Patterns of Individual Change in Response to Reciprocal Imitation Training

Background: Reciprocal Imitation Training (RIT) is a naturalistic developmental behavioural intervention targeting social imitation skills. Previous research has shown RIT to be effective in increasing spontaneous object and gesture imitation. In a recent replication trial of a previous pilot-RCT of RIT, our RIT treatment group demonstrated a significant increase in spontaneous play-based object imitation skills compared with a wait-list control group (Malik, 2016). Alternatively, we did not replicate previously observed gains in elicited imitation skills.

Objectives: The aim of the present study was to evaluate whether individual children in the treatment group demonstrated reliable change in spontaneous and elicited imitation skills, and to further examine response to treatment in relation to child characteristics.

Methods: Twenty-four children with autism participated in the pilot randomized controlled trial. Stratified randomization led to two well-matched groups of 12 children: Treatment and Wait-List Control. Primary change measures included an Unstructured Imitation Assessment (UIA) and a Structured Imitation Assessment (SIA). These measures were administered pre- and post-intervention. Mullen Scales of Early Learning (MSEL) and Autism Diagnostic Observation Schedule (ADOS) were administered to characterize the children. Twenty sessions of RIT were delivered to the Treatment Group.

Results:  Reliable Change Index (RCI) was calculated for both the Treatment and Wait-List groups to explore individual patterns of change that might be associated with RIT. On the UIA, 4 children showed Reliable Change in the Treatment group. Closer inspection of individual data revealed that these were also the only children in the entire dataset who did not exhibit any unusual sensory interests or complex mannerisms on the ADOS. One child exhibited reliable change in spontaneous imitation skills in the Wait-List Control group. No child exhibited skill deterioration in the Treatment group, whereas one child in the Wait-List Control exhibited skill deterioration. One child improved reliably on the SIA in the Treatment group while no child in the Wait-List group demonstrated reliable change on the SIA.

In order to examine child characteristics associated with response to treatment, correlation analyses were conducted between spontaneous imitation change scores (UIA) and chronological age, non-verbal mental age, and verbal mental age (MSEL), and all ADOS domains. In the Treatment group, total spontaneous imitation change scores were negatively correlated with ADOS Reciprocal Social Interaction (r_s = - 0.67, p <0.01), and ADOS Stereotyped Behaviours and Restricted Interests (r_s = - 0.61, p = 0.01). In the Wait-List group none of the child characteristics were associated with change in imitation scores.

Conclusions: A greater percentage of children showed reliable improvements in the RIT treatment group (33%) as compared with the Wait-List Control group (8%). Findings from both RCI and correlation analyses suggest that children who had fewer self-stimulatory behaviours were more likely to respond to RIT. These findings support the efficacy of RIT for children with ASD, but also suggest that children presenting with greater social difficulties and/or greater stereotyped, repetitive behaviours may be less likely to benefit from RIT.