Abstract
OBJECTIVE: Avoidant/restrictive food intake disorder (ARFID) is characterized by limited dietary volume and/or variety due to fear of aversive consequences, lack of interest in eating, and/or sensory sensitivity. Uncontrolled studies suggest that cognitive-behavioral therapy for ARFID (CBT-AR) is beneficial, but neural mechanisms remain unknown. Prior work has shown hyperactivation in regions of the salience network (SN) during the presentation of food stimuli in ARFID compared to healthy controls, and the SN shows reduction from pre- to post-CBT in anxiety disorders. We tested hypotheses that (1) higher pre-treatment food-related activation in the SN (anterior cingulate cortex [ACC], anterior insula, amygdala) would be prospectively associated with greater reductions in ARFID severity at post-treatment, and (2) food-related activation in the SN would decrease from pre- to post-treatment. METHOD: In a validated functional magnetic resonance imaging (fMRI) food cue paradigm, youth with ARFID viewed images of food and non-food in a fasted state at pre-treatment (N = 24) and a subset (N = 14) also completed post-treatment scans. RESULTS: Pre-treatment activation in response to food versus objects was not associated with changes in ARFID severity or remission post-treatment. However, hyperactivation in the right ACC (MNI coordinates: 3, 53, 17; k(E) = 579, p(FWE(corr)) = 0.013) decreased from pre- to post-treatment with a large effect size (d = 1.1). CONCLUSION: These pilot findings raise the possibility that changes in SN neural activity may be candidate mechanisms through which CBT-AR exerts its effects. Randomized controlled trials are needed to replicate and extend these results. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT02963220.