Edited by: Mattie Tops, VU University Amsterdam, Netherlands
Reviewed by: Mattie Tops, VU University Amsterdam, Netherlands; Kristien Aarts, INSERM-IMPACT, France
*Correspondence: Bernhard Hommel, Cognitive Psychology Unit, Institute for Psychological Research, Leiden University, Wassenaarseweg 52, Leiden, 2333 AK, Netherlands e-mail:
This article was submitted to Cognition, a section of the journal Frontiers in Psychology.
†Shared first author.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Stimulus-induced response conflict (e.g., in Simon or Stroop tasks) is often reduced after conflict trials—the Gratton effect. It is generally assumed that this effect is due to a strengthening of the representation of the current intention or goal, which in turn increases the degree of stimulus and/or response control. Recent evidence suggests that the motivational signal driving the Gratton effect might be affective in nature. If so, individual differences in either the strength of affective signals and/or the ability to interpret such signals might explain individual differences in cognitive-control adjustments as reflected in the Gratton effect. We tested this hypothesis by relating individual sizes of the Gratton effect in a Simon task to scores on the affective and the cognitive dimension of the Bermond/Vorst Alexithymia Questionnaire (BVAQ)—which we assumed to assess individual differences in affective-signal strength and ability to interpret affective signals, respectively. Results show that the cognitive, but not the affective dimension predicted control adjustment, while the accuracy of heartbeat detection was only (and only weakly) related to online control. This suggests that the motivation to fine-tune one's cognitive-control operations is mediated by, and may depend on one's ability to interpret one's own affective signals.
Traditional views on the role of motivation in action control have focused on the process of decision-making, that is, on the selection of goals that an agent intends to pursue (e.g., Kahneman,
A prime example is the so-called “conflict-adaptation effect” or “Gratton effect” (the theoretically more neutral term, after Gratton et al.,
From a motivational point of view, the Gratton effect raises two questions. First, why do people need to adapt the degree to which they are controlling their actions at all? If they would keep control at some optimum, adaptation should not be necessary. The fact that they apparently do not suggests that they are either unable to do that, perhaps because cognitive control relies on limited and quickly depleting resources (Baumeister,
The second important motivational question is why people increase the amount of cognitive resources they invest after conflict trials. Conventional motivational accounts would suggest that control adaptation is driven by action errors, because these would imply a mismatch between the selective goal and the actual outcome, but not by mere conflict. This would allow for learning through errors but not for the prevention of errors through the anticipatory adjustment of action-control parameters that the Gratton effect is taken to reflect. According to Botvinick et al. (
Interestingly, all three approaches imply that affective states might be involved in driving control adjustments: Just like real or anticipated negative events induce negative affect, conflict between response tendencies has been assumed to have the exact same effect (Festinger,
If control adjustments would really be driven by negative affect, one would expect that interindividual differences in experiencing or processing affect predict differences in control adjustments—and this was indeed the hypothesis we tested in the present study. We considered two ways in which affect processing might differ.
For one, people may differ with respect to the strength of the signal that affect-related systems send to systems sensing the need for control, such as the ACC. Indeed, Critchley et al. (
For another, people may differ with respect to the degree that they are able to interpret signals sent by affect-related systems. Indeed, several authors have argued that affectively relevant signals might often be difficult to interpret and people may differ with respect to their ability to attribute them to the actual cause. For instance, Schachter (
In the present study, we used the Bermond/Vorst Alexithymia Questionnaire (BVAQ; Vorst and Bermond,
To investigate the adaptivity of cognitive control, we used a standard conflict-inducing task, the Simon task (for reviews, see Lu and Proctor,
Sixty students (8 males; mean age 19.7 years) of the Leiden University participated in the experiment for partial fulfillment of course credit (6.50 Euro). Written informed consent was obtained from all subjects; and the protocol was approved by the local ethical committee (Leiden University, Institute for Psychological Research).
Participants estimated their heartbeat and completed the questionnaire in a counterbalance order. Afterwards, they performed the Simon task.
Heartbeat detection was measured using the Mental Tracking Method (e.g., Tsakiris et al.,
A paper version of the Dutch BVAQ was used (Vorst and Bermond,
The cognitive dimension comprises of three subscales: verbalizing emotions, identifying emotions, and analyzing emotions. The “verbalizing emotions” subscale concerns the degree to which one is able or inclined to describe or communicate about one's emotional reactions (e.g., “I find it difficult to verbally express my feelings”). The “identifying emotions” subscale concerns the degree to which one is able to define one's arousal states (e.g., “When I am distressed, I know whether I am afraid or sad or angry”). The “analyzing emotions” subscale concerns the degree to which one seeks out explanations of one's own emotional reactions (e.g., “I hardly ever go into my emotions”). Subscales have 8 questions each, scored from 1–5. Possible scores thus range from 24–120 (8–42, clearly non-alexithymic; 43–61, modal; and 62–120, alexithymic).
The affective dimension comprises of two subscales: emotionalizing and fantasizing. The “emotionalizing subscale” concerns the degree to which someone is emotionally aroused by emotion-inducing events (e.g., “When something totally unexpected happens, I remain calm and unmoved”). The “fantasizing” subscale concerns the degree to which someone is inclined to fantasize, imagine, daydream, etc. (e.g., “Before I fall asleep, I make up all kinds of events, encounters, and conversations”). Subscales have eight questions each, scored from 1–5. Possible scores therefore range from 16–80 (16–28, clearly non-alexithymic; 29–41, modal; 42–80, alexithymic).
The Simon task was performed on a computer running Windows™, attached to a 17″ color monitor. Viewing distance was about 60 cm. A continuously centrally-displayed small (0.5 cm) dark gray square served as fixation point. Stimuli were green and blue circles (1.5 cm in diameter), presented to the left or right of fixation. The color and location of the stimuli varied randomly and equiprobably. Circles stayed on the screen until 1500 ms had passed or a response was given. Intervals between stimuli varied randomly between 1250–1750 ms, in steps of 100 ms. Feedback on PEs and RTs was provided at the end of each block. Responses were made by pressing the “z” or “?” buttons of the QWERTY computer keyboard with the left or right index finger, respectively. Participants were instructed to react as fast and as accurate as possible to the stimulus-color, but not location.
As a manipulation check (to demonstrate reliable Simon and Gratton effects), mean RTs and PEs from the Simon task were submitted to separate repeated-measures ANOVAs, with compatibility in present trial (compatible vs. incompatible) and compatibility in previous trial (compatible vs. incompatible) as factors.
For the correlation analyses, seven scores were calculated: (1) the size of the Simon effect in RT (RT incompatible in the present trial minus RT compatible in the present trial); (2) the size of the Simon effect in PE (incompatible minus compatible); (3) the size of the Gratton effect in RT (Ci[incompatible trial following compatible trial] – Cc[compatible trial following compatible trial]) – (Ii[incompatible trial following incompatible trial] – Ic[compatible trial following incompatible trial]); (4) the size of the Gratton effect in PE ((Ci – Cc) – (Ii– Ic)); (5) the BVAQ score for the affective dimension (emotionalizing subscale + fantasizing subscale, a higher score indicating more alexithymia); (6) the BVAQ score for the cognitive dimension (verbalizing subscale + identifying emotions subscale + analyzing subscale, a higher score indicating more alexithymia); and (7) an interoceptive awareness score, calculated from the four different heartbeat detection intervals [¼ Σ(1 – (|recorded heartbeats–counted heartbeats|)/recorded heartbeats)]; this score can vary between 0 and 1, with higher scores indicating better heartbeat detection.
Judgment accuracy in heartbeat detection varied between 0.46 and 0.98 (Mean = 0.82,
The BVAQ scores varied between 25 and 78 (Mean = 51,
As the Simon effect is sensitive to the overall RT level (Hommel,
As the correlation analyses had the goal to determine whether alexithymia measures can predict
The BVAQ scores for the 55 participants included in the correlation analyses varied between 25 and 78 (Mean = 50,
Table
Cognitive dimension | 0.207 | −0.112 | −0.370 |
−0.205 | −0.081 |
Verbalizing | 0.267 |
−0.117 | −0.366 |
−0.238 | −0.169 |
Identifying emotions | 0.128 | −0.046 | −0.186 | −0.132 | 0.096 |
Analyzing | 0.012 | −0.091 | −0.282 |
−0.050 | −0.114 |
Affective dimension | 0.048 | −0.022 | −0.061 | −0.190 | −0.051 |
Emotionalizing | 0.128 | −0.129 | −0.078 | −0.124 | −0.175 |
Fantasizing | −0.021 | 0.054 | −0.026 | −0.156 | 0.048 |
Interestingly, we observed only little evidence of a relationship between the Simon effect and alexithymia scores: apart from a just-significant positive correlation between the size of the Simon effect in RTs and the verbalizing subscale of alexithymia (
Also of interest, the heartbeat-detection score did not correlate with any of the alexithymia dimensions (or any subscale,
Finally, we obtained positive correlations between the Simon effects in RTs and PEs (
The aim of our study was to better characterize the motivational signal driving cognitive-control processes to adapt to task demands. Given the evidence that this signal is related to negative affect, we reasoned that people may differ with respect to either the strength of the signal provided by affect-related neural systems (such as the insular cortex; Critchley et al.,
The outcome is clear in showing that individual differences in the Gratton effect are associated with the cognitive, but not the affective BVAQ dimension. This suggests that the ability or preference to analyze and properly interpret affect signals is necessary for engaging in adaptive control. It is important to point out that the BVAQ assesses the subjective experience of difficulties in analyzing and interpreting one's own emotions but does not indicate whether the individual interpretations are correct. Hence, our findings do not rule out the possibility that people engage in the trial-to-trial fine-tuning of cognitive control only because they think they are good in interpreting their feelings while they actually are not. If so, interpretational optimism might be particularly motivating for engaging in the optimization of control, while the actual availability of interpretational skills might not matter so much. And yet, given that life provides numerous occasions to provide objective feedback about the validity of one's interpretation of one's own emotions (Wittgenstein,
One possibility is that the cognitive, perhaps even conscious, interpretation of affective signals is a necessary processing step intervening between the generation or emergence of conflict and operations aiming at reducing or eliminating such conflict in the future. Individuals who can interpret such signals more efficiently, for instance by being better able to discriminate between positive and negative signals, would thus be more effective in adapting cognitive-control settings—as assessed by the Gratton effect. Note that this would not necessarily mean that they can avoid conflict in the first place, even though the positive correlation between verbalizing and the Simon effect in PEs suggests that they might to some degree.
It might be objected that, even though many studies in this area suffer from problematic flaws that make the interpretation of their outcomes difficult (Desender and Van den Bussche,
It is interesting to note that the heartbeat-detection score did not correlate with any of the alexithymia scores. Previous findings have shown that heartbeat detection accuracy is correlated with the same measures as at least some alexithymia subscales (e.g., activity in the insular cortex; see Critchley et al.,
In any case, our findings suggest that the individual motivation to engage in the fine-tuning of cognitive-control processes is related to the perceived and/or objective ability to interpret one's own affective states. This provides further evidence for an important role of emotion in the control of cognition and action, and a deeper insight into the way cognitive control operations are motivated. However, it is important to note that our results can be generalized only to a population showing regular (i.e., positive) Simon and Gratton effects, and cannot be extended to the individuals exhibiting uncommon negative effects.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The Supplementary Material for this article can be found online at: