Discrete Electrophysiological Indices of Anxiety Differentially Predict Anxiety Symptom Reduction Following Group Social Skills Interventions

Background: While anxiety symptom reduction is a frequent target for common autism spectrum disorder (ASD) treatments such as Group social skills interventions (GSSIs; McMahon et al., 2013), individual differences in baseline anxiety can make it difficult to predict such response. Parent and child reports of anxiety symptoms are often highly discrepant and likely measure different anxiety constructs (White et al., 2015). Additionally, there are two established electroencephalography (EEG) indices that measure different manifestations of anxiety; 1) the error-related negativity (ERN), an EEG response to the commission of errors, which reflects endogenous threat sensitivity (Meyer, 2013), and 2) resting-state EEG frontal hemispheric asymmetry (HA) within the alpha frequency band, wherein left- and right-dominant patterns of HA relates to observable-approach motivation and withdrawal/anxiety, respectively (Coan & Allen, 2004; Thibodeau, Jorgensen, & Kim, 2006). However, whether these EEG indices of anxiety relate to discrete subjective reports of anxiety symptom reduction in GSSIs among youth with ASD has yet to be determined.

Objectives: We aim to measure the ERN and HA in ASD youth to predict anxiety symptom reduction in response to GSSI.

Methods: Fifty-three youth (M_age=11.60, SD_age=2.96; 38 male) with IQ≥70 (M_IQ=103.49, SD_IQ=15.40) and ADOS-2-confirmed ASD diagnosis participated in a 10-week GSSI. Pre-post anxiety symptoms were measured via self- and parent-report (MASC-2; March et al., 1997). Baseline ERN responses to the Flanker Task (Eriksen & Eriksen, 1974) were measured, such that more negative ERNs indicated greater threat sensitivity (Figure 1A). Baseline frontal alpha-band (7.5-12.5 Hz) power during eyes-open rest was measured, such that positive HA scores indicated greater relative left frontal alpha power (Figure 1B). ANCOVA-of-change models controlling for baseline anxiety symptoms were used to examine both neural measures’ ability to predict GSSI response.

Results: ERN amplitudes predicted changes in self-reported total anxiety symptoms (Figure 2A), generalized anxiety, social anxiety (including humiliation and rejections), and performance anxiety symptoms (all B>0.39, p<0.04), such that larger ERN amplitudes predicted greater improvements, but smaller ERN predicted attenuated improvements. In contrast, HA predicted changes in parent-reported total anxiety (Figure 2D), generalized anxiety, social anxiety (including humiliation and rejections), and self-reported tense and restlessness symptoms (all B>0.31, p<0.05), such that more right-dominant HA predicted greater improvements, but more left-dominant HA predicted attenuated improvements.

Conclusions: These results suggest that the ERN specifically predicted subjective anxiety symptom changes, which further validates its use in measuring internally oriented threat sensitivity as an anxiety construct in ASD. In contrast, HA predicted parent-observable anxiety symptom changes, consistent with previous work suggesting it relates to overt approach-withdrawal behavior (Coan & Allen, 2004) and parent-reported treatment response (Burnette et al., 2011). Thus, results suggest that individuals with ASD that exhibit a smaller ERN or left-dominant HA may experience attenuated improvement of anxiety symptoms after GSSI, and utilizing these neural indices may be especially useful in social skills treatment planning for youth with ASD and comorbid anxiety.