Individual Differences in Distinct Aspects of Executive Function Predict Theory of Mind in School-Aged Children with ASD

Background: Executive function (EF) and theory of mind (ToM) are often difficult for children with ASD. Previous research has consistently documented theory of mind (ToM) delays in preschoolers with ASD, but investigations examining ToM in school-aged children with ASD and average IQ have produced mixed results. Moreover, research is limited on whether distinct aspects of EF predict different types of ToM in ASD, above and beyond the influence of age and IQ.

Objectives: To examine the relation between individual differences in EF and ToM in school-aged children with ASD and average IQ using large batteries of ToM and EF measures.

Methods: Sixty-four children with ASD between the ages of 7-11 years participated. All children had an ASD diagnosis (ADOS-2, ADI-R, DSM-5) and WASI-2 Full Scale IQ of 80 or higher. The ToM battery included tasks measuring: (1) emotion recognition (TOM Test-Level 1), (2) first-order false belief (Location Change and Unexpected-Contexts False Belief Videos [TOM Videos], and TOM Test-Level 2), (3) second-order false belief (TOM Test-Level 3), and (4) spontaneous social attribution abilities (Social Attribution Task [SAT]). For the TOM Test (Muris et al., 1999) and TOM Videos, children viewed pictures and videos with pre-recorded questions. For the SAT (Klin, 2000), children viewed animated geometrical figures enacting a social scene and answered open-ended questions, which were coded into six index scores representing different aspects of spontaneous social attribution: Pertinence, Salience, Cognition, Affect, Person, and Problem-Solving. The EF battery included tasks measuring: (1) interference suppression (Attention Network Task [ANT] and Stroop task), (2) behavioral inhibition of a dominant response (Change task), and (3) working memory (Backward Digit Span).

Results: A bigger reaction time (RT) difference on the ANT task (more mistakes for the incongruent versus congruent flanker condition) related to lower TOM Videos scores, when controlling for age and VIQ (β = -.31, p = .01). A bigger RT difference on the Stroop task (more mistakes for the incongruent versus congruent condition) related to lower SAT Salience scores, when controlling for age and IQ (β = -.39, p < .01). Decreased verbal working memory on the digit span related to lower TOM Test-level 2 performance, when controlling for age and VIQ (β = .26, p = .04).

Conclusions: Above and beyond the influence of age and verbal IQ, individual differences in EF predicted ToM in school-aged children with ASD. First, children with ASD with decreased interference suppression for incongruent flankers had worse first-order false-belief performance. Second, children with ASD with decreased interference suppression of incongruent color-word pairs had a decreased ability to make salient social attributions about animated geometrical figures enacting a social scene. Finally, children with ASD with decreased working memory had worse first-order false belief performance. Overall, our results indicate that for school-aged children with ASD, interference suppression predicts both spontaneous ToM and first-order false belief, while verbal working memory predicts first-order false belief. Results suggest a clinically relevant link between individual differences in specific aspects of EF and social cognition in ASD.