Neural and Behavioral Responses in an Executive Functioning Task Predict Social Communication Symptoms in Children with Autism Spectrum Disorder

Background:  Underlying many of the social and cognitive symptoms exhibited in individuals on the autism spectrum is a deficit in executive functioning (Robinson, 2009). Executive functioning is defined as the ability to manage complex problems in the service of a goal. This skill involves the inhibition of conflicting information, decision-making, and shifting of attention. Impairments in these cognitive capacities are often observed in ASD, and can have direct implications for the ability to successfully navigate social situations. Little research has been done on whether specific neural components elicited during executive functioning tasks can predict the social deficits exhibited in ASD.

Objectives:  To assess the correlation between inhibitory control, a core domain of executive functioning, and social communication symptoms in children on the autism spectrum, using behavioral and electrophysiological data.

Methods:  72 children with ASD were recruited from the University of Washington and Boston Children’s Hospital for participation in this study. Children ranged from 7 to 11 years old (64 males; 8 females) with a minimum IQ of 80 measured using the WASI-2 (Wechsler, 2011), and diagnosis was confirmed through DSM-V criteria (American Psychiatric Association, 2013) after administration of the ADOS-2 (Gotham, Risi, Pickles, & Lord, 2007) and the ADI-R (Rutter, Le Couteur, & Lord, 2003). Parents completed the Social Responsiveness Scale (SRS), the Social Skills Improvement System (SSIS), and a Vineland-II interview on behalf of their child. To measure inhibitory control, both percent accuracy and reaction times were recorded during a flanker task called the Attentional Network Task (ANT). Electrophysiological recordings were also collected during this task to examine the N2 and P3 ERP components. Difference scores for both ERP amplitude and latency were computed for each of these components by taking ERP cluster averages for congruent trials, and subtracting from the incongruent trial average. Larger difference scores suggest more effortful processing for the more difficult incongruent trials of this task.

Results:  Linear regression analyses were performed controlling for age, IQ, and gender. Severity scores on the ADOS Social Affect subdomain were predicted by percentage of correct incongruent trials (R²_change =.0552, t(72)=-2.094, p=.040), and reaction time for correct congruent trials (R²_change=.049, t(72)=2.029, p=.046), such that more severe social symptoms correlated with poorer performance. Percentage of incongruent trials also predicted scores on the SRS (R²_change =.191, t(65)=-3.846, p=.000), The amplitude of the N2 component predicted scores on the Social Engagement domain of the SSIS (R²_change=.106, t(54)=2.501; p=.016).

Conclusions:  The severity of social symptoms in our population of children with ASD was predicted by both behavioral performance and specific ERP components in an executive functioning task measuring inhibition. Poorer performance and slower reaction times in this task, as well as larger differences for N2 amplitudes between congruent and incongruent trials, was associated with greater deficits in social communication. The ability to inhibit conflicting or interfering information is an aspect of executive functioning vital to communicating and interacting with others in one’s environment, and our findings suggest a neurological source for these social deficits in children with ASD.