The Association of Early Theory of Mind to Two-Year-Later Pretend Play Performance in Children with Autism Spectrum Disorder

Background: Theory of mind (ToM) deficit and poor pretend play are two characteristics of children with autism spectrum disorder (ASD). ToM is a social cognitive ability to infer the mental states of self and others and thereby to make appropriate replies. Via ToM, children can impute others’ beliefs to usual presentations and then can generate new ways to pretend. Therefore, children’s ToM ability may benefit the generation of pretend play. However, the association of early ToM development to later pretend play performance has not been examined.

Objectives: The main purpose of this study was to clarify how early ToM ability is related to two-year-later pretend play in children with ASD.

Methods: A total of 83 children with ASD aged 5-12 years were included for two evaluations. In the first evaluation, the Theory of Mind Task Battery (ToMTB), Verbal Comprehension Index of the Weschler Intelligence Scale for Children, Fourth Edition, (VCI, WISC-IV), and Childhood Autism Rating Scale (CARS) were respectively administered to assess ToM, verbal ability, and symptom severity. Two years later, in the second evaluation, the Child-Initiated Pretend Play Assessment (ChiPPA) was administered to evaluate children’s pretend play in two sessions: conventional imaginative play and symbolic play. In these two sessions, the percentage of elaborate pretend play actions (PEPA), number of object substitutions (NOS), and number of imitated actions (NIA) were respectively scored as measures of the quality, quantity, and creativity of pretend play. Pearson's product-moment correlation coefficients and linear regression models were used for data analysis.

Results: The children’s mean age was 71.98 months (SD = 19.82) at the first evaluation. The earlier ToMTB scores (r=.256, p=.019; r=.217, p=.049; r=-.220, p=.046) had positive associations with the PEPA-conventional imaginative play and total scores, as well as negative association with the NIA-symbolic play scores The earlier VCI (r=.240, p=.029; r=.275, p=.012; r=.300, p=.006) had positive associations with the PEPA-conventional imaginative play, symbolic play and total scores. The earlier CARS scores had positive associations (r=.264, p=.016) with only the NOS-conventional imaginative play scores. The results of the regression models showed that in addition to the earlier VCI and CARS scores, the earlier ToMTB scores could not additionally predict the PEPA-symbolic play scores. Other models were not significant.

Conclusions: Children with ASD who have better early ToM ability are likely to have better quality and creativity of pretend play. However, despite considering early VCI and symptom severity, the early ToM ability could not further predict pretend play. Verbal ability seemed to be the more important predictor of pretend play in children with ASD. ToM ability is known to have close developmental coexistence with verbal ability. Therefore, ToM ability may be associated with pretend play via the mediation of verbal ability in children with ASD. The association of early ToM, verbal ability, and symptom severity to pretend play in children with ASD will require further research with larger sample sizes to provide more solid evidence.