Opinion
Multiple conflict-driven control mechanisms in the human brain

https://doi.org/10.1016/j.tics.2008.07.001Get rights and content

Conflict between competing neural representations is thought to serve as an internal signal for the recruitment of ‘cognitive control’, which resolves conflict by biasing information processing in line with current task demands. Because conflict can occur at different levels of stimulus and response representations, several recent investigations have examined whether conflict-driven cognitive control is domain-general or domain-specific, that is, whether control recruited by one type of conflict affects the resolution of another, but these studies have produced contrary conclusions. I argue here that a critical reading of this literature indicates that the effects of conflict-driven control are domain-specific and are probably mediated by multiple, independent conflict-control loops that can operate in parallel.

Introduction

‘Cognitive control’ describes the ability to generate, maintain and adjust sets of goal-directed processing strategies (task sets). One central question for theories of cognitive control is how the need for reinforcements or adjustments in task sets is determined. The influential ‘conflict-monitoring model’ [1] proposes that this function could be served by an evaluative mechanism monitoring for internal processing ‘conflict’, reflected in the simultaneous activation of mutually incompatible representations in the brain. Specifically, the model describes a regulatory conflict-control loop, consisting of a conflict-monitoring component (the dorsal anterior cingulate cortex) that detects any type of processing conflicts, and which forwards a conflict signal to a strategic control component (the dorsolateral prefrontal cortex) that in turn aims to resolve conflict by reinforcing top-down biasing processes associated with the current task set [1].

The fact that conflict-driven adjustments in performance have been documented across diverse tasks 2, 3, 4, 5, 6, in which conflicts might be generated in different ways, has raised the question whether conflict-driven control mechanisms operate in a conflict-specific manner. In the original model, the conflict-control loop is ‘domain-general’, in that the output of the conflict monitor does not convey the type or origin of conflict, and results in a quantitative but not qualitative change in top-down biasing [1]. For illustration, consider a task set that entails various top-down biasing processes, in a task in which several conflicts can arise from different sources. Assuming a domain-general conflict-control mechanism, the occurrence of any type of conflict would trigger an enhancement of all top-down biasing processes involved in the task set and should, therefore, lead to enhanced resolution of other types of conflict that might arise within the same task. However, if conflict-driven control were domain-specific, a given type of conflict would trigger an up-regulation only in the specific top-down biasing process(es) involved in resolving that particular conflict, leaving the resolution of other conflicts unaffected. If conflict-resolution were indeed conflict-specific, this would raise the additional question of whether conflict-resolution was mediated by a single control resource that flexibly resolves different conflicts in a conflict-specific manner, or by several specialized conflict-control loops that can operate in parallel.

Several recent studies have produced data of relevance to these questions 6, 7, 8, 9, 10, 11, 12, 13, 14, with some authors concluding that conflict-driven control generalizes across different types of conflict 9, 13, others concluding that control acts in a conflict-specific manner 6, 7, 8, 10, 12, and yet others reporting both domain-general and domain-specific effects 8, 11. Here, I argue that the majority of studies cited to support any one of these conclusions suffer from important methodological limitations. By mapping out methodological desiderata for assessing the specificity of conflict-driven control, and by vetting extant studies against these criteria, I conclude that the literature to date indicates that conflict-driven control is domain-specific and probably mediated by multiple independent conflict-control loops.

Section snippets

Establishing independent types of conflict

The main line of empirical support for conflict-driven control stems from the so-called ‘conflict adaptation effect’ [15] (for a recent review, see Ref. [16]) (Box 1) in stimulus-response compatibility (SRC) tasks (Figure 1). Here, the level of interference from task-irrelevant stimulus information (‘conflict’), reflected in slowed responses and decreased accuracy for incompatible relative to compatible stimuli, is found to be reduced after the processing of an incompatible as compared with a

Assessing the specificity of conflict-driven control

Once the independence of two types of conflict has been established, the question arises how to best combine them to assess whether control processes triggered by one type of conflict generalize to the resolution of another. Three approaches can be found in the literature: ‘task-switching’ designs 8, 9, 10, 13, ‘stimulus-switching’ designs 6, 12, 14 and ‘factorial task-crossing’ designs 7, 11, 14 (Figure 2). In task-switching designs, the relevant and irrelevant stimulus characteristics, in

Conflict-specific cognitive control mechanisms

Only a few studies in this literature have employed factorial task-crossing designs. Two experiments factorially combined the Stroop and Simon tasks (Figure 3a), and both found these conflict types to be non-interactive, and that each conflict enhances the resolutions of its own kind, but does neither facilitate nor impair the resolution of the other 7, 14. In another experiment, a factorial combination of the flanker and Simon tasks resulted in conflict-specific, but not conflict-general,

But is it ‘control’?

The fact that two conflict resolution processes can occur in parallel, without affecting each other, could provoke the argument that conflict adaptation does not actually represent an instance of controlled processing because such processing has traditionally been defined precisely through its liability to interference by simultaneous demands on cognitive control [28]. This, however, represents a circular argument because this traditional yardstick for controlled processing presupposes a

Conclusions

I have argued that many of the recent studies seeking to address whether conflict-triggered control mechanisms operate in a conflict-specific fashion, suffer from potential confounds related to non-independence of conflict types and task-switching effects, and that these issues can be overcome by an experimental strategy that combines different SRC tasks in a fully factorial design. The limited literature pursuing this strategy indicates that conflict adaptation processes are domain-specific

Acknowledgements

T.E thanks Chris Summerfield, Amit Etkin and Jim Monti for helpful comments during the preparation of this manuscript.

References (50)

  • E.J. Davelaar

    A computational study of conflict-monitoring at two levels of processing: reaction time distributional analyses and hemodynamic responses

    Brain Res.

    (2008)
  • M.M. Botvinick

    Conflict monitoring and cognitive control

    Psychol. Rev.

    (2001)
  • G. Gratton

    Optimizing the use of information: strategic control of activation of responses

    J. Exp. Psychol. Gen.

    (1992)
  • J.G. Kerns

    Anterior cingulate conflict monitoring and adjustments in control

    Science

    (2004)
  • W. Kunde

    Sequential modulations of stimulus-response correspondence effects depend on awareness of response conflict

    Psychon. Bull. Rev.

    (2003)
  • B. Sturmer

    Control over location-based response activation in the Simon task: behavioral and electrophysiological evidence

    J. Exp. Psychol. Hum. Percept. Perform.

    (2002)
  • D. Fernandez-Duque et al.

    Cognitive control: dynamic, sustained, and voluntary influences

    J. Exp. Psychol. Hum. Percept. Perform.

    (2008)
  • A.L. Freitas

    Contextual adjustments in cognitive control across tasks

    Psychol. Sci.

    (2007)
  • A. Kiesel

    Evidence for task-specific resolution of response conflict

    Psychon. Bull. Rev.

    (2006)
  • W. Kunde et al.

    Sequential modulations of correspondence effects across spatial dimensions and tasks

    Mem. Cognit.

    (2006)
  • W. Notebaert et al.

    Cognitive control acts locally

    Cognition

    (2007)
  • B. Sturmer

    Executive control in the Simon task: a dual-task examination of response priming and its suppression

    Eur. J. Cogn. Psychol.

    (2005)
  • M. Wendt

    Sequential modulations of interference evoked by processing task-irrelevant stimulus features

    J. Exp. Psychol. Hum. Percept. Perform.

    (2006)
  • U. Mayr

    Conflict adaptation effects in the absence of executive control

    Nat. Neurosci.

    (2003)
  • T. Egner

    Congruency sequence effects and cognitive control

    Cogn. Affect. Behav. Neurosci.

    (2007)
  • Cited by (348)

    View all citing articles on Scopus
    View full text