Post-Error Slowing May Not Be Proactive: Electrophysiological evidence favours a disorienting account

<p>Post-error slowing (PES) is widely assumed to reflect strategic response caution, but this assumption is based primarily on computational modelling rather than direct neural evidence. Dutilh and colleagues (2012b) compared five accounts of PES and concluded that slowing results from widened response boundaries reflecting increased caution. We used EEG to test predictions derived from three of the five accounts modelled by Dutilh and colleagues, as well as Notebaert and colleagues' (2009) orienting account. In a Go/No-Go task administered to healthy young adults (N = 94, 24.3 ± 4.8 yrs), ERPs provided temporally specific indices of the cognitive operations implied by these accounts, including stimulus evaluation, attentional processing, and error appraisal. A response-caution account would predict a larger N2 after errors and correlation between N2 amplitude and PES magnitude. Neither prediction was supported. Instead, the results favoured a disorienting account in which errors disrupt ongoing attentional processing. The anterior N1 was significantly attenuated after errors (p = .012), and a less negative post-error N1 was positively correlated with PES magnitude (p = .027). PES in this task was therefore more closely associated with attenuation of early attentional processing than with strategic response caution or the magnitude of error appraisal. This pattern is more consistent with attentional vulnerability than deliberate cognitive control.</p>