Event-Related Potentials Linked to Executive Function Predict Intervention Response and Later Social-Emotional Function Among Children with ASD

Background: Executive function (EF) underlies the ability to strategically approach novel and complex problems. By mid-childhood, it is comprised of inhibition, set-shifting, and working memory. Understanding individual differences in EF and its underlying neural profile is critical because subsequent social function and anxiety are predicted by EF in children without ASD. The N2 event-related potential (ERP) is thought to measure monitoring of conflicting information, enabling inhibition of extraneous information. We have shown that children with ASD have less efficient neural response (i.e., larger overall N2 amplitudes) than children without ASD and larger relative amplitude for more challenging trials related to EF behavior (Faja et al., 2016). The predictive utility of this neural EF signature for changes in social functioning and anxiety in ASD is relatively unexplored.

Objectives: To examine individual differences in ERPs as a predictor of later social function and anxiety. To examine whether ERPs predict social and anxiety changes related to an intervention that targets EF, metacognition, and emotion regulation.

Methods: As part of a clinical trial of EF intervention, 71 children (7 female) with ASD aged 7-11 years with full-scale IQ of M=106 (range: 80-135) participated; 49 provided adequate ERP data at baseline. Thirty-five were randomly assigned to an EF training program and 33 to a waitlist (3 were lost to attrition). Parents responded to the CBCL about emotional and behavioral challenges and the SSIS regarding social functioning. Electrophysiological recordings were collected to examine the N2 during a Flanker task. Mean amplitude was computed for each condition (congruent/incongruent). Regression models controlling for baseline social functioning and anxiety symptoms were used to examine whether baseline N2 predicted changes over a 3-month intervention period.

Results: Regardless of training assignment, baseline N2 amplitudes predicted social function on the SSIS at post-test controlling for baseline social function, such that more inefficient (larger) neural responses to incongruent flankers relative to congruent predicted smaller increases in social function (β = -.21, p=.045). Similarly, CBCL Anxiety at post-testing relative to baseline was predicted by overall N2 amplitude (β = -.24, p=.045), such that more efficient amplitudes predicted lower levels of anxiety. Within the intervention group alone, SSIS social change was no longer significant, but CBCL Anxiety was predicted by overall N2 amplitude at baseline (β = -.33, p=.06). That is, children who had smaller initial N2 amplitudes had a greater reduction in anxiety symptoms after treatment. Data from 5 subjects who are currently in progress will be included for final analyses.

Conclusions: Efficient neural response to extraneous stimuli may be an important predictor of individual differences in social skills growth and anxiety symptom reduction among children with ASD. In particular, overall social function and anxiety symptom improvements appear related to more efficient interference suppression. Moreover, more efficient interference suppression, as measured by the N2, predicted greater anxiety reduction among children in the EF training group. Thus, the N2 may provide a biomarker for predicting best responders to targeted intervention.