Goal-directed vs. habitual instrumental behavior during reward processing in anorexia nervosa: an fMRI study.
Julius StedingIlka BoehmJoseph A KingDaniel GeislerFranziska RitschelMaria SeidelArne DooseCharlotte JaiteVeit RoessnerMichael N SmolkaStefan EhrlichPublished in: Scientific reports (2019)
Previous studies have proposed that altered reward processing and elevated cognitive control underlie the etiology of anorexia nervosa (AN). A newly debated notion suggests altered habit learning and an overreliance on habits may contribute to the persistence of AN. In weight-recovered AN patients, we previously found neuroimaging-based evidence for unaltered reward processing, but elevated cognitive control. In order to differentiate between state versus trait factors, we here contrast the aforementioned hypotheses in a sample of acutely underweight AN (acAN) patients. 37 acAN patients and 37 closely matched healthy controls (HC) underwent a functional MRI while performing an established instrumental motivation task. We found no group differences with respect to neural responses during the anticipation or receipt of reward. However, the behavioral response data showed a bimodal distribution, indicative for a goal-directed (gAN) and a habit-driven (hAN) patient subgroup. Additional analyses revealed decreased mOFC activation during reward anticipation in hAN, which would be in line with a habit-driven response. These findings provide a new perspective on the debate regarding the notion of increased goal-directed versus habitual behavior in AN. If replicable, the observed dissociation between gAN and hAN might help to tailor therapeutic approaches to individual patient characteristics.
Keyphrases
- end stage renal disease
- ejection fraction
- anorexia nervosa
- newly diagnosed
- chronic kidney disease
- magnetic resonance imaging
- prognostic factors
- randomized controlled trial
- computed tomography
- magnetic resonance
- body mass index
- case report
- dna methylation
- contrast enhanced
- prefrontal cortex
- study protocol
- deep learning
- phase iii