatric Bipolar Disorder and Disruptive Mood Dysregulation Disorder

    Julia Linke*, Nancy Adleman, Joelle Sarlls, Andrew Ross, Samantha Perlstein, Heather Frank, Kenneth Towbin, Daniel Pine, Ellen Leibenluft, Melissa Brotman

    National Institutes of Health, NIMH, Bethesda, Maryland, United States

    Background: Longitudinal data showed that chronic pediatric irritability, which affects approx. 10 million youth in the U.S., does not confer the risk to develop bipolar disorder (BD). This led to the creation of the DSM-5 diagnosis of disruptive mood dysregulation disorder (DMDD). Albeit on a different timescale, both DMDD and BD share an elevated proneness to anger, but whereas numerous diffusion tensor imaging (DTI) studies in BD suggested altered myelin plasticity as a key mechanism in this disorder, DTI has not yet been applied in DMDD. Research on how BD and DMDD differ with regard to structural connectivity might guide the development of targeted interventions for these different, but equally impairing phenotypes.

    Methods: We acquired DTI data from 118 participants (BD=36, DMDD=44, HV=38). After preprocessing with TORTOISE applying robust tensor fitting, images of fractional anisotropy (FA), longitudinal diffusivity (LD) and radial diffusivity (RD) were processed with tract based spatial statistics. Next, the three measures were used in an ANCOVA each with age, sex and medication load as nuisance variables and to train Gaussian process classifiers to predict, which group participants belong to. Finally, we determined how results relate to the symptom dimension of irritability.

    Results: In BD vs. HV, we observed widespread reductions in FA and increased RD centered at the corticospinal tract (5206 voxels, pmin=0.0002). In DMDD vs. HV, reductions in FA were confined to the anterior corpus callosum (1494 voxels, pmin=0.007) and negatively associated with the symptom dimension of irritability (rS = -.34, p < .001) and with reduced LD (1453 voxels, pmin=.0002). In BD vs. DMDD, FA in the corticospinal tract and the posterior corpus callosum was reduced. While we observed a negative association between FA in the corticospinal tract and irritability (rS = -.39, p < .001), there was a positive association in the posterior corpus callosum in DMDD only (rS=.55, p=.0001). The Gaussian process classifier could discriminate between BD and HV with an accuracy of 75%, and between DMDD and HV with an accuracy of 68%.

    Conclusions: We replicated findings of widespread reduced FA and increased RD in BD, supporting the hypothesis of altered myelination in BD. Alterations in DMDD were more regionally discrete and related to altered longitudinal diffusivity. Reduced FA in the anterior corpus callosum observed in both patient groups and previously associated with working memory performance and the corticospinal tract relevant in BD only and previously associated with processing speed were associated with elevated levels of irritability. In DMDD only, irritability was positively associated with FA in the posterior corpus callosum. From a clinical perspective, these findings will contribute to the pathophysiological understanding of DMDD and its differentiation from BD.

    Keywords: Irritability, Disruptive Mood Dysregulation Disorder, Bipolar Disorder, Diffusion Tensor Imaging (DTI), Machine Learning Classification

    Disclosure: Nothing to disclose.

    M94. Cortical Thickness Increases With Levomilnacipran in a Randomized Placebo-Controlled Trial in Late-Life Depression

    Beatrix Krause, Lisa Kilpatrick, Prabha Siddarth, Kelsey Laird, Linda Ercoli, Yesenia Aguilar-Faustino, Katherine Narr, Helen Lavretsky*

    Semel Institute for Neuroscience & Human Behavior, Los Angeles, California, United States

    Background: Late-life depression (LLD) is frequently comorbid with medical and cognitive disorders and can results in an inadequate treatment response. More efficacious treatment to improve mood, cognition and quality of life in LLD are urgently needed. Levomilnacipran (LMIL) is a novel anti-depressant whose effects on neuroplasticity have not yet been investigated. We investigated the effect of LMIL on brain cortical thickness in a randomized control trial (RCT).

    Methods: Twenty-nine adults>60 years with major depression (48.3% female; mean age=71.5[SD=5.8] years; mean education=16.0[SD=1.7] years) were randomized to either LMIL or placebo for 12 weeks. Inclusion criteria were a Hamilton Depression Scale (HAMD) >16. T1-weighted images were acquired at baseline and 12-weeks using a Siemens 3 T Prisma system. Freesurfer version 6.0 was used for cortical reconstruction and a whole-brain longitudinal voxel-wise two-stage model was used to compare the between-group symmetrized percent change in cortical thickness across treatment. Age and total intracranial volume were included as covariates. Voxel threshold was set 1.3-5 and Monte-Carlo cluster threshold to 1.3. The smoothing factor was set to 15 full-width half-maximum.

    Results: Of the 29 randomized subjects, 15 subjects (6 LMIL and 9 placebo) completed the study. Dropout rates did not significantly differ between groups. Both groups improved in HAMD (LMIL median change=-14, p=.06; Placebo=-8, p=.004; between-group, p=.5) but LMIL showed greater rates of remission (LMIL=57%; Placebo=33%, p=0.3). In addition, the LMIL group showed larger increases in cortical thickness in the right postcentral gyrus (primary somatosensory) X=44, Y=-26.9, Y=53.8; cluster size=966mm2), the precentral gyrus (primary motor) (X=17.3, Y=-22.1, Z=15.7; 948.9mm2), and the lateral occipital cortex (visual cortex) (X=19.6, Y=-94.9, Z=15.7; 930mm2) while the placebo group decreased over time.

    Conclusions: Both, LMIL and placebo groups showed improvement in depression severity with higher remission rates in the LMIL group. However, the LMIL group demonstrated a gain or preservation of cortical thickness in primary sensorimotor and visual regions in contrast to decreases in thickness in the placebo group. Larger studies are necessary to detect drug/placebo differences in clinical outcomes in relationship to brain neuroplasticity.

    Keywords: Late-life Depression, CNS Clinical Trials, Neuroimaging Biomarkers, Cortical Thickness, Levomilnacipran

    Disclosure: Allergan, Grant

    M95. Obesity is a Predictor of White Matter Microstructure Damage in Cingulate Gyrus of Euthymic Bipolar Patients

    Ramiro Reckziegel, Monise Costanzi, Isadora B. Remus, Leticia S. Czepielewski, Pedro D. Goi, Mireia F. Vianna-Sulzbach, Raffael Massuda, Clarissa Gama*

    Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil, Porto Alegre, Brazil

    Background: Obesity is associated with both structural and functional changes of the central nervous system, and has a high prevalence in BD. It is defined by a body mass index (BMI) higher than 30, which is associated with illness severity, worse functioning status and cognitive impairment. Such deficits are possibly related to microstructural damage in the connective tracts of white matter. Diffusion tensor imaging (DTI) has been employed as a highly sensitive tool to investigate microstructural changes in white matter structure. Fractional anisotropy (FA) is the most commonly used parameter as it is the best estimate of fiber integrity as well as axonal and myelin degeneration. It has been reported an association with BMI in depressed BD patients, but not in euthymia nor in comparison with a control group (CTR). Therefore, our aim was to test the association between body mass index (BMI) and white matter integrity in euthymic patients with bipolar disorder (BD) and healthy controls (CTR) by diffusion tensor imaging study.

    Methods: One hundred and one individuals were assessed for magnetic resonance imaging, consisting in 35 patients with BD and 66 CTR. We acquired high-resolution volumetric T1 and T2 weighted, and DTI data on a 1.5 T Philips scanner. A linear regression models were performed to test the hypothesis controlling for age, sex and skull size.

    Results: BMI predicted fractional anisotropy (FA) of the cingulate gyrus terminal endings in individuals with BD (Left: r² = 0.235, t = -2.792, β = -0.455, p = 0.010, Right: r² = 0.265, t = -2.060, β = -0.329, p = 0.050), but not in CTR. When we analyzed all participants together including a BMI x group interaction, it explained the variation of the average cingulate gyrus FA better than BMI or group alone (r² = 0.217, t = -1.976, β = -1.090, p = 0.052).

    Conclusions: Obesity predicted white matter microstructural damage to the cingulate fibers in patients with BD during euthymia. The structural neurological damage we found in our sample associated with BMI is likely to be a systemic consequence of a pervasive pattern of inflammatory toxicity. The association between obesity and white matter microstructural damage was present regardless acute symptomatology, but only in patients, suggesting a cumulative effect of combined pro-inflammatory states.

    Keywords: Bipolar Disorder, Obesity, White Matter, DTI

    Disclosure: Daiichi-Sankyo, Advisory Board, Janssen-Cilag, Advisory Board

    M96. A Network Model of Resting-State Brain Connectivity Predicts Heterogeneous Symptoms of Major Depression

    Yu Meichen, Kristin A. Linn, Russell Shinohara, Mary Phillips, Desmond Oathes, Maria Oquendo, Melvin McInnis, Maurizio Fava, Madhukar Trivedi, Myrna Weissman, Yvette Sheline*

    University of Pennsylvania, Philadelphia, Pennsylvania, United States

    Background: Major depressive disorder (MDD) is characterized by heterogeneous symptoms, the neurobiological basis of which is still largely unknown. Brain network studies have consistently reported disruptions of several resting-state networks (RSNs) in MDD patients, including hypoconnectivity in the frontoparietal network (FPN), and hyperconnectivity in the default mode network (DMN) and between the DMN and FPN. In this study we aimed to identify the correlations between network abnormalities and symptoms using canonical correlation analysis (CCA), a powerful multivariate correlation approach aiming to seek maximal correlations between two groups of variables.

    Methods: We used a multi-site sample of patients from the EMBARC dataset (n = 189 MDD: n = 39 demographically matched healthy controls). Unmedicated participants were recruited from four clinical sites: Columbia University (CU), Massachusetts General Hospital (MGH), the University of Texas Southwestern Medical Center (TX), and the University of Michigan (UM). All participants provided written informed consent and the institutional review boards from the four clinical sites approved all study procedures. Both patients and healthy individuals were administered the Structured Clinical Interview for DSM-IV-TR, the Hamilton Depression Rating Scale (HAMD), Quick Inventory for Depression Symptomatology (QIDS), Mood and Anxiety Symptom Questionnaire (MASQ), Concise Associated Symptoms Tracking (CAST), Concise Health Risk Tracking (CHRTP-propensity), Concise Health Risk (CHRTP-risk), Childhood Trauma Questionnaire (CTQ), Snaith-Hamilton Pleasure Scale (SHAPS), NEO-Five Factor Inventory (NEO) and Speilberger State Anxiety Inventory (STAI). T1-weighted (T1) images were processed using the ANTS Cortical Thickness pipeline and two resting-state fMRI runs were collected for each participant. Time series data from each participant were processed using the XCP Engine, which uses an optimized confound regression procedure to reduce the influence of subject motion. After standard processing we ran a pipeline for statistical harmonization of data collected from multiple sites to correct for site effects. We then computed within and between network metrics for all of the standard Power networks, resulting in 55 variables. We clustered item level clinical symptoms across the 216 items, producing 4 different clusters. Using CCA, we correlated the 55 network variables with the 4 clinical symptom clusters.

    Results: Within-network connectivity was decreased in several task-positive networks, including the FPN, the dorsal attention network (DAN) and cingulo-opercular network (CON), but increased in the DMN and salience network (SAN). The between-network connectivity was increased in DAN, FPN, DMN and SAN, particularly in the DAN and FPN. The 216 item-level clinical data were clustered into 4 types of depression symptoms: a depression composite (general mood symptoms, anhedonia, anxiety, neuroticism, and suicidal tendency), positive traits (extraversion, openness, agreeableness, conscientiousness and positive childhood experiences), emotional neglect/abuse, and physical abuse. We detected a CCA mode, which linked these depression symptoms to specific brain network connectivity: 1) the depression composite was mainly correlated with between network connectivity of DAN-VAN; 2) emotional neglect/abuse was mainly correlated with between-network connectivity involving the FPN and DAN; 3) positive traits were specifically associated with the network connectivity in subcortical areas and the CON.

    Conclusions: Overall, our study detected associations between functional connectivity and behavioral measures. It proposes a predictive network model to bridge the heterogeneous symptoms and abnormal network architecture in MDD. RSN connectivity patterns may assist in the interpretation of the underlying neurobiological basis for MDD symptoms and serve as a useful biomarker in understanding the disease.

    Keywords: fMRI, Resting State Networks, Canonical Correlation Analysis (CCA)

    Disclosure: Nothing to disclose.

    M97. Sex Hormones, Resting Regional Cerebral Blood Flow With PET, and Resting Functional Connectivity With fMRI in Healthy Women and Women With PMDD

    Shau-Ming Wei*, Erica Baller, Pedro Martinez, Katherine Reding, Philip Kohn, J. Shane Kippenhan, Michael Gregory, Angela Ianni, David Rubinow, Peter Schmidt, Karen Berman

    National Institute of Mental Health, Bethesda, Maryland, United States

    Background: Basic and clinical studies demonstrate that the sex hormones estradiol and progesterone modulate neurocircuitry related to cognition and affective regulation in healthy women and those with premenstrual dysphoric disorder (PMDD). PMDD, a prevalent and debilitating condition that affects 3-15% of women of reproductive age, refers to the appearance of mood and behavioral symptoms in the luteal (post-ovulatory/premenstrual) phase of the menstrual cycle. Abnormal behavioral responsivity to the natural fluctuation of ovarian hormones is implicated in the pathophysiology of PMDD, but the neural substrates of this differential behavioral sensitivity remain poorly characterized. In this study, we used a six-month pharmacological hormone manipulation protocol in healthy women and women with PMDD to investigate the effects of estradiol and progesterone on two measures of resting brain function: regional cerebral blood flow (rCBF) with the oxygen-15 water positron emission tomography (PET) method and resting state functional connectivity (rs-FC) with fMRI.

    Methods: Forty-three healthy, asymptomatic women and 20 women with PMDD underwent two eyes-open resting PET scans (10 mCi of oxygen-15 water/scan, GE Advance [Milwaukee] three-dimensional PET scanner [septa retracted, 4.25 mm slice separation, 35 slices, axial field of view 15.3 cm]) during each of three hormone conditions: ovarian suppression induced by the gonadotropin-releasing hormone agonist leuprolide acetate (Lupron), Lupron plus estradiol replacement, and Lupron plus progesterone replacement. After preprocessing, the two resting rCBF scans of each subject were averaged and entered into a voxel-wise random effects analysis to determine brain regions showing an interaction between diagnosis and hormone condition as well as main effects of hormones at a threshold of p<0.001, uncorrected.

    Additionally, to test for main effects of hormones in rs-FC, seven healthy, asymptomatic women underwent two rs-fMRI scans (GE 3 T scanner, TR/TE=2000/25 ms, 2.5×2.5×2 mm voxels) during the same hormone manipulation protocol. fMRI preprocessing included motion correction, normalization to MNI space, censoring of corrupted volumes, anatomic CompCor, and bandpass filtering. A subgenual cingulate region, which showed a diagnosis-by-hormone interaction in the PET resting rCBF analysis, was used as a seed ROI and a seed-based correlation analysis was performed at an exploratory threshold of p<0.005, uncorrected.

    Results: The resting PET rCBF analysis showed a significant interaction between diagnosis and hormone condition in the subgenual cingulate. Post-hoc analyses revealed that in the healthy cohort who did not have PMDD, there were no hormone condition-related differences in rCBF in this region. In contrast, in women with PMDD, we saw differential modulation by ovarian steroids in this subgenual cingulate region. Specifically, in the Lupron alone (hypogonadal) condition (when PMDD symptoms are typically in remission), resting rCBF was higher compared with estradiol or progesterone replacement (when recurrence of PMDD symptoms has been observed).

    The fMRI study of rs-FC in healthy women revealed several regions in the default mode network that showed differential connectivity with the subgenual cingulate across hormone conditions. These regions included the medial orbitofrontal cortex (mOFC), parahippocampus, and posterior cingulate. Specifically, FC of the subgenual cingulate with the parahippocampus and posterior cingulate was increased during both estradiol and progesterone replacements compared with Lupron alone. FC between the subgenual cingulate and the mOFC was increased during estradiol replacement and decreased during progesterone replacement compared with Lupron alone.

    Conclusions: Our results demonstrate that in PMDD, exposure to experimentally controlled, physiologically relevant levels of ovarian hormones differentially impact resting rCBF in the subgenual cingulate, a region implicated in the pathophysiology of affective disorders. In healthy women, subgenual cingulate rCBF as measured with a gold standard PET technique, was not altered by hormonal condition, but resting state fMRI revealed that functional connectivity of this region with areas in the default mode network was modulated by estradiol and progesterone. These data provide a framework for understanding the impact of ovarian hormones on the neural control of affective state and the propensity for mood destabilization in PMDD.

    Keywords: PET Imaging, Resting-State fMRI, Estradiol, Progesterone, Affective Disorders

    Disclosure: Nothing to disclose.

    M98. Dentate Gyrus Volume is Associated With Childhood Maltreatment and Depression Severity in Adolescents

    Theo van Erp, Robert Jirsaraie, Megan Faulkner, Kiana Scambray, Julia Fong, Derek Taylor, Uma Rao*

    University of California, Irvine, Irvine, California, United States

    Background: Early adverse experiences, including childhood maltreatment (MALTX), are a common risk factor for depression, accounting for up to 50% of the attributable risk for the disorder. Important clinical differences exist with respect to age of onset, symptoms, treatment response and clinical course between depressed individuals with or without MALTX, suggesting that there are two distinct subtypes. Neurobiological differences also exist, with documented differences in hippocampal volume between these two subtypes. We present initial findings from an ongoing study, examining the relationship between hippocampal subregion volumes and dimensional measures of MALTX and depression severity scores in adolescents. Based on prior studies in adults, we hypothesized significant negative relationships between dentate gyrus (DG), Cornu Ammonis 3 (CA3), or subiculum volumes and MALTX and depression severity.

    Methods: Twenty participants [mean age (SD) = 15.5 (1.5); 13 females] were recruited into a multi-modal imaging study examining the main and interactive effects of MALTX and depression on neural circuitry in adolescents. Hippocampal subregion volumes for DG, CA3 and subiculum were extracted from high-resolution T1-weighted and Turbo Spin Echo scans using Automated Hippocampal Subfield Segmentation (ASHS version 2.2.0-118). MALTX and depressive symptom assessments included the Childhood Trauma Questionnaire (CTQ), the Childhood Adversity Interview (CAI), and the Beck Depression Inventory (BDI) for which total scores were computed. Relationships between the regional volumes and symptom severity scores were assessed using linear regression (R’s gam), statistically controlling for age and sex. Initial exploratory analyses examining possible interactive effects of MALTX and depression were performed on all hippocampal subregion volumes [CA1, CA2, ectorhinal, entorhinal, perirhinal, and parahippocampal (PHC) cortexes].

    Results: There were significant negative associations between dentate gyrus volume and CAI and BDI scores. Associations between dentate gyrus volume and the CTQ score showed similar patterns but were not significant. Moreover, initial exploratory analyses showed a significant interaction between CAI and BDI scores on PHC volumes indicating the combination of MALTX and depression severity was associated with greater reduction in PHC volume compared to either condition alone.

    Conclusions: We found, to our knowledge for the first time, significant relationships among dentate gyrus volume, MALTX and depression severity in adolescents. The stronger relationships between dentate gyrus volume for the CAI compared to the CTQ may suggest that interview-based MALTX data (CAI) are more reliable than questionnaire-based data (CTQ). These findings are consistent with observations in adult humans and rodents. We also found initial evidence for interactive effects of MALTX and depression on PHC volumes. While the findings from these initial analyses will need to be confirmed as more study data are collected, they provide, to our knowledge, the first evidence that MALTX and depression are associated with aberrant dentate gyrus development and that they may have interactive effects on PHC development.

    Keywords: Depression, Maltreatment, Adolescent, Imaging, Hippocampus

    Disclosure: Nothing to disclose.

    M99. Cell -Specific Ablation of Microglial Pattern Recognition Receptors Rage and TLR4 Alters Susceptibility to Depressive-Like Behaviors After Chronic Unpredictable Stress

    Tina Franklin*, Ronald Duman

    Yale University School of Medicine, New Haven, Connecticut, United States

    Background: Chronic stress promotes dysregulation of the innate immune system leading to enhanced inflammatory signaling often associated with depressive symptomology. Growing evidence suggests that innate immune cells such as microglia, promote neuroinflammation in response to stress by releasing danger associated molecular pattern (DAMP) molecules leading to increased inflammatory signaling through their ligation to pattern recognition receptors (PRR) such as toll-like receptor 4 (TLR4) and the receptor for advanced glycation end products (RAGE). Our previous studies show that microglial RAGE is upregulated in response to chronic unpredictable stress (CUS) and enhanced microglial RAGE expression coincides with the onset and recurrence of stress-induced depressive-like behaviors. Most importantly, constitutive RAGE KO mice show partial attenuation of stress-induced behavioral effects. These novel findings suggest that stress-induced depressive-like behaviors are partially mediated by enhanced microglial DAMP signaling due to increased receptor availability. In this study, we hypothesize that microglial RAGE or TLR4 deletion will attenuate depressive-like behaviors following stress due to suppressed DAMP signaling. Moreover, we hypothesize that the combined deletion of both microglial RAGE and TLR4 in RAGE/TLR4 double knockout mice (RGTR-DKO) will enhance resilience against stress-induced depressive-like behaviors.

    Methods: To test this hypothesis, we generated RAGEfl/fl:CX3CR1CreERT, TLR4fl/fl:CX3CR1CreERT and RAGEfl/fl: TLR4fl/fl:CX3CR1CreERT mice and utilized tamoxifen-induced Cre recombinase system for cell-specific knockout of microglial RAGE, TLR4 or both, respectively. Cre negative animals are used as controls. Tamoxifen-induced conditional knockout (KO) mice were tested for cognitive, anxiety and depressive-like behaviors at baseline and after CUS exposure using novel object recognition (NOR), forced swim test (FST), open field test (OFT) and three chambers social interaction (SIT). Microglial morphology was assessed immediately following CUS exposure to determine if microglial PRR deficient mice display reduced microglial reactivity following chronic stress exposure compared to littermate controls.

    Results: Microglial TLR4 and RAGE/TLR4 (RGTR) deletion attenuated stress-induced depressive-like behaviors in adult male mice. Stress-induced microglial reactivity was reduced in TLR4 and RGTR knockout animals compared to WT controls. Interestingly, microglia-specific RAGE deletion induced a pro-depressive phenotype in stress-naïve animals. Moreover, microglial RAGE deletion neither enhanced nor prevented depressive-like behaviors following chronic stress exposure.

    Conclusions: These results suggest that microglial RAGE and TLR4 signaling have opposing roles in stress-induced depression risk. Microglial RAGE signaling appears to be important for the attenuation of stress-induced behavioral effects while microglial TLR4 signaling enhances vulnerability to depressive-like behaviors. Together, these data provide novel insights into the role of microglial RAGE and TLR4 signaling in stress-induced microglial reactivity and the development of depressive-like behaviors.

    Keywords: Depression Inflammation Cytokine, Depression-like Behavior, Innate Immunity, DAMPs, Pattern Recognition Receptors

    Disclosure: Nothing to disclose.

    M100. Association of Cerebral Intrinsic Activity With Symptomatology in Drug-free Patients With Major Depressive Disorder

    Xinyu Hu, Lianqing Zhang, Hailong Li, Xuan Bu, Weihong Kuang, John A. Sweeney, Qiyong Gong, Xiaoqi Huang*

    West China Hospital of Sichuan University, Chengdu, China

    Background: Recent neuroimaging studies of major depression (MDD) have highlighted the disrupted topological organization of large-scale functional and structural brain networks in depression (Gong and He, 2015). The relation of different regional brain alterations to different symptom dimensions of MDD remains to be fully characterized. We therefore explore the relations of altered functional cerebral intrinsic activity and symptom dimensions of depression in a relatively large sample of medication-free patients with MDD using regional measures of ALFF and functional connectivity (FC) analysis.

    Methods: The study was approved by the local ethical committee and written informed consent was obtained from all subjects. A total of 144 medication-free MDD patients and 86 age and sex matched healthy control subjects (HCS) were studied. Clinical symptoms were evaluated using the 17-item Hamilton Depression Rating Scale (HAMD).

    The MRI examinations were performed via a 3-Telsa MRI system (Trio; Siemens, Erlangen, Germany) with an 8 channel phase array head coil. The rs-fMR images were obtained via a gradient-echo echo-planar imaging sequence (TR/TE=2000/30msec, flip angle=90°, slice thickness=5 mm with no gap, 30 axial slices, 175 volumes in each run). Subjects were instructed to relax with their eyes closed without falling asleep during MR examination.

    Image preprocessing was performed using DPABI software (Version 2.3, Calculation of amplitude of low-frequency fluctuation (ALFF) maps was implemented in REST ( running under Matlab. Regions with abnormal ALFF values in MDD patients relative to HCS were treated as seed regions for functional connectivity studies. The statistical analyses of ALFF and seed-based FC maps between MDD patients and HCS were performed using SPM8 ( In all image analyses, statistical inferences were made at P < 0.05 using AlphaSim correction to correct for multiple comparisons.

    HAMD-17 items were assigned to the following subscales: ‘Core affective symptoms’ (items 1, 2, 7, 8, 10, 13), ‘Sleep’ (items 4, 5, 6), ‘Activity’ (items 7, 8), ‘Psychic Anxiety’ (items 9, 10), ‘Somatic Anxiety’ (items 11, 12, 13) and “Delusion”(items 2, 5) according to a previous study (Serretti et al, 1999). Correlation analysis was conducted to assess the relationship between these MDD symptomatology dimensions and averaged eigenvalues of brain regions with altered ALFF and associated FC group differences using the eigenvariate option in SPM8.

    Results: Relative to HCS, MDD patients showed lower ALFF in the posterior cingulate gyrus (PCC) and higher ALFF in bilateral orbital frontal cortex (OFC) (P < 0.05, with AlphaSim correction). Seed-based FC analysis revealed elevated functional connectivity between PCC and bilateral dorsal cerebellum and between bilateral OFC and bilateral anterior putamen (AP) in the MDD group, while FC between PCC and medial prefrontal cortex was decreased in MDD patients.

    ALFF in the PCC was negatively correlated with the HAMD total score (r = -0.194; P = 0.02), the “sleep” subscale score (r = -0.208; P = 0.013) and item 16 score for weight loss (r = -0.208; P = 0.013). The increased FC between PCC and the right dorsal cerebellum was positively correlated with “delusion” subscale score (r=0.210, P=0.012).

    Conclusions: This study added new information about regional changes of intrinsic neural activity in MDD, and it associated changes in FC. In addition, we found association of specific symptoms of depression with brain intrinsic activity in PCC, OFC, cerebellum and AP. One important specific finding was the demonstration of a dissociation of brain activity between OFC and PCC in the MDD patients. The alteration of cerebellar function which correlated with “delusion” subscale suggest its involvement in the cognitive aspects of MDD.

    Keywords: Major Depressive Disorder (MDD), Resting-State fMRI, Symptomatology, Functional Connectivity, ALFF

    Disclosure: Nothing to disclose.

    M101. Classifying Suicidal Behavior Using Resting State Functional Connectivity and Structural Neuroimaging: A Machine Learning Approach

    Savannah Gosnell, Christopher Fowler, Ramiro Salas*

    Baylor College of Medicine, Houston, Texas, United States

    Background: About 80% of patients who commit suicide do not report suicidal ideation the last time they speak to their mental health provider, highlighting the need to identify biomarkers of suicide behavior. Our objective was to identify neural suicide behavior biomarkers for building a model to classify psychiatric inpatients.

    Methods: Suicidal and non-suicidal psychiatric inpatients were recruited over 4.5 years as a part of this case-control study. Eighty percent of the sample was used to determine significant differences in structural and resting state functional connectivity measures throughout the brain. These measures were used in a random forest classification model, which was tested on the remaining 20%.

    Inpatients were recruited from the Menninger Clinic, a psychiatric hospital, and neuroimaging was performed close to admittance at the Core for Advanced MRI at Baylor College of Medicine in Houston, Texas.

    Inpatients (n=423 of ~600 approached, not excluded for any psychiatric disorder, with no contraindication for neuroimaging and deemed mentally stable enough for MRI) were included if they met suicide criteria determined using the Columbia Suicide Severity Rating Scale (based on previous attempt and current ideation) or control criteria (no previous suicide attempt and no lifetime ideation). Participants had at least one psychiatric disorder (mainly major depression, anxiety, personality, and substance use disorders), and were predominantly Caucasian.

    Results: The study classified suicidal (n=63, age=30.56 (11.81) years, males=40.5%) and non-suicidal psychiatric inpatients (n=65, age=34.33 (13.3) years, males=64.2%). The model built on 80% of these inpatients resulted in sensitivity=79.4% and specificity=72.3% using a random forest model. This was replicated in an independent sample (n=32) with sensitivity=81.3% and specificity=75.0%.

    From the 57 features that showed significant differences between the two groups and were used for machine learning (in the 80% training sample), 6 were replicated in the 20% testing sample (including resting state functional connectivity between frontal and insular/putamen, habenula and parahippocampal, and amygdala and temporal regions).

    Conclusions: Neuroimaging (an unbiased biomarker) can be used to classify suicidal behavior in psychiatric inpatients without observing any clinical features. In the future, clinicians may use neuroimaging to inform the determination of suicide risk.

    Keywords: Suicide Prediction, Machine Learning Classification, Resting State Functional Connectivity, Habenula

    Disclosure: Nothing to disclose.

    M102. Gender-Based Differences in the Association of C-Reactive Protein (CRP) With Resting State Brain Connectivity in Major Depressive Disorder: Findings From the EMBARC Study

    Cherise Chin Fatt*, Manish Jha, Abu Minhajuddin, Brittany Mason, Tracy Greer, Madhukar Trivedi

    University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, United States

    Background: Elevated levels of c-reactive protein (CRP), an inflammatory biomarker, have been associated with reduced connectivity of striatum with other brain regions in previous neuroimaging studies. While gender significantly moderates the association of peripheral inflammation with depressive symptoms severity, it’s effect on the association of CRP with functional connectivity remains unknown.

    Methods: Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) study participants with plasma samples and completed rsfMRI scans available at baseline (total n=204, male n=67, female n=137) were included. Using magnetic resonance imaging, resting-state functional connectivity were computed after parcellating cortical (pacel n=100) and subcortical (parcel n=21) regions based on previously published atlas. Plasma CRP levels (log-transformed due to skewed distribution) were measured with commercially-available ELISA kits. Separate linear regression analysis was conducted for each connectivity pair (connectivity pairs n=7260) as dependent variable, gender-by-logCRP interaction as primary independent variable, and ethnicity, race, site, body mass index, and age as covariates. Gender-stratified analyses were used to interpret the significant interaction.

    Results: Gender significantly (all p<0.05 after false discovery rate correction) moderated association of logCRP with connectivity of bilateral striatum with left posterior hippocampus and of superior medial frontal region (default mode network) with orbital middle frontal region (executive control network). In all these connectivity pairs, higher levels of CRP were associated with reduced connectivity in females (r= -0.08 to -0.13) but increased connectivity in males (r= 0.37 to 0.48).

    Conclusions: Gender is an important biological factor that differentially affects the association of peripheral inflammation with alterations in resting state brain connectivity.

    Keywords: Resting State Functional Connectivity, CRP, Major Depression Disorder

    Disclosure: Nothing to disclose.

    M103. Quantification of Norepinephrine Transporter Occupancy of Venlafaxine Extended-Release in Patients With Major Depressive Disorder Using Positron Emission Tomography and Radioligand [18 F]FMeNER-D2

    Ryosuke Arakawa, Per Stenkrona, Akihiro Takano, Jonas Svensson, Max Andersson, Sangram Nag, Yuko Asamin, Yoko Hirano, Christer Halldin, Johan Lundberg*

    Karolinska Institutet, Stockholm, Sweden

    Background: Venlafaxine ER is an extended-release (ER) formulation of venlafaxine, with documented clinical effect for treatment for depression. It is categorized as combined serotonin (5-HT) and norepinephrine (NE) reuptake inhibitor (SNRI) based on the in vitro affinity data of 5-HT transporter (5-HTT) and NE transporter (NET) (Ki 82 nM and 2480 nM, respectively) [1]. Due to the large discrepancy of the affinity between 5-HTT and NET, the in vivo effect of clinical doses of venlafaxine on NE reuptake has been questioned, especially in the lower dose range. While 5-HTT occupancy of clinically relevant doses of venlafaxine in the human brain has been reported previously [2,3], there has been no report of in vivo NET occupancy in the human brain.

    [18 F]FMeNER-D2 is a positron emission tomography (PET) radioligand which binds reversibly and selectively to NET. A method for reliable quantification of [18 F]FMeNER-D2 binding in human brain has been described [4]. The applicability of [18 F]FMeNER-D2 PET in clinical occupancy studies has been reported previously.

    The aim of this study was a) to verify that clinically relevant doses of venlafaxine occupies NET in the living human brain, and b) to identify the relationship between oral dose and NET occupancy in patients with major depressive disorder (MDD) using PET and [18 F]FMeNER-D2.

    Methods: This was an open-label, single center, exploratory PET study. Eleven MDD patients (age range, 22 - 65 y; mean ± SD, 37.8 ± 12.1; six females, five males) who had responded to venlafaxine ER treatment were recruited. The dose of venlafaxine ER had been fixed for at least 2 weeks before the PET measurement. The subjects did not take any other antidepressants and psychotropic agents or any other medication that might influence 5-HT and NE transmission for at least 4 weeks before the PET measurement. Nine control subjects (20 - 62 (39.9 ± 14.4) y; three males, six females) were recruited. No medications were allowed for at least 2 weeks before the PET measurement. The study was approved by the Regional Ethical Review Board in Stockholm, Sweden, and the Radiation Safety Committee at the Karolinska University Hospital Solna in Stockholm, Sweden. After thorough oral and written information of the study, and before any study related activity took place, written informed consent was obtained from all participants.

    Each subject participated in one PET measurement with [18 F]FMeNER-D2. For MDD patients, the last administration of venlafaxine ER was approx. 5 h before the radioligand injection. The emission data was collected from 120 minutes to 180 minutes after the radioligand injection using a HRRT system. T1-weighted magnetic resonance imaging (MRI) was also performed for the anatomical reference for PET images.

    For all MDD patients, venous blood samplings were performed to measure the plasma concentration of venlafaxine and O-desmethylvenlafaxine. Total active moiety as sum of both compounds was chosen as the parameter of interest, because O-desmethylvenlafaxine also has affinity for both 5-HTT and NET (Ki: 40.2 nM and 3385 nM, respectively) [5].

    Anatomical regions of interest (ROIs) were delineated on the reoriented MRI image using the Automated Anatomical Labelling (AAL) template. [18 F]FMeNER-D2 binding potential (BPND) was quantified by the area under the curve (AUC) ratio method with thalamus as target and caudate as reference regions: AUCthalamus / AUCcaudate – 1. NET occupancy (%) was calculated as (BPND:baseline – BPND:treatment) / BPND:baseline × 100. The mean BPND of control subjects was used as BPND:baseline for all the MDD patients.

    Results: Injected radioactivity was 278 ± 47 MBq (mean ± SD) and 268 ± 53 MBq for MDD patients and control subjects, respectively. Molar radioactivities was 66 ± 28 GBq/µmol and 68 ± 32 GBq/µmol at time of injection. Injected mass was 1.6 ± 0.6 µg and 1.5 ± 0.6 µg. There was no statistical difference between MDD patients and control subjects (Student’s t-test: p = 0.68 – 0.88).

    Daily doses of venlafaxine ER varied between 37.5 and 225 mg/day, plasma concentrations of total active moiety at time of PET between 108 and 947 ng/mL, and NET occupancy for all MDD patients in thalamus between 17 and 53 %. The NET occupancy increased in a dose and concentration dependent manner.

    Conclusions: This study demonstrates for the first time that clinically relevant doses of venlafaxine ER block the NET in the living human brain. The present data is thus in line with the notion that venlafaxine ER exert its clinical effect via blockade of not only 5-HTT but also NET.

    Furthermore, NET occupancy was identified already at 37.5 mg oral dose (significantly lower BPND compared to control group; p<0.01). This indicates that NET occupancy may be a clinically relevant mechanism of action of venlafaxine already in low clinical doses.

    Results from clinical trials suggest that the clinical effect of a combination of 5-HTT and NET blockade is superior to 5-HTT blockade only [6,7]. Our data confirm that this mechanism may indeed be present in MDD patients treated with venlafaxine.

    Keywords: PET, serotonin and norepinephrine reuptake inhibitor, Major Depressive Disorder (MDD), [18 F]FMeNER-D2

    Disclosure: Nothing to disclose.

    M104. Treatment Effects of Lithium Monotherapy on Serotonin Transporter and Serotonin-1A Binding & Prediction of Clinical Response in Bipolar Depression

    Ramin Parsey*, Mala Ananth, Elizabeth Bartlett, Xuejing Lin, Laura Kunkel, Nehel Vadhan, Gregory Perlman, Michala Godstrey, Daniel Holzmacher, Chuan Huang, R. Todd Ogden, Christine DeLorenzo

    Stony Brook University Medical School, Stony Brook, New York, United States

    Background: Abnormalities of the serotonergic system are thought to play a critical role in mood disorders including bipolar depression (BPD). Lithium is the first line of therapy and one of the few effective treatments in BPD. It has been hypothesized that lithium works by enhancing serotonergic transmission (Carli et al. 1997, Carli and Reader 1997, Bauer et al. 2003). Whether these mechanisms are via increases in serotonin (5-HT) release (through the 5-HT1A receptor) or changes to the 5-HT transporter (5-HTT) uptake remain to be seen. Preclinical studies have shown increases in 5-HTT density (Bmax) in cortical regions following chronic lithium treatment (Carli and Reader 1997), but not in subcortical regions known to be affected in BPD and major depression, including midbrain and amygdala (Oquendo et al. 2007, Miller et al. 2013, Miller et al. 2016). Rodent studies also show that lithium monotherapy decreases 5-HT1A receptor densities in the frontal cortex and hippocampus (Hotta et al. 1986, Mizuta and Segawa 1988, Odagaki et al. 1990, Bauer et al. 2003), but not in the dorsal raphe nucleus (McQuade et al. 2004), suggesting that lithium treatment has a specific effect on post-synaptic 5-HT1A receptors, but not somatodendritic (raphe) 5-HT1A receptors. Here we examined the effects of lithium monotherapy on 5-HTT and 5-HT1A in vivo binding potential in BPD. Additionally, we investigated whether there was a relationship between change in 5-HTT or 5-HT1A binding following lithium treatment and clinical response. Finally, we investigated the potential of 5-HTT or 5-HT1A to predict post-treatment response. We hypothesized that: a) lithium monotherapy would increase 5-HTT cortical binding towards control levels and would decrease 5-HT1A binding post-synaptically; b) lithium-induced increases in 5-HTT or decreases in 5-HT1A binding would associate with better clinical response; c) lower pretreatment 5-HTT binding or higher 5-HT1A binding would predict better clinical response.

    Methods: 19 medication-free patients with BPD currently in a depressive episode were imaged using the 5-HTT tracer [11 C]DASB and the 5-HT1A tracer [11 C]CUMI-101. BPD subjects then received eight weeks of standardized monotherapy with lithium and were scanned a second time following treatment with both tracers. Scans were acquired on a Siemens ECAT HR+ scanner for 90 minutes ([11 C]DASB) or 120 minutes ([11 C]CUMI-101) after tracer injection. Arterial blood samples were collected to calculate the metabolite-corrected arterial input function, or a validated simultaneous estimation algorithm was used to compute the arterial input function using a single venous or arterial blood sample (Ogden et al. 2010). Time activity curves were fit with likelihood estimation in graphical analysis (LEGA) (Ogden et al. 2007) to calculate [11 C]DASB total distribution volume (VT), from which the outcome measure VT/fP was calculated using the tracer plasma free fraction fP, as it has been shown that no brain region is completely devoid of [11 C]DASB binding (Parsey et al. 2006). LEGA was also used to obtain estimates of [11 C]CUMI-101 BPF (Bmax/KD)(Milak et al. 2010). Lithium treatment response was determined using the 24-item Hamilton Depression Rating Scale (HDRS-24) (Hamilton 1960). In [11 C]DASB linear models, midbrain, amygdala and anterior cingulate cortex were examined as a priori regions of interest. In [11 C]CUMI-101 linear models, the raphe nucleus was examined a priori and then 12 post-synaptic regions were considered (Sullivan et al. 2009). The parcellation of these regions are based on separately acquired MRI images.

    Results: The data revealed no significant differences between pre- and post-lithium treatment scans using VT/fP [11 C]DASB binding, or BPF [11 C]CUMI-101 binding (both p>0.05). Additionally, we found no significant relationship between change in 5-HTT binding or change in 5-HT1A binding pre-to-post lithium treatment and treatment response (both p>0.05). We found a significant relationship between both pretreatment 5-HTT [11 C]DASB (all regions p=0.003-0.015) and 5-HT1A [11 C]CUMI-101 (amygdala p=0.016; hippocampus p=0.015; parahippocampal gyrus p=0.013; temporal cortex p=0.039) binding and post-lithium treatment clinical response, where lower pretreatment binding predicted improved clinical response.

    Conclusions: To our knowledge, this is the first report of the effects of lithium on the serotonergic system, in vivo. Our findings indicate that lithium did not significantly alter 5-HTT or 5-HT1A binding post-treatment, and that there was no relationship between change in binding and clinical response for either tracer. These findings suggest that lithium does not act directly on 5-HTT or 5-HT1A to mediate treatment response. Interestingly, we find that pretreatment changes in the serotonergic system are predictive of clinical response following lithium treatment compounding the serotonergic hypothesis of mood disorders and suggesting lithium may act via other serotonergic proteins (e.g. 5-HT1B (Riad et al. 2000, Li et al. 2004, Polter and Li 2011)) to mediate treatment response. Future studies with tracers that target other serotonergic proteins and increased sample sizes are needed to further corroborate these findings and better understand the effect of lithium on the serotonergic system.

    Keywords: PET Imaging, Bipolar Disorder, lithium treatment

    Disclosure: Nothing to disclose.

    M105. Ultra-High Field MRI of Mood-Related Circuit Disturbances in Depression: A Systematic Comparison Between 3-Tesla and 7-Tesla

    Laurel Morris, Prantik Kundu, Sara Costi, Abigail Collins, Molly Schneider, James Murrough*, Gaurav Verma, Priti Balchandani

    Icahn School of Medicine at Mount Sinai, New York, New York, United States

    Background: Ultra-high field 7-Tesla (7 T) MRI has the potential to advance our understanding of neuropsychiatric disorders, including major depressive disorder (MDD). To date, few studies have quantified the advantage of resting state functional MRI (fMRI) at 7 T compared to 3-Tesla (3 T).

    Methods: We conducted a series of experiments that systematically quantified the improvement in temporal signal-to-noise ratio (tSNR) of a multi-echo multi-band fMRI protocol with ultra-high field 7-Tesla, compared to 3-Tesla MRI in healthy controls (HC). We also directly tested the enhancement in ultra-high field 7-Tesla fMRI signal power by examining the ventral tegmental area (VTA), a small midbrain structure that is critical to the expected neuropathology of MDD but difficult to discern with standard 3-Tesla MRI.

    Results: We demonstrate 200-300% improvement in tSNR and resting state functional connectivity metrics provided by ultra-high field 7-Tesla fMRI compared to the same imaging protocol with 3-Tesla, indicating enhanced power for detection of functional architecture. A multi-echo based acquisition protocol and signal denoising pipeline afforded greater gain in signal power at ultra-high field compared to classic acquisition and denoising pipelines. Furthermore, ultra-high field fMRI revealed mood-related neurocircuit disturbances in patients with MDD compared to HC, which weren’t detectable with 3-Tesla fMRI.

    Conclusions: Ultra-high field 7 T fMRI may provide an effective tool for studying functional neural architecture relevant to MDD and other neuropsychiatric disorders.

    Keywords: Human Neuroimaging, 7-Tesla, Depression, Ventral Tegmental Area (VTA)

    Disclosure: Nothing to disclose.

    M106. Impaired Brain Insulin Signaling Moderates Antidepressant Response to Ketamine

    Susannah Tye*, Joshua Price, Mark Frye, Kathryn Cullen, William Bobo

    Queensland Brain Institute, St Lucia, Australia

    Background: Low-dose ketamine is effective in a subset of patients with treatment-resistant depression. Using an animal model of antidepressant-resistance we have identified an association between brain insulin signalling and antidepressant-like behavioural responses to ketamine in the forced swim test. The mechanistic role of insulin signalling in ketamine’s mechanism of action and its potential for use as a clinical biomarker remains to be determined.

    Methods: Rodent Studies: Wistar rats (6-8 wks) were pre-treated with saline (0.9%/d;14d) or adrenocorticotropic hormone-(1-24) (ACTH 100 µg/d;14d; n=64). Animals received ketamine hydrochloride (10 mg/kg) or vehicle saline (0.9%) 1 h prior to behavioral tests (open field test (OFT) and forced swim test (FST). All animals were euthanized 90 mins following treatment. Brain, plasma and peripheral blood mononuclear cells (PBMCs) were collected and snap frozen on dry ice. Brain and PBMC tissue from these animals was harvested and cultured for subsequent ketamine, insulin and glucose assays. Brain and PBMC cell cultures were exposed to ketamine and/or insulin (10 µg, 5 mins) ex vivo and change in mammalian target of rapamycin (mTOR) expression and glucose uptake was determined. Total and phosphorylated levels of insulin signalling pathway proteins in prefrontal (PFC) brain and PBMC tissues was quantified using standard ELISA and western blot procedures. Human studies: Adolescent (n=13) and adult patients (n=12) with treatment resistant depression received serial infusions of ketamine (0.5 mg/kg infused over 100 mins in adults and over 40 mins in adolescent subjects). PBMCs were collected immediately prior to and following the first ketamine infusion. As for animal studies, cells were stimulated ex vivo with insulin (10 µg, 5 mins) and change in mTOR expression was determined using ELISA. Insulin-mediated changes in mTOR were compared between individuals that achieved remission following serial ketamine infusions and those that did not.

    Results: Results demonstrated that ketamine reduced FST immobility in a subset of tricyclic antidepressant-resistant animals (p<0.05). Ketamine’s antidepressant actions in the FST correlated with levels of phosphorylated insulin signaling proteins (Akt: p<0.05; mTOR: p<0.01; and GSK3: p<0.01) in PFC and PBMC tissue of ACTH-treated animals, but not saline treated controls. Correlations between brain insulin signalling pathway protein activation and blood glucose levels was also observed (mTOR: p<0.05; and GSK3: p<0.05). Further, we demonstrated that changes in blood glucose and phosphorylated levels of mTOR and GSK3 in WBCs differentiated ACTH pre-treated ketamine responders from non-responders (p<0.05). Cell culture assays confirmed a ketamine acted directly on brain and PBMC tissue to increase glucose uptake and stimulate release of insulin, concurrent with activation of insulin signalling pathway proteins (p<0.05). In human subjects, the ex vivo insulin stimulation assay further demonstrated that change in PBMC mTOR levels following 5 mins insulin exposure after the first ketamine infusion was significantly greater in both adolescent (p<0.05) and adult (p<0.05) individuals that remitted following serial ketamine infusions. No difference was observed in mTOR levels following exposure to insulin in PBMCs of either adolescent or adult subjects for samples collected immediately prior to initiation of ketamine treatment.

    Conclusions: These findings suggest the antidepressant-like effects of ketamine in treatment-resistant individuals is associated with direct modulation of insulin signalling and glucose uptake. The direct effects of ketamine on insulin signaling and glucose uptake may serve as critical co-regulators of antidepressant response in individuals with dysregulation of metabolic responses. Importantly, cell signaling responses to insulin following first exposure to ketamine may serve as a valuable predictor of long-term clinical response.

    Keywords: Ketamine, Treatment Resistant Depression, Insulin, mTOR

    Disclosure: Nothing to disclose.

    M107. Effects of Early Life Adversities on Acetylcarnitine Deficiency and Insulin Resistance in Patients With Major Depression

    Carla Nasca*, Benedetta Bigio, Katie Watson Lin, Josh Dobbin, Marin Kautz, Ashly Cochran, Thalia Robakis, Alison Myoraku, James Beasley, Tonita Wroolie, Sarah Young, Aleksander Mathe, James Kocsis, James Murrough, Francis Lee, Bruce McEwen, Natalie Rasgon

    Rockefeller University, New York, New York, United States

    Background: The lack of biomarkers to identify target populations greatly limits the promise of precision medicine for major depressive disorder (MDD), a primary cause of disability and risk factor for suicide. Early life adversities (ELA), such as childhood trauma, are risk factors for poor mental health outcomes in adulthood. The biological pathways leading from risk to outcomes are yet-to-be fully elucidated. The endogenously produced molecule acetyl-L-carnitine (LAC) is an epigenetic modulator of central glutamatergic function and a candidate biomarker of insulin resistance (IR). In rodents with depressive-like traits and peripheral IR, LAC levels are markedly decreased and signal decreased acetylation of histones underlying transcriptional regulation of the metabotropic glutamate receptors mGluR2 in the ventral hippocampus and corresponding dendritic plasticity. We evaluated here a role of LAC and IR in in patients with MDD and history of ELA.

    Methods: Study participants were recruited at three independent sites, the Weill Cornell Medicine, the Mount Sinai Icahn School of Medicine and Stanford University. Plasma distribution of LAC and the internal control free-carnitine were determined in 45 healthy controls (HC) and 116 patients with MDD using UPLC-MS/MS and ESI-MS/MS; insulin resistance and sensitive was assessed by fasting plasma glucose (FPG) and insulin (FPI) levels, body mass index (BMI), weight and HOMA, Matsuda index and glucose and insulin responses to oral glucose challenges. Depression severity was assessed with the HDRS-21. Physical, sexual and emotional abuses as well as physical and emotional neglect were assessed with the Childhood Trauma Questionnaire. Two-tailed t-tests, chi-square, Pearson correlations and multiple regression were used as appropriate to specific analyses.

    Results: LAC (and not free-carnitine) is significantly lower in patients with MDD compared to HC (p<0.0001, power 0.99, effect size=0.8), independently of psychotropic drug treatment, as we recently reported. Within the group of patients with MDD, the LAC deficiency was greater with stronger severity, earlier disease onset and history of treatment resistant depression, which was associated with childhood trauma. Our new data in subjects screened for insulin resistance and sensitivity, show that IR was worst with higher reported rates of childhood trauma in that patients who reported higher rates of specific types of childhood trauma had higher levels of FPI and HOMA (FPI: r=0.6, p=0.004; HOMA: r=0.52, p=0.004).

    Conclusions: Our new translational findings suggest that a LAC deficiency and IR may define metabolic endophenotypes of MDD with a vulnerability memory that may be tracked back to ELA. Identifying the role of ELA on metabolic risk factors contributing to behavioral states in patients with MDD can lead to a mechanistic framework for development of a precision medicine approach for preemptive tailored interventions based upon endogenous drug-targets. Studies leveraging the use of brain-derived exosomes are ongoing to evaluate the contribution of central metabolic functions on the peripheral deficiency in LAC and IR in these metabolic endophenotypes of MDD.

    Supported by the Robertson Therapeutic Development Foundation (RTDF), the Hope for Depression Research Foundation (HDRF) to CN and BMC and the Pritzker Neuropsychiatric Disorders Research Consortium, which is supported by the Pritzker Neuropsychiatric Disorders Research Fund L.L.C, and 1R21 MH093948-01A1 (SPO #50260) and ADA7-09-CT-50 (SPO #43365) to NR.

    Keywords: Metabotropic Glutamate Receptor 2 (mGluR2), Epigenetic, Childhood Trauma, Insulin Resistance, Biomarker

    Disclosure: Nothing to disclose.

    M108. Differential Gene Expression in Response to Estradiol Withdrawal in Perimenopausal Depression

    Sarah Rudzinskas*, Jessica Hoffman, David Rubinow, David Goldman, Peter Schmidt

    National Institute of Mental Health, Rockville, Maryland, United States

    Background: Perimenopause, the transition women undergo prior to menopause, has been linked to an increased incidence of depression. Women in the perimenopausal period are 1.5-3x more likely to be diagnosed with major depressive disorder or develop depressive symptoms compared with women who have not entered perimenopause, or women several years past menopause (Cohen et al, 2006). Further, there is evidence that suggests estradiol withdrawal is linked to Perimenopausal Depression (PMD), as 5-10% of women develop depressive symptoms after discontinuing hormone therapy (Ockene et al, 2005). Additionally, peri/postmenopausal women with even minor depression are at an increased risk of cardiovascular mortality (Wassertheil-Smoller et al, 2004). Clinical studies show both the therapeutic benefits of estradiol (E2) in perimenopausal depression (PMD) (Schmidt et al, 2000, Soares et al, 2001) and the symptom-provoking effects of E2-withdrawal in women with past PMD, which are not experienced by those without past PMD (Schmidt et al, 2015). It has been suggested that a heightened sensitivity to changes in ovarian steroids such as E2 may contribute to the onset of PMD.

    Methods: We created lymphoblastoid cell lines (LCLs) derived from blood samples from women with a past PMD (n=8), or asymptomatic controls (AC) (n=9). These LCLs were examined in 3 different experimental conditions: 1) vehicle-treated media, 2) E2-treated media, or 3) E2-treated media which was changed to vehicle-treated media and collected 24 h later, to mimic E2-withdrawal on a cellular level. Levels of E2 in cell culture media were confirmed using High Performance Liquid Chromatography/Tandem Mass Spectrometry. Cells were collected and examined for changes in gene expression levels using whole-transcriptome RNA sequencing on polyAAA selected mRNA. EDGE-R analysis of differential gene expression was used to detect significant transcript expression changes between women with PMD and AC in all three treatment conditions. DAVID, GSEA, and other pathway analyses were applied to the results. Quality control measures and unsupervised clustering analyses were performed for quality control in CLC Bio 11.

    Results: We hypothesized that the differential affective/behavioral responsivity to E2-withdrawal in PMD could be observed on a cellular level. In support of this hypothesis, there were 534 Differentially Expressed Genes (DEGs) in LCLs between AC women and women with PMD after E2-withdrawal with an uncorrected p <0.05. Of these DEGs, the gene CXCL10 which has been previously linked to cardiovascular disease, is significantly upregulated (p<1.55×10-5) in the cells of women with past PMD and had the most extreme increase in transcription in the E2-withdrawal treatment condition. In contrast, a gene coding an enzyme CYP7B1, which is responsible for the metabolism of the steroids DHEA and pregnenolone, is also significantly upregulated (p<5.6×10-4) in PMD, but E2-treatment or withdrawal had no further effect on transcript expression. Furthermore, network analysis with GSEA and DAVID has revealed that there are several molecular pathways that appear to be differentially altered in women with PMD.

    Conclusions: Results support the hypothesis that the differential responsivity to E2-withdrawal in PMD could be linked to dynamic gene expression changes on a cellular level. These findings also suggest a possible contributor to the increased cardiac risk associated with depression and menopause. Current data may suggest that both intrinsic genetic differences as well as differential sensitivity to E2-withdrawal could underlie the behavioral symptomology of PMD. Further studies are underway to replicate these changes in DEG findings in an independent cohort of AC and PMD cases.

    Keywords: Menopause, Neuroendocrinology, Late-life Depression, Estradiol, Perimenopausal Depression

    Disclosure: Nothing to disclose.

    M109. Electrical Field Based Prediction of ECT Induced Clinical Effect: A Multisite Longitudinal Neuroimaging Study

    Miklos Argyelan*, Leif Oltedal, Benjamin Wade, Zhi-De Deng, Georgios Petrides, Anil Malhotra, GEMRIC Consortium, Christopher Abbott

    Hofstra Northwell School of Medicine, Glen Oaks, New York, United States

    Background: Treatment-resistant depression (TRD) is one of the leading causes of disability worldwide (Kessler et al. 2005; Rush et al. 2006). To date, electroconvulsive therapy (ECT) remains the most effective approach for the treatment of TRD, as well as the most established neuromodulatory technique (UK ECT Review Group 2003; Fink & Taylor 2007). Despite intensive research, however, the mechanism of action for ECT remains unknown. No prediction of ECT-induced efficacy or side effects is possible at the individual level. Thus, identifying target mechanisms would not only improve current deployment of bioelectric approaches (ECT, TMS, tDCS, DBS) as part of a precision medicine approach but could also lead to the development of novel therapies (Insel 2014).

    One interesting and not well-understood aspect of ECT is its direct electrical effect on the brain. It is thought that ECT induced seizure is the necessary component of the treatment effect, however recently this dogma has been challenged (Sackeim 2015; Regenold et al. 2015). Besides, electromagnetic, but non-seizure causing neuromodulatory treatments, such as TMS, proved to be efficient in the treatment of depression. Therefore, in this study, we set out to investigate the direct effect of the ECT induced electrical field (EF) in a multisite longitudinal neuroimaging study.

    Methods: We analyzed N=153 (94 F, age: 57.3 +- 17.1) patients who underwent right unilateral (RUL) ECT for depression and had two brain MRIs, first at baseline, second at the end of the ECT course. Patients were followed clinically and MADRS was administered at baseline and the end of the ECT course. The range of the number of ECT was 7 to 15. We estimated ECT induced EF with ROAST v1.1 (Huang et al. 2017). After segmentation of the structural MRI T1-weighted images, ROAST builds the three-dimensional tetrahedral mesh model of the head. The segmentation identifies five tissue types: white and gray matter of the brain, cerebrospinal fluid, skull, and scalp, and assigned them different conductivity values: 0.126 S/m, 0.276 S/m, 1.65 S/m, 0.01 S/m, and 0.465 S/m respectively. ROAST uses finite element methods (FEM, #Logan et al. 2007#) to find the electric potential distribution under 1 A/m2 current density injected into the frontal stimulation electrode. The resulting electric potential distribution was calibrated to correspond to the current injection used in the device (Thymatron 900 mA, MECTA 800 mA). These procedures resulted in a voxel-wise EF distribution map in each individual. Further processing was performed by FreeSurfer version 5.3, and Quarc (Holland and Dale 2011) was used for unbiased estimation of subcortical and cortical volume change. 96 ROIs were identified and volume change between baseline and end of the ECT treatment was calculated. We also calculated the mean EF in each ROI. In the ROI based analyses, we used Bonferroni corrected statistical threshold (p<0.0005).

    Results: There was a strong correlation between EF and volume increase in amygdala and hippocampus both sides (p<0.0005) respectively. These correlations were even stronger when we multiplied EF with the number of ECT (cumulative dose). Due to the asymmetry of the dose distribution of the RUL ECT, this correlation was present on both sides with different scales. In line with our previous reports, volume changes, however, did not correlate with clinical improvement (Oltedal et al. 2018). In a separate analysis, we found that EF correlated with clinical improvement in the cingulate cortex only (voxel-wise analysis, FDR corrected p<0.05), but no volumetric changes were identified in this region.

    Conclusions: The direct electrical distribution of ECT is tightly linked with MRI brain changes in the hippocampus and amygdala. While EF is likely causative in these volume changes, volume changes fail to mediate clinical effect. Volume change is either neutral in these regions or responsible for memory side effects, which needs to be verified in future studies with neurocognitive measures. Based on our results EF exerts its clinical effect at the cingulate cortex. Overall this is an encouraging biological validation of the electrical field modeling technique which could lead to new insights in personalized neuromodulation.

    Keywords: Neurostimulation, Depression, Brain MRI, ECT, Electrical Field Modeling

    Disclosure: Nothing to disclose.

    M110. Discovery of Central Autonomic Dysregulation Mechanisms in Major Depression and Translation of a Novel Neuromodulation Technique for its Treatment

    Ronald Garcia*, Klara Mareckova, Arielle Stanford, Jessica Stowell, Laura Holsen, David Gitlin, Riccardo Barbieri, Vitaly Napadow, Jill Goldstein

    Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, United States

    Background: Cardiac autonomic dysregulation in response to negative stressful stimuli has been implicated in the comorbidity of major depression (MDD) and cardiovascular disease (CVD). We previously argued that this maladaptive physiological response is associated with abnormalities in the development of specific regions in the stress response circuitry (hypothalamus, amygdala, hippocampus, anterior cingulate cortex, orbital and medial PFC, and brainstem regions) that are morphologically and functionally sexually dimorphic and associated with vulnerability for sex differences in MDD and CVD risk. In a previous functional magnetic resonance imaging (fMRI) study we demonstrated the differential and greater negative impact of depressed mood on brain activity deficits and connectivity between stress circuitry regions in women than men (specifically hypothalamus and amygdala with orbital and hippocampal regions). We have now extended this work by investigating sex differences in the associations of these fMRI findings with peripheral autonomic dysregulation (i.e., reduced parasympathetic cardiac tone) in response to negative affective stimuli (Discovery phase). These results were then used in a second study to guide the evaluation of a novel, non-invasive, respiratory-gated auricular vagus afferent nerve stimulation (RAVANS) technique for transcutaneous vagus nerve stimulation (tVNS) on the modulation of the above brain regions and peripheral parasympathetic activity in 20 women with active recurrent MDD (Translational phase).

    Methods: Discovery phase: The sample consisted of 50 subjects (28 females, 22 males), ages 45.5±5.0 years, spanning a range from healthy to major Axis I diagnoses to ensure depressed mood variability. fMRI data were acquired on a Siemens Tim Trio 3 T MRI scanner (TR=2000 ms, TE=40 ms, slice thickness 5 mm). The fMRI task consisted of presentation of images adapted from the International Affective Picture System, comprising a mild visual stress challenge. Average parameter estimates (percent signal change values) and task-related connectivity values (beta weights of psychophysiological interaction regressors) in significant target clusters (negative vs. neutral) were extracted. The relationship between stress response circuitry activity and connectivity and cardiovagal activity [percent change in high frequency power of heart rate variability (HFn)] in response to negative affective stimuli was evaluated using GLM analyses, including interactions with depressed mood and sex.

    Translational phase: Twenty women (30.3±4.7 yrs) with recurrent MDD and in an active episode were included. fMRI data were acquired on a Siemens Tim Trio 3 T MRI scanner (TR=1250 ms, TE=33 ms, slice thickness 2 mm). Women attended two imaging visits within one-week (to ensure early follicular timing across sessions) in which they were exposed to a visual stress challenge that preceded and followed expiratory-gated (eRAVANS) or inspiratory-gated (iRAVANS) stimulation. Stress response circuitry areas evaluated in the discovery phase were used for region of interest group-level analyses with non-parametric permutation tests (eRAVANS vs iRAVANS). In addition, a GLM analysis evaluated effects on cardiovagal modulation (HFn) and depressive symptoms [Beck Depression Inventory (BDI)].

    Results: Discovery phase: Significant relationships between reduction of cardiovagal activity and activation of hypothalamus (p=0.001) and amygdala (p=0.02) during exposure to negative images were observed in women (and not men) with high depressed mood. Analyses also revealed that only women (and not men) with high depressed mood exhibited low connectivity between hypothalamus and right orbitofrontal cortex associated with low cardiovagal activity (lower HFn change) in response to negative affective stimuli (p=0.001).

    Translational phase: Significantly greater activation of nucleus tractus solitarii and raphe nuclei were observed during stimulation with eRAVANS compared to iRAVANS (p<0.01). In addition, a significant greater activation of anterior cingulate and orbitofrontal cortices as well as deactivation of hypothalamus were observed during the stress task (post vs pre RAVANS) when comparing eRAVANS to iRAVANS (p<0.01). Furthermore, a significant acute reduction in BDI values (28.1±6.9 vs 19.9±9.1, p<0.01) and increase in cardiovagal output (HFn: 43.3±9.3 vs 52.2±10.6, p<0.01) were observed after eRAVANS, with no significant differences after iRAVANS.

    Conclusions: The discovery phase in this set of analyses revealed mood and sex-dependent interactions in the association between stress response circuitry activity deficits and reduced cardiovagal modulation in response to negative affective stimuli. In particular, women with depressed mood showed greater inability to regulate brain activity and cardiac autonomic tone during negative emotional arousal than men. These results provide a window into a pathway for understanding sex differences in the comorbidity of MDD and CVD. Subsequently, results from the translational phase evaluating RAVANS tVNS in women with recurrent MDD demonstrated that this neuromodulatory therapeutic approach effectively modulated the stress response circuitry with acute beneficial effects on mood and autonomic regulation. This discovery-translational research design suggests a novel neuromodulatory application to the treatment of comorbid MDD and CVD in women.

    Keywords: Major Depression Disorder, Stress Response Circuitry, Cardiac Autonomic Tone, Transcutaneous Auricular Vagus Nerve Stimulation (taVNS), Sex Differences

    Disclosure: Nothing to disclose.

    M111. Neurotransmitter Sensing via Aptamer-Field-Effect Transistors

    Abstract not included.

    M112. CSF Acetylcholinesterases: Relationship to Peripheral and CSF Pro-Inflammatory Cytokines in Late-Life Major Depression

    Nunzio Pomara*, Davide Bruno, Jay Nierenberg, Chelsea Reichert, Minnie Fu, Katherine Brundage, Henrik Zetterberg, Kaj Blennow

    Nathan S. Kline Institute/New York University School of Medicine, Orangeburg, New York, United States

    Background: Increased levels of peripheral and CSF pro-inflammatory cytokines, such as IL-6, TNF-α, and IL-8 have been reported in depressive disorders, including late life major depression (LLMD), and these elevations may interfere with patients’ responses to conventional antidepressants. However, findings from meta-analyses that examined inflammatory biomarkers in depression, including pro-inflammatory cytokines, also indicate that the results of these studies are conflicting, suggesting that elevations of pro-inflammatory cytokines may occur in some depressed individuals but not others. Thus, there is a need to elucidate clinical and biological factors that may contribute to these differential effects. Several lines of evidence suggest that an increase in cholinergic activity produced by administration of cholinergic agonists such as cholinesterase inhibitors results in a significant attenuation of cell-mediated immune responses, such as those evoked in endotoxin-induced sepsis and other inflammatory conditions. These anti-inflammatory effects of acetylcholine (ACh) have been shown to be mediated by activation of efferent vagal pathways and α-7 nicotinic ACh receptors on macrophages and other inflammatory cells. It has also been suggested that activation of the cholinergic anti-inflammatory reflex can be triggered by localized elevations in peripheral cytokines via vagal afferents or by elevations in circulating cytokines that enter the brain via the circumventricular organs or via active transport across the blood-brain barrier.

    Interestingly, reductions in peripheral butyrylcholinesterase (BChE) activity have been described as one of the earliest manifestations of acute systemic inflammation. These reductions have been associated with decreased release of pro-inflammatory cytokines and better survival, suggesting that a compensatory cholinergic anti-inflammatory response occurs in systemic inflammation. In a preclinical experiment, a single intraperitoneal injection of the lipopolysaccharide endotoxin in rats resulted in significant decline in brain acetylcholinesterase (AChE) and BChE activity consistent with possible upregulation of the central cholinergic anti-inflammatory response. However, the status of the cholinergic system in the context of cytokine changes in individuals with long-standing depression has never been examined. This gap in the literature prompted us to examine peripheral and CSF pro-inflammatory cytokine levels (IL-6, IL-8, TNF-α, IL-1b) in LLMD, a disorder associated with increased risk for Alzheimer’s disease (AD) and their relationship to CSF AChE and BChE activity.

    Methods: 91 participants, aged 60 years and older completed a 3-year longitudinal study. All subjects were cognitively-intact at screening and had a) no evidence of dementia, b) a Mini-Mental State Exam score of at least 28, and c) no gross MRI abnormalities other than white matter hyperintensities. Baseline CSF was obtained from 44 older adults (27 with LLMD and 17 healthy controls). Plasma pro-inflammatory cytokines levels (IL6, IL8, IL1b) and CSF BChE, AChE, and pro-inflammatory cytokines were determined using commercial kits as per manufacturers’ instructions. Nonparametric Mann-U Whitney test and Spearman’s correlations were run.

    Results: CSF BChE was lower in people with LMDD compared to controls (Mann-Whitney U Test p = .02). There was no significant difference between groups for CSF AChE (p = 0.13), but a similar trend was observed with lower AChE in the LLMD group.

    Plasma Findings. In LLMD, lower CSF BChE (rho = -.46, p = .02) and AChE (rho = -.45, p = .02) were associated with higher plasma IL-6. Whereas, in the control group, only lower CSF AChE was associated with higher plasma IL-8 (rho = -.54, p =.03).

    CSF Findings. In the whole group analysis (LLMD and control group), lower CSF BChE was associated with lower CSF IL-8 (rho = 0.60, p < 0.001). However, only in the control group, lower CSF BChE was associated with lower CSF IL-8 (rho = 0.62, p = 0.008).

    Conclusions: CSF BChE activity was lower in the LLMD group compared to controls. A similar trend was also observed for CSF AChE. Plasma cytokines were inversely correlated with CSF AChE or BChE in both LLMD and controls. In contrast, CSF AChE and BChE were positively correlated with CSF cytokines. These preliminary findings, if confirmed, are consistent with a possible compensatory upregulation of the central cholinergic anti-inflammatory pathway in response to elevated plasma pro-inflammatory cytokines, especially in LLMD.

    Keywords: Late-Life Depression, Vagus, Pro-inflammatory Cytokines, Cholinergic Anti-Inflammatory Response

    Disclosure: Nothing to disclose.

    M113. A Significant Dose-Effect and Insignificant Sequence Effect of the Prediction of Suicidal Behavior by Interactions Between Immune-Driven Illnesses and TBI

    Teodor Postolache*, Michael Benros, Lisa Brenner, Olaoluwa Okusaga, Christopher Lowry, John Stiller, Annette Erlangsen

    University of Maryland School of Medicine, Baltimore, Maryland, United States

    Background: Inflammation appears to moderate and immune activation may even mediate certain clinical symptoms and functional trajectories in Traumatic Brain Injury (TBI) and was more recently implicated in suicidal behavior. TBI and specific immune mediated clinical conditions (IMCs: infections, allergy and autoimmune) have been previously predictively associated, individually, in Suicidal Self-Directed Violence (SSDV). We now evaluated the effect of interactions between TBI and IMCs in predicting SSDV in a Nationwide, population-based, prospective cohort study, with a special focus on dose-response and temporal sequence effects.

    Methods: All 7.22 million individuals- 15 years or older living in Denmark between January 1, 1980, and December 31, 2011, were observed during a 32-year follow-up period, with more than 149 million person-years of follow-up. The relative risk for suicide and SSDV (suicide and attempts) was expressed as Incidence rate ratios (IRRs) and accompanying 95% CIs. History of TBI and infections, allergic disease and autoimmune disease, with stratification for mental illness, were coded using ICD 8 and 10. Data was adjusted for sex, age, and time interval, and was analyzed with Cox regression and multivariable logistic regression. We further analyzed lethal vs. non-lethal SSDV, SSDV using violent vs. nonviolent methods, number of admissions for IMCs and TBI and order of interaction (TBI first vs. IMC first).

    Results: There was a significant interaction between TBI and IMCs in predicting SSDV- with synergistic effects for infections and allergy (p<0001), and bellow- additive effects for autoimmune disease (p<0.001). For infections specifically, while a robust dose-response effect has been identified (p<0.001), there was no significant difference in effect size with the sequence of the interaction.

    Conclusions: The results support the possible role of inflammation (and IMCs) in predicting severity of TBI, mediating its SSDV risk elevating effects and perpetuating that risk. Preventing and assertively addressing IMCs may significantly lower the SSDV risk after the TBI.

    Keywords: Suicide, Depression, Traumatic Brain Injury, Inflammation

    Disclosure: Nothing to disclose.

    M114. Temperament, Personality, and Risk for Mood Disorder

    Tiffany Greenwood*, Nancy Downs

    University of California, San Diego, La Jolla, California, United States

    Background: Early detection of vulnerability is an important clinical goal that can have a major impact on future quality of life by enabling timely interventions and effective risk management. However, this requires a better understanding of the environmental and genetic influences, as well as latent personality traits that may serve as key indicators of risk prior to the onset of illness. We, and others, have consistently shown that individuals with mood disorder possess temperament and personality traits that are significantly different from healthy controls, even in euthymia. Based on these observations, we recently began a pilot study to develop prediction models based on behavioral, environmental, and genetic risk factors to facilitate the early identification of mood disorders in transition age youth.

    Methods: In total, 468 undergraduate students aged 18-25 completed an online, anonymous survey during the first year of the study at one public, west coast university with an undergraduate enrollment of approximately 28,000. The survey used the Patient Health Questionnaire to screen for depression, the Mood Disorder Questionnaire to screen for bipolar spectrum disorders, and the Suicide Behaviors Questionnaire-Revised to screen for suicide risk. Additional questions were included to inform lifetime diagnoses of major depression and bipolar disorder according to DSM criteria. The Temperament Evaluation of Memphis, Pisa, Paris, and San Diego Auto-questionnaire (TEMPS-A) was used to assess lifelong, milder aspects of bipolar symptomatology. The Temperament and Character Inventory (TCI) was included to evaluate additional dimensions of personality according to a psychobiological model. The Hypomanic Personality Scale (HPS) and the Barratt Impulsivity Scale (BIS) were also included.

    Results: Of the 441 students with complete mood and personality data, 166 reported no current or lifetime mood symptoms and serve as controls, 54 met criteria for bipolar disorder, and 62 met criteria for depression, 19 of whom also reported significant manic symptoms and are likely bipolar. Additionally, 159 students reported significant mood symptoms but did not meet criteria for a mood disorder due to lack of reported impairment. All temperament and personality traits investigated showed significant group differences. Hypomanic personality, impulsivity, irritable temperament, and novelty seeking were characteristic of individuals with bipolar disorder, while harm avoidance was characteristic of depression. Dysthymic and anxious temperament significantly discriminated healthy controls from those with a mood disorder but did not distinguish bipolar disorder from depression. Only the cyclothymic temperament significantly discriminated between all three groups, with effect sizes of 1.0 and 1.8 for depression and bipolar disorder, respectively, compared with controls. A combination of hypomanic personality and cyclothymic and irritable temperament explained >0.40 of the variance in bipolar disorder.

    Conclusions: These results are consistent with those of others in suggesting that temperament and personality traits can be useful in predicting risk for mood disorder, in combination with other factors. It must be noted, however, that these results are based solely on self-report of mood symptoms in the absence of structured clinical interviews. As mood disorders often emerge during adolescence and young adulthood, this stage of development represents a critical point for early intervention. A prediction model based on latent personality traits in conjunction with genetic and environmental risk factors would facilitate this goal.

    Keywords: Mood Disorder, Temperament, Personality, Biomarkers for Risk Assessment

    Disclosure: Nothing to disclose.

    M115. Data-Driven Subtypes of Depression in a Treatment Resistant Cohort

    Laura Hack*, A. Umair Janjua, Sara Dover, Patricio Riva Posse

    VA Palo Alto MIRECC/Stanford School of Medicine, Palo Alto, California United States

    Background: Initial efforts to parse patients with depression into more clinically homogeneous groups were based on observations of differences in symptomatology. Over the past five decades, these efforts have become data-driven. More recently, there has been renewed interest in this area due to the development of the NIMH’s Research Domain Criteria as well as advances in machine learning approaches and biomarker measurement. The ultimate goal of this work is the development of a more biologically-based understanding of psychiatric diseases that will foster the selection of personalized treatments. Most work in this area has utilized two broad data-driven person-centered analytic approaches, including finite mixture models, particularly latent class analysis, and clustering methods. Both methods are unsupervised in that the algorithms find patterns in the data without being provided with labels. Analyses with these methods have yielded between 2 and 7 latent classes or clusters of depressed patients. Both methods have consistently identified subtypes based on severity, as well as melancholic and atypical features. To our knowledge, no studies attempting to subtype patients with depression have specifically focused on a treatment resistant population.

    Methods: We used data from 319 patients of both sexes who presented for consultation to our Treatment Resistant Depression Center at The Emory Clinic from January 2016 through July 2018. We entered baseline Quick Inventory of Depressive Symptoms (QIDS-SR16) scores into an agglomerative hierarchical clustering algorithm using Ward’s method. To determine the optimal number of clusters, we used the R package NbClust, which provides 30 indices, including the standardly used average silhouette method and the gap statistic, and selects the optimal number based on the majority rule. Using ANOVA, we then tested whether subjects within the 3 clusters differed based on severity and type of depressive symptoms.

    Results: A 3-cluster solution was identified as optimal by the unsupervised machine learning algorithm and the majority rule. Based on the symptom profiles, we labeled the clusters, or subtypes, as “Severe Melancholic” (N=158, 50%), “Mild Melancholic” (N=126, 39%), and “Atypical” (N=35, 11%). The clusters differed significantly on both severity of depressive symptoms (F1 = 42.3, P = 3.6e-10) and symptom type. The “Atypical” cluster had a mean QIDS score of 20.0 with significantly higher scores on the QIDS items “Sleeping too much” (F1 = 11.5, P = 7.7e-4) and “Increased appetite” (F1 = 137.6, P < 2e-16) as compared to the other clusters. The mean QIDS score for the “Severe Melancholic” cluster was 20.4 with significantly higher scores on “Decreased appetite” (F1 = 73.6, P < 2e-16) and “Early morning awakening” (F1 = 21.4, P = 5.5e-6), while the mean QIDS for the “Mild Melancholic” group was 13.6.

    Conclusions: We present a clustering analysis of depressive symptoms in a group of treatment resistant depressed patients that reveals differences in subtypes based on severity and symptom type. This is consistent with many prior studies using data-driven approaches to parse patients with depression, although, to our knowledge, our study is the first to use a treatment resistant cohort. We plan to test for differences among the subtypes on a variety of sociodemographic and biological measures, which may have implications for personalized interventions.

    Keywords: Hierarchical Clustering, Depression Subtypes, Treatment Resistant Cohort

    Disclosure: Nothing to disclose.

    M116. Genome-Wide DNA Methylomic Differences Between Dorsolateral Prefrontal and Temporal Pole Cortices of Bipolar Disorder

    Marin Veldic*, Man Choi (Ada) Ho, Stacey Winham, Sebastian Armasu, Caren Blacker, Vincent Millischer, Catharina Lavebratt, James Overholser, George Jurjus, Lesa Dieter, Gouri Mahajan, Grazyna Rajkowska, Eric Vallender, Craig Stockmeier, Keith Robertson, Mark Frye, Doo-Sup Choi

    Mayo Clinic, Rochester, Minnesota, United States

    Background: Dorsolateral prefrontal cortex (DLPFC) and temporal pole (TP) are brain regions that display abnormalities in bipolar disorder (BD). DNA methylation is an epigenetic mechanism that is heritable and sensitive to environmental influence and hence has received considerable attention in mood disorder studies. Here we hypothesized that DNA methylomic (DNM) differences in DLPFC and TP may play a role in BD. We investigated the genome-wide DNM differences associated with BD within and between the gray matter of DLPFC and TP. Major depressive disorder (MDD) cases were included to identify BD-specific brain DNM differences.

    Methods: Postmortem brains were collected at autopsy at the Cuyahoga County Medical Examiner’s Office, Cleveland, OH. All procedures performed were in accordance with the Declaration of Helsinki and approved by the Institutional Review Boards. Subjects meeting DSM-IV criteria for BD or MDD, and controls without a psychiatric diagnosis, were selected for this study. Subjects with neuropathological or neurological disorders were excluded. Tissues were dissected and rapidly frozen in 2-methylbutane on dry ice without fixation and stored at -80°C. Gray matter tissues of DLPFC and anterior TP from 20 BD, 10 MDD, and 10 control subjects were procured. MDD and control subjects were matched to BD subjects according to the group’s age and sex distributions, while control subjects were age-and-sex matched to MDD subjects individually. Genomic DNA was extracted from ~40 mg tissue and then subjected to bisulfite conversion. DNA methylation levels were determined by the Illumina 850 K Infinium MethylationEPIC BeadChip. Multivariate analyses (principal component analysis and hierarchical clustering) were performed on top variable probes to detect potential confounding factors. Differentially methylated positions (DMPs) were then identified by the R Bioconductor package limma. Paired comparison was used for between-brain-region comparisons in each group and for MDD-control comparison in each brain region. Statistically significant DMP was considered at adjusted p<0.05. Functional enrichment analysis was performed on genes annotated to the lists of significant DMPs that were shared by all groups as well as shared by and distinct to BD and MDD. Pathways were considered statistically significant at Benjamini-Hochberg corrected p<0.05.

    Results: Multivariate analyses detected no significant associations of methylation profile clusters with arrays, sex, age, postmortem interval, or tissue pH. Between DLPFC and TP, 1601 DMPs in the control group, 11954 DMPs in the MDD group, and 39039 DMPs in the BD group reached statistical significance. Substantial proportions of these DMPs overlapped between groups, indicating that some between-brain-region methylation differences were common to all or some subject groups. The genes associated with DMPs shared by BD and MDD (7709 DMPs) were significantly enriched in seven nervous system-associated pathways; while the genes associated with DMPs distinct to the BD group (29846 DMPs) were significantly enriched in 15 nervous system-associated pathways. The top-ranked pathway was “axon guidance signaling” for both analyses. Pathways enriched only in the BD-MDD shared list pointed to GABAergic dysregulation, while those enriched in the BD-only list suggested glutamatergic dysregulation and more impacts on synaptogenesis and synaptic plasticity. No significant DMPs were found in any between-group comparisons in DLPFC or TP, and the pathway analyses of the top 100 DMPs in these pairwise comparisons detected no significantly enriched pathway.

    Conclusions: We identified DNM differences between DLPFC and TP that were associated with BD and MDD and associated with BD only but absent in controls. These differences were related to genes and pathways involved in axon guidance, glutamatergic and GABAergic neurotransmissions, synaptic plasticity, and neurodevelopmental mechanisms. Our findings imply that a methylation imbalance between DLPFC and TP may contribute to BD symptomology. If so, DNM could prove to be an important target for further investigations of vulnerabilities associated with BD. Limitations of this study include the lack of cell type specificity, inability to differentiate between DNA methylation and DNA hydroxymethylation, small sample size, and unbalanced groups. These important issues should be addressed in additional studies.

    Keywords: DNA Methylation, Bipolar Disorder, Postmortem Brain Tissue

    Disclosure: Nothing to disclose.

    M117. Prophylactic Ketamine Protects Against Fear Overgeneralization

    Alessia Mastrodonato*, Randy Martinez, Ina Pavlova, Christina LaGamma, Rebecca Brachman, Alfred Robison, Christine Ann Denny

    Columbia University, Research Foundation for Mental Hygiene, Inc., New York, New York, United States

    Background: Stress exposure is the main cause of mood disorders, such as major depressive disorder (MDD) and post-traumatic stress disorder (PTSD). However, some individuals can successfully adapt to stress, which is known as stress resilience. We previously reported that a single injection ketamine, an NMDA antagonist, prior to stress protects against the development of depressive-like behavior and attenuates learned fear in mice. Recently, we have showed that ketamine induces a prophylactic effect by modulating expression of the transcription factor ΔFosB in ventral CA3 (vCA3). In particular, transcriptional inhibition or overexpression of ΔFosB in vCA3 occludes and mimics, respectively, prophylactic ketamine efficacy, suggesting that ΔFosB expression in vCA3 is necessary and sufficient for ketamine prophylactic efficacy. Here, we sought to identify if prophylactic ketamine protects against fear generalization and which neural alterations correlate to fear generalization.

    Methods: ArcCreERT2 mice were injected with saline or ketamine (30 mg/kg). One week later, mice were administered a pattern separation (PS) paradigm. After an initial training in the context A, where they received a shock, mice were exposed daily to the aversive context A and to a similar, but safe context B, for 10 days. Mice were then sacrificed, and brains were processed. Whole-brain immunolabeling was utilized in order to identify the neural ensembles (e.g., memory traces/engrams) representing fear generalization.

    Results: During PS task, both groups of mice exhibited comparable levels of freezing following one-shock in the aversive context A. However, prophylactic ketamine mice ketamine mice distinguished between the two contexts more rapidly than prophylactic saline mice, froze significantly less in the similar but neutral context B and had significantly higher levels of discrimination between the two contexts. The ketamine prophylactic effect persists up to 6 weeks following a single injection.

    Conclusions: Our data indicate that a single prophylactic injection of ketamine may be able to prevent fear generalization and protect against stress-induced psychiatric disorders (e.g., PTSD and depression) in a long-lasting, self-maintaining vaccine-like fashion. Prophylactic ketamine treatment in human subjects and PTSD patients might decrease fear generalization and alter similar neural circuits identified in the mouse studies.

    Keywords: Fear Generalization, Ketamine, PTSD

    Disclosure: Nothing to disclose.

    M118. Absence of QTc Prolongation with Buprenorphine/Samidorphan Combination: A Thorough QT/QTc (tQT) Study in Healthy Volunteers

    Hong Lu*, Marjie Hard, Robert Kleiman, Wenlei Liu, Tim Bergsma, Arielle D. Stanford, Narinder Nangia, Lisa von Moltke

    Alkermes, Inc., Waltham, Massachusetts, United States

    Background: The fixed-dose buprenorphine/samidorphan (BUP/SAM) combination comprises BUP, a μ-opioid receptor partial agonist, and SAM, a novel μ-opioid receptor antagonist optimized for potency and sublingual bioavailability. The efficacy and safety profile of adjunctive BUP/SAM has been demonstrated in patients with MDD with inadequate response to antidepressant therapy. Clinical development routinely includes evaluation for cardiac ventricular repolarization liability. Here we describe the effect of escalating doses of BUP/SAM on the QT interval and other electrocardiographic parameters to characterize the potential for arrhythmia associated with therapeutic and supratherapeutic doses.

    Methods: Study 213 (NCT0247930) was a phase 1, randomized, double-blind, placebo (PBO)-controlled, parallel study with a nested crossover of BUP/SAM, positive control (moxifloxacin), and PBO in healthy adults. The study consisted of positive control period 1 (Day 1), a titration period (Days 2 to 13), and positive control period 2 (Day 14). Participants were randomized in a 2:1:1 ratio into Treatment Groups 1, 2, or 3, respectively. Group 1 received moxifloxacin-matched PBO during both positive control periods and BUP/SAM during the titration period, with titration from 0.5 mg/0.5 mg to 8 mg/8 mg. Group 2 received moxifloxacin 400 mg during positive control period 1 and BUP/SAM, then moxifloxacin-matched PBO during the titration period and positive control period 2, respectively. Group 3 received moxifloxacin, then BUP/SAM-matched PBO during positive control period 1 and the titration period, respectively, and moxifloxacin 400 mg during positive control period 2. Pharmacodynamic (PD) electrocardiograms (ECGs) were obtained during a 24-hour observation window using a 12-lead digital (Holter) ECG recorder at baseline (Day –1), and on Days 1, 2, 7, 10, 13, and 14. Blood pharmacokinetic (PK) samples were collected at the end of each PD ECG timepoint starting on Day 1. The primary endpoint, ΔΔQTcF, was the time-matched mean difference between BUP/SAM 8 mg/8 mg and PBO in their respective ΔQTcF, which was the time-matched change from baseline to Day 13. The ΔΔQTcF was evaluated using a linear mixed-effects model. Secondary endpoints included the ΔΔQTcF for BUP/SAM 0.5 mg/0.5 mg (Day 2), 2 mg/2 mg (Day 7), and 4 mg/4 mg (Day 10) doses, as well as heart rate (HR), PR interval, QRS interval, and ECG morphology. PK parameters of BUP, SAM, and their respective metabolites were determined using a non-compartmental method. The relationship between the QTcF effect and plasma concentrations of BUP, SAM and metabolites was analyzed using linear mixed-effects modeling. Adverse events (AEs) were recorded.

    Results: Of 128 participants randomized, 111 completed the study. The estimated least squares mean (LSMEAN) ΔΔQTcF ranged from –3.4 to –0.2 msec with all upper bounds of the 2-sided 90% confidence interval (CI) well below 10 msec in the 24 h after BUP/SAM 8 mg/8 mg administration on Day 13 (P < 0.001), indicating an absence of QT prolongation. The LSMEAN ΔΔQTcF for the lower doses were –5.4 to 4.4 msec (0.5 mg/0.5 mg), –1.8 to 3.0 msec (2 mg/2 mg), and –3.2 to 1.5 msec (4 mg/4 mg). No values were associated with upper bounds of 90% CIs exceeding the threshold of 10 msec (P < 0.05), indicating absence of QT prolongation. The QT assay sensitivity was confirmed with lower bounds of 90% CIs above 5 msec following moxifloxacin administration. Mean changes from baseline for HR, PR interval, and QRS interval during the 24-hour period after BUP/SAM administration relative to PBO were small and not clinically meaningful. Plasma exposures for BUP, SAM, and metabolites appeared to increase in a dose-proportional manner. Linear mixed effects modeling of the plasma concentration-QTc relationships showed positive but shallow slopes for each analyte. The incidence of AEs after up to 12 days of BUP/SAM (53.1%) or matched PBO (40.3%) were similar. No serious AEs or deaths were reported. No participants experienced AEs related to ECG morphology findings.

    Conclusions: A daily therapeutic sublingual dose of BUP/SAM 2 mg/2 mg and a daily supratherapeutic sublingual dose of BUP/SAM 8 mg/8 mg did not prolong the QTcF interval above the regulatory threshold of concern. Results indicate that BUP/SAM 2 mg/2 mg does not have an adverse effect on cardiac repolarization. BUP/SAM from 0.5 mg/0.5 mg to 8 mg/8 mg was generally well tolerated in this study.

    Keywords: Thorough QT study, Buprenorphine/samidorphan, Negative QT Study, Safety

    Disclosure: Alkermes, Inc., Employee

    M119. Deconstructing Synaptic and Behavioral Effects of Rapid-Acting Antidepressants With Circuit and Cell Type Specific Resolution

    Marco Pignatelli*, Leonardo Bontempi, Yocasta Alvarez-Bagnarol, Debora Masini, Carlos Zarate Jr., Michael R. Tadross, Gilberto Fisone, Antonello Bonci

    National Institute on Drug Abuse Intramural Research Program/NIH, Baltimore, Maryland, United States

    Background: Preclinical evidence has established that a single sub-anesthetic dose of either (R,S)-ketamine (ketamine) or its metabolite (2 R,6 R)-hydroxynorketamine (HNK) exerts rapid-acting antidepressant effects. While ketamine use is associated with undesirable side effects, such as psychotomimetic effects and addiction liability, HNK exerts its antidepressant actions in mice, while lacking ketamine-related side effects. Whereas ketamine and HNK’s primary molecular target is still uncertain, unequivocal experimental evidence shows that synaptic activation of AMPA receptors (AMPARs) is required for their antidepressant effects. Thus, I started investigating whether AMPAR increase in synaptic strength within the Nucleus Accumbens (NAc), a key brain region involved in motivation and hedonic drive, represents the key “therapeutic” neuroadaptation induced by a single sub-anesthetic dose of ketamine or HNK and if the latter condition might underlie their long-lasting antidepressant effects, which notably outlast drug clearance for days.

    Methods: In order to deconstruct the synaptic and behavioral effects of ketamine and HNK with circuit and cell type specific resolution, I used the mouse as a model system (7-8 males and 7-8 females for each group for each manipulation that follow), and I first explored whether acute intraperitoneal exposure to saline, ketamine (10 mg/kg) or HNK (10 mg/kg) was affecting the activity of glutamatergic neurons that project to the NAc from ventromedial prefrontal cortex (vmPFC), basolateral amygdala (BLA), paraventricular thalamus (PVT) or ventral subiculum (vSub). Toward this goal, I injected the retrograde tracer, CTb, into the NAc and evaluated projection-specific neuronal activation by looking at the neuronal activity marker, cFos, within CTb-expressing projecting neurons, 1 h post injection. Moreover, I collected brain slices containing the NAc and performed whole-cell recordings from the main population of neurons residing within the NAc (D1 and D2 MSNs), while optogenetically interrogating the main monosynaptic inputs to NAc in mice injected with either saline, ketamine (10 mg/kg) or HNK (10 mg/kg) 24 h earlier. To test whether increased AMPAR function within the NAc was required for ketamine or HNK’s rapid-acting antidepressant effects, I delivered the selective AMPAR antagonist, NBQX (40 µM), into the NAc 15 min prior to the forced swim test (FST), a well-described rodent model of behavioral despair, in mice that have been exposed to a systemic injection of either ketamine or HNK24h earlier. Moreover, to determine the postsynaptic sites of action of ketamine and HNK, I employed an intersectional viral approach that allowed me to selectively block AMPAR onto accumbal D1 MSNs or D2 MSNs by means of DART technology. DART (Drugs Acutely Restricted by Tethering) offers an unprecedented opportunity to block AMPAR within genetically specified cells (D1 vs D2 cells). The validated DART AMPAR antagonist, termed YM90K-DART (experimental group) or vehicle (control group), was delivered into the NAc 15 min prior to the FST in animals that have been exposed to a systemic injection of either saline, ketamine or HNK 24 h earlier.

    Results: In this study, I found that a single sub-anesthetic dose of either ketamine or HNK leads to the establishment of input-selective synaptic plasticity, in the form of long-term potentiation, within NAc D1 and D2 medium spiny neurons (MSNs). Moreover, by using a contralateral disconnection procedure in combination with DART, I show that the occurrence of input-selective plasticity and postsynaptic recruitment of AMPAR within genetically specified NAc MSNs is necessary and sufficient to ameliorate behavioral aspects ascribed in the context of anhedonia and behavioral despair.

    Conclusions: These data will help deciphering the detailed mechanisms of ketamine and HNK’s antidepressant action and may lead to novel and safer antidepressant agents with circuit and cell type-selective effects and decreased undesirable side effects, aligned with the spirit of modern precision medicine.

    Keywords: Racemic Ketamine and Metabolites, Nucleus Accumbens, Medium Spiny Neuron, AMPA Glutamate Receptors

    Disclosure: Nothing to disclose.

    M120. The Duration of Antidepressant Effects by Ketamine or the Ketamine Metabolites (2 R, 6 R)-HNK and (2 S, 6 S)-HNK is Extended by Repeated Administration

    Caroline Browne*, Hildegard Wulf, Janet Clark, Carlos Zarate, Irwin Lucki

    Uniformed Services University of the Health Sciences, Bethesda, Maryland, United States

    Background: Major Depressive Disorder (MDD) is one of the leading causes of disability and functional impairment worldwide. Characterized by persistent negative mood and anhedonia, current antidepressant treatments are only modestly effective at producing remission of symptoms. Ketamine is an N-methyl D-aspartate (NMDA) receptor antagonist that demonstrates robust antidepressant effects in treatment-resistant patients at subanesthetic doses within hours of infusion. Unfortunately, the antidepressant effect from a single ketamine treatment lasts for just a few days. However, some recent clinical studies have shown that repeated treatment with ketamine can lengthen the period of its antidepressant effects. The first goal of these studies was to determine whether repeated treatment with ketamine extended the duration of antidepressant-like responsiveness in a rodent model. The second goal of the studies reported here was to corroborate the antidepressant-like effects for (2 R, 6 R)-HNK and (2 S, 6 S)-HNK, two metabolites of ketamine, and to determine whether repeated treatment also extended their period of responsiveness.

    Methods: The behavioral effects of RS-ketamine (10 mg/kg, i.p.) and the ketamine metabolites (2 R, 6 R)-HNK and (2 S, 6 S)-HNK (10 mg/kg, i.p.) were compared with saline in 8 - 12-week-old C57BL/6 J mice. Drugs were administered either once or thrice, with injections spaced 48 h apart. Antidepressant-like activity was measured as decreased immobility in the forced swim test (FST). The FST was administered at 24 h, 7- or 14-days post dosing in separate cohorts of mice. The behavioral effects of HNK were also compared with the motoric and analgesic effects of acute ketamine in assays for locomotor activity and analgesia immediately following the same drug delivery regimen used for the FST.

    Results: Ketamine and the two ketamine metabolites, (2 R,6 R)-HNK and (2 S,6 S)-HNK, exerted an antidepressant-like response when tested 24 h after injection. The effects of ketamine were maintained up to 7 days only when given repeatedly. Repeated treatment with (2 R, 6 R)-HNK, and (2 S, 6 S)-HNK reduced immobility in the FST when measured at 7 and 14 days following repeated injection. Unlike ketamine, (2 R, 6 R)-HNK and (2 S, 6 S)-HNK were inactive in assays for locomotor activity and analgesia, whereas ketamine produces alterations in motoric activity and transient thermal analgesia

    Conclusions: These data confirmed the protracted antidepressant-like activity of ketamine, (2 R, 6 R)-HNK and (2 S, 6 S)-HNK as measured in the mouse FST. Furthermore, repeated treatment extended the duration of antidepressant-like effects for both ketamine and these two metabolites. Finally, in contrast with ketamine’s NMDA mediated effects on locomotor activity and analgesia, the metabolites were inactive in those assays. Ketamine, and its metabolites (2 R, 6 R)-HNK and (2 S, 6 S)-HNK, have potential as novel, rapidly-acting antidepressants that may serve to fill an important treatment gap for MDD.

    Keywords: Depression, Ketamine, Metabolites

    Disclosure: Nothing to disclose.

    M121. (2 R,6 R)-Hydroxynorketamine Acts Presynaptically to Rapidly and Selectively Enhance Hippocampal Glutamatergic Transmission

    Lace Riggs*, Yasco Aracava, Panos Zanos, Jonathan Fischell, Edna Pereira, Edson Albuquerque, Scott Thompson, Todd Gould

    University of Maryland School of Medicine, Baltimore, Maryland, United States

    Background: Ketamine is a rapid-acting antidepressant that is promptly converted into several metabolites in vivo, including (2 R,6 R)-hydroxynorketamine (HNK). Preclinical studies have shown that (2 R,6 R)-HNK retains the rapid and sustained antidepressant actions of ketamine but lacks its dissociative-like properties and abuse potential. While N-methyl-D-aspartate receptor (NMDAR) antagonism has long been credited for the antidepressant effects of ketamine, (2 R,6 R)-HNK is one of several compounds to show preclinical antidepressant actions in the absence of NMDAR antagonism. Previous studies suggest that—regardless of their precise mechanism of action—putative rapid-acting antidepressants exert their effects by restoring the balance between excitation and inhibition in affect-regulating synapses. Understanding how (2 R,6 R)-HNK exerts its effects will allow for a fuller appreciation of the diverse mechanisms underlying rapid antidepressant action and will help to promote the development of therapeutic options that possess superior efficacy, safety, and tolerability than what is currently available. In the current study, we examined the acute actions of (2 R,6 R)-HNK on hippocampal excitatory synaptic transmission.

    Methods: Ex-vivo slice electrophysiology recordings of field excitatory postsynaptic potentials (fEPSPs) and whole-cell patch clamp recordings of miniature excitatory postsynaptic currents (mEPSCs) were used to examine the effects of (2 R,6 R)-HNK in the CA1 region of the hippocampus. A dual recording configuration was used to assess the effects of (2 R,6 R)-HNK concurrently among subregions of the apical dendritic field, and between hippocampal afferent pathways.

    Results: Bath application of (2 R,6 R)-HNK enhanced Schaffer collateral-CA1 fEPSPs in hippocampal slices derived from male and female rats. We found that while the (2 R,6 R)-HNK-induced potentiation is completely eliminated by α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) blockade, these effects were independent of NMDAR activity. Instead, (2 R,6 R)-HNK enhanced AMPAR-mediated synaptic transmission through a concentration-dependent increase in the probability of glutamate release, which occurs uniformly throughout the dorsal-ventral axis of the hippocampus. Additionally, (2 R,6 R)-HNK increases release probability across dendritic subfields that receive Schaffer collateral innervation, but not when stimulating electrotonically distinct temporoammonic afferents that synapse onto the same CA1 cell population. Whole-cell recordings of hippocampal mEPSCs revealed that (2 R,6 R)-HNK increases the frequency, but not amplitude, of AMPAR-mediated mEPSC events, consistent with a presynaptic site of action.

    Conclusions: (2 R,6 R)-HNK augments AMPAR-mediated excitatory synaptic transmission in the hippocampus by acting presynaptically to enhance glutamate release. This suggests that the mechanism underlying the acute effects of (2 R,6 R)-HNK may be distinct from the sustained adaptations in synaptic efficacy that they give rise to, and that are shared among other rapid-acting antidepressant compounds.

    Keywords: Depression, Ketamine, Hydroxynorketamine, Hippocampus, Glutamatergic Transmission

    Disclosure: Nothing to disclose.

    M122. Enhancement of the Firing Activity of Dopamine and Norepinephrine Neurons in Rats Following Repeated Administration of a Subanesthetic Dose of Ketamine

    Chidiebere Michael Iro, Mostafa El Mansari, Pierre Blier*

    University of Ottawa Institute of Mental Health Research, Ottawa, Canada

    Background: Ketamine is a non-competitive NMDA channel blocker which acts by binding to its PCP site. Clinical studies have repeatedly demonstrated the rapid antidepressant effects of subanesthetic doses of ketamine (within 2 to 24 h post administration) [1,2]. In addition to these, previous studies have shown that there is a rapid enhancement (within 2 h) of ventral tegmental area (VTA) dopamine (DA) and locus coeruleus (LC) norepinephrine (NE) activity following ketamine administration, which is α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-dependent [3,4]. The current study investigated the presence of these changes 24 h after a single administration, as well as after repeated administration. We hypothesized that the effects of ketamine on these neurons can be prolonged with repeated administration.

    Methods: Experiments were performed on male Sprague Dawley rats (250 – 320 g). Rats were anesthetized with chloral hydrate (400 mg/kg intraperitoneal (i.p) and mounted in a stereotaxic frame. Recordings were carried out in the dorsal raphe nucleus (DRN), VTA and LC. Electrophysiological recordings were conducted: a) 24 h after a single administration of ketamine (10 mg/kg; i.p) or b) 24 h after repeated administration of ketamine given thrice weekly for two weeks. Neurons were analyzed for changes in action potential frequency and changes in the percentage of bursts occurring per spike, as well as changes in the number of spontaneously active neurons per trajectory through the VTA.

    Results: The previously observed increase in VTA DA neuron activity was absent 24 h after acute administration. However, it was present after repeated administration for two weeks (spikes occurring in bursts: controls 24 ± 4 %, ketamine 32 ± 3 %; p=0.02), as well as an increase in spontaneously active DA neurons per tract (controls: 1.8 ± 0.1, n=61 neurons; ketamine: 3.1 ± 0.6, n=94 neurons; p=0.02; N=5 rats per group). Similar results were obtained for LC NE neurons: there was no change in firing 24 h after a single administration, but an increase after repeated administration (controls: 1.6 ± 0.1 n=49 neurons; ketamine: 2.4 ± 0.1 Hz n=57neurons; p<0.001). No change was observed in the firing activity of DRN serotonin neurons under these conditions.

    Conclusions: The present study supports previous findings which show that ketamine can increase the firing activity of NE and DA neurons. However, this was not observed 24 h after a single dose, but present after two weeks of repeated administration. These effects may thus be contributory to the sustained therapeutic actions of ketamine in MDD when given repeatedly.

    Keywords: Ketamine, Dopamine, Norepinephrine

    Disclosure: Nothing to disclose.

    M123. HCN Channel Inhibitor Induces a Rapid and Sustained Reversal of Social Deficit in a Chronic Social Defeat Stress Model of Depression

    Yingbo Zhu, Mary Regis Shanley, Stacy Ku, Carole Morel, Hongxing Zhang, Yuan Shen, Allyson Friedman, Ming-Hu Han*

    Icahn School of Medicine at Mount Sinai, New York, New York, United States

    Background: Current antidepressant medications for major depressive disorder (MDD) require several weeks to achieve therapeutic effects, and medication optimization takes months to reach treatment efficacy suitable for MDD patients. Such delay for drug efficacy not only prolongs distress and impairment for depressed patients but is also life threatening for suicidal MDD patients. Recent clinical studies, including deep brain stimulation, ketamine and scopolamine therapies, have begun to change the field due to their ability to regulate mood states within minutes to hours. Preclinical studies show that optogenetic control of dopamine neuron activity in the ventral tegmental area (VTA) reward circuit can rapidly regulate social avoidance, a behavioral deficit seen in susceptible mice in the chronic social defeat stress (CSDS) model of depression. Further evidence shows that hyperpolarization-activated cyclic nucleotide-gated (HCN) channels play a key role in the regulation of VTA dopamine neurons, and more recent work supports that HCN channels are involved in mediating ketamine’s rapid antidepressant effects. However, it is unknown whether directly targeting HCN channels can achieve ketamine-like rapid and sustained antidepressant efficacy.

    Methods: C57Bl/6 J mice were subjected to 10-day CSDS. Those exhibiting depressive-like symptoms (susceptible mice) received a VTA microinfusion or systemic intraperitoneal (i.p.) injection of HCN channel inhibitor cilobradine (DK-AH269). In vitro brain slice recordings were employed to assess HCN-mediated Ih currents and firing rate of VTA dopamine neurons post-cilobradine treatment.

    Results: We first show that mice susceptible to CSDS spend significantly less time with a social target as compared to stress-naïve controls and display significantly higher firing activity and increased Ih currents of VTA dopamine neurons when compared to controls. Utilizing this robust model, we further show that bath application of cilobradine in brain slice preparation obtained from susceptible mice significantly reduces Ih currents and dose-dependently decreases pathophysiological firing activity of VTA dopamine neurons. Additionally, our in vivo studies show that acute intra-VTA infusion and systemic administration of cilobradine normalizes social avoidance behavior within one hour. Strikingly, both acute intra-VTA infusion and systemic i.p. injection of cilobradine produce a sustained 13-day duration of treatment efficacy. Moreover, our electrophysiological recording data determined that the pathological hyperactivity of VTA dopamine neurons in susceptible mice was significantly decreased during this 13-day period.

    Conclusions: This study provides a new line of evidence that directly targeting HCN channels induces acutely acting, long-lasting antidepressant effects.

    Keywords: Major depression, HCN, Ventral Tegmental Area, Rapid Depression Treatment, Ketamine

    Disclosure: Nothing to disclose.

    M124. AGN-241751, an Orally Bioavailable Positive NMDA Receptor Modulator, Exhibits Rapid and Sustained Antidepressant-Like Effects in Rodents

    John Donello*, Yong-Xin Li, Pradeep Banerjee, Kirk Bertelsen, Yuan-Xing Guo, Xiao-Lei Zhang, Patric Stanton, Roger Kroes, Amanda Gross, Joseph Moskal

    Allergan, Inc., Irvine, California, United States

    Background: Positive modulation of glutamatergic synaptic transmission by N-methyl-D-aspartate receptors (NMDARs) can produce rapid and sustained antidepressant effects. Rapastinel, a positive NMDAR modulator, has been shown to produce rapid and sustained antidepressant effects in animals and MDD patients. However, rapastinel is a tetrapeptide and is administered IV. Therefore, an orally bioavailable compound like AGN-241751 with rapastinel-like pharmacological properties offers distinct advantages over rapastinel for the treatment of depression. In this study, the pharmacological properties of AGN-241751 were characterized in various in vitro and in vivo assays.

    Methods: NMDAR activity was measured by a [3 H]MK-801 potentiation assay in membrane extracts prepared from human NR2A-D subtype-expressing HEK cells. NMDA induced intracellular calcium increases with or without AGN-241751 were analyzed with calcium indicator Fluo-4 in cultured rat brain cortical neurons. All test articles were dissolved in Mg2+-free extracellular medium (containing TTX and NBQX) and applied to the cells using a customized rapid drug application system that exhibits a 90% solution exchange time of less than a second. NMDAR mediated synaptic currents (Inmda) and synaptic plasticity were recorded in rat medial prefrontal cortical (mPFC) slices. Pharmacokinetic parameters of AGN-241751 were evaluated in male Sprague Dawley rats following oral dose administration. Concentrations of AGN-241751 in rat plasma, CSF and brain samples were determined by fit-for-purpose LC-MS/MS method. The antidepressant-like effects of AGN-241751 were tested in the rat forced swimming test (FST).

    Results: Relative to the maximal level of glutamate co-agonism by 1 mM glycine, AGN-241751 (1 fM to 0.1 µM) increased [3 H]MK-801 binding by 35% to 78%. Median effective concentration (EC50) values ranged between 0.1 pM at NR2C receptors to 0.6 nM at NR2B receptors. In cultured rat cortical neurons, AGN-241751 and rapastinel exhibited a similar modulation of NMDAR activity, but AGN-241751 was approximately 30X more potent. NMDA (10 µM) produced a small but significant increase in intracellular calcium in the absence of exogenous D-serine or glycine. While AGN-241751 (0.1-1000 nM) alone did not increase intracellular calcium influx, co-application of low concentrations of AGN-241751 (0.3 to 10 nM) with 10 µM NMDA produced ~30% potentiation of the NMDA-induced calcium signal. Similar to rapastinel, the potentiation of NMDAR signaling by AGN-241751 is independent of the NMDAR glycine site. In mPFC, AGN-241751 (20 -100 nM) increased Inmda and enhanced the magnitude of long-term potentiation (LTP).

    Following a single oral dose of AGN-241751 in male Sprague Dawley rats, plasma, brain, and CSF concentrations were quantified. AGN-241751’s oral bioavailability was >95%, and it demonstrated dose proportional increases in plasma, CSF, and brain exposure up to doses of 10 mg/kg. CSF/plasma ratios were ~0.15. In the rat FST assay, a single oral dose of AGN-241751 (3 - 100 µg/kg) produced a dose-dependent and rapid antidepressant-like effect which lasted for 7-14 days. A single antidepressant dose of AGN-241751 also induced a dose dependent increase in mPFC metaplasticity.

    Conclusions: AGN-241751 belongs to a novel class of positive NMDAR modulators that act independently of the glycine site but act as modest glutamate co-agonists. In vitro NMDAR-dependent functional assays demonstrate that AGN-241751 positively modulates NMDARs similar to rapastinel but is significantly more potent (~ 30×). Consistent with its modulation of mPFC NMDARs, AGN-241751 also increases mPFC LTP and metaplasticity. Oral administration of AGN-241751 produces dose dependent increases in plasma, brain, and CSF concentrations which yields dose dependent anti-depressant-like activity. The beneficial properties of AGN-241751 represent a significantly improved therapeutic profile for the development of therapeutics for major depressive disorders.

    Keywords: Depression, NMDA Receptor, AGN-241751, Rapastinel, Rapid-acting Antidepressant

    Disclosure: Allergan, Inc., Employee

    M125. Intrinsic Dopaminergic Activity Underlying Sex-Specific Differences in the Stress Response

    Mary Shanley, Christopher Guevara, Allyson Friedman*

    Hunter College, New York, New York, United States

    Background: Sex differences in the incidence of depression frequently emerge during adolescence. This suggests that female stress susceptibility may be linked to surges in reproductive hormones that occur during puberty. Interestingly, the β isoform of the estrogen receptor is expressed in a population of mesolimbic dopaminergic neurons in the ventral tegmental area (VTA) which, in males, have been linked to changes in social stress induced depressive behaviors. These VTA neurons exhibit a range of activity modes, varying in frequency and degree of burst firing. In females, in vivo VTA dopamine neuron bursting activity is affected by the phase of the estrous cycle. Functionally, these differences in activity have been shown to contribute to the release of dopamine. Previous work has shown that a single pulse activation of these neurons releases twice as much dopamine in high estrogen females in the VTA- nucleus accumbens (NAc) pathway, compared to both males, and low estrogen females. We explored the physiological role of estrous cycle on dopaminergic regulation, as well as dysregulation, following social stress. We demonstrate that underlying dopamine neuron activity changes are intrinsic neuroadaptations that are driven by the fluctuations in reproductive hormones. Further, we propose that these intrinsic changes in channel function affect the female social stress response.

    Methods: We utilized electrophysiological cell-attached and whole-cell recording of VTA neurons from freely cycling female C57BL/6 J mice. We used a series of voltage-clamp protocols, in combination with pharmacological agents, to isolate A-type (fast) and M-type (slow) potassium channel function which influence DA neuron bursting profiles. To determine the effect of estrous cycle hormone fluctuations on neuronal activity we used vaginal cytology and estradiol 17β ELISA measurements to correlate hormone levels with channel function. In in vivo pharmacological experiments, female mice were subjected to local infusions to the VTA of bio-available estradiol or ICI 182,780 (estrogen receptor antagonist) prior to social stress. Mice were then tested on social interaction and novelty suppressed feeding 24 h post-social stress.

    Results: The in vitro firing rate of VTA neurons demonstrated minimal variance across the estrous cycle in stress naïve mice. Interestingly, we found significant differences in K+ channel function that correlated with the phases of the estrous cycle. Despite these channel function differences in stress naïve mice we found no influence on social interaction and novelty suppressed feeding behaviors. Conversely, 24 h following an acute social stress, behavioral testing revealed that the estrogen dependent intrinsic differences mediated significant differences in the behavioral responses. Next, we determined that estradiol was in part acting through the VTA by directly infusing either estradiol or an estrogen receptor antagonist into the VTA prior to the social stress and testing 24 h later.

    Conclusions: Throughout the estrous cycle the reward system responds differently to salient stimuli, an effect that could increase the probability of social stress inducing dopamine dysregulation. Our findings demonstrate that estradiol modulates VTA neuron physiology, which may contribute to sexual dimorphism in social stress induced depression vulnerability.

    Keywords: Acute Stress, Estradiol, Dopamine

    Disclosure: Nothing to disclose.

    M126. Affective Dysregulation and Dopamine Hypofunction in Female Rodents During the Early Postpartum Period

    Millie Rincón-Cortés*, Anthony Grace

    University of Pittsburgh, Pittsburgh, Pennsylvania, United States

    Background: Background: The onset of motherhood is often accompanied by alterations in mood, including anxiety and depressive-like symptoms, in humans and rodents (Fernández et al., 2014; O’Hara et al., 2014). The highest rates of anxiety and depression occur during the first few weeks, months or year postpartum: most new mothers experience mild transient affective dysregulation (i.e. postpartum blues) during the 1-2 weeks postpartum (Pawluski et. al 2017). In accordance, animal models of postpartum depression have also reported time-dependent effects on depressive-like behavior and anhedonia (Brummelte and Galea, 2010), with negative effects being greater in early/mid postpartum females (Haim et al., 2014). Yet there is little data regarding the neural adaptations that occur in response to parity (i.e. the condition of having borne offspring) and shortly after giving birth that may be associated with these affective changes.

    The dopamine (DA) system has traditionally been associated with anhedonia, the inability to derive pleasure from normally rewarding stimuli, and has been repeatedly implicated in the pathophysiology of depression. A causal link between a hypofunctioning DA system (i.e. decreased DA neuron activity) and stress-induced depression-related behaviors (i.e. anhedonia, despair, anxiety) has been demonstrated in animal models (Tye et. al, 2013; Chang and Grace, 2014), with females showing greater effects (Rincón-Cortés and Grace, 2017). However, little is known about the DA system in females following reproductive experience, including postpartum DA function.

    Methods: The elevated plus maze (EPM) and the three-chambered social approach test (SAT) were used to compare anxiety-like behavior and social motivation in virgin and postpartum female rats across different timepoints (i.e. 1d-PP, 3d-PP, 1week-PP; n=9-16 per group). A separate cohort of animals was tested during late postpartum (i.e. 22-24d-PP; n=14-18 per group). Single-unit recordings of VTA DA neurons (n=7-12 animals per group) were conducted the day following behavioral testing and 3 parameters were measured: i) the number of spontaneously active DA cells (i.e. population activity or cells/track), ii) basal firing rate and iii) firing pattern (i.e. the percentage of spikes firing in bursts). Comparisons of 3 or more groups were analyzed using one-way ANOVA, whereas comparisons between 2 groups were analyzed using t-tests.

    Results: 1d-postpartum female rats exhibited increased anxiety-like behavior in the EPM, as indexed by reduced open arm entries (p <0.01) and time spent in the open arms (p <0.001) compared with virgin or 1wk-postpartum rats. Early (1d-,3d-) postpartum female rats exhibited reduced social motivation (p < 0.05), as indexed by decreased social sniff time and fewer number of crossings into the social chamber (p < 0.05) compared with virgins. Early (1d-,3d-) postpartum female rats females exhibited an attenuation of VTA population activity, as indexed by a decrease in the number of active DA cells per electrode track, compared with virgin rats (p < 0.01) but no differences in firing rate or the percentage of spikes occurring in bursts. At 1-week postpartum, a split was observed in which approximately half the dams exhibited high neurobehavioral dysregulation (i.e. low social motivation, DA hypofunction), suggesting a link between impaired social motivation and attenuated DA function. None of these behavioral or electrophysiological effects were observed in late (22-24d) postpartum female rats compared with virgins.

    Conclusions: Collectively, our findings suggest that parity-driven changes in affect and VTA DA neuron activity vary across the postpartum period and that the early postpartum period is a time of enhanced affective dysregulation, which is accompanied by reduced DA activity, in female rats.

    Keywords: Postpartum, Dopamine, Mood and Anxiety Disorders, Electrophysiology, Ventral Tegmental Area (VTA)

    Disclosure: Nothing to disclose.

    M127. Human Experimenter Sex Modulates Mouse Behavioral Responses to Stress and to the Antidepressant Ketamine

    Polymnia Georgiou*, Panos Zanos, Jaclyn Highland, Carleigh Jenne, Brent Stewart, Danielle Gerhard, Ronald Duman, Todd Gould

    University of Maryland School of Medicine, Baltimore, Maryland, United States

    Background: Lack of replicability of experimental results in different laboratories may be due to unexpected experimental variables that are unrecognized, and therefore not appropriately controlled. The sex of the human experimenter is rarely considered as a biological variable that can affect experimental results and is not usually reported or required in the experimental methods. However, there is some evidence that rodents may be able to differentiate the sex of human experimenters who handle them, and this discrimination could have measurable effects on their biological/behavioral responses. In particular, exposure of rodents to the scent of male experimenters was previously shown to result in increased anxiety and stress-induced analgesia in mice compared with an exposure to the scent of female experimenters.

    Methods: We investigated the effects if the sex of the experimenter on baseline and stress-induced behaviors in mice, and the reversal of such behaviors by antidepressant drugs. In addition, we assessed the involvement of distinct stress systems on experimenter sex-induced differences on mouse stress-related behaviors. The effects of the sex of human experimenter was also investigated on electroencephalographic (EEG) and biochemical outcomes. All experimental procedures were approved by the University of Maryland, Baltimore Animal Care and Use Committee and were conducted in full accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Results were analyzed by a two-way ANOVA followed by Holm-Sidak post-hoc test when a significant interaction effect was observed.

    Results: We show that mice manifest differential behavioral and stress responses to human odors (n=8 experimenters/group; n=16 mice/group). Administration of the antidepressant drug ketamine to mice by male experimenters reversed stress-mediated behavioral responses, while such responses were absent following injection of the drug by a female experimenter (n=4 pairs of experimenters; n=10/group). Similar sex-dependent experimenter effects were identified with the ketamine metabolite (2 R,6 R)-hydroxynorketamine (n=8 experimenters/group; n=16 mice/group), but not with other mechanistically distinct antidepressants (1 pair of experimenters; n=10 mice/group). The nearby presence of a female experimenter was sufficient to block antidepressant actions of male-administered ketamine (1 pair of experimenters; n=10 mice/group). Non-antidepressant behavioral actions of ketamine where present regardless of the sex of the experimenter (1 pair of experimenters; n=10 mice/group). Ketamine administration induced both overlapping and distinct EEG and biochemical profiles dependent upon the sex of the experimenter (n=3 experimenters/group; n=16/group). Manipulation of different mouse stress systems prior to treatment yielded differential responses to male and/or female-administered ketamine (n=5 experimenters/group; n=9-10 mice/group).

    Conclusions: We establish that male scent is necessary to induce ketamine’s antidepressant effects in some model systems in a manner dependent upon specific stress hormones. Overall, these findings demonstrate the importance of the sex of the experimenter on a subset of experimental outcomes. Our data argue that the sex of the experimenter may affect replicability and should be considered as a potential experimental variable in neuropsychopharmacology research involving rodent behavioral models.

    Keywords: Ketamine, Hydroxynorketamine, Reproducibility, Design, Replication, Replication, Sex Differences

    Disclosure: Nothing to disclose.

    M128. Ovarian Hormones Causally Contribute to the Prophylactic Efficacy of (R,S)-Ketamine and (2 R,6 R)-Hydroxynorketamine in Female Mice

    Briana Chen, Christina LaGamma, Rebecca Brachman, Xiaoming Xu, Shi-Xian Deng, Donald Landry, Christine Ann Denny*

    Columbia University, New York, New York, United States

    Background: Exposure to stress is a major risk factor for mood disorders, such as major depressive disorder (MDD) and post-traumatic stress disorder (PTSD). However, stress does not universally cause disease. Stress resilience, the ability to adapt to stress without developing psychopathology, varies in individuals. For example, although women experience trauma at significantly lower rates than men, they are twice as likely to develop MDD. We previously reported that a single injection of (R,S)-ketamine prior to stress protects against the onset of depressive-like behavior and buffers against a deleterious fear response in male mice. However, the prophylactic efficacy of ketamine and ketamine metabolites in female mice remain largely unknown. Therefore, the goal of this study was to determine whether prophylactics can be developed for use in females and whether metabolites of (R,S)-ketamine can have the same prophylactic efficacy as their precursor.

    Methods: Female 129S6/SvEv mice were administered (R,S)-ketamine, (2 R,6 R)-HNK, or (2 S,6 S)-HNK at various doses 1 week before one of a number of stressors, including contextual fear conditioning (CFC), learned helplessness (LH), and chronic immobilization stress (CIS). Prophylactic efficacy was validated using the forced swim test (FST). In a separate set of experiments, we examined whether sex hormones influenced the efficacy of prophylactic compounds.

    Results: We found that (R,S)-ketamine and (2 R,6 R)-HNK, but not (2 S,6 S)-HNK, significantly reduced immobility in the FST compared to saline controls. Interestingly, in females, ketamine was prophylactic at a lower dose than previously shown in males. Moreover, (2 R,6 R)-HNK was prophylactic at a significantly smaller dose and at a faster rate than its precursor (R,S)-ketamine. Moreover, we determined that the prophylactic efficacy of these compounds may be mediated by ovarian-derived hormones.

    Conclusions: Overall, these data indicate that (R,S)-ketamine and (2 R,6 R)-HNK are effective prophylactics against a variety of stressors in females and that their sex-specific effects may be modulated by gonadal hormones. To our knowledge, this is the first demonstration that (2 R,6 R)-HNK may possess the same prophylactic properties of its precursor (R,S)-ketamine. Our findings offer insights into the prevention of stress-related impairments in a susceptible population and may further elucidate underlying sex-specific neuropathology contributing to the onset of MDD.

    Keywords: Racemic Ketamine and Metabolites, Depression, Sex Differences, Sex Hormones

    Disclosure: Provisional Patent, Patent

    M129. Cholinergic Innervation of the Hippocampus Regulates Stress-Induced Behaviors

    Yann Mineur*, Cristiana Pineda, Laura Vanopdenbosch, Ashraful Islam, Tenna Mose, Richard Crouse, David Thompson, Matthew Bentham, Wengliang Zhou, Marina Picciotto

    Yale University School of Medicine, New Haven, Connecticut, United States

    Background: Acetylcholine (ACh) acts through multiple cholinergic receptors to regulate neurobiological systems controlling behaviors related to depression, anxiety and stress resilience. Clinical imaging studies suggest that nicotinic acetylcholine receptor (nAChR) occupancy by ACh is increased in the brains of human subjects with major depression or bipolar disorder during a depressive episode. Conversely, blockade of either nicotinic or muscarinic ACh receptors can have antidepressant effects in human subjects and in preclinical models. Stress can increase ACh levels in specific brain areas and alter the activity of acetylcholinesterase (AChE), the primary enzyme degrading ACh. Further, blocking AChE systemically leads to depression-like symptoms in humans and in mouse models, while increasing ACh signaling locally in the hippocampus is sufficient to induce behaviors related to anxiety and depression. Thus, it is critical to determine how the cholinergic innervation of the hippocampus is involved in behaviors related to anxiety, depression, and stress resilience.

    Methods: We first examined the changes of AChE (as a proxy for ACh concentration) in several brain regions, including the hippocampus and the medial septum, following restraint stress in mice. We then combined multiple strategies including optogenetics, chemogenetics, immunohistochemistry, pharmacology, and electrophysiology in wild-type and transgenic mice expressing Cre in cholinergic neurons (ChAT-Cre) to determine how modulating the cholinergic neurons of the hippocampus or of the septo-hippocampal pathway could alter physiology and behavior at baseline and in a hypercholinergic model of anxiety- and depression-like states. Most experiments were performed in two independent ChAT-Cre lines (BAC ChAT-Cre and ChAT-IRES-Cre) to account for potential line effects.

    Results: Stress had significant region-dependent effects on AChE activity, with the greatest regulation occurring in ventral hippocampus. Few behavioral effects were observed when cholinergic neurons innervating the hippocampus were silenced or activated by infusing AAV-floxed-Gi-DREADD or AAV-floxed-Gq-DREADDs, respectively, into the medial septum of ChAT-Cre mice. However, the increases in stress-related behaviors induced by physostigmine were attenuated by inhibiting ChAT-Cre positive neurons, confirming that cholinergic inputs from the septo-hippocampal pathway are activated in response to stress, and that this pathway mediates the increased anxiety- and depression-like behaviors induced by cholinesterase antagonism. The lack of effect seen following DREADD stimulation of the medial septum in the absence of physostigmine could be due to projections that are not selective for the hippocampus. We therefore activated neurons projecting to the hippocampus selectively via optogenetic stimulation of septo-hippocampal terminals but did not selectively activate cholinergic terminals and found that stimulated mice exhibited less exploration of an anxiogenic environment and more immobility in tests of antidepressant efficacy, along with decreased interaction in a social defeat paradigm. To reinforce that these behavioral changes were due to activation of cholinergic inputs to the hippocampus, a retro-AAV-floxed-Gq-DREADD was infused into the hippocampus of ChAT-Cre mice. Consistent with optogenetic experiments in wild type mice, CNO-treated ChAT-Cre mice exhibited increased anxiety- and depression-like phenotypes and decreased stress resilience. Finally, stimulating the sparse population of ChAT-Cre-positive cells intrinsic neurons in the hippocampus using AAV-floxed-Gq-DREADD increased anxiety-like behaviors, immobility in tests of antidepressant efficacy and interaction in the social defeat paradigm, but only in BAC ChAT-Cre mice. These effects were partially reversed by locally infusing the broad nicotinic antagonist mecamylamine into the hippocampus, suggesting that this effect was nAChR-dependent.

    Conclusions: Taken together, these results demonstrate that stress can alter expression of the primary degradative enzyme for ACh, thereby modulating cholinergic signaling in the hippocampus. Activating ChAT-positive neurons in the hippocampus, or specifically modulating the cholinergic inputs from the medial septum/diagonal bands of Broca, can induce behaviors related to anxiety and depression in the mouse. Conversely, the results also suggest that activating non-hippocampal septal projections could have opposite effects on these stress-related behaviors.

    Keywords: Acetylcholine, Hippocampus, Medial Septum, Depression, Stress

    Disclosure: Nothing to disclose.

    M130. Pilot Study of an Intracranial Electroencephalography Biomarker of Co-Morbid Depression in Epilepsy

    Katherine Scangos*, Lowry Kirkby, Kristin Sellers, Alia Shafi, Heather Dawes, Andrew Krystal, Edward Chang

    University of California, San Francisco, Sacramento, California, United States

    Background: Adult patients with epilepsy have an increased prevalence of major depressive disorder (MDD) and other psychiatric co-morbidities. MDD in epilepsy is associated with worse outcome and quality of life. However, it continues to be underdiagnosed and untreated and further attention to this comorbidity is critical. Intracranial electroencephalography (iEEG) captured during extended inpatient monitoring of patients with treatment-resistant epilepsy who are potential surgical candidates offers a particularly promising method to study MDD networks in adult epilepsy, offering both high temporal resolution and spatial precision. Despite the enormous potential of iEEG, there are no studies to date that examine the neurophysiological signatures of network dysfunction in mood and anxiety disorders in patients with epilepsy. Such studies are critical in order to better understand the etiology of co-morbid MDD and could lead to the development of novel personalized therapies. In this pilot study high-density intracranial electroencephalography (iEEG) in humans was utilized to investigate the neural activity patterns within corticolimbic structures that reflect the presence of co-morbid MDD in epilepsy.

    Methods: We examined 24 h of resting state iEEG recordings from a preliminary sample of 14 human subjects undergoing surgical treatment for medication refractive epilepsy. Eight patients (57%) had co-morbid MDD [Patient Health Questionnaire-9 (PHQ-9) ≥10] and 6 had mild/no MDD (no-MDD, PHQ-9 < 10). All patients had neural recordings from 5 brain regions: orbitofrontal cortex, amygdala, hippocampus, cingulate, and insula. Standard iEEG/ECoG pre-processing techniques were used. Continuous waveform transformation using the Morlet transform wavelet method was performed in 30 s intervals to obtain power spectra in 5 standard frequency bands. Relative power was calculated by dividing the power of each frequency band by the total power for each electrode. Principal component analysis was performed on the z-scored mean relative power across 24 h of iEEG recordings and across electrodes within a region and combined across subjects. Stepwise linear and logistic regressions were performed with PHQ-9 score or presence of moderate-severe MDD as the outcome variable respectively, and principle components (PCs) of power in a particular frequency band as the independent variables. A logistic classifier model was used to determine the ability of the biomarker to correctly classify patients with and without MDD. The standard “leave-one-out” cross-validation method provided a preliminary indication of whether results would generalize to an independent dataset.

    Results: A forward stepwise linear regression model was significant with 4 PCs accounting for 70% of the variance in the PHQ-9 score (p = 0.0178, R2 = 0.70, ‘leave-one-out’ cross-validation: R2 = 0.25). PC1 was found to account for the majority of the variance in correlation with the PHQ-9 score (R2 = 0.29, p = 0.046, coeff = 0.82), with the 3 other PCs contributing about equally to the remainder. The 4 PCs were heavily dependent on beta and theta power, with greater MDD symptom severity associated with greater beta power throughout the entire 5 region network, higher subcortical theta power and lower cortical theta power.

    A forward stepwise logistic regression model was significant with 2 PC’s (PC1,PC2) accounting for the majority of the difference between the MDD and no-MDD groups (R2 = 0.40, p = 0.023; ‘leave-one-out” cross-validation R2 = 0.17) with an odds ratio of 1.5 for PC1 and 2.0 for PC2. A logistic classifier model using these two PC’s correctly classified 79% of patients (sensitivity 75% and specificity of 83%, leave-one-out cross validation).

    Consistent with the results from the PCA analysis, mean beta power across the network was significantly greater by ~14% in the MDD group than in the no-MDD group (p = 0.009). Similarly, an analysis of variance revealed a significant group (MDD vs. no-MDD) by location (subcortical, cortical) interaction effect for theta frequency power (p = 0.0133, F(1,66) = 6.13). Cortical theta power was significantly lower in the depressed group (p = 0.003, ttest), while subcortical theta was numerically, although non-significantly, higher in this group. Using these three features alone (mean beta power, subcortical theta power, and cortical theta power) a logistic classifier model correctly classified 12/14 patients correctly (86% accuracy, sensitivity 88%, specificity 83%, leave-one-out cross validation).

    Conclusions: In this pilot dataset, we have identified a putative biomarker of co-morbid MDD in medication refractive epilepsy patients that is correlated with the severity of their MDD symptoms and can correctly classify patients with and without MDD with 86% accuracy. This finding is the first of its kind to address circuit level activity within the overlapping circuits involved in epilepsy and co-morbid psychiatric conditions. Further research is indicated to examine what aspects of this activity and its time-course contribute most strongly to the differences observed.

    Keywords: Depression, Biomarker, Intracranial EEG, Epilepsy

    Disclosure: Nothing to disclose.

    M131. Effect of Social Stimulation on Ethanol Intake and Gene Expression in Mice

    Lei Yu*, Arthur Tomie, Shivani Mehra, Usman Toor, Minjie Bao, Akhila Manthena

    Rutgers University, Piscataway, New Jersey, United States

    Background: Social stimulation may play a role in ethanol consumption. There is considerable evidence that socializing stimulates ethanol drinking in humans and in animals. Humans drink more ethanol in the presence of social interactions than in the absence of such interactions. Previous studies in rodents have shown that, when put in social situations, mice change their behavior and drinking habits. Interestingly, we noticed that, although mice display a similar overall pattern of change for social interaction-mediated ethanol consumption, the extent of change in ethanol consumption varies substantially among individual mice, suggesting genetic factors at play. In particular, differences were noted with different gender pairing.

    Methods: To investigate the effect of social stimulation on ethanol intake and gene expression, mice were given a free choice of water and ethanol, while in the presence of companion mice of the same or opposite gender.

    Results: Some mice drank a significantly higher amount of ethanol than others, depending on the gender pairing of the social companion mouse. Brain tissues were collected, and the steady-state levels of candidate gene messengers were measured. Our results suggest that certain dopamine receptor genes, as well as selected GABA receptor genes, appear to display significant correlation with social stimulation-mediated effect.

    Conclusions: Social stimulation plays a role in the extent of free choice ethanol intake. Gender pairing during social stimulation is a factor, and dopaminergic and GABAergic neurotransmission pathway genes appear to be affected.

    Keywords: Ethanol, Social, Gender, Gene Expression

    Disclosure: Nothing to disclose.

    M132. Human-Based Exome Capture Performed in Rhesus Macaques Identifies a GABRA6 Polymorphism That Predicts Individual Differences in Neonatal Imitation Behavior

    Ravi Kumar Sindhu*, Carlos Driscoll, Qiaoping Yuan, Stephen G. Lindell, Annika Paukner, James Kohler, David Goldman, James Dee Higley, Stephen J. Suomi, Pier Francesco Ferrari, Christina Barr

    National Institute of Health, Rockville, Maryland, United States

    Background: We have been examining genetic factors that contribute to variation in the neurobiological systems that influence reward pathways, sociality, impulsivity and stress reactivity, as these systems may also influence alcohol use and addiction vulnerability in modern humans. In order to interrogate these systems, we wanted to sequence the exomes of rhesus macaque subjects that were selected based on variation in temperament. The major advantage of using Whole Exome Sequencing (WES) is that DNA extracted from whole blood can be used without the need for extracting DNA from the tissue of interest. WES allows for the capture of large insertion/deletions and SNPs and is best for either non-synonymous or frameshift mutations. While there are no commercially available reagents for performing WES in nonhuman primates, a human whole exome platform is available.

    Methods: Exonic sequences were enriched using the Agilent SureSelect all exon capture array targeting ~38 Mb and were annotated using Ensemble genes, refGene and refSeq xenoRefGene for H. sapiens & M. musculus from UCSC. Sequence alignment was performed using CASAVA, Recalibration & SNP calling using GATK, Annotations, filtering, counts and calculations using custom scripts, and visualizations using R. Missense variants’ functionalities were inferred using Polyphen. PolyPhen posterior probability scores associated with false positive rate (FPR) below 5% are characterized as ‘probably’ damaging and those below 10% are determined to be ‘possibly’ damaging under the HumDiv classifier model. Focusing on candidate genes relating to risk for alcohol use disorders (1536 marker Addictions Array, LNG), we identified a number of non-synonymous SNPs, some of which were predicted to be “damaging” or “potentially damaging” by in silico analysis (PolyPhen). Among these was a SNP in the GABARA6 gene, a gene confined in its expression to the cerebellum and which has known roles in alcohol response in both rodents and humans.

    Given a putative role of the cerebellum in mirroring behavior, we wanted to determine if Tyr28Phe genotype predicted individual differences in neonatal imitation. On postnatal days 1-14, infants were presented with two different stimuli: 1) a lipsmacking gesture (LPS) which is a common affiliative gesture in macaque and consists of the rapid opening and closing of the mouth; and 2) a nonsocial control condition (a white plastic disk with black/red or green/yellow orthogonal stripes slowly rotated clockwise and counter-clockwise). The neonatal imitation study was conducted on nursery-reared infants, raised at the Laboratory of Comparative Ethology at the National Institutes of Health in eight cohorts between 2007 and 2014. All procedures described below adhered to the NIH Guide for the Care and Use of Laboratory Animals and were approved by the NICHD Animal Care and Use Committee.

    Results: The human exome capture kit was effective for sequencing rhesus macaque exomes. Using this approach, we obtained good coverage at capture sights. Also bearing on the validity of the SNP detection pipeline, the transition/transversion ratio of human and macaque SNPs was non-random (2:1). We observed approximately 25,000 missense variants, among which was a nsSNP in the GABRA6 gene.

    We found that GABRA6 genotype predicted the degree to which infant macaques imitated the behavior of a human caregiver. Carriers of the 28Phe allele exhibited higher levels of lip-smacking imitation (an affiliative gesture in macaques) than did infants homozygous for the ancestral allele (P < 0.01).

    Conclusions: We used a human capture kit to effectively sequence the genome of rhesus macaques, which are 25 mya diverged from humans. In so doing, we identified a damaging nsSNP in the GABRA6 gene. Studies in humans and rodents have demonstrated that variation at the GABRA6 gene predicts individual difference in alcohol response. As the GABRA6 gene, which encodes the α6 subunit, is highly expressed in cerebellar granule cells, it is not surprising that functional variation at this gene influences alcohol-induced ataxia. One other system that is under control by the cerebellum and that is also reduced in its function in various psychiatric disorders is the mirror neuron system. The cerebellum is involved in the action-perception mechanism and is likely linked to imitative behaviors and probably other social cognitive skills. Notably, some motor impairments in children with autism -and their limited capacity to build an internal motor representation of actions- are thought to be based on a motor loop between the traditional parietal-premotor mirror circuit and the cerebellum. Here, we demonstrate that a non-synonymous SNP at the GABRA6 gene predicts individual differences in neonatal imitation behavior in rhesus macaques. This might suggest that functional variation at GABRA6 in humans could moderate risk for psychiatric and neurodevelopmental disorders.

    Keywords: Whole Exome Sequencing, GABRA6, Next Generation Sequencing, Alcoholism

    Disclosure: Nothing to disclose.

    M133. Regional Brain Creatine Relates to Cognitive Function: A Multi-Voxel 1H-MRS Study in First Episode Psychosis and Clinical High Risk for Psychosis

    Jeremy Watts*, Sofia Chavez, Cory Gerritsen, Huai-Hsuan Tseng, Michael Kiang, Michael Bagby, Romina Mizrahi

    University of Toronto, Centre for Addiction & Mental Health, Toronto, Canada

    Background: Brain energy metabolism is thought to be disrupted in schizophrenia and despite evidence that brain energy systems, such as creatine-phosphocreatine, are involved in cognition and are altered in schizophrenia, few in vivo studies have investigated the relationship between regional brain creatine and cognitive deficits in schizophrenia.

    In vivo imaging studies in psychosis report lower rates of phosphocreatine formation in first episode psychosis (FEP). Further, development of neuropathologies of brain metabolism may precede the onset of frank psychotic symptoms, as observed in individuals at clinical high risk of developing psychosis (CHR). We aim to investigate regional brain creatine and cognition in psychosis by measuring neurocognitive performance and regional total creatine+phosphocreatine (hereafter, ‘creatine’) levels using 1H-magnetic resonance spectroscopy (1H-MRS) in 114 participants comprising FEP, CHR and healthy controls (HC).

    Methods: FEP patients (n=31) and CHR (n=43) were excluded if concurrent diagnoses better explained the symptoms for psychotic disorders or CHR criteria. HC (n=40) had no past or current DSM IV axis I disorder. All participants underwent 1H-MRS scans for one or more of the five regions of interest, for a total of 388 individual voxels.

    Single voxel 1H-MRS spectra were obtained using a 3 T scanner and a standard sequence using point-resolved spectroscopy. Voxels were positioned on the left hippocampus, left dorsal caudate, left dorsolateral prefrontal cortex (DLFPC), bilateral supra-genual anterior cingulate cortex (sgACC) and bilateral PMC. Binary voxel masks were applied to segmented T1 images in order to correct metabolite concentrations for tissue composition.

    We used the Repeatable Battery for the assessment of neuropsychological status; ‘RBANS’) to assesses immediate and delayed memory, attention, visuospatial ability and language. Participants were antipsychotic-free at the time of the study except for six FEP patients and ten CHR taking antipsychotic drugs at low doses (i.e., below 100 mg chlorpromazine equivalents).

    Data were analyzed using a generalized linear model with effect of group, creatine and their interaction, with RBANS total score as the dependent variable. Interactions were followed by between-group contrasts of regression β values. Main effects were followed up by a GLM for RBANS subscales.

    Results: Our preliminary analysis found that lower creatine in PMC was associated with lower total cognitive scores (F(1,53)=13.669, p=.001) and interacted with group (F(2,53)=8.873, p=.0005) such that creatine was more strongly associated with total cognitive score in HC (β=11.32; p=.027) and FEP (β=15.54; p=.01) as compared to CHR (β=-0.27). In contrast to observations in PMC, total cognitive scores were not significantly associated with creatine in hippocampus (F2,61=0.278, p=.759), striatum (F(2,59)=0.286, p=.752), DLPFC (F(2,95)=0.185, p=.832), or sgACC (F(2,96)=0.261, p=.771).

    Lower PMC creatine was associated with poorer immediate memory (F(1,53)=8.873, p=.0005), and this effect differed by group (F(2,53)=8.873, p=.0005) such that creatine was more strongly associated with immediate memory score in HC (β=10.89; p=.009) and FEP (β=21.21; p=.0003) compared to CHR subjects (β=-4.42). Lower PMC creatine was significantly associated with better visuospatial ability (F(1,39)=8.607, p=.006) across all groups, and this effect did not differ between groups. No associations were observed between creatine and other cognitive domains in PMC or other regions.

    Conclusions: Preliminary results of this study support a role for PMC total creatine levels in overall cognition and immediate memory in HC and FEP but not in CHR. This result is in line with the PMC’s high basal energy consumption and its role as a major network hub involved in memory. That these effects were not observed in CHR may result from the high clinical heterogeneity of this group. Thus, the primary result of this study suggest that the domain of immediate memory may be particularly sensitive to total creatine levels in the PMC.

    Keywords: psychosis, clinical high risk for psychosis, Cognition, imaging, energy metabolism

    Disclosure: Nothing to disclose.

    M134. Imaging Alpha7 Nicotinic Acetylcholine Receptors in PTSD: Preliminary Findings and Sex Differences

    Ansel Hillmer*, Irina Esterlis, Maggie Davis, David Matuskey, Henry Huang, Richard Carson, John Krystal, Kelly Cosgrove

    Yale University School of Medicine, New Haven, Connecticut, United States

    Background: Posttraumatic stress disorder (PTSD) affects over 7% of Americans, yet the underlying neurochemical factors are poorly understood, and limited pharmacotherapies are available. The α7 nicotinic acetylcholine receptor (nAChR) is present on glutamatergic neurons connecting the medial prefrontal cortex (mPFC) with the amygdala and hippocampus. These receptors modulate plasticity in these important circuits implicated in PTSD. Here, we aimed to measure α7 nAChRs in PTSD and controls with positron emission tomography (PET) brain imaging, with a focus on possible sex effects in people with PTSD.

    Methods: Study participants included eleven individuals (5 females, 6 males) with a SCID-5 diagnosis of PTSD and 6 age-matched healthy control males. Data collected from all subjects included MRI and PET scans and mood and PTSD symptom assessments. Individuals with PTSD exhibited moderate to severe PTSD symptomatology (PCL-5 Scores: females = 52±7; males = 53±16). The α7 nAChR specific radioligand [18 F]ASEM was injected as a 338±55 MBq bolus. Imaging data and metabolite-corrected arterial blood samples were collected for 120 min. α7 nAChR availability was indexed by the total distribution volume (VT), estimated using multilinear analysis in brain regions of amygdala, hippocampus, and mPFC. Cohen’s d effect sizes were estimated for comparisons among three groups: males with PTSD, females with PTSD, and male controls.

    Results: For all regions, group average VT values followed the rank order of: PTSD males < PTSD females < control males. When comparing PTSD males with PTSD females, medium effect sizes of 0.7, 0.7, and 0.5 were observed in amygdala, hippocampus, and mPFC, respectively. Comparing PTSD males with control males revealed very large effect sizes of 1.6 and 1.2 in amygdala and hippocampus, with a moderate effect size of 0.6 in mPFC. We did not observe any associations between receptor availability and symptom severity in any of the groups.

    Conclusions: These preliminary findings suggest altered α7 nAChR availability in men with PTSD. Although a healthy control female comparison group is not available at this time, our data suggest possible sex differences in α7 nAChR availability in brain regions associated with emotional processes. Work is ongoing to determine the role of α7 nAChR in PTSD, specifically by sex.

    Keywords: PTSD, Alpha-7 Nicotinic Acetylcholine Receptor, PET Imaging, Sex Differences

    Disclosure: Nothing to disclose.

    M135. Fiber Cluster Topography Analysis of the Pattern of Structural Connectivity in Frontostriatal Circuitry in Healthy Subjects

    James Levitt*, Fang Zhang, Mark Vangel, Marek Kubicki, Martha Shenton, Lauren O’Donnell

    VA Boston Healthcare System, Brigham and Women’s Hospital, Harvard Medical School, Brockton, Massachusetts, United States

    Background: Alterations in brain connectivity may underlie neuropsychiatric conditions such as schizophrenia or autism. We here assess the pattern of structural connectivity between the prefrontal cortex (PFC) and striatum in 100 healthy subjects (HSs) from the Human Connectome Project (HCP); age: 22 to 35; sex: 46 females and 54 males. We propose a novel method, using fiber clustering of dMRI tractography to assess the pattern of frontostriatal structural connectivity, which allows us to quantify the degree of pattern deviation from a topographic, parallel, arrangement.

    Methods: To enable the identification of fiber tract parcels from the prefrontal cortex (C) and the striatum (S), we used a data-driven fiber clustering atlas (Zhang et al, 2018) that allows for a whole brain tractography parcellation into 2000 fiber clusters according to the white matter (WM) anatomy (i.e., fiber geometric trajectory). Then, fiber clusters of interest (i.e., from C to S) from the whole brain WM were identified according to their connected anatomical brain regions. We studied multiple Freesurfer PFC regions including orbital, lateral and medial PFC regions and the caudate. We identified 17 fiber clusters that connect C and S. To quantify the topographical relationship of these fiber clusters, we measured the mean distances between the endpoints of the fiber clusters within the prefrontal cortex (i.e., cortical distance) and the mean distances between the endpoints of the corresponding fiber clusters terminating in the striatum (i.e., striatal distance).

    Results: We plotted the data in 2 ways. First, we generated a plot (not shown) based on the 17 fiber clusters (with 136 pairs of fiber clusters, yielding 136 data points), showing the relationship between the cortical distances and the corresponding striatal distances of the obtained fiber cluster pairs that connect the prefrontal cortex and the caudate. A 2-term exponential model was fit to the data points which was superior to a linear model. We showed that the PFC-striatal WM streamline projection pattern was non-linear, which was driven by the results from 10 cluster pairs. Certain clusters, for example cluster 6, representing a cluster originating in pars orbitalis PFC, were significantly over-represented in these 10 cluster pairs. Second, we generated plots (not shown) for each of the 17 cluster pairs. For each cluster, we fit a least squares line for predicting striatal distance from cortical distance, for the distance from that cluster to each of the other 16. Then we performed two-tailed binomial tests for the smaller of the number of 16 points with striatal < cortical distance and the number with cortical < striatal. Adjusting for the 17 comparisons, clusters 6 and 8 (both originating in pars orbitalis PFC), and cluster 10 (originating from both medial and lateral orbitofrontal PFC) showed significant deviation from chance with striatal < cortical, i.e., in a convergent, non-parallel, pattern (adjusted p-value 0.009). Further, clusters 1 (originating in rostral middle frontal gyrus PFC) and 7 (originating in lateral orbitofrontal PFC) trended towards significance (adjusted p-value = 0.071), again, in a convergent, non-parallel, pattern.

    Conclusions: Using dMRI fiber cluster topography analysis in HSs, we show that the PFC projection pattern between the PFC and the striatum deviates from a topographic, parallel, organization, due to convergence, driven by specific anatomic clusters, in particular, those emanating from orbital and lateral PFC. This approach will allow us to test for variation in the pattern of frontostriatal structural connectivity in other neuropsychiatric conditions such as schizophrenia and autism.

    Keywords: Diffusion Weighted Imaging, Frontostriatal Circuitry, Brain Structural Connectivity

    Disclosure: Nothing to disclose.

    M136. The Effects of Acute Dopamine Depletion on Resting-State Functional Connectivity and Striatal Glutamate in Healthy Humans

    Fernando Caravaggio*, Yusuke Iwata, Alex Barnett, Shinichiro Nakajima, Carol Borlido, Eric Plitman, Yukiko Mihashi, Julia Kim, Philip Gerretsen, Ariel Graff

    Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada

    Background: Understanding the interplay between the neurotransmitters dopamine and glutamate in the striatum has become the highlight of several theories of neuropsychiatric illnesses, such as schizophrenia. However, it is unclear how acute dopamine depletion in healthy humans may affect the functional connectivity of the striatum and glutamate concentrations therein.

    Methods: Sixteen healthy participants (8 female, age [mean ± SD] = 27.4 ± 9.3) participated in the study. Participants provided both resting-state MRI (rsMRI) and 1H-MRS scans (3 T GE Discovery MR750; 8-channel head coil) before and after acute dopamine depletion. Dopamine depletion was induced by oral metyrosine intake over 25 h (64 mg/kg over 24 h).

    Results: Metyrosine administration resulted in increased self-reported fatigue (t(15)=-3.66, p=0.002) and reduced vigor (t(15)=3.95, p=0.001). Compared to baseline, dopamine depletion reduced functional connectivity between the caudate and medial prefrontal cortex, as well as between the ventral tegmental area and hippocampus. However, acute dopamine depletion did not alter striatal glutamate concentrations measured with 1H-MRS.

    Conclusions: Acute dopamine depletion in healthy humans may reduce the functional connectivity between the dorsal striatum and prefrontal cortex. This work has important implications for theorizing about neurochemical alterations in neuropsychiatric diseases and, in turn, our understanding of the functioning of the basal ganglia in humans.

    Keywords: Dopamine, Resting State Functional Connectivity, Resting-state fMRI, 1H-MRS, Glutamate

    Disclosure: Nothing to disclose.

    M137. Increasing Minority Participation and Diversity in ACNP Meetings and Membership: 2018 Status and Recommendations

    Jack Henningfield*, Lawrence Brown, Sandra Comer, Albert Garcia-Romeu, Dorothy Hatsukami, Debra Furr-Holden

    Johns Hopkins University School of Medicine, Bethesda, Maryland, United States

    Background: This poster will provide an update on data-based evaluation of progress, issues, and thinking on efforts to increase underrepresented minority (URM) membership, annual meeting attendance, and participation in committees since the Henningfield et al. poster presented at the 2016 ACNP meeting. The core premises remain the same as discussed in the 2016 poster and are consistent with ACNP organizational goals and values. Namely, that neuropsychopharmacology science, its application in disease prevention and medicine, and drug regulation and policy, should address and benefit the diversity of the population, and that increasing its own diversity will support these goals. The 2016 poster included data from the College on Problems of Drug Dependence (CPDD) which, through, its UnderRepresented Populations (URPop) committee, shares these same core goals with ACNP. It was proposed then that both ACNP and CPDD might be more effective in promoting underrepresented minority representation through more active collaboration, and indeed increased communication and sharing of ideas between ACNP and CPDD since 2016 has been helpful in addressing these issues. Adopting epidemiological approaches to data collection was discussed to better understand minority participation in ACNP and CPDD and guide interventions to increase participation.

    Methods: Most surveys by ACNP and CPDD address the following specific populations of priority to enhance diversity: “African American”, “American Indian or Alaskan Native”, and “Hispanic”. Some surveys have also counted “Asian”, and “other”. These surveys have also included “Caucasian” or “white” and “male” and “female” gender. The most recent ACNP data represent racial and gender self-identification by ACNP members in 2017.

    Neither organization systematically collects information reflecting potential barriers to participation or information that might guide the development of more effective inclusion strategies by the organizations and external funders such as the National Institute on Drug Abuse. A novel approach that was piloted at the 2016 meeting will also be conducted at the 2018 meeting to extend the expert poll to ACNP meeting visitors to the poster.

    Results: The most recent data from ACNP are from its 2017 membership poll, whereas the most recent CPDD data are from a poll of meeting participants following its 2017 annual meeting. Although using somewhat different approaches, the overall data are remarkably similar in population representation. Specifically, African Americans represented 0.7% of ACNP members and 2.5% of CPDD meeting participants who responded to the poll (note it appears that approximately 25% of CPDD attendees responded to the poll – a more accurate estimate should be available for the poster); Hispanic self-identification was 3% of ACNP members and not counted by CPDD but may have been included among the 7.7% who self-identified as “other”; Pacific Islander self-identification was 0.1% of ACNP members and 0.3% of CPDD attendees; Asian self-identification was 8% of ACNP members and not counted by CPDD but may have been included among the 7.7% who self-identified as “other”. Caucasian/white self-identification was 79% of ACNP members and 82% of CPDD attendees who responded to the poll.

    Gender self-identification indicated that approximately 28% of ACNP members are female, and 60% of those who responded to the CPDD meeting poll were female (again note that only about 25% of meeting attendees responded to the CPDD poll).

    Conclusions: Gender self-report data suggest promising trends for women with their ACNP membership increasing from approximately 5% in the early 1990s to approximately 28% in 2017. Although the CPDD estimates for meeting participation are limited by low poll response rates, the self-identification of approximately 60% of meeting participants as female are also encouraging. In contrast, racial self-identification data are far from acceptable in absolute numbers and trends for African American, Hispanic, and Pacific Islander/Native American individuals. As will be discussed in the poster, the filling of all available minority travel awards indicates that these programs are functioning at least to ensure some level of minority participation in meetings that would otherwise likely be lacking without such programs. However, the limited membership data available for CPDD and trends in membership based on ACNP data do not support the conclusions that these programs are substantially contributing to the diversification of membership for either organization.

    Clearly more needs to be done and done more rapidly as neither organization is keeping pace with emerging trends in racial ethnic makeup of the US. Unfortunately, neither organization is currently utilizing modern epidemiologically-based surveillance methods to capture the true diversity of meeting participants and membership beyond crude historical census-based types of categories, nor is either organization collecting the kinds of qualitative data vital to better understand the barriers and provide a data-based foundation for more effective interventions to increase diversity. The poster will include such ideas for going forward, and its poll to be conducted at the meeting is expected to provide additional options for consideration by ACNP and CPDD.

    Keywords: Racial Ethnic Minority, Epidemiology, Ethical Issues

    Disclosure: Pinney Associates, Consultant

    M138. Practical Experiences in Advanced Data Discovery: The Databridge for Neuroscience Project

    Howard Lander, Arcot Rajasekar, Jessica Turner, Lei Wang*

    Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States

    Background: The DataBridge for Neuroscience (DBfN) project is an NSF funded effort to introduce the benefits of advanced data discovery to the Neuroscience community. In this pilot project, we exploited the technical capabilities of the DataBridge to provide discoverability to a collection of clinical data sets housed in the SchizConnect ( virtual archive.

    Methods: We exploited existing work in harmonization of the data sets to create an actionable RDF ontology representing relationships among entities represented in the data sets. Next, we used the ontology to create a DataBridge “signature” for each data set. We then selected the well-known Hammings Distance algorithm for our similarity measure. Finally, we used the DataBridge to perform a pairwise comparison among the selected corpus of SchizConnect data to produce a searchable similarity matrix.

    Results: The project produced several versions of an owl formatted ontology representing a data harmonization for research instruments utilized by studies in the SchizConnect archive. These are available at The project also produced searchable matrices representing estimates of the underlying relationships among the studies.

    Conclusions: The pilot project succeeded in the basic objective of applying the DataBridge to a small set of Neuroscience data. The produced matrices seem “reasonable” to domain experts. Further evaluation will require larger amounts of data.

    Keywords: Data Sharing, Deep Indexing, Data Discovery, Data Similarity, Open Neuroscience

    Disclosure: Nothing to disclose.

    M139. Inaccurate Prescribing Warnings in Electronic Medical Record Systems: Results From an American Society for Clinical Psychopharmacology Membership Survey

    Katharine Phillips*, Leslie Citrome

    Weill Cornell Medical College, Cornell University; New York-Presbyterian Hospital; Alpert Medical School of Brown University, New York, New York, United States

    Background: Electronic prescribing is becoming widespread. All states allow it, some states require it, and many institutions now mandate electronic prescribing. Many electronic prescribing systems use computerized decision support algorithms that give automated warnings or alerts at the time of prescribing if a potential prescribing error is identified -- for example, regarding dosing or contraindications. Some studies suggest that electronic prescribing alerts may reduce prescribing errors and can be clinically useful, but others caution that the warnings may have substantial limitations and that clinicians often consider them clinically irrelevant. Harmful unintended consequences of such prescribing alerts have been described. Despite this topic’s importance, few studies have examined the accuracy of automated prescribing warnings in electronic prescribing systems; to our knowledge no study has focused primarily on the accuracy of systems regarding prescribing of psychotropic medications. To examine this issue, we surveyed members of the American Society for Clinical Psychopharmacology (ASCP) regarding automated warnings generated by electronic prescribing systems.

    Methods: The authors developed a 30-item survey that obtained information about prescribers’ experience with prescribing warnings in electronic medical record prescribing systems. The survey included questions on respondent and practice characteristics, electronic prescribing systems used, perceived errors in automated electronic prescribing warnings, ability to override prescribing warnings, and related questions. The survey was hosted by Survey Monkey, a free online survey tool, and was sent to all current active ASCP members (n=1223). Summary statistics were calculated.

    Results: A total of 118 ASCP members from 33 states completed the survey (9.6% response rate). A majority of respondents (56.8%) were age 55-74; 72.4% were male. An electronic prescribing system was used by 78.0% of respondents; 15.9% used more than one system. Overall, respondents used a total of 31 different electronic prescribing systems. Use of such a system was mandatory in the state, institution, or practice setting of 43.2% of respondents. Regarding the number of electronic prescriptions written in a year, about one third wrote more than 900. Among those who electronically prescribe, 83.1% reported that their electronic prescribing system provides automated warnings if the system detects a potentially problematic prescription. Among these individuals, one third believed that their system has provided incorrect warning information, and one third of this latter group believed that warnings were inaccurate 50% of the time or more. Types of information in prescribing alerts that clinicians considered inaccurate were: dosing range (54.2% of respondents), drug interactions (50%), contraindications (41.7%), dosing frequency (37.5%), dosing time (12.5%), indications (12.5%), and other (8.5%). Examples of erroneous prescribing alerts included: the maximum daily dose of fluoxetine is 40 mg/day, fluoxetine should not be combined with clonazepam, three doses a day of venlafaxine XR 37.5 mg exceeds the recommended maximum dose of one a day, prescribing bupropion and fluoxetine concurrently is contraindicated, aripiprazole is contraindicated for patients ages 6 through 18, and SSRIs are contraindicated for patients under age 18. Among respondents who perceived some warnings to be inaccurate, 95.8% stated that their system allows them to explain their rationale for the prescription or to override the warning, thereby enabling them to prescribe the desired medication despite the warning. However, a majority (76.2%) reported being unable to alert the system that the warning was incorrect. Regarding the burden of managing inaccurate prescribing data alerts, only 4.2% indicated that such warnings were not at all burdensome; 45.8% reported slight burden, 45.8% reported moderate burden, and 4.2% reported severe burden.

    Conclusions: These results indicate that a substantial proportion of prescribing clinicians with an interest in psychopharmacology believe that their electronic prescribing system has provided incorrect prescribing warnings. It is particularly puzzling and problematic that some warnings do not reflect product labeling information – for example, regarding maximum dose or contraindications. Such errors potentially have profound consequences. For example, the erroneous warning that SSRIs and aripiprazole are contraindicated for children and adolescents may cause inadequate treatment of potentially life-threatening conditions in this age group. It is also concerning that most respondents reported being unable to alert the system about the inaccuracy of a prescribing warning. Nonetheless, automated electronic prescribing alerts are potentially useful; a limitation of our survey is that it did not assess the perceived usefulness of alerts or the balance of perceived benefits versus perceived risks. Other limitations include the low survey response rate and the small number of responses for some questions. We also do not know whether all alerts considered erroneous were actually erroneous. Additional studies of this topic are needed, especially given increasing use of electronic prescribing and potentially detrimental clinical consequences of inaccurate prescribing warnings.

    Keywords: Clinical Psychopharmacology, Prescribing, Electronic Prescribing Warnings

    Disclosure: Nothing to disclose.

    M140. Kappa Opioid System Regulates Motivation to Pursue a Cold Environment

    Manish Madasu, Tayler Sheahan, Jordan McCall, Ream Al-Hasani*

    St. Louis College of Pharmacy/ Washington University in St. Louis, St. Louis, Missouri, United States

    Background: The kappa opioid system plays a critical role in the modulation of motivated behaviors, which regulate disease states such as depression and addiction. Importantly, kappa opioid receptors are well known to have analgesic efficacy despite the induction of negative affect. A role for KOR has been well elucidated in decreasing sensitivity to heat, however less is known about their role in cold sensation. Studies have demonstrated that transient receptor potential (TRP) A1 channels facilitate the perception of noxious cold at the level of dorsal root ganglia (DRG), where expression of KOR has also been reported. Here we investigate the role of the KOR in motivation to pursue cold environments and whether the presence/activation of TRP channels modulates such an effect.

    Methods: To measure cold responsivity, we used the cold plate and cold plantar tests. For the cold plate we habituated male WT mice in plexiglass boxes for 1 h prior to treatment with: 1) U50 (5 mg/kg, i.p.), 2) norBNI (KOR antagonist,10 mg/kg, i.p.), 3) saline (10 ml/kg, i.p.). Post-treatment mice were placed on the cold plate for 5 minutes and nocifensive response was calculated. For the cold plantar assay, 30 mins following saline (10 ml/kg) or U50 (10 mg/kg) administration, we acclimated male and female WT mice on a glass plate and applied a cold stimulus to the hind-paw and measured the latency to withdraw from the cooled glass. We assessed the role of KOR in the motivation to purse hot or cold environments using an operant plantar thermal assay. To determine the calcium dynamics, we cultured DRG neurons from WT mice and treated the cultures with a TRPA1 agonist, mustard oil (MO) (100 µM), U50 (10 µM), and a combination of both at 1, 3, 5 mins respectively and quantified the change in the intensity of the Ca2+ indicator.

    Results: Mice injected with KOR agonist U50 showed a significant potentiation in the number of jumps on the cold plate compared to controls. U50-induced nocifensive responses were attenuated in mice injected with norBNI. In the cold plantar assay, the latency to withdraw the hindpaw to cold stimulus appeared to decrease in males, but not in females. Mice showed increased motivation to avoid the cold environment when injected with U50. In the calcium imaging experiments, simultaneous application of MO, and U50 yielded a potentiated Ca2+ response, while U50 alone failed to elicit a calcium response, suggesting crosstalk between receptors.

    Conclusions: Here we show that activation of KOR not only increases cold sensation in male and female mice, but these mice are actively motivated to avoid the cold environment. We further show that KOR mechanism of action is likely in conjunction with TRPA1 as the Ca2+ response is potentiated when both are activated. These findings reveal a distinct role for KOR in motivation to pursue a cold environment suggest a role for KOR in physiological and pathological cold transduction and cold-triggered pain.

    Keywords: Kappa Opioid Receptors, Transient Receptor Potential (TRP) A1, Temperature Regulation, Motivation

    Disclosure: Nothing to disclose.

    M141. Cortical Involvement in RMTg-mediated Aversive Signaling

    Elizabeth Glover*, E. Margaret Starr, Wesley Wayman, John Woodward, L. Judson Chandler

    University of Illinois at Chicago, Chicago, Illinois, United States

    Background: The rostromedial tegmental nucleus (RMTg) is a small GABAergic nucleus that exerts inhibitory control over midbrain dopamine neurons. Exposure to aversive stimuli is associated with enhanced activity within the RMTg and activation of the RMTg facilitates aversive responding. The medial prefrontal cortex (mPFC) provides input to the RMTg and activity within this region has also been implicated in many of the same endpoints. However, little is known about the anatomy and function of this projection.

    Methods: To better characterize the density of mPFC input to the RMTg anatomically, the distribution and number of tracer labeled neurons was assessed in adult male Long-Evans rats injected with the retrograde tracer cholera toxin B (CtB) into the RMTg (n=6). To determine the functional role of the mPFC-RMTg projection, we measured the behavioral response to stimulation of mPFC terminals in the RMTg using in vivo optogenetics in combination with real-time place preference (n=5-7/group). In addition, we quantified cFos induction in RMTg-projecting mPFC neurons following exposure to neutral and aversive stimuli (n=6-9/group). Finally, we evaluated stimulus-induced plasticity in RMTg-projecting mPFC neurons by measuring evoked firing following exposure to aversive stimuli in both male and female Long-Evans rats using whole-cell patch-clamp slice electrophysiology (n=4-6/group). Data were analyzed using t-tests and ANOVAs where appropriate.

    Results: Rats sacrificed two weeks following CtB injection into the RMTg exhibited dense cell body labeling throughout the entire medial and orbital walls of the PFC as well as the entire rostrocaudal extent of the region. Labeling was restricted primarily to layer V of the mPFC, though a small number of cells were also consistently observed in the deepest portion of layer VI. Quantification of CtB labeled neurons relative to NeuN labeling found that approximately 14.21% ± 0.38 of layer V prelimbic neurons projected to the RMTg – a proportion that is substantially greater than many other subcortical projections from the prelimbic cortex including the amygdala, ventral tegmental area, and periaqueductal gray. Stimulation of RMTg-projecting mPFC inputs produced significant avoidance behavior, the magnitude of which was similar to that produced by stimulation of lateral habenula inputs to the RMTg (p ≤ 0.01) indicating that activation of this pathway can induce aversive responding. Rats presented with a shock or a tone predictive of shock exhibited a significant increase in cFos expression in RMTg-projecting mPFC neurons compared to rats exposed to neutral stimuli (p ≤ 0.05). In addition, exposure to a single episode of 10 consecutive foot shocks resulted in a significant decrease in the frequency of evoked firing in RMTg-projecting prelimbic neurons compared to unshocked controls (p ≤ 0.0001).

    Conclusions: Together, these results uncover a remarkably dense projection from the mPFC to the RMTg that drives behavioral responding to aversive stimuli. We further reveal significant plasticity as a result of exposure to aversive stimuli. Alterations within this neural circuit may be critically involved in neuropsychiatric illnesses associated with disruption of the balance between signaling of rewarding and aversive outcomes including addiction and mood disorders.

    Keywords: Aversion, RMTg, mPFC

    Disclosure: Nothing to disclose.

    M142. fMRI-Measured Brain Network Connectivity Changes After Acute Cannabis Challenge

    Michael Stevens*, Nicolas Ward, Erwin Boer, Diana King, Catherine Boyle, Ashley Bullock, Vince Calhoun, Godfrey Pearlson

    Yale University School of Medicine, Hartford, Connecticut, United States

    Background: The timecourse of acute cannabis effects on brain function and how these changes relate to behavior is important for understanding the ways in which marijuana use might confer public health risk or therapeutic benefit. One of the least understood phenomena is how cannabis alters the way in which widely-distributed brain regions co-engage to form functional neural networks underlying task performance. Here, we used functional connectivity analyses to describe how complex network characteristics were changed by smoking cannabis and how those changes evolved over the course of several hours.

    Methods: As part of an ongoing NIDA-funded study (R01DA038807), fMRI data were collected from 15 regular marijuana users after administering NIDA herbal 13.4% THC, 5.9% THC, or placebo cannabis via a paced inhalation protocol using a vaporizer. On each visit, participants were given the drug/placebo by 9:00 a.m., then underwent fMRI 3 separate times post-dose in a randomized, double-blind design. All participants were scanned at 0.5 h following dosing, but the other two fMRI sessions occurred either at 2.5 h, 4 h, or 5.5 h after drug use in a counter-balanced design that provided a greater number of timepoints at which to observe whether the effects of the drug had resolved. Two fMRI task probes were used that prior research has validated are sensitive to both neural and behavioral effects of smoked cannabis -- A time estimation task and a cognitive set shifting paradigm. fMRI BOLD timeseries were prepared for analysis using Human Connectome Project protocols, then examined using graph theory-based connectivity metrics to characterize connectivity abnormalities after drug dosing relative to placebo. A repeated-measures ANOVA identified brain regions with linear or curvilinear changes in the degree of connectivity strength over time for all 391 cortical and subcortical HCP parcels. For any HCP parcel-localized effect p<.05 or less, we inspected mean plots to understand the profile of regional connectivity abnormalities relative to placebo over time.

    Results: Linear effects that met the p<.05 significance threshold had eta-squared effect sizes ranging from 0.32 to 0.60 (mean 0.42) for the time estimation task, and from 0.35 to 0.53 (mean 0.49) for the set shifting task (all “large” effects). Network connectivity abnormalities induced by cannabis showed evidence of task-dependency. Most network connectivity changes observed immediately after cannabis inhalation during set shifting task performance took the form of excessive cortical connectivity, whereas there was widespread evidence for hypo-connectivity soon after dosing on the time estimation task. No brain regions were commonly impaired across both tasks. The most notable finding was that brain connectivity remained abnormal compared to placebo 5.5 h after cannabis use. While the majority of initial abnormalities had waned by 2.5 h, evidence emerged that regional connectivity disturbances seen immediately after using cannabis then “overshot” their baseline levels after several hours. That is, by 5.5 h many regions showing cortical hyper-connectivity became hypo-connected, or vice versa. There also were brain regions that failed to show any initial difference from placebo in their local connectivity strength, but that over time showed more connectivity changes.

    Conclusions: These preliminary study findings suggest that cannabis does not simply have a localized effect in regions with high density of CB receptors, but rather it can induce persistent, widespread effects on functionally-integrated neural networks. This effect appears not only to be context-specific, but also to be biphasic over time, with initial abnormalities changing dramatically after a few hours to a longer, persistent state with quantitatively and qualitatively different connectivity abnormalities. These findings prompt different possible interpretations. One possibility is that the immediate effects of cannabis on network connectivity might reflect the limited duration of acute intoxication. Another is that the network abnormalities that emerge later reflect a form of neural compensation for the initial brain connectivity impairment. There were more examples of late-occurring “overshoot” abnormalities than quickly-waning network connectivity disturbances, raising questions about the nature and effect of these longer-lasting brain network effects of cannabis use. These different classes of drug-induced effects on distributed brain function suggest future functional neuroimaging research should consider lengthier periods of post-drug recovery when attempting to understand how cannabis-induced brain changes might influence cognition and behavior.

    Keywords: Cannabis Use, Functional Connectivity, fMRI, Drug Response

    Disclosure: Nothing to disclose.

    M143. Critical, Sex-Dependent Impacts of a Vasopressin Model of Preeclampsia on Neurodevelopment and Behavior in Mouse Offspring

    Serena Gumusoglu*, Akanksha Chilukuri, Salma Haider, Sabrina Scroggins, Jeremy Sandgren, Dark Santillan, Donna Santillan, Justin Grobe, Hanna Stevens

    University of Iowa Carver College of Medicine, Iowa City, Iowa, United States

    Background: Preeclampsia is an often severe, gestational hypertensive condition linked to abnormal child neuropsychiatric outcomes (e.g. autism spectrum disorder, learning delays). This connection between preeclampsia and offspring neurodevelopment is poorly understood. One unique animal model of preeclampsia involves continuous maternal administration of arginine vasopressin (AVP) throughout pregnancy. This exposure results in pregnancy-specific hypertension, proteinuria, and an altered inflammatory profile in dams and offspring. Here, we examine critical impacts of this model of preeclampsia exposure on neurodevelopment and behavior in mouse offspring.

    Methods: C57Bl/6 J mouse dams were implanted with a subcutaneous osmotic minipump (Alzet), filled with AVP or saline (24 ng/hr), 3 days prior to mating. Juvenile and adult offspring were tested for behavioral phenotypes using the Y-Maze, rotarod, social approach, and elevated plus maze assays (n=12-14 per group per sex). Offspring brains were collected embryonically and postnatally to examine cortical morphogenesis histologically (n=4-6 per group per sex per age). Additionally, RNA from embryonic cortical tissues (n=4 per group per sex) was extracted and sequenced (Illumina HiSeq 4000). Reads were pooled across two lanes (approximate depth: 30,000), aligned and summed using Kallisto, and then statistically compared and analyzed using Sleuth. Hits were functionally annotated and analyzed for enrichment of functional clusters/pathways using PANTHER and DAVID. Differences that survived correction for multiple comparison were validated with qPCR.

    Results: Preeclampsia-exposed adult males exhibited increased anxiety-like behavior and social approach while adult females exhibited impaired procedural learning and working memory. Despite these sex differences, preeclampsia-exposed male and female offspring showed similar volume reductions in dorsal forebrain and cortical plate late in embryonic and early in postnatal development only. Interestingly, preeclampsia-exposed offspring also exhibited compensatory increases in embryonic cell packing density, maintaining the same total number of dorsal forebrain cells and preserved cortical layers. RNAseq analyses of embryonic anterior cortex revealed a single gene differentially expressed by preeclampsia exposure across males and females—ebf2—which was upregulated and plays a critical role in cortical neurogenesis. Sex-specifically, 49 genes were differentially-expressed by preeclampsia exposure in male offspring (select GO pathways: translation, transcriptional regulation) and 31 in females (GO pathways: development, signal transduction). Of the top five altered genes in male embryonic cortex—gdf1, capn11, slc26a4, fgf15, and tfap2d—three have been implicated in cortical patterning, neuronal differentiation, and neurogenesis. Of the top five altered genes in female embryonic cortex—cped1, gprc5c, postn, lox, and mrc1—four are implicated in development and response to injury.

    Conclusions: Within an AVP-based model of preeclampsia, we found that exposed offspring exhibited sex-specific deficits in behavior. Corticogenesis was disrupted across both males and females, with altered development of neuronal precursors and upregulation of the cortical neurogenesis regulating gene, ebf2. Sex-specific alterations in embryonic cortex with preeclampsia exposure most significantly disrupted genes implicated in processes of protein translation in male embryonic brain while female offspring showed the most substantial alterations in signal transduction genes. The origins of sex differences in corticogenesis and behavior were not apparent, but they suggest distinct pathways by which preeclampsia may disrupt brain development.

    Keywords: Neurodevelopment, Pregnancy, Hypertension, RNA Sequencing, Animal Models

    Disclosure: Nothing to disclose.

    M144. Developmental Exposure to Pyrethroid Pesticide Causes an Autism-Related Phenotype and Dopamine System Alterations in Mice

    James Burkett*, Rachel Branco, Rohan Dhamsania, Emily Winokur, Shannon Gourley, Gary Miller

    Emory University Rollins School of Public Health, Atlanta, Georgia, United States

    Background: Pyrethroids are a class of pesticide whose use has broadly expanded in recent years, as it is considered the “safe” alternative to harmful organophosphates and organochlorides. Recent epidemiological studies, however, have suggested that residential exposure to pyrethroid pesticides during pregnancy increases the risk of Autism Spectrum Disorders (ASD) in the unborn child. Our lab’s previous work has shown that developmental exposure to the pyrethroid deltamethrin, in doses considered safe by the EPA, causes an ADHD-related phenotype and disruptions in the dopamine system in mice.

    Methods: In this study, we exposed multiple independent cohorts of female mice (total N=60) to a low dose of deltamethrin (3 mg/kg PO Q3D) during pre-conception, pregnancy and weaning. We tested the offspring on a broad behavioral battery representing all three diagnostic domains of ASD, as well as related symptoms and comorbid disorders.

    Results: Consistent with our prior publications, we found evidence of hyperactivity and repetitive behaviors. In addition, we found decreased ultrasonic vocalizations, increased rigidity, and learning deficits. No deficits in social behavior were found. We also found increased striatal dopamine, increased dopamine transporter, and other dopamine-related disruptions.

    Conclusions: Additional studies are urgently needed to determine whether exposure to this common class of pesticides contributes to ASD diagnosis in humans.

    Keywords: Neurodevelopmental Disorders, Toxins, Autism Spectrum Disorders, Pesticides

    Disclosure: Nothing to disclose.

    M145. Larger Numbers of Glial and Neuronal Cells in the Periaqueductal Gray Matter of μ-Opioid Receptor Knockout Mice

    Abstract not included.

    M146. Increased Calcium Influx in Response to Activity in Human Induced Pluripotent Stem Cell-Derived Neural Progenitors Lacking Fragile X Mental Retardation Protein

    Ulla-Kaisa Peteri, Venkat Swaroop Achuta, Maija Castrén*

    University of Helsinki, Helsinki, Finland

    Background: The absence of FMR1 protein (FMRP) results in fragile X syndrome (FXS) and disturbed FMRP function is implicated in several forms of human psychopathology. FMRP is essential for normal neurogenesis and functional maturation of synapses and neuronal networks. Exaggerated group 1 mGluR function is found to be a fundamental abnormality in FXS. Responsiveness of FXS neurons is increased and correlates with hyperarousal, sensory hypersensitivity, and susceptibility to epilepsy in the FXS phenotype. We have functionally characterized differentiating mouse and human FXS neural progenitors at the early stage of differentiation [1]. The studies have revealed enhanced differentiation of glutamate-responsive cells and augmented intracellular calcium responses to activation of mGluR in FXS progenitors. Increased responses to activity promoted us to explore the role of L-type voltage-gated calcium (Cav) channels in altered differentiation of FXS progenitors.

    Methods: Human neural progenitors were differentiated in neurospheres from four FXS iPS cell lines and three control cell lines that are previously characterized [1,2]. Mouse neural progenitors were generated from the brain of wild type and Fmr1 knockout mice. We used fura2-AM based calcium imaging to study intracellular calcium responses and RT-PCR for expression studies and to confirm results of the Affymetrix array analysis.

    Results: We observed augmented responses to depolarization with high extracellular calcium in both human and mouse neural progenitors. Nifedipine treatment prevented the increase in calcium influx indicating that L-type Cav channels contributed to the exaggerated responses in FXS progenitors. Ratio of the pore-forming alfa1 subunits of the L-type channel/T-type channel expression was increased and correlated with enhanced differentiation of FXS progenitors, but no significant changes were detectable in the expression of single subunits. Transcriptome analysis of human iPS cell-derived FXS progenitors revealed alternative splicing of the CAST gene that encodes calpastatin, suggesting contribution of cytoplasmic factors to augmented intracellular calcium responses.

    Conclusions: Our results show a critical role for increased calcium influx through L-type Cav channels in FXS progenitors and address impact of altered calcium homeostasis during neuronal development in FXS brain.

    Keywords: Voltage-Gated Calcium Channel, Fragile X Syndrome, Neural Progenitor Cells, Calcium, In Vitro Neuronal Differentiation

    Disclosure: Nothing to disclose.

    M147. Genetic Variants of OCD Phenotypes and Comorbid Conditions

    Vera Nezgovorova*, Nichole Rigby, John Battles, David Krause, Naomi Fineberg, Michael Van Ameringen, Eric Hollander

    Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, New York, United States

    Background: Obsessive-compulsive disorder (OCD) is characterized by recurrent and intrusive thoughts, impulses or images (obsessions) that cause discomfort and are accompanied by repetitive behaviors (compulsions) to alleviate discomfort. OCD is often associated with significant psychiatric comorbidity which is frequently underdiagnosed and undertreated. Comorbid disorders include mood and anxiety disorders as well as obsessive-compulsive spectrum disorders (OCSDs), including anorexia nervosa, bulimia nervosa, binge-eating disorder, BDD, compulsive shopping, kleptomania, pathological gambling, intermittent explosive disorder, hypersexual disorder, tic disorder and autism spectrum disorders. Comorbidity in OCD may influence treatment decisions and may require specific pharmacotherapy augmentation. The Genecept Assay® (Genomind, King of Prussia, Pennsylvania) provides information to clinicians that can be used for personalized treatment decisions on psychotropic medications based on individual patient’s genetic, pharmacokinetic and pharmacodynamic profile. It also might help to develop management strategies for different OCD comorbidities, influencing treatment selection.

    Methods: We evaluated 130,629 patients (61.09% female) who underwent the Genecept Assay phenotyping protocol. 1,882 patients (1.44%) had OCD as the primary diagnosis and 2,763 patients (2.12%) had OCD as comorbid diagnosis. All patient data was deidentified for analysis purposes. DNA sample collection occurred as part of the patient’s standard of treatment. Consent to perform the Genecept Assay was obtained by study physician. All analyses were performed using R software, v.3.3.3 (R Foundation for Statistical Computing, Vienna, Austria). We performed a descriptive analysis of the studied cohort data using nonparametric tests and logistic regression analysis while controlling for comorbid depressive disorders, bipolar disorders, ADD/ADHD, anxiety and panic disorders phenotypes.

    Results: In the studied cohort, 21.79% of patients with primary OCD diagnosis were diagnosed with comorbid depressive disorder, 19.29% with anxiety, 14.24% with ADD/ADHD, 3.77% with panic disorders, 3.13% with bipolar disorder and 1.75% with PTSD. Over 80% of patients with OCD were reported as Caucasian, and over 20% were between 15-19 years old. Patients with primary OCD diagnosis had significantly higher frequencies of comorbidities with depressive disorders and ADD/ADHD, as compared to patients without an OCD-related diagnosis (p<0.05). Comorbid OCD and bipolar disorders phenotypes were associated with COMT, OPRM1 and GRIK1 genotypes. Comorbid OCD and depressive disorders phenotypes were associated with OPRM1 and CYP3A4/5 genotypes (p<0.05). Comorbid OCD and ADD/ADHD were associated with MC4R and 5HT2C genotypes (p<0.05). Comorbid OCD and panic disorders were associated with CACNA1C genotypes (p<0.05). Comorbid OCD and anxiety were associated with CYP3A4/5 variants in OCD patients (p<0.05).

    Conclusions: Our analysis constitutes a first attempt to use the Genecept assay to elucidate genetic profiles in OCD patients and relate these to comorbidities and treatment response. Although more research is needed, a personalized approach for management of treatment-resistant OCD cases associated with mental health comorbidities based on common gene variants could potentially improve patients’ quality of life and reduce treatment costs. Future studies of sequential next steps in treatment-resistant OCD management including use of gene variants as predictors of response to personalized treatments are needed.

    Keywords: Obsessive-Compulsive Disorder (OCD) Phenotypes, Genecept Assay, Pharmacogenomics, Obsessive-Compulsive Spectrum Disorders (OCDS), Personalized Medicine

    Disclosure: Nothing to disclose.

    M148. Pharmacokinetic and Pharmacodynamic Properties of the Investigational AMPA Receptor Positive Allosteric Modulator TAK-653 After Single and Multiple Rising Doses in Healthy Volunteers

    Mahnaz Asgharnejad*, Stephen Zicha, Lin Xu, Don Cheng, Sergio Iadevaia, Andy Dorner, Venketesha Murthy, Emiliangelo Ratti

    Takeda Development Center Americas, Inc., Cambridge, Massachusetts, United States

    Background: Antagonism of the N-methyl-D-aspartate (NMDA) receptor by ketamine at subanesthetic doses to patients with treatment-resistant depression (TRD) is associated with rapid onset of antidepressant effects. This antidepressant action appears to be mediated via increased levels of brain-derived neurotrophic factor (BDNF) and increased mammalian target of rapamycin (mTOR) signaling. This effect can be replicated by the activation of post-synaptic alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) receptors downstream of NMDA inhibition. Activation of AMPA receptors triggers increases in BDNF release; raising BDNF levels has been described as a potential biomarker of antidepressant activity. Therefore, selective activation of AMPA receptors can be considered a novel mechanism for antidepressant effect. We have discovered TAK-653, a selective and potent AMPA positive allosteric modulator that was efficacious in nonclinical models of depression. In this study, we describe the results from the first-in-human study with TAK-653 where we tested the safety and tolerability in a single rising dose (SRD) and multiple rising dose (MRD) settings in healthy volunteers. In addition, for a preliminary understanding of the antidepressant potential of TAK-653, plasma and serum BDNF levels were measured to determine if TAK-653 has any effect on this depression-related biomarker.

    Methods: This study (NCT02561156) was a randomized, double-blind, placebo-controlled, combined SRD and MRD in healthy volunteers. Subjects with a history of or a risk factor for seizures were excluded. The study enrolled 8 subjects (6 active and 2 placebo) per cohort in 6 SRD and 5 MRD cohorts. Each cohort was orally administered TAK-653 in tablet form sequentially to ensure adequate safety and tolerability before administering the next dose level. In SRD cohorts, a single dose of TAK-653 (0.3, 1, 3, 5, 9, or 18 mg) or matching placebo was administered on Day 1 (fasting conditions) followed by safety, tolerability, pharmacokinetic (PK), and pharmacodynamic (PD) assessments. The starting doses were selected based on predicated human PK and no-observed-adverse-effect level in preclinical species. In MRD cohorts, a single dose of TAK-653 (0.3, 1, 3, 6, or 9 mg) or matching placebo was administered on Day 1 followed by 5-day safety, tolerability, PK, and PD assessments. TAK-653 was then administered once daily (QD) for 13 days. In the 3 mg MRD cohort, a single cerebrospinal fluid (CSF) sample was collected from each subject on Day 12 to estimate average central concentrations of TAK-653 at approximately steady state. Safety assessments were treatment-emergent adverse event (TEAE) monitoring; vital signs; safety laboratory tests, electrocardiograms, and electroencephalograms; and the Columbia-Suicide Severity Rating Scale.

    Results: TAK-653 was safe and well tolerated in all cohorts. All TEAEs were considered mild or moderate. No dose relationship in the incidence of any TEAEs was observed in either SRD or MRD cohorts.

    TAK-653 was absorbed rapidly in plasma after SRD or MRD administrations of TAK-653 tablet formulation under fasted conditions. TAK-653 plasma concentrations generally reached maximal levels within median 1.25 to 5 h after dosing. Mean Cmax ranged from 3.63 to 126 ng/mL following single doses ranging from 0.3 to 18 mg, and from 7.86 to 243 ng/mL following repeated QD doses ranging from 0.3 to 9 mg. Mean overall drug exposure during the dosing interval (AUC24) ranged from 151 to 4598 h*ng/mL for QD doses of 0.3 to 9 mg. Observed interindividual variability was low to moderate (coefficient of variation <45%) for both Cmax and AUC. TAK-653 accumulation after QD dosing was 3- to 4-fold, which is consistent with its 33.1 to 47.8 h half-life. The mean CSF/plasma ratio of TAK-653 was 0.056, which is similar to the unbound plasma concentration.

    Plasma BDNF values were extremely variable and prone to factors such as hemolysis and differences in time sitting on ice. Since it is well known that platelets are a source of BDNF in the blood, only serum values were analyzed. In the SRD part of the study, only the 3 mg dose caused a significant 40% increase in serum baseline-normalized BDNF that was maintained for up to 96 h postdose. All the other doses caused a transient ~25% decrease in BDNF levels that recovered to baseline levels after 12 h. The increase in BDNF with the 3 mg dose of TAK-653 was not replicated in the MRD part of the study. None of the other dose levels had any significant elevation compared to baseline levels of BDNF.

    Conclusions: TAK-653 was well tolerated in healthy subjects at all doses tested, both in the SRD and MRD cohorts. Exposure to TAK-653 was approximately dose proportional. Because of its long half-life, 3- to 4-fold accumulation of TAK-653 was observed after QD administration. The comparable TAK-653 concentration between CSF and unbound plasma suggests its rapid brain penetration. Although BDNF is hypothesized to be a potential biomarker of antidepressant activity, measuring serum levels in healthy volunteers, who had normal levels of BDNF, did not detect an increase in BDNF (PD activity of TAK-653). This was likely due to a ceiling effect as well as the fact that BDNF levels are not reduced, as they are in TRD patients.

    Keywords: TAK-653, treatment-resistant depression, AMPA receptor, positive allosteric modulator, antidepressant effect

    Disclosure: Nothing to disclose.

    M149. Neurophysiological Correlates of Sensorimotor Cortex in Humans Affected by Fragile X Syndrome

    Ernest Pedapati*, Craig Erickson, Lauren Ethridge, John Sweeney, Donald Gilbert

    Cincinnati Children’s Hospital, Cincinnati, Ohio, United States

    Background: Hyperarousal and other abnormalities in cortical excitability have been well documented in animal and slice models of Fragile X syndrome (FXS). However, the degree that these abnormalities are present, including any clinical correlates, in humans with FXS in the sensorimotor cortex remains an open question in the literature.

    Methods: Edit per original abstract feedback for review: We have now collected 22 affected Fragile X subjects and are currently recruiting healthy controls. We anticipate by the meeting in December we will have approximately 15 -20 control subjects completed for a robust comparison. This includes TMS data as well as resting EEG data.

    Original abstract submission:

    We present early analysis of 14 males affected by full mutation Fragile X Syndrome with historical match controls. The primary technique is paired pulse transcranial magnetic stimulation (TMS) presented with data resting state 128-lead EEG data and behavioral when available.

    Results: We have completed analysis on 14 subjects with a short interval cortical inhibition (SICI) of 0.75 +/- SD 0.59 and intracortical facilitation (ICF) of 1.26 +/- 0.49. Of particular interest, we have identified that ICF is highly correlative with EEG resting alpha power (-0.783, p = 0.013).

    This pilot data represents a small sample without adequately matched controls. TMS measures indicate variability within the FXS population, with early indication of correlation with resting state theta and gamma power.

    Conclusions: We apologize for the rather preliminary nature of this abstract but anticipate a more complete set of data with recruited matched controls by December 2019.This study provides feasibility and initial evidence to encourage further investigation into the sensorimotor cortex to complement ongoing investigations of the sensory cortex in humans and sensorimotor physiology in animal models. The quantification and correlation of these deficits to identify individual sources of variability can aid in the development of physiological biomarkers, which may differentiate endotypes of disease or speed drug discovery.

    Keywords: TMS, Fragile X Syndrome, Quantitative Electroencephalography (qEEG)

    Disclosure: Nothing to disclose.

    M150. Inflammatory Biomarkers and Anxiety Traits in the Development of Persistent Post-Traumatic Headaches Following Mild Traumatic Brain Injury in a Patient Cohort: A Pilot Study

    Smriti Iyengar*, Jie Xie, Youngsok Kim, Jared Smith, Christopher Carey, Hyun Kang, Jonathan Saxe, Kelly Naugle, Fletcher White

    Indiana University School of Medicine, Carmel, Indiana, United States

    Background: Post traumatic headache (PTH) is the most frequent symptom following head injury/traumatic brain injury (TBI), manifesting both as a new-onset headache as well as a worsening of preexisting headache. The International Classification of Headache Disorders (third edition, ICHD-3 beta) defines PTH as a secondary headache with no defining clinical characteristics, that starts within 7 days of injury. If headache persists for greater than 3 months’ duration it is considered to be persistent. PTH is frequent and common in both military and civilian populations. PTH appears to be more likely to develop following mild TBI (concussion) compared with moderate or severe TBI (with a prevalence of 47–95%). Although PTH is highly prevalent, the underlying mechanisms leading to PTH and its maintenance are less clear. Sensitization of the trigeminal pain network may be one mechanism underlying the chronification of PTH following head injury. Inflammatory mediation of peripheral and central sensitization has also been implicated. Persistent activation of peripheral immune cells which give rise to a network of immune-to-brain signaling may influence behavior following trauma. Thus, the purpose of this prospective pilot study was to evaluate inflammatory biomarkers at 1-week, 1 month and 4 months post injury, with a view to predict the persistence of post-traumatic headaches at 4 months post injury in mTBI patients.

    Methods: Adult mTBI patients recruited from a Level I Emergency Department Trauma Center completed study sessions at 1-week (visit 1), 1 month (visit 2) and 4-months post mTBI (visit 3). Peripheral blood mononuclear cells (PBMCs) derived from individual PTH patients and sex and age matched controls were evaluated for primed innate immune responses to lipopolysaccharide (LPS) stimulation of PBMCs in vitro. Resultant supernatant samples were assayed using ELISA for the presence of IL-10, IL-1□, TNF□ and the chemokine CCL2. Participants also completed (1) a 0-10 numeric pain rating scale (NRS) to measure average intensity of headache pain; (2) the Center for Epidemiologic Studies (CES-D) depression scale with cutoff scores of 16 and greater for identifying risk for depression; (3) the State-Trait Anxiety Inventory (STAI) for anxiety symptoms. Patient-derived cytokine protein levels were correlated with pain severity and anxiety and depression states, at visits 1, 2 and 3. Patients with self-report of previous concussion events and use of long-term pain medications were excluded. Over the counter NSAIDS or Tylenol was allowed for episodic use but not on days of assessments.

    Results: A total of 11 of 16 patients completed all tests and visits. Overall, in 8 of the 11 patients, headache persisted at Visit 3 (4 months after the mTBI incident) while in 3 patients there was no persistence of headache at Visit 3. Across individual patients, trends suggested that: (1) Patients showed high elevations in CCL2 levels at visit 1 that persisted in several patients at visit 3 and these patients had ongoing headaches at visit 3. (2) IL-1b levels showed increases at visit 1 and persisted in several patients at Visit 3. (3) TNFa levels showed no significant increases in any patient. (4) Increases in IL-10 and headache persistence were variable. One patient with elevated IL-10 levels on visit 3 did not report headache, while headache persisted at Visit 3 in one patient with IL-10 increases at visit 3 and in one patient with IL-10 increase at visit 1. (5) Anxiety traits stayed steady or showed only slight decreases in patients who had persistent headaches, but showed decreased traits in two of three patients that showed no headache at Visit 3. (6) Depressive symptoms were variable with regards to persistence of headache at visit 3 e.g., two patients with no depressive symptoms still had persistent headache at visit 3, three patients with decreasing depression had persistent headaches at Visit 3, three patients with steady or increased depressive symptoms had persistent headache, while three patients with no headache had decreasing depressive symptoms at Visit 3.

    Conclusions: This pilot study suggests that persistent headache occurred in 8 of 11 mTBI patients. Overall, increases in the chemokine CCL2 and persistence of anxiety traits appeared to be associated with persistence of headache in mTBI patients. These observations need further confirmation in a clinical study with larger cohorts of mTBI patients.

    Keywords: Chronic Pain, Inflammation, Biomarker

    Disclosure: Nothing to disclose.

    M151. Effects of Selective Melatonin MT2 Receptor Ligands in the Treatment of Neuropathic Pain: Modulation of Descending Brainstem Antinociceptive Pathways and Opioid Interactions

    Abstract not included.

    M152. Biased Signaling Agonists of D3R Modulate D1R-D3R Crosstalk in Levodopa Induced Dyskinesia

    Abstract not included.

    M153. Blast Exposure in Veterans and an Animal Blast Model Reveal Inter-Related Acute and Chronic Blood-Brain-Barrier Disturbances Associated With Chronic Neuropathology

    Elaine Peskind*, James Meabon, Aric Logsdon, William Banks, C. Dirk Keene, Desiree Marshall, Donna Cross, Todd Richards, Daniel Perl, Cynthia Mayer, Abigail Schindler, Murray Raskind, David Cook

    VA Puget Sound Health Care System, Seattle, Washington, United States

    Background: We have reported structural and functional brain neuroimaging abnormalities in Iraq and Afghanistan war Veterans > 4 years after their last blast mild traumatic brain injuries (mTBI) [1,2]. In an mTBI mouse model of repetitive blast mTBI we have demonstrated region-specific blood-brain-barrier (BBB) disruption in brain regions vulnerable to chronic blast mTBI neuropathology [3–5]. Collectively, these findings raise the possibility that early-occurring BBB disruptions (which often resolve in <24 h), may trigger latent pathologic processes that give rise to chronic neuropathology. To address this possibility, we have found evidence of: (i) blast-induced peripheral immune cell infiltration into the mouse brain, (ii) distinct microglial/macrophage neuropathology in both mice and postmortem brains from Veterans with blast-related mTBI in cerebrospinal fluid (CSF) from Veterans with blast mTBI (iii) chronically elevated immune cell chemo-attractant proteins, as well as (iv) liver-derived circulating peripheral proteins.

    Methods: Blood and CSF were collected from 55 Iraq/Afghanistan Veterans with blast mTBIs (mean 21, range 1-102) and 22 deployed Veterans with no lifetime history of TBI. CSF and plasma proteins were measured using a validated MesoScale platform assays. Postmortem brain tissue was accessioned from 3 mTBI Veterans and 3 controls. Male 3-4-month-old C57Bl/6 mice received blast exposure using a shock tube that accurately models battlefield-relevant blast exposure in mice [3–5]. To measure BBB disruption in mice, brain concentrations of peripherally injected [99Tc]-albumin, which does not normally cross the BBB, were measured in blast and sham control animals. Well-established flow cytometry methods were used to measure blast-induced brain infiltration of CD3+/CD4+ T-cells and CD45hi/Ly6C/6+ monocytes in mice.

    Results: In mice, within 24 h after blast, CNS [99Tc]-albumbin levels in the pons/medulla had returned to normal (sham control levels), but the BBB again reopened with significantly abnormal entry of radiolabeled albumin into the pons/medulla (p<0.05). Importantly, 3 days post-TBI is a time frame previously established as a key temporal window for peripheral immune cell infiltration that subsequently abates over days. In keeping with this we found that repetitive blast exposure also caused a significant elevation in T-cell and peripheral monocyte infiltration into the pons/medulla (p<0.006, p<0.018). While it is not possible to perform such experiments in our living mTBI Veteran cohort, we found that Monocyte Chemo-attractant Protein 1 (MCP-1) and Macrophage-Derived Chemokine (MDC) were elevated in CSF from mTBI Veterans compared to deployed controls (p≤0.05 and p≤0.058, respectively, one-tailed). MCP-1 and MDC are well-characterized chemokines that mediate trafficking of monocytes, T-cells, and other peripheral immune cells into the brain. In addition, serum amyloid A (SAA), a protein produced selectively by the liver and released into the circulatory system, was elevated in CSF from mTBI Veterans versus deployed controls (p<0.058). In mice we found distinctive, clustered IBA1+ microglia/macrophages in the brainstems of blast exposed mice (p<0.001) compared to controls. Prompted by this finding, we also examined pons and medulla of postmortem brains from Veterans with blast-related mTBI and found very similar appearing IBA1+ and CD68+ clustered microglia/macrophages.

    Conclusions: These data indicate that disturbed BBB function contributes to both short and long term neuroimmune-related neuropathologic and structural abnormalities associated with blast exposure and implicate the importance of interactions between the CNS and peripheral immune systems in the pathologic cascades triggered by blast exposure.

    All experimental protocols in animal studies were approved by the VA IACUC and were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals.

    These data indicate that disturbed BBB function contributes to both short and long term neuroimmune-related neuropathologic and structural abnormalities associated with blast exposure.

    All experimental protocols in animal studies were approved by VA IACUC and were conducted in accordance with NIH Guide for Care and Use of Laboratory Animals.

    Keywords: TBI, Blast, CSF Biomarkers, Animal Models, Neuropathology

    Disclosure: Nothing to disclose


    1. Peskind ER, Brody D, Cernak I, McKee A, Ruff RL, Military- and sports-related mild traumatic brain injury: clinical presentation, management, and long-term consequences. J. Clin. Psychiatry, 2013. 74: 180-8; quiz 8.

    2. Petrie EC, Cross DJ, Yarnykh VL, Richards T, Martin NM, Pagulayan K, Hoff D, Hart K, Mayer C, Tarabochia M, Raskind M, Minoshima S, Peskind E, Neuroimaging, Behavioral, and Psychological Sequelae of Repetitive Combined Blast/Impact Mild Traumatic Brain Injury in Iraq and Afghanistan War Veterans. J. Neurotrauma, 2014. 31: 425-36.

    3. Meabon JS, Huber BR, Cross DJ, Richards TL, Minoshima S, Pagulayan KF, Li G, Meeker KD, Kraemer BC, Petrie EC, Raskind MA, Peskind ER, Cook DG, Repetitive blast exposure in mice and combat veterans causes persistent cerebellar dysfunction. Science translational medicine, 2016. 8: 321ra6.

    4. Logsdon AF, Meabon JS, Cline MM, Bullock KM, Raskind MA, Peskind ER, Banks WA, Cook DG, Blast exposure elicits blood-brain barrier disruption and repair mediated by tight junction integrity and nitric oxide dependent processes. Scientific reports, 2018. 8: 11344.

    M154. Drug Metabolizing Enzyme Associations With Response to Agents Used in Neuropsychopharmacology

    Abstract not included.

    M155. RDoC Working Memory Constructs Spanning Levels From Disability to Structural MRI

    Robert Bilder*, Agatha Lenartowicz, Jesse Rissman, Sandra Loo, Jean Baptiste Pochon, Holly Truong, Kristen Enriquez, Carrie Bearden

    Semel Institute for Neuroscience & Human Behavior, Los Angeles, California, United States

    Background: The Research Domains Criteria (RDoC) initiative identified a series of constructs including “working memory,” each of which is putatively specified by features at different levels or “units of analysis” including genetic, molecular, cellular, physiological, neurocognitive, and symptom levels. Specification of these constructs implies causal links between levels of analysis, but so far, most research has examined only two levels at a time, and the nature of associations of findings across multiple levels of analysis remains sparse. Further, most studies have focused on examining individuals from a discrete diagnostic group in comparison to relevant control groups, but this design may be suboptimal if the aim is to identify transdiagnostic characteristics that will better identify future diagnostic markers and targets for treatment. We ascertained relevant samples and examined them using MRI, EEG, neurocognitive, symptom, and disability assessments putatively relevant to the RDoC Working Memory domain.

    Methods: To avoid biases associated with recruiting diagnostic groups, we examined two groups of individuals comprising both men and women: (1) Care-Seeking (CS, n=106) individuals from the community, recruited through clinics and clinical web-sites, if they were seeking help for the way they were thinking or feeling; and (2) Non-Care-Seeking (NCS, n=44) individuals from the same community, who were not seeking care and had not done so within the year preceding enrollment. All participants had the same battery of clinical, cognitive, EEG and MRI assessments. Clinical assessments included MINI 7.0 neuropsychiatric interview, multiple symptom rating scales, and the WHODAS 2.0 disability assessment. The neurocognitive tests on working memory included both “complex multifactorial” measures (e.g., Digit Span, Letter-Number Sequencing, Symbol Span, Spatial Addition, Automated Operation Span) and “cognitively specific” measures (e.g., Lateralized Change Detection [LCD], Spatial Working Memory Capacity [SCAP], Dot Pattern Expectancy [DPX], Switching Stroop Color Word [SSCW], Recent Negatives [RN], Delayed Face Recognition [DFR]). We acquired EEG during an eyes open/eyes closed rest condition, and during 4 of our cognitive tasks (LCD, DPX, SCAP, and DFR). MRI measures included: structural MRI (sMRI) acquired using MPRAGE sequences; fMRI measures at rest from which graph theory metrics of functional connectivity were extracted; and fMRI in an activation paradigm using the DFR task to identify load-sensitive effects at supra-capacity levels of span. We used a combination of dimension reduction strategies and general linear models to examine both the structure of measures within levels of analysis and relations across levels of analysis.

    Results: CS and NCS groups were well matched on age, sex, personal and parental education, but the CS group had lower personal and family income. The CS group had more categorical mental disorders than the NCS group, the most prominent of which were current Major Depressive Disorder in 50% of CS compared to 21% of NCS, and Anxiety Disorders in 25% of CS compared to 14% of NCS). Psychotic disorders were found in 7.5% of CS and 0% of NCS individuals. Disability and symptom data were explained by 6 or 7 factors comprising “generalized psychopathology”, “depression-anxiety”, “psychoticism”, “agitation-mania”, and “somatic concern.” Neurocognitive data were best explained by models comprising “general cognitive ability”, “complex span”, “relational-long term encoding/retrieval”, “goal maintenance”, “interference control”, and “storage capacity.” Disability was well explained by the clinical measures, which accounted for 75% of variance in WHODAS scores. Neurocognitive, EEG, and MRI measures did not add to the prediction of disability beyond clinical assessments, and among the more “basic” measures most did not share more than 5% of the variance in functional disability, but sMRI measures did account for 13% of the variance in disability, and total gray matter volume alone explained 7% of variation in disability across groups. In contrast, there were more robust associations among neurocognitive, EEG and MRI variables, with typical multivariable associations sharing 10% of variance between sets of variables within each level.

    Conclusions: The results suggest that “multi-level” models of RDoC dimensions face several challenges, including limits in the strength of associations across levels. While we observed relatively robust associations of basic indicators like total gray matter volume reduction with the severity of disability, this effect was not obviously mediated via EEG, fMRI or cognitive measures. There do exist patterns of association among neurocognitive, EEG and fMRI measures that may define meaningful functional neuroanatomic dimensions, but so far their relations to higher level symptoms, syndromes, and disability remain less clear.

    Keywords: Neurocognitive Functioning, Disability, Research Domain Criteria (RDoC), EEG, MRI

    Disclosure: Nothing to disclose.

    M156. Effects of Transcranial Infrared Stimulation on Neural Information Flow in Healthy Volunteers

    Elisa Kallioniemi*, Tyrell Pruitt, Xinlong Wang, Mustafa Husain, Hanli Liu

    University of Texas Southwestern Medical Center, Dallas, Texas, United States

    Background: Transcranial infrared brain stimulation (TIBS) is a novel non-invasive neuromodulation method which uses high power infrared light to alter neural activity via intracellular metabolism. So far, TIBS has been observed to induce changes in neural activity and behavior. For instance, in healthy controls TIBS has been shown to change both reaction time and emotional state. Furthermore, some evidence exists that TIBS might have therapeutic effects in neurological and psychiatric disorders. Therefore, in this study our aim was to gain more understanding on how TIBS affects the neural function in healthy volunteers.

    Methods: We measured thirty healthy volunteers (24=males, age range 21-30 years). The protocol included both an active and sham TIBS. Each participant underwent both conditions which were separated by 14 days to avoid any potential after effects from TIBS. In both conditions, the stimulation was given with the same FDA approved laser (Model CG-5000 Laser, Cell Gen Therapeutics LLC, Dallas, TX, USA). In the sham condition, a black cap was inserted in between the laser aperture and the participant to prevent any infrared light exiting from the device. TIBS was administered at a wavelength of 1064 nm for 11 minutes and was targeted to the right forehead just below the Fp2 electrode. This target was chosen, as previous studies have shown TIBS to enhance cognitive performance when targeted here. The order of the conditions was randomized, and the participants were blind to whether they got active or sham TIBS at each session. Throughout the experiment, the participant and the researcher wore protective goggles.

    To evaluate the neural effects of TIBS, we acquired high-density electroencephalography (EEG) with 64-active electrodes using the BioSemi ActiveTwo EEG system (BioSemi BV, Amsterdam, Netherlands). EEG was recorded with a sampling frequency of 256 Hz and the ground and reference electrodes were located near the inion. The measurement contained three EEG parts: a 2-min baseline, an 8-min laser stimulation and a 3-min post-stimulation period.

    The offline processing of EEG data was done using the EEGLAB toolbox and custom Matlab scripts. Effects of TIBS on EEG activity were evaluated with Granger causality test on the correlation coefficients between EEG data at each electrode with a custom RStudio script. Granger causality measures the directional information flow, i.e., flow of EEG activity, between EEG electrodes and can thus, estimate how TIBS influences the information processing in a resting brain. Granger causality was evaluated at the alpha frequency band. Information flow estimates with Bonferroni corrected p-values smaller than p=0.05 were considered.

    Results: Active TIBS was able to significantly (p≤0.05) alter the information flow in the alpha band of a resting brain. During TIBS, information spread from the frontal areas to the right premotor cortex. Interestingly, TIBS also caused information to flow from the left hemisphere to the right premotor cortex. No changes in sham condition were observed. Instead, the information flow in the sham condition resembled of that observed in the default mode network. This network is the basic functional network measured in a resting brain when the brain is not under the influence of any intervention.

    Conclusions: TIBS is a promising neuromodulation method able to alter neural information flow (EEG activation) in a resting brain. Furthermore, TIBS was found to be safe and feasible in healthy volunteers.

    Keywords: Neuromodulation, Infrared, EEG

    Disclosure: Nothing to disclose.

    M157. Elucidation of the Role of Serotonergic IL-1R Signaling in the CNS

    Paula Gajewski-Kurdziel*, Elizabeth Potts, Matthew Robson, Nicole Baganz, Xaioyu Lui, Ning Quan, Randy Blakely

    Florida Atlantic University, Jupiter, Florida, United States

    Background: Serotonergic circuits are involved in many complex behaviors where dysfunction is believed to support multiple neuropsychiatric disorders, including anxiety, depression, autism, and schizophrenia. Although often treated monolithically, serotonergic projections arise from subpopulations of neurons that ultimately confer specific behavioral outcomes following activation. One subpopulation of serotonergic neurons that has drawn recent attention, but for which little is understood, are those that support the transduction of peripheral immune system activation into changes in discrete behaviors. Prior RNA expression studies suggest non-uniform expression of the interleukin 1 receptor (IL-1R), indicating that specific pathways dictate the serotonergic contributions to the behavioral actions of the inflammatory cytokine IL-1β (Okaty et al., 2015). Additional evidence points to a subpopulation of serotonergic neurons in the dorsal raphe interfascicular region that respond to stimulation of the peripheral immune system, suggesting that disruptions in immune system function could have effects on serotonergic neurotransmission (Hollis et al., 2006). Furthermore, peripheral administration with the bacterial mimetic, lipopolysaccharide (LPS), results in rapid, significant increases in activity of the serotonin transporter, supporting serotonergic neurons as a mediator of the CNS and behavioral actions of peripheral immune system activation (Zhu et al., 2010). Here we report evidence that serotonergic IL-1Rs mediate the responsiveness of serotonergic neurons to peripheral immune activation and contribute significantly to CNS IL-1β production.

    Methods: Until recently, detection of expression of IL-1Rs at the protein level has proved difficult, and as such IL-1R modulated 5-HT circuits have remained poorly defined. Using transgenic mouse lines in which we specifically eliminate IL-1Rs in serotoninergic neurons or restore serotonergic expression of IL-1Rs on an IL-1R knockout (KO) background, we explored both the serotonergic distribution and the role of serotonergic IL-1Rs in mediating immune-system related changes. To assess whether the peripheral immune system’s effects on serotonergic neurons are mediated by IL-1Rs, we acutely administered LPS (0.2 mg/kg i.p.) or saline to adult (8-12 wks) male ePet1:Cre:IL-1Rflx/flx or ePet1:Cre:IL-1Rr/r mice. One cohort was then sacrificed by transcardial perfusion with PBS and 4% PFA one hour after treatment, followed by brain extraction and post-fixation of the brain in 4% PFA for 24 h before being moved to PBS until sliced on a compresstome at 100 µm. Slices containing dorsal raphe were immunolabeled with 5-HT (Immunostar #20079, 1:1000) and cFos (Abcam #ab190289, 1:5000) and cleared using an established glycerol gradient (Selever et al., 2011), with cell counts obtained by a blinded observer. Another cohort was sacrificed by rapid decapitation one hour after treatment, followed by dissection of the midbrain and hippocampus, which were processed for RNA isolation. qPCR was conducted using the Taqman gene expression assay, with Taqman probes for 18 S (HS99999901) and IL-1β (Mm00434228). Additionally, in vivo chronoamperometry was utilized to determine the effect of local IL-1β (2 ng) on 5-HT clearance in wild type mice.

    Results: We have gathered evidence to support the existence of subsets of serotonergic neurons that express IL-1Rs, with enrichment evident in the dorsal and lateral subsections of the dorsal raphe, and with the capacity to respond to IL-1β via IL-1Rs. Acute treatment with LPS (0.2 mg/kg) caused an upregulation of IL-1β expression in the midbrain of wildtype mice as monitored by qPCR. A constitutive KO of IL-1Rs completely eliminated midbrain IL-1β expression after LPS treatment. Strikingly, selective elimination of IL-1Rs in 5-HT neurons also significantly reduced LPS-mediated increases in IL-1β. Additionally, re-expression of IL-1Rs in 5-HT neurons on a constitutive KO background significantly rescued LPS-induced IL-1R expression in the midbrain, effects also seen in the hippocampus. Our initial studies demonstrate that the same acute treatment with LPS increases serotonergic neuronal activation as measured by cFos, and in mice that lack serotonergic expression of IL-1Rs, the number of cFos-positive 5-HT neurons is significantly decreased. Finally, local application of IL-1β in the CA3 subregion of the hippocampus resulted in increased 5-HT clearance.

    Conclusions: These data provide the first evidence, to our knowledge, of 5-HT neurons as contributing significantly to CNS immune activation, specifically IL-1β induction, in response to peripheral immune system activation. Our anatomical studies point to a subset of 5-HT neurons with high IL-1R expression that we propose contribute significantly to IL-1β induction and LPS induced 5-HT neuron activation. Our current efforts are aimed at determining requirements for serotonergic IL-1R expression for the physiological and behavioral actions of IL-1β, elucidating the projections and inputs of IL-1R expressing 5-HT neurons, and establishing the necessity of serotonergic IL-1R expression for IL-1β driven behavioral responses. We see the broader significance of our work as elucidating a mechanistic connection between observations of elevations in inflammatory markers in psychiatric disorders and the role played by 5-HT in modulating circuits that support disease risk.

    Keywords: Serotonin, Immune System, Interleukin 1beta, Central Nervous System

    Disclosure: Nothing to disclose.

    M158. Individual-Level EEG Connectomics Reveals Differential Functional Connectivity Patterns in Different Subtypes of Resting-State

    Jacob Van Doorn, Rael Cahn, Arnaud Delorme, Olusola Ajilore, Alex Leow*

    University of Illinois at Chicago, Chicago, Illinois, United States

    Background: As part of a study on the effects of different techniques of meditation on electroencephalographic (EEG) activity, this pilot study aimed to determine EEG functional brain connectivity signatures in two forms of resting state activity among nonmeditators by leveraging computational connectomics.

    Methods: Overview: EEGs were recorded from individuals without prior meditation experiences, who were instructed to observe and focus on their breathing or engage in emotionally neutral autobiographical thinking as a form of mind wandering for twenty minutes each. Functional connectomes of each individual were formed using weighted phase lag index (WPLI). For each individual, dynamic EEG connectomes as a function of time were then analyzed using a recently developed state-space manifold learning procedure, the Graph Dissimilarity Embedding procedure (GDE), followed by visualization via Minimum Spanning Tree (MST) and Isomap dimensionality reduction. Due to the intra-individual nature of the recovered state-space manifold, we then devised a permutation testing procedure to determine whether these two ‘subtypes’ of resting exhibit statistically significant differences.

    Subject recruitment: 32 healthy control subjects of both sexes were recruited and data collection took place at the Meditation Research Institute in Rishikesh, India. Subjects were compensated and reimbursed for travel, accommodation, and meals. The project was approved by the local Indian ethical committee and the IRB of the University of California San Diego. Control subjects were chosen for inclusion in this study based on age and gender and their absence of a meditation practice. For the “meditation condition,” they were instructed to focus on the sensation of their breath for the duration of the recording period (“keep paying attention to the sensations of the breath, both of the inhalation and the exhalation. If your mind starts to wander, please bring it back to your breath.”)

    Data collection and preprocessing: We recorded data using a 64 + 8 channel Biosemi Active-Two amplifier system and a 10–20 Headcap standard 64-channel cap from the same company. Using external electrodes, we also recorded right and left mastoid electrodes as well as vertical and horizontal electrooculogram (EOG) by placing two periocular electrodes above and below the left eye and two electrodes at both the left and right outer canthi. Data processing was done using the EEGLAB open source software version 12 running on Matlab R2009b under a Linux operating system. EEG data were first referenced to the right mastoid and down-sampled from 1024 Hz to 256 Hz. The first 10 minutes of the meditation block (MED) are considered a preparatory period helping to relax and deepen the meditation practice. Subjects practiced breath awareness throughout the whole MED block. For consistency, we kept the length of the Instructed Mind-Wandering block (IMW) equal to the length of the meditation block. When analyzing the data, we compared the last 10 minutes of the breath-focused MED block to the last 10 minutes of the IMW block. We then applied a high-pass filter at 1 Hz using an infinite impulse response (IIR) filter with a transition bandwidth of 0.3 Hz and an order of 6. We automatically removed portions of the signal presenting non-stereotyped artifacts using EEGLAB’s pop_rejcont function. The data were first segmented in 1-second epochs with 0.5-second overlap. Segments of 8 contiguous epochs in which the 0–10 Hz frequency band and the 35–128 Hz frequency band had an amplitude higher than 17 and 14 decibels respectively were labeled as artifactual. We used this rejection procedure to ensure that artifact rejection was uniform for all subjects. Rejection of low-frequency segments helped remove signals related to head and body movements. Rejection of high frequency activity helped reject data portions of muscular activity. The data was then checked visually for any potential remaining artifact. We then manually identified and removed bad electrodes (from 0 to 18 bad electrodes per subject, average of 5 electrodes removed per subject).

    Results: 16 subjects out of 32 were included for analysis based on the number of good electrodes after preprocessing (usable number of electrodes 53 to 70). After computing the dynamic EEG connectomes associated with each resting subtype (the connectome at each time point is an electrode x electrode x frequency 3D matrix; frequency range 2-110 Hz) we constructed individual-level GDE, yielding the pairwise geodesic between two connectomes. GDE, visualized by both MST and Isomap, show qualitative separation between the two states. Using the average inter-class geodesic minus the average intra-class geodesic as the test statistic we sampled the null distribution (i.e. connectomes do not differ between states) by randomly assigning all connectome labels (to either MED or IMW). Using this procedure, we demonstrated significant connectome differences between MED and IMW in all of 16 participants (p-value < 0.001) and an effect size ranging from 10 to 120 standard deviations.

    Conclusions: In sum, our pilot analyses suggested distinct differential EEG functional signatures between these two forms of resting state activity – instructed mind wandering and meditative engagement with the sensations of the breath.

    Keywords: Connectomics, Electroencephalography, Resting State Functional Connectivity

    Disclosure: Nothing to disclose.

    M159. A Randomized, Double-Blind, Placebo-Controlled Crossover Trial of the NMDA Receptor Antagonist Memantine in People Receiving Chronic Prescription Corticosteroid Therapy: Effects on Hippocampal Subfields and Memory

    E. Sherwood Brown*, Alexandra Kulikova, Erin Van Enkevort, Alyson Nakamura, Elena Ivleva, Nicholas Tustison, Jared Roberts, Michael Yassa, Changho Choi, Alan Frol, David Khan, Miguel Vazquez, Traci Holmes, Kendra Malone

    University of Texas Southwestern Medical Center, Dallas, Texas, United States

    Background: Corticosteroid excess is associated with decreases in memory performance and hippocampal volume in humans as well as in animal models. In animal models, the CA3 region of the hippocampus appears to be particularly sensitive to the effects of stress or corticosteroids, although changes in the CA1 region are also reported. Hippocampal changes with corticosteroids, such as dendritic shortening, in animal models are prevented by NMDA-receptor antagonists. In this report, the effect of the NMDA receptor antagonist memantine on memory and hippocampal subfield volume was examined in people receiving chronic prescription corticosteroid therapy.

    Methods: Outpatient adults receiving chronic oral corticosteroid therapy were randomized to memantine and placebo for 24 weeks each using a crossover design with a four-week washout period between treatment conditions. Declarative memory was assessed using the Hopkins Verbal Learning Test (HVLT) and structural magnetic resonance imaging (MRI) was obtained at baseline and after each medication condition. Neuroimaging was performed on Philips 3 T scanner at the Advanced Imaging Research Center, UT Southwestern Medical Center. Sagittal T1–weighted MPRAGE scans (TE/TI/TR=3.8/875/1360 ms, 256×256×160 mm3 field of view, 160 slices, voxel size 1 ×1 ×1 mm3) were collected and used for hippocampal subfield segmentation. The segmentation was performed using a consensus labeling approach with multi-spectral atlas-based registration based on a well-validated atlas composed of T1 and T2-weighted high-resolution structural scans from cognitively normal subjects that were manually labeled using a highly reliable anatomical protocol used in prior published work for segmentation of hippocampal subfields. To propagate a weighted consensus labeling from the expertly labeled atlas set to the study scans, we spatially normalized the atlas set to the study scans and applied the joint label fusion technique. Advanced Normalization Tools (ANTs) package was used for both spatial normalization and consensus-based labeling (i.e., joint label fusion). First, the intra-subject atlas T1/T2 rigid transforms were calculated. To minimize total number of deformable registrations, a “pseudo-geodesic” approach to align the data was used. This required construction of an optimal T1-weighted template representing the average shape/intensity information of the T1 component of the atlas set. Deformable transformations between each T1-weighted image of the study cohort and the T1 atlas template were calculated. Transformation between the atlas labels and unlabeled study cohort image was then computed by concatenating the T1 atlas /T2atlas rigid transformation, the T1atlas /T1 template deformable transformation, and the T1 template/and T1subject deformable transforms. Once the atlas set was normalized to the unlabeled subject, regional labeling was determined using weighted averaging where the weighting takes into account the unique intensity information contributed by each atlas member. After visual quality assessment to confirm the output of the labeling procedures, voxels within the labeled regions were counted and multiplied by the voxel resolution to calculate volumes in cubic millimeters. Utilizing a mixed-model approach, structural data were examined by an ANOVA, and memory was assessed with a multi-level longitudinal model.

    Results: A total of 46 participants (n=28 women) were randomized and received memantine and placebo. The majority of participants were receiving chronic corticosteroid therapy to prevent renal transplant rejection. The mean dose of corticosteroid therapy was 7.6±6.4 mg/day and the mean duration was 58.8±67.3 months. The mean age of the participants was 43 years. Controlling for baseline volume, the left DG/CA3 region was significantly larger (~ 2%) following 24 weeks of memantine as compared to placebo (p=0.011). A trend in the same direction was observed with the right CA1 subfield (p=0.055). HVLT scores did not differ significantly between the memantine and placebo conditions. Memantine was well tolerated.

    Conclusions: Memantine was associated with a larger CA3/DG volume than placebo when given to corticosteroid-treated patients. The findings translate preclinical research on the ability of NMDA receptor antagonists to block the effects of corticosteroids on the hippocampus. Furthermore, the findings suggest memantine may have potential as a neuroprotective agent in people receiving prescription corticosteroids therapy or with elevated levels of endogenous cortisol.

    Keywords: corticosteroids, Hippocampal subfields, Memory, Memantine, Corticosteroids

    Disclosure: Otsuka, Grant, Allergan, Advisory Board

    M160. Cerebellar [18 F]FDG Uptake is Positively Correlated With Time Elapsed Since Blast-Related Mild Traumatic Brain Injury in Veterans

    Garth Terry*, Mark Muzi, Donna Cross, James Meabon, Murray Raskind, David Cook, Elaine Peskind

    VA Puget Sound Health Care System and University of Washington, Seattle, Washington, United States

    Background: Blast-induced mild traumatic brain injury (mTBI) is the “signature injury” of the wars in Iraq (OIF) and Afghanistan (OEF). Blasts account for 70-88% of mTBIs sustained in OEF/OIF, with affected Servicemembers usually experiencing multiple blast mTBIs. Even years after returning from deployments, many Veterans with blast-related mTBI report postconcussive cognitive symptoms, behavioral symptoms, and somatic symptoms, leading to substantial disability and interference with job and family relationships.

    Our group’s initial [18 F]FDG-PET study in a sample of 12 male OIF Veterans with mTBI and 12 cognitively and medically healthy community controls (7 male, 5 female) demonstrated a consistent pattern of less [18 F]FDG uptake in infratentorial structures and medial temporal cortex in mTBI Veterans compared to the controls. In our follow-up study of 33 Veterans with blast mTBI and 16 deployed control Veterans with no lifetime history of mTBI, voxelwise analysis demonstrated hypometabolism in bilateral parietal lobes, left somatosensory cortex, and right visual cortex, as well as in parahippocampus for Veterans with 20 or more blast mTBI. There was a strong correlation between glucose metabolism in the cerebellum and number of blast exposures.

    We sought to identify if our prior findings are present in our much larger cohort, and assess if hypometabolism previously described is affected by time since injury.

    Methods: The local institutional review boards approved all procedures and all the participants provided written informed consent before enrollment into the study. Veterans with blast-related mTBI (n=79; average blast mTBI exposures 22), with history of military deployment and no lifetime history of any TBI (n=29), and community controls with no history of TBI or deployment (n=12) underwent thorough history and clinical characterization as part of our ongoing longitudinal study. A standard clinical 15-minute PET acquisition was performed 60 minutes after injection 8-10 mCi of [18 F]FDG on either a GE Advance scanner or Philips Gemini PET/CT, and data underwent OSEM reconstruction. Three 5-minute frames were averaged following motion correction, smoothed 8×8×8 mm, and transformed directly into MNI standardized space using SPM12. Individual images were scaled for intensity using a VOI applied to parenchyma, yielding fractional uptake (unitless) as the measured value in PMOD. The AAL VOI library (119 regions) was applied for regional analysis, and data tested with 2-sided t-test with Bonferroni correction of multiple comparisons (p<0.004). Replication regression analysis of blast x fractional uptake in the cerebellum was conducted with a priori p<0.05. Images were separately analyzed on a voxelwise basis using SPM12 with 6 mm smoothing in the nonparametric toolbox (SnPM) which implements a single tailed test, with age as a covariate, and adjusted significance value (pFWE-corrected<0.05, equivalent to t-score >5).

    Results: Compared to deployed or community controls, Veterans with mTBI demonstrated less fractional uptake in areas of frontal and temporal cortex, and greater fractional uptake in deep gray of globus pallidus and putamen by both VOI and voxelwise analysis and correction for multiple comparisons (left globus pallidus p<5×10-10). Similar to our previous finding, fractional uptake in cerebellum (specifically, regional cerebellum 7-9 and vermis 6-9) was negatively correlated with log transformation of number of blast mTBI (i.e., more blast exposures was associated with less [18 F]FDG uptake; vermis7 r2=0.09, p<0.008). However, fractional uptake was positively correlated with time since last blast mTBI (i.e., individuals imaged at longer intervals after exposure had greater [18 F]FDG uptake; left cerebellum9 r2=0.20, p<0.00003).

    Conclusions: Veterans with history of blast mTBI demonstrate less fractional [18 F]FDG uptake in cortex and cerebellum compared to individuals with no history of TBI, and greater fractional uptake in regions of basal ganglia. Fractional uptake in cerebellum was negatively correlated with number of blast mTBI exposures, and was positively correlated with time elapsed since blast exposure.

    Keywords: Mild Traumatic Brain Injury, fluorodeoxyglucose, Positron emission tomography, combat veteran

    Disclosure: Nothing to disclose.

    M161. Impact of Resiliency Training on Mental Health Symptoms and Resting State Connectivity for First-Semester College Students

    Robin Aupperle*, Elisabeth Akeman, Tim McDermott, Masaya Misaki, Ashley Clausen, Namik Kirlic, Lisa Cromer

    Laureate Institute for Brain Research, Tulsa, Oklahoma, United States

    Background: College students experience substantially higher rates of anxiety and depression (30% prevalence) than the general population, which can have deleterious effects on academic progress and success (Ibrahim et al., 2013; Regehr et al., 2013). This highlights the importance of identifying preventative interventions that can be used widespread across college campuses. Given the known time of stressor onset and high risk for experiencing mental health symptoms, university settings also offer a unique opportunity for development, evaluation, and widespread dissemination of programs targeting anxiety and depression. While meta-analyses indicate that cognitive-behavioral (CBT) or mindfulness interventions can be beneficial for college students, these have often been long (13-14 session) interventions (Conley et al., 2015). The current project sought to identify the potential mental health benefits a brief, scalable intervention delivered during college orientation classes. In addition, we aimed to identify potential neurobiological mechanisms (neural connectivity) for the resiliency training, thus informing further refinement of intervention targets or personalized medicine approaches.

    Methods: A total of 126 undergraduate students completed four weeks of resiliency training as part of orientation classes during the first semester of college (73 female). Training focused on value identification, goal-building, mindfulness, and the growth mindset. A sample of 126 students from orientation classes that did not receive training and who were matched on baseline depression scores were identified as the comparison group (83 female). All participants complete NIH PROMIS® anxiety and depression measures and the Connor-Davidson Resilience Scale pre and post training and at the end of the semester (e.g., “finals week”). The Frieburg Mindfulness Inventory (FMI) and a cognitive-behavioral skills questionnaire were completed pre, weekly during training, and post training. A subset of participants (23 training; 17 no-training) completed functional magnetic resonance imaging (Discovery MR750 3 T) with a seven-minute resting state scan (96×96 matrix, 2×2mm2 in-plane resolution, 35 axial slices, 2000ms TR, 207 volumes, SENSE factor R=2). Linear mixed models were used to examine training by time interaction effects on self-report survey scores (covarying for gender and college [e.g., Arts and Sciences, Engineering]). Connectome-wide association analysis was conducted voxel-wise with resting-state fMRI data using multivariate distance matrix regression (MDMR) based on the Euclid distance matrix. MDMR offers a data-driven methodology for identifying regions for which connectivity with the rest of the brain is impacted by conditions of interest (Shehzad et al., 2014), which included the time by training effect in the current study. To correct for multiple comparisons, permutation testing was utilized to identify the required cluster size (896 mm3) for voxel-wise p < .005.

    Results: Resiliency training resulted in significant reductions in depression symptoms compared to the control group (F(2,452)=4.08, p=.018), with effects most robust at the end of the semester (t=-2.85, p=.005; Cohen’s d = .26). There were also trend effects on anxiety (F(2,448)=2.51, p=.083). This was accompanied by within-training improvements in mindfulness (F(6,1151)=6.89, p<.001) and cognitive-behavioral skills (F(6,1147)=4.91, p<.001) but not resilience (F(2,44)=.77, p=.464). MDMR analysis of resting connectivity scans revealed regions of the bilateral dorsolateral prefrontal cortex (dlPFC; BA 9), left precuneus (BA 31), and left superior parietal cortex (BA 40) in which whole-brain connectivity was impacted by training.

    Conclusions: Results indicate that a brief, four-session resiliency training course that is incorporated into first-semester undergraduate classes can have a significant, beneficial impact on mental health over the first semester of college. Changes in cognitive-behavioral and mindfulness-based skills may be important psychological mechanisms for the observed benefits, while self-beliefs regarding one’s resilience may be more difficult to change. Such resiliency training may also lead to changes in whole-brain connectivity within regions involved in executive function and decision-making (dorsolateral PFC), attention processing (superior parietal), and self-relevant processing (precuneus). These findings have implications for how colleges may approach the problem of rising mental health concerns and offer insights into potential mechanisms of preventative-based resiliency programs.

    Keywords: resilience, Depression, Functional MRI (fMRI), Cognitive Behavioral Therapy

    Disclosure: Nothing to disclose.

    M162. How Does Electroconvulsive Therapy Suppress Self-Injurious Behavior Associated With Intellectual and Developmental Disabilities

    Alec Stepanian, Melissa Staley, Cindy Jiang, Sanmit Adhikari, Andrew Chang, Jay Baraban, Irving Reti*

    Johns Hopkins University, Baltimore, Maryland, United States

    Background: Electroconvulsive therapy (ECT) is highly effective for suppressing severe intractable self-injurious behavior (SIB) in many patients with intellectual and developmental disabilities (IDD). However, in contrast to depressive patients who respond to ECT and can typically be successfully weaned off ECT with the aid of antidepressant medications, IDD patients with SIB remain dependent on long-term frequent maintenance ECT, often starting at a young age, the long-term effects of which are unknown. Accordingly, it is important to learn more about why ECT is effective in these patients in order to generate long-term effective non-convulsive treatments.

    It is not known why IDD patients being treated with ECT for SIB seem to have a different pattern of response from depressive patients being treated with ECT. However, it suggests that ECT’s antidepressant mechanism of action may be different from the ECT response mediating suppression of SIB. Moreover, we have found that the GABAergic agonist class of drugs, benzodiazepines (which do not act as antidepressants), are sometimes effective in suppressing SIB in IDD patients, especially when they have concomitant intermittent catatonia, which occurs frequently. These observations raise the possibility that ECT may act through the GABAergic system to mediate suppression of SIB in IDD. In fact, numerous lines of evidence suggest that ECT enhances GABAergic tone, including increased cortical GABA concentration following a course of ECT.

    Methods: To further characterize how ECT suppresses SIB in IDD, we are using mouse models that display stereotyped excessive self-grooming, mimicking the SIB seen in IDD. We are using the conditional knockout mouse, Viaat-Mecp2-/y, which displays Rett Syndrome phenotypes, excessive self-grooming and reduced GABA levels in the forebrain (Chao et al, 2010). We have found that over the course of a week following a single electroconvulsive seizure (ECS), excessive self-grooming progressively declines in these mice reaching significant suppression between days three and five, before returning to baseline by day eight after treatment (Chang et al., 2016). We are also using a genetically distinct line of mice that also displays stereotyped excessive self-grooming, namely the Shank3B-/- mice (Peca et al., 2011), but that does not demonstrate the GABA dysfunction observed in Viaat-Mecp2-/y mice. In addition to monitoring self-grooming following ECS in these mice, we are also monitoring both tissue and extracellular GABA levels in the cortex, hippocampus and striatum, the latter by in vivo microdialysis and high-performance liquid chromatography (HPLC).

    Results: In contrast to the Viaat-Mecp2-/y mice, where stereotyped self-grooming was quite sensitive to ECS, we have found that ECS does not suppress the excessive self-grooming of Shank3B-/- mice one, three or five days after either a single ECS or after three ECS treatments spaced 48 h apart.

    In ongoing studies, we are monitoring extracellular GABA levels in the striatum before and after ECS by microdialysis and HPLC. In a mixed cohort of Shank3B-/- and wild-type mice, we found that after an initial spike in GABA levels following implantation of the microdialysis probe, levels come down quickly and stabilize within about 30 minutes. In preliminary results in this mixed cohort we have found that ECS appears to produce a marked increase in GABA release in the striatum within 30 minutes of a seizure.

    We have also begun monitoring GABA levels by immunohistochemistry and in preliminary results have observed more intense GABA cell body staining in the striatum of wild-type mice euthanized 60 minutes after ECS compared with mice exposed to sham stimulation.

    Conclusions: We have found differing responses of stereotyped excessive self-grooming to ECS in two genetically distinct mouse models, namely the Viaat-Mecp2-/y and Shank3B-/- lines. Whereas the excessive self-grooming of the Viaat-Mecp2-/y mice is suppressed by a single ECS, ECS is not effective in the Shank3B-/- line. These data raise the possibility that the response of GABA levels to ECS may differ between these two mouse models, which we are presently assessing. In particular, GABA release following ECS may be more marked in Viaat-Mecp2-/y than in Shank3B-/- mice. These studies could help shed light on how ECT suppresses SIB associated with IDD. They may also shed light on why some of these patients may respond better to ECT than others.

    Keywords: Self-Injurious Behavior, Electroconvulsive Therapy, GABA, Mouse Models

    Disclosure: Nothing to disclose.

    M163. Delineating the Hippocampal Circuitry Underlying Pair Bonding in Prairie Voles

    Jayme Temple, Michaela Best, Hannah Dougherty, Zoe Donaldson*

    University of Colorado Boulder, Boulder, Colorado, United States

    Background: Pair bonds are long-lasting social attachments that form between mating partners. While pair bonding is common among humans, the majority of mammals, including laboratory rats and mice, do not exhibit this trait. Instead, socially monogamous prairie voles, which form life-long pair bonds, provide an excellent model for studying attachment in adults. To date, the study of pair bonding in prairie voles has focused on neuromodulatory systems, including oxytocin, vasopressin, dopamine and endogenous opiates, while the specific neural circuits that modulate attachment remain largely unexplored. The ventral hippocampus has been implicated in regulation of emotions and social memory across a variety of rodent species. Thus, we asked whether the ventral hippocampus is required for a selective partner preference, a behavioral indicator of pair bonding.

    Methods: We used ibotenic acid to bilaterally lesion the ventral hippocampus in female prairie voles. After two weeks of recovery, estrogen-primed females were paired and mated with a male. Twenty-four hours later, we performed a partner preference test.

    Results: Animals with sham injections spent significantly more time huddled with their partner than a novel individual (n = 6, p = 0.002), indicating that they had formed a bond. In contrast, animals with ventral hippocampal lesions did not spend more time huddled with their partner (n = 10, p = 0.45).

    Conclusions: Our data indicates that an intact ventral hippocampus is required for partner preference. However, because lesions result in permanent tissue disruption, it remains unclear whether an intact hippocampus is required for preference formation, expression of an existing bond, or both. Likewise, it is unclear which neuronal subsets within this region modulate partner preference. Thus, our ongoing work is examining the specific role of different hippocampal projections during pair bond formation versus expression. This research will provide a novel dissection of the role of hippocampal systems in pair bonding behavior, potentially providing valuable insights into how disruption of these circuits contributes to social deficits.

    Keywords: Pair Bond, Hippocampus, Social Behavior, Prairie Voles

    Disclosure: Nothing to disclose.

    M164. Altered Functional Connectivity of the Spatial Network in Children With Nonverbal Learning Disability, Specific Learning Disability in Reading, and Typically Developing Children

    Amy Margolis*, Sarah Banker, David Pagliaccio, Lauren Thomas, Rachel Marsh

    Columbia University Medical Center/NYSPI, Brooklyn, New York, United States

    Background: Nonverbal Learning Disability (NVLD) is characterized by deficits in visual-spatial processing in contrast to strengths in verbal reasoning, whereas reading disorder is characterized by deficits in verbal reasoning and strengths in spatial processing. The current study compared resting state functional connectivity of the spatial network among children with NVLD, children with reading disorder (RD), and typically developing (TD) children. We hypothesized that children with NVLD would have altered connectivity within the spatial network relative to the other groups, consistent with the phenomenology of the disorder.

    Methods: Resting State fMRI data (rsfMRI) were acquired from 86 children. Thirty-six children were excluded for excessive motion (less than 5 minutes of usable data with FD < 0.5 mm), leaving a total of 54 children in the analysis (15 TD, 19 NVLD, 20 RD). Using the CONN toolbox, we performed ROI-ROI analyses among 12 seeds of a spatial network previously defined by Arnold [Journal of Cognitive Neuroscience, 26(2), 380 (2014)] and evaluated group differences while controlling for age, sex, and mean motion. We evaluated associations between network connectivity and functional impairment on the Child Behavior Checklist across all subjects while controlling for age, sex, mean motion, and diagnostic group.

    Results: Two ROI-ROI connections in the spatial network showed significant group differences. Connectivity from left posterior cingulate (PCC) to right retrolimbic area (RA) was reduced in children with NVLD relative to typically developing children and children with RD. Across all children, such connectivity was inversely associated with internalizing problems (b=-14.550, t=-2.741, p=0.009), social problems (b=-6.438, t=-2.331 p=0.024), thought problems(b=-9.904, t=-2.575, p=0.013), and total problems (b=-12.408, t=-2.338, p=0.024), and positively associated with school competence (b=7.307, t=2.110, p=0.040).

    Connectivity from left PCC to right cerebellum was reduced in children with NVLD and children with RD relative to TD children. Given that both clinical groups showed reduced connectivity relative to TD children, the two groups were combined for behavioral analyses. Connectivity from left PCC to right cerebellum was positively associated with activities competence (b=19.37, t=2.02, p=.05) and total competence (b=16.0, t=1.99, p=.05).

    Conclusions: Reduced cortico-cortical connectivity characterized children with NVLD and associated with behavioral impairment consistent with NVLD but not RD (internalizing and social problems) pointing to the specificity of these findings to NVLD. Altered connectivity in the spatial network may contribute to the psychological deficits that accompany NVLD, perhaps providing a novel target for treatment.

    Reduced cortico-cerebellar connectivity characterized both groups of children with learning disabilities and associated with reduced competence in activities including sports, hobbies, and chores, possibly reflecting the motor deficits often reported in both NVLD and RD. In addition, reduced cortico-cerebellar connectivity associated with their overall sense of competence, possibly reflecting a domain general phenomenon experienced by children with learning disabilities. Such findings of reduced connectivity between the default mode network (PCC) and cerebellum in children with learning disabilities provide direction for future studies of children with learning disabilities.

    Keywords: Learning Disability, Resting State Functional Connectivity, Visuospatial Ability

    Disclosure: Nothing to disclose.

    M165. Stress and the Brain’s Hemodynamic Response: A New Perspective on Stress Response Regulation in Humans

    Immanuel Elbau*, Immanuel Elbau, Benedikt Brücklmeier, Manfred Uhr, Janine Arloth, Darina Czamara, Victor Spoormaker, Michael Czisch, Klaas Enno Stephan, Elisabeth Binder, Philipp Sämann

    Max Planck Institute of Psychiatry, Munich, Germany

    Background: Ample evidence links dysregulation of the stress response to risk for psychiatric disorders. Yet, we lack an integrated understanding of mechanisms that are adaptive during the acute stress response but potentially pathogenic when dysregulated. One mechanistic link emerging from rodent studies is the interaction between stress effectors and neurovascular coupling, a process that adjusts cerebral blood flow according to local metabolic demands.

    Methods: Here we use functional magnetic resonance imaging in healthy humans (two cohorts, together spanning n=98, ~ 50 % female) during an acute psychosocial stress intervention and parallel multi-level assessment of the stress response (on the endocrine, autonomous nervous system and subjective experience level) to investigate the effect of stress on shape-characteristics of the brain’s hemodynamic response function (HRF). The HRF models how local neural activity elicits cerebral blood flow changes and reflects several biophysical processes including neurovascular coupling.

    Results: We show that acute psychosocial stress rapidly impacts the peak latency of the hemodynamic response function (HRF-PL) in temporal, insular and prefrontal regions in two independent cohorts of healthy humans. These latency effects occurred in the absence of amplitude effects and were moderated by regulatory genetic variants of KCNJ2, a known mediator of the effect of stress on vascular responsivity. Further, hippocampal HRF-PL correlated both with cortisol response and genetic variants that influence the transcriptional response to stress hormones and were associated with risk for major depression.

    Conclusions: We conclude that acute stress modulates hemodynamic response properties as part of the physiological stress response and suggest that HRF indices could serve as endophenotype of stress-related disorders.

    Keywords: Acute Stress, Functional MRI (fMRI), Hemodynamic Response Function, Neuroendocrine Responses, Neurovascular Coupling

    Disclosure: Nothing to disclose.

    M166. Sexual Dimorphic Effects of Restraint Stress on Anxiety and Recency Memory, Limbic System Activation and Synaptic Plasticity

    Aggeliki Velli, Kleanthi Chalkiadaki, Aikaterini Chatzaki, Kyriaki Sidiropoulou*

    University of Crete, Heraklio, Greece

    Background: Sex differences are present in several neuropsychiatric disorders. Women are more vulnerable to anxiety disorders and major depression while men exhibit worse cognitive symptoms in several psychiatric disorders, such as schizophrenia. A common environmental factor that precipitates or worsens the pathology of the above disorders is stress. Stressor exposure affects men and women differentially (Bangasser et al, 2017, Physiology and Behavior). However, we do not understand the neurobiological mechanisms involved. Our aim in this work was to investigate the sexual dimorphic effects of stress on males and females on the function of the prefrontal cortex (PFC), the amygdala and the hypothalamus.

    Methods: Adult male and female were initially subjected to two hours of restraint stress (RS), or left in their home-cage (NR), and then tested in the light-dark test (after 5 min), followed by the temporal order object recognition (TOR) test (1 h later) (n=9 for male NR, n=8 for male RS, n=9 for female NR and n=8 for female RS). In the light-dark test, mice were placed in the dark compartment of chamber and they were allowed to explore both the dark and the light compartments for 5 min. In the TOR task, mice are allowed to explore two identical copies of object A for 5 min and after 25 min two identical copies of object B for another 5 min. In the test session (25 min after the last one), mice were allowed to explore object A (old object) and object B (recent object). The mice brains were removed immediately after the end of the TOR test and prepared for immunofluorescence detection of the c-Fos protein. Brain regions of the limbic system, including the PFC, the amygdala, the hypothalamus and the dentate gyrus, were analyzed for the number of c-Fos-expressing cells. In a different cohort of animals, field excitatory postsynaptic potentials (fEPSPs) were recorded in PFC layer II, using the brain slice preparation, following the 2 h restraint stress. Tetanic stimulation resulted in long-term synaptic potentiation (LTP) of the fEPSP. All experiments were approved by the Institutional Animal Care and Use Committee of the University of Crete and were conducted in accordance with the European Union ethical standards. Statistical significance was ascertained by two-way ANOVA or t-test.

    Results: When mice were tested in the light-dark test, RS resulted in significant differences in the latency to exit the dark side (two-way ANOVA, F(3, 32) =3.4, p=0.03). Significant effects were observed both in the stress variable (NR vs RS) and the sex*stress interaction. Post-hoc LSD tests showed that the latency of female RS mice was significantly longer compared to female NR mice (p=0.004), male NR mice (p=0.02) or male RS mice (p=0.03). In the TOR task, male and female NR mice, as well as female RS mice, spent significantly more time exploring the recent object compared to the older object (t-tests, p=0.03, p=0.01, p=0.02, respectively), indicating that mice in these groups could successfully discriminate the temporal order of object presentation. However, there was no significant difference between recent and old object exploration in male RS mice (p=0.3), demonstrating that male RS mice exhibit impaired recency memory.

    Expression of restraint stress-induced c-Fos expression was examined in the cingulate cortex, the prelimbic cortex, the dentate gyrus of the hippocampus, the amygdala and the hypothalamus of RS male and female mice. Fos expression was enhanced in all of the above regions in RS female and male mice, compared to their respective NR mice. Fos expression was enhanced in female RS mice, compared to male RS mice, as well as compared to male and female NR mice in the cingulate cortex, the hypothalamus and the amygdala (4 brains in each groups tested so far, two-way ANOVA, F(3,13)=4.1, p=0.02, F(3,13)=3.5, p=0.04, F(3,13)=3.8, p=0.03, respectively).

    In addition, electrophysiological recordings in the PFC showed that both adult male and female NR mice could enhance the fEPSP response following tetanic stimulation, thus, resulting in LTP (Konstantoudaki et al, 2018, Journal of Neurophysiology). However, LTP (measured at 35-50 min post tetanic stimulation) in the PFC was significantly reduced in male RS (n=5) compared to male NR (n=6), female RS mice (n=5), and female NR mice (n=7) (two-way ANOVA, F(3, 20)=3.8, p=0.02).

    Conclusions: Restraint stress results in enhanced anxiety in female mice only and reduced recency memory in male mice. These behavioral effects are due to enhanced amygdala activation in female stressed mice and reduced LTP in the PFC of male stressed mice, respectively.

    Keywords: Prefrontal Cortex, Amygdala, Hypothalamus

    Disclosure: Nothing to disclose.

    M167. Inhibition of the Endocannabinoid Anandamide Degradation Restores Stress-Mediated Impairments of Contextual Fear Memory Extinction and Hippocampal In Vivo Long-Term Synaptic Plasticity

    Maria Morena*, Roberto Colangeli, Andrei S. Nastase, Alessia Santori, G. Campbell Teskey, Matthew Hill

    University of Calgary, Calgary, Canada

    Background: It is well established that stress has a profound impact on memory functions and the endocannabinoid system is a crucial modulator of stress-related behavioral responses.

    Acute stress reduces hippocampal endocannabinoid anandamide (AEA) levels and impairs long-term potentiation (LTP) at CA3-CA1 synapses in the hippocampus.

    Studies have shown that direct cortical input to the CA1 through the temporoammonic (TA) pathway is required for long-term synaptic plasticity and memory formation.

    Based on this evidence, we examined the effects of acute stress on cued and contextual fear memory dynamics and on long-term synaptic plasticity at the TA-CA1 synapses in the dorsal hippocampus and the role of AEA in modulating these effects.

    Methods: To test the effects of stress on fear memory dynamics, male adult Sprague Dawley rats were subjected to a forced swim stress 30 min before being tested for fear memory recall/extinction training in either an auditory or contextual fear conditioning paradigm (n = 11-12 per group). To test the effects of stress on synaptic plasticity, different groups of animals were subjected to the forced swim stress and anesthetized for in vivo extracellular field recordings at TA-CA1 synapses (n = 5-8 per group). To examine the role of AEA in the modulation of stress effects on fear memory and plasticity, separate cohorts of rats received either bilateral intra-CA1 injections of the AEA hydrolysis inhibitor URB597 30 min before the acute stress and tested for contextual fear memory dynamics, or systemic URB597 injections before the forced swim stress and then anesthetized for in vivo extracellular field recordings (n = 5-10 per group). All experimental procedures were in compliance with protocols approved by the University of Calgary Animal Care Committee and guidelines from the Canadian Council on Animal Care.

    Results: Our results show that acute stress given before fear memory recall/extinction training session did not affect memory dynamics in the cued fear conditioning paradigm, but strongly reduced contextual fear memory recall, as rats in the stress group showed a marked reduction of freezing behavior when re-exposed to the fear conditioning context 24 h after conditioning, as compared to their control group (p = 0.03). Moreover, the exposure to the forced swim stress robustly impaired the acquisition of contextual fear memory extinction, as shown by increased freezing levels in stressed rats during the extinction recall session as compared to the non-stressed control group (p = 0.01). Consistently with these behavioral results, in vivo extracellular field recordings also revealed that acute stress strongly impaired LTP at TA-CA1 synapses. Very interestingly, pharmacological inhibition of AEA hydrolysis reverted the detrimental effects of stress on both fear memory extinction and LTP.

    We are currently investigating possible mechanisms underlying such effects.

    Conclusions: Our results show that acute stress impairs the acquisition of contextual, but not cued, fear memory extinction and this might be likely mediated by the stress-induced suppression of long-term synaptic plasticity at TA-CA1 synapses in the dorsal hippocampus. Moreover, pharmacological inhibition of AEA degradation completely reverts these stress impairing effects.

    Taken together, our results shed light on the mechanisms underlying stress detrimental effects on memory and synaptic plasticity, opening the avenue to investigate new potential endocannabinoid-based tools to treat stress-related psychopathologies, such as Post-Traumatic Stress Disorders.

    Keywords: Endocannabinoids, Fear Extinction, Long Term Potentiation, Hippocampus, Acute Stress

    Disclosure: Nothing to disclose.

    M168. Differential Valuation and Learning From Social and Non-Social Cues in Borderline Personality Disorder

    Abstract not included.

    M169. Imaging the Endocannabinoid System in Borderline Personality Disorder: A [11 C]-CURB Positron Emission Tomography Study

    Nathan Kolla*, Shelley McMain, Pablo M. Rusjan, Sylvain Houle, Michael Bagby, Karolina Karas, Christine Wang, Romina Mizrahi, Isabelle Boileau

    University of Toronto, Toronto, Canada

    Background: Borderline personality disorder (BPD) is a poorly understood psychiatric condition with few biomarkers to guide development of novel treatments. Peripheral abnormalities of the endocannabinoid system have been described in BPD, but central measures have not yet been investigated. We used the novel radiotracer [11 C]-CURB and positron emission tomography (PET) to quantify levels of fatty acid amide hydrolase (FAAH) in multiple brain regions. FAAH is an enzyme that degrades anandamide (AEA), an endocannabinoid that binds to CB1 receptors with high affinity. We hypothesized that FAAH would be elevated in the prefrontal cortex (PFC) and amygdala of BPD relative to healthy controls.

    Methods: Twenty females with BPD and 20 healthy subjects were recruited. All of the BPD participants were medication-free. None had a comorbid major depressive episode. All were currently self-harming. None of the healthy controls had a psychiatric history. All study participants screened negative for illicit substance use and were non-smokers. Each participant underwent one [11 C]-CURB PET scan and a proton density magnetic resonance imaging scan for coregistration. Scans were analyzed as previously described (Kolla et al, 2016). Subjects were also genotyped for a common polymorphism that affects binding of FAAH. For the primary hypothesis, PET data were analyzed using multivariate analysis of variance (MANOVA) with PFC and amygdala FAAH levels as the dependent variables and diagnosis and genotype as fixed factors. Effects in each region, analyzed by univariate ANOVA, were considered significant after Bonferroni correction (p < 0.025).

    Results: MANOVA revealed that BPD individuals had greater FAAH binding in the PFC relative to controls (p = 0.019). There was no significant difference in amygdala FAAH levels between groups (p = 0.13). Among the BPD patients, there was a positive correlation between resentment scores on the Buss-Durkee Hostility Inventory and FAAH levels in the PFC (r = 0.53, p = 0.019).

    Conclusions: This study is the first to report an elevation of FAAH in any psychiatric condition. To the best of our knowledge, it is also the only PET study of BPD since 2016. Results indicate that the endocannabinoid system is altered in BPD. Elevated levels of FAAH could lead to lower AEA levels, resulting in lower CB1 receptor neurotransmission. Increased FAAH may play a role in specific BPD symptomatology, as PFC FAAH levels were positively correlated with a measure of anger. Our results suggest that FAAH inhibitors could be tested for their utility in treating BPD with high anger.

    Keywords: Borderline Personality Disorder, Positron Emission Tomography Imaging, Endocannabinoid System, Fatty Acid Amide Hydrolase, Anger

    Disclosure: Nothing to disclose.

    M170. Exploring Familial Relationships Between Borderline Personality Disorder and Autism and Asperger’s Spectrum Disorders: New Perspectives for Treatment

    Valerie Porr*, Nicolette Mallios, Isak Leon, Da Ae Ban, Fauve Arthur

    TARA for Borderline Personality Disorder Association, New York City, New York, United States

    Background: BPD has a lifetime prevalence rate of 5.9%. It is characterized by impairment in interpersonal functioning, poor empathy and problems with trust and intimacy. People with BPD have been found to be hyper-vigilant, rejection sensitive, negatively biased and self-referential. They have difficulties in both cognitive and emotional empathy. Autism, characterized by social and communication difficulties, narrow interests and repetitive behaviors, has an estimated prevalence of 1%, predominately in males. Very few studies have examined the relationship between these disorders. People with BPD and ASD have difficulties with interpersonal relationships and in understanding, regulating and responding to emotions their own and others. These disorders seem to overlap in difficulty processing facial expressions, high sensory sensitivity, and difficulty in mentalizing. Hofvander found a 68% overlap between ASD and personality disorder (42 out of 62) while Ryden found 6 out of 41 BPD subjects (14.6%) met criteria for ASD. Lugnegard found 26 out of 54 (48%) ASD subjects had at least one personality disorder. In a 2017 study of the Overlap Between ASD and BPD, Dudas (2017), found that people with BPD have as high levels of autistic traits as people with ASD. He suggests that some ASD in females with BPD is not easily detected, perhaps owing to their “apparent competence.” This study highlights the frequent misdiagnosis of these disorders. No study to date has looked at the incidence of BPD and ASD in family members nor on the cognitive and social behaviors within the two populations.

    Methods: TARA developed an online survey, Exploring the relationship between BPD, Autism Spectrum Disorder and Asperger’s Syndrome, to determine if there is a relationship between these two greatly misunderstood disorders. A literature review of Borderline Personality Disorder, Autism Spectrum Disorder and Asperger’s syndrome was conducted to gain an in depth understanding of the symptoms or traits characteristic of each disorder. The survey included questions pertaining to social cognition, cognitive deficits, sensory processing disorder, and learning disorders such as ADHD and Dyslexia. We sought to identify symptoms that overlap between all three disorders and those specific to each. We then developed a symptom scale expressed in layman’s terms regarding the following categories: social interactions, cognition, perception, experiencing emotions, and biologically based reactions. The survey was posted on the TARA4BPD website. The survey responses were predominantly from the BPD community. There were 322 responses to the survey.

    Results: The survey results found 130 people identified themselves as having BPD; 27% of them identified as having Autism themselves or in a family member (1 out of 4). Out of 322 (N) responders, 48 participants reported having Autism in self and family, representing 15% of the autism survey population. Out of 322 responders, 94 participants reported having Asperger’s in self and family, representing 29% of the ASD survey population. Of these 94 participants, 57 participants, reported having a relative with BPD or 61% of the ASD survey population. Of 205 people who had BPD within their family, 58% had a relative with Autism while 75% had a relative with ASD. Of 205 people with BPD family members, 16% had a child with Autism, 20% had a child with ASD; 12% had a niece/nephew with Autism, !2 % of their niece/nephews reported having ASD. In the case of siblings, 5% of BPD family had a sibling with autism and 10% had an ASD siblings. We also found interesting results for sensory processing disorder, dyslexia, ADHD, auditory disorder.

    Conclusions: The relationships between BPD and Autism Spectrum Disorders is a neglected area of research. TARA’s online survey was available to the public and therefore cannot be compared to a study done in a clinical setting with reliable clinical assessments. Despite these shortcomings, our preliminary findings seem to point to an apparent relationship between these disorders. Our extensive experience with family members of those with BPD has led us to believe that there are early signs of BPD in young children, such as emotional and sensory hypersensitivity, social interaction problems, concrete/rigid thinking, difficulty in facial recognition, early sleep dysregulation, and problems with empathy. These early signs are generally ignored or trivialized by most pediatricians and child psychiatrists. These symptoms overlap with ASD symptoms. As early intervention training with Autistic/Asperger children has been shown to be effective, early recognition of similar BPD symptoms might benefit from similar training.

    Keywords: Borderline Personality Disorder, Autism, Asperger’s Spectrum Disorder

    Disclosure: Nothing to disclose.

    M171. Effectiveness of an Emotional Working Memory Training in BPD

    Annegret Krause*, Julia-Caroline Walther, Susanne Schweizer, Stefanie Lis, Martin Bohus, Christian Schmahl

    Leiden University, Leiden, Netherlands

    Background: Borderline Personality Disorder (BPD) is a severe mental disorder, characterized by emotion dysregulation, interpersonal problems, and stress-related cognitive disturbances. While several specialized evidence-based treatments exist, waiting times are often long and affective symptoms persist after successful treatment completion. Online trainings targeted at specific BPD symptoms, such emotion dysreguation and interpersonal distubances, may be cost-effective add-on interventions for individuals waiting for treatment. Findings from previous research in healthy individuals suggests that an ‘Emotional Working Memory Training’ (EWM), developed by Schweizer and colleagues [1], may show beneficial effects on affective-cognitive control. This computerized training involves different dual-n-back-tasks with emotional distractors (pictures, words) and increasing difficulty (adaptive testing). So far, no study has tested the effectiveness of this EWM training in BPD: The purpose of this study was to evaluate its effects on working memory and emotion regulation capacities in BPD.

    Methods: In a randomized control trial, 60 female BPD patients were assigned to either an EWM training group (dual-n-back-task with emotional distractors, n=30) or a placebo training (feature match WM task, n=30). Training duration and success were recorded online. Before the training (T1) and after 26-day training (T2), participants performed two tasks measuring cognitive control of disturbing emotional material: 1) an adapted Sternberg working memory paradigm (with vs. without emotional pictures as distractors), and 2) an emotion regulation (cognitive re-appraisal) task. Behavioral performance and heart rate variability (HRV) were assessed as outcome measures before and after the training and evaluated within/between groups, using a repeated measure analysis of variance design.

    Results: After the training (T2), both training groups showed significant improvements in working memory during the Sternberg Task compared to T1 (F(1,46)=5.73, p=0.021, η²=.11). In the Emotion Regulation task, patients who had performed the EWM training showed significantly better down-regulation of negative emotions at T2 compared to T1, in comparison to the Placebo Training group (interaction effect: F(1,40)=11.33, p=0.002, η²²=0.16). In addition, significant correlations between training success and lower arousal ratings (r= -.585, p=0.003, R²=.34) as well as higher HRV (r= -.577, p=0.005, R²=.33) during the Emotion Regulation task were observed in the EWM training group (but not in the Placebo group). Results were similar for completers and intent-to-treat.

    Conclusions: These novel un-published findings provide preliminary evidence for a beneficial effect of an Emotional Working Memory training on emotion regulation capacities in BPD. Computerized trainings, such as the EWMt, may be a cost-effective add-on intervention, e.g., for patients who are on the waiting list for a treatment. More research is needed to evaluate whether these effects can be replicated in larger samples. Future studies may further improve the acceptance of the training in the EWM group. Further implications for the clinical setting and future research are discussed.

    Keywords: Borderline Personality Disorder, Working Memory, Emotional Regulation, Heart Rate Variability, Computerized Cognitive Training

    Disclosure: Nothing to disclose.


    1. Schweizer, S., Grahn, J., Hampshire, A., Mobbs, D. & Dalgleish, T. (2013). Training the emotional brain: improving affective control through emotional working memory training. Journal of Neuroscience, 33(12), 5301–5311. doi:10.1523/JNEUROSCI.2593-12.2013.

    M172. Cortical RNA Profiles of Adult Offspring of Prenatal Restraint Stressed (PRS) Mouse Dams Identifies Differentially Expressed Genes

    Dennis Grayson*, Erbo Dong, Eleonora Gatta, Vikram Saudagar, Jeff Dunn, Alessandro Guidotti

    University of Illinois at Chicago, Chicago, Illinois, United States

    Background: Prenatal or early life exposure to adverse events including stress, infection, malnutrition, hormones, and the use of illicit drugs, may have a profound effect on adult brain structure and function. Mice born from dams stressed during pregnancy (PRS mice), manifest various behavioral and biochemical endophenotypes when they reach adulthood. The behavioral endophenotypes include hyperactivity, stereotypic and compulsive behaviors, abnormal social interactions, prepulse inhibition, fear conditioning, object recognition, and hypersensitivity to NMDA receptor blockers.

    Methods: Pregnant dams were subjected to daily episodes of restraint stress (45 min, 3 times daily with bright light) from days 7 through birth. Control dams were left in their cages. Following birth, pups were kept with their moms and were allowed to nurse. After weaning, mice were group housed and left alone until early adulthood (90 days) at which time, animals were sacrificed and the prefrontal cortex (PFC) was dissected. RNA was isolated from the prefrontal cortices of 5 PRS and 5 non-stressed (NS) mice. RNA quality was determined using an Agilent Bioanalyzer and all samples had a RIN greater than 7.5.

    RNA-seq libraries were prepared using the TruSeq Stranded mRNA Library Prep Kit (Illumina) at the High-Throughput Sequencing Unit of the Roy J. Carver Biotechnology Center, University of Illinois. Quantity and size distribution of each library was observed using an Agilent 2100 Bioanalyzer using the High Sensitivity DNA kit. Pooled libraries were sequenced as single-end 100 nt reads on two lanes for 101 cycles using an Illumina HiSeq2500 platform using a TruSeq SBS sequencing kit. FastQ files were evaluated for read quality and trimmed. TopHat2 was used to align sequences to the mouse genome. Raw gene counts were then analyzed in R using STAR in edgeR. 5,084 genes without at least 1 count per million mapped reads in at least 3/6 samples were removed from analysis, leaving 13,936 genes. TMM normalization (20) was used to correct for any RNA composition bias and differential expression between N and S groups was tested using a replicate-paired design and a generalized linear model appropriate for count-based data. Principle Components Analysis showed a clustering of samples into two well separated groups. voom transformed values and weights were tested for differential expression using an empirical Bayes method. For gene ontology, gene symbols and fold change for all mRNAs with adjusted p< 0.05 were submitted to ConsensusPathDB, DAVID, and Gene Set Enrichment Analysis (GSEA).

    Results: There were a total of 1,619 differentially expressed (DE) genes in the FC of adult (PD80) PRS mice at an adjusted p value ≤ 0.05. Of these significantly different RNAs, 665 were increased and 955 were decreased relative to the NS samples. The volcano plot, shows the relationship between the statistical test p-values and the log10 (Fold Change) of the difference in expression between PRS samples and NS groups. Differentially Expressed (DE) genes (with adjusted p values < 0.05) were analyzed using gene ontology software. When these transcripts were input into DAVID, top annotation terms included ribosome, mitochondrial function, immune system, and response to stress. Many of these were predominantly up- or down-regulated. For example, of the 63 genes associated with ribosomal proteins, all were down-regulated in PRS mice relative to NS mice. Similarly, of the 125 genes associated with the DAVID term Mitochondrion, 111 mRNAs were decreased while only 13 were increased. The ribosomal protein mRNAs significantly different in PRS mice are shown, as are those mRNAs encoding proteins linked to mitochondrial function and oxidative phosphorylation. We’ve validated and replicated many genes associated with each functional gene category.

    Conclusions: Our data support the hypothesis that prenatal stress induces long-lasting changes in gene expression in the prefrontal cortex of mice. Many of the genes that exhibit altered expression profiles are likely associated with the observed behavioral anomalies. The reduced expression of genes associated with ribosomal proteins are likely responsible for the recently noted reduction in dendritic spines on prefrontal cortical neurons in PRS mice and may be associated with the spine loss observed in psychiatric disorders. Numerous studies have linked the gene ontology terms mitochondrial function and immune process with schizophrenia (SZ) and autism spectrum disorder (ASD). We looked for overlap between our DE genes (FDR<0.05) and genes associated with SZ, ASD, and bipolar disorder (BP). We found a pronounced overlap (73 genes) between our DE genes and genes associated with SZ. There were 20 DE genes that overlap with ASD associated genes and 10 genes that are also altered in BP. Collectively, offspring of PRS mice exhibit behavioral and molecular endophenotypes that are reminiscent of the endophenotypes associated with various psychiatric disorders.

    Keywords: Prenatal Stress, Mouse Model, Epigenetics, Schizophrenia

    Disclosure: Nothing to disclose.

    M173. Stimulus-Specific Adaptation and Deviance Detection in Primary Auditory Cortex of the Awake Non-Human Primate

    Tobias Teichert*

    University of Pittsburgh, Pittsburgh, Pennsylvania, United States

    Background: Mismatch negativity is the key biomarker of altered short-term memory function and/or predictive coding in schizophrenia. Mismatch negativity has been sub-divided into two components, stimulus-specific adaptation and deviance detection. Deviance detection, i.e., increased neural activity to stimuli that violate expectations, can facilitate the identification of informative events in the environment. Stimulus-specific adaptation is a complementary process that reduces responses to repeated stimuli that are less likely to be informative. While stimulus-specific adaptation can most likely be explained by short-term presynaptic plasticity, it remains unclear how the brain computes expectations and deviations from expectation required for deviance detection. Efforts to understand the underlying neural mechanisms of deviance detection have been hampered by the fact that it has not been observed in the non-human primate, an important model system.

    Methods: To better understand the phenomenology of mismatch negativity in the monkey we recorded neural responses in primary auditory cortex (A1) of two awake macaque monkeys while they listened to 50 ms long 60 dB SPL loud pure tone bursts presented in the context of a modified roving standard paradigm. The transition probabilities between different tones in the paradigm were manipulated to create a regular condition in which time and identity of upcoming tones was predictable and a random condition in which it was not.

    Results: Following standard conventions, deviance detection was operationalized as stronger responses for deviants in the regular compared to the random condition. Based on this definition both animals exhibited deviance detection in addition to stimulus-specific adaptation. This is the first report of deviance detection in this species. Deviance detection was found in supra-granular, granular and infra-granular layers. Deviance detection was present in the first wave of cortical activity (20-45 ms) and was even stronger in a later period from 70 to 125 ms. Deviance detection was observed independent of the stimulus-onset asynchrony over the entire range of values tested (250 to 8000 ms). We present a computational model that can explain key aspects of deviance detection using a simple yet effective local circuit architecture and biologically realistic synaptic function.

    Conclusions: In addition to stimulus-specific adaptation, primary auditory cortex of the macaque monkey also exhibits deviance detection. This finding suggests a closer homology between the two species than previously assumed, thus paving the way for additional studies to dissect the neural mechanism. The delayed timing of deviance detection relative to stimulus-specific adaptation suggests that it might be fed back from higher auditory regions.

    Keywords: Mismatch Negativity, Deviance Detection, Macaque Monkey, Auditory Short-Term Memory, Primary Auditory Cortex

    Disclosure: Nothing to disclose.

    M174. Identification of Functional Role of Medial Prefrontal Cortical Neurons Co-Expressing D1 and D2 Receptors

    Clementine Quintana, Jivan Khlghatyan, Jean-Martin Beaulieu*

    University of Toronto, Toronto, Canada

    Background: The D2 receptor (Drd2) is a direct or indirect target of antipsychotics and mood stabilizers. Medial prefrontal cortical (mPFC) functions of Drd2 are of interest, considering the involvement of cortical dopamine in cognitive functions and emotional processing in schizophrenia and mood disorders. However, low mRNA expression of Drd2, selectivity of antibody, drug and reporter systems has prevented the specific detection and study of Drd2 expressing neurons. To overcome previous technical limitations hindering reliable mapping of Drd2 neurons particularly in cortex we used a TRAP (Translational Ribosome Affinity Purification) approach using mice expressing a Cre activated RiboTag, specifically in Drd2 positive cells (Drd2+ cells). These mice allow for a highly sensitive detection of Drd2+ cells and provide the possibility to perform cell-specific translational profiling to identify the molecular determinants of Drd2+ cells in mPFC. Interestingly, this model allows for the detection of other dopamine receptors, such as Drd1, in mPFC Drd2 cells+.

    Methods: To investigate the impact of co-expression of Drd2 and Drd1 in mPFC neuronal function, we used somatic CRISPR/Cas9 mediated knockout and chemogenetic modulation revealing their involvement in the regulation of behaviors. Immunohistochemistry and translational profiling were used to identify the molecular determinants of Drd2+ cells in mPFC. To investigate behavioral regulation, we used CRISPR/Cas9 and DREADD approaches modulating directly Drd2+ cells activity in mPFC in adult mouse brain.

    Results: By extraction of mRNA binding HA tagged ribosome in mPFC Drd2 cells+ we observed the co-expression of various GPCRs in Drd2+ neurons. Interestingly, expression profiling revealed a high level of Drd1 mRNA expressed in Drd2 cells+ in cortical neurons. This was confirmed by an immunofluorescent analysis showing a high density of Drd2 cells+ that co-express D1 receptor in cortical layer II and V. Ultimately, using different approaches, we modulated directly the protein synthesis or the activity of these neurons to reveal their involvement in behavioral regulation.

    Conclusions: This comprehensive analysis of mPFC Drd2 neurons provides indications for its functional implications in healthy and disease conditions. Heterogeneity of dopamine receptor expression in mPFC has to be taken into consideration during pharmacological intervention and assessment of functional and behavioral data.

    Keywords: D1 Dopamine Receptors, Dopamine 2 Receptor, Medial Prefrontal Cortex

    Disclosure: Nothing to disclose.

    M175. Enhanced NGR/p75/KAL9 Signaling Influences Dendritic Morphogenesis in a Schizophrenia-Relevant Manner

    Melanie Grubisha*, Zachary Wills, Susan Erickson, Cassandra Helmer, Gregg Homanics, Peter Penzes, Robert Sweet

    University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, United States

    Background: Kalirin (KAL) is a Rho GEF that is highly involved in regulation of cytoskeletal morphology within dendrites. The KAL9 isoform activates RhoA downstream of p75 when p75 interacts with Nogo receptor (NGR). This NGR/p75/KAL9 complex acts to restrict dendritic morphogenesis. Increased levels of Nogo mRNA as well as elevated levels of KAL9 protein have been described in schizophrenia, suggesting enhanced activity of this pathway may contribute to the impairments in dendritic morphogenesis in disease. We evaluated a missense mutation in KALRN (KALRN-PT), located near the RhoA GEF in KAL9, and found that it acts as a gain of function mutation for RhoA activation. We hypothesized that enhanced activity in the NGR/p75/KAL9 pathway impairs dendritic morphogenesis in pyramidal cells (PCs) across development.

    Methods: RhoA activation assays were performed using transient expression of a RhoA sensor in in vitro dissociated cortical cultures overexpressing either KALRN-WT or KALRN-PT. CRISPR/Cas9 gene editing was used to insert the human KALRN-PT mutation at the endogenous locus of the C57/Bl6J strain. Golgi staining was performed on cortical sections from 4 and 12-week old mice encompassing primary auditory cortex (A1), and full dendritic reconstructions of A1 Layer 3 PCs were performed in NeuroLucida software. Spine density analyses and cortical volume measurements were performed on Golgi stained material using StereoInvestigator software.

    Results: KALRN-PT confers enhanced RhoA activation compared to KALRN-WT. L3 PCs from A1 in homozygous KALRN-PT mice demonstrate reduced dendritic length and complexity at 12-weeks, but not at 4-weeks. There is no observed change in spine density along secondary apical dendrites between 12-week old KALRN-WT and KALRN-PT mice. Similarly, there is no significant change in cortical volume between genotypes at 12-weeks, although KALRN-PT mice showed a trend towards a 5-6% volume reduction (p=0.24).

    Conclusions: The increased RhoA activity arising from the PT mutation results in increased NGR/p75/KAL9 signaling and subsequently leads to reduced dendritic length and complexity in L3 PCs in A1 in early adulthood. Interestingly, this change is not present during the pre-adolescent period and presumably emerges during adolescence, consistent with the timing of onset of clinical symptoms of schizophrenia in humans. This change in dendritic structure is not accompanied by any statistically significant change in spine density or cortical volume, suggesting that KALRN-PT provides a useful model for isolating mechanisms of impaired dendritic morphogenesis as they relate to schizophrenia.

    Keywords: Auditory Deficits in Schizophrenia, NGR/p75, Dendritic Morphogenesis

    Disclosure: Nothing to disclose.

    M176. Increased Protein Insolubility in Brains From a Subset of Patients with Schizophrenia

    Leslie Nucifora, Matthew MacDonald, Brian Lee, Matthew Peters, Alexis Norris, Benjamin Orsburn, Kun Yang, Kelly Gleason, Russell Margolis, Jonathan Pevsner, Carol Tamminga, Robert Sweet, Christopher Ross, Akira Sawa, Frederick Nucifora*

    Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

    Background: The mechanisms leading to schizophrenia are likely to be diverse. However, there may be common pathophysiological pathways for subsets of the disease. In the present study, we hypothesized that disruption of protein quality control can lead to protein insolubility for a subset of patients with schizophrenia.

    Methods: Prefrontal cortex or superior temporal gyrus from autopsy brains provided by the University of Pittsburgh, University of Texas Southwestern, and Harvard were subjected to cold sarkosyl fractionation, separating proteins into soluble and insoluble fractions. All pellet samples were analyzed to quantify insoluble protein levels and ubiquitin reactivity, normalized to total homogenate protein. We then performed mass spectrometry analysis to identify the contents of the insoluble pellets. The potential biological relevance of the detected proteins was assessed using Gene Ontology Enrichment Analysis and Ingenuity Pathway Analysis.

    Results: A subset of patients with schizophrenia showed an increase in protein insolubility and ubiquitination in the insoluble protein fraction. Mass spectrometry of the insoluble fraction revealed that cases with increased insolubility and ubiquitination showed a similar pattern of peptide clustering by principal component analysis plot and heatmap analysis. The proteins that were significantly altered in the insoluble pellet were enriched for terms relating to axon target recognition as well as nervous system development and function.

    Conclusions: This study suggests a pathological process related to protein insolubility for a subset of patients with schizophrenia. Understanding the molecular mechanism of this subtype of schizophrenia could lead to a better understanding of the pathways, circuitry, and symptoms seen in some patients with major mental illness and could lead to improved nosology and novel therapeutic targets.

    Keywords: Protein Insolubility, Ubiquitination, Schizophrenia

    Disclosure: Nothing to disclose.

    M177. Laminar- and Cell Type-Specific Transcriptomic Analysis of Electron Transport Chain Complexes in the Prefrontal Cortex of Schizophrenia Subjects

    Jill Glausier*, John Enwright, Dominique Arion, David Lewis

    University of Pittsburgh, Pittsburgh, Pennsylvania, United States

    Background: Accumulating evidence indicates that a key mitochondrial function, ATP synthesis via oxidative phosphorylation (OXPHOS), is lower in the prefrontal cortex (PFC) of schizophrenia subjects. Pathway analyses of transcriptome data from PFC layers 3 and 5 pyramidal cells, and layer 3 parvalbumin (PV) interneurons, indicate electron transport chain (ETC) and OXPHOS as top affected gene pathways. Despite the abundance of data indicating less ATP production via OXPHOS, the nature of these alterations in the disease is unknown. Two distinct mechanisms can affect OXPHOS: cellular demand for ATP and dysfunction of the ETC complexes. The first mechanism regulates the level of ATP produced by OXPHOS in mitochondria in response to cellular energetic demands, whereas the second mechanism impairs OXPHOS production of ATP in mitochondria. Determining which of these mechanisms, lower energetic demand or impaired energetic capacity, contributes to disease findings is paramount since each possibility carries very different implications for therapeutic targets.

    Lower cellular ATP demand and impaired OXPHOS have different effects on the expression levels of ETC complexes. Synthesis of ATP via OXPHOS only occurs upon demand, and the most energetically-demanding processes involve support of neuronal action potentials and synaptic signaling. Accordingly, persistent decreases in neuronal firing lowers ATP demand, eliciting a correlated reduction in the expression of ETC complexes in these cells. Thus, if ATP demand is lower in PFC layers 3 and 5 pyramidal cells and layer 3 PV interneurons, then we would expect to find lower expression of ETC complexes of similar magnitude within each cell population. Impaired OXPHOS, on the other hand, has an unequal effect on the expression of ETC complexes, and preferentially affects cells with high firing and energy demands. For example, PV interneurons have substantially higher firing rates and ATP demand than pyramidal cells. Accordingly, perturbations of OXPHOS enzyme function, including induction of oxidative stress by a variety of experimental manipulations, preferentially affects PV cells. Thus, if OXPHOS is dysfunctional, then we would expect to find a larger reduction in the levels of ETC complexes in PV relative to pyramidal cells and non-correlated changes in the expression of ETC complex expression in each cell population. To distinguish between these alternatives in schizophrenia, we compared transcript expression levels of each ETC complex in layers 3 and 5 pyramidal cells and layer 3 PV interneurons.

    Methods: We used microarray data from previous reports of gene pathways altered in schizophrenia. Briefly, frozen tissue sections containing PFC area 9 from 36 matched pairs of schizophrenia and unaffected comparison subjects were stained with thionin for Nissl substance to identify pyramidal cells or labeled using immunoperoxidase for aggrecan to identify PV cells. Pyramidal and PV somata were dissected using a Leica laser microdissection (LMD) system. Transcriptome profiling was performed by microarray using Affymetrix GeneChips specific to the human genome. A Z-score was calculated for each subunit comprising an ETC complex (Complex I: 37 subunits, Complex II: 4 subunits, Complex III: 9 subunits, Complex IV: 10 subunits, Complex V: 16 subunits), then all subunit Z-scores within a complex were averaged to obtain a composite Z-score for each subject.

    Results: Expression levels of all five ETC complexes were 14-21% lower in layer 3 pyramidal cells, 17-22% lower in layer 5 pyramidal cells, and 4-27% lower in layer 3 PV cells of schizophrenia subjects relative to unaffected comparison subjects. The expression levels among ETC complexes within all three cell populations was highly correlated in unaffected comparison subjects (layer 3 pyramidal cells: mean r=0.9; layer 3 PV cells: mean r=0.8; layer 5 pyramidal cells, mean r=0.7), and these correlations were completely retained in all three cell populations in subjects with schizophrenia.

    Conclusions: Our laminar- and cell type-specific transcriptomic analyses indicate that expression of all five ETC complexes is not more affected in PV cells relative to pyramidal cells, and that correlated expression across ETC complexes is preserved in the disease state. Together, these data suggest that lower measures of OXPHOS likely reflect lower ATP demand, and not impaired OXPHOS capacity, in the PFC of subjects with schizophrenia.

    Keywords: Postmortem Brain Tissue, Parvalbumin Neurons, Oxidative Phosphorylation, Schizophrenia, Pyramidal Cell

    Disclosure: Nothing to disclose.

    M178. Comprehensive Analysis of Regional Specific RNA-Editing Differences in Brain Across the Lifespan of Normal Subjects and Individuals With Schizophrenia

    HyeonJin Cho*, Tae Young Hwang, Carrie Wright, Jae-Kyung Won, Emily Burke, Andrew Jaffe, Thomas Hyde, Joel Kleinman, Daniel R. Weinberger, Joo Heon Shin

    Lieber Institute for Brain Development, Baltimore, Maryland, United States

    Background: RNA editing is a co/post-transcriptional process that alters the nucleotide sequence of primary transcripts. ADAR (adenosine deaminases acting on RNA) enzymes catalyze A-to-I RNA editing by deamination of adenosine, converting stable A:U base pairs to less stable I:U mismatches. These RNA sequence alterations also often influence structure as well as recoding of amino acids. Especially in the mammalian nervous system, it has been shown that RNA editing contributes to the function of diverse neuronal receptors and immune-homeostasis, suggesting that abnormalities in RNA editing are involved in several pathological conditions including psychiatric diseases. Indeed, previous studies have shown multiple links between RNA editing and certain human brain disorders, including schizophrenia, mood disorder, autism and Alzheimer’s disease. However, these studies have been limited by small sample sizes and have largely only evaluated a limited number of editing sites. To better understand the relationship between schizophrenia and A-to-I RNA editing in an unbiased way, we compared RNA editing profiles of a large number of postmortem neurotypical brain samples and samples with schizophrenia using RNA-seq data of several brain regions.

    Methods: We used RNA-seq data from 254 control and 184 schizophrenia from dorsolateral prefrontal cortex (DLPFC), 319 control and 133 schizophrenia from hippocampus (HIPPO), which were generated by BrainSeq Consortium and also from the dentate gyrus (DG) of the hippocampal formation isolated by laser capture microdissection from 93 control and 75 individuals with schizophrenia. We profiled 4,282 editing sites previously identified in DLPFC human postmortem brain data set that had high median numbers of sequencing reads in our current sample and were associated with genes. We calculated editing ratios for each site by dividing the number of alternate sequence counts with the total number of counts (both reference matching and alternate) that were filtered with multiple quality controls in the RNA-editing pipeline.

    For each brain region, a linear regression analysis was performed to compare editing ratios of each site between the diagnostic groups using the empirical Bayes method of the Bioconductor Limma package in R. RNA-seq quality metrics, as well as age and gender were included as covariates. In addition, multi-dimensional scaling (MDS) components determined from genotype data were included in the models for each region to account for the ethnic diversity of our samples. Significance values were corrected for multiple testing using the Benjamini Hocheberg method and a threshold of 0.05 was used to determine significance.

    Results: No editing sites had significantly different ratios between the diagnostic groups for the DG and the HIPPO; however, 17 sites were significantly different for the DLPFC, one of which was identified as a candidate risk gene in the current GWAS in PGC2 loci about its relationship with schizophrenia.

    Conclusions: Previous studies suggest a link between editing rates of glutamate receptor subunits and serotonin 2 C receptor subtype in schizophrenia, but these studies have had limited sample sizes and have not performed global assessments of other editing sites. Furthermore, previous work has often not accounted for possible effects of age, race, gender and RNA quality on ADAR editing rates. We plan to perform targeted re-sequencing for validation of the discovered sites to confirm the differences in ADAR editing. Further research is necessary to better understand how the differences in editing rates between the diagnostic groups of these editing sites may influence the function of these genes and their potential role in pathogenesis of schizophrenia.

    Keywords: A-to-I RNA-editing, Schizophrenia, Adenosine Deaminases Acting on RNA

    Disclosure: Nothing to disclose.

    M179. A Missense Mutation in SLC39A8, a Manganese Transporter Linked to Schizophrenia, is Associated With Specific Changes in Plasma N-Glycosylation

    Robert Mealer*, Richard Cummings, Mark Daly, Sylvain Lehoux, Jordan Smoller

    Harvard Medical School, Massachusetts General Hospital, Boston, Massachusetts, United States

    Background: Recent Genome Wide Association Studies (GWAS) identified a strong association between the disorder and a missense mutation (A391T) in the zinc/manganese transporter SLC39A8 (p < 8×10-15). Manganese (Mn) is a critical cofactor for many enzymes including certain glycosyltransferases; enzymes that attach sugar molecules to proteins and lipids to regulate their function. Patients with severe mutations in SLC39A8 (not A391T) present with symptoms of a type II congenital disorder of glycosylation, and treatment with Mn or precursors of galactosylation reverses deficient transferrin glycosylation and some clinical phenotypes. The purpose of our study is to determine the link between the SLC39A8 A391T missense mutation linked to schizophrenia and the glycosylation pathway.

    Methods: Using the Partners Biobank, a biorepository of 80,000 samples linked to the electronic medical record and genomic data, we optimized a search algorithm with 65% accuracy to identify patients with schizophrenia. After confirming the diagnosis via chart review, controls and patients with schizophrenia were stratified into groups based on SLC39A8 genotype, and samples of serum and plasma were obtained from the Biobank for study. Levels of Mn and Zn were measured in serum using Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Plasma N-glycans were release by PNGaseF cleavage, permethylated, and purified prior to analysis using MALDI-TOF.

    Results: We confirm that the A391T mutation results in decreased serum manganese with no effect on zinc. Analysis of plasma N-glycans shows a reduction in high molecular weight structures by ~25% (p<0.05) in A391T mutation carriers, suggesting reduced activity the Mn-dependent enzyme β(1,4)-galactosyltransferase. Analysis of plasma N-glycans from patients with schizophrenia independent of SLC39A8 genotype identified a 20% increase in fucosylation and 30% reduction of sialylation relative to controls, though these modifications can also be affected by inflammation, environmental exposures, and medications that may be present in the disease group. Investigation of a subset of individuals with schizophrenia and the A391T mutation are ongoing.

    Conclusions: We demonstrate that plasma N-glycosylation is altered in individuals harboring the SLC39A8 A391T mutation and hypothesize that similar changes in the developing brain result in the increased risk of developing schizophrenia in mutation carriers. We also find changes in the plasma N-glycans of individuals with schizophrenia, though the significance of these differences are yet to be determined. We ultimately hope to develop a scalable blood biomarker of risk based on genetics, Mn levels, and glycosylation patterns, in addition to generating novel insights in the neurodevelopmental changes that underlie schizophrenia.

    Keywords: N-Glycosylation, Schizophrenia Genetics, SLC39A8

    Disclosure: Nothing to disclose.

    M180. Inhibition of Brain and Liver Kynurenine Aminotransferase II Activity by N-Acetylcysteine in Rodent, Pig and Human

    Korrapati Sathyasaikumar, Tonali Blanco-Ayala, Robert Schwarcz*

    Maryland Psychiatric Research Center, Baltimore, Maryland, United States

    Background: Kynurenic acid (KYNA), a metabolite of the kynurenine pathway of tryptophan degradation, is increasingly understood to play an important role in the mechanism(s) underlying normal and abnormal cognitive processes, most likely by acting as an antagonist of α7 nicotinic and NMDA receptors. Specifically, elevated KYNA levels are detrimental in psychiatric diseases such as schizophrenia. KYNA is synthesized from its immediate precursor kynurenine - either by non-enzymatic oxidation or through irreversible enzymatic transamination by kynurenine aminotransferases. In the mammalian brain, kynurenine aminotransferase II (KAT II) is the principal enzyme responsible for the neosynthesis of rapidly mobilizable KYNA. KAT II therefore constitutes an attractive target for pro-cognitive interventions (Schwarcz et al., 2012). N-acetylcysteine (NAC), a brain-penetrant drug with pro-cognitive efficacy in humans, including individuals with schizophrenia, has been proposed to exert its actions by increasing the levels of the endogenous anti-oxidant glutathione (GSH) in the brain (Steullet et al., 2016). We now examined a possible alternative mechanism of NAC action, namely KAT II inhibition.

    Methods: Using a well-established procedure (Sathyasaikumar et al., 2011), we first tested the effect of NAC on KAT II activity in liver and brain tissue homogenates from mice, rats, pigs and humans in vitro. Next, we investigated the ability of NAC to affect the activity of pure human recombinant KAT II protein. Finally, we examined the effect of NAC on the neosynthesis of KYNA in vivo using microdialysis in the medial prefrontal cortex (mPFC) of unanesthetized adult rats. To this end, NAC (20 mM) was locally applied by reverse dialysis for a period of 6 h. 120 min after starting the perfusion, kynurenine was administered systemically (50 mg/kg, i.p.) while NAC perfusion continued for the remaining 4 h. In separate animals, NAC (500 mg/kg, i.p.) was administered systemically 120 and 60 min before the peripheral administration of kynurenine.

    Results: In all tissue homogenates, NAC inhibited KAT II activity with IC50 values in the high micromolar to the low millimolar range. Using human recombinant KAT II, NAC inhibited enzyme activity with an IC50 of ~500 μM, while GSH was approximately 40 times less potent (IC50 >20 mM). In vivo, NAC reduced the de novo production of KYNA from its immediate precursor kynurenine by ~45% and ~50%, respectively, in the two experimental paradigms used.

    Conclusions: Our results raise the possibility that NAC exerts its neurobiological effects at least in part by reducing cerebral KYNA levels via KAT II inhibition.

    Keywords: Cognition, Kynurenine, Kynurenic Acid, Schizophrenia, N-acetylcysteine

    Disclosure: Nothing to disclose.

    M181. Interaction Between Anticholinergic Medication Burden and Treatment Effects of Targeted Cognitive Training in Schizophrenia Patients Mandated to Long-Term Locked Inpatient Care

    Yash Joshi*, Michael Thomas, Andrew Bismark, William Hochberger, Emily Treichler, Melissa Tarasenko, John Nungary, Laura Cardoso, Alexa Sison, Juan Molina, Neal Swerdlow, Gregory Light

    University of California, San Diego School of Medicine, San Diego, California, United States

    Background: Targeted cognitive training (TCT) is an emerging computerized intervention aimed at remediating auditory information processing deficits in schizophrenia (SZ). TCT increases the fidelity of low-level auditory discrimination in SZ through adaptive and intensive exercises in a durable way and is associated with improved psychosocial outcomes. The underlying brain circuitry which mediates TCT-gains is thought to depend on appropriate brain cholinergic signaling. While patients with SZ are known to be vulnerable to elevated anticholinergic load stemming from psychotropic medications, the relationship between brain cholinergic signaling in the context of TCT trials conducted in a treatment-refractory SZ population has not been well characterized. Here we report findings from a randomized clinical trial investigating the effectiveness of TCT in SZ patients mandated to long-term locked inpatient care and the interaction of anticholinergic burden on treatment effects.

    Methods: Patients with SZ were randomized to treatment as usual (TAU) or treatment as usual + TCT. No significant differences were present in demographic, clinical or cognitive variables at baseline. The Anticholinergic Cognitive Burden Scale (ACCBS) was used to calculate anticholinergic burden using medication administration records. Auditory discriminability, cognitive functioning and symptoms scores were assessed at baseline and after a full course of TCT. Data were analyzed using linear mixed effects models. No group differences in ACCBS scores were present at baseline or follow up. Both groups had similar total antipsychotic load.

    Results: TCT improved auditory discriminability (d=0.6), verbal learning and memory (d=0.8), and positive symptoms (d=-0.6). Baseline ACCBS score negatively correlated with change in verbal learning (TAU, r=-.61, p<0.02; TCT, r =0.14, NS) and auditory discriminability (TAU, r=-.53, p<0.05; TCT, r = 0.41, NS) in TAU but not TCT group.

    Conclusions: TCT improved auditory discrimination, cognitive functioning and symptoms in SZ patients in locked residential care. Anticholinergic burden negatively correlated with measures of auditory discrimination and cognition in the TAU group, but not in those receiving TCT. These data suggest that patients with SZ who experience a high anticholinergic burden can benefit from TCT. These results also imply that TCT may protect against anticholinergic-burden associated impairment in SZ. Future studies are needed to identify strategies that optimize brain cholinergic signaling in SZ to better support or augment cognitive training in SZ patients.

    Keywords: Computerized Cognitive Training, Cholinergic System, Human Clinical Trial, Auditory Deficits in Schizophrenia, Psychotropic Medications

    Disclosure: Nothing to disclose.

    M182. Influence of Olanzapine Dosing and History of Cannabis Use on Treatment Response in Acute Schizophrenia

    Dagmar Koethe*, David Meyrath, Carola Schaefer, Cathrin Rohleder, F. Markus Leweke

    Brain and Mind Centre, The University of Sydney, Camperdown, Australia

    Background: In acute schizophrenia higher dosages of antipsychotics are often administered early and previous cannabis use is most common. The influence of the early antipsychotic dosage and/or cannabis use in response to treatment are not well investigated. Here, we performed a clinical trial starting with a low dose of olanzapine with controlled increase where needed. Thereby, we studied the impact of olanzapine dosage and previous cannabis use on clinical response and plasma olanzapine.

    Methods: To address these questions, we analyzed the data from a 6-week clinical trial in 37 acute schizophrenia patients. Olanzapine was stepwise increased by 2.5 mg/day weekly from 5 mg/day up to 15 mg/day based on an improvement of symptoms to the previous week (BPRS improvement >2, no increase). 18/37 patients received additional oxcarbazepine (1,500 mg/day) with no significant effect on plasma levels or improvement of symptoms. The history of cannabis use and psychopathology (PANSS) were assessed. Previous cannabis use was determined by self-report (no (n=13; NC), moderate (n=6; MC), and high (n=16; HC) cannabis use.

    Results: Patients were in average 29.5 ys, 23/14 male/female, 62% smokers. Average decrease in tPANSS to baseline was 25.4±23.0 (MV±SD). The mean maximum dose of olanzapine during the trial was 9.85±3.3 mg/day. 24/37 patients received a maximum dose of ≤10 mg/day of olanzapine, 7/37 remained on the initial dose of 5 mg/day. Cannabis history did not influence response, but HC use was associated with lower olanzapine plasma levels.

    Conclusions: In this design, 65% of patients benefited from low dose olanzapine. The effects of pharmacological treatment were independent of previous cannabis use although the latter influenced olanzapine plasma levels significantly. Our results suggest that initial low dose treatment in acute schizophrenia is justified and does not lead to a less favorable outcome seems to justify the low-dose start of olanzapine in acute schizophrenia where possible.

    Keywords: Cannabis Use, Psychosis, Treatment Response, Olanzapine, Clinical High-Risk State For Psychosis

    Disclosure: Nothing to disclose.

    M183. KarXT, a Combination of the M1/M4 Cholinergic Receptor Agonist Xanomeline and Trospium for the Treatment of Psychosis and Cognitive Impairment in Schizophrenia: Phase I Studies

    Stephen Brannan*, Andrew Miller, Steven Paul, Alan Breier

    Karuna Pharmaceuticals, Boston, Massachusetts, United States

    Background: Xanomeline is a M1 /M4 preferring muscarinic agonist that has demonstrated antipsychotic and pro-cognitive effects in both schizophrenia (Shekar, et al, 2001) and Alzheimer’s disease (Bodick et al, 2007). However, its peripheral cholinergic side effects, including nausea, vomiting and diarrhea, have prevented its continued clinical development. A previous Phase I clinical trial demonstrated that the addition of trospium, a peripheral cholinergic receptor antagonist that does not cross the blood brain barrier, to xanomeline substantially improved its tolerability by reducing peripheral cholinergic side effects (nausea, vomiting, diarrhea, excess sweating and salivation). We now report the results of a Phase 1, multi-dose safety study aimed at optimizing the combination of xanomeline with trospium using a new BID co-formulation.

    Methods: 69 healthy volunteers participated in the phase 1 multiple ascending dose (MAD) study of KarXT. Objectives included: assessing the PK characteristics of the new proprietary BID co-formulation (and comparing results to the two components given separately) and to explore the tolerability of higher doses of both xanomeline and trospium respectively. The study design was comprised of a 2-day titration period of either placebo or a KarXT dose of 50 mg xanomeline + 20 mg trospium followed by a 5-day treatment period. The doses (all BID) assessed were: xanomeline 100 mg, 125 mg and 150 mg in combination with trospium 20 mg or 40 mg. For comparing cohorts, tolerability assessment was focused on the pre-specified cholinergic adverse events of nausea, vomiting, diarrhea, excessive sweating and salivation; Safety and tolerability were also assessed using spontaneous reports of adverse events, labs, vital signs, Bristol stool scale, pupillometry and saliva volume.

    Results: The 2-day titration of 50/20 was well tolerated in all cohorts. Doses of 100 and 125 BID of xanomeline were also well tolerated when paired with 20 mg and 40 mg BID of trospium, respectively. Across the cohorts, cholinergic adverse events were correlated with xanomeline dose. Increasing trospium dose ameliorated cholinergic events, and lead to the observance of some anticholinergic adverse events as cohorts tested on 40 mg trospium BID reported some signs of anticholinergic effects (i.e., dry mouth), particularly in the cohort receiving 125 mg BID of xanomeline. The total number of subjects reporting cholinergic adverse events (ChAEs) in cohorts on xanomeline of 100 mg (39%) or 125 mg (33%) was similar to that seen in the KarXT arms in the previous KAR-001 study (34%). Most ChAEs occurred within the first few days of starting or increasing the study drug. The majority of these TEAEs at 100 mg and 125 mg xanomeline-dose levels were mild and transient in nature, with only a few adverse event reports lasting more than 3 h in the 100/20 mg cohort.

    None of the cohorts showed meaningful changes in orthostatic heart rate or obvious differences in BP between placebo and KarXT compared to placebo. All cohorts receiving KarXT showed placebo-adjusted increases in mean resting HR in subjects, consistent with past studies with xanomeline where short-term increases in resting heart rate were observed that normalized to baseline over time. Both trospium and xanomeline exposures (AUCs) and variability were comparable to KAR-001 where the compounds were given using separate formulations.

    Conclusions: The new KarXT co-formulation performed well in humans and will be carried forward in future studies with KarXT. A range of combination doses was identified for further investigation and xanomeline-doses equivalent to, and higher, that what demonstrated efficacy in previous xanomeline clinical studies will be carried forward into a planned Phase II study in schizophrenia. Longer terms studies will provide further data around the safety and tolerability of KarXT. Importantly, the tolerability observed in this healthy volunteer study may not be representative of schizophrenia patients, who tolerate currently marketed antipsychotic medicine better than healthy volunteers. No new safety signals were reported in the present study. The timing and duration of AEs were related to peak drug levels (Cmax) and suggest that there is a potential for increased tolerability over time. In addition, all AEs rapidly ameliorated to baseline levels upon dosing discontinuation. Consistent with our previous studies, KarXT is substantially better tolerated than xanomeline alone.

    Keywords: Short-Term Clinical Studies, Schizophrenia Novel Treatment, Muscarinic Acetylcholine Receptor, Advantages of The Combination Of Therapies

    Disclosure: Karuna Pharmaceuticals, Employee

    M184. Maintenance of Long-Term Clinical Stability in Patients With Schizophrenia Converted From Oral Antipsychotic Medications to Monthly Extended-Release Risperidone (RBP-7000)

    James Graham*, John Csernansky, Maurizio Fava, John Newcomer, Anne Andorn, Sunita Shinde, Gilbert Muma, Christian Heidbreder

    Indivior Inc., Richmond, Virginia, United States

    Background: RBP-7000 (risperidone in ATRIGEL® Delivery System) is a once-monthly, extended-release depot formulation recently approved for the treatment of schizophrenia in adults. RBP-7000 delivers clinically relevant plasma levels of risperidone on the first day of use without requiring loading or supplemental dosing. The efficacy, safety and tolerability of RBP-7000 were established in an 8-week, double-blind, placebo-controlled trial (NCT02109562) conducted in adults with acute exacerbation of schizophrenia. Maintenance of clinical stability was examined in the open-label study reported here.

    Methods: This 52-week, Phase III, open-label study (NCT02203838) included completers (“rollovers”) from the previously conducted 8-week double-blind study, as well as de novo patients with a DSM-IV-TR diagnosis of schizophrenia and a Positive and Negative Syndrome Scale (PANSS) total score ≤70. Rollover patients—those who had received 2 injections of placebo or RBP-7000 (90 mg or 120 mg) in the double-blind study—received 11 additional injections of RBP-7000 120 mg in the current study. De novo patients, who were required to be on oral antipsychotic treatment prior to enrollment, received up to 13 injections of RBP-7000 120 mg after titration or conversion to oral risperidone (3 or 4 mg/day). Mean changes from Baseline to end of study in PANSS total score, PANSS Positive, Negative, and General Pathology subscale scores, and Clinical Global Impression-Severity of Illness (CGI-S) score were summarized descriptively. No statistical testing was performed for this open-label study. As the primary objective of this study was long-term safety (results reported elsewhere), all analyses of clinical outcome data (secondary objective) were conducted using the safety population, defined as all those who received ≥1 dose of open-label RBP-7000 120 mg.

    Results: Of 500 patients in the safety population (rollover, n=92; de novo, n=408), a total of 234 (46.8%) completed the study. Most patients were male (67.8%) and black/African-American (70.8%), with a mean age of 45.1 years. De novo patients were stable at enrollment [mean (SD) PANSS total score = 58 (8.3)] and remained stable throughout the study [mean (SD) change from Baseline to end of study = -0.4 (8.7)]. Rollover patients continued to show improvement in PANSS total score for the duration of the trial (baseline to end of treatment): placebo, -20.2 (15.6); RBP-7000 90 mg, -12.5 (15.5); RBP-7000 120 mg, -10.9 (13.2). PANSS subscale scores generally remained stable throughout the study among de novo patients, with mean (SD) changes from Baseline of -1.3 (3.3) for PANSS Positive Scale, 1.3 (3.6) for PANSS Negative Scale and -0.4 (5.0) for PANSS General Psychopathology Scale. CGI-S scores remained stable throughout the study for both rollover and de novo patients. Decreases in each PANSS subscale scores were observed in each rollover group, including those who had received placebo in the double-blind study: mean (SD) changes from Baseline on the PANSS Positive Scale score were -7.8 (5.5), -3.7 (4.1) and -3.7 (3.5) among those previously treated with placebo, RBP-7000 90 mg or RBP-7000 120 mg, respectively. For the PANSS Negative Scale score, mean (SD) changes from Baseline were -4.0 (5.9), -4.1 (4.6) and -0.9 (3.8) for placebo, RBP-7000 90 mg and RBP-7000 120 mg, respectively. For the PANSS General Psychopathology Scale score, mean (SD) changes from Baseline were -8.3 (7.0), -4.7 (9.0) and -6.4 (7.9) for those previously treated with placebo, RBP-7000 90 mg or RBP-7000 120 mg, respectively.

    Conclusions: In de novo patients, efficacy measures remained stable over 12 months of treatment once treatment with RBP-7000 120 mg was initiated. Additionally, rollover patients from the double-blind study continued to improve over the course of this long-term open-label study. While pre-study antipsychotic medications and dose equivalents varied among the de novo patients included in this study, the results strongly suggest that RBP-7000 120 mg is associated with maintenance of clinical stability.

    Keywords: Antipsychotic, Schizophrenia, Extended-Release Depot, Clinical Stability, Positive and Negative Syndrome Scale

    Disclosure: Indivior, Employee

    M185. Effects of Roluperidone on Two Dimension of the Negative Symptoms Factor Score: Reduced Emotional Experience and Expression

    Philip Harvey*, Jay Saoud, Michael Davidson, Remy Luthringer

    Miller School of Medicine, University of Miami, Miami, Florida, United States

    Background: Everyday functioning is impaired in people with schizophrenia, with these impairments spanning domains of vocational, social, and everyday activities. Recent research has suggested that negative symptoms (NS) can be considered in terms of two different dimensions: diminished expression (expressive deficit) and reduced experience (experiential deficit). Expression includes displays of facial affect (referred to as blunted affect), reduced vocal inflection, and reduced vocal output. Experience includes the motivation to engage in potentially pleasurable activities and the subjective experience of enjoyment when engaging in reinforcing activities. Reduced experience has also been referred to as reflecting avolition and apathy, as a descriptor of the observable consequences of deficits in experience.

    Roluperidone, a compound with high affinities for 5 HT2A and sigma2 receptors, has previously shown superiority over placebo on improving NS in prospectively designed study in patients with schizophrenia. The objective here is to explore the effect of roluperidone compared to placebo, on the 2 domains of the Negative Symptoms Factor Score (NSFS): Reduced Experience and Expression in patients with stable symptoms of schizophrenia.

    Methods: This was a multi-national Phase 2b trial that enrolled 244 patients diagnosed with schizophrenia who were symptomatically stable for ≥ 3 months prior to entering the trial and had baseline scores ≥ 20 on the NS subscale of the PANSS. Patients were randomized to daily monotherapy with roluperidone 32 mg, roluperidone 64 mg, or placebo in a 1:1:1 ratio. The primary endpoint was the PANSS negative symptom score based on the 5-factors (pentagonal) model. The Mixed-Effect Model Repeated Measure (MMRM) was used for analyzing the 2 domains of NSFS.

    Results: All enrolled patients (N=244) were Caucasian, and 137 (56%) were male. Median age was 41 years and range from 18 to 60. The 3 treatment groups were balanced on all demographic and illness-related baseline characteristics. Completion rates for randomized patients in this 12-week study were as follows: roluperidone 62 mg = 69%, roluperidone 32 mg = 69% and placebo = 65%.

    The NSFS mean scores at baseline were 25.0 ± 3.98 for the placebo group, and 25.2 ± 4.00 for the 2 roluperidone groups. Both doses of roluperidone were superior to placebo on both domains: Experience (p ≤ 0.006 for the 32 mg; p ≤ 0.001 for the 64 mg) with persistent significant superiority beginning at Week 2 for the 64 mg dose and at Week 8 for the 32 mg dose. Expression (p ≤ 0.003 for the 32 mg; p ≤ 0.001 for the 64 mg) with similar persistent significant superiority.

    Conclusions: While both doses of roluperidone improved negative symptoms and showed good tolerability in stable schizophrenia patients, the post hoc analysis reported here found the drug to work on all aspects of negative symptoms as quantified by the PANSS in general, as well as the reduced emotional experience and reduced emotional expression subscales recently empirically derived from the PANSS.

    Keywords: Negative, Symptoms, Expression, Experience, Schizophrenia

    Disclosure: Minerva Neurosciences, Employee

    M186. Reduction in Peripheral C-Reactive Protein Levels With Canakinumab Administration is Related to Reduced Positive Symptom Severity in Patients With Schizophrenia and Inflammation

    Thomas Weickert*, Cynthia Shannon Weickert

    University of New South Wales, Randwick, Australia

    Background: Schizophrenia is characterized by positive and negative symptoms and cognitive impairments that are related to functional disability. All present approved treatments targeting small molecule neurotransmitter receptors are limited in their effectiveness and leave many patients with residual symptoms and debilitating side effects. Psychotic symptoms, cognitive impairment, and treatment response are variable in schizophrenia, highlighting heterogeneity in the etiology and presentation of the illness. Thus, there is a critical need for novel treatments targeting subgroups related to underlying biology that can be identified by biomarkers. We have shown that a substantial subgroup (approximately 40%) of people with schizophrenia can be distinguished by inflammation in both peripheral blood and in brain, e.g., the cytokine interleukin 1-beta (IL-1β) is elevated in schizophrenia. IL-1β mRNA and protein levels are significantly increased in serum, plasma, white blood cells, cerebrospinal fluid and brain tissue in chronically ill patients with schizophrenia and in first episode psychosis. Inhibition of excessive immune response is possible by interrupting the cascade of cytokine responses which can be achieved by administering canakinumab: an approved human anti-interleukin 1β monoclonal antibody that interferes with the bioactivity of IL-1β. For the approved indications, canakinumab is effective for up to 4 to 8 weeks following a single injection. However, the extent to which adjunctive IL-1β blockade by canakinumab can reduce peripheral markers of an overactive immune system (e.g., high sensitivity C-reactive protein: hsCRP) and bring about amelioration of psychotic symptoms in schizophrenia is unknown.

    Methods: We conducted a randomized, placebo-controlled, double-blind, parallel group trial of the monoclonal antibody canakinumab to block IL-1β in schizophrenia. Twenty-seven chronically ill patients with schizophrenia or schizoaffective disorder who had elevated peripheral inflammation markers (IL-1β, IL-6, hsCRP and/or neutrophil to lymphocyte ratio) were randomized to a one-time subcutaneous injection of canakinumab (150 mg) or placebo (normal saline) as an adjunctive treatment to their usual antipsychotic. Peripheral hsCRP levels were measured at baseline (prior to injection) and at 1, 4 and 8 weeks after a single injection of canakinumab or placebo. Positive and Negative Syndrome Scale scores were assessed at baseline and at 4 and 8 weeks after injection of canakinumab or placebo.

    Results: Separate t-tests comparing canakinumab or placebo treatments at weeks 1, 4, and 8 to baseline showed significant reductions in peripheral hsCRP levels at all time points (all p’s < .02) in the canakinumab treatment group only. There were no significant changes in hsCRP levels in the placebo group. Based on separate t-tests, we also found a significant reduction in positive symptom severity scores at week 8 (p = 0.05) in the canakinumab group and at week 4 (p = 0.02) in the placebo group. There was a trend for low peripheral hsCRP levels to be strongly correlated with low positive symptom severity scores (r = .57, p = .07) only at week 8 in the canakinumab treatment group only. There were no significant reductions in negative or general psychopathology symptom severity scores in either canakinumab treatment or placebo groups. Given that canakinumab was administered via subcutaneous injection, all participants were compliant and the dropout rate (at only 7%), was low.

    Conclusions: Preliminary results from our clinical trial suggest that blockade of the cytokine IL-1β by the monoclonal antibody canakinumab can significantly reduce peripheral hsCRP serum levels and that these reductions may be related to a reduction in positive symptom severity in chronically ill patients with schizophrenia who received adjunctive canakinumab. It is important to note that the groups were selected such that only those patients who initially display elevated peripheral markers of inflammation were recruited. The effects of canakinumab on the reduction of peripheral hsCRP would potentially be of benefit to general health in schizophrenia. The relatively strong relationship between the effects of canakinumab on the biological marker hsCRP and positive symptom severity after 8 weeks of treatment supports the effect of canakinumab on positive symptom severity. Given that treatment with a monoclonal antibody is a novel and substantial advance in the potential treatment of psychotic symptom severity in schizophrenia, future studies should consider increased and/or top-up doses with longer follow up assessments to confirm the benefit of adjunctive canakinumab treatment in schizophrenia.

    Keywords: Schizophrenia Novel Treatment, Positive Symptom Factor, Antibody, Inflammation, IL-1b

    Disclosure: Nothing to disclose.

    M187. Replicated Associations Between Common Variants and Neurocognitive Functioning in Adult- And Childhood-Onset Schizophrenia

    Tim Bigdeli, Keith H. Nuechterlein, Catherine Sugar, Kenneth L. Subotnik, Thomas Kubarych, Michael Neale, Kenneth Kendler, Robert Asarnow*

    UCLA, Los Angeles, California, United States

    Background: The aggregation of neurocognitive deficits among both the non-psychotic first-degree relatives of adult- and childhood-onset schizophrenia patients suggests that there may be a common etiology for these deficits in childhood- and adult-onset illness. Our comprehensive assessment of a wide range of neurocognitive domains in adult- and childhood-onset schizophrenia families yielded comparable estimates of familial transmission for some dimensions, as well as preliminary evidence of partially divergent multifactorial architectures for attention, working memory, and verbal learning. However, whether specific genetic factors influencing these neurocognitive domains have comparable effect sizes in adult- and child-onset schizophrenia is not known.

    Methods: The Consortium on the Genetics of Schizophrenia (COGS) single nucleotide polymorphism (SNP) chip includes 1,536 common variants and provides excellent coverage of 94 functionally relevant genes including AKT1, CHRNA7, COMT, DAO, DAOA, DISC1, DTNBP1, ERBB4, GRM3, GSK3B, NOS1AP, NRG1, PAFAH1B1, PPP3CC, PRODH, RELN, and RGS4. Using the COGS custom array, we genotyped probands and relatives from the UCLA Family Study (N=458), including N=109 from 39 childhood-onset schizophrenia families, N=148 from 67 adult-onset schizophrenia families, and N=201 from 66 age-matched community control pedigrees. We used the variance-component association module of MERLIN (v1.1.2) to test for association between SNPs and empirically derived composite scores representing attention, working memory, verbal learning, verbal retention, and memory for faces.

    Results: We attempted replication of selected genotype-phenotype associations based on analogous neurocognitive assessments reported by Greenwood et al. (2011) and established an experiment-wide multiple-testing threshold based on 15 a priori hypotheses (P<10e-3). We successfully replicated associations between the verbal learning factor and GRIN2B and NRG1 in adult- and childhood-onset schizophrenia, respectively; and replicate an association between working memory and GRID2 in both adult- and childhood-onset schizophrenia. Associations between variants in ERBB4 and verbal learning were successfully replicated in our analysis of childhood-onset schizophrenia but fell short of our experiment-wide significance threshold in the adult-onset analysis (P<0.001). Comparing observed directions of allelic effects for significant gene-trait associations from Greenwood et al., we observed the greatest degree of consistency with adult-onset schizophrenia findings for the verbal learning factor (P=0.0195). Comparing association results for childhood- and adult-onset schizophrenia, we observed an excess of same-direction effects for the verbal learning factor (P<0.01).

    Conclusions: We successfully replicate several previously reported associations between neurocognitive performance and biologically relevant loci in adult- and childhood-onset schizophrenia. While diverging evidence of replication in adult- and childhood-onset analyses may reflect limited statistical power to detect associations, the observed pattern of findings may also reflect their partially divergent multifactorial architectures.

    Keywords: Childhood-Onset Schizophrenia, Genetic Association Study, Neurocognitive Functioning, First Episode Schizophrenia

    Disclosure: Nothing to disclose.

    M188. Abnormal Effective Connectivity During Eye Gaze Processing in Schizophrenia

    Open Board.

    M189. Engagement of the Temporo-Parietal Junction and Posterior Superior Temporal Sulcus (TPJ-PSTS) by Naturalistic Stimuli in Individuals at High-Risk for Developing Schizophrenia

    Gaurav Patel*, Nahian Chowdhury, Juan Sanchez-Peña, Cheryl Corcoran, Daniel Javitt

    Columbia University, New York, New York, United States

    Background: Deficits in social functioning are a major cause of psychosocial disability in schizophrenia patients (SzP). Many of the brain networks involved in social functioning have nodes in the temporoparietal junction and posterior superior temporal sulcus (TPJ-pSTS), which may act as an integrative hub for the cognitive processes underlying social functioning. This region has previously been shown to only be engaged by naturalistic stimuli requiring integration of information over long time-scales (>30 seconds). This integration in the TPJ-pSTS has recently been shown by Lerner et al. 2018 to be disrupted in patients with first episode schizophrenia. It remains unclear, however, whether this disruption is the result of the development of schizophrenia or precedes its onset.

    To answer this question, we examined the engagement of TPJ-pSTS and other cortical areas by a naturalistic auditory stimulus in individuals at high risk for developing schizophrenia (CHR) compared to a sample of healthy controls (HC) and chronic schizophrenia patients (SzP). CHRs in our sample are individuals who already have some positive symptoms and meet criteria for Attenuated Positive Symptom Syndrome (APSS). Subjects listened to a 7-minute radio story (“Pieman”, adapated from Lerner et al. 2011) while BOLD-fMRI data was collected. They also listened to scrambled versions of this story, scrambled by paragraph, sentence, and word. We then compared the intersubject correlation (ISC) of these areas between the three populations.

    We predicted that, like the first-episode patients, the SzP in our sample would demonstrate integration deficits in the TPJ-pSTS as compared to the HC. For CHRs, we predicted that the ISC for these areas would be between that of HCs and SzPs.

    Methods: 12 CHRs, 19 SzPs, and 16 HC performed the Pieman task while BOLD-fMRI data were collected. CHR patients met criteria for attenuated positive symptom syndrome (APSS, >=1 positive symptom >= 1×/week in the past month) with worsening of positive symptoms ≥1 pt in the past year as rated by the SIPS/SOPS. SzPs met DSM-IV criteria for schizophrenia or schizoaffective disorder. HCs were age matched to the SzPs and CHRs and did not meet criteria for any major DSM-IV Axis 1 disorders. Subjects lay in the NYSPI 3.0 T GE MR750 MRI while listening to the Pieman stimuli through noise insulated headphones; subjects were simply instructed to passively listen to the story. A separate BOLD run of about 7.5 minutes was collected for each stimulus. BOLD-fMRI data were collected with a Human Connectome Project (HCP) compatible functional sequence: a multi-band fMRI sequence with a TR=850 ms and a spatial resolution of 2 mm isotropic and were processed with the HCP processing pipelines. Additional processing was performed to remove spurious artifacts related to movement and other sources, including global signal regression and censoring of high movement frames with a framewise displacement >.2 mm. Parcels from Gordon et al. 2014 were used to parcellate the time-series data from each subject. ISC was then calculated for each parcel by correlating the timecourse of activity evoked by the Pieman stimuli to the average of all of the HCs for each parcel. HCs were correlated to a leave-one-out average of all other HCs. These individual ISC maps were then used to calculate the significance of ISC for each group and for the pairwise comparison of each group.

    Results: As expected, HCs demonstrated significant ISC of TPJ-pSTS, mid and anterior STS, and auditory cortex in both hemispheres (p<.0001). In addition, anterior medial prefrontal default mode areas in both hemispheres also demonstrated significant ISC. Also, as expected, SzPs demonstrated significantly decreased ISC of posterior TPJ parcels corresponding to the default mode network as well as a number of STS areas in both hemispheres, along with anterior medial prefrontal default mode areas (p<.05). CHRs, however, largely resembled the ISC pattern of HCs as opposed to SzPs, with significantly stronger ISC in TPJ-pSTS, mid/anterior STS, and dorsomedial prefrontal cortex compared to SzP. Only anterior-medial prefrontal cortex demonstrated any significantly decreased ISC in CHRs vs. HCs (p<.05).

    Conclusions: In the present study, we have found that engagement of the TPJ-pSTS and other cortical areas by a naturalistic stimulus is largely similar in CHRs vs. HCs, even though this sample of CHRs has already begun to experience some positive symptoms. This suggests that the onset of schizophrenia causes some degeneration of the integrative functions in the TPJ-pSTS and other cortical areas. Our findings not only demonstrate the utility of using naturalistic stimuli to study psychiatric disorders, but also helps to establish a timeline for the cortical changes that accompany the onset of schizophrenia.

    Keywords: Language, Functional MRI (fMRI), Default Mode Network (DMN)

    Disclosure: Pfizer Inc, Employee

    M190. Translational Studies of Neuroscience-Informed Cognitive Training in Schizophrenia

    Rogerio Panizzutti*, Gerson Guercio, Linda Scoriels, Melissa Fisher, Sophia Vinogradov

    Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

    Background: Cognitive impairments are related to deficits in primary auditory and visual sensory processes in schizophrenia. These impairments can be remediated by neuroscience-informed computerized cognitive trainings that target auditory and visual processes. However, it is not clear which modality results in more transfer of benefits to non-trained tasks and clinical outcomes. In this study, we investigated the impact of training auditory or visual cognitive processes in parallel experiments with schizophrenia subjects and mice model of schizophrenia.

    Methods: Seventy-nine schizophrenia participants were randomly assigned to either 40 h of auditory or visual computerized training. Auditory and visual exercises were chosen to be dynamically equivalent and difficulties increased progressively during the training.

    Results: Participants who received the visual training showed significant improvements in global cognition compared to the auditory training group. The visual training significantly improved attention and reasoning and problem-solving, while the auditory training improved reasoning and problem-solving only. Schizophrenia symptoms improved after training in both groups, whereas quality of life remained unchanged. Interestingly, there was a correlation between attention and schizophrenia symptoms improvements for the visual training group. In mice, auditory training induced improvements in the pre-pulse inhibition of the startle response.

    Conclusions: We conclude that the visual training and the auditory training are differentially efficient at remediating cognitive deficits and symptoms of clinically stable schizophrenia patients. Cognitive training can also induce transfer of benefits to untrained tasks in an animal model of schizophrenia.

    Keywords: Cognition, Schizophrenia Novel Treatment, Computerized Cognitive Training

    Disclosure: NeuroForma, Stock / Equity

    M191. Emotion Identification Bias in Youths With Psychosis-Spectrum Symptoms From the Philadelphia Neurodevelopmental Cohort

    Sunny Tang*, Lauren White, Tyler Moore, Ran Barzilay, Monica Calkins, Raquel Gur, Ruben Gur

    University of Pennsylvania, Philadelphia, Pennsylvania, United States

    Background: Individuals with psychosis-spectrum (PS) disorders and clinical risk are more likely to exhibit an attribution style that is biased toward perceiving hostility and external causality. Biased attribution style in PS is associated with increased persecutory ideations and impaired functioning. We examine whether PS youths over-identified threat-related emotions in an emotion recognition task compared to non-PS youths.

    Methods: The Philadelphia Neurodevelopmental Cohort is a large non-clinical sample including 9,292 youths aged 8 – 21 years who underwent structured evaluation of psychiatric symptoms and a computerized neurocognitive battery (CNB). The CNB includes the Penn Emotion Identification Task (ER40) where participants are asked to identify the emotion corresponding to 40 faces (angry, fearful, happy, sad, or neutral). Accuracy, speed, and efficiency (combination of accuracy and speed) were collected. A bias score for misidentifying faces as being threat-related (angry or fearful) was calculated in contrast with bias toward non-threat-related emotions (happy, sad, neutral). Scores were standardized and adjusted for age where appropriate. Linear and logistic regressions were used.

    Results: PS youths were less accurate for threat (p = 0.003, Std β Coef = -0.01), nonthreat (p < 0.001, Std β Coef = -0.29) and overall items (p <0.001, Std β Coef = 0.05) when covarying for sex and socioeconomic status. There were no sex by threat or nonthreat accuracy interactions. Accounting for overall poorer performance, efficiency in identifying non-threat items was more impaired in PS (p = 0.007, Std β Coef = -0.01), but efficiency in identifying threat items was preserved (p = 0.37, Std β Coef = 0.004). Accounting for overall accuracy, PS youths were also more likely to score correctly for angry faces (p = 0.04, Std β Coef = 0.009); this was also true when comparing accuracy on angry faces to a dimensional measurement of psychosis severity (p = 0.02, Std β Coef = 0.02). Increased threat bias (p < 0.001, β Coef = 1.8 ×10-4) but not bias toward non-threat emotions (p = 0.12, β Coef = -8.5 ×10-5) was correlated with older age; both were correlated with male sex, and not related to race or socioeconomic status. Greater overall cognitive impairment (p < 0.001, Std β Coef = 0.009) and use of an illicit substance (p = 0.02, Std β Coef = 0.009) were correlated specifically with greater threat bias and were not related to bias toward non-threat emotions.

    Conclusions: Youths with PS symptoms perform worse on both threat and nonthreat emotion items. When accounting for overall differences in performance, threat-related emotion identification is relatively preserved in PS. Threat bias is correlated with older age, male sex, poorer overall cognitive performance and substance use. Effect sizes are generally small. Future directions include replication in a sample with more severe psychosis spectrum pathology and specifically in young people who endorse threshold and subthreshold persecutory ideations.

    Keywords: Clinical High-Risk State for Psychosis, Social Cognition, Facial Emotion Processing, Early Illness Schizophrenia, Threat Faces

    Disclosure: Nothing to disclose.

    M192. White Matter Integrity and Verbal Memory Performance in Schizophrenia Compared to Healthy Controls: Differences Between Episodic Memory and Working Memory

    Leticia Czepielewski*, Clarissa Gama

    Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil

    Background: Cognitive impairments are core symptoms in schizophrenia. Among them, memory deficits seem to be strongly related to several outcomes, such as psychosocial functioning and accelerated aging. Nevertheless, the neurobiological mechanisms associated with memory dysfunction are still unclear. One hypothesis would be related to white matter microstructural abnormalities, since neuronal connectivity underlies cognitive processing. More specifically, uncinate fasciculus, cingulum cingulate and angular bundles, and superior longitudinal fasciculus-temporal terminations seem to be relevant to memory performance. They connect the medial temporal lobe to other cortical regions and the cingulate to the prefrontal cortex. Therefore, we aimed to study the relationship between white matter integrity in relevant tracts using diffusion tensor imaging (DTI) and episodic memory and working memory in individuals with schizophrenia (SZ) compared to healthy controls (HC).

    Methods: This was a cross-sectional study. We included 100 participants, 43 individuals with SZ that were medicated according to guidelines and stable for at least 6 months, and 57 unaffected individuals. All participants were informed about study procedures and signed consent before assessment. Research protocol was conducted in accordance with the Helsinki Declaration. The Institutional Review Board approved the study protocol. Participants underwent cognitive assessment with Hopkins Verbal Learning Test for episodic memory and Letter-Number Sequencing subtest from WAIS-III for working memory. Diffusion tensor imaging was performed on a Philips Achieva 1.5 T scanner. We performed linear regression models separately for patients and controls with the sum of left and right fractional anisotropy (FA) of the abovementioned relevant tracts predicting memory performance, controlling for age, sex, and years of education.

    Results: Individuals with SZ had worse episodic and working memory performances compared to HC (t(96) = 7.088, p < .001 and t(97) = 4.614, p < .001, respectively), however groups did not show differences among FA of white matter tracts (p > .05). Episodic memory was not significantly predicted by uncinate fasciculus, cingulum-cingulate bundle, cingulum-angular bundle, and superior longitudinal fasciculus-temporal terminations FAs in both HC and SZ. Working memory was not significantly predicted by uncinate fasciculus, cingulum-cingulate bundle, and superior longitudinal fasciculus-temporal terminations FAs. However, cingulum-angular bundle showed a significant main effect predicting working memory in SZ (t = 2.303, p = .027, Beta = .307; model: F = 5.209, p = 0.002, R2Adj = .286), but not in HC (t = .383, p = .704, Beta = .046; model: F = 7.369, p < .001, R2Adj = .329).

    Conclusions: Decreased white matter integrity in cingulum-angular bundle, which are fibers that connect frontal, parietal and temporal lobes, was related to worse working memory performance in schizophrenia. This highlights the importance of neuronal connectivity abnormalities to higher-order cognitive processes in psychosis. The cingulum microstructure is described as being related to cognitive control and executive function, and thus could be a focus for new therapeutic strategies focusing on cognitive symptoms.

    Keywords: Verbal Episodic Memory, Working Memory, Diffusion Tensor Imaging (DTI)

    Disclosure: Nothing to disclose.

    M193. Integrated EEG-fMRI Analysis of Reward Processing During Slot Machine Play in Schizophrenia

    Susanna Fryer*, Brian Roach, Judith Ford, Michael Jacob, Daniel Mathalon

    University of California, San Francisco, San Francisco, California, United States

    Background: Negative symptoms of schizophrenia, such as anhedonia and amotivation, signal an absence of typical hedonic and motivational outputs, and may reflect underlying alterations in the way the brain processes rewarding stimuli. The goal of this study was to directly combine neuroimaging modalities to enable a more comprehensive assessment of spatial (fMRI) and temporal (EEG) features of reward-related brain responses to help characterize reward processing dysfunctions in schizophrenia.

    Methods: EEG and fMRI were recorded in separate sessions while participants (patients with schizophrenia (SZ)=50; healthy controls (HC) = 45) completed the same slot machine gambling task. The reward positivity (RewP), an event-related potential reflecting reward outcome evaluation, was measured from the EEG data by creating a difference wave of wins – losses from a time window established in the literature by prior meta-analysis (Sambrook and Goslin, 2015). Reward-related spatial activations contrasting brain responses to wins vs. losses were modeled using standard voxelwise parametric methods. Next, these EEG and fMRI signals of interest were entered into a Joint Independent Components Analysis (jICA) to establish how task-specific temporal patterns in EEG related to spatial fMRI activation patterns, and to interrogate group differences in the relationship between these electrophysiological and hemodynamic reward-related signals.

    Results: RewP (gain – loss) event-related potentials and fMRI activations to rewards vs. losses were equivalent across SZ and HC groups. Joint analysis revealed two components of interest showing distinct patterns of association between EEG and fMRI signals. First, a component that reflected the temporal signature of the RewP in the EEG domain (peaking between 250-300 ms) showed strong, but equivalent, covariation in both groups with fMRI activations from regions including bilateral anterior cingulate, insula, and distributed visual cortex. Second, a later evaluative component (peaking ~400 ms) distinguished the groups, with patients showing significantly weaker co-activation, relative to HCs, who showed strong, positive covariance with activations in regions including bilateral brodmann area 39, and left brodmann areas 6 and 40 (p=.03).

    Conclusions: These data suggest that even in the context of our slot machine paradigm, which intentionally placed minimal demands on decision making and response planning, early reward evaluation signals appear preserved in schizophrenia, but abnormalities emerge at later stages of reward evaluation that likely reflect more extensive and/or higher-order processing of reward outcome information.

    Keywords: Prediction Error, Early Illness Schizophrenia, Frontomedial Negativity, Theta

    Disclosure: Nothing to disclose.

    M194. Spite Sensitivity and Persecutory Ideation in a Transdiagnostic Sample of Patients: An fMRI Study of Social Decision-Making Under Uncertainty

    Angus MacDonald*, Rebecca Kazinka, Danielle Pratt, Anita Kwashie, Saaraa Ameri, Krista Wisner, Sylia Wilson

    University of Minnesota, Minneapolis, Minnesota, United States

    Background: Persecutory ideation is a psychiatric symptom with devastating consequences. Recently, the neural mechanisms underlying persecutory ideation have been examined using decision-making tasks amenable to neuroimaging, such as the Minnesota Trust Game, that make it possible to distinguish between rational concerns about others’ selfishness versus concerns about others’ spitefulness. To evaluate the relationship between persecutory ideation and suspicious decision-making, we examined decision-making during fMRI in a sample of patients recruited for early-episode psychosis.

    Methods: To date, the two-player social decision-making task, Minnesota Trust Game (MTG), has been completed by a transdiagnostic sample of patients (N=28) during fMRI at 3 T. By varying financial risk and the partners’ temptation to defect, MTG provides a psychometric function that differentiates suspiciousness (fear of a partner being spiteful when the partner would lose from defection) from rational mistrust (fear of a partner being selfish because the partner gains from defection). We examined persecutory ideation symptoms, trusting behaviors and brain activation in cognitive control, mentalizing and reward networks.

    Results: Individuals who reported more persecutory ideation tended to be less trusting in the suspiciousness condition, but not the rational mistrust condition. The manipulation of financial risk by partner’s temptation to defect was associated with differential activity in the lateral orbitofrontal cortex (OFC) and nucleus accumbens (NAcc). Bilaterally, greater OFC activity was associated with increasing risk in the suspiciousness condition. This pattern was also observed in the medial frontal cortex (anterior default mode network, or DMN) and posterior cingulate (posterior DMN). Greater NAcc activity was associated with decreased risk, but only in the rational mistrust condition. Patients reporting more persecutory ideation showed less activity in posterior DMN as risk increased (r=-0.40, p≤.04).

    Conclusions: Using a novel paradigm in a transdiagnostic sample, we demonstrated the utility of studying spite sensitivity to advance neuroscientific understanding of persecutory ideation. We replicated previous behavioral findings using this task and have found a similar set of regions sensitive to interpersonal risk and reward as previously reported. While future analyses will examine the crucial hypothesis that dysfunctional synchrony between top-down executive and midline affective networks increase risk for persecutory ideation, the current findings highlight to role of mentalizing networks, such as the default mode network, in these symptoms.

    Keywords: Persecutory Ideation, Early Psychosis, Reward-Based Decision-Making, Paranoia

    Disclosure: Nothing to disclose.

    M195. Intra-Individual Cognitive Variability in Psychosis Spectrum Disorders: Association With Structural Brain Measures

    David Roalf*, Raquel Gur, Andrew Crow, Ruben Gur

    University of Pennsylvania, Philadelphia, Pennsylvania, United States

    Background: Neurocognitive deficits are a core feature of psychosis {Goldman-Rakic, 1994 #320; Andreasen, 1997 #321; Harvey, 2006 #322}, are strong predictors of functional outcome {Harvey, 2006 #322}, and are associated with alterations in brain structure {Shepherd, 2012 #287} and function {Gur, 2010 #323}. Hence, recent neurocognitive studies have focused on individuals at clinical high-risk (CHR) for developing psychosis. The profile of neurocognitive deficits in CHR largely mimics the pattern seen in psychosis {Carrion, 2015; Brewer, 2005; Brewer, 2006; Gur, 2015}, and recent evidence suggests there may be little progression of these deficits in CHR individuals after conversion to psychosis {Carrion, 2015}. As such, cognitive deficits likely confer vulnerability to developing psychosis making them a salient marker of incipient illness. Yet, the preponderance of neurocognitive research has emphasized group differences in mean neurocognitive performance. This approach largely ignores intra-individual variability (IIV) in neurocognitive performance, which likely confers unique predictive information about cognitive functioning beyond mean performance {Jensen, 1992 #318; Hultsch, 2004 #334; MacDonald, 2006 #319}

    In general, IIV reflects within-person fluctuations on a trial-by-trial basis {MacDonald, 2006 #319}. The inability to maintain consistent responding across tasks is associated with difficulty in tuning cortical circuitry and in the failure to optimize signal-to-noise in prefrontal/anterior cingulate cortex {MacDonald, 2009 #39} and distinct control networks in the brain {Fair, 2007 #432}. This disruption in communication may be associated with loss of gray matter (GM) or disorganization of white matter (WM) {MacDonald, 2006 #319}, which are typically disrupted in psychosis and in CHR. Yet, there is limited investigation of IIV in individuals at-risk for developing psychosis and associated structural brain phenotypes.

    Methods: Young adults and adolescents were recruited into one of two groups: 1) Psychosis Spectrum (PS; n=32), which included clinical high-risk youth and those meeting criteria for DSM-IV psychotic disorder; and 2) Typically Developing (TD) youth who were symptom free, without an Axis II Cluster A diagnosis or a family history of psychosis (n = 28). All individual underwent neurocognitive testing the Variability Toolbox. The Variability Toolbox (see Figure 1) is a set of eight computerized tasks designed to measure variability in reaction time during neurocognitive performance (~35 minutes). This Toolbox is specially designed to index patterns of moment-to-moment inconsistency during speeded cognitive testing. The tasks include: Simple Motor Reaction Time, Attention & Working Memory (2 tasks), Simple and Complex Choice Reaction Time, Simple and Complex Processing Speed and Abstract Matching. Most individuals also underwent 3 T MRI, which include T1-weighted imaging and diffusion-weighted imaging.

    Results: Overall, PS performed more poorly than TD in accuracy (F(1,53) = 13.44, p < .001, η2 = .16) and speed (F(1,53) = 5.93, p = .018, η2 = .06) measures. IIV was significantly higher across all tasks in PS as compared to TD. (F(1,53) = 10.59, p = .002, η2 = .21). The largest IIV effects were found in simple reaction time and processing speed, followed by complex processing speed and reasoning. Small (Pearson rs = 0.25), but positive associations between IIV and diffusion-weighted scalar measures were also found, particularly within the corpus callosum.

    Conclusions: Our work indicates that youth with psychosis spectrum symptoms exhibit subtle, but significant, abnormalities in cognitive performance variability (IIV). Preliminary associations between IIV and brain white matter organization suggest important association, particularly within the corpus callosum. Given the clinical evidence implicating abnormal neurodevelopment in the pathogenesis of schizophrenia, and the potential utility of cognitive and neuroimaging measures to predict illness vulnerability, this approach holds promise for understanding neurodevelopmental contributions to psychosis pathophysiology. These abnormalities may reflect aberrant brain development that predisposes an individual to the illness, and quantification of these features will likely enhance our ability to detect those truly at risk for developing psychosis.

    Keywords: Cognition, Clinical High Risk for Psychosis, Diffusion Tensor Imaging (DTI)

    Disclosure: Nothing to disclose.

    M196. Interaction Between Neurological Soft Signs, Extrapyramidal MotorSymptoms and Antipsychotic Medication in Schizophrenia Spectrum Disorders

    Dusan Hirjak*, Stefan Fritze, Fabio Sambataro, Alina Bertolino, Cristina Topor, Katharina Kubera, Robert Christian Wolf

    The Central Institute of Mental Health, Mannheim, Germany

    Background: In the last two decades, neurological soft signs (NSS) have garnered increasing attention in neuroimaging research on genuine motor abnormalities (GMA) in schizophrenia spectrum disorders (SSD). However, the majority of these studies examined patients receiving antipsychotic treatment and therefore it still remains unclear whether NSS levels might have been confounded by antipsychotics. Here, we address the question whether NSS scores are associated with extrapyramidal motor symptoms (EPMS) and antipsychotic treatment in SSD?

    Methods: A comprehensive motor assessment was performed to examine NSS and EPMS by means of standardized instruments in 105 patients with SSD. Antipsychotic dose equivalence estimates were determined by the classical mean dose method (doses equivalent to 1 mg/d olanzapine). Relationships between NSS, EPMS and antipsychotic medication were determined using correlational analysis and multiple linear regression analyses.

    Results: NSS and EPMS scores were not significantly associated with olanzapine equivalents. NSS were not significantly associated with other EPMS scales (p=0.003; Bonferroni-corrected for multiple comparisons). There was a significant association between NSS motor coordination and global parkinsonism (p=0.002, Bonferroni-corrected for multiple comparisons).

    Conclusions: Our results support the view that NSS and the majority of EPMS are not significantly modulated by antipsychotic medication in SSD. The significant relationship between NSS and global parkinsonism supports the genuine rather than medicationdependent origin of particular motor symptoms in SSD.

    Keywords: Schizophrenia, Neurological Soft Signs, EPMS, Antipsychotic Medication

    Disclosure: Nothing to disclose.

    M197. Brain Insulin Resistance and Altered Brain Glucose are Related to Memory Impairments in Schizophrenia

    Andrea Wijtenburg*, Dimitrios Kapogiannis, Stephanie Korenic, Roger Mullins, Joyce Tran, Frank Gaston, Maja Mustapic, Elliot Hong, Laura Rowland

    Maryland Psychiatric Research Center, University of Maryland School of Medicine, Baltimore, Maryland, United States

    Background: Memory impairments are among the most robust cognitive impairments observed in schizophrenia (SZ) and are related to functional outcome. Altered brain glucose (Glc) metabolism has been associated with memory dysfunction, and studies in SZ show lower Glc metabolism using PET. Glc hypometabolism in Alzheimer’s disease (AD) is associated with insulin resistance (IR), which is believed to be an underlying cause of the illness. Recent work has identified biomarkers of neuronal-specific IR in the setting of AD by isolating blood exosomes of neuronal origin and measuring levels of phosphorylated insulin signaling effector molecules. Complementary to exosome IR biomarkers, measures of brain Glc levels may provide further insight into brain Glc metabolism. Currently, magnetic resonance spectroscopy (MRS) is the only non-invasive technique that does not use ionizing radiation to study cerebral Glc. In this study, we examine the relationships between biomarkers of brain IR, brain Glc levels using MRS, and memory function in SZ.

    Methods: Twenty-two adults with SZ and 24 healthy controls participated in this study. Verbal and visual spatial memory were assessed using the HVLT and BVMT from the MATRICS battery. We isolated exosomes from fasting plasma samples and enriched them for neuronal origin with immunoprecipitation for neuronal cell adhesion molecule L1-CAM. We quantified insulin signaling markers Akt, pAkt (phosphorylation of the Ser473 residue), GSK3B, pGSK3□ (phosphorylation of the Ser9 residue), p70S6K and pp70S6K (phosphorylation of the Thr421/Ser424 residue). Participants were scanned on a Siemens Trio 3 T MR system (Erlangen, Germany) using a 32-channel phased array head coil housed at the CBIR. A short TE PR-STEAM sequence (TR/TE=2000/6.5 ms, 128 excitations, VOI ~ 24 cm3) was used to acquire Glc data from the occipital cortex. Spectra were analyzed with LCModel and were corrected for the proportion of gray matter, white matter, and CSF within the voxel using in-house Matlab code. T-tests with significance set at p<0.05 were performed to examine group differences, and correlation analyses were performed to examine relationships between brain IR, brain Glc, and memory function.

    Results: Exosome biomarkers, Akt and p70S6K, showed trends for being lower in adults with SZ compared to controls (p = 0.097 and p = 0.112, respectively). MRS Glc was higher in adults with SZ compared to HC (p = 0.002), and BVMT and HVLT scores were lower in adults with SZ compared to HC (p = 0.002 for both). Correlation analyses revealed several trend-level relationships only in adults with SZ such that MRS Glc was correlated with exosome biomarkers p70S6K (r = -0.432, p = 0.095), and pGSK3β (r = -0.477, p = 0.062), as well as with BVMT scores (r = -0.448, p = 0.071). Exosome biomarker p70S6K was correlated with HVLT (r = 0.541, p = 0.017). There were no significant relationships between exosome biomarkers, MRS Glc, and memory in HC.

    Conclusions: Our results suggest that higher levels of brain Glc, reflective of poor glucose utilization, and brain IR are evident in schizophrenia. For the first time, we show that MRS Glc and brain IR biomarkers are related to each other, as well as with memory impairments in SZ. Therefore, brain IR may play a role in the pathogenesis of memory dysfunction in SZ, and treatments targeting brain IR may be useful in alleviating memory deficits in SZ.

    Keywords: Schizophrenia, Brain Insulin Resistance, Magnetic Resonance Spectroscopy, Exosomes of Neuronal Origin, Memory Function

    Disclosure: Nothing to disclose.

    M198. Effects of Acute Olanzapine Exposure on Central Insulin-Mediated Regulation of Whole Body Fuel Selection and Feeding

    Abstract not included.

    M199. Clinical, Neurocognitive, and Genetic Characteristics of Youths at Familial High Risk for Psychosis

    Jerome Taylor*, Nana Asabere, Monica Calkins, Tyler Moore, Sunny Tang, Rose Xavier, Daniel Wolf, Laura Almasy, Ruben Gur, Raquel Gur

    Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States

    Background: Youth with a first-degree relative with psychosis (FHR-P) are at increased risk for psychosis. FHR-P youth also have increased psychopathology and cognitive deficits compared to other youth. In this study, we sought to characterize clinical, neurocognitive, and genetic features of FHR-P youths in the Philadelphia Neurodevelopmental Cohort (PNC).

    Methods: For 8,929 youths, ages 8-21, the PNC assessment included the Family Interview for Genetics Studies (FIGS) Screen and the FIGS Psychosis Supplement if the Screen found that first-degree relatives of the youth had symptoms that could possibly be due to psychosis. Familial high risk for psychosis (FHR-P) status was based on the FIGS done with the youth and caregiver, and we compared youths with (i.e. FHR-P) and without family histories of psychosis.

    Results: Family history of psychosis was reported for 5.6% (n=500) of youths. FHR-P youths had lower global assessment of functioning scores (p<0.001), increased mood, psychotic, externalizing, and anxious symptoms (all p<0.001) and similar neurocognitive scores relative to youths without a family history of psychosis. FHR-P youths were also more likely to be African-American, Hispanic, live in lower SES neighborhoods, and have trauma experiences (all p<0.001); however, functioning and clinical differences persisted after adjusting for these factors. Similar impairments were seen in FHR-P youths, whether the reported family history was a schizophrenia spectrum disorder, psychosis due to substance abuse or medical causes, or psychotic symptoms without a clear etiology. In a subsample of Caucasians (n=4,433), FHR-P youth had higher polygenic risk scores for schizophrenia (PRS) (p=0.02), and FHR-P status predicted psychopathology and functioning in youths even after adjusting for PRS.

    Conclusions: Self- and caregiver-reported family history of psychotic symptoms is an important risk factor for greater psychopathology and poorer functioning. We also found that FHR-P youths tended to live in more adverse environments and had increased genetic loading for psychosis, but these did not fully account for the clinical impairment. Given that FHR-P youth have increased risk for developing psychiatric disorders, future research should study FHR-P youth longitudinally to identify mechanisms by which familial psychosis impacts outcome and characterize both risk and resiliency factors for developing severe mental illness.

    Keywords: Polygenetic Risk Score, Children and Adolescents, Psychosis Risk

    Disclosure: Nothing to disclose.

    M200. Associations Among Prenatal History, Postnatal Brain Development, and Psychosis Risk in the Initial ABCD Data Release

    Kevin Dowling, Eren Sipahi, Gianni Petrozzino, Hamdi Eryilmaz, Ezra Susser, Joshua Roffman*

    Massachusetts General Hospital, Charlestown, Massachusetts, United States

    Background: Convergent evidence implicates the fetal milieu in postnatal brain development and risk for serious mental illness, but longitudinal and mechanistic studies are needed to develop tractable early interventions. The Adolescent Brain Cognitive Development (ABCD) study, a longitudinal multi-site investigation of clinical, cognitive, and brain imaging data, provides a new opportunity to relate prenatal history to risk for psychotic and other emergent symptoms in youths through MRI measures of neurodevelopment. Following the initial release of baseline data, we related obstetrical history to psychosis spectrum symptoms and cortical thickness at age 10 in the ABCD cohort.

    Methods: Baseline data from n=3,785 participants (age mean/SD 10.0/0.6 years, 47% female) were included in the present analysis. We excluded 739 participants whose structural scans failed the ABCD quality control pipeline or who were missing data. Included participants were enrolled across 20 US-based sites and scanned on one of 5 3 T magnet platforms. Obstetric history questionnaires were administered to the primary caregiver. Psychotic spectrum symptoms were measured using the Prodromal Questionnaire – Brief Child version (PQ-BC) and participants were grouped according to total distress score percentile (1-50, 51-60, 61-70, 71-80, 81-90, 91-99 %ile). Covariates of interest included maternal age, planned/unplanned pregnancy, vaginal/caesarian delivery, single/multiple live births, presence of pregnancy or birth complications, birth before/after 37 weeks’ gestation, and sex of child. Potential confounders also included in the model were caregiver education and income; child’s age, race, and ethnicity; and presence of a parental partner and any older sibling. Ordinal regression determined whether dependent variables predicted psychosis grouping, controlling additionally for study site. Cortical thickness measurements were derived from standardized T1 scan data; global mean thickness was the primary outcome, and for post hoc analyses, the cortex was divided into 12 bilateral parcels based on genetic correlation of surface area (Chen et al, 2012). General linear models examined the relation of dependent variables to cortical thickness, including any significant predictors of psychosis grouping from the previous ordinal regression analysis, as well as scanner type.

    Results: C-section (p=.003), younger maternal age (p=.004), unplanned pregnancy (p=.008) and pregnancy complications (p=.02) independently predicted increased psychosis scores after controlling for all other covariates. Of these, only unplanned pregnancy associated with reduced global cortical thickness (p=.016); further, pregnancy planning status interacted significantly with psychosis group to influence cortical thickness (p=.032), with strongest effects of unplanned pregnancy on reduced thickness in the 61-70 and 81-90 %ile psychosis groups (p<.05). Post hoc testing of regions-of-interest indicated significant interactive effects of unplanned pregnancy and psychosis group on thickness in bilateral posterior/lateral and superior temporal cortex, left dorsomedial frontal cortex, left precuneus, right inferior parietal cortex, right anterior/medial temporal cortex, and right pars opercularis.

    Conclusions: While echoing previous associations of obstetric complications to psychosis risk, the present results newly relate psychosis spectrum symptoms in youth to unplanned pregnancy. Further, they suggest that the well-established relationship between thinner cortex and psychosis risk may be modified by pregnancy planning status. Importantly, pregnancy planning and related maternal health decisions are complex, reflecting multiple behavioral and biological determinants, and specific mediating factors on brain development and psychosis remain to be elucidated. Other potential limitations include retrospective obstetrical history and unrecognized confounders. However, future data releases from ABCD will provide the opportunity to replicate and test longitudinal associations among the prenatal milieu, postnatal brain development, and psychosis risk.

    Keywords: Early Psychosis, Cortical Thickness, ABCD Study, Pregnancy

    Disclosure: Nothing to disclose.

    M201. Phenome Wide Association Study (PheWAS) of Schizophrenia Risk Genes and Associated Diagnosis in Patients From the Mount Sinai Biome Database

    Whitney McFadden*, Laura Huckins, Andrew Chess, Eli Stahl, Marie Verbanck, Lisheng Zhou

    The Mount Sinai Hospital, New York, New York, United States

    Background: Schizophrenia is a complex neuropsychiatric disorder that presents with psychosis, paranoid ideation, hallucinations, and flat affective symptoms. Large well powered Genome Wide Association Studies (GWAS) have identified over 255 genome-wide significant loci that confer associated risk for disease, however their function is poorly understood. In addition, the complexity of phenotypes is not fully captured and heterogeneity in clinical presentation can make our understanding of genetic risk and specific function of risk loci unclear.

    Phenome-wide association studies (PheWAS) help us to identify unbiased genetic associations with traits and diagnoses to better understand disease mechanisms and discover variants that confer risk for multiple diseases with similar underlying genetic causes. Conversely to GWAS, which compare allele frequencies between very well defined ‘cases’ and ‘controls’, PheWAS seeks to understand the impact of a given genotype, across all phenotypes within our medical system. For example, we might ask which traits and disorders occur in individuals with a specific schizophrenia-associated genotype, or differences in expression of a schizophrenia risk gene in relevant tissues.

    These analyses can identify other disorders and traits affected by schizophrenia risk variants and genes and could provide clues to their biological mechanisms and functions.

    Here, we ran PheWAS on well-studied schizophrenia-associated genes and tested for disease associations in a large metropolitan biobank.

    Methods: We used electronic health records (EHRs) and genotyping information from the Mount Sinai BioMe biobank. We applied “prediXcan” to impute genetically regulated gene expression (GREX) across multiple tissue types for 11,094 patients in BioMe that had genotypic and EHR records available at the time. By modeling genetically regulated gene expression in multiple tissues, we can observe relevant disease affected tissues without having to sample the tissues from patients. We separated biobank individuals into three groups, according to self-defined ancestry; European American (n=1,349), Hispanic American (n=3,083), and African American(n=2,252). We ran our PheWAS analysis separately for each population-specific group, and meta-analyzed results across groups using an inverse-variance based approach in metal.

    We defined traits and disorders within our biobank according to ICD9 codes. We included only traits or disorders occurring in at least 20 individuals within our biobank. Individuals with missing data (i.e. no history of a specific diagnosis) were treated as controls. In total, we tested 1,358 ICD9 codes.

    Our initial PheWAS analysis includes two sets of genes. First, 15 validation genes were used to establish consistency with previous PheWAS studies. Second, we tested five well-studied schizophrenia risk genes (FURIN, SNAP91, CLCN3, TSNARE1, CNTN4), identified in the CommonMind Consortium (CMC) post-mortem brain transcriptome analysis.

    We applied two thresholds for multiple testing. First, a gene-specific threshold, correcting for all ICD9 codes tested only. Second, an experiment-wide threshold, correcting for all genes and all ICD9 codes tested.

    Results: We first tested 15 genes with strong, replicating associations from PheWAS in other biobanks. In line with these previous studies, our PheWAS results revealed robust gene-endophenotype associations with diabetes, psoriasis, Crohn’s and rheumatoid arthritis. CMC derived genes reveal 12 associations reaching our per-gene threshold (p<8.4e-07), including SNAP91 with hearing loss (ICD9 389 p=5.2e-5) and CLCN3 with alcoholism (ICD9 317, 317.1 p<8.2e-5). SNAP91 reaches experiment wide significance with hearing loss (p=5.2e-5).

    Conclusions: The results we present here are the first stage of our analysis, applying PheWAS to psychiatric risk genes. As such, our results constitute a proof-of-concept: that PheWAS studies are a promising approach for studying psychiatric risk genes. We believe that PheWAS provides the opportunity to study the impact of psychiatric risk genes in the general public; this will lead to a deeper understanding of the biological mechanisms underlying these diseases.

    To date, we have validated a small subset of risk genes from the CMC study. Our next steps will incorporate schizophrenia and bipolar disorder risk genes from our recent large scale PrediXcan studies. We hypothesize that schizophrenia risk genes are associated with diagnoses and endophenotypes that have common mechanistic foundations.

    Keywords: Human Genetics, DNA Sequencing, Multiplex Families, Schizophrenia, Bipolar Disorder, Genetics, Genome Wide Association Study

    Disclosure: Nothing to disclose.

    M202. Convergent Phenotypic and Genomic Analyses of Symptom-Level Data in the Multi-Ancestry Genomic Psychiatry Cohort (GPC) Study of Schizophrenia

    Tim Bigdeli*, Jacquelyn Meyers, Helena Medeiros, Roseann Peterson, Jorge Valderrama, Giulio Genovese, Conrad Iyegbe, James A. Knowles, Kenneth Kendler, Carlos Pato, Michele Pato, Nina Schooler, Ayman Fanous

    SUNY Downstate Medical Center, Brooklyn, New York, United States

    Background: Schizophrenia is a complex and clinically heterogeneous syndrome, with individual patients varying widely in terms of symptomatic presentation, course, outcome, and severity. Whether categorical subtypes describe variation in clinical features is unclear, and subtype specifiers were removed from DSM-5 given a lack of corroborating field evidence. Dimensional constructs such as positive, disorganization and negative, and affective symptoms represent alternative conceptualizations of demonstrated clinical significance. However, the potential promise of clinical features in the prediction of course, outcome, and response to treatment has not yet been realized. Defining dimensional constructs with large data bases and linking these constructs to genetic information represents an important step in achieving this goal.

    Methods: We performed exploratory factor analysis (EFA) of symptom-level endorsement and interviewer rating data for N > 10,000 schizophrenia and schizoaffective disorder patients from the Genomic Psychiatry Cohort (GPC). The GPC is a large cosmopolitan sample of newly ascertained cases and screened controls with considerable representation of African and Latino ancestries. Participants enrolled as probable cases were interviewed by mental health professionals using the Diagnostic Interview for Psychosis and Affective Disorders (DI-PAD), a semi-structured clinical interview incorporating questions developed for the Diagnostic Interview for Genetic Studies (DIGS) (Nurnberger et al. 1994). Genome-wide single nucleotide polymorphism (SNP) data was available for 7,289 cases with complete DI-PADs, and we constructed polygenic risk scores using summary statistics from the Psychiatric Genomics Consortium (PGC) study of schizophrenia (PGC Schizophrenia Working Group, 2014).

    Results: Our EFA yielded a six (6) factor model representing delusions, hallucinations, and depressive, manic, disorganized and negative symptoms. We retained items with loadings of 0.4 or greater in a subsequent confirmatory factor analysis (CFA), which indicated good model fit (RMSEA = 0.043; SRMR = 0.040; CFI = 0.943). Next, we tested for associations between individual-level factor scores and polygenic risk of schizophrenia; we observed the strongest evidence of association for a combined negative/disorganized symptom factor in African-American participants (n = 3181) (R2 = 0.40%; P = 1.02×10e-4), successfully replicating our previous finding for European participants from the PGC study (n = 8000) (R2 = 0.39%; P = 6.39×10e-9).

    Conclusions: These analyses represent the largest empirical analysis of symptom-level clinical data for schizophrenia to date. In polygenic risk scoring analyses, we generate trans-ancestry replication support for an association between negative/disorganized symptom severity and increased polygenic burden, showing a similar relationship between a dimensional construct and polygenic risk in African-American and European participants with schizophrenia Ongoing analyses seek to further parse associations between clinical heterogeneity and polygenic risk of schizophrenia, bipolar disorder, and depression.

    Keywords: DNA, Whole-Genome, Sequencing, Schizophrenia, Negative Symptoms, Diversity, Clinical Subtypes

    Disclosure: Nothing to disclose.

    M203. Polygenic Risk Scores and Cognitive Performance Before and After Antipsychotic Treatment in First Episode Psychosis

    Jeffrey Bishop*, Bin Guo, Scot Hill, James Stevenson, Leah Rubin, Sarah Keedy, Rebekka Lencer, James Reilly, Konasale Prasad, Vishwajit Nimgaonkar, Matcheri Keshavan, Baolin Wu, John Sweeney

    University of Minnesota, Minneapolis, Minnesota, United States

    Background: Cognitive performance before and after antipsychotic treatment in psychotic disorders vary across individuals. Untreated patients early in their course of disease with little prior treatment exposure studied before and after treatment may be useful in identifying genetic and pharmacogenetic relationships.

    Methods: We examined 95 untreated first episode patients (65% male, age 23.5+/- 6.9, schizophrenia n=69 bipolar disorder with psychotic features n=16 and, major depressive disorder with psychotic features n=10) who had little to no prior antipsychotic exposure in a 6-week pharmacogenomic study of antipsychotic treatment response. Risperidone was the preferred antipsychotic (n=67) with others chosen as secondary options when clinically preferred. Participants were genotyped using the Affymetrix Genome-Wide Human SNP Array 6.0 followed by imputation using the 1000Genomes reference panel. Polygenic risk scores (PRS) were computed based on the most recent GWAS summary data for schizophrenia (SCZ), major depressive disorder (MDD) and bipolar disorder (BD) from the Psychiatric Genomics Consortium (PGC). We divided the data into two groups for analysis (European n=46 and Non-European n=49) and calculated the PRS for SCZ, MDD and BD under 10 different p-values thresholds (p = 0.5, 0.1, 0.05, 0.01, 10-3, 10-4, 10-5, 10-6, 10-7, 5*10-8). Cognitive performance was assessed with a neuropsychological battery including the California Verbal Learning Test, Digits forward/back, digit symbol, Trails A/B, and verbal fluency. A composite cognition z-score (CogZ) was computed for each participant using performance on these measures relative to demographically matched controls. CogZ before treatment and changes after 6 weeks of antipsychotic therapy were examined as quantitative traits. We then fitted models with PRS, using the largest R^2 across different p-value cutoffs for each disease state. PRS were modeled individually and together along with genetic ancestry and dose (posttreatment only) in relation to CogZ before and after 6 weeks of antipsychotic treatment. The significance of R^2 was calculated based on 1000 permutations for each outcome.

    Results: PRS for SCZ, but not MDD or BD, were negative predictors of cognitive performance before treatment (p < 0.01), but not change after 6 weeks of antipsychotic therapy in Caucasians. SCZ PRS accounted for a maximum of 25% of the variance in CogZ before treatment in Caucasians using the Caucasian-derived PRS (permutation p<0.001), but only 2% of the variance in non-Caucasians (p=0.90). Collectively examining SCZ+MDD+BD PRS in relation to CogZ did not significantly improve prediction models compared to SCZ alone.

    Conclusions: To our knowledge this is the first examination of PRS for psychiatric illnesses in relation to cognitive performance in untreated first episode psychosis. Our findings indicated that higher SCZ PRS were strongly and negatively associated with cognition before treatment are important given the relatively lower or absence of association of PRS with cognition in schizophrenia in previous studies of chronically treated patients. This may indicate that genetic factors are important determinants of cognition early in illness while the relative importance of disease progression or treatment may increase as illness progresses to reduce cognition-genetics associations. Additionally, that the genetic determinants of cognition change after treatment may differ from those related to disease risk. The lack of association of PRS with cognition in non-Caucasian participants underscores the racial specificity of these PRS measures which were developed from large Caucasian study cohorts.

    Keywords: Antipsychotic-Naïve First-Episode Schizophrenia, Polygenic Risk Score, Treatment-Response

    Disclosure: Nothing to disclose.

    M204. Identification and Prioritization of Gene Sets Associated With Schizophrenia Risk by Co-Expression Network Analysis in Human Brain

    Abstract not included.

    M205. Identification of a Novel Neuropeptide in Mammalian Brain Using a Ribosome Profiling and Peptidomic Approach

    Ravi Tharakan, Simion Kreimer, Ceereena Ubaida Mohien, Joelle Lavoie, Robert Cole, Koko Ishizuka, Akira Sawa, Leslie Nucifora*

    Johns Hopkins School of Medicine, Batimore, Maryland, United States

    Background: Intracellular communication in the central nervous system (CNS) is chemically mediated. Neurotransmitters composed of amino acids and their metabolites were first identified as these chemical communication agents. Later, evidence emerged that numerous peptides also serve as messengers in the brain. Neuropeptides are now recognized as the largest and most diverse signaling molecule in the CNS.

    Neuropeptides are 3 – 100 amino acids residues in length and act as neurotransmitters, neuromodulators, hormones or growth factors. They have been implicated with a myriad of physiological and behavioral processes. The antagonist to one of the first neuropeptides discovered, neurokinins A and B, has been approved by the Food and Drug Administration for treatment of chemotherapy-induced emesis and has also shown clinical efficacy in the treatment of major depression. This exemplifies the important role of neuropeptides and their receptors with respect to psychiatric disease and the potential for new therapies.

    Although neuropeptides possess increased structural complexity, binding affinity, and selectivity compared to the smaller classical neurotransmitters, their concentration levels are much lower in the brain. This makes detection of novel neuropeptides using traditional proteomic methods developed for the detection of larger globular proteins challenging. Continued neuropeptide discovery necessitates the development of proteomic methods with increased sensitivity.

    Methods: Peptide extraction was performed on whole brain from black 6 male mice. Capture and detection of extracted peptides was achieved by integration of an online strong cation exchange (SCX) separation into a nanoRPLC-MS/MS system. A combined approach utilizing ribosome profiling method and mass spectrometry-based peptidomics were utilized for the identification of the novel peptide. TMT labeling was used to detect peptide abundance over the course of mouse adolescence.

    Results: In this study, we introduce and validate a new methodology for neuropepetide discovery in mammalian brain. This methodology utilizes a combined ribosome profiling and mass spectrometry-based peptidomics approach with sufficient sensitivity to observe low abundance signaling neuropeptides. Using this framework, we have identified a novel neuropeptide in mouse whole brain. The peptide was also detected in two independent cultured brain cells confirming that the peptide is derived from brain and not contamination of peripheral blood. Insignificant changes in expression levels of this neuropeptide was observed over the course of adolescence. An examination of the expression of the underlying transcript using the BioGPS database, indicates highest expression of this transcript in mouse brain when compared to other tissues. The highest expressing brain regions include the dorsal root ganglia, hypothalamus, and amygdala. The hypothalamus and amygdala have roles in the regulation of behavior, suggesting a potential functional role for this peptide as a behavior regulating neuropeptide.

    Conclusions: This study reports a novel neuropeptide in mammalian brain that has been identified utilizing a combined ribosome profiling and peptidomic approach. Although we have yet to identify its functional role, it has been previously detected in peripheral tissue and serves as a component of the mRNA decapping complex (D’Lima et al., 2016). Continued efforts into neuropeptide discovery and the understanding of their functional roles in mammalian brain may provide insight into psychiatric disease.

    Keywords: Neuropeptides, Ribosome Profiling, Peptidomics, Mass Spectrometry

    Disclosure: Nothing to disclose.

    M206. Indices of Neuro-Inflammation in the Superior Temporal Gyrus in Chronic Schizophrenia

    Margaret Niznikiewicz*, Elsiabetta DelRe, Kevin Spencer, Hiu Liao, Simone Merugulama, Paul Nestor, Robert McCarley, Nicolas Bolo, Alexander Lin

    Harvard Medical School, Brockton, Massachusetts, United States

    Background: Schizophrenia (SZ) is associated with both glial abnormalities and with neuro-inflammation. In the current study, we have used resting state magnetic resonance spectroscopy (rs-MRS) to investigate myo-inositol, a marker of glial proliferation, choline, a marker of demyelination, and glutathione, a marker of neuro-inflammation, in the three voxels of left and right superior temporal gyrus (STG), given its involvement in schizophrenia pathology and medial anterior cingulate cortex (ACC), given its role in SZ symptomatology and prior reports of neuro-metabolite abnormalities in this region in first episode SZ (FESZ). While metabolites in the medial prefrontal cortex, ACC and medial temporal gyrus have been examined in both chronic SZ (CSZ) and FESZ, we are not aware of published reports of metabolic abnormalities in the STG in either CSZ or FESZ. Based on existing literature, no group differences were predicted in CSZ in ACC, and higher levels of glutathione (GSH), myo-inositol (mI) and choline (Cho) were predicted in the STG in CSZ group as reflective of neuroinflammation.

    Methods: Fourteen CSZ patients and 14 healthy controls (HC) matched for age and gender were tested. Myo-inositol, choline and glutathione levels were acquired using short echo spectroscopy (TR/TE=2000/30, PRESS, 128 avgs, 24 cm3) at 3 Tesla (Siemens Skyra, 32 channel head coil). All metabolite values were corrected for the intra-voxel content (white and gray matter and cerebrospinal fluid) for each voxel and for each subject individually and metabolite concentrations were obtained (LCmodel). For all metabolites including GSH, were fitted with CRLB <20%. ANOVA was used to examine myo-inositol, choline, and glutathione in the left and right STG with group as a between factor, and region (left and right STG) as a within factor. One-way ANOVA was used to examine the metabolite levels in the ACC. In addition, correlations were conducted between white matter volume for each voxel and each metabolite in both subject groups. All analyses were corrected for age.

    Results: Group differences were observed in the STG bilaterally but not in the ACC for glutathione (p=0.037), myo-inositol (p=0.035) and choline (p=0.006). All three metabolite levels were higher in the CSZ group relative to HC group. Both myo-inositol and choline in the left STG were negatively correlated with the white matter volume in the left STG in HC but not in the CSZ group (myo-inositol: HC: r=-0.62, p=0.018; CSZ: r=-0.29, p=0.36; choline: HC: r=-0.65, p=0.012; CSZ: r=-0.3, p=0.32).

    Conclusions: These results suggest that both neuroinflammatory processes and microstructural glial abnormalities were present in CSZ in the STG voxel while they were not evident in the ACC. The lack of correlations of both choline and myo-inositol with the STG white matter underscores the presence of neuro-pathological processes in the STG in the CSZ group. The findings of the presence of the neuro-inflammatory and glial integrity marker abnormalities in the left STG offer novel evidence for the role of this brain region in the pathology of schizophrenia. Differences in the profile of abnormalities in neuro-metabolites dependent on the stage of illness, and the voxel examined, have been reported in several studies focusing on FESZ and CSZ (Jellen et al., 2018). The absence of group differences in the ACC in this clinical group is in line with previously published studies (Jellen et al., 2018) and suggests that the profile of abnormalities in neuro-metabolites may dependent on the stage of illness and the voxel examined.

    Keywords: Neurochemistry, Schizophrenia, Choline, Myo-Inositol, STG

    Disclosure: Nothing to disclose.

    M207. Neural and Behavioral Effects of Oxytocin Administration During Theory of Mind in Schizophrenia and Matched Controls

    Lize DeCoster, Lisa Lin, Joshua Woolley*, Daniel Mathalon

    San Francisco Veterans Affairs Medical Center, San Francisco, California, United States

    Background: Social cognitive impairments, including theory of mind (ToM), in schizophrenia more strongly predict functional outcomes than psychotic symptoms or non-social cognitive deficits. Despite their clinical importance, current medications do not improve these deficits. The current study investigated the hypothesis that oxytocin, a neuropeptide implicated in social behavior, would normalize neural abnormalities in schizophrenia during ToM, and that this normalization would mediate improvement in ToM behavior.

    Methods: In this cross-over, double-blind, placebo-controlled functional magnetic resonance imaging study, a single dose of 40 IU of oxytocin was administered via nasal spray to male schizophrenia patients (n = 23) and healthy controls (n = 25). Participants completed two ToM tasks in the scanner, the False belief and Person description tasks.

    Results: During both tasks, on placebo day, schizophrenia was associated with reduced accuracy, hypo-activity in the rTPJ, and hypo-connectivity between the rTPJ and medial prefrontal cortex (mPFC) compared to healthy controls. Oxytocin, relative to placebo, significantly increased accuracy (Belief I FBT: M=75.54%, SD=14.29%; Thought in PDT: M=82.95%, SD=7.98%) and rTPJ activation for ToM but not control stories in schizophrenia. Furthermore, a significant positive correlation was found between oxytocin-induced increases in rTPJ activity and accuracy (Belief in FBT: r-=.44, p=0.04). Oxytocin also significantly improved connectivity between rTPJ and mPFC in schizophrenia.

    Conclusions: Schizophrenia is associated with reduced rTPJ activity during ToM. Oxytocin improved rTPJ activity in schizophrenia during ToM, and this improvement predicted behavioral improvement. rTPJ activity during ToM might be a potential neural target for the treatment of social cognitive deficits in schizophrenia.

    Keywords: Theory of Mind, Oxytocin, Schizophrenia, Functional Magnetic Resonance Imaging, Right Temporo-Parietal Junction

    Disclosure: Nothing to disclose.

    M208. Striatal Connectivity in Breakthrough Psychosis on Antipsychotic Maintenance: Preliminary Results From the BAMM Study

    Jose Rubio*, John Kane, Christina Fales, Todd Lencz, Anita Barber, Anil Malhotra

    Feinstein Institute for Medical Research, Glen Oaks, New York, United States

    Background: Despite antipsychotic drugs being highly efficacious in preventing relapse in psychosis, (1) approximately 15% of treatment adherent patients do relapse during the maintenance phase. (2) This phenomenon is known as “breakthrough psychosis on antipsychotic maintenance medication” (BAMM). (3) The biological mechanisms of psychotic relapse in treatment adherent individuals are poorly understood.

    Recent data from our own and other groups indicate that individuals with greater treatment response to antipsychotics demonstrate lower striatal functional connectivity before treatment onset as compared with poor responders. Furthermore, symptom improvement has been associated with increases in specific striatal connections (4-6). This literature suggests that the striatum would be involved in the pathophysiology of psychosis responsive to antipsychotics, whereas poor responders could have extra-striatal pathophysiology.

    The purpose of the “BAMM Study” is to generate hypotheses on the striatal connectivity status for individuals with psychosis who have a history of response to antipsychotics yet fail to maintain such response despite confirmed continuous treatment with long acting injectable (LAI) antipsychotics.

    Methods: The BAMM study aims to compare the functional connectivity between subjects with BAMM and antipsychotic non-adherent controls using resting state functional magnetic resonance imaging. The BAMM group consists of subjects with a psychotic disorder treated with LAI antipsychotics and history of clinical response to that trial confirmed by collateral, that present with acute psychotic symptoms at the time of the scan (defined as ≥4 in BPRS in at least one of the psychotic items). The control group consists of subjects with a psychotic disorder and history of good treatment response to antipsychotics who are non-adherent yet acutely psychotic (≥4 in BPRS in at least one of the psychotic items) at the time of the scan. Drug adherence status is confirmed with antipsychotic serum level determination, and history of previous treatment resistance or clozapine treatment excludes subjects from study participation. For these preliminary analyses we studied group differences in connectivity maps of striatal seeds as in Di Martino et al. (7). For this, data were preprocessed using the Human Connectome Project pipeline and cleaned with the ICA-FIX procedure, which removed motion artifacts and nuisance variables (white matter and CSF signal), but not global signal. Outputs from this procedure underwent high pass (0.01 Hz), and low pass (0.01 Hz) filtering. Time-series for each striatal seed were extracted and correlated with those of voxels in the whole brain. The resulting z-transformed maps for each subject were then used in the group analysis in FSL. Given the exploratory nature of these analyses, we used very low thresholds (voxelwise z=1.64, with cluster threshold p=0.10). Two-sample t-tests were run to compare group differences in clinical variables.

    Results: Within 5 months, analyzable imaging data for 8 BAMM and 9 control subjects has been collected. The mean age of this sample was 39.4 (SD=13.3) years in the BAMM group and 35.4 (11.6) in the control group. Two subjects (25%) were male in the BAMM group compared with 3 (33.3%) in the control group. The number of previous hospitalizations was 4.9 (SD=3.0) in the BAMM group, compared with 3.3 (SD=1.9) in the control group. Mean duration of illness was 10.7 (SD=7.0) years for the BAMM group compared with 8.0 (SD=4.3) for the control group. The clinical global impression severity score was 4.75 (SD=1.1) in the BAMM group and 4.9 (SD=0.6) in the control group. The BPRS total score was 43.0 (SD=6.5) for the BAMM group, and 42.7 (SD=6.2) for the control group, while the psychotic sub-component score (sum of score of the items “unusual thought content”, “hallucinations”, and “conceptual disorganization”) was 11.5 (SD=2.7) and 14.9 (SD=2.9) respectively. The BAMM group had been treated in average over a year with a LAI antipsychotic with mean dose of 1.4 (SD=0.6) defined daily doses (standard defined daily dose=1) and had received their last dose a mean of 11.0 (SD=7.3) days before the scan. No group differences were found in the clinical variables between groups. No group differences were found in the functional connectivity of any of the 6 bilateral striatal seeds compared with the rest of the brain between the BAMM and the control group.

    Conclusions: The preliminary analyses of the first 17 participants of the BAMM study did not find group differences in striatal connectivity between subjects with confirmed antipsychotic adherence who relapsed versus controls who were non-adherent with antipsychotics at the time of relapse. If these results stand through the completion of the study, they will serve to motivate the hypothesis that the pathophysiology of relapse in antipsychotic adherent individuals is mediated through extra-striatal mechanisms. However, given the limited power to detect group differences with this sample, further collection of data is required.

    Keywords: Psychosis, Relapse Biomarkers, Medication Adherence, Striatal Pathways, Resting State Functional Connectivity

    Disclosure: Lundbeck, Consultant

    M209. Activation of Internal Performance Monitoring Circuitry is Reduced in Psychosis Spectrum Youth

    Tess Levinson, Greer Prettyman, Theodore Satterthwaite, Lauren White, Tyler Moore, Monica Calkins, Kosha Ruparel, Raquel Gur, Ruben Gur, Daniel Wolf*

    University of Pennsylvania, Wynnewood, Pennsylvania, United States

    Background: Self-directed performance monitoring is a critical contributor to cognitive performance and general functioning and is impacted by a variety of psychiatric symptoms and personality traits. We have previously shown that ventral striatum (VS) activates more strongly to correct than incorrect responses during cognitive tasks where no explicit feedback is required, and that this “intrinsic reinforcement” response is reduced in schizophrenia. Here we examined this phenomenon in youth from the Philadelphia Neurodevelopmental Cohort (PNC) performing a working memory fMRI task and related it to a reaction time measure of performance monitoring and to psychosis spectrum status. We hypothesized that VS would respond to internal correctness monitoring while classic “salience network” regions such as dorsal anterior cingulate cortex (ACC) and anterior insular cortex (AIC) would reflect internal error monitoring. We expected that both behavioral and imaging measures of performance monitoring would be reduced in youth with subclinical psychosis spectrum features.

    Methods: A sample of 951 youths ages 8-21 from the PNC (227 typically developing, TD; 285 psychosis spectrum, PS) were included. 3 T fMRI BOLD data from the n-back working memory task was examined, focusing on the hardest (2-back) level and comparing fMRI response to correct vs. incorrect trials. fMRI analysis focused on a priori ROIs in VS, ACC and AIC, along with whole-brain voxelwise analyses. Behavioral performance monitoring was assessed in a separate out-of-scanner task without feedback (progressive matrices, PMAT), by correlating across trials within each participant the item-wise reaction time (RT) and difficulty (based on item-response theory analysis).

    Results: Across the full sample, a strikingly selective pattern was observed for correct vs. incorrect trials, despite the fact that no accuracy information was provided during the task. As expected, VS showed robust correct>incorrect responses (peak Z=10.6), with positive responses to correct and negative responses to incorrect. In contrast, ACC and AIC showed the opposite pattern with robust incorrect>correct differential activation (peak Z’s: ACC -13.9, AIC -14.2), although activating to both correct and incorrect trials. This accuracy monitoring pattern did not simply reflect greater task activation on correct trials: VS was not robustly task activated, task-active regions ACC and AIC were more activated for errors, and other strongly task-activated working memory regions such as dorsolateral prefrontal cortex and posterior parietal cortex did not show robust differences between correct and incorrect trials. Correct-vs-incorrect responses in VS were only weakly correlated with the ACC and AIC responses (r=-0.05), suggesting distinguishable circuitry for correctness vs. error processing. In the out-of-scanner PMAT, RT was slower for items that were more difficult providing evidence of internal performance monitoring; the average RT-difficulty correlation was r=0.39 across all participants. This RT-difficulty correlation measure was significantly correlated with VS correct>incorrect fMRI responses (r=0.16, p<.0001). This correlation was stronger than that seen for overall performance on the out-of-scanner cognitive battery (r=.09), slightly stronger than the relationship seen for accuracy in the fMRI working memory task itself (r=.14), and not accounted for by either of these. Older age was associated with both greater VS correct>incorrect activation and with greater PMAT performance monitoring but did not explain their association. Compared to TD youth, PS showed reduced correct>incorrect VS responses, and reduced AIC incorrect>correct responses (cluster-corrected p <0.05 within regions showing differential correct vs. incorrect response), without significant group differences in ACC incorrect>correct activation. Across all participants, there was a small but significant correlation between reduced VS correct>incorrect activation and greater negative/disorganized symptoms (r=-0.07, p=0.04), and trends for externalizing symptoms and global psychopathology (r’s=0.06, p’s=0.1); no specific symptom correlations were significant in the PS group alone. In a subset of PS (n=81) with self-report data for subjective experience in the cognitive battery, there was a significant correlation between VS correct>incorrect activation and caring about doing well on the tasks and finding them interesting (r’s=0.27,0.28; p’s=0.01).

    Conclusions: We provide evidence for both error monitoring and correctness monitoring involving distinct brain regions, occurring in the absence of accuracy feedback. We further show a reduction of this internally-generated feedback or monitoring response in psychosis spectrum youth. This impairment likely has multiple contributors, including various symptoms associated with particular disorders as well as trans-diagnostic characteristics such as low motivation and impulsivity. Further work is needed to disentangle these contributors and develop a better understanding of the bidirectional relationship between cognitive performance and aspects of psychiatric syndromes or personality that are not typically considered “cognitive”. Addressing such features through pharmacological or non-pharmacological interventions could facilitate cognitive rehabilitation and efforts to improve social, academic and occupational function.

    Keywords: Psychosis Risk, Functional MRI (fMRI), Ventral Striatum, Error Processing

    Disclosure: Nothing to disclose.

    M210. Neurite Orientation Dispersion and Free Water Alterations in Unmedicated Patients With Schizophrenia are Not Affected by Short Term Treatment With Risperidone

    Nina Kraguljac*, Thomas Anthony, Frank Skidmore, William Monroe, Jon Marstrander, David White, Meredith Reid, Adrienne Lahti

    University of Alabama at Birmingham, Birmingham, Alabama, United States

    Background: Advanced diffusion imaging analytics are designed to provide more specific characterization of white matter microstructure, but only few studies have taken advantage of these techniques in schizophrenia. Results are suggestive of altered neurite morphology and neuro-inflammation as pathophysiological processes in the illness, but because all studies enrolled medicated patients, it is not clear to which extent these alterations can be attributed to antipsychotic medication effects as opposed to intrinsic illness characteristics.

    The goal of this study was to apply NODDI to investigate white matter neurite configuration and extracellular free water content in unmedicated patients with schizophrenia and to assess the effects of six weeks treatment with risperidone on these measures.

    Methods: We enrolled 42 unmedicated patients (thirty were medication-naïve) with schizophrenia in a longitudinal trial with risperidone. Symptom severity was assessed with the Brief Psychiatric Rating Scale (BPRS). We obtained diffusion weighted images before medication was started, and after six weeks of treatment. Healthy controls matched 1:1 on age, gender, and parental socioeconomic status were also scanned twice six weeks apart. 30 diffusion sampling directions spanning the whole sphere were acquired twice and concatenated (in plane resolution 2.2 mm, slice thickness 2.2 mm, b-value 1000 s/mm2, 5 b0 images). After visual inspection of raw images we used TORTOISE for correction of bulk motion, eddy currents and susceptibility artifacts using a single interpolation step in DIFF_PREP. Gradient tables were rotated along with motion correction. To obtain a summary measure of motion, the root-mean-square (RMS) was calculated both for absolute (RMSabs) and relative (RMSrel) movement. For estimation of extracellular free water and orientation dispersion index maps, we used the NODDI toolbox. To spatially normalize diffusion images, we implemented an optimized non-linear image registration using a modified version of 3dQwarp in AFNI. To assess whole brain voxel-wise group differences and changes over time in diffusion indices used AFNI’s 3dttest++ (age, sex, and RMSrel as covariates) with clustsim, a bootstrapping method used to correct for multiple comparisons.

    Results: Mean age of patients was 26.62 years, 62% of subjects were male. Of the 42 patients included here, 33 completed the study. At baseline, voxelwise analyses demonstrated decreased orientation dispersion in the posterior limb of the internal capsule, and whole brain analysis revealed a trend level increase extracellular free water in patients compared to controls. Longitudinal analyses showed no changes in whole brain extracellular free water or orientation dispersion in healthy controls over time, and no changes in patients after six weeks of treatment with risperidone.

    Conclusions: Our results demonstrate circumscribed fiber geometry alterations in the posterior limb of the internal capsule and a trend-level whole brain extracellular free water increase in unmedicated patients, indicating that reported alterations in advanced diffusion indices are not merely a confound of antipsychotic treatment. Furthermore, we found no overall changes of diffusion indices after treatment but did report a positive relationship between change in extracellular free water and medication dose at endpoint.

    Keywords: Diffusion Weighted Imaging, Schizophrenia, Antipsychotic Treatment

    Disclosure: Nothing to disclose.

    M211. Dopamine Receptor Density and White Mater Integrity: 18F-Fallypride Positron Emission Tomography and Diffusion Tensor Imaging Study in Healthy and Schizophrenia Subjects

    Monte Buchsbaum*, Serge Mitelman, Bradley Christian, Brian Merrill, Bradley Buchsbaum, Jogeshwar Mukherjee, Douglas Lehrer

    University of California, San Diego, San Diego, California, United States

    Background: Dopaminergic dysfunction and changes in white matter integrity are among the most replicated findings in schizophrenia. A modulating role of dopamine in myelin formation and white matter integrity has been proposed in animal models and healthy human brain but has not yet been systematically explored in schizophrenia.

    Methods: We used diffusion tensor imaging and 18F-fallypride positron emission tomography in 19 healthy and 25 never-medicated schizophrenia subjects to assess the relationship between gray matter dopamine D2/D3 receptor Binding Potential and white matter fractional anisotropy in each diagnostic group. MANOVA comparisons of the groups have previously been published. Patients were neuroleptic naïve or almost neuroleptic naïve, negative for drugs of abuse on urine screen on the scan day. The groups were approximately matched on handedness (16/19 right handed in the healthy group and 23/25 in the patient group (chi-square 0.65, p=0.41) One patient had an earlier diagnosis of alcohol abuse and patients with other comorbidity were excluded. Analysis of Functional NeuroImages (AFNI) regions of interest were acquired for 42 cortical Brodmann areas and subcortical gray matter structures, as well as stereotaxically placed in representative white matter areas implicated in schizophrenia neuroimaging literature. Four separate frame images were collected for fallypride and the odd and even frames separately quantified using the same algorithm in a second sample to confirm highly significant reliability (0.8-0.9 correlations) for cortical and subcortical regions.

    Results: Healthy subjects displayed an extensive pattern of predominantly negative correlations between 18F-fallypride binding across a range of cortical and subcortical gray matter regions and fractional anisotropy in rostral white matter regions (internal capsule, frontal lobe, anterior corpus callosum). Positive correlations tended to cluster at the dorsal white matter regions closer to the cortex and predominated only in the temporal white matter. These patterns were disrupted in subjects with schizophrenia, who displayed significantly weaker overall correlations, as well as comparatively scant numbers of significant correlations with the internal capsule and frontal (but not temporal) white matter, especially for dopamine receptor density in thalamic nuclei. Permutational analysis was used to confirm correlation coefficient matrix differences between groups.

    Conclusions: Dopamine D2/D3 receptor density and white matter integrity appear to be interrelated phenomena, and their decreases in schizophrenia may stem from dysregulation of dopaminergic impact on axonal myelination.

    Keywords: Positron Emission Tomography Imaging, dopamine, fluorodeoxyglucose

    Disclosure: Nothing to disclose.

    M212. Reduced Stress-Induced Dopamine Release in Medial Prefrontal Cortex in Chronic Cannabis Users at Clinical High Risk for Psychosis

    Christin Schifani*, Jens Pruessner, Huai-Hsuan Tseng, Naren Rao, Abanti Tagore, Alan A. Wilson, Sylvain Houle, Pablo M. Rusjan, Romina Mizrahi

    Centre for Addiction and Mental Health, Toronto, Canada

    Background: Dopamine hyperactivity in striatum is associated with psychotic symptoms and both stress and cannabis are risk factors for the development and relapse to psychosis. We have previously reported that subjects at clinical high risk for psychosis (CHR) exhibit a higher striatal dopamine response to acute stress compared to healthy volunteers and that chronic cannabis use blunts this response. However, it is unknown if this abnormal dopamine response to stress and cannabis extends to the prefrontal cortex (PFC), a brain region involved in regulating cognition and stress response, and how it relates to symptoms and cannabis use history. Therefore, we investigated the effect of acute psychosocial stress on PFC dopamine signaling in CHR with and without chronic cannabis use using [11 C]FLB457 positron emission tomography (PET) and a validated stress challenge.

    Methods: Thirty-three participants matching in demographics completed two [11 C]FLB457 PET scans (14 CHR without cannabis use, 8 CHR chronic cannabis users (CHR-CU) and 11 healthy volunteers), one while performing a Sensory Motor Control Task (control) and another while performing the Montreal Imaging Stress Task (stress). Binding potential (BPnd) was estimated using the Simplified Reference Tissue Model. Dopamine release was defined as percent change in BPnd between control and stress scans (ΔBPnd). To assess the physiological stress response, saliva samples were taken throughout the scans and cortisol levels (AUCi) determined. The salivary cortisol response was defined as difference in AUCi between control and stress scans (ΔAUCi). Furthermore, attenuated psychotic symptoms were assessed before and after each PET scan session using an abridged version of the Scale of Prodromal Symptoms (SOPS). Drug history was assessed using a semi-structured interview.

    Results: Our preliminary analysis revealed that stress-induced dopamine release (ΔBPnd) was significantly different between groups in medial PFC (F(2,30)=5.40, p=0.010) with CHR-CU exhibiting lower ΔBPnd compared to CHR (Bonferroni-corrected p=0.008) but not compared to healthy volunteers (Bonferroni-corrected p=0.29). Similarly, salivary cortisol response was significantly different between groups (F(2,29)=5.08, p=0.013) with CHR-CU exhibiting lower ΔAUCi as compared to CHR (Bonferroni-corrected p=0.018) but not compared to healthy volunteers (Bonferroni-corrected p=1.0). Furthermore, CHR-CU participants were observed to have increased attenuated psychotic symptoms following the stress task as compared to their score before the stress task (paired t-test; t=4.58, df=7, p=0.0025). The attenuated psychotic symptoms following the stress task were negatively associated with ΔBPnd in medial PFC (ß=-1.49, SE=0.71, F(1,18)=4.44, p=0.049) and dorsolateral PFC (ß=-1.70, SE=0.59, F(1,18)=8.43, p=0.009) in the combined high-risk sample (CHR and CHR-CU) when controlling for group, suggesting that high-risk participants with more severe attenuated psychotic symptoms had lower PFC dopamine release in response to the stress task. Moreover, in CHR-CU participants, length of chronic cannabis use was negatively associated with stress-induced dopamine release in medial PFC, when controlling for current weekly cannabis use (β=-0.19, SE=0.57, F(1,5)=10.89, p=0.021), suggesting that participants with longer chronic cannabis use had lower medial PFC dopamine release in response to the stress task.

    Conclusions: These findings provide first evidence of an altered response to stress in PFC dopamine signaling in CHR cannabis users which is directly related to attenuated psychotic symptoms and cannabis use history. This study is especially relevant in the light of the global trend to legalize cannabis, as it highlights the effects of chronic cannabis use on cortical dopamine function in high-risk youth.

    Keywords: Cannabis, Positron Emission Tomography, Dopamine, Prefrontal Cortex, Clinical High Risk for Psychosis

    Disclosure: Nothing to disclose.

    M213. Cerebellar Theta-Frequency Transcranial Pulsed Stimulation Increases Frontal Theta Oscillations in Patients With Schizophrenia

    Krystal Parker*

    University of Iowa, Iowa City, Iowa, United States

    Background: Cognitive dysfunction is a pervasive and disabling aspect of schizophrenia without adequate treatments. A recognized correlate to cognitive dysfunction in schizophrenia is attenuated frontal theta oscillations. Neuromodulation to normalize these frontal rhythms represents a potential novel therapeutic strategy. Here, we evaluate whether noninvasive neuromodulation of the cerebellum in patients with schizophrenia can enhance frontal theta oscillations with the future goal of targeting the cerebellum as a possible therapy for cognitive dysfunction in schizophrenia

    Methods: Using transcranial pulsed current stimulation (tPCS), a non-invasive transcranial direct current that can be delivered in a frequency specific manner, we targeted the midline cerebellum with a single 20-minute session of theta frequency stimulation in 9 patients with schizophrenia (cathode on right shoulder with delta frequency as a control for frequency specific effects). EEG signals from midfrontal electrode Cz were analyzed before and after cerebellar tPCS while patients estimated the passage of 3 and 12 second intervals.

    Results: Theta oscillations were significantly larger following theta frequency cerebellar tPCS in the midfrontal region which was not seen with delta frequency stimulation. As previously reported, patients with schizophrenia showed a reduction in accuracy estimating 3 and 12 second intervals as compared to control subjects. Timing accuracy was not modified by theta or delta frequency cerebellar tPCS.

    Conclusions: These results suggest that theta frequency cerebellar tPCS may modulate task-related oscillatory activity in the frontal cortex in a frequency-specific manner. This finding warrants further investigation to evaluate whether a full treatment course of cerebellar stimulation may impact cognitive performance and thus have therapeutic implications for schizophrenia.

    Keywords: Cerebellum, Transcranial Direct Current Stimulation, Medial Prefrontal Cortex

    Disclosure: Nothing to disclose.

    M214. The Effect of Ethnicity and Immigration on Treatment Resistance in Schizophrenia

    Abstract not included.

    M215. Preclinical and Early Clinical Pharmacological Profile of Basmisanil, a GABA-A α5 Receptor Negative Allosteric Modulator, Currently in a Phase 2 Clinical Trial for Cognitive Impairment Associated With Schizophrenia

    Maria-Clemencia Hernandez, Jörg Hipp, Michael Hoener, Robert Comley, Michael Derks, Frederic Knoflach, Sian Lennon-Chrimes, Nicholas Seneca, Gerhard Trube, Theresa Ballard-Yardy, Rodolfo Gasser, Tania Wallace, Eric Prinssen, Daniel Umbricht*

    F. Hoffmann-La Roche Ltd, Basel, Switzerland

    Background: Genetic and pharmacological studies have demonstrated the importance of the GABA-A alpha5 subunit-containing receptors in cognitive processes. Here we describe the preclinical and early clinical pharmacological profile of basmisanil, a potent and highly selective negative allosteric modulator of GABA-A alpha5 receptors, which currently is in Phase 2 clinical development for the treatment of cognitive impairment associated with schizophrenia.

    Methods: Radioligand binding and voltage-clamp electrophysiology experiments were conducted in vitro on GABA-A receptors expressed in HEK293 cells and Xenopus oocytes to demonstrate binding and functional selectivity for the GABA-A alpha5 vs. alpha1/2/3 subunit-containing receptors. Cognitive effects were assessed in rats and monkeys using the Morris water maze and the object retrieval task, respectively. Pro-convulsant and anxiogenic potentials were evaluated in mice and rats. In vivo receptor occupancy was determined using [3 H]-RO0154513 in rats and [11 C] RO0154513 in a PET study conducted in 10 healthy volunteers at 3 timepoints: baseline, 3, and 9 h following 1 of 4 doses of basmisanil (2×15, 2×60, 3×130, 3×1250 mg). A separate EEG study in 12 volunteers measured at baseline, midazolam (5 mg), and 14 days of basmisanil treatment (240 mg, bid).

    Results: Basmisanil bound to recombinant human GABA-A alpha5 receptors with 5 nM affinity and with more than 90-fold selectivity versus alpha1, alpha2, and alpha3 subunit-containing receptors. At saturating binding concentrations, basmisanil inhibited the GABA-induced current in cells expressing GABA-A alpha5 yet had little or no effect at the other receptor subtypes. In vivo, basmisanil exhibited concentration-dependent occupancy of GABA-A alpha5 receptors and reversed diazepam-induced impairment of spatial learning of rats in the Morris water maze. Basmisanil also significantly improved performance in the object retrieval task in non-human primates. At plasma concentrations shown to improve cognition, basmisanil lacked anxiogenic and proconvulsive activity in rodent paradigms.

    In human, the PET study demonstrated basmisanil selective GABA-A a5 target engagement, mainly in temporal and frontal cortical regions, nucleus accumbens and amygdala, consistent with known expression patterns of this receptor. A tight exposure-occupancy relationship was observed, providing an excellent basis for dose selection for clinical trials. In addition, Basmisanil showed modulation of large-scale brain function reflected in characteristic changes of EEG spectral power, in particular power increase in theta to alpha, and decrease in beta frequency ranges at dose associated with ~90% occupancy. This EEG signature was qualitatively opposite to that of midazolam, a non-selective positive allosteric modulator of GABA-A receptors. Importantly, Basmisanil was safe and well tolerated, and no treatment-emergent epileptiform abnormalities were observed.

    Conclusions: These data suggest basmisanil is a promising candidate drug with a unique pharmacological and safety profile for further clinical testing in conditions associated with associated cognitive impairment such schizophrenia.

    Keywords: GABA-A receptors, Cognition, Schizophrenia, Pharmacology, PET Imaging

    Disclosure: F. Hoffmann - La Roche, Employee, Stock / Equity

    M216. Persistent Firing in Parvalbumin-Expressing Inhibitory Interneurons of the Neocortex: A Newly Discovered Property With Potential Implications for Psychopathology

    Brian Theyel*, Rachel Stevenson, Barry Connors

    Brown University, Providence, Rhode Island, United States

    Background: A basic property of axons involves their ability to propagate action potentials in both orthodromic (“forward”) and antidromic (“backward”) directions, e.g. (Brain Res Rev. 21:42, 1995). While spikes are typically generated in axon initial segments and propagate orthodromically, in certain cases spikes have been observed initiating in distal regions of axons and their terminals before propagating antidromically; these are often called ‘ectopic’ action potentials. Recently, ectopic spiking has been demonstrated primarily in NPY-expressing interneurons of the hippocampus (Nat Neurosci. 14:200, 2011), whereas few parvalbumin (PV)-expressing interneurons generated ectopic spikes.

    Methods: To investigate ectopic spiking behavior in cortical interneurons we used standard whole-cell patch clamp techniques to patch neurons in layers 2/3 or 4 of orbitofrontal or somatosensory cortex in 300 micron slices. 5 male and 17 female PV-Cre x Ai14 mice, postnatal day 20 to 35 were used. Statistical analyses are pending completion of experiments.

    Results: We have observed that a large majority (>95%) of PV+ cells in neocortex (both orbitofrontal cortex and primary somatosensory cortex) are not only capable of generating ectopic action potentials after they’ve been sufficiently activated but do so in varying patterns for up to tens of seconds. We also found that somatostatin positive interneurons and even a few pyramidal cells were capable of rare, sparse ectopic spiking, but only after intense stimulation unlikely to occur in vivo. PV+ cells, often referred to as “fast-spiking” interneurons, had a much lower initiation threshold for ectopic spikes, requiring a few hundred spikes over the course of tens of seconds, and often (~75% of the time) fired trains of ectopic action potentials rather than just one or a few.

    Conclusions: PV+ cells generate strong inhibition in excitatory cells and other interneurons, and they play a key role in generating network gamma rhythms, so ectopic spiking may have significant implications for network activity and cognitive processing. Several studies have demonstrated abnormalities in PV+ interneurons in schizophrenia (J Neurosci, 29:8, 2344-2354; Trends Neurosci, 35(1), 57-67) and autism spectrum disorder (PLoS One, 10(3), e0119258; Mol Psychiatry, 20(10), 1161-1172). Our future work will focus on whether ectopic spiking behavior changes in PV+ interneurons of animal models of both disorders.

    Keywords: GABAergic Interneurons, Cortical Circuit Function, Schizophrenia, Cortical Circuit Function, Autism, Inhibition

    Disclosure: Nothing to disclose.

    M217. Plasma Nitrite is Associated With Dyslipidemia in Medicated Obese Schizophrenia Patients

    Olaoluwa Okusaga*, Dietmar Fuchs, Teodor Postolache

    MEDVAMC/Baylor College of Medicine, Houston, Texas, United States

    Background: Obesity and cardiovascular disease (CVD) are highly prevalent in patients with schizophrenia leading to premature death, with 15–20 years shorter life expectancy than the general population. Obesity and CVD are related to antipsychotic medication use. We recently linked plasma nitrite (NO2-) a metabolite of nitric oxide (NO), to obesity and components of metabolic syndrome including dyslipidemia, in a non-psychiatric sample but this association has not been evaluated in schizophrenia. We have now assessed the relationship between plasma NO2-, body mass index (BMI) and plasma lipids in a sample of medicated patients with schizophrenia.

    Methods: Methods: 106 medicated patients with schizophrenia were recruited. Body mass index (BMI), fasting plasma NO2-, total cholesterol (TC), triglycerides (TG), low-density lipoprotein (LDL), high-density lipoprotein (HDL), neopterin (a marker of immune activation/inflammation) and cotinine (a metabolite of nicotine) were measured in all the participants. Non-HDL Cholesterol (NHDL-C) was calculated by subtracting HDL from TC. The patients were categorized into normal weight (BMI 18.5 to <25) and overweight/obese (BMI >25) groups. T-test compared plasma NO2- between the two groups. Pearson’s and partial correlations (adjusting for age, sex, race, education, cotinine, and neopterin) between NO2-, BMI and plasma lipids were calculated in the total sample and in the two patient groups (i.e. normal weight and overweight/obese) respectively.

    Results: Plasma NO2- did not differ between normal weight and overweight/obese patients (geometric mean 21.38µmol/L vs 23.98 µmol/L, p=0.398) and there was no correlation between NO2- and BMI. In the unadjusted analysis, plasma NO2- correlated positively with TC, TG, LDL and NHDL-C but negatively with HDL in the total sample but only TG correlated with NO2- in the adjusted analysis. In the unadjusted subgroup analyses, NO2- correlated with TG and NHDL-C in normal weight and overweight/obese groups but also with TC and HDL in the overweight/obese group. In the subgroup analyses, the only correlation that withstood adjustment for potential confounders was between NO2- and TG (partial r= 0.75, p=0.001) and this was only in the overweight/obese group.

    Conclusions: The observed correlation between NO2- and TG observed in the total sample is probably driven by this association in the overweight/obese group since no such association was found in the normal weight group. Elucidation of the link between antipsychotic medication-related obesity, dyslipidemia, metabolic syndrome and plasma NO2- may lead to novel treatments to reduce premature death from CVD in patients with schizophrenia.

    Keywords: Nitrites, Obesity, Dyslipidemia, Schizophrenia, Antipsychotics

    Disclosure: Nothing to disclose.

    M218. Neural Mechanisms Underlying Higher Rates of Psychotic and Mood Symptoms in Females With Schizophrenia

    Laura Tully*, Ana-Maria Iosif, Mary Blendermann, Lauren N. Zakskorn, Anna K. Moody, Bryn M. Ritter, Raisa Rahim, Amanda Guyer, Cameron Carter

    University of California, Davis School of Medicine, Sacramento, California, United States

    Background: Females with schizophrenia (SZ) report more psychotic and mood symptoms. Further, in females, psychotic symptom severity relates to mood symptom severity, indicating a link between psychosis and emotional responses not observed in SZ males. One potential mechanism is cognitive control of emotion (CCoE): the self-regulation of emotions and their influence on behavior, mediated by a frontal-limbic network involving the lateral and medial prefrontal cortices and amygdala. Sex differences in CCoE in healthy individuals have not been examined, nor have relationships to sex differences in SZ symptomatology. This study seeks to identify sex differences in the frontal-limbic network during CCoE and their relationship to symptoms in females with SZ.

    Methods: During fMRI, participants complete the Social Evaluation Task (SET), in which participants react to or down-regulate their emotional responses to negative social evaluation (NSE) and self-report their negative emotion on a 1-to-5 scale (1=not at all negative; 5=very negative). Participants also complete psychosocial symptom assessments and measures of emotion regulation (cognitive emotion regulation questionnaire [CERQ], the emotion amplification and reduction scales [TEARS]), internal attribution styles (Internal, Personal and Situational Attributions Questionnaire [IPSAQ]), and perspective taking (interpersonal reactivity index [IRI]). At time of submission, 21 healthy (11 female) and 17 SZ (10 female) participants have completed the study. We used independent sample T tests (p < 0.05 two-tailed) to conduct preliminary behavioral analyses of differences in reactivity to and regulation of NSE within group (HC females vs. HC males; SZ females vs. SZ males) and within sex (HC females vs. SZ females; HC males vs. SZ males). Effect sizes are reported as standardized mean differences.

    Results: Preliminary behavioral results on the SET suggest that compared to healthy males (n=10), healthy females (n=11) demonstrate greater reactivity to (standardized mean difference = 1.45) and greater regulation of (standardized mean difference = 1.18) NSE (ps < 0.05). Preliminary data in SZ participants were in the same direction, although differences did not reach statistical significance: Compared to SZ males (n=7), SZ females (n=10) demonstrate greater reactivity to (standardized mean difference = 0.66) but reduced regulation of NSE (standardized mean difference = -0.30) (ps > 0.05). Preliminary within-sex comparisons indicate the following: Compared to healthy females, SZ females demonstrate a non-significant pattern of reduced reactivity to (standardized mean difference = -0.32; p < 0.05) but significantly reduced regulation of NSE (standardized mean difference = -1.64; p = 0.001). Compared to healthy males, SZ males demonstrate a pattern of lower reactivity (standardized mean difference = -0.33) and similar regulation of NSE (standardize mean difference = -0.03); differences were not significant (ps < 0.05). Preliminary whole-brain fMRI results demonstrate healthy controls activate expected regions of the frontal-limbic network during CCoE, including the dorsolateral prefrontal cortex (DLPFC), medial prefrontal cortex (MPFC) (p < 0.001 uncorrected, k = 10 voxels).

    Conclusions: Females may be more reactive to NSE, and SZ females may have impaired regulation of NSE compare to healthy females and SZ males. Future analyses will examine associated frontal-limbic activity and relationships with symptoms and psychosocial measures. This work directly examines how sex as a biological variable contributes to individual differences in the symptomatology of SZ and could inform the development of sex-specific treatments.

    Keywords: Emotion Regulation, Cognitive Control, Early Psychosis

    Disclosure: Nothing to disclose.

    M219. Elevated TNFα Levels in Males but Not in Females in Cerebrospinal Fluid of Patients With Schizophrenia

    Juan Gallego*, Robert McNamara, Emily Blanco, Santiago Castaneda, Todd Lencz, Anil Malhotra

    Weill Cornell Medical College, White Plains, New York, United States

    Background: Elevated levels of pro-inflammatory cytokines have provided evidence in support of the inflammatory hypothesis of schizophrenia. Most studies of cytokines in schizophrenia have reported cytokine levels in peripheral blood and some have used cerebrospinal fluid (CSF). However, most of these studies have not examined the impact of sex on cytokine levels in detail. Prior studies with non-psychiatric patients have described a potential protective effect of estrogens in inflammatory conditions such as multiple sclerosis. Whether this effect is also present in schizophrenia is unclear, but it could partially explain better outcomes observed among females with this disorder as compared to males with schizophrenia, including a later onset of the illness and an overall better prognosis. Therefore, our aim was to study the relationship between sex and a panel of cytokines in cerebrospinal fluid of patients with schizophrenia and healthy volunteers.

    Methods: Thirty-three patients with schizophrenia-spectrum disorders and 23 healthy volunteers underwent a lumbar puncture. 15-25 cc of CSF were obtained from each subject. CSF cytokine (IL-1β, IL-2, IL-4, IL-6, IL-8, TNFα) concentrations were determined in duplicate by enzyme-linked immunosorbent assay (ELISA) and a high-sensitivity MilliplexTM Multiplex kit (HSTCMAG-28SK-06, Millipore, Billerica, MA) per manufacturer’s instructions. Raw cytokine values were log-transformed. Comparisons in cytokine levels between groups were performed using either t-tests for normally distributed variables or Wilcoxon rank-sum tests for non-normally distributed variables. Logistic regression analyses were conducted using subject type (patient vs. control) as the dependant variable and age, sex and cytokines as the independent variables.

    Results: The mean age was 36.6 years (SD=11.7) in patients and 38.1 years (SD=10.1) in controls. 24/33 (72.3%) of the patients and 14/23 (60.1%) of the healthy volunteers were male. Patients in the SSD group were clinically stable and taking antipsychotic medications. In males, levels were statistically significantly higher in patients vs. controls for TNFα (mean=6.88 pg/ml [SD=3.08] vs. 3.44 pg/ml [2.03], p=0.0004]) and IL-4 (12.21 pg/ml [8.91] vs. 6.65 pg/ml [3.98], p=0.02). There were no significant differences in IL-6 (9.65 pg/ml [2.89] vs. 8.96 pg/ml [2.09], p=0.42) or IL-8 (48.88 pg/ml [10.97] vs. 47.07 pg/ml [10.52], p=0.67). In females, none of the cytokines studied were statistically significantly different between patients and controls. Multivariate logistic regression analysis stratified by sex and adjusting by age showed that TNFα was significantly associated with schizophrenia (OR=9.7 [95% CI: 2.2-42.0], p=0.003) in the male group. In contrast to the findings from the bivariate analysis, IL-4 was dropped from the model due to a p-value >0.05. In females, there were no significant associations between cytokines and schizophrenia.

    Conclusions: TNFα, a pro-inflammatory cytokine and a key participant in the acute phase response of the inflammatory cascade, is elevated in CSF of male patients with schizophrenia but not in females. IL-4 levels were also significantly higher in patients in bivariate analysis but were not significant in multivariate regression analysis. Other cytokines studied did not show this sex effect and were not differently found in patients vs. controls despite prior studies having reported elevated levels of other proinflammatory cytokines such as IL-6.

    Keywords: Cerebrospinal Fluid, TNF-Alpha, Cytokines, Psychosis

    Disclosure: Nothing to disclose.

    M220. The Prism Project: A Quantitative Approach to Neuropsychiatry

    Celso Arango*, Amy Bilderbeck, Gerard Dawson, Brenda Penninx, Nic Van der Wee, Inge Winter-van Rossum, Bernd Sommer, Alex Schubert, Hugh Marston, Niels Jongs, Martien Kas

    Hospital Gregorio Marañon, Madrid, Spain

    Background: Current nosology for the diagnosis of neuropsychiatric disorders, such as Schizophrenia and Autism, separates each into non-overlapping diagnostic categories. This separation is not based on their underlying etiology but on convention-based clustering of qualitative symptoms of the disorder. While these diagnostic categories are sufficient to provide the basis for general clinical management, they do not describe the underlying neurobiology that gives rise to individual symptoms. The ability to precisely link these symptoms to underlying neurobiology would not only facilitate the development of better treatments, it would also allow physicians to provide patients with a better understanding of the complexities and management of their illness.

    Methods: The PRISM project focuses on schizophrenia (SZ), Alzheimer’s disease (AD), and major depression (MD), disorders that share in part common symptomatologies, including social withdrawal and certain cognitive deficits, such as attention, working memory and sensory processing. Innovative technologies (e.g. EEG, cognitive tasks, (f)MRI, smartphone monitoring, genetics and epigenetics) are used to deep phenotype clinical cohorts of SZ and AD patients. These data is then combined with existing clinical data sets from major European and global disease cohorts. The aim is to derive a set of quantifiable biological parameters from these data that allow to cluster and differentiate SZ, AD and MD patients who are, or are not, socially withdrawn.

    Results: One of PRISM objectives is to identify quantitative biological parameters that allow assessment of social withdrawal levels across different neurological and psychiatric diagnosis. Preliminary analysis, using self-rated minus researcher-rated WHODAS scores, revealed that Alzheimer (p < 0.001) and schizophrenia (p = 0.049) patients score significant lower based on the WHODAS score, suggesting that the patients are overestimating their levels of social activities. Using objective BEHAPP smartphone monitoring (Mulder et al., 2018), initial analysis indicated that Alzheimer patients spent significantly (p = 0.001) more time at home as compared to healthy controls (BEHAPP analysis was corrected for age).

    Conclusions: These results will inform future clinical trials assessing negative symptoms such as social withdrawal in different brain disorders. The use of self rated assessments does not seem to be valid based on preliminary results from our study using a digital monitoring.

    Keywords: Dimensions of Psychosis, Social Withdrawal, Alzheimer’s Disease, Endophenotypes, Diagnostic Boundaries

    Disclosure: Acadia, Advisory Board, Ambrosetti, Advisory Board, Gedeon Richter, Consultant, Janssen Cilag, Consultant, Lundbeck, Advisory Board, Otsuka, Consultant, Roche, Advisory Board, Servier, Consultant, Shire, Consultant, Schering Plough, Consultant, Sumitomo Dainippon Pharma, Advisory Board, Sunovion, Advisory Board, Takeda, Advisory Board

    M221. Deficient Hippocampal Novelty Response and Habituation in Early Psychosis

    Suzanne Avery*, Baxter Rogers, Maureen McHugo, Kristan Armstrong, Neil Woodward, Jennifer Blackford, Stephan Heckers

    Vanderbilt University Medical Center, Nashville, Tennessee, United States

    Background: Deficits in hippocampal function are among the most consistent and replicable findings in schizophrenia. There is evidence that deficits in hippocampal function also exist at the earliest stages of psychotic illness, although in a subtler form, suggesting hippocampal function may be a useful model for psychosis staging. However, functional markers sensitive to hippocampal pathology, such as interneuron loss and progressive excitatory/inhibitory imbalance, are needed. Animal models have linked novelty response and habituation of neural signal—the decrease in response to repeated sensory stimuli—with neural excitatory/inhibitory balance, suggesting the temporal dynamics of hippocampal response may be a sensitive marker of neuropathology. We previously showed that hippocampal habituation is reduced in chronic schizophrenia. Here we investigate whether hippocampal habituation deficits can also be detected during the earliest stages of psychotic illness.

    Methods: We measured hippocampal activity in 65 early psychosis patients (< 2 years illness) and 70 healthy control participants using fMRI. Novelty response and habituation to repeated face and object images were computed for the anterior and posterior hippocampus separately. Analysis of variance tested for between-group differences by stimulus type. Effects of gender were examined and correlations with memory, cognitive, and clinical symptoms were performed.

    Results: Early psychosis patients had lower novelty response to faces in the anterior hippocampus compared to control participants (FWE corrected p = .04) but did not show altered habituation to faces. In contrast, in response to objects, early psychosis patients showed reduced habituation in the anterior hippocampus compared to control participants (FWE corrected p = .003) but did not show altered novelty response. There were no associations between hippocampal response and gender, years of education, IQ, cognitive function, or current clinical symptoms.

    Conclusions: These results provide evidence that both hippocampal novelty response and habituation are deficient in the early stage of psychosis. Hippocampal deficits were specific to stimulus type, suggesting examining the combination of social and nonsocial information may elicit a clearer model of pathophysiology. Importantly, our results provide preliminary evidence that deficits in hippocampal temporal response may progress with illness, although longitudinal studies are necessary. These findings further our understanding of the pathophysiology of schizophrenia and suggest