A Second Look At Imitation: Imitative Errors in Children with Autism Spectrum Disorders

Background: Although imitation appears to be an innate and effortless ability for typical individuals, children with autism spectrum disorders (ASD) are often severely impaired in this capacity (Rogers & Williams, 2006; Williams, Whiten, & Singh, 2004). Despite documented impairment, research suggests children with ASD demonstrate effective imitative skills under certain conditions (Ingersoll, 2008; Ingersoll, Schreibman, Tran, 2008; Hobson & Lee, 1999; Want & Harris, 1998). These findings have led to a recent focus on imitative errors. Thus, rather than assessing whether a child passes or fails an imitative tasks, researchers are beginning to attend to the unique ways in which individuals with ASD imitate. However, error analyses are relatively new and only two known studies have aimed to specifically assess imitative errors (Rogers, Young, Cook, Giolzetti, & Ozonoff, 2010; Vanvuchelen, Roeyers, & Weerdt, 2007). 

Objectives: This study assessed the prevalence of imitative errors across three types of imitative tasks (i.e., object, facial-object, and facial tasks) in children with ASD and a typically developing control group. In addition, this study aimed to assess the impact of autism severity and verbal IQ on the frequency of imitative errors.

Methods: Thirty-one children with ASD and eighteen typically developing children completed an imitation battery comprised of three types of imitative tasks. Trained observers used a behavioral coding system to assess the prevalence of six error types (i.e., the need for multiple attempts, spatial errors, failure to attend, mirroring, non-compliance, and no-response). Multilevel modeling (MLM) was used to examine differences in error rates across groups and task types.

Results: The overall frequency of errors across tasks was significantly greater for the ASD group (p <.05). Furthermore, the error rates of typically developing children did not differ significantly from zero in any of the error by task combinations (e.g., multiple attempt errors in the facial task). Children with ASD made significantly more errors across each outcome as compared to the typically developing group, except for non-compliance errors in the object and object-facial tasks. As predicted by previous research, the facial task posed the most difficulty for the ASD group, with higher levels of errors in this task as compared to other tasks (p <.05).

Conclusions: The findings in this study expand on previous error analyses by assessing error rates in two new task types: object-facial tasks and facial tasks. Furthermore, the results of this study appear to support an ASD specific imitation deficit, as ASD status, rather than IQ predicted error frequency. Although it is difficult to compare the results of the present study with other error analyses conducted (due to differences in tasks and methodology), these findings appear to support the notion that individuals with ASD emit an atypical level of imitative errors.