Gray Matter Network Property Differences in Children with Autism Spectrum Disorder

Background: Structural brain imaging studies have revealed differences between individuals with autism spectrum disorders (ASD) and Typical Development (TD), particularly in frontal brain regions. Functional connectivity studies have shown differences in correlated activity and functional network properties in ASD. However, there are relatively few investigations on detailed measures of grey matter comparing ASD and TD children, and no one has examined structural brain connectivity using graph theoretical network analysis applied to cortical thickness measures in ASD.

Objectives: To examine structural brain connectivity as measured by regional correlations in cortical thickness (CT) across the brain in ASD versus TD.  

Methods: We present preliminary data from 30 ASD and 36 TD control children as part of the ‘NeuroDevNet ASD project’, an ongoing multi-site study on brain and behavioral development in ASD.  Ethical approval was granted by the Montreal Neurological Institute Research Ethics Board. The groups were matched on age (from 6-16 years old), and all subjects had an IQ above 70 (except 2 ASD individuals).  Vertex-wise CT values were generated from T1 structural MR images using the CIVET pipeline.  Correlations of mean CT among 74 cortical regions were calculated separately for the ASD and TD groups, and then tested for differences while controlling the false discovery rate. Cortical connectivity analyses were conducted on the thresholded correlation matrix in order to calculate binary network graphs, as well as network global and local efficiency measures separately in ASD and TD.

Results: Examination of interregional correlations revealed greater connectivity in the ASD group in frontal, parietal and temporal lobes. Increased connectivity was mostly within hemispheres. Diminished connectivity in ASD was found in frontal, parietal and temporal lobes, and mostly between hemispheres. In comparing overall network connectivity using graph theory, we found significantly enhanced local efficiency in the ASD brain, combined with a trend toward diminished global efficiency.

Conclusions: We provide novel evidence for atypical structural brain connectivity in ASD children using graph theoretical network analyses of cortical thickness. These results are consistent with the idea that short-range connectivity is often enhanced, and long-range connectivity diminished, in ASD.