Localization and diagnosis of abnormal channels in children with ASD based on WMSSE and ASI

The purpose of this study is to explore the differences of resting EEG in children with autism spectrum disorders and then analyze the sensitive channels with significant differences, to provide support for the accurate differential diagnosis of autism spectrum disorder (ASD).

Based on the weighted multi-scale sample entropy (WMSSE) algorithm and amplitude synchronization index (ASI) algorithm of EEG, this paper comprehensively evaluates the brain state of ASD children from the two aspects of brain function complexity and brain function synchronization connectivity. Further, by combining the support vector machine (SVM) classification model to explore the location of abnormal channels of ASD children and realize the diagnosis of ASD children.

The WMSSE of the ASD group was lower than that of the healthy group. Furthermore, there was a significant difference in the F3/F4 channels and F7/F8 channels (P < 0.05), and the synchronization of the brain in the ASD group was also lower than that of the healthy group in Delta, Theta, Alpha, Beta band. Finally, combined with the WMSSE and ASI features of the F3/F4 channels (posterior frontal lobe) and F7/F8 channels (anterior temporal lobe), the classification accuracy and AUC value of ASD patients calculated by the SVM classification model were 82.7 ± 3.2%/ 0.795 (F3 / F4 channels), 89.8 ± 1.7%/ 0.812 (F7 / F8 channels).

It avoids the one-sided problem of single analysis complexity and synchronous connectivity, and provides a research basis for the comprehensive evaluation of ASD brain function.

WMSSE and ASI can be used as effective potential biomarkers for ASD diagnosis, and F7 and F8 channels can be preliminarily located as abnormal sensitive channels for ASD brain regulation. It also proves that the feature analysis of comprehensive complexity and synchronous connectivity is more conducive to the abnormal diagnosis of ASD patients.