Neural Response to Food Cues in Avoidant/Restrictive Food Intake Disorder

Importance: The neurobiology of avoidant/restrictive food intake disorder (ARFID) is poorly understood. Objective: To evaluate whether individuals with ARFID exhibit disruptions in fear, appetite, and disgust brain regions compared with healthy control (HC) participants when shown images of food and objects. Design, Setting, and Participants: In this case-control study conducted from July 2016 to January 2021, children, adolescents, and young adults completed structured interviews and a validated functional magnetic resonance imaging (fMRI) food cue paradigm. The study was conducted at a single academic medical center. Data analysis was conducted from April 2023 to August 2024. Exposures: Presence vs absence of ARFID and its phenotypes (ARFID-fear, ARFID-lack of interest in eating, ARFID-sensory sensitivity); pictures of food vs objects during fMRI food cue paradigm. Main Outcomes and Measures: Blood oxygenation level-dependent activation in regions of interest (ROIs; amygdala, hypothalamus, insula, anterior cingulate cortex [ACC]) and the whole brain. Results: Participants were 110 children, adolescents, and young adults with full or subthreshold ARFID (75 participants; mean [SD] age, 16.2 [3.8] years; 41 [55%] female) and age-matched HC participants (35 participants; mean [SD] age, 17.3 [4.0] years; 27 [69%] female) recruited for studies of the neurobiology of ARFID and restrictive eating disorders. Participants with ARFID demonstrated greater activation than HC participants of the ACC (mean difference, 0.48 [95% CI, 0.19 to 0.77]; P = .009), sensory association cortex (mean difference on left side, 0.54 [95% CI, 0.29 to 0.79]; P = .005; right side, 0.52 [95% CI, 0.28 to 0.76]; P = .02), and supplementary motor cortex (mean difference, 0.81 [95% CI, 0.47 to 1.15]; P = .04). The ARFID-fear group showed greater amygdala activation vs HC (mean difference, 0.49 [95% CI, 0.16 to 0.82]; P = .04), and greater lack of interest was associated with lower hypothalamus activation in the ARFID-lack of interest group (r = -0.38 [95% CI, -0.69 to -0.11]; P = .03). The ARFID-sensory sensitivity group did not show greater insula activation vs HC but showed greater activation of the ACC (mean difference, 0.48 [95% CI, 0.22 to 0.74]; P = .005) and somatosensory cortex (mean difference on left side, 0.60 [95% CI, 0.33-0.87]; P = .001; right side, 0.54 [95% CI, 0.29 to 0.80]; P = .03). Conclusions and Relevance: Results indicate generalized hyperactivation of ACC, sensory association cortex, and supplementary motor cortex in response to visual food stimuli in children, adolescents, and young adults with ARFID, suggesting a novel neurobiological circuit associated with this disorder. Activation appears consistent with ARFID phenotypic rationales for food avoidance, with hyperactivation of fear regions in ARFID-fear and hypoactivation of appetite regions with increasing ARFID-lack of interest severity.