Shared and differential cortical functional abnormalities associated with inhibitory control in patients with schizophrenia and bipolar disorder

Schizophrenia (SZ) and bipolar I disorder (BD-I) share genetic risk factors and cognitive impairments, but these conditions may exhibit differences in cortical functioning associated with inhibitory control. We measured hemodynamic responses during a stop-signal task using near-infrared spectroscopy (NIRS) in 20 patients with SZ, 21 patients with BD-I and 18 healthy controls (HCs). We used stop-signal reaction time (SSRT) to estimate behavioural inhibition. Compared with HCs, patients with either SZ or BD-I exhibited significantly reduced activation in the bilateral inferior, middle and superior frontal gyri. Furthermore, patients with BD-I showed inactivation of the right superior temporal gyri compared with patients with SZ or HCs. Patients with SZ or BD-I demonstrated significant negative correlations between SSRT and hemodynamic responses of the right inferior frontal gyrus. Moreover, patients with SZ exhibited correlations in the middle and superior frontal gyri. Our findings suggest that right inferior frontal abnormalities mediate behavioural inhibition impairments in individuals with SZ or BD-I. Differential patterns of orbitofrontal or superior temporal functional abnormalities may reflect important differences in psychopathological features between these disorders.

Cognitive activation within groups. HCs showed significantly increased mean oxy-Hb levels between the pre-task baseline and the inhibitory control task period in 45 of 52 channels ('ch') (ch1-5, ch8-9, ch11-16, ch18-20, ch22-30, ch32-41 and ch43-52; t = −8.27 to −2.28; maximum false discovery rate (FDR)-corrected P < 0.05, corrected for 52 channels). Thus, widespread frontotemporal cortical activation related to oxy-Hb was induced by the inhibitory control task in HCs. Patients with SZ showed significantly increased mean oxy-Hb levels in 14 of 52 channels (ch19, ch29-30, ch34-36, ch38-39, ch44-47 and ch49-50; t = −4.22 to −2.76; FDR-corrected P < 0.05, corrected for 52 channels). By contrast, patients with BD-I did not show significant changes in oxy-Hb levels in any channel (FDR-corrected P < 0.05, corrected for 52 channels). Figure 1 summarises the cognitive activation results for all three groups.  Fig. 2). Thus, both patient groups produced significantly smaller task-related changes in oxy-Hb levels in the bilateral middle frontal, inferior frontal and superior frontal gyri compared with HCs (Table 2). Furthermore, patients with BD-I demonstrated significantly smaller task-related changes in oxy-Hb Figure 1. P-value significance map of cognitive activation among patients with SZ or BD-I and HCs (false discovery rate [FDR]-corrected P < 0.05, corrected with 52 channels). Blue-coloured circles represent significantly increased mean oxy-Hb changes from pre-task baseline to the inhibitory control task period in schizophrenia, bipolar I disorder and healthy controls. Channels that showed no significant correlations are highlighted in grey.
In addition, analysis of covariance (ANCOVA) revealed that task-related oxy-Hb changes differed significantly among the SZ, BD-I and HC groups in 9 channels (ch5, ch34-35, ch43-46, ch49 and ch51; F = 5.41-11.59; FDR-corrected P < 0.05, corrected for 52 channels). The BD group exhibited significantly smaller task-related oxy-Hb changes in 9 channels in comparison to HCs (ch5, ch34-35, ch43-46, ch49 and ch51; P = 0.000-0.010) and in 3 channels in comparison to the SZ group (ch43-45; P = 0.004-0.021). However, ANCOVA did not show any significance between the SZ and HC groups after FDR correction. Figure 3 summarises the results of regional differences in task-related changes in oxy-Hb levels in patients with SZ or BD-I. In brief, reduced hemodynamic responses in the bilateral middle frontal, inferior frontal and superior frontal (including orbitofrontal region) gyri are a shared feature in SZ and BD-I. However, the reduction in the right superior temporal gyrus is specific to patients with BD-I. Correlation analysis. Significant negative correlations between individual SSRT and task-related oxy-Hb changes were observed in the SZ group at four channels (ch36 and ch46-48; r = −0.71 to −0.64; FDR-corrected P < 0.05, corrected for 52 channels) located in the right middle frontal, inferior frontal and left superior frontal gyri, as well as in the BD-I group at one channel (ch45; r = −0.70; FDR-corrected P < 0.05, corrected for 52 channels) located in the right inferior frontal gyrus. There were no significant correlations between SSRT and task-related oxy-Hb changes in HCs. Figure 4 summarises the correlation analysis in patients with SZ or BD-I.
We observed significant negative correlations between the Global Assessment of Functioning (GAF) scale score and task-related oxy-Hb changes in the BD-I group at two channels (ch43-44; r = −0.72 to −0.68; FDR-corrected P < 0.05, corrected for 52 channels) located in the right middle and superior temporal gyri. No significant correlations were observed between task-related oxy-Hb changes and any other clinical variable in the SZ, BD-I, or HC groups, including antipsychotic or antidepressant dosage. No significant correlations were observed between task-related oxy-Hb changes and HAMD or YMRS scores in the BD-I group.

Discussion
To our knowledge, this is the first NIRS study to directly compare regional cortical hemodynamic responses associated with inhibitory control among patients with SZ or BD-I and HCs. Patients with SZ or BD-I exhibited functional abnormalities associated with inhibitory control in the right inferior frontal gyrus. The differential abnormalities associated with behavioural inhibition were observed in the right middle and superior frontal gyri in patients with SZ and in the right superior temporal gyrus of patients with BD-I. Our findings suggest that differential patterns of frontotemporal functional abnormalities may reflect an important difference in psychopathology between schizophrenia and bipolar disorder.
Compared with HCs, patients with SZ or BD-I showed significantly reduced hemodynamic responses in the bilateral inferior frontal, middle frontal and superior frontal gyri. These results are consistent with previous SZ or BD-I neuroimaging studies that reported functional abnormalities in these regions 7,12-16 . The right inferior frontal gyrus is known to play a general role in attentional control, rapidly adapting inhibitory control responses to current salient stimuli [25][26][27] . Previous neuroimaging studies of HCs reported activation of the right inferior frontal and superior temporal gyri during a response inhibition task 28 . By contrast, studies have reported significant associations between poor inhibitory control and the volume of damage to grey matter in the right inferior frontal gyrus in patients with lesions of the right frontal lobe 29 and reduced white matter fractional anisotropy in this region in patients with methamphetamine dependence 30 . Our results indicate that poor inhibitory control is associated with reduced hemodynamic responses in the right inferior frontal gyrus of patients with SZ or BD-I. These individuals have specific patterns of cerebral alterations in the right inferior frontal gyrus; this region may therefore be a neural substrate of impaired behavioural inhibition. Our findings suggest that right inferior frontal abnormalities mediate deficits in inhibitory control that are shared across patients with either SZ or BD-I.
In addition, the results of the current study suggest that patients with SZ exhibited deficits in inhibitory control associated with reduced hemodynamic responses in the right middle frontal and left superior frontal gyri, including the orbitofrontal region (channels [47][48]. Previous studies on individuals with SZ indicated that structural or functional abnormalities in the orbitofrontal cortex were strongly associated with impulsive aggression or suicidal behavior [31][32][33] . Furthermore, alterations of orbitofrontal areas associated with SZ have been linked to positive and negative symptom profiles in previous structural neuroimaging studies 34,35 . Taken together, these studies suggest that functional abnormalities in the orbitofrontal region impair inhibition during impulsive or aggressive expression of symptoms in patients with SZ. Further, the alteration in this region may be related to the specific psychopathology of SZ. Our findings suggest that orbitofrontal functional abnormalities related to impaired behavioural inhibition are a neural substrate of the impulsive or aggressive clinical features associated with SZ.  By contrast, patients with BD-I demonstrated reduced hemodynamic responses in the right superior temporal gyrus, which was unique to this group. Moreover, patients with BD-I exhibited negative associations between SSRT and hemodynamic responses in this region. The superior temporal cortex (gyrus and sulcus) is part of a complex face-processing system involved in the perception of emotions 36 and regulating responses to negative visual social stimuli 37 . Recent structural neuroimaging studies in patients with BD-I have found abnormally reduced thickness of the right superior temporal cortex 38 or abnormally reduced fractional anisotropy in the right temporal lobe 39 . One investigation suggested that reduced activation in the superior temporal sulcus may indicate disturbances in affect-processing circuitry, leading to abnormalities in mood and social cognition 40 . These findings suggest that the superior temporal gyrus is an important neural substrate of emotion regulation and reward processing, and that alterations in this region may produce some of the cognitive symptoms and social impairments associated with BD-I. Moreover, functional abnormalities in this region may result in impaired emotional control inhibition in individuals with BD-I. This hypothesis is partly supported by the results of the current study, which showed an association between impaired global functioning and abnormal right temporal region hemodynamic responses in patients with BD-I. Our findings suggest that the functional abnormalities related to inhibitory control in the right superior temporal gyrus are characteristic of BD-I.
In addition, patients with SZ showed significantly longer SSRTs compared to patients with BD-I or HCs. This is consistent with previous evidence associating SZ with response inhibition deficits 2,5 . By contrast, no such difference in mean SSRT was observed patients with BD-I or HCs, consistent with the notion that neurocognitive functioning in individuals with BD-I is impaired relative to HCs but better than in patients with SZ 41,42 . An alternative explanation for this difference between SZ and BD-I groups is that the current mood state in patients with BD-I influenced SSRT, whereas the deficits persisted across all phases of BD-I 6,7,43 . A further explanation is that neuroimaging techniques, including the NIRS signal, are more sensitive than SSRT at detecting differences between the BD-I and HC groups, especially given small sample sizes. This observation has been reported in previous neuroimaging studies 17,18 . In the current study, we considered observed differences in hemodynamic responses among patients with SZ or BD-I and HCs to be activated by the stop-signal task.
There are several potential limitations to our study. First, our sample size was small, which may have limited the statistical power to produce reliable findings, i.e. ANCOVA did not show any significant differences between the SZ and HC groups after FDR correction. This could also explain why associations between SSRT and task-related oxy-Hb changes were not observed in HCs. In addition, the NIRS signal might be more sensitive, allowing greater detection of cortical functional abnormalities in patients with psychiatric diagnoses. This speculation is partly supported by previous NIRS studies that did not detect brain regions associated with task  23 . Further studies with larger sample sizes are needed to verify our findings. Second, we cannot exclude the possibility of medication effects on hemodynamic responses in patients with SZ and BD-I. However, it should be noted that the dosage of antidepressants or antipsychotics was not correlated with hemodynamic responses in any channels, suggesting a minimal confounding effect of medication. In addition, previous NIRS studies have reported no significant effect of psychotropic medications on neural activity in patients with SZ or BD [19][20][21][22][23] . Lastly, NIRS has lower spatial resolution to detect cortical activity from the scalp surface than other neuroimaging techniques such as fMRI. However, this limitation should be within an acceptable range because differences in hemodynamic responses in patients with SZ or BD-I and HCs were clearly observed in our study, as they have been in other NIRS studies 23,24,44 . In addition, we used a virtual spatial registration method to define the spatial information for each channel using data [45][46][47] . This method was used in most recent NIRS studies and may be useful for replicating our findings.

Conclusion
This study demonstrated shared and differential patterns of hemodynamic responses in the frontotemporal cortex among patients with SZ or BD-I and HCs. Our findings suggest that differential patterns in frontotemporal functional abnormalities may reflect important differences in the psychopathological features of SZ or BD-I. NIRS is a non-invasive neuroimaging modality with easy applicability and high ecological validity. This makes NIRS particularly suitable for use with patients with psychiatric diagnoses who may be afraid of enclosed spaces or exhibit motor restlessness that interferes with motion-sensitive imaging methods 45 . Therefore, NIRS has the potential to be a powerful and specific diagnostic tool for use among individuals with SZ or BD-I.

Methods
Participants. The diagnosis of SZ or BD-I was made according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition 48 criteria using the Mini International Neuropsychiatric Interview (MINI) 49 . HCs were also screened using the MINI and excluded if there was any history of psychiatric disorders or heritable neurological diseases among first-or second-order relatives. Exclusion criteria for the study groups were as follows: a history of head trauma with loss of consciousness for more than 5 min, current or previous neurological disease, current or previous endocrine disease, a history of electroconvulsive therapy and/or alcohol/substance abuse or addiction.
Psychiatric symptoms were assessed using the Positive and Negative Syndrome Scale (PANSS) 50 . Global functioning was assessed with GAF 48 . Depression severity was evaluated using the 17-item Hamilton depression rating scale administered using a structured interview guide 51 . Manic symptoms were assessed using the Young mania rating scale (YMRS) 52 . Premorbid IQ was estimated using the Japanese version of the National Adult Reading Test 53 . Handedness was evaluated according to the Edinburgh Inventory 54 . The chlorpromazine-equivalent dose of antipsychotics was calculated for each patient 55 .
After providing a complete description of the study, written informed consent was obtained from all participants. This study complied with the Declaration of Helsinki and was approved by the Ethics Committee of the Kindai University Faculty of Medicine. All data generated or analysed during this study are included in this article.

NIRS methodology.
We used a 52-channel NIRS device (ETG-4000 Optical Topography System; Hitachi Medical Co., Tokyo, Japan) to measure relative changes in oxy-Hb and deoxygenated haemoglobin (deoxy-Hb) at two wavelengths (694 and 830 nm) of near-infrared light (indicated as mM) based on the modified Beer-Lambert law 56 . The NIRS probes were fixed using 3 × 11 thermoplastic shells with 17 emitters and 16 detectors. The distance between each source and detector probe was 30 mm and the area analysed between the probes was defined as a 'channel' . The probes of the NIRS device were placed on the frontotemporal region of each participant, with the lowermost probes located along the T3-Fp1-Fpz-Fp2-T4 line, in accordance with the International 10-20 Placement System used for electroencephalography. The NIRS device can measure Hb values bilaterally from the prefrontal and temporal surface regions at depths of 20-30 mm from the scalp, which correspond with the surface of the cerebral cortex. The spatial information for each channel was estimated using data from the Functional Brain Science Laboratory at Jichi Medical University, Japan 45,46 . According to the LONI Probabilistic Brain Atlas (LPBA40) 47 , NIRS channels can record functional hemodynamics within the bilateral frontal, temporal and parietal cortices. We anatomically labelled NIRS channels only after the LPBA40 region of highest probability was determined (Fig. 5).
We recorded relative mean changes in Hb concentrations from baseline in mM because NIRS cannot measure the absolute path length from the emitter to the detector. NIRS signals were acquired with a time resolution of 0.1 s. We set the moving average window to 5 s to remove high frequency noise such as heartbeat and small bodily movements. Channel records with low signal-to-noise ratios or motion artefacts were excluded, and there was no difference in the number of samples at each channel, including the channels that indicated significant differences among study groups.
In the present study, we focused on changes in oxy-Hb. Animal studies indicate that oxy-Hb is the most sensitive indicator of regional cerebral blood flow because the direction of change in deoxy-Hb is determined by the degree of change in venous blood oxygenation and volume 57 . Pre-task baseline was determined as the mean across the last 10 s of the 30 s pre-task segment. The obtained oxy-Hb data were averaged for each participant during each baseline and task period. The task-related oxy-Hb changes (task minus baseline) were used for statistical analysis.
Stop-signal task. We used a variation on the classic stop-signal task 4 (SST) to assess inhibitory control 58 .
The original task design has been described in detail 58,59 . In brief, across the task period, participants viewed the shape of a 'go' stimulus (circle or square) that appeared on screen in rapid succession. Occasionally, an auditory stop signal was presented at a short, variable delay after the onset of the go stimulus.
Procedure. The cognitive activation task used in this study had a one-block design and included 1) a pre-task baseline period and 2) a stop-signal task period (Fig. 6A). During a 30 s pre-task period, participants were instructed to alternatively tap with their right and left index fingers every 1 s as accurately as they could, using Overall task period. During the pre-task period, subjects were instructed to alternatively tap with their right and left index fingers every 1 s as accurately as they could. (B) During the stop-signal task, subjects were instructed to respond as quickly and accurately as possible to the go stimulus ('•' , '■') in no-stop-signal trials. In stopsignal trials, participants were instructed to try to withhold their response as required until the auditory stop signal occurred.
SCieNTiFiC RepoRtS | (2018) 8:4686 | DOI:10.1038/s41598-018-22929-y the left and right buttons on a keyboard. The task consisted of a block of 64 trials; stop-signal trials (75% of trials) and non-stop-signal trials (25% of trials) were randomly presented. Each trial started with the presentation of a fixation sign on a computer monitor, which was replaced by the primary task stimulus after 250 ms. Stop-signal delay was initially set at 250 ms and continuously adjusted with the staircase tracking procedure (range of 250-1050 ms). When inhibition was successful, the stop-signal delay increased by 50 ms; when inhibition was unsuccessful, it decreased by 50 ms. Response registration continued during stop-signal presentation. In stop-signal trials, the go stimulus was followed by a stop signal (750 Hz, 75 ms) after a variable stop-signal delay. In both no-stop-signal trials and stop-signal trials, the go stimulus remained on screen until subjects responded or the maximal reaction time (RT; 1,250 ms) had elapsed. The inter-trial interval was 2,000 ms and independent of RT. Participants were instructed to respond as quickly and accurately as possible to the go stimulus in non-stop-signal trials. In stop-signal trials, participants were instructed to try to withhold their response as required until the auditory stop signal occurred (Fig. 6B).
Statistical analysis. Demographic and clinical variables were compared among the study groups using the χ 2 test, t-test and one-way ANOVA followed by Tukey's post hoc test. Statistical significance was assumed at P < 0.05 (two-tailed).
To identify regional differences in frontotemporal hemodynamic activation associated with the inhibitory control task, we compared mean oxy-Hb changes during the pre-task (baseline) and task periods for every channel within individual subjects using the paired t-test. Task-related oxy-Hb changes were compared among the study groups using ANOVA followed by Tukey's post hoc test. Because Tukey's post hoc test focused on NIRS channels where the hemodynamic response was shown to be significant after FDR correction, no further correction for multiple comparisons was applied for the post hoc test and responses were considered significant at P < 0.05 23 . Complementary analyses were performed to identify between-group differences in activation associated with inhibitory control using ANCOVA with age, gender (dummy parameterised, male = 0, female = 1) and SSRT as covariates. Furthermore, to examine the relationships between task-related oxy-Hb changes and clinical variables, we calculated Pearson's correlation coefficients. We set the value of q (i.e. FDR) to 0.05 such that the false-positive rate was not greater than 5% on average when processing oxy-Hb data obtained from multiple channels 60 .