Objectives: The goals of the current study are (1) to examine the associations between physiological and behavioral indices of response monitoring, and (2) to examine the mediating role of self-perceived locus of control in the association.
Methods: Preliminary data are presented on 37 (32 male) higher functioning (IQ> 70) children with autism (9 – 16 years) and a comparison sample of 34 age-, IQ-, and gender-matched typically-developing children. EEG was collected continuously from 22 scalp sites as children completed a modified version of the Eriksen Flanker Task. Electrophysiological measures of response monitoring were measured at midline sites as the most negative peak 25 to 150 ms following error (ERN) versus correct (CRN) responses. Post-error slowing (RT on trial after error – RT on trial after correct) served as a behavioral index of response monitoring. Self-perceived Locus of Control was assessed using the Behavior Assessment System for Children – Self Version (BASC-2).
Results: A 2 (diagnostic group: HFA vs. comparison) x 4 (cortical site: Fz, FCz, Cz, Pz) x 2 (response type: error vs. correct) repeated measures ANCOVA controlling for error rates revealed that all participants had larger amplitude responses on error versus correct trials. This effect was generalized across all midline sites in the HFA sample, but localized to frontal sites (Fz, FCz) in the comparison sample, F(3, 204) = 3.31, p = .02. Within the comparison sample, enhanced amplitude ERN responses predicted more post-error slowing, r(33) = -.40, p = .02. In contrast, ERN amplitude did not predict post-error slowing in the HFA sample, r(37) = -.02, ns. Diagnostic group differences in the associations between physiological and behavioral measures of response monitoring appeared to be partially accounted for by differences in self-perceived locus of control. For the comparison sample an enhanced amplitude ERN was associated with a more internal locus of control, r(32) = .61, p < .001, whereas for children in the HFA sample an enhanced amplitude ERN was associated with a more external locus on control, r(37) = -.40, p = .02.
Conclusions: These results suggest that children with autism may have less well developed response monitoring tendencies as indicated by a lack of frontally-localized ERN responses. Unlike children in the comparison sample, enhanced physiological reactions to errors did not predict more behavioral compensation for errors. For children with autism, enhanced physiological reactions to errors in conjunction with feelings of a lack of personal control over their environment, may lead to behavioral and attentional disengagement in the context of errors. These results will be discussed in terms of the importance of using multi-method approaches to further integrate existing neurobiological and psychological models of autism.