Elsevier

NeuroImage

Volume 20, Issue 2, October 2003, Pages 1132-1139
NeuroImage

Regular article
A midline dissociation between error-processing and response-conflict monitoring

https://doi.org/10.1016/S1053-8119(03)00334-3Get rights and content

Abstract

Midline brain activation subsequent to errors has been proposed to reflect error detection and, alternatively, conflict-monitoring processes. Adjudicating between these alternatives is challenging as both predict high activation on error trials. In an effort to resolve these interpretations, subjects completed a GO/NOGO task in which errors of commission were frequent and response conflict was independently varied by manipulating response speeds. A mixed-block and event-related fMRI design identified task-related, tonic activation and event-related activations for correct and incorrect trials. The anterior cingulate was the only area with error-related activation that was not modulated by the conflict manipulation and hence is implicated in specific error-related processes. Conversely, activation in the pre-SMA was not specific to errors but was sensitive to the conflict manipulation. A significant region by conflict interaction for tonic activation supported a functional dissociation between these two midline areas. Finally, an intermediate, caudal cingulate area was implicated in both error processing and conflict monitoring. The results suggest that these two action-monitoring processes are distinct and dissociable and are localised along the midline.

Introduction

The smooth control of human behaviour requires individuals to be responsive to fluctuations in task demands. Such responsiveness might be triggered by the detection of errors in performance (Rabbitt, 1966) or by continuous monitoring of the demand placed by a task upon a person's limited cognitive resources (Botvinick et al., 2001). The importance of these monitoring processes, coupled with evidence of their altered function in several clinical groups Gehring and Knight, 2000, Carter et al., 2001, Alain et al., 2002, has increased interest in identifying their neuroanatomical bases. The empirical data suggest that midline cortical areas, with a particular focus on the anterior cingulate cortex (ACC), may be critical for these processes. An ERP component, the error-related negativity, has been identified with error processing and has been localised to the ACC Gehring et al., 1993, Dehaene et al., 1994. Functional brain imaging investigations have identified substantial error-related midline activations, thereby confirming these electrophysiological findings Kiehl et al., 2000, Menon et al., 2001, Garavan et al., 2002.

A more recent, alternative theory holds that error-related activations identified with the ACC may reflect the processing of increased response conflict rather than the processing of errors per se Carter et al., 1998, Botvinick et al., 2001. This theory posits a role for the ACC in the continuous monitoring of conflict created by the coactivation of competing response tendencies as typified by Stroop, flanker, or GO/NOGO tasks. Errors in tasks such as these might be assumed to occur on trials with relatively high response conflict; that is, the inappropriate response can be assumed to have been sufficiently activated to overwhelm the appropriate response and therefore to have created high interresponse conflict. Consequently, greater ACC activation for errors relative to control trials may be consistent with either an error processing or a response conflict monitoring role for this structure.

Recent evidence suggests that the error-detection and conflict-monitoring functions may be subserved by distinct midline regions. Functional imaging data Braver et al., 2001, Kiehl et al., 2000, Ullsperger and von Cramon, 2001 suggest that error detection may engage more anterior areas of the anterior cingulate while response conflict produces activation more posteriorly in the cingulate extending superiorly into the pre-SMA. Some support for this has been provided by human lesion and monkey unit recording studies Swick and Turken, 2002, Schall et al., 2002. However, to our knowledge, no midline statistical dissociations between these action-monitoring functions have yet been reported with the result that much controversy still remains regarding both the localization of these critical executive functions and, indeed, their very existence as separate cognitive processes. Consequently, in an attempt to resolve these various accounts of the midline's role in action monitoring, we manipulated response conflict in a GO/NOGO task that was designed to yield both successful inhibitions and errors of commission. The factorial design enabled us to separate error-related and conflict-related activations.

Section snippets

Subjects and task design

Sixteen right-handed subjects (10 female; mean age 31, range: 18–46), reporting no history of neurological or psychological impairment, participated after providing written informed consent. The GO/NOGO task, based on our earlier work Garavan et al., 1999, Garavan et al., 2002, required multiple responses and occasional response inhibitions. Subjects were presented with a 1 Hz serial stream of alternating letters, X and Y, and were required to make a button press response to each letter except

Task performance

The conflict manipulation proved successful with GO response times being faster in the HIGH CONFLICT condition relative to the LOW CONFLICT condition [326 ms vs 352 ms; t(15) = 4.8, P ≤ 0.0002]. The HIGH CONFLICT condition also produced more errors of commission [46% vs 17%; t(15) = 7.2, P ≤ 0.0001], and more errors of omission [4% vs 1%; t(15) = 3.1, P ≤ 0.007].

Tonic task-related activation

Areas tonically activated during the task included the pre-SMA (located on the medial frontal gyrus and rostral to the anterior

Error processing and response conflict

The results suggest a midline separation of neuronal activity for error-processing and response conflict monitoring functions. As shown in Fig. 3, anterior areas of the ACC showed significant responses to errors of commission but were not sensitive to the task's changing conflict demands. At the pre-SMA, this pattern was reversed and a significant conflict effect only was observed. The pre-SMA was also more active tonically in the higher conflict condition and this effect was statistically

Conclusion

The present results suggest that two action-monitoring functions, error-detection and response conflict monitoring, are distinct and can be identified with separate midline structures. These results speak to an emerging topography of executive functions wherein the control of our behaviour is accomplished by the coordinated operation of distinct brain regions performing distinct control tasks. However, irrespective of their separate anatomical locations, it is likely the case that these two

Acknowledgements

This work was supported in part by USPHS Grants DA09465, DA14100, and GCRC M01 RR00058. The assistance of Shelley Canevit and Michael SanFillipo in data collection is gratefully acknowledged.

References (37)

  • M. Ullsperger et al.

    Subprocesses of performance monitoringa dissociation of error processing and response competition revealed by event-related fMRI and ERPs

    NeuroImage

    (2001)
  • V. van Veen et al.

    Anterior cingulate cortex, conflict monitoring, and levels of processing

    NeuroImage

    (2001)
  • B.D. Ward et al.

    Non-linear regression for fMRI time series analysis

    NeuroImage

    (1998)
  • P.J. Whalen et al.

    The emotional counting STROOP paradigma functional magnetic resonance imaging probe of the anterior cingulate affective division

    Biol. Psychiatry

    (1998)
  • C. Alain et al.

    Neurophysiological evidence of error-monitoring deficits in patients with schizophrenia

    Cereb. Cortex

    (2002)
  • M.W. Botvinick et al.

    Conflict monitoring and cognitive control

    Psychol. Rev.

    (2001)
  • T.S. Braver et al.

    Anterior cingulate cortex and response conflicteffects of frequency, inhibition and errors

    Cereb. Cortex

    (2001)
  • S.A. Bunge et al.

    Prefrontal regions involved in keeping information in and out of mind

    Brain

    (2001)
  • Cited by (0)

    1

    Current address: NIDA-IRP, 5500 Nathan Shock Drive, Baltimore, MD 21224, USA.

    View full text