Introduction: Previous evidence suggests that improvisational music therapy (IMT) can be effective at improving social skills of children on the autism spectrum (CoS). To date, no published research exists regarding the effectiveness of specific IMT techniques. Given the improvised nature of IMT such research is important to understand the effective ingredients of the method. The purpose of this study is to examine the effectiveness of three IMT techniques with minimally verbal CoS: (1) exact imitation, (2) imitation with elaboration, and (3) contingent response.

Method: Seven children aged 40 to 56 months (M = 50.1) participated in the study.Inclusion criteria included minimally verbal children (producing fewer than 10 spontaneous words or phrases) between three and five years of age diagnosed with autism spectrum disorder. Each child participated in nine 30-minute sessions designed toinclude one of the three treatment techniques. Hence, three therapy sessions targeted each therapy technique. Sessions were video-recorded, analyzed and coded for treatment fidelity and child engagement.

Results: Analysis revealed that there was a significant overall effect of therapist technique on engagement state of a child, Hotelling’s Trace = 4.061, F(10, 14) = 2.843, p = 0.037. The overall pattern suggests that exact imitation and imitation with elaboration may result in better overall engagement than contingent response.

Discussion: Results suggest that using IMT imitation techniques may improve engagement. Since engagement is often a clinical goal, and is important in the process of social learning, music therapists are encouraged to consider employing imitation techniques during IMT sessions with CoS.