Neuroanatomic Correlates of Dyspraxia in Children with Autism Spectrum Disorders

Background: Impairments in performance of skilled motor gestures suggestive of developmental dyspraxia, are among the most consistently observed abnormalities on motor examination of children with autism.  In the adult literature, apraxias are well-described, and are generally associated with impairments in left-hemisphere premotor and parietal association areas or the arcuate fasiculus connecting these regions.  However, the neural underpinnings of observed praxis impairments in the developmental context of autism are not well understood. 

Objectives: To examine volumetric neuroanatomic correlates of limb praxis in children with autism and typically developing (TD) children. 

Methods: Thirty-five TD children (6 girls; M_age=10.6 +/-1.2) and 21 children with high-functioning autism (3 girls; M_age=10.7 +/-1.8) completed a standardized praxis examination modified for children (i.e., to include gestures learned in early and later childhood: waving good-bye, using scissors). High-resolution MPRAGE images were acquired for each subject, and analyzed using automated surface-based methods within the program, Freesurfer.  Cortical volume measurements of regions of interest (ROI) in premotor and parietal regions were chosen based on their described relation to praxis.

Results: Groups were similar on demographic factors, with the exception of VIQ (children with autism M_VIQ=110.3 +/-20.4; TD M_VIQ=121.8 +/-12.0; t_{(1, 52)}=2.6,p=.012).  Therefore, VIQ was used as a covariate in regression analyses examining interaction effects. 

Children with autism (M_{Total Errors}=0.83 +/-0.09) performed more poorly than TD children (M_{Total Errors}=0.69 +/-0.14) on praxis exam (t_{(1, 52)}=4.4,p<.0001).  Following Bonferroni correction (p<.01), significant negative correlations were observed in the autism group between praxis performance and left-hemisphere white matter volume in the supramarginal area (r=--0.674, p=.002); a positive correlation was observed with the left-hemisphere gray matter volume in the caudal anterior cingulate (r=0.625, p=.004).  There were no significant correlations in the TD group.  Additionally, there was a significant interaction between diagnosis and praxis performance with the left hemisphere white supramarginal (p=.001) and left hemisphere gray caudal anterior cingulate (p<.0001) areas, with both interactions being driven by the highly significant correlation observed in the group of children with autism.

Conclusions: The significant correlation observed between praxis performance and the supramarginal and anterior cingluate regions in the autism group is consistent with the classical schema of apraxia.  A positive association with gray matter volume was observed (i.e., more gray matter associated with improved performance), suggesting that dyspraxia in autism is associated with decreased gray matter volume in the caudal anterior cingulate, important for guiding motor responses.  The opposite association was observed with white matter volume in the supramarginal region, important for acquisition and storage of the perceptual representations of movement (i.e., increased supramarginal white matter volume correlated with poorer performance).  Prior studies have revealed that children with autism show overgrowth of localized white matter connections and that for children with autism, increased primary motor white matter volume predicts basic motor skill impairment.  Similarly, the present findings suggest that dyspraxia in autism may be associated with overgrowth of localized white matter connections in the supramarginal region, possibly reflecting fewer long-range white matter fibers connecting this region with premotor association areas.