Second-Order False Belief Reasoning, Recursive Language Competencies and Working Memory in Children with ASD

Background: Second-order (SO) false belief (FB) competency is an important component of Theory of Mind (TOM). In first-order ToM development, links between language and false belief reasoning have been firmly established and proposed as potential factors in explaining the success of children with ASD who succeed on TOM tasks. However, SOFB development is less studied, and the links between language and SOFB reasoning, as well as the potential developmental role of language, remain unclear.

Objectives: We investigate the relationship of SOFB and several language measures, and in particular, whether competency in linguistic recursion (sentential complements) predicts the ability to reason about other's SOFBs. In addition, we investigate the role of working memory in SOFB development.

Methods: The sample consists of Danish speaking children with ASD without ongoing language delays and with working memory within the normal range. Four types of standard SOFB tasks are given. The standardized language measures include the Verbal Comprehension index from WISC-IV and TROG. Linguistic recursion is investigated using a new Danish language tool that measures comprehension of sentential complements. The tool was developed and validated for this study (Cronbach’s alphas 0.56). Working memory is measured by the working memory index from WISC-IV. We are using these measures in an ongoing correlation study. To date, 18 children with ASD, with mean age 12.61 (SD = 3.07), have been tested. ASD was diagnosed by psychiatrists, using the ICD-10 criteria.

Results:  The Pearson correlation analyses of the three language measures and working memory against three types of dependent variable yielded the following findings: Total SOFB correlates with Verbal Comprehension (r=.482, p=.043) and Working Memory (r=.598, p=.009). SOFB without justification correlates with Verbal Comprehension (r=.641, p=.004), Grammar comprehension (r=.558, p=.016) and Working Memory (r=.621, p=.009). Justification-only correlates with working memory (r=.571, p=.006) but none of the language measures.

Conclusions: First, the correlational analysis suggests that verbal concept formation skills (which are influenced by semantic knowledge) are related to SOFB reasoning. However this is not the case for syntactic skills, whether general grammar or recursive sentential complements comprehension. This result points to the influence of semantic and conceptual aspects of language, which are sometimes referred to as “tool for thinking”. Furthermore, our correlation data suggests that justification responses cannot be accounted for by language skills, and therefore does not support the claim that language plays an expressive role in SOFB reasoning. Second, the correlational analysis suggests that working memory is related to SOFB reasoning, including justification skills, and thus provides further support to “the complexity-only position” (Miller 2009), claiming that information processing skills predict SOFB development. By the time of IMFAR 2017 the size of our sample will enable us to draw conclusions about how much of the variation can be explained by relative contribution of each language measure and working memory based on multiple regression analysis.

Miller, S. A. (2009). Children’s understanding of second-order mental states. Psychological bulletin, 135(5), 749.