ERN As a Predictor of Treatment Response to Social Skills Interventions in ASD

Background: Larger neural responses to error commission are a biomarker for increased anxiety symptoms (Meyer et al., 2015) and are measured by error-related negativity (ERN) amplitudes, an event-related potential (ERP) component measured via electroencephalogram (EEG). However, the extent to which this is true in youth with ASD is still unclear (Boulter et al., 2014; vs. Henderson et al., 2006; 2016). Moreover, while anxiety, when measured by informant-report, has been shown to moderate treatment outcomes in ASD (e.g., Pellecchia et al., 2015), physiological indices of anxiety, such as the ERN, have not been examined in this manner. Therefore, we aimed to examine the relation of the ERN to 1) anxiety symptoms in youth with ASD, and 2) anxiety and ASD-related treatment outcomes in youth with ASD following Social Skills Intervention (SSI).

Objectives: This study examined whether differences in neural responses of ERN amplitudes at baseline index anxiety in ASD and predict treatment response to SSI.

Methods: Thirty-seven youth (M_age=12.11, SD_age=2.91; 29 male) with IQ≥70 (M_IQ=104.68, SD_IQ=15.84) and ADOS-2-confirmed ASD diagnosis participated in a 10-week SSI. At pre- and post-treatment, parents completed ratings of psychopathology, anxiety and ASD symptomatology, respectively (CASI-5, Gadow & Sprafkin, 2016; SRS-2, Constantino & Gruber, 2005), and participants’ rated general and social anxiety symptoms (MASC-2; March et al., 1997; SAS, La Greca & Lopez, 1998). At baseline, unstandardized residual scores of the mean ERN amplitudes were computed by saving the variance leftover in a regression equation wherein the Correct Response Negativity (CRN) was regressed on the ERN (Meyer et al., 2016).

Results: Residualized ERN amplitudes correlated with parent reported symptoms of general anxiety disorder at baseline on the CASI-5 (r= -0.322, p= 0.052). Residualized ERN amplitudes predicted change (via ANCOVA-of-change models) in self-reported treatment outcomes of total social anxiety symptoms, as measured by the SAS (B=0.400, p=0.001) and anxiety symptoms, as measured by the MASC-self (B= .887, p=0.008), wherein smaller ERN amplitudes predicted greater treatment-related improvements, but larger ERN predicted attenuated improvements. MASC-self effects were driven by general anxiety disorder, humiliation and rejection, performance fear, social anxiety, obsession and compulsions, panic, and physical symptoms (all B > .608, p < .041), but not separation anxiety, tension or restlessness, or harm avoidance (all p>.241). No effects of ERN on treatment outcomes of ASD related symptoms on the SRS were found.

Conclusions: Results suggest that the ERN relates to concurrent parent-reported anxiety symptoms in youth with ASD, replicating previous work in children with anxiety disorders (Meyer et al., 2015) and ASD (Boulter et al., 2013, but see Henderson et al., 2006, 2016). We found that the direction of change on multiple self-report anxiety measures was dependent on baseline ERNs, such that smaller ERNs predicted augmented – but larger ERN predicted attenuated – treatment effects. This pattern of results suggests that the ERN may be a marker for subjective anxiety treatment response, but not parent-observable anxiety symptom or ASD symptom treatment response after SSI. Thus, the ERN may be a useful predictor of individual differences regarding self-reported anxiety symptom reduction following an SSI.