Associations among acute exercise response and eating disorder-related measures in a sample of young women with and without restrictive eating pathology
Year: 2026
Type: other
Abstract: Compulsive exercise (CE) is a common behavioral feature of eating disorders (EDs), defined as excessive, driven, or rigid physical activity intended to prevent feared consequences or escape distressing emotions. Despite negative sequelae associated with CE, including worse treatment outcomes1, increased relapse risk2, and suicidality3, relatively little is known about the biobehavioral mechanisms that maintain this symptom. Prevailing psychological theories posit that CE is maintained through both negative reinforcement (e.g., reduction of negative affect, mitigation of body weight/shape concerns) and positive reinforcement (e.g., increased positive affect, body satisfaction4,5). In a recent pilot investigation6 (N = 40), we developed and tested a novel in-laboratory exercise assessment paradigm comprised of two tasks—a Prescribed exercise task (30-minute moderate-intensity cycling at 70–75% HRmax) and a Self-Paced exercise task (≤30-minute self-selected intensity cycling)—to evaluate acute psychobiological response to exercise among young women with and without restrictive eating pathology. This pilot study served as preliminary data for a funded NIMH R21 (1R21MH131787-01A1; PIs: Schaumberg, Gorrell), which specified two primary aims: Aim 1 (internal validation) to characterize variability in psychobiological response to in-lab exercise across ED and HC, and Aim 2 (concurrent validity) to examine associations between acute exercise response and validated exercise and eating pathology self-report measures (compulsive exercise, exercise dependence, eating disorder severity) as well as accelerometer-measured physical activity. In the published pilot, these aims were operationalized as: feasibility and safety (Aim 1), group differences in exercise effects on affect and body image (Aim 2), and correlations among exercise response variables and exercise and eating pathology self-report measures (e.g., compulsive exercise, exercise dependence, eating disorder severity) within each group (Aim 3). Regarding Aim 3, pilot findings revealed a pattern of moderate-to-large correlations between acute exercise response (i.e., slopes of change in affect and body image during exercise) and clinical, behavioral, and physiological variables of interest in the ED group. Notably, several correlations reached substantive effect sizes (|r| > 0.30) within the ED group. Body image response during exercise was strongly associated with BMI (r = −0.65 for average BISS during Self-Paced; r = −0.48 during Prescribed). Self-paced exercise intensity showed robust associations with affective response, including r = 0.57 for Karvonen intensity × perceived exertion and r = −0.55 for maximum heart rate × crummy. Exercise dependence was consistently associated with more enthusiastic responses (r = 0.43 Self-Paced, r = 0.42 Prescribed). Average zero-order r² associations between exercise response variables among controls was 0.14 and in the ED group was 0.19, with little evidence of within-group collinearity. Further, 14% of correlations among exercise response variables in the ED group and 6% in the control group reached the nominal p < .05 threshold, exceeding the 5% false positive rate expected under a strict null hypothesis and suggesting greater coherence in response patterns among individuals with EDs. While these pilot findings are promising, the small sample size (n = 20 per group) limited statistical power to detect significant correlations and precluded drawing firm conclusions about the strength and specificity of associations. The current study addresses the central research question of whether acute exercise response—as measured by in-laboratory changes in affect and body image during exercise—is associated with validated self-report measures of compulsive exercise, exercise dependence, eating disorder severity, and real-world physical activity in a larger, independent sample of young women (N = 75; 45 ED, 30 HC) recruited and assessed under the R21 grant (1R21MH131787-01A1; PIs: Schaumberg, Gorrell). This is an entirely new sample with no participant overlap with the pilot. Participants in the current sample completed the same exercise paradigm (Prescribed and Self-Paced Tasks) and a largely overlapping battery of self-report measures. The current study uses the Functional Assessment of Driven Exercise (FADE7) to assess exercise functions, rather than the Functional Assessment of Maladaptive Behaviors (FAMB) used in the pilot; both measures assess the same four reinforcement function subscales (automatic positive reinforcement, automatic negative reinforcement, social positive reinforcement, and social negative reinforcement) but differ in item content and scaling. The current assessment battery also includes several additional measures (Difficulties in Emotion Regulation Scale [DERS], Generalized Anxiety Disorder 7-item [GAD-7], Patient Health Questionnaire 9-item [PHQ-9], Eating Pathology Symptoms Inventory [EPSI], Exercise Flexibility Questionnaire [EFQ], Behavioral Inhibition System/Behavioral Activation System scales [BIS/BAS], Sensitivity to Punishment and Sensitivity to Reward Questionnaire [SPSRQ]) that enable a broader evaluation of concurrent validity across domains of psychopathology, emotion regulation, and reinforcement sensitivity—constructs central to the theoretical model of compulsive exercise maintenance through reward and threat pathways.
Source: Open Science Framework
Language: English
Subfield: Clinical Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7152004222