Causal Network Analysis Reveals Key Brain Regions Associated With Severity in Children With Autism Spectrum Disorder

ABSTRACT This study aims to investigate the relationship between gray matter (GM) changes and severity in children with Autism Spectrum Disorder (ASD). We examined 113 ASD children aged 2–8 years (17 mild cases, 56 moderate cases, and 40 severe cases), as well as 110 age and sex‐matched healthy controls (HC). Voxel‐based morphometry (VBM) was used to compare GM density (GMD) changes between ASD and HC groups. Additionally, structural covariance network analysis quantified the cross‐regional synchronous changes in GM among ASD children, and causal analysis described the pattern of changes in the GM network related to symptom severity in ASD children. The results indicated that ASD children exhibiting mild symptoms have an enlarged parahippocampal gyrus, and as the severity of ASD increases, the range of GMD changes expands ( p < 0.05, FDR correction). Granger causality (GC) analysis revealed that the parahippocampal gyrus may function as a central hub within ASD‐related directional networks, exerting causal effects on other brain regions ( p < 0.05). These findings were validated by external datasets. Our results provide preliminary insights into the role of the parahippocampal gyrus in ASD and promote the application of dimensional models.