A Brief, Low-Cost Intervention for Prosodic Phrasing in High-Functioning ASD: Translating Science into Treatment

Background:   Atypical prosody is frequently reported in autism spectrum disorder (ASD), and deficits have significant negative consequences for social and occupational functioning. They are also persistent: Even individuals who attain average expressive and receptive language skills (per standardized assessment) often have atypical prosody. Though novel methods have been developed (e.g., iPad applications), there is limited empirical evidence for these or other interventions for expressive prosody in ASD; this evidence points to the efficacy of explicit training on the nature and use of supra-linguistic cues, including prosody. 

Objectives:   We examined spontaneous (“baseline”) production of prosodic cues, and compared these to productions following a brief intervention involving explicit discussions of prosody and opportunities for structured imitation and practice.

Methods:    Participants included adolescents with ASD and typical development (TD; n=15 per group). Groups did not differ in age, IQ (>80), or gender, p’s > .50. Participants were told to instruct a “partner” (a research assistant) to perform actions using small toys (Snedeker & Yuan, 2008). Eight critical trials had syntactically ambiguous phrasing; to provide a clear instruction, subjects had to rely on prosodic cues (pause lengthening, clause duration). We asked whether subjects were aware of the syntactic ambiguity, and what strategies they employed for disambiguation. Subjects then participated in a five-minute intervention, consisting of visual slides with a voiceover, that gave direct instruction to clarify the presence of the ambiguity, provided explicit instruction about prosodic phrasing, and prompted subjects to imitate prosodic phrasing. Finally, subjects completed “post-intervention” trials containing the same syntactic ambiguity (with trial-unique stimuli). Trials were recorded and analyzed using Praat. In addition, naïve college student raters listened to recordings and chose a syntactic match for each.

Results:    Interestingly, there were no mean group differences in acoustic variables, at baseline; naïve raters were at chance in interpreting the ambiguity. Following the intervention, all participants increased pause length and clause duration, and raters were significantly better than chance at interpreting the ambiguity. While there were no mean group differences, there was a subset of participants, in the ASD only, who were unaware of the ambiguity. Following the intervention, naïve raters were above chance in interpreting productions of the “unaware” group. Also, a subset of subjects across groups (“low distinctiveness”) was initially ineffective at producing prosodic cues; this subset was indistinguishable from the “high distinctive” group following intervention (Figure 1).

Conclusions:   Results suggested intact groupwise performance for expressive syntactic prosody, consistent with reports that atypical prosody is not universal in ASD (Shriberg et al., 2001). Acoustic analyses indicated no group differences at baseline or after intervention. However, a subset of participants with ASD was initially unable to use prosodic cues effectively; they made significant gains in prosodic phrasing following intervention. Participants in both ASD and TD groups demonstrated more effective expressive prosodic phrasing following a brief, targeted intervention. Future research must examine generalizability of these findings to lower-functioning individuals and to other syntactic phrases, and efficacy of long-term retention. These promising results suggest that a low-cost, brief intervention may yield clinically-meaningful effects.