Examining Executive Functioning in Adolescents and Young Adults with Autism Spectrum Disorder Using the Nih Toolbox

Background:

Adolescents with autism spectrum disorder (ASD) often exhibit impairments in executive functioning (EF; Hill, 2004; Ozonoff et al., 2004, Solomon et al., 2017). However, previous work reporting these deficits has been inconsistent, and it remains unclear in which specific cognitive domains deficits exist. Furthermore, little is known about how the cognitive profiles of those with ASD change over time, particularly in the transition from adolescence into early adulthood, as well as the impact of EF impairments on pivotal outcomes, including adaptive functioning.

Objectives:

1) To characterize the EF profiles of adolescents and young adults with ASD and typical development (TD) in a large IQ-matched sample using well-validated measures of cognition and 2) to examine associations between EF profiles and adaptive outcomes.

Methods:

Participants were 66 youth with TD and 66 age, gender and FSIQ matched individuals with ASD (ages 12 – 22), who were qualified using gold standard ASD diagnostic measures (Figure 1). Participants completed the NIH Toolbox Cognition Battery (NTCB), a well validated suite of computer tasks consisting of measures of fluid intelligence, including cognitive flexibility (Dimensional Change Card Sorting; DCCS), inhibitory control (Flanker), episodic memory (Picture Sequence Memory; PSM), list-sorting working memory (LSWM), and processing speed (Pattern Comparison Processing Speed; PCPS), as well as crystallized intelligence, including picture vocabulary (PV) and oral reading (ORT; Akshoomoff et al., 2014). Participants also completed an updated version of the Global Social Global Role (GSGR), a clinician-administered assessment of social competence and engagement in age-appropriate roles (Cornblatt et al. 2007).

Results:

The ASD group performed significantly worse on Flanker, DCCS, and PCPS (p < .001, all BF₁₀ > 9000). Participants with ASD also performed worse on PSM (p = .018), but evidence was insubstantial (BF₁₀ < 2) and differences were not observed in LSWM. There were no differences in measures of crystallized intelligence (Figure 2). These findings generalized to the composites, with significant differences in fluid (p < .001) but not crystallized composite scores. Stepwise discriminant function analysis was used to examine the combination of NTCB measures that best predicted diagnosis. The extracted function consisted of PCPS and DCCS, and accounted for 34% of the variance in each group’s cognitive performance (Wilks λ = .757; Chi-square = 35.048, df = 2, p < .001). Additionally, higher fluid intelligence composite scores were positively associated with GSGR social functioning in ASD (r_s = .29, p < .05).

Conclusions:

Compared to a well IQ-matched TD sample, adolescents and young adults with ASD exhibited specific weaknesses in measures of fluid intelligence, specifically processing speed, cognitive flexibility, and inhibitory control. Furthermore, EF was positively associated with higher social competence, suggesting that the social impairments observed in ASD may be exacerbated or driven by EF impairments. Finally, discriminant function analysis suggests that reduced cognitive flexibility and processing speed may uniquely contribute to the cognitive profiles of individuals with ASD. These results further our understanding of the cognitive profiles of ASD and their influence on the adaptive outcomes observed in adolescence and early adulthood.