Imitation-Dependent Visuomotor Sequence Learning in ASD

Background: Motor imitation deficits, hypothesized to be a key feature of autism spectrum disorder (ASD), affect learning of communicative gestures (e.g. waving) and may contribute to social impairments associated with the disorder.  There have been inconsistent findings regarding visuomotor sequence learning in ASD.  However, due to the nature of imitation deficits in ASD, visuomotor sequence learning may be particularly impaired in imitation-dependent contexts.

Objectives: To examine the effect of imitation on visuomotor sequence learning in autism.

Methods: Experiment 1 - 10 children with ASD and 8 typically developing (TD) children completed a standard Serial Reaction Time Task (“Standard SRTT”) in which subjects used the fingers of their right hand to push one of four buttons corresponding to red squares on a computer screen.  Experiment 2 - 11 children with ASD and 17 TD children completed an imitation-dependent version of the SRTT (“Imitation SRTT”), in which subjects used the fingers of their right hand to push one of four buttons in imitation of a video of a left hand facing the subject. The tasks in Experiment 1 and 2 included 7 blocks of 80 trials; blocks 2-5 and 7 comprised an implicit 10-trial repeated sequence; blocks 1 and 6 were random.  For both experiments, learning was assessed using a repeated measures ANOVA (RM-ANOVA) to examine for a decrease in reaction time across blocks 2-5. 

Results: Experiment 1 - RM-ANOVA revealed a significant effect of block across the two groups of subjects (F = 9.9, p = .0009). There was no significant effect of diagnosis across blocks (F = .63, p = .44).  There was no significant interaction effect of block with diagnosis (F = .28, p = .84); follow-up within-group analyses revealed a main effect of block for TD (F = 18.7, p = .0038) and ASD (F = 5.1, p = .035), indicating that both groups showed evidence of learning.  Experiment 2 - RM-ANOVA revealed no effect of block across the two groups of subjects (F = 1.7, p = .2).  There was no significant effect of diagnosis across blocks (F = .07, p = .8), nor was there an interaction effect of block with diagnosis (F = 1.5, p = .2).  Within-group analyses, however, revealed the HFA group showed a distinctly different pattern (much flatter learning curve) than the TD group, such that there was a main effect of block for TD children (F = 4.4, p = .02), but not for children with ASD (F = .04, p = .99); this suggests that while TD children showed learning similar to that seen with the Standard SRTT, children with ASD did not. 

Conclusions: The findings suggest that children with ASD exhibit impaired motor sequence learning under conditions involving action understanding, such as imitation.  This would suggest that, for children with autism, abnormalities in circuits comprising inferior parietal (important for perceptual representations of movement), premotor (important for preparing motor action) and superior temporal sulcal (important for self-other mapping) regions may contribute to difficulties with acquisition of skilled gestures, including those important for socialization/communication.