Abstract
Autonomic dysfunction is implicated in the manifestation of binge eating (BE) behaviors. Aberrant functional connectivity among brain regions involved in reward processing, affect regulation, and cognitive control have been the focus of resting state functional connectivity studies in BE. However, the influence of brain regions involved in central autonomic control and the relationship between autonomic balance and BE behavior remains unclear. The current study examines whether effective connectivity (EC) patterns in neural networks supporting the afore-mentioned cognitive processes moderates the relationship between heart rate variability (HRV), a physiological marker of autonomic balance, and BE behaviors. Resting-state fMRI data from 158 healthy adults (BE = 30; Control = 128) were analyzed using spectral dynamic causal modeling within a priori-defined eating regulation network. Leave-one-out cross-validation (LOOCV) assessed the predictive utility of EC patterns. Although 20 EC connections were associated with BE (posterior probability > 0.99), EC from the left posterior cingulate cortex to the hypothalamus predicted group membership above chance (AUC = 0.622, 95% CI [0.508, 0.731]). Approximate entropy was a significant predictor of binge eating behavior (ß = -6.625; p < 0.05) in a subset of 114 participants with usable photoplethysmography data (BE = 24; Control = 90). This relationship was significantly moderated by ventral tegmental area to hypothalamus EC (ß = -6.529; p < 0.05). These findings uncover novel directional connectivity patterns and identify brain-autonomic interactions that may underlie BE behavior, offering promising targets for future intervention strategies.