Cognitive control adjustments are dependent on the level of conflict

Abstract

The congruency sequence effect (CSE) is one of the most investigated effects in the cognitive control literature. The conflict monitoring theory suggests that the CSE is the result of adjustments in cognitive control based on perceived conflict. A recent paper by Zhang and colleagues, has investigated whether the manipulation of conflict level by changing distractor incompatibility in a flanker task affects the amount of adjustments in cognitive control. Their study produced mixed findings, somewhat supporting the original conflict monitoring theory, but also suggesting other explanations, such as the repetition expectancy account. We replicated the experimental design in a multisite online study (N = 347), with Hungarian, Italian, and Singaporean participants. Our results supported the prediction that changes in the level of conflict trigger conflict adaptation, revealing that increasing conflict levels induced stronger adaptive control. Bayesian hypothesis testing indicated a monotonic reduction in congruency effects as a function of previous conflict strength. This finding is in line with the extension of the traditional conflict monitoring theory, as well as other theories like affective signaling and expected value of control, implying that the relationship between conflict and interference effects is gradual, rather than a binary function.

Figure 1

The predicted patterns in RT based on the repetition expectancy and the extended conflict monitoring theories. In the repetition priming model suggested by Zhang and colleagues, conflict type differences are dependent on the expected repetition of the previous conflict level, regardless of the relative conflict level of the previous trial: RTs decrease when the current and previous conflict levels are the same, while increase when current and previous conflict levels are different. In contrast, the extended conflict monitoring account predicts RTs to decrease in current low-, and high-conflict (incongruent) trials with the level of previous conflict, while RT to increase on current no-conflict (congruent) trials with the previous level of conflict.

Figure 2

Timings and stimulus types of a trial.

Figure 3

Reaction time for different levels of conflict in the current trial.

Figure 4

Interaction between previous and current trial congruency type. Mean RT-s on current congruency × previous congruency. Error bars represent the standard error of the mean.