T1w/T2w ratio maps identify children with autism spectrum disorder and the relationships between myelin-related changes and symptoms

Shujun Zhang,Liping Jiang,Zhe Hu,Wenjing Liu,Hao Yu,Yao Chu,Jiehuan Wang,Yueqin Chen

DOI: https://doi.org/10.1016/j.pnpbp.2024.111040

IF: 5.201

2024-05-28

Progress in Neuro-Psychopharmacology and Biological Psychiatry

Abstract:Background Modern neuroimaging methods have revealed that autistic symptoms are associated with abnormalities in brain morphology, connectivity, and activity patterns. However, the changes in brain microstructure underlying the neurobiological and behavioral deficits of autism remain largely unknown. Methods we characterized the associated abnormalities in intracortical myelination pattern by constructing cortical T1-weighted/T2-weighted ratio maps. Voxel-wise comparisons of cortical myelination were conducted between 150 children with autism spectrum disorder (ASD) and 139 typically developing (TD) children. Group differences in cortical T1-weighted/T2-weighted ratio and gray matter volume were then examined for associations with autistic symptoms. A convolutional neural network (CNN) model was also constructed to examine the utility of these regional abnormalities in cortical myelination for ASD diagnosis. Results Compared to TD children, the ASD group exhibited widespread reductions in cortical myelination within regions related to default mode, salience, and executive control networks such as the inferior frontal gyrus, bilateral insula, left fusiform gyrus, bilateral hippocampus, right calcarine sulcus, bilateral precentral, and left posterior cingulate gyrus. Moreover, greater myelination deficits in most of these regions were associated with more severe autistic symptoms. In addition, children with ASD exhibited reduced myelination in regions with greater gray matter volume, including left insula, left cerebellum_4_5, left posterior cingulate gyrus, and right calcarine sulcus. Notably, the CNN model based on brain regions with abnormal myelination demonstrated high diagnostic efficacy for ASD. Conclusions Our findings suggest that microstructural abnormalities in myelination contribute to autistic symptoms and so are potentially promising therapeutic targets as well as biomarkers for ASD diagnosis.

pharmacology & pharmacy,neurosciences,psychiatry,clinical neurology