Cognitive Flexibility Among Individuals with Autism: The Influences of Chronological Age Vs. Mental Age

Background: Cognitive flexibility, the ability to switch sets in thinking, is a component of executive function that appears to be impaired among children with autism (Corbett et al., 2009; Russo et al., 2007; Sanders et al., 2008). Although the ability to complete tasks of cognitive flexibility is generally linked to higher verbal abilities (Ardila, Pineda & Rosselli, 2000; Jacques & Zelazo, 2005), the relative non-verbal strengths among persons with autism (Lincoln et al., 1995; Mayes & Calhoun, 2003) allows for a unique opportunity to study the relative contributions of verbal and non-verbal abilities on the Flexible Item Selection Task (FIST; Jacques & Zelazo, 2001), a widely used test of cognitive flexibility.   

Objectives: The objective of the current study was to examine the relative influences of chronological age (CA), performance mental age (PMA), and verbal mental age (VMA) on performance on the FIST among a group of children and adolescents with autism. 

Methods: Fourteen individuals with autism, ranging in CA from 53 months to 206 months (M=133.1 mos, SD=47.1 mos), participated in this study. The participants were administered the PPVT-III (Dunn & Dunn, 1997) to determine VMA and the Leiter-R (Roid & Miller, 1997) to determine PMA. They had a mean VMA of 51.5 (SD=20.86) based on scores from the PPVT-III and a mean PMA of 76.1 (SD=22.72) based on scores from the Leiter-R. The FIST, a picture based task that required matching three objects on two different dimensions, was used as a measure of cognitive flexibility.

Results:  Pearson correlations for PMA and both first r (14) = .62, p <.05 and second pair match r (14) = .69, p <.01 on the FIST were significant, while VMA was only significantly correlated with FIST second pair match r (14) = .59, p <.05. First pair matching assesses abstraction abilities, whereas second pair matching assesses cognitive flexibility. Chronological age was not significantly correlated with FIST performance. A stepwise regression analysis on second pair selections revealed that PMA was the best predictor of cognitive flexibility (F (1,12) = 10.99, p <.01, adjusted R²=.44).  

Conclusions: Chronological age was not found to have any discernible relationship with cognitive flexibility, supporting the premise that MA is a better indicator of cognitive development than CA (Burack et al., 2001, 2004; Iarocci et al., 1997). Despite previous evidence that verbal intelligence predicts performance on tasks of cognitive flexibility, including the FIST, we found that non-verbal intelligence (PMA) was the better predictor among children with autism. As individuals with autism typically display patterns of strength in non-verbal cognitive abilities and weaknesses in verbal areas (Lincoln et al., 1995), the findings may indicate a tendency for people with autism to rely on non-verbal abilities to complete tasks that require cognitive flexibility, unlike typically developing children who appear to rely more on verbal abilities.