Reinforcement Learning in Eating Disorders: Clinical, Cognitive, and Metabolic Correlates

Buyuktopcu E., Yucel B.

EUROPEAN EATING DISORDERS REVIEW, 2026 (SSCI, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1002/erv.70164
  • Dergi Adı: EUROPEAN EATING DISORDERS REVIEW
  • Derginin Tarandığı İndeksler: Social Sciences Citation Index (SSCI), Scopus, CINAHL, EMBASE, MEDLINE, Psycinfo, Academic Search Ultimate (EBSCO), Biomedical Reference Collection: Corporate Edition (EBSCO), Health Research Premium Collection (ProQuest), Psychology & Behavioral Sciences Collection (EBSCO)
  • İstanbul Üniversitesi Adresli: Evet

Özet

Objective Reinforcement learning has been implicated in eating disorders, but findings remain inconsistent. This study compared reinforcement learning performance in anorexia nervosa (AN), bulimia nervosa (BN), and healthy controls (HC) using the Probabilistic Reward Task (PRT), and explored clinical, cognitive, and metabolic correlates.Method The sample included 31 AN, 20 BN, and 25 HC participants. PRT performance, working memory, and pre-task blood glucose were assessed. Group differences were tested using ANOVA and repeated-measures GLM. Exploratory correlations and follow-up regression examined correlates of response bias indices.Result Groups did not differ in total response bias or change in response bias, and no block or block & times; group effects emerged for response bias. Discriminability changed across blocks, but similarly across groups. Accuracy and reaction times did not differ. In AN, blood glucose was positively associated with change in response bias, and this association remained significant after correction. Follow-up regression showed that blood glucose explained 46.7% of the variance in change in response bias.Discussion Findings do not support generalized diagnostic differences in reward-based response bias in AN or BN. However, the association between blood glucose and response bias change in AN suggests that acute metabolic context may influence reward-related task performance.