Neurobehavioral Association Between Macroscopic Network Decoupling and Reduced Eating Impulsivity: A Longitudinal fMRI Study in Bulimia Nervosa

BACKGROUND: Fluoxetine is an established pharmacological intervention for bulimia nervosa (BN); however, the macroscopic functional network remodeling underlying its efficacy in suppressing eating impulsivity remains poorly understood. METHODS: Fifty-eight patients with BN and 58 healthy controls assessed at both baseline and follow-up underwent resting-state functional magnetic resonance imaging (rs-fMRI). Network-Based Statistic (NBS) analysis with edgewise paired comparisons was used to identify whole-brain connectivity changes in patients with BN. The resulting BN-derived component was subsequently examined in both groups using a Group × Time analysis. Exploratory brain-behavior associations within BN were examined using Spearman and partial Spearman correlations. RESULTS: During follow-up, patients with BN showed significant improvements in depressive symptoms and externally driven eating behavior. Within-group NBS analysis identified a 23-edge component showing reduced functional connectivity in BN at follow-up, with network-level family-wise error correction (p < 0.05). This component primarily involved dorsal attention, ventral attention/salience, and somatomotor networks. Whole-brain NBS analysis in healthy controls did not identify a significant longitudinal component. When mean connectivity of the BN-derived component was extracted in both groups, connectivity decreased more strongly in BN than in healthy controls (Group × Time β = -0.053, p = 0.027). CONCLUSIONS: The follow-up period in patients with BN was accompanied by clinical improvement and reduced functional connectivity among attention-, salience-, and sensorimotor-related systems. These findings provide preliminary evidence of treatment-period-associated functional network reconfiguration in BN. TRIAL REGISTRATION: ChiCTR2200066885.