Large-scale exploration of whole-brain structural connectivity in anorexia nervosa: alterations in the connectivity of frontal and subcortical networks

Abstract

BackgroundAnorexia nervosa (AN) is characterized by disturbances in cognition and behavior surrounding eating and weight. The severity of AN combined with the absence of localized brain abnormalities suggests distributed, systemic underpinnings that may be identified using diffusion-weighted MRI (dMRI) and tractography to reconstruct white matter pathways. MethodsdMRI data acquired from female patients with AN (n = 147) and female healthy controls (HC; n = 119), aged 12-40 years, were combined across five studies. Probabilistic tractography was completed, and full cortex connectomes describing streamline counts between 84 brain regions generated and harmonized. Graph theory methods were used to describe alterations in network organization in AN. The network-based statistic tested between-group differences in brain subnetwork connectivity. The metrics strength and efficiency indexed the connectivity of brain regions (network nodes), and were compared between groups using multiple linear regression. ResultsIndividuals with AN, relative to HC, had reduced connectivity in a network comprising subcortical regions and greater connectivity between frontal cortical regions (p < 0.05, FWE corrected). Node-based analyses indicated reduced connectivity of the left hippocampus in patients relative to HC (p < 0.05, permutation corrected). Severity of illness, assessed by BMI, was associated with subcortical connectivity (p < 0.05, uncorrected). ConclusionsAnalyses identified reduced structural connectivity of subcortical networks and regions, and stronger cortical network connectivity, amongst individuals with AN relative to HC. These findings are consistent with alterations in feeding, emotion and executive control circuits in AN, and may direct hypothesis-driven research into mechanisms of persistent restrictive eating behavior.