Charting the Development of Exploration in Adolescent Bulimia Nervosa: A Neurocomputational Approach

PROJECT SUMMARY/ABSTRACT Bulimia nervosa (BN) is a common and devastating illness that often emerges in adolescence, a time of dynamic neural maturation and cognitive development. Dieting and impulsivity are implicated in the onset and maintenance of BN. Yet, the few existing first-line treatments for adolescent BN primarily target dieting and produce suboptimal outcomes. Impulsivity is a complex, multi-faceted process, and the neural circuits subserving each facet change with development. The precise facets involved in adolescent BN remain unknown, hampering targeted interventions. Suboptimal exploration is a neurocognitive process that may underlie impulsivity and be salient to understanding adolescent BN. The overarching aim of this K23 Mentored Patient-Oriented Career Development Award is to test the hypothesis that adolescent BN is characterized by the altered development of exploration strategies and associated neural circuitry using an accelerated longitudinal design. Cohorts of early (aged 12-15), middle (aged 14-17), and late (aged 16-20) adolescent girls with BN and group-matched controls will be recruited. Each cohort will enroll 40 participants (20 BN/20 controls, N = 120). Adolescents will complete a bandit task that uses computational modeling to parse exploration strategies during fMRI scanning to decipher the neural correlates of exploration strategies. At 1- and 2-year follow-ups, adolescents will remotely complete the bandit task and clinical measures to examine developmental shifts in exploration strategies and their cross- sectional and prospective associations with BN symptoms. This design will test the following hypotheses: 1) adolescents with BN will not show shifts in exploration strategies over time, whereas controls will rely more on directed exploration and less on value-free random exploration over time (Aim 1); and 2) adolescents with BN will show altered neural activation and effective connectivity patterns during exploration that will be similar across BN age cohorts, but neural patterns will differ across control age cohorts (Aim 2). Links between exploration strategies and associated neural activation and effective connectivity and BN symptoms at baseline and over time will be examined (Exploratory Aim). The research will serve as a platform for mentorship and hands-on training in concepts and skills to support the candidate’s transition to independence. Specifically, the study will provide training in neurodevelopmental methods (e.g., adolescent recruitment and retention, longitudinal analysis, Training Aim 1) and computational modeling of brain and behavior data (Training Aim 2). The candidate will receive conceptual training in translational, transdiagnostic reward processing models to help her develop more precise models of adolescent BN (Training Aim 3). The research will support the candidate’s first R01 by yielding preliminary data on the neuro-computational and -developmental bases of BN and generating insights into developmentally sensitive clinical targets. Ultimately, the candidate’s research program will inform mechanistic, age-appropriate interventions that promote lasting recovery from this pernicious illness.