Evidence That Anomalous Patterns of Imitation-Dependent Visuomotor Sequence Learning Is Specific to Autism

Background: There is increasing evidence that autism is associated with anomalous motor development (Gidley Larson and Mostofsky, 2006), including impaired imitation and execution of goal-directed skills (“dyspraxia”).  Development of these skilled gestures, including those involving social communication as well as tools use, likely depends on intact visuomotor learning (Haswell et al., 2009), in particular learning that involves imitating others’ actions. However, children with autism spectrum disorder (ASD) commonly demonstrate impaired action imitation  (Williams, Whiten, and Singh, 2004); as such, they may demonstrate a pattern of imitation-dependent visuomotor sequence learning that is anomalous and specific to ASD.

Objectives: To examine whether children with ASD show anomalous patterns of imitation-dependent visuomotor learning as compared with typically developing (TD) children and to determine the specificity of the findings through comparison with a separate Attention Deficit/Hyperactivity Disorder (ADHD) control group.

Methods: Seventeen children with ASD (3 female), 16 children with ADHD (8 female), and 34 typically developing (TD) children (11 female), completed an imitation-dependent version of a serial reaction time task (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 task included seven blocks of 80 trials.  Blocks 2-5 and 7 consisted of an implicit 10-item repeated sequence; blocks 1 and 6 were random. Visuomotor sequence learning was assessed using repeated measures ANOVAs (RM-ANOVA) to examine a change in the mean reaction times (RT) over the repeated sequenced blocks 2 through 5.

Results: Initial analyses revealed that the groups were matched on gender, race, handedness, age, socioeconomic status, and performance IQ.  RM-ANOVA of all three groups revealed a significant main effect of block (F = 11.170, p < .001), and a trend-level interaction effect of block by diagnosis (F = 1.987, p = .069).  Follow-up two-group RM-ANOVA examining TD and ADHD children revealed no significant interaction effect of block by diagnosis (F = 1.012, p = .389), whereas RM-ANOVA examining TD and ASD  children did (F = 2.790, p = .043).  RM-ANOVA within groups showed that TD and ADHD children displayed significant decrease in RT over blocks 2-5 (TD: F = 13.382, p < .001; ADHD: F = 6.727, p = .001).  However, the RT of children with ASD did not change significantly (F = .197, p = .898).

Conclusions: Children with ASD show a different pattern of learning than both children with ADHD and TD children.  In contrast, differences in learning were not seen between ADHD and TD children.  The findings suggest that ASD is associated with differences in mechanisms underlying imitation-dependent visuomotor sequence learning, consistent with previous findings (Gidley Larson and Mostofsky, 2009; Mostofsky et al, 2000).  Furthermore, the similarity of the patterns of learning between children with ADHD and TD children suggests that these differences in motor learning may be specific to autism.