Academic journal article Memory & Cognition

Central Interference in Error Processing

Academic journal article Memory & Cognition

Central Interference in Error Processing

Article excerpt

This study dealt with capacity limitations in error processing. Participants classified digits into three arbitrary categories (initial response). Half were required to correct their errors if an error was detected (correction response), and half were required to produce a second response, regardless of the correctness of the initial response (approval response). Auditory interference was introduced before, during, or after the initial response. Interference stimuli were to be recalled later and were, thus, considered to involve central processes. Results for before showed that although correction responses were elongated, approval responses given after erroneous initial responses were shortened. For during, both correction and approval responses were elongated. On the basis of our findings, we argue that the error process is generated before the erroneous response is given and that it is a central process in terms of being subjected to capacity limitations in the presence of other central processes.

One productive way in which to explore the functional characteristics of the human information-processing system is to focus on the processing capacity of various parts of the system. Although the capacity of the earliest stages of processing is high, the capacity of some later stages, such as response selection, is much lower (see, e.g., Pashler, 1994; Pashler & Johnston, 1998, cf. Navon & Miller, 2002). The goal of this study was to improve our understanding of capacity limitations in the error-processing system.

This question has important implications. For example, driving a car seems to require constant processing of steering errors. At least in some situations, such as heavy traffic or narrow roads, rapid indication and correction of an error are essential to our survival. If the error-processing activity is capacity limited, any activity that constitutes a response selection, even one as simple as choosing between the wiper control stick and the indicator control stick, might cause a delay in the detection of an error, and such delays are potentially fatal.

Various studies have established the existence of a central system for processing errors. Rabbitt (1966a), in a pioneering experiment, showed that participants were able to indicate when they erred without any error feedback from the experimenter. Later studies revealed that detection of an error is possible even without feedback from the participant's own peripheral neural system (Angel, Garland, & Fischler, 1971; Higgins & Angel, 1970). These findings led to the conclusion that the error information originates from the same system that generated the error. Later psychophysiological studies showed that when a participant makes an error in a choice response time (RT) task, a component of the event-related potential (ERP), called error-related negativity (ERN), is observed (Falkenstein, Hohnsbein, Hoormann, & Blanke, 1991; Gehring, Goss, Coles, & Meyer, 1993). The ERN is a sharp negative deflection in the ERP, which is monitored when participants commit an error. The ERN peaks around 100-150 msec following EMG activity onset (approximately 50-80 msec after the keypress response; Falkenstein et al., 1991; Gehring et al., 1993). Although it is not entirely clear whether the ERN is related to error detection or error correction, the mere fact that it originates almost simultaneously with response onset accords with the assumption regarding a central error-monitoring system that uses the same information as that used for response selection.

Although the existence of a central system for error monitoring has been well established, the exact way in which errors are processed for detection and subsequent correction is far from clear. ERN literature supplies several explanations for its behavioral significance, diverging in the nature and complexity of the process it manifests. Falkenstein et al. (1990, 1991) interpreted the ERN as reflecting error detection-that is, a mismatch of a rather simple process in which the actual response is compared with the required response. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.