Quantifying Social-Communicative Function in ASD Via a Structured Social Attribution Task

Background:  With the move to an all-encompassing diagnostic classification – Autism Spectrum Disorders – from previously unsuccessful subtyping attempts, there will be renewed need for quantification measurements of the disability that have clinical and prescriptive validity. Individuals with ASD exhibit great difficulty in social cognitive tasks that require intuitive understanding of ambiguous social stimuli, these being less mediated by language, rote knowledge or explicit rules. Capitalizing on a classic social animation created by Heider & Simmel (1944), we previously quantified spontaneous narratives given after viewing the cartoon - in which geometric shapes enact a social story - and demonstrated significant social attribution deficits in higher functioning adolescents with ASD relative to matched controls (Klin, 2000), deficits that were not apparent in more traditional tests of social and mental state attribution. However, this procedure – the Social Attribution Task – required laborious codification of open-ended narratives and was impractical for more expansive usage. In this study, we tested the clinical utility of a simplified procedure in which 19 scenes from the original cartoon were isolated and participants provided answers in a multiple-choice format.

Objectives:  To examine the diagnostic discriminative utility of the Social Attribution Task – Multiple Choice (SAT-MC), and to assess its differential predictive power relative to social- communicative adaptive function, independent of verbal skill levels.

Methods:  The SAT-MC was administered to a heterogeneous group of children with ASD characterized with standardized diagnostic procedures (N=23; Age range 4.5 to 12 years; VIQ range 62 to 146), and to a control group matched on chronological age and Verbal IQ (N= 57). Adaptive skills in the areas of Communication, Socialization and Daily Living Skills were assessed with the Vineland Adaptive Behavior Scales. Correct answers on 19 items yielded a global SAT-MC score. Between-group comparisons on the SAT-MC score were performed to assess diagnostic discriminative utility. For the ASD group, correlational analyses between SAT-MC score and Vineland scores in the Communication and Social domains were performed to assess differential predictive utility, relative to the unrelated construct of Daily Living Skills. We performed correlations with Age and Verbal IQ in order to ensure that the SAT-MC taps on a developmental skill and is relatively independent from verbal skill level, respectively. Finally, individual SAT-MC items were analyzed for their specific discriminative power.

Results:  Performance on the SAT-MC differed significantly between the ASD and TD groups. SAT-MC scores were positively correlated with age (r= .474) and were independent from Verbal IQ (r= .236). SAT-MC scores were strongly correlated with Vineland Communication (r= .464) and Socialization (r= .482) standard scores but not with Vineland Daily Living scores (r= .116). Item analyses revealed variability of individual items’ diagnostic discriminative power.

Conclusions:  The SAT-MC was shown to discriminate a heterogeneous group of children with ASD from matched controls, and to differentially predict levels of social and communicative adaptive skills independently from verbal function. This initial study corroborates previously demonstrated deficits in social attribution and holds promise for quantification of skills that are essential for successful adaption in real life.