Pragmatic performance, its relationship with symptom severity, and early clinical predictors of pragmatics in 5 ~ 6-year-old children with autism spectrum disorder.

Background: Pragmatic language refers to using spoken language to convey messages effectively across diverse social communication contexts. However, minimal longitudinal research has focused on defining early predictors of pragmatic development in children with autism spectrum disorder (ASD).

Methods: In the present study, 71 children with ASD and 38 age- and gender- matched 24- to 30-month-old typically developing (TD) children were enrolled. Social-communication, language, and parent‒child interaction measures were collected for the ASD group at baseline. Three years later, all subjects were assessed for pragmatic ability via the Chinese version of the Language Use Inventory (LUI-Mandarin). First, the differences in pragmatic performance between the ASD and TD groups at follow-up were analyzed. Second, pragmatic performance was correlated with autism symptomatology at follow-up, as well as the structural language difficulties and joint engagement (JE) levels at baseline in the ASD group. Furthermore, hierarchical multiple regression analyses and machine learning techniques were performed to explore the effects of early potential predictors of pragmatic development in the ASD group.

Results: First, our results revealed that performance was significantly lower in the ASD than in the TD group with respect to the LUI-Mandarin total scores and subscale scores (except for subscale C). Second, correlation analysis revealed that more severe symptoms of ASD at follow-up were associated with lower LUI-Mandarin total scores and better language performance on the Gesell Developmental Schedules (GDS). Additionally, increased proportions of supported JE (SJE) states were associated with higher LUI-Mandarin total scores. In contrast, increased proportions of unengaged (UE) states were associated with lower LUI-Mandarin total scores in the ASD group. Third, hierarchical multiple regression analyses and machine learning techniques indicated that the proportions of SJE during parent‒child interactions, as well as the degree of social symptoms and structural language impairments, were significant contributors to pragmatic development for the ASD group in the prediction models.

Conclusion: In summary, our findings suggest that pragmatic language difficulties are present in children with ASD as early as preschool age. Additionally, given the close correlation between the LUI-Mandarin score and symptom severity on the ADOS/ADI-R, the LUI-Mandarin might be a good way to triage children who need to wait a long time for a more extensive evaluation. Furthermore, this study provides new insights into potential targets for pragmatic interventions, and interventions can be designed to promote SJE between caregivers and children in future work.