Exploring the disorders of brain effective connectivity network in ASD: A case study using EEG, transfer entropy, and graph theory

Many people worldwide suffer from Autism Spectrum Disorder (ASD) which is a neurodevelopmental disorder. It severely degrades the subjects' communication skills. The earlier diagnosing of ASD, The higher probability to prevent the severity of ASD symptoms. In the recent decade, brain connectivity studies on ASD subjects have converged to the theory of under-connectivity as a biomarker of ASD. Most of these studies have used fMRI data rather than EEG/MEG data and investigated functional connectivity rather than effective connectivity. There are few EEG/MEG studies which investigated the effective connectivity disorders in ASD subjects. Also, to the best of our knowledge there is no published study to investigate the disorders of brain effective connectivity networks in ASD subjects using EEG data, nonlinear effective connectivity measures and graph theory. In this paper, we aim to start filling this gap. We used EEG data, transfer entropy with self-prediction optimality, and four graph theoretic parameters to compare the effective connectivity networks of ASD youths with those of healthy controls (HCs) during a passive face processing task. Our results showed a significant difference in average degree (p<;0.05) between ASD and HC groups which is consistent with the under-connectivity theory of ASD. On the other hand we detected no significant changes in total clustering coefficient, average path length, and longest path length.