Visual-Motor Functional Connectivity Relates to Autism Severity

Background:  Autism Spectrum Disorder (ASD) affects many areas of skill development, including motor skills. Research suggests that motor, communication, and social skill learning may share a common brain-basis. ASD-associated deficits in imitating others’ actions, dependent on visual-motor integration, likely impact both motor and social skill acquisition; however, it is unclear what brain mechanisms contribute to these deficits.

Objectives: To investigate the relationship between visual-motor functional connectivity (FC) and both imitation ability and autistic trait severity in children with ASD.

Methods: Resting state (rs) fMRI scans were collected from 100 children (50 ASD; 50 typically developing [TD]). Data were interpolated to account for slice acquisition order, and rigid body realignment parameters were estimated with respect to the first functional volume to adjust for motion. Anatomical images were registered to the first functional volume and then to the MNI template. The estimated rigid body and non-linear spatial realignment transformations were then applied to the rs data. Linear trends were removed from the rs data and a Gaussian filter (6-mm FWHM kernel) was applied. Voxel timeseries were intensity-normalized to have a mean of 100. All 100 rs scans were combined to estimate visual and motor networks using independent component analysis. Participant-specific spatial maps and timeseries were back-reconstructed from the group-level components. To estimate visual-motor FC, Pearson correlations between each pair of participant-specific motor and visual network timeseries were computed and converted to z-scores using Fisher’s transform. Brain-behavior relationships were assessed by regressing visual-motor FC with imitation and autistic trait severity scores. Imitation ability was assessed using the Florida Apraxia Battery; scores reflected the number of imitative gestures performed correctly. Autistic trait severity was assessed using the Social Responsiveness Scale (SRS); higher scores indicated more severe autistic traits.

Results:  Two motor networks (dorsal [DM] and ventral [VM]) and three visual networks were identified. Two visual networks included early visual processing areas (BA 17 and 18), while the third (VC3) included higher-order visual areas involved in perspective processing. In children with ASD, VM was more negatively correlated with VC3 (-.12) compared to TDs (.08, p < .05 corrected), and the strength of this coupling was inversely correlated with total SRS score in the ASD group (R=-.42, p=.004). In TDs, stronger VM-VC3 FC was associated with better imitation (R=.41, p=.01) and better overall performance of gestures on praxis examination (R=.61, p<.001). No relationship was observed between visual-motor FC and imitation ability in ASDs.

Conclusions: Children with ASD exhibited significantly stronger negative functional coupling between motor and higher-order visual areas compared to their TD peers, and the strength of this negative coupling was associated with the severity of autistic traits. Children with stronger negative coupling between motor and higher-order visual areas demonstrated more severe autistic traits. In TD children, motor-visual FC strength was correlated with imitation performance; children with stronger positive visual-motor coupling were better imitators. The findings suggest that visual-motor connectivity is associated with the ability to imitate others actions, and that for children with ASD, decreased visual-motor connectivity may contribute to impaired social skill development