Food for thought: how aversive valence shapes decisions about what to eat in anorexia nervosa

About This Grant

PROJECT SUMMARY Anorexia nervosa (AN) is a devastating eating disorder that has one of the highest mortality rates of any psychiatric illness. AN is characterized by persistent restriction of food intake below the body’s needs, which can be understood as a series of maladaptive food choices. Choices are guided by neural value signals that reflect an individual’s preference, encoded in the ventromedial prefrontal cortex (vmPFC). Individuals with AN assign low value to high-fat/calorie foods they consider to be unhealthy, and are consequently unwilling to eat these foods. The mechanisms underlying this disturbance in value assignment are unclear, though the significant eating-related distress experienced by individuals with AN suggests involvement of brain systems signaling aversive valence. This hypothesis has not been directly tested, which precludes targeting aversive valence systems in treatments for AN. This study will examine whether aversive valence is assigned to energy- dense foods in AN, and whether this is a mechanism of restrictive eating pathology. Patients with AN (n=50) and healthy controls (HC, n=50) will complete a food valuation task in which they rate their preference to eat a range of foods (including those typically highly disliked and liked by HC), while undergoing fMRI scanning and autonomous arousal (skin conductance response) assessment. After scanning participants will rate food items across 20 attributes, including aversive valence, healthiness, and nutritious and sensory features (e.g., fat content, texture). Using computational and multivariate fMRI analyses, we will examine 1) associations between ratings of aversive valence and food healthiness, 2) activation of aversive valence representations in the amygdala by foods considered unhealthy, 3) coupling of regions encoding aversive valence (amygdala) with those encoding healthiness (orbitofrontal cortex) and subjective preference (vmPFC). Comparing AN and HC groups will allow us to identify 1) whether assigning aversive valence to energy-dense foods is a mechanism of restrictive eating in AN; 2) regions encoding aversive valence of energy-dense foods in AN. Exploratory analyses will examine the mechanisms by which information about additional food attributes is integrated into value computations, to identify additional disturbances in food choice valuation processes in AN. Associations between the activation and coupling of brain regions encoding aversive valence, and clinical characteristics, will be explored among patients. In completing the planned research and training, the applicant will develop expertise in cognitive and computational neuroscience as applied to eating behavior. This expertise combined with study findings will 1) form the foundation of a unique research program examining disturbances in food choice valuation processes in eating disorders, and 2) support an R01 proposal to examine the malleability of aversive valence representations in AN, and mechanisms of/lack of change.