Motor Learning in Children with Autism Spectrum Disorder

Background: In addition to experiencing difficulties with social communications, individuals with an autism spectrum disorder (ASD) frequently experience problems in other domains of functioning such as motor control and learning. Past studies have generally reported that individuals with ASD show improved performance on motor tasks at a rate similar to that of healthy non-ASD individuals.  Additional behavioral and functional neuroimaging research (e.g., Gidley Larson et al., 2008; Müller et al., 2003), however, suggests that the neurocognitive processes underlying motor learning in individuals with ASD may differ from those associated with motor learning in typically developing individuals.

Objectives: The goal of the current study was to further advance our understanding of ASD-related differences in motor movement control and learning. To this end, we isolated and examined the ballistic and corrective submovements associated with performance of a rapid aimed limb movement in a sample of children with ASD and a demographically-matched group of typically developing children without ASD.

Methods: A sample of 25 males with ASD ranging in age from 8.15 to 18.09 years (M = 13.01, SD = 7.17) and a comparison group of 34 typically developing males without ASD ranging in age from 8.08 to 18.25 years (M = 12.94, SD = 8.74) participated. A 3D motion tracking system was used to record hand position while participants performed a rapid aimed limb movement (Abrams & Pratt, 1993).  Participants were instructed to move their right hand as quickly as possible from a starting position on the right to a target position on the left. Participants completed 10 blocks consisting of 10 trials per block (100 total movements).

Results: Data analysis revealed an overall main effect of trial block on task performance, with movement duration decreasing with increased practice, F(4,228) = 5.2, p = .001, partial eta² = .08.  In addition, the rate of this improvement did not differ significantly between the ASD group and non-ASD group, t(57) < 1, p = .38. The main effect of diagnosis was also not significant, with the ASD and control groups taking comparable time to complete the aimed limb movement, F(1, 57) = 2.2, p = .15, partial eta² < .04 in both instances.  Analysis of movement subcomponents, however, revealed a significant group difference. For the non-ASD group, repeated practice of the task was accompanied by an increase in the proportion of overall movement time devoted to ballistic as compared to corrective submovements, F(4,132) = 9.7, p = .001, partial eta² = .23.  In contrast, no such change in the proportions of submovements was observed for the ASD group, F(4,96) = 1.5, p = .21, partial eta² = .06.

Conclusions: Analysis of movement subcomponents revealed significant ASD-related differences in motor learning that were not otherwise evident by inspection of overall measures of task performance (e.g., reaction time, movement time).  Taken together, these findings support the hypothesis that, while motor learning per se is not impaired in ASD, individuals with ASD utilized different strategies in learning as compared to healthy non-ASD individuals.