Abnormal Functional Connectivity Underling Social-Communicative Impairments in Autism Spectrum Disorder

Background: Researchers have gained interest on aberrant functional connectivity (FC) in Autism Spectrum Disorder (ASD) over the past decade. However, it is still unclear whether and how abnormal FC between different brain networks is associated with their behavioral impairments. Two key brain networks, default mode network (DMN) and fronto-parietal control network (FPC), are proposed to be associated with social-communicative functions in typical populations. Abnormal FC within these two networks was found recently.

Objectives: Here, we investigated whether the FC between these two networks relates to the social-communicative impairments in ASD by using both group comparison and correlation of individual differences.

Methods: We selected 112 subjects (56 with ASD and 56 typically developing (TD), mean age 19.16 years) from The Autism Brain Imaging Data Exchange (ABIDE) with strict criteria (e.g., matched age, gender and IQ) . We analyzed the resting-fMRI dataset and calculated three brain measures, including a) the amplitude of low frequency fluctuation (ALFF), which reflected the local neuronal activity; b) regional homogeneity (ReHo), which reflected the local FC, c) and FC, which reflect interregional FC strength, in both groups. We compared both groups on these brain measures, and estimated to what extent the degree of FC correlates with the individual differences of social-communicative scores, assessed by ADOS.

Results: The main findings are as follows: (1) we observed increased ReHo in the right middle frontal gyrus (RMFG) in the ASD group. Within the ASD group, individual differences of ReHo in the RMFG (r = 0.43, p < .005), right superior frontal gyrus (RSFG) (r = 0.38, p < .005) and right middle temporal gyrus (RMTG) (r = 0.36, p < .005) were positively correlated with the individual differences of social-communicative deficit. In contrast, the ReHo in the PC (r = -.45, p < 0.001) showed a negative correlation with the social deficit in the ASD group. (2) We further found that the FC between PC, RSFG (two key regions in the DMN) and the lateral RMFG (a key brain region in the control network) were associated with social impairment. Specifically, the FC between the RMFG and the PC was negatively related to the social symptoms in ASD (r = -.35, p < .01). (3) We did not find any ALFF group differences. However, individual differences of ALFF in the RMFG (r = 0.44, p < 0.001), RSFG (r = 0.39, p < .005), RMTG (r = 0.4, p < 0.001), and posterior cingulate (PC) (r = -.42, p = .001) showed significant correlations with the individual differences of social deficit for the ASD group.

Conclusions: Our findings support our hypothesis that abnormal inter-network FC was showed in the ASD group as compared to the TD group. Specifically, the abnormal social-communication functions in autism were highly related to the functional connectivity between the DMN and CN, two key brain networks supporting social and communicative behaviors in typical populations. In summary, our study provided further evidence for the neurophysiological basis of the social barriers in ASD.