Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 May 6;59(9):1983–1991. doi: 10.1002/eat.70081

Clinical Phenotypes and Neurobiology of Youth With Avoidant/Restrictive Food Intake Disorder (ARFID) Across Sex

Ethiopia D Getachew 1,2,3, Marie‐Louis Wronski 1,3,4, Avery L Van De Water 3, Lauren Breithaupt 3,5,6, Kendra Becker 3,5,6, Helen Burton‐Murray 3,5,6,7, P Evelyna Kambanis 3,5,6, Madhusmita Misra 1,3,8, Kamryn Eddy 3,5,6, Franziska Plessow 1,2,3,9, Nadia Micali 10,11,12, Laura M Holsen 2,3,13,14, Jennifer J Thomas 3,5,6, Elizabeth A Lawson 1,2,3,✉
PMCID: PMC13374036  NIHMSID: NIHMS2190378  PMID: 42089273

ABSTRACT

Objective

Avoidant/restrictive food intake disorder (ARFID) is an eating disorder characterized by persistent avoidant/restrictive eating unrelated to body image. Sex differences in clinical phenotypes and neurobiology of ARFID are understudied. We hypothesized that females with ARFID would have a greater frequency of the lack of interest and fear of aversive consequences phenotypes, higher levels of anorexigenic hormones, and greater fMRI activation in brain regions associated with cognitive control during a food cue paradigm. We further hypothesized that males with ARFID would have a greater frequency of the sensory sensitivity profile, higher levels of orexigenic hormones, and greater fMRI activation in reward processing brain regions.

Method

We recruited 96 children and adolescents with ARFID and sub‐threshold ARFID (49% female) from two studies on the neurobiology of ARFID and low weight eating disorders from 2016 to 2022. We analyzed ARFID clinical phenotypes; appetite‐regulating hormones, ghrelin, cholecystokinin (CCK), peptide YY (PYY), and oxytocin; and fMRI activation of cognitive control and reward‐related brain regions during a food cue paradigm using frequentist and Bayesian statistical analysis.

Results

Contrary to our hypothesis, there were no sex differences in frequency of ARFID clinical phenotypes, appetite‐regulating hormone levels, or brain activation.

Conclusions

This is the first study to provide empirical evidence that ARFID has a similar clinical and neurobiological presentation in males and females.

Keywords: appetite, ARFID, CCK, eating disorders, neuroimaging, sex differences

Summary

  • This is the first study to examine sex differences in ARFID clinical phenotypes, appetite‐regulating hormones, and brain activity.

  • We found no sex differences in ARFID profiles, hormone levels (ghrelin, peptide YY, cholecystokinin, and oxytocin), or activation of cognitive control and reward‐related brain regions.

  • Findings provide strong evidence that ARFID presents similarly in males and females and underscore the importance of unbiased screening for ARFID across sex.

1. Introduction

Avoidant/restrictive food intake disorder (ARFID) is an eating disorder introduced over a decade ago in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, (DSM‐5) to provide diagnostic specificity for individuals who have avoidant/restrictive eating behavior unrelated to body image and display at least one of three core clinical phenotypes: lack of interest in eating or food, fear of aversive consequences related to food intake (e.g., choking, vomiting), and avoidance based on sensory characteristics of food (e.g., taste, texture, smell) (American Psychiatric Association 2022). Inability to meet nutritional needs in ARFID leads to significant complications and is associated with several psychiatric and medical comorbidities including delayed onset of puberty and depressive and anxiety symptoms (Seetharaman and Fields 2020).

Epidemiological findings suggest that ARFID occurs approximately equally in males and females, distinguishing it from other eating disorders, such as anorexia nervosa (AN), bulimia nervosa (BN), and binge‐eating disorder (BED), which are markedly more prevalent in females (versus males) (American Psychiatric Association 2013; Eddy et al. 2015; Fisher et al. 2014; Katzman et al. 2021; Kinasz et al. 2016; Nagl et al. 2016). In these latter disorders, sex differences have been well‐documented, not only in prevalence but also in clinical presentation and biological profiles. For instance, females with AN present with greater drive for thinness, whereas males typically present with greater drive for muscularity (Karazsia et al. 2017; McCreary and Sasse 2000). Moreover, neuroimaging studies have revealed sex‐specific biological correlates with one study showing widespread reductions in connectivity from reward‐related regions in males compared to more localized prefrontal reductions seen in females with BED (Murray et al. 2023). These findings illustrate that sex can influence both behavioral expression and underlying biology across eating disorders. However, whether similar sex‐related differences exist in ARFID remain unclear.

Emerging evidence suggests possible sex differences across ARFID phenotypes, though findings have been inconsistent (Katzman et al. 2021; Reilly et al. 2019; Zickgraf et al. 2019). For example, one large scale pediatric surveillance study in Canada showed that females had higher rates of “eating but not enough” while males had significantly higher rates of “meal refusal due to sensory sensitivity characteristics” (Katzman et al. 2021). These observations are consistent with the clinical impression—supported by some literature, but largely anecdotal among practitioners—that females are more likely to present with the fear of aversive consequences phenotype, whereas males more commonly present with the sensory sensitivity phenotype. Such differences could reflect broader epidemiologic patterns of a higher prevalence of anxiety disorders in females and a higher prevalence of neurodevelopmental disorders and associated traits among males (McLean et al. 2011; Zeidan et al. 2022). Indeed, in this cohort, we have previously demonstrated that severity in the fear profile is associated with higher likelihood of comorbid anxiety‐related disorders, while severity in the sensory sensitivity profile is associated with greater likelihood of comorbid neurodevelopmental disorders (Kambanis et al. 2020). This observational parallel may point toward partially overlapping pathways that influence the expression of avoidant or restrictive eating behaviors.

The neurobiology underlying appetite regulation offers a useful framework for examining these potential sex‐linked mechanisms. In BED, where, similar to ARFID and unlike other EDs, sex ratios are less stark, emerging research has begun to identify meaningful sex differences in neuroendocrine pathways (Murray et al. 2023). Understanding their role in ARFID may provide insight into why and how males and females develop and maintain the disorder differently (Culbert et al. 2021). Clinically, this knowledge could inform sex‐sensitive strategies for accurate and unbiased diagnosis, targeted interventions, and personalized treatment strategies.

Recent work in ARFID has identified alterations in appetite‐regulating pathways that may contribute to early satiation, high inter‐meal satiety and post‐prandial fullness (Becker et al. 2021; Burton Murray et al. 2022). Specifically, ghrelin, an orexigenic (appetite‐stimulating) hormone, has been found inappropriately low in females with ARFID compared to weight‐matched females with AN, suggesting reduced hunger signaling in the context of undernutrition status (Becker et al. 2021). In the same study, levels of peptide YY (PYY), an anorexigenic (appetite‐suppressing) hormone, did not differ between low‐weight females with ARFID and healthy controls, indicating that appetite suppression signaling may not adapt appropriately to a low weight state (Becker et al. 2021). Across the weight spectrum, children and adolescents with full/subthreshold ARFID displayed significantly higher levels of cholecystokinin (CCK), another anorexigenic hormone, compared to healthy controls (Burton Murray et al. 2022). Though the functional significance of this elevation remains speculative, we suspect that higher CCK in ARFID may contribute to reduced intake, particularly in individuals with a lack of interest profile. Finally, fasting and postprandial levels of oxytocin, another anorexigenic hormone, were previously found to be elevated in youth with ARFID versus healthy controls (Aulinas et al. 2023). Collectively, these findings point toward dysregulation of both hunger and satiety signals, though whether these patterns differ by sex remains unknown. In studies of healthy adolescents and adults, females had higher plasma oxytocin levels than males and largely showed a lack of sex differences in fasting ghrelin and PYY levels, with limited data on CCK showing variable results (Asarian and Geary 2013; Horner and Lee 2015; Marazziti et al. 2019). Understanding whether similar or distinct patterns emerge in ARFID may help clarify mechanisms underlying its phenotypic variability.

Neuroimaging studies in healthy individuals using functional magnetic resonance imaging (fMRI) have also demonstrated sex‐dependent activations in response to visual food cues: females, relative to males, showed higher activation in the orbitofrontal cortex (OFC; involved in valuation of hedonic foods and modulating reward pathways) and lateral prefrontal cortex (LPFC, associated with cognitive control), and lower activation in the right hippocampus (which plays a role in reward processing as well as memory for food‐related cues) (Cornier et al. 2010; Frank et al. 2010; Killgore and Yurgelun‐Todd 2010; Luo et al. 2019). Though there is nascent research on activation of food motivation circuitry in ARFID, one study showed significant activation in the OFC and anterior insula (involved in food cue valuation) in response to high‐calorie foods in individuals with ARFID with overweight or obesity versus those with normal weight (Kerem et al. 2022). Another recent study identified greater frontal cortical thickness (involved in executive functioning) among children with ARFID symptoms compared to those without, emphasizing the relevance of neural circuits in ARFID and the need to explore sex‐related differences in these circuits (Sader et al. 2025).

Taken together, these findings highlight gaps in our understanding of whether sex differences exist at multiple levels, including phenotypes, appetite‐regulating hormones, and neural activation. Addressing this gap is essential for identifying sex‐related biomarkers that may guide treatment response and for improving early identification by linking sex‐specific biological markers with phenotypic presentation. Accordingly, we investigated sex differences in (1) ARFID clinical phenotypes, (2) appetite‐regulating hormone (ghrelin, CCK, PYY, and oxytocin) levels, and (3) activation in brain appetite regions (OFC, LPFC, and right hippocampus). We hypothesized that, consistent with clinical observations, females would have both higher continuous scores and exhibit a greater frequency of the lack of interest and fear of aversive consequences phenotypes, while males would have higher scores and a greater frequency of the sensory sensitivity profile. Further, given that trends in ARFID phenotypes demonstrate higher rates of reduced food intake in females (Katzman et al. 2021; Reilly et al. 2019; Zickgraf et al. 2019), we hypothesized that we would observe lower levels of orexigenic ghrelin and higher levels of anorexigenic hormones (PYY, CCK, and oxytocin) in females compared to males. We further hypothesized that in response to a visual food cue fMRI paradigm in a fasted state, females (vs. males) with ARFID would demonstrate greater activation in response to high‐calorie food (vs. non‐food) images in the LPFC, a cognitive control region. Conversely, we hypothesized that males (vs. females) would demonstrate greater activation in the reward processing regions (OFC and right hippocampus).

2. Method

2.1. Participants

We recruited 96 adolescents and young adults with ARFID (n = 88) and subthreshold ARFID (n = 12) from an NIH‐funded longitudinal study on the neurobiology of ARFID and low‐weight eating disorders, with females comprising 49% of the cohort. We recruited participants to achieve approximately equal numbers of males and females, as well as balanced representation across age groups. This cohort has been described in prior publications from the larger study but has not been analyzed for sex differences as presented in this study (Aulinas et al. 2023; Becker et al. 2021; Burton Murray et al. 2022; Kambanis et al. 2020; Kerem et al. 2022; Thomas et al. 2025). The majority of participants identified as white (92%) and non‐Hispanic (90%). There was no difference between females and males in age ([mean ± SD] 15.9 ± 3.9 years; range 9–23 years) nor body mass index (BMI; participants > 18 years) or BMI z‐score (participants < 17.9 years; Table 1). All participants met criteria for ARFID on the Eating Disorder Assessment for DSM‐5 (EDA‐5) (Sysko et al. 2015) or endorsed ARFID symptoms on the Kiddie Schedule for Affective Disorders and Schizophrenia‐Present and Lifetime (KSADS‐PL) (Kaufman et al. 1997). Participants completed the Pica, ARFID, and Rumination Disorder Interview (PARDI) (Bryant‐Waugh et al. 2019) and were assigned a profile score ranging from 0 (No symptoms) to 6 (Extreme severity) for each subtype. Individuals who endorsed ARFID symptoms on the PARDI but not at the severity required by the PARDI algorithm (e.g., endorsing clinical impairment at a severity of 2/6 rather than 4/6) were assigned a diagnosis of subthreshold ARFID. We identified those who met criteria for the lack of interest and sensory sensitivity ARFID phenotypes using the suggested cut‐offs by Cooper‐Vince and colleagues and for the fear of aversive consequences phenotype using the natural break in the distribution (Cooper‐Vince et al. 2022). We excluded participants with psychosis history, active substance‐use disorder, pregnancy, breastfeeding, or systemic hormonal use within 8 weeks prior to the study visit, gastrointestinal tract surgery history, hematocrit < 30%, and a history of other feeding or eating disorders. The study was approved by the Mass General Brigham Human Research Committee and conducted in accordance with the Declaration of Helsinki. Study clinicians obtained written informed consent or assent (for participants < 18 years). Subjects completed outpatient screening at the Massachusetts General Hospital (MGH) Translational and Clinical Research Center (TCRC) and the main visit at the Athinoula A. Martinos Center for Biomedical Imaging. The main study visit took place within 3 months of the screening visit.

TABLE 1.

Demographic and clinical characteristics of study participants and fasting appetite‐regulating hormones by sex.

Male Female
No of participants (%) 49 (51.0) 47 (49.0)
Age, mean (SD) 15.3 (3.5) 16.6 (4.2)
Race
American Indian/Alaskan Native 0 (0) 0 (0)
Asian/Asian American 0 (0) 1 (1)
Black/African American 1 (1) 1 (1)
Native Hawaiian/Pacific Islander 0 (0) 0 (0)
White/Caucasian 48 (50) 40 (42)
More than one race 0 (0) 5 (5)
Ethnicity
Hispanic 2 (2) 8 (8)
Non‐Hispanic 27 (28) 39 (41)
BMI kg/m2, mean (SD)* (n = 28) 20.7 (6.6) 25.3 (7.2)
BMI z‐score, mean (SD)* (n = 68) −0.79 (1.6) −0.73 (1.2)
ARFID phenotypes (categorical yes/no)**
Lack of interest (%) 33 (67) 31 (66)
Fear of aversive consequences (%) 10 (20) 14 (30)
Sensory sensitivity (%) 35 (71) 40 (85)
ARFID phenotypes (continuous scores)**
PARDI lack of interest severity 2.0 (1.5) 2.1 (1.6)
PARDI fear of aversive consequences severity 0.4 (0.7) 0.5 (0.9)
PARDI sensory sensitivity severity 1.4 (1.1) 2.0 (1.4)
ARFID severity 2.3 (0.8) 2.4 (0.9)
Hormones**
Ghrelin pg/mL, mean (SD) 5418 (2452) 5164 (2384)
Ln‐transformed ghrelin 8.5 (0.5) 8.4 (0.5)
PYY pg/mL, mean (SD) 79.9 (45.5) 88.8 (49.9)
Ln‐transformed PYY 4.2 (0.6) 4.3 (0.6)
CCK pg/mL, mean (SD) 317.9 (338.3) 174.3 (90.1)
Ln‐transformed CCK 5.5 (0.8) 5.1 (0.4)
Oxytocin pg/mL, mean (SD) 566.2 (280.8) 566.0 (221.4)
Ln‐transformed OXT 6.3 (0.4) 6.3 (0.3)
Open in a new tab

Note: Groups were compared using two‐sample tests.

Abbreviations: BMI = body mass index; CCK = cholecystokinin; PYY = peptide YY.

*

We utilized BMI for young adults ages 18–23 years and BMI z‐scores for children and adolescents younger than 18 years.

**

Adjusted p values for anorexigenic hormones (PYY, CCK, and oxytocin) and PARDI phenotypes were calculated using the Bonferroni correction for multiple comparisons (adjusted α = 0.0167).

We measured participants' height and weight at the screening visit on a wall‐mounted stadiometer and electronic scale in triplicate and calculated BMI. We defined male/female sex as biological (birth) sex as noted by the participant. Trained study personnel administered the KSADS‐PL to assess current and lifetime psychiatric diagnoses, including history of psychosis and substance or alcohol use disorder within the past month; and Eating Disorder Examination Questionnaire (EDE‐Q) (Fairburn and Beglin 2008) to assess clinically significant disordered eating in the past 28 days (defined as Global Score > 4.0). Participants arrived for the main visit following a 10 h fast and underwent a fasting fMRI scanning session at approximately 7:30 a.m. and a fasting blood draw around 8:45 a.m.

2.2. Appetite‐Regulating Hormones

We assessed appetite‐regulating hormones in a subset of 92 participants (46 female) for whom sufficient blood samples were available. Consistent with standard laboratory protocols, blood samples were immediately placed on ice, then processed in a chilled centrifuge to separate plasma, and stored at −80°C until analysis. We used enzyme‐linked immunosorbent assays to assess serum PYY levels (EMD Millipore, Burlington, MA; intra‐assay CV 17%–18% and inter‐assay CV 12%–18%; lower limit of detection 10.0 pg/mL), plasma total ghrelin (EMD Millipore, Burlington, MA; intra‐assay CV 1.32% and inter‐assay CV 6.62%; lower limit of detection 50.0 pg/mL), plasma total CCK (RayBio, Peachtree Corners, GA; intra‐assay CV < 10% and inter‐assay CV < 15%; lower limit of detection 0.2 pg/mL), and unextracted serum oxytocin levels (Enzo Life Sciences, Farmingdale, NY; intra‐assay CV 10.2%–13.3% and inter‐assay CV 11.8%–20.9%; lower limit of detection 15.0 pg/mL).

2.3. Neuroimaging Protocol

Participants completed fMRI imaging in a Food and Drug Administration (FDA)‐approved 3T Skyra scanner while viewing images of foods, non‐food items, and fixation stimuli, using a well‐established and validated visual food cue paradigm (Siemens/Erlangen/Germany). A full description of the sequence parameters and food cue paradigm is provided in Thomas et al. (2025).

3. Statistical Analyses

3.1. Clinical Phenotype Analyses

We performed descriptive statistics, reporting means ± SD for continuous variables and frequencies (percentages) for categorical variables. We confirmed data normality using Shapiro‐Wilks tests. We used independent‐samples t‐tests to compare male and female participants on continuous PARDI subscale scores for each ARFID phenotype (i.e., sensory sensitivity, lack of interest, and fear of aversive consequences) to assess severity. We compared categorical phenotype presence based on the defined cutoffs above using chi‐square tests and computed effect sizes (Cohen's d). We applied a Bonferroni‐adjusted significance threshold of p < 0.0167 across the three phenotypic comparisons. In addition to the frequentist (traditional) analysis framework, we performed Bayesian general linear models to follow up on nonsignificant group differences to evaluate evidence in favor of the null hypothesis. For Bayesian testing, we used default weakly informative priors (R package “rstan”) and 10 chains á 5000 iterations (warm‐up period of 1000 iterations). Bayes Factor > 3 and Bayes factor > 10 respectively indicate moderate and strong evidence in favor of the null hypothesis.

3.2. Hormone Analysis

Hormone distributions were non‐normal; therefore, we applied a natural logarithm transformation to the hormone data to approximate a normal distribution and used single‐value imputation for hormone values below the lower limit of detection. Extreme outliers (> 3 standard deviations from the mean of hormone concentration) were identified in the hormones CCK (one male, one female), oxytocin (two males) and PYY (one male, one female) and excluded. We calculated Cohen's d to estimate effect size and applied a Bonferroni correction (p < 0.0167) across anorexigenic hormones CCK, oxytocin, and PYY comparisons. As above, in addition to frequentist analyses, we conducted Bayesian statistical models to assess evidence for or against sex differences in hormone levels. We conducted all statistics for participant characteristics and hormone analysis using R version 4.1.2 (Vienna, Austria).

3.3. Neuroimaging Analyses

We analyzed fMRI data using Statistical Parametric Mapping software (SPM12; Wellcome Trust Centre for Neuroimaging). We realigned and unwarped volumes using phase correction from the fieldmap, slice‐time corrected, and co‐registered to a bias‐corrected mean image matched to a skull‐stripped, segmented T1‐weighted protocol. Following this, we performed normalization to the Montreal Neurological Institute (MNI) 152 brain template, and the data were re‐sampled to 3 mm isotropic and smoothed with a 6 mm Gaussian kernel. We used Artifact Detection Tools (ART) to detect outliers in global mean image time series (threshold: 3.5 SD) and movement (threshold: 0.8 mm, scan‐to‐scan movement) and entered as nuisance regressors in the single‐subject level General Linear Model (GLM) (“NITRC: Artifact Detection Tools (ART): Tool/Resource Info,” NITRC n.d.). We used masks excluding voxels outside the brain to ensure that voxels in regions with signal dropout were not inadvertently excluded. Eighty‐two participants completed the full set and had useable, non‐motion degraded fMRI data as defined above.

For the block design, we modeled each stimulus using a boxcar function convolved with a canonical hemodynamic response function. We selected high‐calorie foods vs. objects as the contrast of interest based on prior findings using similar contrasts of sex differences in reward and cognitive control regions and our prior findings of reward circuitry in individuals with ARFID and overweight/obesity (Cornier et al. 2010; Kerem et al. 2022). We then used linear contrasts and SPM t‐maps from the first‐level analysis and submitted these contrasts for second‐level random effects group analysis. At the group level, we examined effects of interest using independent sample t‐tests (males vs. females at premeal). We analyzed region of interest (ROI) analyses using small volume correction with a priori regions including the LPFC, right hippocampus, and OFC. The right hippocampus and OFC ROIs were selected defined anatomically using the Automated Anatomic Labeling (AAL)‐3 atlas44. The LPFC ROI was based on the coordinates reported by Cornier et al. (2010) (x, y, z = 48, 39, 15) and defined functionally using a 10 mm sphere centered around these coordinates. We controlled for multiple comparisons using a combination of cluster extent (k ≥ 20) and p < 0.05 Family‐Wise Error (FWE)‐corrected thresholds and calculated effect sizes (Cohen's d). We additionally performed secondary exploratory whole‐brain analyses. Finally, we extracted average parameter estimates (betas) for each anatomically pre‐defined ROI from peak activation clusters using the Region of Interest Extraction Toolbox (REX) and exported to SPSS (v19, Chicago, IL) for graphical depiction.

4. Results

4.1. ARFID Clinical Phenotypes in Males Versus Females

Contrary to our hypotheses, males did not differ significantly from females on the severity of any ARFID phenotypes including sensory sensitivity (p = 0.019, Cohen's d = 0.50), lack of interest (p = 0.866, Cohen's d = 0.04), or fear of aversive consequences (p = 0.246, Cohen's d = 0.25), as measured by continuous PARDI subscales (Table 1). Females showed a trend toward higher PARDI sensory sensitivity scores compared to males, but this difference was not statistically significant after Bonferroni correction (p = 0.019). Furthermore, contrary to our hypothesis, when we applied categorical cut points to each ARFID phenotype, males did not differ significantly from females on the frequency of any phenotype. Bayesian analysis corroborated these results for the sensory sensitivity (Bayes factor = 1.37), lack of interest (Bayes factor = 23.0) and fear of aversive consequences (Bayes factor = 11.0) profiles. Across both sexes, sensory sensitivity was the most common phenotype (affecting 78%), followed by lack of interest (67%), with fear of aversive consequences being the least common (25%) (Table 1).

4.2. Appetite‐Regulating Hormone Levels in Males Versus Females

In contrast to our hypothesis, there were no statistically significant differences between females and males in levels of the orexigenic hormone ghrelin (p = 0.608, Cohen's d = 0.11). Among the anorexigenic hormones, we similarly found no significant sex differences after Bonferroni correction in PYY (p = 0.348, Cohen's d = 0.20), CCK (p = 0.018, Cohen's d = 0.51), and oxytocin (p = 0.730, Cohen's d = 0.08); although CCK showed a trend toward higher levels in males compared to females, in disagreement with our hypothesis that females would have higher levels of anorexigenic hormones (Table 1, Figure 1). Bayesian statistics yielded strong evidence in favor of the absence of sex differences in ghrelin (Bayes factor = 19.7), PYY (Bayes factor = 15.8), and oxytocin (Bayes factor = 22.5). The trend toward significance of CCK was corroborated with Bayes analysis (Bayes factor = 1.43).

FIGURE 1.

FIGURE 1

Open in a new tab

Fasting appetite‐regulating hormones in female and male youth with avoidant/restrictive food intake disorder. Note: Line plots display mean concentrations and standard deviations (ln‐transformed) of appetite‐regulating hormones (A) ghrelin, (B) peptide YY (PYY), (C) cholecystokinin (CCK), and (D) oxytocin.

4.3. Neural Activation to Food Cues in Males Versus Females

Contrary to our hypotheses, there were no significant sex differences in the appetitive neural circuitry during the food cue paradigm. Specifically, there were no differences in our ROI (OFC: p[FWE‐corrected] > 0.50, Cohen's d = 0.11; R LPFC: no suprathreshold clusters, or R hippocampus: p[FWE‐corrected] = 0.29, Cohen's d = 0.15) or in the secondary whole‐brain analysis (cluster p[FWE‐corrected] > 0.5 for the male > female contrast and no suprathreshold clusters for the contrast of female > male) (Table 2). Complementing frequentist results, Bayesian generalized linear models indicated substantial evidence (Bayes Factors > 3 [threshold]) in favor of the absence of sex differences in all examined brain regions.

TABLE 2.

Activation of reward and cognitive control brain regions to high‐calorie food stimuli compared with nonfood objects (secondary exploratory whole‐brain analysis).

Model Contrast Region of interest BA MNI x MNI y MNI z Cluster size (voxels) t p Cohen's d BF01
Two‐sample t‐test Female > male Lateral prefrontal cortex/BA 10 No suprathreshold clusters NA 12.69
Male > female Orbitofrontal cortex 9 35 −22 10 2.82 > 0.500 0.11 19.88
Right hippocampus 24 −37 8 13 2.7 0.290 0.15 17.84
Open in a new tab

5. Discussion

This was the first study to evaluate potential sex differences in the clinical phenotypes and neurobiological underpinnings of ARFID. Contrary to our hypothesis, we found no sex differences in the severity or presence of ARFID phenotypes. Similarly, we found no sex differences in levels of orexigenic (ghrelin) or anorexigenic (PYY, CCK, oxytocin) hormones, or in activation of regions related to reward and cognitive control. Bayesian analyses showed the same pattern of findings, thus providing substantial evidence for the frequentist negative results.

The high degree of similarity in clinical phenotype and neurobiology of ARFID across sex is notable among eating disorders. Other feeding and eating disorders—particularly AN, BN, and BED—are significantly more common in females than males and often have sex‐specific clinical presentations, with females scoring higher than males on many indices of eating disorder psychopathology (Zayas et al. 2018). In contrast, prior epidemiological work has shown similar ARFID prevalence among males and females. Our findings build on this literature by demonstrating that, within a sex‐balanced clinical cohort, ARFID clinical phenotypes likewise do not differ by sex. Indeed, contrary to our hypotheses about the differential sex distribution of the ARFID phenotypes, we found no sex differences in the presence of ARFID profiles, underscoring the importance of a broad and unbiased approach to ARFID screening and diagnosis in clinical settings. Notably, there was a trend toward higher sensory sensitivity scores in females vs. males, which was contrary to our hypothesis based on clinical observations. Recent literature shows increasing recognition of neurodiversity in females, reflecting both true prevalence and improved detection in females (Harrop et al. 2024). Our findings, along with emerging literature, highlight that the historical assumption of higher male sensory sensitivity due to association with increased prevalence of neurodevelopmental disorders in males is overly simplistic. Future studies will be needed to further elucidate these patterns and to explore etiologies underlying sensory‐based aversions in ARFID, such as potential sex differences in primary sensory and association areas of the brain.

Contrary to our hypothesis, we also did not identify any sex differences in orexigenic or anorexigenic hormones, or in the neural circuitry examined. Though ghrelin, PYY, CCK, and oxytocin have emerged as important hormones in the pathophysiology of ARFID, it does not appear that they differ across sex within an ARFID sample. We observed, however, a borderline trend of higher levels of CCK in males versus females that was contrary to our hypothesis, as we expected higher rates of the lack of interest profile in females to be associated with greater levels of anorexigenic hormones. Similarly, contrary to our predictions, ARFID patients did not demonstrate sex differences in activation of brain regions, specifically the OFC, LPFC, and right hippocampus, that show sexually dimorphic responsivity to food cues in healthy individuals.

Strengths of this study include a balanced male–female sample of individuals with full and subthreshold ARFID and integration of clinical, hormonal and neuroimaging data. We utilized both traditional statistical methods as well as Bayes analysis to increase confidence in study results. Limitations include the heterogeneity introduced by psychotropic medication use and the wide age range of participants. Additional factors such as pubertal status (e.g., Tanner stage), body fat percentage and menstrual cycle were not fully accounted for and could further influence hormonal and neural mechanisms. However, approximately half of the participants in our study were male, and half of the female participants were prepubertal, which likely reduced the potential impact of menstrual status on the hormonal and neural findings. Future studies with larger sample sizes are needed to further validate these findings, especially given our trend level results for CCK (male > female) and sensory sensitivity profile (female > male), which could potentially reach statistical significance in a larger sample. Additionally, future studies should consider other neurobiological mechanisms such as the anorexigenic hormone leptin and neural circuits involving fear and sensory processing, as these may elucidate the role of sex in ARFID and its clinical phenotypes further.

6. Conclusion

This is the first study to investigate sex differences in appetitive regulation pathways that may contribute to the etiology and/or maintenance of ARFID and provide empirical evidence against sex differences in the ARFID profiles. We found no sex differences in the severity or frequency of ARFID clinical phenotypes, nor levels of orexigenic ghrelin, anorexigenic PYY, CCK, and oxytocin, nor in appetite regulating neural circuitry. These findings are largely consistent with epidemiological data that show minimal sexual dimorphism in ARFID, underscoring the similarity of clinical presentation and neurobiology across sex in the only feeding/eating disorder that appears to affect males and females at equal rates.

Author Contributions

Ethiopia D. Getachew: conceptualization, investigation, writing – original draft, formal analysis, writing – review and editing, validation, visualization, methodology, software. Marie‐Louis Wronski: writing – original draft, formal analysis, writing – review and editing, validation, visualization, methodology, software. Avery L. Van De Water: formal analysis, validation, visualization, methodology, software, data curation. Lauren Breithaupt: writing – review and editing, formal analysis, data curation. Kendra Becker: writing – review and editing, data curation. Helen Burton‐Murray: writing – review and editing. P. Evelyna Kambanis: conceptualization, writing – review and editing. Madhusmita Misra: writing – review and editing. conceptualization. Kamryn Eddy: conceptualization, writing – review and editing. Franziska Plessow: conceptualization, writing – review and editing, data curation. Nadia Micali: writing – review and editing. Laura M. Holsen: conceptualization, investigation, funding acquisition, writing – review and editing, formal analysis, supervision, resources, data curation. Jennifer J. Thomas: conceptualization, investigation, funding acquisition, writing – review and editing, formal analysis, supervision, resources, data curation. Elizabeth A. Lawson: conceptualization, investigation, funding acquisition, writing – review and editing, formal analysis, supervision, resources, data curation.

Lived Experience Involvement Statement

No specific efforts were undertaken to involve persons with lived experience in the study design or execution, or in the preparation of this manuscript.

Funding

This work was supported by the National Institute of Health: R01MH108595 (J.J.T., E.A.L., N.M.); K24MH120568 (E.A.L.); K24MH135189 (J.J.T.); P30DK04056 (Nutrition Obesity Research Center at Harvard); K23DK131334 (H.B‐M.).

Conflicts of Interest

E.A.L. receives grant support and research study drug from Tonix Pharmaceuticals and receives royalties from UpToDate. E.A.L. and/or immediate family member holds/recently held stock in Thermo Fisher Scientific, Zoetis, Danaher Corporation, Intuitive Surgical, Merck, West Pharmaceutical Services, Gilead Sciences, and Illumina. F.P. and E.A.L. are inventors on PCTUS2025/030536 entitled, “Oxytocin‐based therapeutics to improve cognitive control in individuals with attention deficit hyperactive disorder” filed on May 22, 2025. J.J.T. receives royalties from Harvard Health Publications and Hazelden for the sale of her book on anorexia nervosa. J.J.T. and K.E. receive royalties from Cambridge University Press for the sale of their books on ARFID. J.J.T., K.E., and P.E.K. receive consulting fees from Equip Health. J.J.T. and H.B‐.M. receive royalties from Oxford University Press for the sale of their book on rumination syndrome. M.M. receives royalties from UpToDate, has served as a consultant for Regeneron, and receives study drug donation from Amgen and from Tonix Pharmaceuticals. The other authors declare no conflicts of interest.

Acknowledgments

We thank our study participants and staff at the Massachusetts General Hospital (MGH) Translational and Clinical Research Center (TCRC) and the Athinoula A. Martinos Center for Biomedical Imaging.

Use of AI was not involved in any component of preparation of this manuscript.

Getachew, E. D. , Wronski M.‐L., Van De Water A. L., et al. 2026. “Clinical Phenotypes and Neurobiology of Youth With Avoidant/Restrictive Food Intake Disorder (ARFID) Across Sex.” International Journal of Eating Disorders 59, no. 9: 1983–1991. 10.1002/eat.70081.

Action Editor: Ruth Weissman

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

References

  1. American Psychiatric Association . 2013. Diagnostic and Statistical Manual of Mental Disorders. Vol. 5. American Psychiatric Association. [Google Scholar]
  2. American Psychiatric Association . 2022. Diagnostic and Statistical Manual of Mental Disorders 5‐TR. Vol. 5. American Psychiatric Association. [Google Scholar]
  3. Asarian, L. , and Geary N.. 2013. “Sex Differences in the Physiology of Eating.” American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 305, no. 11: R1215–R1267. 10.1152/ajpregu.00446.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Aulinas, A. , Muhammed M., Becker K. R., et al. 2023. “Oxytocin Response to Food Intake in Avoidant/Restrictive Food Intake Disorder.” European Journal of Endocrinology 189, no. 2: 149–155. 10.1093/ejendo/lvad087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Becker, K. R. , Mancuso C., Dreier M. J., et al. 2021. “Ghrelin and PYY in Low‐Weight Females With Avoidant/Restrictive Food Intake Disorder Compared to Anorexia Nervosa and Healthy Controls.” Psychoneuroendocrinology 129: 105243. 10.1016/j.psyneuen.2021.105243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bryant‐Waugh, R. , Micali N., Cooke L., Lawson E. A., Eddy K. T., and Thomas J. J.. 2019. “Development of the Pica, ARFID, and Rumination Disorder Interview, a Multi‐Informant, Semi‐Structured Interview of Feeding Disorders Across the Lifespan: A Pilot Study for Ages 10‐22.” International Journal of Eating Disorders 52, no. 4: 378–387. 10.1002/eat.22958. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Burton Murray, H. , Becker K. R., Harshman S., et al. 2022. “Elevated Fasting Satiety‐Promoting Cholecystokinin (CCK) in Avoidant/Restrictive Food Intake Disorder Compared to Healthy Controls.” Journal of Clinical Psychiatry 83, no. 5: 21m14111. 10.4088/JCP.21m14111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cooper‐Vince, C. E. , Nwaka C., Eddy K. T., et al. 2022. “The Factor Structure and Validity of a Diagnostic Interview for Avoidant/Restrictive Food Intake Disorder in a Sample of Children, Adolescents, and Young Adults.” International Journal of Eating Disorders 55, no. 11: 1575–1588. 10.1002/eat.23792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cornier, M. A. , Salzberg A. K., Endly D. C., Bessesen D. H., and Tregellas J. R.. 2010. “Sex‐Based Differences in the Behavioral and Neuronal Responses to Food.” Physiology & Behavior 99, no. 4: 538–543. 10.1016/j.physbeh.2010.01.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Culbert, K. M. , Sisk C. L., and Klump K. L.. 2021. “A Narrative Review of Sex Differences in Eating Disorders: Is There a Biological Basis?” Clinical Therapeutics 43, no. 1: 95–111. 10.1016/j.clinthera.2020.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Eddy, K. T. , Thomas J. J., Hastings E., et al. 2015. “Prevalence of DSM‐5 Avoidant/Restrictive Food Intake Disorder in a Pediatric Gastroenterology Healthcare Network.” International Journal of Eating Disorders 48, no. 5: 464–470. 10.1002/eat.22350. [DOI] [PubMed] [Google Scholar]
  12. Fairburn, C. G. , and Beglin S. J.. 2008. “Eating Disorder Examination Questionnaire (EDE‐Q 6.0).” International Journal of Eating Disorders, 309–313. Guilford Press. [Google Scholar]
  13. Fisher, M. M. , Rosen D. S., Ornstein R. M., et al. 2014. “Characteristics of Avoidant/Restrictive Food Intake Disorder in Children and Adolescents: A “New Disorder” in DSM‐5.” Journal of Adolescent Health 55, no. 1: 49–52. 10.1016/j.jadohealth.2013.11.013. [DOI] [PubMed] [Google Scholar]
  14. Frank, S. , Laharnar N., Kullmann S., et al. 2010. “Processing of Food Pictures: Influence of Hunger, Gender and Calorie Content.” Brain Research 1350: 159–166. 10.1016/j.brainres.2010.04.030. [DOI] [PubMed] [Google Scholar]
  15. Harrop, C. , Tomaszewski B., Putnam O., Klein C., Lamarche E., and Klinger L.. 2024. “Are the Diagnostic Rates of Autistic Females Increasing? An Examination of State‐Wide Trends.” Journal of Child Psychology and Psychiatry 65, no. 7: 973–983. 10.1111/jcpp.13939. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Horner, K. , and Lee S.. 2015. “Appetite‐Related Peptides in Childhood and Adolescence: Role of Ghrelin, PYY, and GLP‐1.” Applied Physiology, Nutrition, and Metabolism 40, no. 11: 1089–1099. 10.1139/apnm-2015-0050. [DOI] [PubMed] [Google Scholar]
  17. Kambanis, P. E. , Kuhnle M. C., Wons O. B., et al. 2020. “Prevalence and Correlates of Psychiatric Comorbidities in Children and Adolescents With Full and Subthreshold Avoidant/Restrictive Food Intake Disorder.” International Journal of Eating Disorders 53, no. 2: 256–265. 10.1002/eat.23191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Karazsia, B. T. , Murnen S. K., and Tylka T. L.. 2017. “Is Body Dissatisfaction Changing Across Time? A Cross‐Temporal Meta‐Analysis.” Psychological Bulletin 143, no. 3: 293–320. 10.1037/bul0000081. [DOI] [PubMed] [Google Scholar]
  19. Katzman, D. K. , Spettigue W., Agostino H., et al. 2021. “Incidence and Age‐ and Sex‐Specific Differences in the Clinical Presentation of Children and Adolescents With Avoidant Restrictive Food Intake Disorder.” JAMA Pediatrics 175, no. 12: e213861. 10.1001/jamapediatrics.2021.3861. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kaufman, J. , Birmaher B., Brent D., et al. 1997. “Schedule for Affective Disorders and Schizophrenia for School‐Age Children‐Present and Lifetime Version (K‐SADS‐PL): Initial Reliability and Validity Data.” Journal of the American Academy of Child and Adolescent Psychiatry 36, no. 7: 980–988. 10.1097/00004583-199707000-00021. [DOI] [PubMed] [Google Scholar]
  21. Kerem, L. , Van De Water A. L., Kuhnle M. C., et al. 2022. “Neurobiology of Avoidant/Restrictive Food Intake Disorder in Youth With Overweight/Obesity Versus Healthy Weight.” Journal of Clinical Child and Adolescent Psychology 51, no. 5: 701–714. 10.1080/15374416.2021.1894944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Killgore, W. D. , and Yurgelun‐Todd D. A.. 2010. “Sex Differences in Cerebral Responses to Images of High Versus Low‐Calorie Food.” Neuroreport 21, no. 5: 354–358. 10.1097/WNR.0b013e32833774f7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kinasz, K. , Accurso E. C., Kass A. E., and Le Grange D.. 2016. “Does Sex Matter in the Clinical Presentation of Eating Disorders in Youth?” Journal of Adolescent Health 58, no. 4: 410–416. 10.1016/j.jadohealth.2015.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Luo, S. , Alves J., Hardy K., et al. 2019. “Neural Processing of Food Cues in Pre‐Pubertal Children.” Pediatric Obesity 14, no. 2: e12435. 10.1111/ijpo.12435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Marazziti, D. , Baroni S., Mucci F., et al. 2019. “Sex‐Related Differences in Plasma Oxytocin Levels in Humans.” Clinical Practice and Epidemiology in Mental Health 15: 58–63. 10.2174/1745017901915010058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. McCreary, D. R. , and Sasse D. K.. 2000. “An Exploration of the Drive for Muscularity in Adolescent Boys and Girls.” Journal of American College Health 48, no. 6: 297–304. 10.1080/07448480009596271. [DOI] [PubMed] [Google Scholar]
  27. McLean, C. P. , Asnaani A., Litz B. T., and Hofmann S. G.. 2011. “Gender Differences in Anxiety Disorders: Prevalence, Course of Illness, Comorbidity and Burden of Illness.” Journal of Psychiatric Research 45, no. 8: 1027–1035. 10.1016/j.jpsychires.2011.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Murray, S. B. , Alba C., Duval C. J., Nagata J. M., Ganson K. T., and Jann K.. 2023. “Sex Differences in Functional Connectivity From Reward‐Based Regions in Pre‐Adolescent Binge Eating Disorder.” Psychiatry Research 324: 115186. 10.1016/j.psychres.2023.115186. [DOI] [PubMed] [Google Scholar]
  29. Nagl, M. , Jacobi C., Paul M., et al. 2016. “Prevalence, Incidence, and Natural Course of Anorexia and Bulimia Nervosa Among Adolescents and Young Adults.” European Child & Adolescent Psychiatry 25, no. 8: 903–918. 10.1007/s00787-015-0808-z. [DOI] [PubMed] [Google Scholar]
  30. NITRC . n.d. “Artifact Detection Tools (ART): Tool/Resource Info.” https://www.nitrc.org/projects/artifact_detect/.
  31. Reilly, E. E. , Brown T. A., Gray E. K., Kaye W. H., and Menzel J. E.. 2019. “Exploring the Cooccurrence of Behavioural Phenotypes for Avoidant/Restrictive Food Intake Disorder in a Partial Hospitalization Sample.” European Eating Disorders Review 27, no. 4: 429–435. 10.1002/erv.2670. [DOI] [PubMed] [Google Scholar]
  32. Sader, M. , Harris H. A., Waiter G. D., Jansen P. W., Williams J. H. G., and White T.. 2025. “Neural Correlates of Children With Avoidant Restrictive Food Intake Disorder Symptoms: Large‐Scale Neuroanatomical Analysis of a Paediatric Population.” Journal of Child Psychology and Psychiatry 66, no. 6: 785–795. 10.1111/jcpp.14086. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Seetharaman, S. , and Fields E. L.. 2020. “Avoidant/Restrictive Food Intake Disorder.” Pediatrics in Review 41, no. 12: 613–622. 10.1542/pir.2019-0133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Sysko, R. , Glasofer D. R., Hildebrandt T., et al. 2015. “The Eating Disorder Assessment for DSM‐5 (EDA‐5): Development and Validation of a Structured Interview for Feeding and Eating Disorders.” International Journal of Eating Disorders 48, no. 5: 452–463. 10.1002/eat.22388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Thomas, J. J. , Holsen L., Van De Water A. L., et al. 2025. “Neural Response to Food Cues in Avoidant/Restrictive Food Intake Disorder.” JAMA Network Open 8, no. 2: e2460101. 10.1001/jamanetworkopen.2024.60101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Zayas, L. V. , Wang S. B., Coniglio K., et al. 2018. “Gender Differences in Eating Disorder Psychopathology Across DSM‐5 Severity Categories of Anorexia Nervosa and Bulimia Nervosa.” International Journal of Eating Disorders 51, no. 9: 1098–1102. 10.1002/eat.22941. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Zeidan, J. , Fombonne E., Scorah J., et al. 2022. “Global Prevalence of Autism: A Systematic Review Update.” Autism Research 15, no. 5: 778–790. 10.1002/aur.2696. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Zickgraf, H. F. , Lane‐Loney S., Essayli J. H., and Ornstein R. M.. 2019. “Further Support for Diagnostically Meaningful ARFID Symptom Presentations in an Adolescent Medicine Partial Hospitalization Program.” International Journal of Eating Disorders 52, no. 4: 402–409. 10.1002/eat.23016. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Articles from The International Journal of Eating Disorders are provided here courtesy of Wiley

RESOURCES