The Relationship of Social Cognition, Language, and Executive Function to Theory of Mind in Children and Adults with ASD

Background:  Whereas it is generally accepted that individuals with autism spectrum disorders (ASD)  have deficits in theory of mind (ToM), or the ability to understand that other people have thoughts and to infer or predict what those thoughts might be, the relationship of this deficit to other aspects of ASD, particularly social cognition, language, and executive function is still debated. Previous studies have been limited by only examining one or two of these cognitive abilities in relationship to ToM, in a limited number of participants, and/or in only including children or adults in the sample.  

Objectives:  The current study examined the relationship of measures of social cognition and language/verbal ability to measures of ToM in a large sample of 272 children and adults with ASD ranging in age from 6 to 53 years.  In addition, large subgroups of children and adults within this sample had received the same executive function measures of working memory and cognitive flexibility making it possible to examine relationships between these variables to ToM in cohesive set of individuals.

Methods:  The primary dataset contained 272 participants with ASD (166 individuals ages 5 to 15 years and 106 individuals ages 16 to 53 years) mean age 16.2 (SD = 7.9; range 6.0 to 53.8). It included 239 males and 33 females with mean VIQ 104.7 (SD = 14.5; range 73 to 141), mean PIQ 106.3 (SD = 14.5; range 66 to 144), and mean FSIQ 106 (SD = 13.9; range 73 to 146).  Three hypotheses of the relationship between the variables were examined using two different versions of three different hierarchical multiple regression models. The ToM measure was the outcome or predicted variable.  Two different ToM measures were used as dependent variables: 1) an aggregated measure from the traditional Sally and Anne, John and Mary, and Peter and Jane tasks; and, 2) the Reading the Mind in the Eyes Test-Revised. Independent variables included measures of social cognition (from the ADOS & ADI-R), language ability, verbal and spatial working memory, and cognitive flexibility, which were entered at different steps of the analyses based on the predictive model that was being tested. Age was entered as a control variable in all of the models. 

Results:  A series of hierarchical linear regression models indicated that the strongest relationship occurred between language ability (as measured by Verbal IQ) and performance on both the traditional ToM measures and the Eyes test.  Social cognition abilities were a greater predictor for ToM in children than adults with ASD, but these abilities were not the primary predictor.  No executive functioning measures were revealed to be significant predictors of ToM ability in ASD for either the children or adults.

Conclusions:  In both children and adults with ASD, ToM abilities appeared to be related to overall language abilities rather than a more generalized ability in social cognition or executive function. This relationship occurred both for ToM as derived from traditional measures which are more language dependent and for ToM as derived from the Eyes test, a task which is less language-dependent.