Introduction

As the obesity epidemic progresses, with 42% of the adult U.S. population being obese, rising healthcare costs of over 700 billion dollars annually have been observed1,2,3. Studies have shown associations between obesity and abnormal ingestive behaviors, primarily “food addiction”, in 40% of individuals seeking bariatric surgery4,5. Food addiction describes an addictive response in some individuals, where unintended overeating, the increased intake of ultra-processed/hyperpalatable foods beyond homeostatic needs or eating primarily for pleasure occur despite negative consequences6,7,8. The Yale Food Addiction Scale (YFAS) is a validated and psychometrically sound measure that uses the DSM-IV diagnostic criteria for substance abuse to operationalize food addiction9,10. While it is believed that food addiction is distinct from other behavioral eating disorders, it does share the characteristics or withdrawal, tolerance, impulsivity, and emotional reactivity seen with substance-use disorders and other addictive behaviors6,7,11.

Research depicting sex differences in obesity and food addiction have gained momentum to increase treatment efficacy12. Although prevalence rates in obesity are similar between the sexes, females report nearly double the rates of food addiction compared to males (12.2–6.4% respectively), with obese females being more likely to encounter loss of control while eating12,13. This stems from the increased frequency and cravings towards food as well as the increased reactivity to food cues experienced by females with obesity compared to males12,14,15.

Past literature has used magnetic resonance imaging (MRI) as a tool to understand the underlying neurobiology of both obesity and food addiction16,17,18. Individuals with obesity show greater activation in the reward and salience networks, particularly in the basal ganglia, in response to visual stimuli compared to normal weight individuals19,20. This has been associated with an increased drive for cravings and greater food consumption, which could explain the prevalence of overeating in obese individuals21,22. Studies have also indicated brain connectivity differences related to food addiction between regions of the reward network due to its role in controlling voluntary behavior23,24,25. Altered connectivity in the basal ganglia occurs as a result of the decreased expression of dopamine (D2) receptors which mediate reward-seeking behavior4,26,27. The reduction in the availability of D2 receptors creates a perpetual hypodopaminergic state in affected individuals leading to decreasing reward sensitivity, as greater amounts of dopamine must be released to receive the same stimulatory effect25. Similar to other addictive behaviors, individuals with food addiction exhibit increased alterations in the regions of the reward network in response to food cues28,29,30,31. Decreased sensitivity of the reward system translates into decreased inhibitory control and cognitive modulation in food addiction4,26,27. Dysregulation of the reward network in conjunction with persistent activation of the dopaminergic pathway supports the compulsive, unregulated food consumption of palatable and high calorie foods seen in obese individuals with food addiction32,33.

Sex differences have also been observed in the brain signatures of individuals with obesity. Compared to males, females display an increased reactivity to visual food cues and higher incidences of food cravings, and greater activations in brain regions associated with visual stimuli identification, such as the fusiform gyrus34. Additionally, females with obesity showed positive associations between increased BMI and lower connectivity of core reward network regions with cortical and emotion regulation regions35,36. This has been linked to the decreased ability to control the physiological response to negative-emotion-inducing stimuli leading to increased emotional eating and food addictive behaviors in females with obesity compared to their male counterparts37,38. Sex differences in food addiction have implications for treatments outcomes, as sex differences to naltrexone responsivity have also been observed in other substance abuse disorders, with women generally reporting greater levels of nausea, and poorer treatment adherence and outcomes39,40.

Although past literature has addressed the association between food addiction and brain responses to stimuli, few, if any, studies have investigated sex differences in brain signatures, in individuals with and without food addiction. In this study, we aim to examine the neural substrates of food addiction using resting state fMRI and determine sex differences in network connectivity in order to test the following hypotheses: (1) Individuals with food addiction show greater connectivity with regions of the reward network, such as the basal ganglia, compared to individuals with no food addiction. (2) Females with food addiction show greater connectivity with regions of the brainstem, reward, and emotional regulation networks, but lowered connectivity with regions of the brainstem, sensorimotor, and central executive regions compared to males with food addiction. (3) Higher scores on questionnaires measuring altered ingestive eating behaviors are associated with greater connectivity of reward and emotional regulation regions, especially in females.

Materials and methods

Subjects

A total of 150 participants (male: N = 47, female: N = 103), ages 18–55 years old, were recruited with the use of flyers and community advertisements and enrolled in the study at the G. Oppenheimer Center for Stress and Resilience. All participants had a BMI greater than 25 kg/m2 (obese/overweight; referred to as high BMI henceforth). Participants were excluded for the following: major medical/neurological conditions, current or past psychiatric illness, comorbidities (vascular disease and diabetes), weight loss/abdominal surgeries, pregnancy or breastfeeding, substance use, extreme strenuous exercise (> 8 h of continuous exercise per week), substance use, tobacco dependence (half a pack or more daily), and metal implants. Participants taking medications that interfere with the central nervous system or regular use of analgesic drugs were also excluded. No participants exceeded 400lbs due to magnetic resonance imaging scanning weight limits. Since female sex hormones such as estrogen are known to effect brain structure and function, only premenopausal females were included. All procedures were in compliance with institutional guidelines and were approved by the Institutional Review Board at UCLA’s Office of Protection for Research Subjects. All participants provided written informed consent.

Questionnaires

Participants were asked to fill out the Yale Food Addiction (YFAS) questionnaire, a 25-item scale developed to measure “food addiction” by assessing signs of substance-dependence symptoms in eating behavior41. The Yale Food Addiction Scale (YFAS) has been the commonly utilized measure of food addiction to highly palatable (high fat and high sugar) foods, as these foods have been linked with excess consumption and lowered appetite modulation4,41. This scale is based upon the substance dependence criteria found in the DSM-IV42 (e.g., tolerance [marked increase in amount; marked decrease in effect], withdrawal [agitation, anxiety, physical symptoms], and loss of control [eating to the point of feeling physical ill])41. A symptom count of ≥ 3 together with endorsement of clinically significant impairment or distress on the YFAS denotes diagnostic criteria food addiction (FA). Clinically significant impairment or distress was defined as having a at least one positive response to the following two questions in the YFAS questionnaire: “My behavior with respect to food and eating causes significant distress” and “I experience significant problems in my ability to function effectively (daily routine, job/school, social activities, family activities, health difficulties) because of food and eating,” similar to previously published works10. Based on a 353-respondent exploratory survey, the YFAS has displayed a good internal reliability α = 0.8641. With these measures, participants were placed into one of the following groups based on diagnostic criteria of food addiction (FA): 1. High BMI and with FA (N = 40) 2. High BMI and without FA (N = 110). 3. Females with high BMI and with FA (N = 30), 4. Females with high BMI and without FA (N = 73), 5. Males with high BMI and with FA (N = 10), and 6. Males with high BMI and without FA (N = 37).

Another questionnaire that measures abnormal ingestive behavior was also administered to participants. The General Food Cravings Questionnaire—Traits (GFCQT-r), comprised of 15 items, measures the frequency and intensity of food cravings as a way to study eating patterns and behavior43. Higher scores on the GFCQT-r are positively associated with eating pathology, low dieting success, BMI, and increased food cravings43. The GFCQT-r has shown high internal consistency (α = 0.94) and reliability in its ability to assess food cravings as a trait44.

The Hospital Anxiety/Depression Scale (HADS) is a 14-item questionnaire that attempts to identify both possible and probable cases of both anxiety disorders and depression among patients in a non-psychiatric setting45. When applied to samples of primary care, psychiatric, and somatic patients, the Hospital Anxiety/Depression Scale has shown strong internal reliability and good concurrent validity45.

Functional magnetic resonance imaging acquisition

Whole brain resting-state functional data was acquired using a 3.0 T Siemens Prisma MRI scanner (Siemens, Erlangen, Germany). Detailed information on the standardized acquisition protocols, quality control measures, and image preprocessing are provided in previously published studies35,46,47,48,49,50. Resting-state scans were acquired with eyes closed and an echo planar sequence with the following parameters: TE/TR = 28 ms/2000 ms, flip angle = 77 degrees, scan duration = 8m6s–10m6s, FOV = 220 mm, slices = 40 and slice thickness = 4.0 mm, and slices were obtained with whole-brain coverage. Preprocessing and quality control was done using Statistical Parametric Mapping-12 (SPM12) software and involved bias field correction, co-registration, motion correction, spatial normalization, tissue segmentation, and Fourier transformation for frequency distribution. Data was then spatially normalized to the Montreal Neurological Institute (MNI) template using the structural scans, and then smoothed using a 4 mm isotropic Gaussian kernel.

Functional network construction

Functional brain networks were constructed as previously described in35. To summarize, measures of region-to-region functional connectivity (Fisher transformed Pearson’s correlations) were computed using the CONN toolbox and the aCompCor method in Matlab51. Confounding factors such as white matter, cerebrospinal fluid, the six motion realignment parameters, and the root mean squared (RMS) values of the detrended realignment estimates were regressed out for each voxel using ordinary least squares (OLS) regression on the normalized, smoothed resting-state images52. Subjects with RMS values over 0.25 were not included. Images were then filtered using a band-pass filter (0.008/s < f < 0.08/s) to reduce the low- and high-frequency noises. Although the influence of head motion cannot be completely removed, CompCor has been shown to be particularly effective for dealing with residual motion relative to other methods53. Regions of interest were segmented with the Harvard–Oxford Subcortical atlases, the Schaefer 400 cortical atlas, and the Ascending Arousal Network brainstem atlas54,55. These atlases parceled into a total of 430 brain regions. The ROI-ROI functional connectivity between the brain regions was indexed by a matrix of Fisher Z transformed correlation coefficients reflecting the association between average temporal BOLD time series signals across all voxels in each brain region. The magnitude of the Z value represents the weights in the functional network. Permuted statistical values from ROI-to-ROI analyses were further corrected using the false discovery rate (FDR) to measure significance with p(FDR) < 0.05.

Statistical analyses

Descriptive statistical analyses were performed on the 150 subject dataset using SPSS Statistics software on the four groupings listed earlier as well as a combined FA (N = 40) and no FA (N = 110) groups. The General Linear Model (GLM) procedure in SPSS was utilized to analyze the variance and significance between the four groups across all the clinical variables. Second level connectivity analyses were run on CONN to discern sex differences in functional connectivity using GLMs. For all analyses, the disease-dependent analyses (FA vs. no FA) were controlled for age and sex, and all sex difference analyses were controlled for age. The following 5 planned contrasts were used in the statistical analyses for both the clinical variables and connectivity analyses: FA > No FA, Females with FA > Males with FA, Females with FA > Females with no FA, Males with FA > Males with no FA, Females with no FA > Males with no FA.

Results

Subject characteristics

All participants had a BMI ≥ 25 kg/m2 and were divided into two groups based on diagnostic criteria of FA by their YFAS Symptom Count scores cut off ≥ 3 together with endorsing clinically significant impairment or distress. A complete summary of all the group differences in clinical variables are summarized in Tables 1 and 2.

Table 1 Summary of Study Demographics and Clinical Behavioral Measures.
Table 2 Group differences in demographics and clinical behavioral measures.

Individuals with high BMI and food addiction had higher reported scores on the General Food Craving Questionnaire (GFCQT) (p < 0.05), and higher anxiety (p = 0.02) and depression (p = 2.00E−3) compared to individuals with no food addiction.

Compared to females with food addiction, males with food addiction had higher scores on the following YFAS subscales: Given Up (p = 4.00E−03), Loss of Control (p = 9.00E−03), and Clinical Significant Impairment components (p = 0.03).

Within females, females with food addiction had higher scores on all components of the GFCQT (p < 0.05) and higher anxiety (p = 0.03) and depression (p = 0.04) compared to females without food addiction. Compared to males without food addiction, males with food addiction had higher scores on all components of the GFCQT (p < 0.05), and depression (p = 0.02) scales.

Food addiction dependent effects on brain connectivity

Results are detailed in Table 3 and summarized in Table 5, and depicted in Fig. 1.

Table 3 Resting state pairwise connections in individuals with food addiction compared to individuals with no food addiction.
Figure 1
figure 1

Connectogram depicting food addiction dependent differences in brain connectivity (food addiction vs. no food addiction). Demonstrates significant differences in functional connectivity between individuals with food addiction and individuals with no food addiction. Analysis was performed Harvard–Oxford Subcortical atlases, the Schaefer 400 cortical atlas, and the Ascending Arousal Network brainstem atlas. Labels on the diagram are Destrieux, Harvard–Oxford Subcortical atlases, and the Ascending Arousal Network brainstem atlas equivalents. Red lines between two networks indicate greater functional connectivity, and blue lines indicate lowered functional connectivity. All connections are significant q < 0.05. Light Green: SMN (Sensorimotor Network); Black: BG (Basal Ganglia); Purple: DMN (Default Mode Network); Yellow: SAL (Salience); Red: ERN (Emotional Regulation Network); Dark Green: CAN (Central Autonomic Network); Orange: CEN (Central Executive Network); Blue: OCC (Occipital); Gray: CeB (Cerebellum); Brown: BST (Brain Stem). Bst, Brainstem; CAN, Central Autonomic Network; CEN, Central Executive Network; DMN, Default Mode Network; ERN, Emotional Regulation Network; InfFS, Inferior frontal sulcus; MFG, Middle frontal gyrus; MRF, Mesencephalic reticular formation; MTG, Middle Temporal Gyrus; OrG, Orbital gyri; PosCG, Postcentral Gyrus; SMN, Sensorimotor Network.

Brainstem connectivity

Individuals with food addiction had greater connectivity between the brainstem and middle frontal gyrus (q = 0.02), and with bilateral orbital gyri (Left and Right q = 0.02) in compared to those without food addiction.

Emotional regulation network connectivity

Individuals with food addiction had lowered connectivity between the inferior frontal sulcus and posterior central gyrus compared to those without food addiction (q = 0.04).

Sensorimotor network connectivity

Lowered connectivity was found between the postcentral gyrus and middle temporal gyrus in those with food addiction compared to those without food addiction (q = 0.03).

Sex differences in the food addiction dependent effects on brain connectivity

Results are detailed in Table 4 and summarized in Table 5, and depicted in Fig. 2.

Table 4 Resting state pairwise connections in females with food addiction compared to males with food addiction.
Table 5 Summary of all group comparisons in network connectivity (food addiction and sex).
Figure 2
figure 2

Connectogram Depicting Sex Differences in the Food Addiction Dependent Effects on Brain Connectivity (Females with Food Addiction vs. Males with Food Addiction). Demonstrates significant differences in functional connectivity between females with food addiction and males with food addiction. Analysis was performed Harvard–Oxford Subcortical atlases, the Schaefer 400 cortical atlas, and the Ascending Arousal Network brainstem atlas. Labels on the diagram are Destrieux, Harvard–Oxford Subcortical atlases, and the Ascending Arousal Network brainstem atlas equivalents. Red lines between two networks indicate greater functional connectivity, and blue lines indicate lowered functional connectivity. All connections are significant q < 0.05. Light Green: SMN (Sensorimotor Network); Black: BG (Basal Ganglia); Purple: DMN (Default Mode Network); Yellow: SAL (Salience); Red: ERN (Emotional Regulation Network); Dark Green: CAN (Central Autonomic Network); Orange: CEN (Central Executive Network); Blue: OCC (Occipital); Gray: CeB (Cerebellum); Brown: BST (Brain Stem). AngG, Angular gyrus; CEN, Central Executive Network; CS, Central sulcus (Rolando's fissure); DMN, Default Mode Network; ERN, Emotional Regulation Network; InfFGTrip, Triangular part of the inferior frontal gyrus; IntPS_TrPS, Intraparietal sulcus(interparietal sulcus) and transverse parietal sulci; MFG, Middle frontal gyrus; MOcS_LuS, Middle occipital sulcus and lunatus sulcus; MPosCgG_S, Middle-posterior part of the cingulate gyrus and sulcus; PaCL_S, Paracentral lobule and sulcus; PosCG, Postcentral Gyrus; PRCG, Precentral gyrus; PrCun, Precuneus; ShoInG, Short insular gyri; Tpo, temporal pole; SAL, Salience Network SbPS, Subparietal sulcus; SMN, Sensorimotor Network; SupFG, Superior frontal gyrus; SuMarG, Supramarginal gyrus; SupTGLp, Lateral aspect of the superior temporal gyrus.

Within sex results detailed in Supplemental Tables S1, S2, and depicted in Supplemental Figures S1, S2.

Brainstem connectivity

Compared to females without food addiction, females with food addiction had greater connectivity between the brainstem and middle-anterior part of the cingulate gyrus/sulcus (q = 0.04).

Compared to males without food addiction, males with food addiction had greater connectivity between the brainstem and bilateral middle frontal gyrus, orbital gyrus, and middle occipital gyrus (q = 0.04), but lower connectivity between the brainstem and parahippocampal gyrus (q = 0.03), brainstem and precentral gyrus (q = 0.04), and the brainstem and anterior traverse collateral sulcus (q = 0.04).

Emotional regulation network connectivity

Compared to males with food addiction, females with food addiction had greater connectivity between the inferior frontal gyrus and paracentral lobule (q = 0.04).

Compared to females with no food addiction, females with food addiction had lower connectivity between the inferior frontal sulcus and subcentral gyrus (q = 0.01) and between the inferior frontal gyrus and superior parietal lobule (q = 0.03).

Salience network connectivity

Compared to males with food addiction, females with food addiction had greater connectivity between the short insular gyrus and paracentral lobule (q = 0.01) and between the middle/posterior part of the cingulate gyrus and superior temporal gyrus (q = 0.03).

Compared to males without food addiction, males with food addiction had greater connectivity between the opercular part of the inferior frontal gyrus and the inferior part of the precentral sulcus (q = 0.03).

Sensorimotor network connectivity

Compared to males with food addiction, females with food addiction had greater connectivity between the paracentral lobule/sulcus and the middle frontal gyrus (q = 0.01–0.03), and between the central sulcus, precentral gyrus and superior temporal gyrus (q = 0.03–0.04).

Compared to females without food addiction, females with food addiction had lowered connectivity between the inferior frontal gyrus and subcentral gyrus (q = 0.02) and between the precentral gyrus and lingual gyrus (q = 0.04).

Compared to males with no food addiction, males with food addiction had lowered connectivity between the paracentral lobule and middle frontal gyrus (q = 0.04).

Central autonomic network connectivity

Compared to females without food addiction, females with food addiction had greater connectivity between the anterior cingulate gyrus and sulcus and paracentral lobule (q = 3.00E-03).

Central executive network connectivity

Compared to males with food addiction, females with food addiction had lowered connectivity between the middle frontal gyrus and superior temporal gyrus (q = 0.03), intraparietal sulcus and supramarginal gyrus (q = 0.04), and superior frontal gyrus and superior temporal gyrus (q = 0.03–0.04).

Compared to females without food addiction, females with food addiction had greater connectivity between the intraparietal sulcus and orbital gyrus (q = 0.03).

Default mode network connectivity

Compared to males with food addiction, females with food addiction had lowered connectivity between the supramarginal gyrus and the precuneus (q = 0.01), precuneus and superior temporal gyrus (q = 0.03), and superior temporal gyrus and angular gyrus (q = 0.03).

Compared to males without food addiction, males with food addiction had greater connectivity between the supramarginal gyrus and the precuneus (q = 0.02) and between the temporal pole and orbital sulcus (q = 0.03).

Associations between brain connectivity and clinical variables

Food addiction associations

Results are summarized in Table 6.

Table 6 Associations between significant functional connectivities and clinical variables.

For individuals with no food addiction, connectivity between the brainstem (mesencephalic reticular formation) and central autonomic network (orbital gyrus) was positively associated with BMI (r = 0.35, q = 0.01) and negatively associated with the Intention component of the GFCQT (r = − 0.38, q = 0.04).

Sex difference associations

Results are summarized in Table 6.

In females with food addiction, connectivity between the brainstem (locus coeruleus) and emotional regulation network (middle-anterior part of the cingulate gyrus and sulcus) was positively associated with BMI (r = 0.61, q = 0.04).

Discussion

The goal of this study was to identify sex-related differences in the connectivity of brain networks in individuals meeting diagnostic criteria for food addiction. The main findings of the study were: 1) Food addiction was associated with greater connectivity among the reward regions and between the brainstem and central autonomic networks, and lower connectivity among the emotional regulation, sensorimotor, and default mode networks compared to those with no food addiction. 2. Sex differences were observed with females showing greater connectivity in the emotional regulation and salience networks and lower connectivity in the brainstem, central executive, and default mode networks. Our results support the hypothesis that altered connectivity in reward regions could increase the risk for addictive ingestive behaviors, resulting in the uncontrollable overeating patterns seen in individuals with clinically significant impairment or distress with food addiction. Additionally, our results indicate greater connectivity in particular resting-state networks, such as the emotional regulation and salience networks, which could explain the higher rates of emotional eating and food addiction seen in females. To our knowledge, this is the first study to investigate sex-related differences in resting-state connectivity in individuals with food addiction.

Food addiction dependent effects on brain connectivity

Uncontrollable eating seen in food addiction, can be explained by the reward deficiency hypothesis which states that a decreased availability of dopamine receptors, specifically D2 receptors, creates a less responsive reward system that is susceptible to addictive pathologies35,56,57,58,59,60. Drug-addiction studies have associated drug-dependence with changes in the dopamine receptor availability, with individuals with decreased receptors seeking greater and more frequent reward stimulation57. In individuals with food addiction, high sugar/high fat foods act as this source of stimulation, as such these ultra-processed foods serve as potent reward triggers that increase synaptic dopamine concentration overriding internal satiety cues57,61. Hence, the constant hypodopaminergic state of these individuals results in increased levels of food cravings and overindulgence of ultra-processed foods as a compensatory attempt to derive the euphoric effects of reward network stimulation62,63. Studies have also suggested that disruptions in the mesolimbic pathways associated with addiction behaviors effects both the DA reward circuits and DA pathways that lead to increased stress reactivity and disruption in interoceptive awareness29,57,64.

Consistent with the reward deficiency hypothesis, our results indicated greater connectivity between the brainstem and reward regions in individuals with food addiction. This could be associated with dopaminergic dysregulation, as the increased intake of ultra-processed foods, in those with food addiction leads to a more persistent stimulation of the reward pathway as a compensatory mechanism for the decreased receptor sensitivity and availability19,33,65. On the other hand, in those with no food addiction, a greater connectivity between the brainstem and the central autonomic network (orbital gyrus) was negatively associated with GFCQT scores, a measure of food cravings. Perhaps, in individuals with food addiction, increases in brainstem connectivity could be a result of an increased frequency of dopamine release in response to select salient inputs, such as ultra-processed foods, causing heightened reactions and cravings to consume larger quantities of such foods66,67. This increased brainstem connectivity, could be a counterbalancing mechanism for the reduction in dopamine receptors in the reward circuit per the reward deficiency hypothesis26,32,57,68.

In addition to decreased reward sensitivity, hyperactivation of the locus coeruleus could also contribute to the habitual overconsumption of ultra-processed foods seen in individuals with food addiction. The locus coeruleus plays a major role in an individual’s response to external stressors, with greater activation being linked to consistent higher levels of norepinephrine seen in individuals with anxiety69. Individuals in a sustained state of anxiety, consistently have elevated levels of norepinephrine which result in cortical atrophy leading to reduced cognitive and attentional control70. Thus, greater activation of the locus coeruleus could translate to a greater susceptibility to maladaptive, habitual behaviors such as food addiction due a combination of decreased internal regulation capabilities and perpetual anxiety in individuals with food addiction69,70.

Our results mirrored these predicted differences, with individuals with food addiction displaying significantly higher rates of anxiety, particularly in females. Additionally, greater connectivity between the locus coeruleus and the emotional regulation network was positively associated with BMI in females with food addiction. These results suggest the role of the locus coeruleus in weight management may contribute to the food addictive behaviors in an attempt to cope with anxiety.

Sex differences in the food addiction dependent effects on brain connectivity

Compared to males with food addiction, females with food addiction displayed significantly higher rates of emotional overeating37,38. Emotional overeating is related to cognitive alterations in two aspects: the inability to regulate emotional states and the inability to limit consumption of especially ultra-processed foods when in a compromised emotional state71. According to the self-medication hypothesis, individuals with difficulties in self-regulation turn to specific actions or external substances to help relieve the negative emotional impact of stimuli, resulting in substance abuse disorders72,73. In the case of food addiction, females with food addiction utilize high sugar/high fat foods as their sources of stimulation, in an attempt to lessen their emotional load57. The second aspect of emotional overeating, the reduced capability to limit intake of ultra-processed foods, is a result of lowered cognitive inhibitions created by heightened reactivity to food cues and attenuated satiety responses74,75,76 .

According to the incentive salience model, the motivational value of a “reward” is based on trigger cues, such as the sight, smell, and taste of ultra-processed foods77. In cases of addiction, other rewards are perceived to have a diminished incentive value relative to the drug causing compulsive behavior, reorientation of attentional resources, and downregulation of cognitive control regions57,78. This creates an increased reward being placed on ultra-processed foods, resulting in individuals seeking out these types of food beyond their basic homeostatic needs and even when their consumption leads to negative consequences57. In addition, due to the chronic consumption of high sugar/high fat foods, satiety signals in these individuals are compromised, as the palatability of ultra-processed food stimuli overrides an individual’s energy need79. These altered satiety cues in conjunction with increased response to food cues translate into more frequent and greater overeating behaviors.

Our results support this emotional overeating model, with females with food addiction exhibiting greater connectivity between the emotional regulation and salience networks compared to males with food addiction. Greater activation of the emotional regulation network aligns with the self-medication hypothesis, as females with food addiction engage in uncontrollable eating behaviors as an artificial coping mechanism to manage their emotional response to negative stimuli, similar to that of a drug46,80,81,82,83,84. The tendency of females with food addiction to actively seek out and consume ultra-processed foods in response to negative emotional stimuli, could refer to the relationship between greater emotional instability and increased food-seeking behavior as predicted in our model67,85.

The greater connectivity observed in the salience network, particularly between the salience and default mode network, is similarly consistent with our hypothesis and the incentive salience model77,86. The salience network adjusts attentional resources to salient sensory stimuli, with the observed greater activation suggesting an increased attentional value on food cues12,81,87,88,89,90. As females with food addiction place greater value and attention towards ultra-processed foods, the “incentive value” of this stimuli increases, resulting in a greater subconscious food focus and food cravings77,86,91. Additionally, this greater preoccupation towards ultra-processed foods, seen in females with food addiction, may also cause inhibition of cognitive regions in areas of the default-mode network leading to increased decisional impulsivity and lowered inhibitory responses in response to salient stimuli92,93.

Compared to males with food addiction, females with food addiction had greater connectivity among the sensorimotor network, particularly in the regions controlling evaluation of external stimuli94,95,96,97. These regions, namely the precentral gyrus and paracentral lobule, have been linked with altered explicit memory and inappropriate cognitive evaluations, potentially resulting in an increased motivational reward placed on food-related stimuli in females with food addiction98,99.

Additionally, when compared to females with no food addiction, those with food addiction exhibited lower connectivity in the sensorimotor network in regions involved in inhibition and attentional control, such as the inferior frontal gyrus100. Hypoactivation in the inferior frontal gyrus suggests increased impulsivity and the reinforcement of habit-forming systems, leading to weakened attempts to disengage with compulsive eating behaviors101.

Consistent with our model, females with food addiction had lowered connectivity within the default mode network compared to males with food addiction, potentially explaining the reduced cognitive control and greater preoccupation with food-related stimuli predicted in females with food addiction102. The default mode network plays a major role in self-generated and subconscious thought, displaying greater activation when resting103,104. Specifically, the dorsal medial prefrontal cortex subsystem has shown to preferentially activate during subconscious decision-making processes regarding one’s present mental state or situation105,106. Lower activation in this subsystem could translate to decreased subconscious inhibitions and altered attentional processing as a result of the top-down inhibition on cognitive control regions102,107,108. Weakened cognitive control circuits, upon interacting with the reward and emotional regulation system, reorient attentional resources in those with food addiction and create a vicious cycle of chronic overconsumption, impaired appetite regulation, and heightened food focus109. Additionally, consistent with the incentive salience model, our results showed lowered connectivity between the default mode network and the central executive network in females with food addiction, indicating that the lowered awareness of the present state results in less sensitive internal body cues and chronic food consumption behaviors as predicted in our hypothesis74,75,76.

Limitations and future research

Due to the cross-sectional design of the study, we were unable to address questions of causality between functional connectivity differences and food addiction. Future longitudinal studies are needed to determine if the observed connectivity differences between brain networks in individuals with food addiction are a premorbid state, or if they are a consequence of food addiction and associated metabolic changes. The integration of systemic inflammatory markers and metabolites derived from gut microbiota as mediators can help gain a more comprehensive understanding of food addiction in future mechanistic studies. While this study focused on resting-state connectivity differences, future research should consider investigating connectivity differences in response to ultra-processed food cues in individuals with food addiction. In order to be able to combine data to obtain a large sample size for subgroup analyses, all participants in this study completed the earlier version of the YFAS vs. the updated YFAS 2.0 questionnaire which is not only more psychometrically sound but has a stronger threshold of FA associated with obesity. This may have contributed to a reduced sensitivity in observing significant results, which will need to be validated in future studies using the YFAS 2.0.

Conclusions and clinical implications

When viewed together with previous findings, our results show greater connectivity in reward regions, indicative of the altered function of the reward circuit in individuals with food addiction. Sex differences in functional connectivity reveal that females with food addiction engage in more emotional eating behaviors while males with food addiction exhibit greater cognitive control and homeostatic processing. Since connectivity differences differ among males and females, this study contributes to the understanding of the nuances driving the sex-specific pathophysiology of food addiction. These mechanistic pathways may have clinical implications for understanding the variability in response to diet interventions and the need for more effective, targeted treatments for those with food addiction, especially females. Most clinical trials do not report sex differences related to treatment responses, but a few existing reports suggest that women are less likely to complete treatment, tend to lose less weight than men, have greater unsuccessful attempts to maintain weight loss resulting in the well-known “YoYo” diet phenomenon, and have limited responses to pharmacological treatments12,110,111. Therefore, treatments need to focus on the different sex-related eating patterns, such as women gaining weight via eating ultra-processed foods more frequently during emotional or stressful times, as compared to men, who gain weight via the consumption of larger meals. Therapeutic approaches would need to target different nodes of the brain related to food addiction, and individualization of treatments would need to be based on sex-related differences in order to improve greater clinical benefits13,112.