Abstract

Motivation: With the rising incidence of autism spectrum disorder (ASD) and the absence of clear diagnostic biomarkers, there's an urgent need to facilitate early diagnosis and intervention for affected children. Goal(s): We aimed to investigate microstructural disparities within the corpus callosum of ASD. Approach: We extracted diffusion parameters in the corpus callosum's genu, body, and splenium. Logistic Regression and Linear Discriminant Analysis models were constructed to assess the diagnostic potential of each parameter. Results: Significant distinctions in diffusion and white matter tract integrity metrics were observed, and machine learning models revealed the effectiveness of metrics in ASD diagnosis. Impact: DKI data can be used to evaluate the abnormalities in the microstructure of the corpus callosum in children with ASD and provide objective measurements to diagnose children with ASD.