Elsevier

Consciousness and Cognition

Volume 31, January 2015, Pages 35-45
Consciousness and Cognition

Unconscious context-specific proportion congruency effect in a stroop-like task

https://doi.org/10.1016/j.concog.2014.09.016Get rights and content

Highlights

  • We investigate the masked Stroop effect.

  • We associated a congruency percentage to each mask.

  • Behavioral and electrophysiological measures (EEG) were used.

  • We found the classic context specific proportion of congruency effect (CSPCE).

  • The amplitude of a frontal N2 and a parietal P3 were influenced by the CSPCE.

Abstract

Cognitive control is a central topic of interest in psychology and cognitive neuroscience and has traditionally been associated with consciousness. However, recent research suggests that cognitive control may be unconscious in character. The main purpose of our study was to further explore this area of research focusing on the possibly unconscious nature of the conflict adaptation effect, specifically the context-specific proportion congruency effect (CSPCE), by using a masked Stroop-like task where the proportion of congruency was associated to various masks. We used electrophysiological measures to analyze the neural correlates of the CSPCE. Results showed evidence of an unconscious CSPCE in reaction times (RTs) and the N2 and P3 components. In addition, the P2 component evoked by both target and masks indicated that the proportion of congruency was processed earlier than the congruency between the color word and the ink color of the target. Taken together, our results provided evidence pointing to an unconscious CSPCE.

Introduction

In a Stroop task, participants have to name the ink color of color words (Stroop, 1935, experiment 2). In a more recent version of the task, two different conditions are usually presented: congruent and incongruent. In the congruent condition, the ink color matches the color name of the word (e.g., RED in red ink). In the incongruent condition, by contrast, the ink color is different from the color name of the word (e.g., BLUE in red ink). To perform the task in the incongruent condition, participants must avoid the automatic process of reading the word and name the color of the ink in which the word is printed by using control strategies. In the congruent condition, the ink color matches the color name of the word, so avoidance of the automatic reading process is not necessary. Consequently, reaction times (RTs) are longer in the incongruent condition than in the congruent one. This additional time is known as Stroop interference or Stroop effect.

Stroop interference is affected by the relative proportion of congruent and incongruent trials. Behavioral studies have shown a reduction in RTs in incongruent trials presented in a context of low congruency (Logan & Zbrodoff, 1979). In other words, when the proportion of incongruent trials is higher than that of congruent trials, the Stroop effect is smaller than in situations in which the proportion of congruent trials is predominant. It seems as if frequent experience with conflicting stimuli or response features facilitates the resolution of interference. In general terms, this facilitation is known as the conflict adaptation effect (Botvinick, Braver, Barch, Carter, & Cohen, 2001).

Over the years, this improvement in control has been attributed to different factors. The main ones are the overall proportion of congruency at the list level (proportion of congruency effect, PCE) and a more specific and online processing of the information either at the item level (item-specific proportion congruency effect, ISPCE) or at the context level (context-specific proportion congruency effect, CSPCE). Proponents of the PCE (Logan, Zbrodoff, & Williamson, 1984) argue that in a Stroop task subjects become aware of the contingency between the color and the word and adapt their strategies to attend to the word when the proportion of congruent trials is higher than that of incongruent ones and ignore the word when the proportion of congruent and incongruent trials is reversed. Consequently, the Stroop effect is larger when the proportion of congruent trials is predominant because the color word is likely to be attended to in incongruent trials. By contrast, when the proportion of congruent trials is lower than that of incongruent trials, the Stroop effect is smaller than in the previous case because the color word is mostly ignored in incongruent trials. According to this approach, the proportion of congruent items influences performance at a list-wide level by switching participants’ attention to one of the two dimensions depending on which one is relevant for the task.

At computational and neural levels of analysis, Botvinick et al., 2001, Botvinick et al., 2004, McDonald et al., 2000 implemented these ideas in what they termed the conflict-monitoring hypothesis. According to these authors, the Stroop task is just a particular case of conflict resolved by the interplay of a set of neural structures. The dorsal anterior cingulate cortex (dACC) acts as a detection mechanism that responds to the occurrence of a conflict situation (i.e., an incongruent trial in a Stroop task). The conflict signal triggers strategic adjustments in the dorsolateral prefrontal cortex (DLPFC) and other posterior and subcortical structures that serve to prevent conflict in subsequent performance. Accordingly, in a list where incongruent trials outnumber congruent ones, strategic adjustment mechanisms will be highly operative, thus minimizing the influence of the color word on performance and reducing the Stroop effect. For the purposes of the present study, it should be underlined that PCE proponents understand control as a strategic and conscious process.

Proponents of the ISPCE contend that the proportion congruency effect acts at the item level. Jacoby, Lindsay, and Hessels (2003) chose six colors and their corresponding color words and divided them into two sets of equal size. To produce mostly congruent (MC) items, words in one set were presented in their congruent colors in eighty percent of trials and in another color from that set in the remaining twenty percent of trials. These rates were reversed in words from the other color set to produce mostly incongruent (MI) items. The proportion of congruency at the list level was fifty percent. This manipulation of proportions prevents subjects from developing a general list-level strategy because the proportion of congruent and incongruent items is the same. However, Jacoby et al. found that MC items showed a larger Stroop effect than MI items. From a computational approach, Blais, Robidoux, Risko, and Besner (2007) showed that Botvinick’s conflict-monitoring model could not explain the ISPCE and proposed a modified model according to which the DLPFC exerts control at the specific item level rather than at the general list level. Results similar to those of Jacoby et al. were obtained by Crump, Gong, and Milliken (2006) in a situation in which the likelihood of congruency was associated to a task-irrelevant location context rather than to a particular color word. According to these authors, the ISPCE belongs to a larger class of effects known as context-specific proportion congruency effect (CSPCE) that is driven by the relationship between the context and the likelihood of congruency. In recent years, the CSPCE has been reported in a range of Stroop-like effects (Bugg et al., 2008, Crump et al., 2008; for a review, see Bugg & Crump, 2012) and in different types of contexts, including social categories (Cañadas, Rodríguez-Bailón, Milliken, & Lupiáñez, 2012). An interesting characteristic of the CSPCE is that the type of control involved cannot be assumed to be a strategic, deliberate and conscious process. Instead, this type of control seems to be automatic (Jacoby et al., 2003) and unconscious. In fact, the effect has been shown to be independent of participants’ awareness of the proportion of congruency manipulation (Crump et al., 2008).

The main purpose of the present research was to further explore the likely unconscious character of the conflict adaptation effect in a Stroop task by asking subjects to name the color of a color patch that was preceded by a masked color word. We used two masks (one previous to the color word and one following the color word) and calibrated the duration of the color word individually for each participant until it was unnoticed. Under similar masking conditions to ours, this variable has been used in some studies in which the proportion of congruency was manipulated in an effort to find qualitative differences between conscious and unconscious processes (Daza et al., 2002, Merikle and Joordens, 1997). In those experiments, the usual Stroop effect was observed under masked color word conditions. Yet, the effect was reversed when the color word was easily seen, as participants exhibited faster RTs in incongruent trials than in congruent ones. The PCE effect has also been explored under masked conditions in other experiments. Klapp (2007, Experiment 3) used a spatial Stroop-like task involving successively presented arrows pointing in either the same direction (i.e., compatible condition) or the opposite direction (i.e., incompatible condition). The first arrow was presented for 32 ms and was immediately followed by the mask and the second arrow, so that the first arrow was unnoticed to the subjects. Results revealed a significant interaction. Specifically, the difference in RTs between compatible and incompatible conditions was greater when the arrows pointed in the same direction than when they pointed in opposite directions. However, in Klapp’s experiment (2007), participants received feedback for incorrect responses, which makes it difficult to interpret his results. A possible explanation could be that subjects adapted their responses to conscious error rates rather than to the unconscious frequency of congruent and incongruent trials. Heinemann, Kunde, and Kiesel (2009) explored the CSPCE under masked conditions using a task in which subjects had to categorize target numbers as being larger or smaller than five. The target was preceded by a masked number (i.e., the prime) that could be congruent or incongruent with the target. At the beginning of each trial, a colored rectangle was presented as background simultaneously with the fixation cross. The color of the rectangle was associated with a particular congruency context that could be either 80% or 20%. Results showed that the difference in RTs between congruent and incongruent trials was significantly higher in the 80% than in the 20% congruent trial condition but only when subjects were able to see the prime; no differences were observed when they could not see it. (Heinemann et al., 2009). The authors concluded that the CSPCE requires conscious representation of the conflicting information, namely prime, target and context (Kunde, Reuss, & Kiessel, 2012).

It seems that the unconscious conflict adaptation effect has not been clearly established. First, studies of the PCE under unconscious conflict conditions are difficult to interpret due to the presence of confounding factors. Second, the CSPCE seems to occur only when the conflicting information is consciously perceived. We considered that it would be interesting to explore the conflict adaptation effect by associating the proportion of congruent and incongruent trials to different masks and using electrophysiological recordings at the same time to study the brain response to both conflict adaptation and the proportion of congruency.

We partly followed a procedure developed by Blais, Tudela, and Bunge (2007) and used two different types of masks. One type of mask was associated with mostly congruent trials and a second type of mask was associated with mostly incongruent trials. Each mask appeared an equal number of times within each block. Yet, one mask was followed by 33% congruent and 66% incongruent color words the other mask was followed by the opposite proportion of congruent and incongruent trials. Thus, the total number of congruent and incongruent trials was kept the same in each block.

The use of brain activity-related measures to study the unconscious conflict adaptation effect has been highly recommended (Desender & Van den Bussche, 2012). Such measures can provide additional information about the brain response not only to targets and masked primes but also to the processing of the congruency relation between prime and target and of the proportion of congruency. In an influential study, Dehaene, Artiges, et al., 2003, Dehaene et al., 2003 used functional magnetic resonance imaging (fMRI) and a Stroop-like task in which subjects were asked to decide whether target numbers were larger or smaller than five. Targets were always preceded by another number (i.e., the prime) that could be congruent with the target (i.e., both numbers greater or smaller than five) or incongruent (i.e., one number larger and the other one smaller than five). In addition, the numbers could be either visible or made invisible by masking. They found that normal subjects showed a clear activation of the anterior cingulate cortex (ACC) when the prime could be consciously processed; by contrast, no ACC activation was observed when the prime was masked. Interestingly, they also observed impaired conscious priming but normal subliminal priming in patients with schizophrenia who exhibited a hypoactivation of the ACC. These authors concluded that subliminal conflicts are resolved without ACC involvement. It seems logical to conclude that if the control structures of the brain do not play a role in subliminal priming, they are not likely to participate in conflict adaptation effects. However, in a recent experiment using fMRI, Blais and Bunge (2010) found that many of the regions involved in cognitive control in the Stroop task, including the ACC and DLPFC, were operative at a local item level.

In the literature on electroencephalography (EEG), ACC activation has been associated with the N2 component, a negative deflection in the averaged event-related potential (ERP) wave with a fronto-central scalp distribution that peaks around 250–350 ms after stimulus presentation (Ridderinkhof, Ullsperger, Crone, & Nieuwenhuis, 2004). Consistent with the role of the ACC in the conflict-monitoring theory, the amplitude of this component has been found to be more negative in incompatible trials compared to compatible trials (Clayson and Larson, 2011, Wendt and Luna-Rodríguez, 2009). Moreover, the P3 component, a positive deflection in the averaged waveform with a central–parietal distribution that occurs 350–500 ms after stimulus presentation, has been associated with response conflict and has been found to be more positive in incongruent trials compared to congruent trials (Clayson and Larson, 2011, Frünholz et al., 2011). Research conducted by Clayson and Larson (2011) is of particular interest for the present study. In a flanker task, these authors studied the sequential effect of the previous trial on the N2 and P3 components as a function of its congruency (Gratton, Coles, & Donchin, 1992). This effect is considered to reflect conflict adaptation and seems to be closely related to the proportion congruency effect (Botvinick et al., 2001). Clayson and Larson’s results showed that in both components (i.e., N2 and P3), the difference in amplitude between an incongruent and a congruent trial was lower when the previous trial was incongruent than when it was congruent. As for the ISPCE effect, Shedden, Milliken, Walter, and Monteiro (2012) analyzed various ERPs in a recent study and failed to find a scalp signature directly related to the conflict adaptation effect both in a global–local task and in a standard Stroop task. However, these authors observed that subjects distinguished the proportion congruency category before the congruency of the specific stimulus as early as 100 ms post-stimulus onset in the global–local identification task and 150 ms in the Stroop task.

To the extent that unconscious conflict adaptation may also involve the activation of control structures in the brain, we expected to find a conflict adaptation effect in a late (i.e., N2 or P3) event-related component in our experiment. In a recent experiment, Jiang, van Gaal, Bailey, Chen and Zhang (2013) reported an unconscious block-wise proportion congruency effect (PCE) in N2 and P3 amplitudes using a meta-contrast masked priming task. However, to our knowledge, no unconscious ISPCE or CSPCE has been observed yet. On the other hand, if proportion congruency is processed earlier than the congruency of the specific stimulus, as reported by Shedden et al. (2012), we should expect a similar effect in our experiment since participants are usually unaware of the proportion of congruency manipulation even under regular non-masked conditions.

In summary, the goals of our current research were to explore the Stroop effect under unconscious conditions and to observe how the proportion of congruency modulates this effect when it is associated to a particular context. We also intended to study the time course of neural activation related to processing of the proportion of congruency, the congruency between the color word and the color of the target, and the interaction between item congruency and proportion of congruency that defines the CSPCE. We focused on analyzing the unconscious situation by using an extended set of color-word masked trials. However, we tried to maintain an expectation of the presence of a color word between masks by calibrating the duration of the color word individually for each participant until it was unnoticed and by including a trial block in which the color word could be easily seen at the end of each session. It has been recently reported (Martens, Ansorge, & Kiefer, 2011) that subliminal semantic priming can be modulated by attentional task set. In this experiment we tried to keep the Stroop task set constant while the processing of the color word remained unconscious.

Section snippets

Participants

Twenty-nine students of the University of Granada (21 females and 8 males) participated in a two-session experiment. Sessions were separated by a one-week interval. Participants reported having normal or corrected-to-normal vision and were not aware of the purpose of the experiment. Mean age was 23 years (SD = 4.26). Participants received course credit in exchange for their participation and signed a consent form approved by the local Ethics Committee.

Stimuli and apparatus

Three different types of stimuli were

Threshold setting

The average presentation time of the masked words at threshold value was 15.97 ms (SD = ± 5.53 ms) in the first session and 16.66 ms (SD = ± 8.03 ms) in the second session. A comparison between these two threshold values showed no significant differences between sessions, t = .586, p = .562.

RT analysis

Overall accuracy was very high (99%) and no difference in accuracy was observed between experimental conditions. Only correct responses were considered in the RT and ERPs analyses. RTs shorter than 300 ms or longer than

Discussion

The main purpose of the present research was to explore whether or not the proportion of congruency modulated the Stroop effect when the color words were masked. Both RT and ERP measures provided an affirmative answer to this question. RTs in congruent and incongruent trials significantly changed depending on the percentage of incongruent trials. When this percentage was low, congruent trials were responded to faster than incongruent trials; by contrast, when the percentage of incongruent

Conclusions

Our results provide evidence of the unconscious character of cognitive control when a masked Stroop-like paradigm is used. The early P2 component showed amplitude differences as a function of percentage of congruency, displaying larger amplitude for mostly incongruent than for mostly congruent trials. This seems to be evidence of a type of implicit learning across trials given that, when masks were analyzed, the P2 tended to be higher in the second session compared to the first one.

The

Acknowledgments

We thank María Ruz and our reviewers for their careful reading of our manuscript and their many insightful comments and suggestions.

This research was supported by the Spannish Ministerio de Educación y Cultura with a Grant (FPU-AP2007-00313) to the first author and a research Grant P09-HUM-5422 from the Andalusian Autonomous Government to Pío Tudela.

References (48)

  • P.S. Wong et al.

    Event-related brain correlates of associative learning without awareness

    International Journal of Psychophysiology

    (2004)
  • C. Blais et al.

    Behavioral and neural evidence for item-specific performance monitoring

    Journal of Cognitive Neuroscience

    (2010)
  • C. Blais et al.

    Item-specific adaptation and the conflict-monitoring hypothesis: A computational model

    Psychological Review

    (2007)
  • C. Blais et al.

    Proportion congruency and the neural correlates of consciousness

    (2007)
  • M.M. Botvinick et al.

    Conflict monitoring and cognitive control

    Psychological Review

    (2001)
  • Bugg, J. M. & Crump, M. J. C. (2012). In support of a distinction between voluntary and stimulus-driven control: A...
  • J.M. Bugg et al.

    Converging evidence for control of colour-word Stroop interference at the item level

    Journal of Experimental Psychology: Human Perception and Performance

    (2013)
  • J.M. Bugg et al.

    Multiple levels of control in the Stroop task

    Memory & Cognition

    (2008)
  • E. Cañadas et al.

    Social categories as a context for the allocation of attentional control

    Journal of Experimental Psychology

    (2012)
  • M.J.C. Crump et al.

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

    Psychonomic Bulletin & Review

    (2006)
  • M.T. Daza et al.

    Perception without awareness: Further evidence from a Stroop priming task

    Perception & Psychophysics

    (2002)
  • S. Dehaene et al.

    Conscious and subliminal conflicts in normal subjects and patients with schizophrenia: The role of the anterior cingulate

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

    (2003)
  • S. Dehaene et al.

    A neuronal network model linking subjective reports and objective physiological data during conscious perception

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

    (2003)
  • K. Desender et al.

    Is consciousness necessary for conflict adaptation? A state of the art

    Frontiers in Human Neuroscience

    (2012)
  • Cited by (12)

    • Cortical and subcortical contributions to context-control learning

      2019, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      A corresponding psychological distinction between proactive versus reactive forms of control has been formalized in Braver’s dual mechanisms of control (DMC) framework (Braver, 2012; Braver et al., 2009). The proposal that the dACC is involved in gauging the need for up-regulating control has also been supported in studies of control learning effects in the context of longer-term temporal variations in cognitive demand, that is, in fMRI studies of the PC and ISPC effect (e.g., Blais and Bunge, 2010; Carter et al., 2000; Grandjean et al., 2012; Wilk et al., 2012) as well as in electrophysiological studies of the CSPC and ISPC effects (e.g., Blais et al., 2016; Panadero et al., 2015; Whitehead et al., 2017). For instance, Carter et al. (Carter et al., 2000) employed a Stroop color-naming task and manipulated the proportion of congruent trials between blocks (20% vs. 80%).

    • Unconscious influence over executive control: Absence of conflict detection and adaptation

      2018, Consciousness and Cognition
      Citation Excerpt :

      However, evidence regarding unconscious modulation of the executive control system remains unclear (Ansorge et al., 2014; Desender & Van den Bussche, 2012). Specifically, concerning unconscious conflict detection and the subsequent adaptation, researchers have tackled different approaches and interpretations and provided an overall positive picture regarding these effects (Diede & Bugg, 2017; Hasegawa & Takahashi, 2014; Jiang, Zhang, & van Gaal, 2015; Panadero, Castellanos, & Tudela, 2015; Van Opstal, Calderon, Gevers, & Verguts, 2011; Wang, Xiang, & Li, 2013). Nonetheless, some studies still raise doubts regarding such possibility, by showing either partial or no adaptation to conflict (e.g., Ansorge, Fuchs, Khalid, & Kunde, 2011; Atas, Desender, Gevers, & Cleeremans, 2016; Schouppe, de Ferrerre, Van Opstal, Braem, & Notebaert, 2014), or even failing to observe conflict detection (e.g., Bruchmann, Herper, Konrad, Pantev, & Huster, 2010; Wu et al., 2015).

    • Context-specific attentional sampling: Intentional control as a pre-requisite for contextual control

      2016, Consciousness and Cognition
      Citation Excerpt :

      Awareness of conflict between the relevant and irrelevant dimensions has also been assessed as a pre-requisite for CSPC effects. For example, CSPC effects can be produced in masked-prime procedures where subjects are not aware of the source of response conflict (Heinemann et al., 2009; Panadero et al., 2015; Reuss et al., 2014; but see, Schouppe, de Ferrerre, Van Opstal, Braem, & Notebaert, 2014). Reuss et al. (2014) embedded the contextual cue within the masked prime and found CSPC effects even when the prime was below a perceptible threshold.

    • The semantic origin of unconscious priming: Behavioral and event-related potential evidence during category congruency priming from strongly and weakly related masked words

      2016, Cognition
      Citation Excerpt :

      The processing of the response congruency relation between prime and target has been previously associated with the N200 ERP component, a negative deflection with a fronto-central distribution that peaks around 250–350 ms after stimulus presentation (e.g., Ridderinkhof, Ullsperger, Crone, & Nieuwenhuis, 2004). The amplitude of the N200 component has been found to be more negative in incongruent (or incompatible) trials compared to congruent (or compatible) trials (e.g., Clayson & Larson, 2011; Wendt & Luna-Rodríguez, 2009; see also Panadero, Castellanos, & Tudela, 2015). Furthermore, unconscious response priming has been shown to modulate occipito-parietal ERPs in the N200 time range, consistent with the notion of a rapid activation of visuo-motor representation within the dorsal visual pathway (e.g., Jaśkowski, Skalska, & Verleger, 2003; Martens, Ansorge, & Kiefer, 2011; Zovko & Kiefer, 2013).

    View all citing articles on Scopus
    View full text