Dynamic brain network transitions of Children with autism in spontaneous brain activity

Autism Spectrum Disorder (ASD) is a neurodevelop-mental disorder characterized by learning difficulties like social interaction and communication. Numerous studies have indicated that dynamic transitions in brain networks are crucial feature of neural activity and are associated with the learning functions of our brain. However, dynamic brain network transitions of autistic children is still unclear due to limited research and immaturity of methodologies. To solve this challenge, this paper collected resting-state magnetoencephalography data from 11 autistic children to explore their main features of brain networks. Semi-supervised classification-driven methods and the HMM-MAR analysis framework were utilized to identify brain network states in source space and then identified networks states by comparing with standard brain networks. In addition, we analyzed the time characteristics of transitions between certain states during the conversion process. We found out that a) the Somatomotor network of autistic children was relatively weak and less pronounced compared to other rest networks; b) the network with the longest dwelling time of autistic children is the Default Mode network ; c) the Visual network is easy to switch and has the shortest dwelling times. These findings advocate for a tailored educational framework that considers the unique neural characteristics and cognitive needs of autistic children, thereby promoting more effective learning and integration strategies.