Resting State Functional Network Organization and Topological Properties in Autism Spectrum Disorder

Background:  Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by impairments in social communication and restricted, repetitive behaviors. Resting-state functional connectivity MRI (fcMRI) measures subtle correlations in low-frequency blood oxygenation level-dependent (BOLD) signal changes between brain regions while individuals are “at rest” or not otherwise engaged in a directed cognitive task.  Whereas findings from past studies suggest that functional connectivity is altered in individuals with ASD as compared to typically developing individuals without ASD, the nature of this aberration has not been clearly delineated.

Objectives: In the present study, we used advanced graph theory methods to examine resting-state fcMRI data from a large sample of individuals with and without ASD in hopes of gaining additional insight into how the topological properties of brain functional networks differ based on factors such as diagnosis (ASD vs non-ASD) and age.

Methods:  Resting state fcMRI data was collected from 28 adolescents and adults with ASD (mean age = 14.9 years) and an age- and gender-matched comparison group of 35 typically developing individuals without ASD (mean age = 15.6 years).  Following pre-processing and anatomical parcellation into 90 cortical and subcortial regions of interest (ROIs), functional partial correlation matrices were generated. The matrices for the participants in each group were averaged, thresholded to obtain undirected binary matrices, and then visualized as mean networks. Overall differences in network membership and structure were evaluated. Lastly, topological parameters were calculated to evaluate group differences in both local and global network properties.

Results:  The ASD and control group mean networks differed in node membership and organization.  Overall, 73 connections comprise the ASD group network whereas the control group network contains 67 connections.  The ASD and non-ASD networks were similar in that both were characterized by numerous long-distance connections (both anterior-to-posterior and left-to-right).  However, the ASD network was characterized by a higher number of local connections.

Following network visualization, topological parameters were calculated for each participant and compared between groups.  The two groups did not differ significantly in global network efficiency, t(63) = 1.40, p = 0.17.  Local efficiency (i.e., the efficiency with which information is transferred at a local level), however, was greater for the ASD group as compared to the non-ASD group, t(63) = 1.93, p = 0.05.

Conclusions: The results of the present study indicate potential alterations in function network organization and topological properties in ASD.  Further analysis is needed to fully characterize these differences within the contexts of both age and symptom severity.