Intra-individual cortical networks in Anorexia Nervosa: Evidence from a longitudinal dataset

Objective: This work investigates cortical thickness (CT) and gyrification patterns in Anorexia Nervosa (AN) before and after short-term weight restoration using graph theory tools.Methods: 38 female adolescents with AN underwent structural magnetic resonance imaging scans at baseline and after - on average - 3.5 months following short-term weight restoration while 53 age-matched healthy controls (HCs) were scanned once. Graph measures were compared between groups and longitudinally within the AN group. Associations with clinical measures such as age of onset, duration of illness, BMI standard deviation score (BMI-SDS), and longitudinal weight changes were tested via stepwise regression.Results: Cortical thickness graphs of patients with acute AN displayed lower modularity and small-world index (SWI) than HCs. Modularity recovered after weight gain. Reduced global efficiency and SWI were observed in patients at baseline compared to HCs based on gyrification networks. Significant associations between local clustering of CT at admission and BMI-SDS, and clustering/global efficiency of gyrification and duration of illness emerged.Conclusions: Our results indicate a shift towards less organised CT networks in patients with acute AN. After weight recovery, the disarrangement seems to be partially reduced. However, longer-term follow-ups are needed to determine whether cortical organizational patterns fully return to normal.

Automated summary

This study investigated the cortical thickness and gyrification patterns in patients with Anorexia Nervosa (AN) before and after short-term weight restoration using graph theory analysis. The results showed that patients with acute AN exhibited less organized cortical thickness networks compared to healthy controls, but these networks partially recovered after weight gain. The baseline gyrification networks in patients also showed reduced global efficiency and small-worldness compared to healthy controls. Associations were found between local clustering of cortical thickness and BMI-SDS, as well as clustering/global efficiency of gyrification networks and duration of illness.