Ketone Supplementation Attenuates Prediction Error Taste Response in Eating Disorders: A Proof-of-Concept Study

Emerging evidence suggests that eating disorder (ED) pathophysiology involves metabolic alterations, and preliminary data indicate that ketogenic interventions may reduce symptom severity. We tested whether exogenous ketone supplementation modulates neural responses in individuals with anorexia nervosa (AN) and bulimia nervosa (BN) during a task previously shown to elicit altered prediction-error signaling in ED populations.

Twenty-two participants across the AN and BN spectrum (AN: n = 8; OSFED-AN: n = 9; BN: n = 4; OSFED-BN: n = 1) spanning a wide range of ages (19.0-52.1 years) and body mass index (BMI: 17.2-46.2) were studied in a randomized, double-blind, placebo-controlled crossover design. On separate days, participants consumed 10 g beta-hydroxybutyrate (BHB) or placebo prior to completing a taste-based conditioning task during electroencephalography (EEG). Trial-level neural sensitivity to prediction error during unexpected taste receipt and omission was indexed using the reward positivity (RewP/FRN) and late positive potential (LPP) event-related potentials (ERPs), cortical markers of outcome salience linked to dopaminergic and noradrenergic systems.

Active BHB increased blood ketone levels. Compared to placebo, BHB was associated with a significant reduction in LPP sensitivity to unsigned prediction error (p = 0.028), controlling for diagnosis, comorbidity, and medication use. Poorer self-esteem was associated with a smaller ketone-related reduction in LPP sensitivity (p = 0.002).

These findings suggest that EEG-based measures can capture prediction-error processing during cue-taste learning and that exogenous ketones may attenuate neural sensitivity to salient taste outcomes in EDs. Further research is needed to determine whether such modulation translates into clinically meaningful changes in illness behavior.