Belief Reasoning in ASD: The Role of Language and Executive Functions

Background:  Deficits in theory of mind are central in autism spectrum disorders (ASD).  False belief (FB) tasks are often used to assess theory of mind.  Children with ASD show significant delays, not solving FB tasks until age nine  (Happé, 1996); typically developing children do so at age four on average (e.g. Wellman et al., 1990). Evidence suggests that language may be a necessary precursor to belief reasoning (e.g. de Villiers, 2005) because “inner speech” provides an on-line representation of another person’s perspective.  Newton and de Villiers (2007) tested this hypothesis by suppressing inner speech during FB.  A non-verbal interference task had no effect, but verbal interference interfered with FB.  An alternative explanation suggests that inner speech impacts FB performance via executive functions (EF; Baddeley et al., 2001; Lidstone et al., 2010) including working memory (Hermer-Vazquez et al., 1999). Perhaps language mediates EF via EF’s role in belief reasoning (Apperly & Butterfill, 2009).  Language deficits are central to ASD; this permits an examination of relationships among language, EF and theory of mind. The present study examines the role of inner speech in FB in ASD. People with autism may rely more heavily on visuo-spatial than linguistic processes (e.g. Williams et al., 2012); verbal loads have a lesser performance impact (e.g. Holland & Low, 2010). 

Objectives:  Our goals were to test the effects of a) verbal load versus spatial load on FB in ASD and b) verbal load on FB specificity versus EF generality (both false and true belief conditions).  We hypothesized that verbal load would disrupt belief reasoning less than visuo-spatial load in ASD, and that load tasks would affect executive functions generally, impacting both true and false belief performance.

Methods:  Adolescents ages 12-17 with ASD (n = 10; final sample size will be n = 15 per group) and typical development (n = 9) matched on age, NVIQ, and VIQ, p’s > .28 viewed silent location change FB and true belief videos, and simultaneously performed verbal load (VL) or a spatial load (SL) task. Participants also completed EF tasks. 

Results:  Results indicated no main effects of group, belief condition, and no interactions, all p’s > .50, with one exception: a significant main effect of load, p = .048, such that performance was worse on VL versus SL across groups and belief conditions. Better VL task performance predicted higher EF scores, p’s < .05.

Conclusions:  Results of this study suggest that inner speech suppression interferes with belief reasoning equally for individuals with and without ASD, and further, that effects of inner speech suppression are not unique to false belief.  Verbal load interfered similarly with performance on false and true belief tasks.  Both false and true belief videos required working memory for complex representations in mind.  Findings suggest that inner speech is involved in EF tasks, including belief reasoning, and that individuals with worse baseline EF may be particularly susceptible to effects of inner speech suppression.