Functional connectivity signatures in fMRI-derived connectome for the diagnosis of autism spectrum disorder

Abstract

Rearrangement of synaptic connectivity is believed to be involved in shaping neuronal connections during brain developmental in early childhood, which determine mental health and behaviour of an individual. We here compare two types of the functional connectivity patterns: i) changes of healthy infants (3.7–32.6 months) to adults (25–35 yrs), reflecting normal developmental pattern, and ii) changes of healthy infants to autism spectrum disorder (ASD) patients (9.3–35.2 yrs), which encapsulate both developmental and ASD-specific pathological pattern. Using graph-based network parameters in the fMRI-derived connectome, we quantified changes and calculated average and median percent differences of various brain lobes. The connectome-to-connectome comparison suggests that synaptic rewiring is primarily concentrated in intra-thalamic and inter-lobe connections with the thalamus, and the ASD patients were characterised by significant thalamo-cortical hyperconnectivity. The average percent difference of the functional connectivity between ASD patients and adults for thalamus were as high as 82.58 % (degree), 72.8 % (betweenness centrality) 17.41 % (clustering coefficient) and 10.77/15.57 % (global/local efficiency). Regression models were built for normal brain development using functional connection data of healthy infants, child, adolescents and adults for each brain lobe. These regression curves suggest linearly increasing trends of functional connections from infant to adult in all brain lobes except in thalamus. Functional connections of ASD data are significantly different from this trends and characterized by significant overconnectivity. The distinct functional network signatures have the potential to serve as diagnostic markers of ASD. Towards this end, a method has been developed. The method can distinguish ASD subjects from their typically developing peers and healthy individuals with reasonable accuracy and specificity.