Band-specific atypical functional connectivity pattern in childhood autism spectrum disorder.

Abstract

Altered brain connectivity has been theorized as a key neural underpinning of autism spectrum disorder (ASD), but recent investigations have revealed conflicting patterns of connectivity, particularly hyper-connectivity and hypo-connectivity across age groups. The application of graph theory to neuroimaging data has become an effective approach for characterizing topographical patterns of large-scale functional networks. We used a graph approach to investigate alteration of functional networks in childhood ASD. Magnetoencephalographic signals were quantified using graph-theoretic metrics with a phase lag index (PLI) for specific bands in 24 children with autism spectrum disorder and 24 typically developing controls. No significant group difference of PLI was found. Regarding topological organization, enhanced and reduced small-worldness, representing the efficiency of information processing, were observed respectively in ASD children, particularly in the gamma band and delta band. Analyses revealed frequency-dependent atypical neural network topologies in ASD children. Our findings underscore the recently proposed atypical neural network theory of ASD during childhood. Graph theory with PLI applied to magnetoencephalographic signals might be a useful approach for characterizing the frequency-specific neurophysiological bases of ASD.