Elsevier

Acta Psychologica

Volume 142, Issue 2, February 2013, Pages 203-210
Acta Psychologica

Dissociating proportion congruent and conflict adaptation effects in a Simon–Stroop procedure

https://doi.org/10.1016/j.actpsy.2012.11.015Get rights and content

Abstract

Proportion congruent and conflict adaptation are two well known effects associated with cognitive control. A critical open question is whether they reflect the same or separate cognitive control mechanisms. In this experiment, in a training phase we introduced a proportion congruency manipulation for one conflict type (i.e. Simon), whereas in pre-training and post-training phases two conflict types (e.g. Simon and Spatial Stroop) were displayed with the same incongruent-to-congruent ratio. The results supported the sustained nature of the proportion congruent effect, as it transferred from the training to the post-training phase. Furthermore, this transfer generalized to both conflict types. By contrast, the conflict adaptation effect was specific to conflict type, as it was only observed when the same conflict type (either Simon or Stroop) was presented on two consecutive trials (no effect was observed on conflict type alternation trials). Results are interpreted as supporting the reactive and proactive control mechanisms distinction.

Highlights

► Proportion congruent effects cannot be explained by the same mechanism than conflict adaptation effects. ► Sustained effects of proportion congruent, which impact performance on a posterior phase in which proportion is not manipulated. ► Proportion congruent effects can be unspecific to conflict type while conflict adaptation effects are specific to conflict type.

Introduction

This manuscript focuses on the construct of cognitive control. The main function of cognitive control is to ensure that behaviour unfolds in line with task goals. In many contexts goal-driven behaviour requires responses that are driven by the selection of relevant sources of information amidst competing sources of distraction. For this reason, the mechanisms underlying distractor interference have become a central issue in the study of cognitive control. Generally speaking, the aim of such studies is to understand the control processes that modulate the processing of irrelevant sources of information.

Two behavioural effects in particular have played an important role in the cognitive control literature; conflict adaptation effects (Gratton, Coles, & Donchin, 1992) and proportion congruent effects (Lowe & Mitterer, 1982). An important issue in the study of cognitive control is whether these two effects have the same or different causes. It isn't clear whether proportion congruency effects are caused by a mechanism different from the mechanisms that could produce conflict adaptation effects. The primary aim of this manuscript is to address this outstanding issue.

A common strategy used to study cognitive control is to measure performance in interference tasks. For example, in the classical Stroop color-naming task (for a review see Macleod, 1991) participants are required to name the color in which color words are displayed. Response times (RTs) are reliably slower for trials where the name of the printed word is incongruent with its color (e.g. the word RED printed in green) compared to trials where the word and color are congruent (e.g., the word RED printed in red). This difference in performance provides a measure of the contribution of irrelevant word reading to performance, with greater amounts of word reading leading to larger differences in performance between congruent and incongruent trials (i.e., larger interference). Similarly, in the Simon task people are required to respond to a non-spatial dimension of target stimuli (e.g. color), which are presented, say, to the left or right of fixation by pressing response keys lateralized to the left or right (Simon and Craft, 1972, Simon et al., 1973, Simon and Small, 1969). Although target location is irrelevant for the task, people respond more quickly and accurately to targets appearing on the same side as the response location (e.g., a left target requiring a left hand response) than to targets appearing on the side opposite the response location (e.g., a left target requiring a right hand response). In this case, the difference in performance for these two trial types provides a measure of the contribution of irrelevant spatial processing to performance, with greater amounts of spatial processing leading to a larger Simon effect.

The fact that Stroop and Simon effects are driven by the processing of task-irrelevant stimulus dimensions suggests that the processing of these irrelevant stimulus dimensions occurs automatically and is difficult to prevent. However, these effects can be subjected to modulation, as conflict adaptation (CA) and proportion congruent (PC) effects reveal.

The CA effect is defined by congruency effects that are smaller on a current trial when preceded by an incongruent trial than when preceded by a congruent trial (Gratton et al., 1992), and have been explained by different approaches. According to the “Control Monitoring Theory” (Botvinick, Braver, Barch, Carter, & Cohen, 2001) the detection of conflict on incongruent trials by the anterior cingulate cortex (ACC), leads to the recruitment of control by the activation of the dorsolateral prefrontal cortex (DLPFC), which will recruit other areas responsible to implement control by reducing the interference from an irrelevant distractor. Consequently, when the following trial is also incongruent, the cognitive control system is already activated and prepared to handle interference from the upcoming irrelevant distractor, with the outcome of a reduced congruency effect.

However, this theory has been challenged by several authors who explain CA effects as the result of mere priming and/or learning processes (Hommel et al., 2004, Mayr et al., 2003, Nieuwenhuis et al., 2006). Specifically, they pointed out that congruent followed by congruent trials (cC transitions) or incongruent followed by incongruent trials (iI transitions) can include complete repetitions of the full event (target + response). In contrast, this does not occur in mixed transitions (i.e. incongruent followed by congruent trials (iC) or congruent followed by incongruent trials (cI)), where only partial repetitions can occur. Note that CA effects are usually due to improved performance for cC compared to iC transitions and for iI compared to cI transitions. Therefore, taking into account that responses are faster in complete repetitions, according to the priming and/or learning processes CA could be due to repetition priming speeding up cC and iI trials. In fact, when considering the contribution of all those 4 possible transitions, Schlaghecken and Martini (2012) showed that CA effects were driven mostly by faster responses on cC trials relative to iC, and less by faster responses on iI trials relative to cI.

In an attempt to reconciliate both reinforcement learning and conflict monitoring accounts, hybrid learning-conflict models have been proposed (Davelaar and Stevens, 2009, Verguts and Notebaert, 2008, Verguts and Notebaert, 2009), highlighting the idea that cognitive control results from interactions between binding processes and conflict detection, for example, by conflict indicating where binding should be applied (Verguts and Notebaert, 2008, Verguts and Notebaert, 2009).

Another well known effect reflecting cognitive control adjustments to task demands is the PC effect. This effect is measured by manipulating the relative proportions of congruent and incongruent trials within an experimental block. The magnitude of the congruency effect varies with the proportion of congruent trials, being larger in the context of a high proportion of congruent trials than in the context of a low proportion of congruent trials (e.g. Carter et al., 2000, Logan and Zbrodoff, 1979, Lowe and Mitterer, 1982, West and Baylis, 1998). This modulatory effect is most commonly attributed to the adoption of a sustained strategy or task set, probably implemented after having experienced the level of conflict encountered on the first few trials in a block. This task set produces tonic changes in processing by, for example, altering the ‘weighting’ of word-reading relative to color-naming (e.g. Cohen, Dunbar, & McClelland, 1990).

Similar to CA effects, alternative accounts of the PC effects have been proposed. For example, Risko, Blais, Stolz, and Besner (2008), in line with the priming and binding proposals described beforehand for CA effects (Hommel et al., 2004, Mayr et al., 2003, Nieuwenhuis et al., 2006), suggested that the different ratio of complete repetitions/alternations to partial repetitions of features, also present in high vs. low proportion congruent blocks of trials explained the observed PC effect.

Furthermore, recent studies have shown that PC effects can also be observed under conditions in which a sustained strategy for all trials within a block cannot possibly explain the effect. In these studies, proportion congruent is manipulated independently for two sets of stimuli or contexts that are intermixed at random within a block of trials (Crump et al., 2006, Crump et al., 2008, Jacoby et al., 2003). Critically, in these studies the overall proportion of congruent and incongruent trials is kept at .50. Even though, particular items (Jacoby et al., 2003) or contexts (Cañadas et al., 2012, Crump et al., 2006) entail a high (or low) proportion of congruent trials, whereas other items or contexts entail a low (or high) proportion of congruent trials. The key result is again larger congruency effects for the items or contexts associated with a high proportion of congruent trials. Interestingly, as the type of item or context is manipulated randomly from trial to trial, the effect cannot be explained on the basis of a general, sustained, strategic attentional control set. Instead, the effect must be explained by reference to processes initiated at onset of the stimulus that perhaps cue the retrieval of prior memory episodes involving similar items/contexts. In turn, the cognitive control operations engaged on the current trial may be those that are retrieved as part of the memory representation of similar prior trials (see also Mayr et al., 2003).

A still unsolved question is whether the two above described forms of cognitive control adjustments (CA and PC effects), are reflections of the same or different control mechanisms. According to the above-described conflict monitoring account there is no need for sustained strategic control to explain PC effects. Instead, PC effects can be explained by the same transient or reactive mechanism proposed to account for CA effects. By this view, in a high conflict condition (low proportion of congruent trials), incongruent trials are very frequent, both as current and previous trials. Consequently, the number of iI transitions (incongruent trials preceded by incongruent trials) increases, leading to an overall reduction of interference effects.

An alternative view is that CA and PC effects tap separate control mechanisms (De Pisapia and Braver, 2006, Dosenbach et al., 2008), one responsible for transient shifts of control and the other responsible for more sustained forms of control. For example, De Pisapia and Braver (2006) propose the Dual Mechanisms of Control framework, a computational model with two separate mechanisms. On the one side, they propose a reactive mechanism responsible of CA effects through transient activation of prefrontal cortex (PFC) based on conflict detected in ACC over a short-time scale (in the order of milliseconds). A second mechanism would be responsible for PC effects, and is characterized by the sustained active maintenance of task-set information in a separate PFC module, which is driven by long time-scale conflict detected in a separate ACC unit (in the order of several seconds or minutes).

In agreement with this approach, Funes, Lupiáñez, and Humphreys (2010b), recently reported a behavioural dissociation between CA and PC effects in the context of conflict tasks, and concluded that they reflect different mechanisms. Concretely, they used a paradigm in which participants were to respond to up/down pointing arrows by pressing a left or right key. Arrows could be randomly presented either in the horizontal axis (to the left or to the right of fixation, in which case Simon interference was observed), or in the vertical axis (above or below fixation, in which case Spatial Stroop was observed). They manipulated the proportion of congruent trials for just one conflict type (Simon), leaving neutral (i.e., 50% congruent and 50% incongruent) the overall congruency ratio on the other conflict type (Spatial Stroop). The Simon and Stroop conflict trials were randomly presented within the same block of trials and participants had to perform the same task on all of them (i.e., to press a left-right key depending on whether the arrows pointed up or down). Importantly, a PC effect was observed for both Simon and Spatial Stroop despite manipulating proportion congruent for Simon trials only. At the same time, the CA effect was present only when the same type of conflict repeated on consecutive trials (i.e., they were observed from Simon to Simon and from Stroop to Stroop trials, but not from Simon to Stroop or from Stroop to Simon trials). This conflict-type specificity of CA effects has been consistently observed with different paradigms (Egner et al., 2007, Funes, Lupiáñez and Humphreys, 2010a, Funes, Lupiáñez and Humphreys, 2010b, Kiesel et al., 2006, Verbruggen et al., 2005, Wendt et al., 2006).

Thus, PC effects generalized in Funes et al. study across different conflict types combined within a single task, while CA effects seems to be conflict type specific. This dissociation was interpreted as evidence of different mechanisms underlying the two effects. However, Funes et al. study (2010b) measured PC effects under proportion of congruency manipulations, that is, under over-represented cC transitions for the high proportion congruent condition and iI transitions for the low proportion congruent condition. Furthermore, they measured PC effects on-line with the same trials with which the proportion of congruency was manipulated. Therefore, although it is assumed that PC effects are the results of the application of a sustained control set, there is no evidence for their sustained nature in this study.

Therefore, in the present experiment we define two main goals. The first one is to show whether we can find evidence of PC effects reflecting a genuine sustained control strategy. In that sense, if a PC effect reflects the activation of a control set that can be sustained over a substantial time period, then it ought to be possible to measure the generalization of such a PC effect beyond the block of trials in which PC is manipulated, in a subsequent block of trials on which the PC is kept neutral. Thus, we investigated whether this across block transfer effect occurs, and for how long it lasts.

Our second goal is to test further whether PC and CA effects can be dissociated in a more qualitative manner, thus reflecting different mechanisms at their basis. Concretely, we study whether the across block transfer of the PC effect will be general, generalizing to both conflict types, in contrast to the CA effect, which will be conflict-specific. Furthermore, given that there is no actual manipulation of the proportion congruency in the transfer block, any observed PC effect (as manipulated in the previous block) could not be explained on the basis of different proportion of transitions of cC, cI, iI and iC for the high vs. low proportion of congruent trial conditions, or to differences in the ratio of complete repetitions/alternation vs. partial repetitions associated to the different types of transitions.

To achieve these two goals, we introduced a modification of Funes et al's paradigm (2010b). In particular, the experiment had three phases; pre-training, training and post-training. In the critical training phase, only Simon conflict type trials were presented (i.e., only trials on the horizontal axis, left or right of fixation). Proportion congruent was manipulated for these trials, with half of the participants receiving mostly congruent and the other half receiving mostly incongruent trials. In the pre-training and the post-training phases, both Simon and Stroop conflict types (i.e., trials on both the horizontal and vertical axes) were presented, but without a manipulation of proportion congruent. With this procedure we were able to break our goals into contrastable questions: 1) Will there be a sustained component to the PC effect that transfers from the training to the post-training phase?; 2) If such a transfer occurs, how long will it persist into the post-training phase?; and 3) Will the PC effect that transfers to the post-training phase be specific to the conflict type manipulated in the training phase or will it generalize to both conflict types?

Section snippets

Participants

Thirty-six undergraduate psychology students (4 males) from the University of Granada participated in the experiment. Their ages ranged from 18 to 29 years (with a mean age of 21.72 years). Four of the participants were left handed. All of them had normal or corrected to normal color vision and were naive as to the purpose of the experiment. They all participated voluntarily and received credits for their Psychology courses.

Apparatus and stimuli

Participants were tested on a Pentium computer running E-prime software (

Results

Several analyses were conducted on mean reaction times (RTs) and error rates. For the mean RTs, error trials and trials with RT either shorter or longer than 2.5 standard deviations from the mean were excluded, which constituted 6.0% and 2.8% of the overall trials, respectively. Furthermore, trials following an error and the first trial of each block were also excluded, which eliminated a further 6.2% of the trials. For the analysis of error rates only the first trial of each block and trials

Discussion

The current experiment is to our knowledge the first demonstration that PC effects can be observed as transferring from one block where proportion congruency is manipulated to a following block where proportion congruency is actually not manipulated. Therefore the main conclusion of the present study is that PC, at least under the conditions used here, can be the reflection of a genuine sustained mode of cognitive control. Consequently, we provide further evidence that PC and CA effects can be

Conclusion

The broad implication of the present research is that cognitive control is expressed in performance by a set of distinct mechanisms. This conclusion follows from the fact that more than a single control mechanism is needed to explain both conflict adaptation and proportion congruency effects. It is a question of future research to investigate which mechanism is particularly involved in controlling the way we adapt to the different demands of control. In this endeavour we believe that apart from

Acknowledgements

This study was supported by a research position (FPU grant; AP2008-04006) to Maryem Torres-Quesada, and research grants funded by the Spanish Ministerio de Ciencia y Tecnología (PSI2008-04223PSIC, PSI2011-22416, and CONSOLIDER-INGENIO2010 CS).

References (45)

  • G.C. Burgess et al.

    Neural mechanisms of interference control in working memory: effects of interference expectancy and fluid intelligent

    PLoS ONE

    (2010)
  • E. Cañadas et al.

    Social Categories as a Context for the Allocation of Attentional Control

    Journal of Experimental Psychology General

    (2012)
  • C.S. Carter et al.

    Parsing executive processes: Strategic vs. evaluative functions of the anterior cingulated cortex

    Proceedings of the National Academy of Sciences of the United States of America

    (2000)
  • J.D. Cohen et al.

    On the control of automatic processes: A parallel distributed processing model of the Stroop effect

    Psychological Review

    (1990)
  • A. Correa et al.

    The two sides of temporal orienting: Facilitating perceptual selection, disrupting response selection

    Experimental Psychology

    (2010)
  • M.J.C. Crump et al.

    The context-specific proportion of congruent Stroop effect: Location as contextual cue

    Psychonomic Bulletin & Review

    (2006)
  • E.J. Davelaar et al.

    Sequential dependencies in the Eriksen fnaker task: A direct comparison of two competing accounts

    Psychonomic Bulletin & Review

    (2009)
  • T. Egner

    Right ventrolateral prefrontal cortex mediates individual differences in conflict-driven cognitive control

    Journal of Cognitive Neuroscience

    (2011)
  • D. Fernández-Duque et al.

    Cognitive control: Dynamic, sustained, and voluntary influences

    Journal of Experimental Psychology. Human Perception and Performance

    (2008)
  • A.L. Freitas et al.

    Contextual adjustments in cognitive control across tasks

    Psychological Science

    (2007)
  • M.J. Funes et al.

    Analyzing the generality of conflict adaptation effects

    Journal of Experimental Psychology. Human Perception and Performance

    (2010)
  • M.J. Funes et al.

    Generalization of the proportion congruency effect across conflict types. How general is it?

  • Cited by (67)

    • A process-specific approach in the study of normal aging deficits in cognitive control: What deteriorates with age?

      2022, Acta Psychologica
      Citation Excerpt :

      Therefore, the overall reduction in congruence effects might simply be the sum of all the SC effects that have taken place within the high conflict context. However, recent studies have dissociated PC and SC effects by showing that, while SC effects are typically specific to conflict type, PC effects can generalize across conflict type (Funes, Lupiáñez, & Humphreys, 2010; Torres-Quesada et al., 2013; see Aschenbrenner & Balota, 2019). More importantly, a previous study in our laboratory showed PC effects in the absence of SC effects (Torres-Quesada, Lupiáñez, Milliken, & Funes, 2014).

    • The role of dopamine in action control: Insights from medication effects in Parkinson's disease

      2021, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      The differential impact of medication status between the studies by Duthoo et al. (2014) and Ruitenberg et al. (2019) may pertain to the type of conflict adaptation that was examined in each study. Whereas the CSE as studied by Duthoo et al. (2014) is thought to reflect transient conflict adaptation involving trial-by-trial fluctuations in attentional demand, the proportion congruency effect as studied by Ruitenberg et al. (2019) is thought to reflect more sustained conflict adaptation in which attentional settings were optimized based on the overall congruency context of an entire block (e.g., Funes et al., 2010; Torres-Quesada et al., 2013). The findings related to the former type of conflict adaptation (as in Duthoo et al., 2014) seem to be in line with those for error processing (cf. Seer et al., 2017), suggesting that the use of dopaminergic medication may lead to a reduced ability to efficiently process trial-by-trial fluctuations in errors or conflicting information.

    • The role of task similarity for ego depletion: A registered report

      2021, Journal of Experimental Social Psychology
    • Measuring Adaptive Control in Conflict Tasks

      2019, Trends in Cognitive Sciences
    View all citing articles on Scopus
    View full text