Dynamic Neural Computations Underlying Cognitive Control in Bulimia Nervosa

PROJECT SUMMARY/ABSTRACT
Bulimia nervosa (BN) is a disabling and often chronic eating disorder associated with medical complications and
premature death. There is an acute need to identify the mechanisms that maintain BN and that may serve as
targets for new treatments. Most studies have assumed that BN is perpetuated by stable, trait-like factors,
including self-reported impulsivity and deficits in the ability to exert cognitive control. However, the out-of-control
excessive intake and compensatory behaviors characteristic of BN are episodic and tend to alternate with
periods of restricted intake or even fasting, suggesting intermittent oscillations in control. In addition, emerging
work in behavioral economics and cognitive neuroscience suggests that failures to ultimately exert cognitive
control depend on numerous neural computations, including: updating predictions about whether control will be
needed in the next moment, and deciding whether trying to exert control is worth the costly cognitive effort. The
overarching goal of this R01 is to clarify how these cognitive control computations may fluctuate across fasted
and fed states in BN, potentially maintaining the cyclical nature of the disorder. Specifically, the proposed study
combines functional magnetic resonance imaging (fMRI), computational modeling, and real-time mobile
assessments to examine whether food consumption abnormally impairs frontostriatal function and associated
control-related updating and effort-valuation processes in BN. We propose that the impairment of these
processes in the fed state and relative improvement of these processes in the fasted state perpetuate out-of-
control binge/purge episodes alternating with periods of dietary restriction. We will compare behavioral and
neural responses of women with BN to those of group-matched healthy women during an inhibitory control
paradigm to assess control-related prediction updating (Aim 1) and during a cognitive effort discounting paradigm
to assess control-related effort valuation (Aim 2) in two states: after a 16-h fast and after a standardized meal.
Aim 3 will use multi-modal symptom measures to examine main effects and potential interactions of state-specific
updating and valuation processes on binge eating, purging, and restriction severity at baseline and at 6-month
follow-up. The research team includes experts in BN-focused research, neuroimaging, computational
neuroscience and psychiatry, advanced statistical methods, and ecological momentary assessment. The project
innovatively 1) applies a neurocomputational framework to examine the roles of understudied subcomponents
of cognitive control in BN; 2) assesses the influence of metabolic state (fasted, fed) on these subcomponents;
and 3) relates the dynamics of these subcomponents in the laboratory to real-world, state-specific experiences
and symptoms at baseline and over time. Data from this project will pinpoint altered elements of the control
process that may represent prognostic biomarkers. Results will also identify the potential optimal patient state
(i.e., fasted or fed) for control-focused interventions. Therefore, this study will yield vital data to inform urgently
needed, precisely targeted treatments for cycles of binge eating, purging, and restriction.

Grant Number: 5R01MH132786-03
NIH Institute/Center: NIH
Principal Investigator: Laura Berner