This crossover trial examined gastrointestinal (GI) interoception in 62 weight-restored females with restrictive anorexia nervosa (AN) vs. 57 healthy comparators (HCs), using a vibrating ingestible capsule, EEG, computational modeling, and peripheral physiology, with 6-month follow-up for relapse and symptom severity.
Participants with AN showed significantly lower GI perceptual accuracy (d = −0.98) and higher miss rates (d = 1.02) vs. HCs. At 6-month follow-up, relapse was predicted by maladaptive prior beliefs (OR 3.82), response bias (OR 5.37), and stomach unpleasantness (OR 5.73); symptom severity was predicted by miss rate and interoceptive precision shifts.
All participants were female, limiting generalizability to males and other AN subtypes. Sample size for follow-up was reduced to 54 (of 62), which may limit statistical power for relapse prediction. The single-site design and relatively short 6-month follow-up window may not capture longer-term relapse trajectories.
Ingestible vibrating capsules paired with computational modeling can identify GI interoceptive markers that predict relapse risk in weight-restored AN patients — consider these tools as adjuncts to guide relapse prevention. Stronger negative prior beliefs about gut signals and high stomach unpleasantness ratings at discharge may flag patients who need closer monitoring.
Explore related topics
