Project Summary/Abstract
Anorexia nervosa (AN) is a psychiatric illness characterized by low weight, extreme food restriction,
intense fear of weight gain, and disturbance in body self-evaluation. AN is markedly persistent, and there are
no empirically supported treatments for adults with this deadly illness. Indeed, over half of individuals with
AN do not reach full remission during treatment, and approximately one third of individuals experience
at least one relapse after discharge. There is great need for identifying remission-promoting targets to adapt
and create optimized interventions for this population. The period of partial remission in AN (i.e., weight-
restored individuals who continue to endorse significant cognitive-affective AN pathology) represents a critical
time for identifying mechanisms underlying remission. Specifically, unaddressed cognitive-affective AN
symptoms (e.g., extreme fear of weight gain) at this stage represent the greatest barrier to achieving lasting full
remission, and thus, mechanisms underlying the persistence vs. reduction of these symptoms are critical to
study at this stage. AN is highly comorbid with anxiety disorders and shares several maintenance factors,
including high levels of fear, anxiety, and harm avoidance. Despite considerable overlap with anxiety
disorders, fear and avoidance have not been tested as mechanisms inhibiting remission in weight-
restored AN. The proposed study will test how fear-based learning (e.g., fear extinction, fear renewal, fear
approach, fear avoidance) relates to AN pathology in the naturalistic environment in a sample of 45 individuals
meeting criteria for AN in partial remission. This study will use both behavioral assessment (i.e., Approach-
Avoidance Task, Behavioral Approach Task, associative fear-learning paradigm) and momentary assessment
(i.e., smartphone-based intensive longitudinal data collection) of fear, avoidance, and approach behaviors to
conceptualize how these mechanisms of interest operate in the ‘real world’. First, we will test whether
behavioral approach, avoidance, fear extinction, and fear renewal are correlated with real-time approach,
avoidance, fear, and AN symptoms (Aim 1). Second, vector autoregressive models using the intensive time-
series data will elucidate prospective associations across fear-based processes and AN symptoms between
subjects (Aim 2) and within individuals (Aim 3) to identify group-level and intraindividual targets for care during
this critical stage in recovery. In the between-subjects models, we will test behavioral fear extinction and
renewal as moderators (Aim 2). The opportunities, resources, and expert consultation afforded by this award
would support my training goals to collect and analyze behavioral data, learn new analytic approaches with
intensive longitudinal data, and gain experience working with clinical AN participants. Overall, this proposed
study will implement advanced statistics and multimethod research design, conducted fully remotely,
at an important stage of AN. These results will inform precision relapse prevention interventions and
step-down care, improving outcomes for this impairing, persistent, and costly disorder.