A Single-Blind RCT of an Evidence-Based Social Skills Intervention for Youth with ASD: Effects on Behavioral and Neural Indices of Social Functioning

Background: While social skills interventions (SSIs) for youth with autism spectrum disorder (ASD) evince modest effects, it is not clear whether widely-reported parent-report effects of SSIs represent true change, or parent expectations for change (Gates et al., 2017). Furthermore, no study has used rigorous component-control designs to examine whether less subjectively-determined effects are specific to particular elements of those SSIs. The SSI socio-dramatic affective-relational intervention (SDARI; Lerner et al., 2011) uses specific game-like activities putatively designed to provide opportunities for intensive practice of social competence elements (e.g., interpreting tone of voice) in a supportive environment. While SDARI has shown effects on parent-report social skills, observed social behavior, peer sociometrics, and neural indices of social perception (e.g., Lerner & Mikami, 2012; Kang et al., 2017), it is unknown whether SDARI-specific activities (versus nonspecific supportive environment) or parent expectations are responsible for these effects.

Objectives: To examine whether SDARI exerts unique effects on social behavior, peer regard, and neural metrics of social perception. To examine whether effects on parent-report are attributable to expectancy effects.

Methods: 55 youth (M_age=12.40, SD_age=2.92; 40 male), IQ≥70 and ADOS-2 (Lord et al., 2012)-confirmed diagnoses, were randomly assigned to SDARI or an attention control (8 groups total; 1.5 hours/week, 10 week), which followed a schedule identical to SDARI but used structured recreational activities not inherently social (e.g. art projects). As both conditions were likely to produce some benefit, participants and parents were held unaware of condition. Parents completed a questionnaire battery (e.g., SSIS; Gresham et al., 2010) and indicated the group to which they believed their child had been assigned. Participants completed assessments including EEG-indexed evaluation of social processing (e.g., N170 ERP latency to face; N100 amplitude to voice; Lerner et al., 2013). Participants provided sociometric ratings of one another during the first and final sessions; their prosocial behaviors in unstructured free play for 20 minutes during the middle of each session were coded (Bauminger, 2002) by reliable (ICC(1,2) = .829) raters blinded to time and condition. Generalized Estimating Equations (accounting for nesting within groups) were used.

Results: Parent-report effects on social skills were attributable to the condition to which parents believed their child was assigned, not actual condition (Wald’s X²=6.256, p<.05; Figure 1A). The SDARI group showed a relative acceleration in N170 latency and increase in N100 amplitude (both X² >6.00, p≤ .05; Figure 1B-C). While those in SDARI became reciprocated friends more quickly, the attention control caught up (X²=6.522, p<.05; Figure 1D). The same pattern was seen for interactions with peers (SDARI: X²=35.87, p<.001; condition*time: X²=4.87, p<.05).

Conclusions: This blinded RCT supports specificity of SDARI activities on neural indices of social processing, sociometrics, and observer-rated social behavior, but suggests the latter may be accomplished – albeit more slowly – without SDARI activities, a pattern found in a previous RCT (Lerner & Mikami, 2012). It also suggests that parent-report effects are attributable to expectations of condition, not SDARI-specific. Future work should employ such designs to better elucidate what aspects of other SSIs are responsible for their effects.