Citation

Coles MGH, Scheffers MK, Fournier L. Acta Psychol. 1995; 90(1-3): 129-144.

Affiliation

Department of Psychology, University of Illinois at Urbana-Champaign 61820, USA.

Copyright

(Copyright © 1995, Elsevier Publishing)

DOI

unavailable

PMID

8525868

Abstract

Human performance is seldom perfect, and even when an overt response is correct it may be accompanied by partial-error activity that does not achieve the level of a complete incorrect response. Partial errors can be detected in measures of the lateralized readiness potential, of the electromyogram, and of response force. Correct responses accompanied by partial errors tend to have slower reaction times than "clean" correct responses (because of response competition), and condition differences in reaction time can, on some occasions, be explained in terms of differences in the incidence of partial errors. In two-choice reaction time tasks, partial errors are more frequent when the imperative stimulus contains information that favors both responses, than when it contains information that favors only one response. The non-random nature of partial errors supports the inference that partial information about the stimulus is used to guide responses. A similar inference is supported by the observation that, in hybrid choice Go/No-go tasks, the kinds of partial errors that follow a No-go stimulus represent activation of the response that would have been correct had the stimulus been a Go stimulus. Finally, we note that the human processing system is capable of monitoring its own behavior and of initiating remedial actions if necessary. The activity of an error-detection system, as revealed by measures of the error-related negativity, is related to the degree to which responses are slowed after errors.

Language: en