Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Action information contributes to metacognitive decision-making


Metacognitive abilities allow us to adjust ongoing behavior and modify future decisions in the absence of external feedback. Although metacognition is critical in many daily life settings, it remains unclear what information is actually being monitored and what kind of information is being used for metacognitive decisions. In the present study, we investigated whether response information connected to perceptual events contribute to metacognitive decision-making. Therefore, we recorded EEG signals during a perceptual color discrimination task while participants were asked to provide an estimate about the quality of their decision on each trial. Critically, the moment participants provided their confidence judgments varied across conditions, thereby changing the amount of action information (e.g., response competition or response fluency) available for metacognitive decisions. Results from three experiments demonstrate that metacognitive performance improved when first-order action information was available at the moment metacognitive decisions about the perceptual task had to be provided. This behavioral effect was accompanied by enhanced functional connectivity (beta phase synchrony) between motor areas and prefrontal regions, exclusively observed during metacognitive decision-making. Our findings demonstrate that action information contributes to metacognitive decision-making, thereby painting a picture of metacognition as a process that integrates sensory evidence and information about our interactions with the world.


The ability to monitor and evaluate the quality of our decision-making is crucial for adept behavior. For instance, when driving a car for a long time it is important to have a reliable estimate about the adequacy of one’s driving performance to avoid unsafe situations. However, not much is known how our brain constructs such an estimate, or what exactly is being monitored and evaluated. In lab settings, perceptual or memory tasks have been frequently used to probe the mechanisms that underpin metacognitive performance1,2,3. In such studies, first-order task performance generally correlates with second-order (metacognitive) decisions, leading to the intuitive assumption that metacognitive decisions are largely based on the same information that governs first-order decision-making4,5,6.

In recent years, however, dissociations between objective task performance and subjective ratings, and dissociations between sources of information supporting first- and second-order decisions have been observed7,8,9,10,11,12. Typically, metacognitive decisions are provided after first-order responses, thereby allowing certain sources of information to become available during second-order decision-making. Recent findings suggest that metacognition can be supported by ‘embodied’ processes, such as interoception or response information that become available for metacognitive decision-making after a first-order decision has been made7,13,14,15,16. For instance, manipulation of neural activity via transcranial magnetic stimulation over premotor cortex resulted in altered confidence judgments during a perceptual task8. Critically, stimulation of premotor areas reduced metacognitive capacity without changing visual discrimination performance. Further, it has been shown that the order of rating confidence (before or after the response) influenced metacognitive performance on an anagram problem-solving task17. From a computational perspective, Pasquali and colleagues explored neural network architectures aimed at capturing the complex relationships between first-order and second-order (metacognitive) performance in a range of different cognitive tasks and suggested that metacognitive judgments are rooted in learned redescriptions of first-order error information rather than in the relevant first-order information itself18. This is broadly consistent with Fleming and Daw’s perspective, in which they offered to unify the above observations in a single framework in which confidence operates as a second-order computation about one’s own performance19. In this framework, samples of sensory evidence that support first- and second-order decisions are coupled yet distinct. Interestingly, their second-order model of confidence computation incorporates knowledge about the reliability of actions towards perceptual events.

Here, in three experiments, we aimed to elucidate whether and in what way action information informs metacognitive judgments. We therefore constructed a color discrimination task in which we varied the amount of available action information (i.e., response strength and fluency of response execution) at the moment a metacognitive judgment had to be provided. Our design enabled us to contrast metacognitive decisions based on purely perceptual information (uninformed by action processes) with metacognitive decisions having access to both perceptual and motor action information. We recorded electroencephalographic signals to investigate whether functional connectivity between motor regions and prefrontal cortex could serve as a mechanism to convey relevant action information (e.g., response competition or response fluency) during metacognitive decision-making.

Previously, beta oscillations have been intimately linked to sensory and motor processing20. Recently, however, beta-band power (de)synchronization in motor regions has been shown to provide insight into the dynamics underlying perceptual decisions21 and response uncertainty22. Beta oscillations have repeatedly been shown to predict first-order decisions22,23,24, to support maintenance of persistent activity25,26,27 to mediate long-range communication, and to play an important role in the preservation and ‘awakening’ of endogenous information28. Here, we focused on beta phase synchrony between motor regions and prefrontal cortex9. Specifically, we expected both functional connectivity (beta phase synchrony) and metacognitive performance to increase when response information about first-order decisions would be accessible during metacognitive decision-making.



To determine whether action processes (i.e., response competition, ‘ease’ of action preparation29) contributed to the quality of metacognitive judgments, we varied the amount of first-order action information present at the moment metacognitive decisions had to be provided (see Fig. 1a). We constructed three conditions that differed in the moment participants had to provide their metacognitive judgment (see methods). In the first condition, participants provided verbal metacognitive judgments after the response cue and after the first-order response (ACT condition). In the second condition, metacognitive judgments were provided before the first-order response but after the presentation of the response cue (PRE_ACT condition). In the third condition, participants provided metacognitive judgments before presentation of the response cue and execution of the first-order response (PRE_CUE condition). We performed three repeated measures ANOVAs (the three conditions as levels) on first-order task performance (da), metacognitive sensitivity (meta da) and metacognitive efficiency (meta da - da), respectively (see methods). Metacognitive sensitivity quantifies (in units of da) how well a participant can discriminate correct from incorrect decisions on a first-order task. Metacognitive efficiency is the ability to discriminate between correct and incorrect decisions relative to different levels of first-order task performance. Because of the known influence of first-order task performance on metacognitive performance (meta da), metacognitive efficiency is a measure of metacognitive performance that is more independent from variability in first-order performance30.

Figure 1

(a) Task design experiment 1. Participants had to decide whether the majority of randomly moving dots were red or green by pressing a left or right key. The key that mapped onto a ‘red’ or ‘green’ answer was signaled by a response cue on each trial. Verbal confidence ratings were recorded either at the end of each trial (ACT), or directly preceding the first-order response (PRE_ACT), or directly following stimulus presentation (PRE_CUE). In this way, in each condition a different amount of first-order action information was available at the moment metacognitive decisions were provided. (b) Behavioral results. Participants’ metacognitive efficiency decreased when action information was not available, while first-order performance remained unaltered. Error bars represent between-subjects standard error of the mean.

We found a significant effect of condition, specifically for metacognitive efficiency (F(2, 28) = 4.04 p = 0.0029, η2 = 0.224). For both da (F(2, 28) = 0.631 p = 0.540, η2 = 0.043) and meta da (F(2, 28) = 1.882 p = 0.171, η2 = 0.118) no significant effects were observed. Next, we performed (one-tailed) t-tests to find out whether metacognitive efficiency decreased when response information was reduced. Results demonstrate that ACT and PRE_CUE significantly differed from each other (t(14) = 2.45, p = 0.014, d = 0.663, BF+0 = 4.75), while no significant differences were observed between ACT and PRE_ACT (t(14) = 1.65, p = 0.061, d = 0.426, BF+0 = 1.47) and PRE_ACT and PRE_CUE (t(14) = 1.45, p = 0.085, d = 0.374, BF+0 = 1.13), see Fig. 1b. These findings suggest that participants’ capacity to distinguish accurate from inaccurate decisions improved when first-order response information was fully available (the ACT condition) compared to when such information was entirely unavailable (the actual response and response preparation). We did not observe differences in the average confidence level between the conditions (all ts < 0.753, ps > 0.464).

We aimed to prevent the influence of prolonged evidence accumulation introduced by differences in time between stimulus offset and response as much as possible by introducing a blank of 1 second after stimulus presentation in all three conditions31. However, it could still be possible that a longer time window to reflect on the perceptual decision could nonetheless influence performance independently of action information. In such a scenario of prolonged evidence accumulation we would expect the da to be higher in the ACT condition compared to the PRE_CUE and PRE_ACT conditions since evidence had more time to accumulate. To assess whether our experimental design was successful in preventing effects due to prolonged evidence accumulation, we post-hoc tested differences between da scores in ACT and PRE_ACT, and in ACT and PRE_CUE respectively (see Fig. 1b). We did not observe any significant da differences (ACT vs. PRE_ACT: t(14)=0.56, p = 0.584, BF10 = 0.301; ACT vs. PRE_CUE: t(14)=1.00, p = 0.334, BF10 = 0.403). These findings indicate that the presented blank after stimulus offset most likely eliminated effects of prolonged evidence accumulation.

EEG results

In order to examine the neural mechanisms that support communication between motor areas and prefrontal regions during metacognitive decision-making, we assessed differences in interregional functional connectivity (beta phase synchrony) between the central frontal electrode Fz9(see methods) and motor channels C3 or C4 (depending on the hand that responded) in the 500 ms time window preceding participants’ metacognitive judgment. There was a significant effect of condition for changes in beta phase synchrony (Greenhouse-Geisser corrected: F(1.29,18.19) = 8.434, p = 0.006, η2 = 0.376). Because oscillatory activity in the alpha band has also been closely linked to action mechanisms32, we explored whether differences between conditions in alpha phase synchrony could be observed. No effects were found for changes in alpha phase synchrony between conditions (F(2,28) = 1.483, p = 0.244, η2 = 0.096); see Fig. 2. We found higher functional connectivity (beta phase synchrony) in ACT compared to PRE_ACT (t(14) = 3.89, p = 0.002, d = 1.004, BF10 = 25.437) and PRE_CUE (t(14) = 2.446, p = 0.028, d = 0.632, BF10 = 2.405). No differences were observed between PRE_ACT and PRE_CUE (t(14) = 1.20, p = 0.250, d = 0.310, BF10 = 0.482).

Figure 2

Functional connectivity. Functional connectivity (beta phase synchrony) between motor cortex and prefrontal cortex was higher in ACT where response information was available during metacognitive decision-making compared to PRE_ACT and PRE_CUE. No effects were observed for alpha phase synchrony. Shaded areas represent within-subjects standard error of the mean. Time zero refers to the onset of the metacognitive question (see Fig. 1).

Next, we investigated whether functional connectivity changes (beta phase synchrony) were accompanied by changes in beta power in the central frontal channel Fz. Beta power was higher in ACT compared to PRE_ACT (t(14) = 2.765, p = 0.015, d = 0.714, BF10 = 3.957), while no differences were found between ACT and PRE_CUE (t(14) = 1.364, p = 0.194, d = 0.352, BF10 = 0.011); see Fig. 3a.

Figure 3

Time frequency results of experiment 1 (a), control experiment (b) and experiment 2 (c). In contrast to the functional connectivity results, we observed a similar pattern of enhanced beta power in all three experiments (including the control experiment), indicating that these beta power effects are unspecific to metacognitive decision-making. Time zero refers to the onset of the metacognitive question (see Fig. 1).

Control experiment

In our EEG analyses, we attempted to minimize the effect of the mere presence of a motor response (the act of moving your finger) by focusing on the last 500 ms preceding the metacognitive judgment (see Fig. 1a). Nonetheless, EEG results observed in the first experiment could still be influenced by epiphenomenal/lingering motor activity caused by pressing a button in ACT versus not having pressed a button in PRE_ACT and PRE_CUE. We thus repeated the first experiment (ACT and PRE_ACT) while replacing the verbal confidence judgment with a verbal report of a random letter (see Fig. 4). In this way, we were able to find out whether the observed beta effects (phase synchrony/power) were related to epiphenomenal motor activity or whether this was instead specifically linked to metacognitive judgments. In the control experiment, no differences in first-order performance (da) between the two conditions were observed (t(18) = 0.164, p = 0.872, d = 0.038, BF10 = 0.240; Mean da condition 1 = 0.99, SD = 0.45; Mean da condition 2 = 0.97, SD = 0.44). In contrast to the first experiment, we did not observe a significant difference in functional connectivity between ACT and PRE_ACT (beta phase synchrony: t(18) = 0.475, p = 0.641, d = 0.109, BF10 = 0.263; alpha phase synchrony: t(18) = 0.511, p = 0.615, d = 0.117, BF10 = 0.267), see Fig. 5a. Similarly to the first experiment, however, we did observe a difference in beta power between ACT and PRE_ACT (t(18) = 5.098, p < 0.001, d = 1.201, BF10 = 311.7), see Fig. 3b. These findings indicate that the increase in functional connectivity (beta phase synchrony) between frontal and motor areas is not merely caused by epiphenomenal first-order response activity, but seems instead to be connected to the metacognitive processes that follow first-order responses. In contrast, beta power differences between the conditions in the current experiments seem to be non-specific to what happens after the first-order response: we observed beta power differences when a metacognitive judgment had to be provided as well as when a random letter had to be reported.

Figure 4

Task design control experiment. In the control experiment we replaced the metacognitive decision with a verbal response of a letter, while keeping the rest of the design identical to ACT and PRE_ACT of the first experiment.

Figure 5

(a) Functional connectivity differences of beta (left) and alpha (right) phase synchrony. Similar to Fig. 2, we observed enhanced functional connectivity (beta phase synchrony) between motor cortex and central frontal cortex in ACT where response information was available during metacognitive decision-making compared to PRE_ACT. This effect was not observed in the control experiment where participants were not engaged in a metacognitive task. In all three experiments, no alpha phase synchrony differences were observed. Shaded areas represent within-subjects standard error of the mean. (b) Direct comparisons of the observed beta phase synchrony differences in all three experiments show that the effect is specific to settings in which metacognitive decisions are required. Error bars represent between-subjects standard error of the mean. Time zero refers to the onset of the metacognitive question (see Fig. 1).

Experiment 2

To find out if we could replicate the findings from the first experiment and to investigate whether the strength of the stimulus-response mapping influenced the strength of the observed behavioral and EEG effects, we recorded behavioral data and EEG signals during a second experiment in which we omitted the response cue (see Fig. 6a). As such, the experiment was similar to the first experiment with the exceptions that the stimulus-response mapping was kept stable across the entire experiment, and that the PRE_CUE condition was no longer present.

Figure 6

(a) Task design experiment 2. In the second experiment we omitted the response cue, while keeping the rest of the design similar to experiment 1. (b) Behavioral results. We replicated our findings from the first experiment and observed that metacognitive efficiency decreased when action information was absent, while first order performance remained unaffected. Error bars represent between-subjects standard error of the mean.


We performed (one-tailed) t-tests to find out whether metacognitive efficiency decreased when action information was absent. We replicated findings from the first experiment (though the statistical effect is small) and found increased metacognitive efficiency when response information was available (ACT) compared to PRE_ACT in which this information was absent (t(18) = 2.134, p = 0.023, d = 0.490, BF+0 = 2.89). No significant differences were observed between conditions for da scores (t(18) = 0.713, p = 0.758, d = 0.164, BF+0 = 0.151) or meta da scores (t(18) = 1.622, p = 0.061, d = 0.372, BF+0 = 1.337), see Fig. 6b. In this experiment, we did observe a consistent lower level of confidence in ACT (mean = 2.63, SD = 0.433) compared to PRE_ACT (mean = 2.70, SD = 0.431), t(18) = 2.999, p = 0.012, d = 0.642, BF10 = 4.17.

Although our design is not well suited to investigate differences in reaction times (due to the 1 s blank that preceded each first-order response, see Fig. 6a), we nonetheless tested whether RT differences existed between ACT (Mean = 491 ms, SD = 61) and PRE_ACT (Mean = 502 ms, SD = 70) that could accompany the observed difference in confidence level between the conditions. We found no differences in RT between both conditions (t(18) = 1.970, p = 0.064, d = 0.452, BF10 = 1.163).

EEG results

In the second experiment we repeated the analyses from the first experiment by focusing on functional connectivity differences between ACT and PRE_ACT. We replicated our previous findings and observed higher functional connectivity (beta phase synchrony) in ACT compared to PRE_ACT (t(15) = 4.038, p = 0.001, d = 1.009, BF10 = 36.003; alpha phase synchrony: t(15) = 0.881, p = 0.392, d = 0.22, BF10 = 0.358), see Fig. 5a. We also observed higher beta power in ACT compared to PRE_ACT (t(15) = 2.639, p = 0.019, d = 0.660, BF10 = 3.269, see Fig. 3c), however, due to a similar beta power effect observed in the control experiment, it is highly unlikely that the beta power effects are the result of our experimental manipulation.

General results

In order to determine the overall effect of action processes on metacognitive efficiency, we grouped the data from the first and second experiment together (see methods) using Bayesian statistics, which make it possible to meaningfully aggregate subjects and/or experiments in a post-hoc manner. We therefore grouped PRE_ACT and PRE_CUE from experiment 1 so as to create two conditions, as in experiment 2. We observed strong evidence for higher metacognitive efficiency (BF+0 = 19.151, see Fig. 7) when action information was available during metacognitive judgments. Note that the combined effect is much stronger than the weak behavioral effects observed in each individual study, suggesting the need for large enough sample size. Future studies investigating changes in metacognitive performance could benefit from such a larger sample size, and from using a longer lasting staircase procedure for second-order performance as well as first-order performance, preventing the exclusion of participants.

Figure 7

Combined results. When combining the data from experiment 1 and 2 we find strong evidence for increased metacognitive efficiency when action information is available during metacognitive decision-making. Similarly, strong evidence is observed for increased functional connectivity (beta phase synchrony) between motor channels and central frontal regions when action information is available during metacognitive decision-making.

To test whether functional connectivity differences between ACT and PRE_ACT differed between the experimental and control experiment, we directly compared ACT and PRE_ACT differences with each other33 using independent sampled t-tests. In all experiments we subtracted values from PRE_ACT from ACT. Again we averaged PRE_ACT and PRE_CUE from experiment 1 and subtracted that from the ACT condition. We observed significantly greater differences in the experimental conditions compared to the control condition (first experiment vs. control experiment: t(32) = 2.904, p = 0.007, d = 1.003, BF10 = 6.901; second experiment vs. control experiment: t(33) = 4.057, p < 0.001, d = 1.377, BF10 = 87.51), see Fig. 5b. When examining the combined data from the first and second experiment with respect to functional connectivity, we find strong evidence for greater beta phase synchrony between motor and central frontal regions when action information is available at the moment of metacognitive decision-making (BF+0 = 1127.912, see Fig. 7 & 8).

Figure 8

Topoplot of the combined functional connectivity effect (ACT vs. PRE_ACT). For illustration purposes we plotted beta phase synchrony differences between ‘seed’ electrode C3/C4 and other electrodes to show the spatial distribution of the observed effect.


Decision-making is typically accompanied by an estimate about the quality of one’s choices, actions or performance. Adequate metacognition is not only important in everyday life settings (e.g., whether you can assess whether you are still able to drive safely on a long trip, or knowing what you know while studying for an exam), but can even be critical in certain situations (e.g., in case of medical decisions, or decisions made by a flight controllers). Despite its importance, it remains unclear how metacognition emerges, and what kind of information is used to determine the quality of our decisions.

Here, we investigated whether first-order action information could inform second-order (metacognitive) decisions. Specifically, we studied whether reducing available first-order response information at the moment second-order decisions had to be provided affected metacognitive performance in a color discrimination task. Further, we investigated whether functional connectivity between motor regions and prefrontal cortex could be a candidate to convey action information during metacognitive decision-making. Results demonstrate that metacognitive efficiency slightly decreased when first-order action information was reduced at the moment metacognitive decisions had to be provided. We replicated our findings in a second experiment and showed that the effect was small but robust to changes in the experimental design (see Figs. 1b, 6b & 7). Similarly, we found converging electrophysiological evidence that functional connectivity between motor areas and prefrontal cortex increases during metacognitive decision-making when action information is available (see Fig. 2 & 5). In a control experiment, we demonstrated that this effect was not related to lingering response activity, but in fact specific to metacognitive processes following first-order decisions (Fig. 5). Combined analyses of the three experiments provide converging evidence for the contribution of action information in metacognitive decision-making.

Models of metacognitive decision-making

In lab settings, metacognition is typically studied by asking participants to make a decision about a stimulus (e.g., the motion direction of a cloud of moving dots, the orientation of a grating), after which they are asked to provide the level of confidence in their decision being correct. Previously, it has been shown that manipulating stimulus parameters (evidence strength and evidence reliability) affects confidence judgments34 during perceptual decision-making, suggesting similar (sensory) evidence processing mechanisms support first- and second-order decision-making. Similarly, in signal-detection-like models, the distance of the decision variable from a criterion represents a level of confidence4,6,35,36. The time between the decision and presentation of sensory evidence could in such cases result in discrepancies between first- and second-order decisions, due to prolonged accumulation of evidence10,37,38. Alternatively, different sources or quality of information could contribute to first- and second-order decisions39,40, resulting in different first- and second-order performance12. With respect to the latter, we previously demonstrated that sensory evidence contributing to first-order decision-making does not similarly support metacognitive decision-making. Variance in first-order performance was driven by different stimulus features compared to variance in metacognitive performance. These findings indicated that sensory evidence used for first-order performance differed from information used for metacognitive judgments9. Maniscalco and Lau recently compared models describing discrepancies between first- and second-order decisions during a visual masking task. They compared models which depict first- and second-order decision-making as supported by similar sources of information (single channels models) with dual channel models, which describe two processing streams giving rise to first- and second-order task performance; and hierarchical models, which presume that a late processing stage monitors the state of sensory processing. Their results demonstrated that dissociations between first- and second-order performance are best captured by hierarchical models. Hierarchical models of metacognition propose that sensory evidence used for first-order performance can become susceptible to accrual of noise and signal decay over time and due to further processing12. As such, the experimental design itself can be important as the first-order response is typically given closer in time to stimulus offset compared to the second-order response. Over time, various factors can contribute to a loss in strength of the sensory signal. For instance, further neural processing of the sensory signal could result in the accumulation of noise when arriving at the stage at which this information is being used by the metacognitive system12,39. Therefore, our design not only manipulated the amount of available “action information” but additionally also manipulated the (potential) level of accumulated noise/signal decay. However, in our design the effect of signal decay and noise should counter any beneficial effect of action information available at a later processing stage: On the one hand additional information becomes available for the metacognitive system at a later processing stage, but on the other hand the sensory evidence has most likely become degraded9,12. In the current experiments, we observed slight improvements of metacognitive efficiency when the metacognitive judgment was made with more time in between stimulus offset and the second-order response. In order to tease these different factors apart it would be interesting to combine our previously used experimental design9 with an adaptation of the current design in order to investigate signal decay/noise accumulation in combination with the contribution of action information.

Another factor that has to be taken into account reflects observations indicating that the level of confidence is mainly driven by response-congruent evidence, and appears to be less sensitive to response-incongruent evidence41,42. From such a perspective, a confidence judgment made prior to the first-order decision could be based on the strength of evidence of each response alternative, whereas a confidence judgment made after to the first-order decision would be dominated by response-congruent evidence. In our task, we instructed our participants to provide a level of confidence of the to-be-made decision, thereby stimulating a commitment to one decision alternative prior to the second-order decision. However, we did not assess the exact moment of commitment to the perceptual decision directly, leaving it an open empirical question how information from different response alternatives contributes to confidence judgments when shifting the order within a trial.

Fleming and Daw19 recently put forward a framework in which confidence operates as a second-order computation about one’s own performance. While first-order models are able to reproduce the above-described relationship of confidence and stimulus parameters, their second-order model accommodates the present findings that action information influences metacognitive performance and metacognitive bias. The second-order framework predicts that action affects confidence ratings, in the sense that it decreases overall confidence and enhances metacognitive performance. In the current experiments we observed this pattern in our behavioral results. In two experiments, we demonstrated that metacognitive efficiency increased when first-order action information became available for second-order decision-making. In addition, we observed a (somewhat counterintuitive) decrease in confidence when metacognitive judgments followed first-order responses in the second experiment, as predicted by the second-order model19. We did not observe differences in overall confidence in the first experiment. It could be that trial-by-trial alternations of stimulus-response mappings in the first experiment tampered the effect on metacognitive bias shifts. Previously, it was found that participants’ metacognitive bias shifted when they learned motor sequences in a blocked design compared to when sequences were interleaved43. These findings suggest that the current ease of stimulus-response mappings affected metacognitive bias. In that sense, it would be interesting for future experiments to assess whether/how manipulation of ease or the integrity of first-order responses influences metacognitive behavior.

Beta oscillations

Beta oscillations are classically linked to sensory and motor processing20,28. During preparation and execution of movements, beta band activity typically decreases initially, followed by an increase in beta power44. For instance, an upcoming action could be reliably predicted several seconds prior to response execution, based on lateralization of beta band activity in motor regions, linking beta band activity to the unfolding of an action21. It has been suggested that beta activity reflects the maintenance of an existing motor set whilst weakening processing of new actions45. Interestingly, beta synchronization has been associated with the correctness of an action and has been shown to follow motor errors or after observing the motor errors of others46,47. Recently, the importance of beta oscillations has been demonstrated beyond the sensorimotor domain, extending to visual perception27,48,49, working memory50, long-term memory51, and decision-making21,24,52. It has been proposed that beta oscillations support long-range neuronal interactions25,53,54, thereby maintaining a current cognitive set, sensorimotor state or the so-called ‘status quo’26. In this way, the up or down regulation of beta depends on whether the ‘status quo’ is prioritized over novel incoming signals. Recently, Spitzer and Haegens28 extended the role of beta oscillations further, advocating a role of beta in the awakening of a (endogenous) cognitive set, depending on current task demands.

In the current study, we found increased phase synchrony in the beta band between motor channels and central frontal regions (electrode Fz) specifically when a metacognitive decision followed the first-order response. Critically, when task demands changed and a metacognitive judgment was not required, beta phase synchrony differences between conditions disappeared. In line with the above-proposed role of beta oscillations, our beta phase synchrony findings indicate that task demands (the metacognitive task) resulted in the maintenance of first-order action information (e.g., response fluency, response competition strength). It would be interesting to investigate what role explicitly asking for a metacognitive judgment has on beta band activity. If we assume that decisions are naturally accompanied by an estimate about the quality of an action or choice, it could be that by explicitly asking for such an estimate after a short time interval we could have prolonged or boosted a naturally occurring more transient event (for a similar discussion in consciousness research55). Indeed, beta phase synchrony effects in the control condition initially seem to mimic those observed in the other two experiments, only starting to deflect in the period preceding the metacognitive judgment. It would be interesting to test ‘naturally occurring’ metacognitive processes in future experiments, thereby using observed neural markers of explicitly probed metacognitive processes9,56,57,58,.

Motor activity and metacognition

The present results indicate a contribution of first-order motor response information in metacognitive decision-making. Previously, Wenke and colleagues29 demonstrated that participants were sensitive to conflicting motor activity (response competition) induced by subliminal information. In their study the “ease” or “smoothness” of action selection in a visual reaction-time task was manipulated by presenting a subliminal response prime that was congruent to one out of two action possibilities. Results demonstrated that action priming influenced the sense of control over action consequences following the response. Other work indicates that metacognitive experience of response competition is crucial for triggering cognitive adaptation59,60. Further, it has been shown in the memory domain that the experience of motor fluency is used as a cue that affects metamemory61,62.

Recently, it has been shown that perceptual decisions were biased by the amount of motor effort it took for participants to make the response62. In this study, participants’ decision was biased towards the least effortful motor response. These findings demonstrate that the ease to act on a decision might influence the decision itself. However, it seems that metacognitive awareness of effort or of task demands is necessary for the development of such a decision bias63. In the current experiments, results indicate that participants could be sensitive to response competition, the fluency or ease of the first-order response60,64 when computing an estimate about the quality of the decision.

Alternatively, motor activity could provide insight into the mechanisms of the unfolding perceptual decision. Recent studies demonstrated that evidence accumulation processes ‘echo’ in activity in motor regions21. As such, perceptual and cognitive states could be reflected in the motor system65,66 and be used to inform metacognitive decisions.

Prefrontal cortex and metacognition

Previous work demonstrated that lesions to prefrontal cortex affect metacognitive performance without altering first-order decision-making67,68. Similarly, disrupting prefrontal activity via theta burst stimulation has been shown to selectively alter metacognitive performance69,70,71 (but see72,73). The detection of erroneous behavior, a key aspect of metacognition5,19, has been strongly linked to a rapidly emerging central frontal negativity in the EEG signal (error-related negativity74), thought to reflect coordinated theta oscillatory mechanisms75,76,77,78,79. In addition, theta has been implicated in learning, feedback processing, and action monitoring77,80,81,82,83,84,85. Recently, fluctuations in prefrontal theta band activity has been linked to fluctuations in metacognitive performance9,57,86. Taken together, these findings suggest that frequent exposure to external feedback, learning from one’s correct and incorrect decisions induces a shift in which error detection, initially elicited by external feedback (or observing the consequences of our decisions), is shifting towards the use of internal simulations of stimulus-response contingencies. This internally processing of the probabilities of our actions towards outside events and their most likely outcomes5,87,88,89 could be used to adapt future behavior. In such a way, metacognition could be seen as an internalization of external feedback processing and error monitoring, employing similar neural mechanisms57,90,91.

It has been previously proposed that next to perceptual evidence, inferences about “the state of the decider” (i.e., one’s own actions19, and prior or global estimates of performance92,93) are important for metacognitive decision-making. In addition, to adequately compute an estimate about the quality of a decision it is necessary to know the broader task context or infer “the state of the world” (i.e., value for an action at a certain state of the (task) environment) at the moment of the decision94,95,96,97. Recently, the orbitofrontal cortex has been linked to inferring such “states of the world” during decision-making95,98. As such, central frontal regions and anterior frontal areas could play distinct roles in metacognitive decision-making71. Figure 9 illustrates how sensory, action and interoceptive signals could be integrated in central frontal regions, interacting with anterior prefrontal regions providing inferences about the state of the world94,95 and the state of the decider19,93 when computing an estimate about the quality of a decision.

Figure 9

Sensory, interoceptive and action signals are read out in central frontal cortex. Anterior prefrontal cortex provides predictions about the “state of the world” and the “state of the decider” when a decision is made. Central frontal theta oscillations serve as a mechanism to broadcast the need for control in response to the estimate about the quality of the decision.


In the current study, we focused on functional connectivity changes between motor and prefrontal regions. However, the current neural measurements (EEG) lack spatial specificity to make strong claims about neural sources. It would therefore be necessary to replicate our findings using alternative methods (e.g. fMRI) that have a higher spatial resolution.

We used a staircase performance prior to the experimental blocks to determine appropriate task settings. Despite our efforts we had to exclude participants based on first- and second-order task performance. In future studies it might be useful to use a longer staircase period to eliminate learning effects, and employ a staircase procedure for second-order task performance in addition to first-order performance.


Monitoring and evaluation of one’s own performance is crucial for adept behavior. However, how metacognition emerges is still hotly debated10,12,19. In a series of three experiments, we demonstrated that manipulations of available action information affected metacognitive performance. Concurrent EEG recordings showed that functional connectivity between prefrontal regions and motor areas increased after a first-order response, specifically when a metacognitive judgment was required. Together with previous findings9,17,67, our results paint a picture of metacognition as a second-order process that integrates sensory and action information.

Materials and Methods


Twenty-five participants (15 females, mean age = 21.1, SD = 4.82) took part in experiment 1, twenty-nine participants (18 females, mean age = 22.1, SD = 2.65) in experiment 2, and twenty (13 females, mean age = 21.6, SD = 3.87) in the control experiment. Participants received financial compensation for their participation in this experiment. All participants had normal or corrected-to-normal vision and were naïve to the purpose of the experiment. All procedures complied with international laws and institutional guidelines and were approved by the local Ethics Committee of the Université Libre de Bruxelles, department of Psychology. All participants provided their written informed consent prior to the experiment.

Task design

A field of 600 green and red moving dots was centrally presented (250*250 pixels) on a Dell 17 monitor with a refresh rate of 60 Hz. The monitor was placed at a distance of ~57 cm in front of each participant so that the collection of moving dots subtended a visual angle of 6.6°. Crucially, on each trial a majority of the 600 dots (on average 315.11 dots, SD = 6.76) was either green or red. Participants were instructed to determine what color (red or green) was predominant on each trial by pressing a left (~) or right (/) key. The level of difficulty was determined for each participant individually by using a one-up- two-down staircase procedure in steps of 0.5% of total number of dots before the start of the experiment. After two consecutive correct responses, the difference between the total number of red vs. green dots was reduced by 0.5% (3 dots). During the staircase procedure, each participant performed a total of three blocks (one block of each condition in experiment 1, and one block of each condition in the control and second experiment plus a block randomly picked between condition 1 and 2) in order to assess the level of difficulty that resulted in a stable level of performance set at 71% correct. The stimulus was presented for 800 ms, and at any moment during stimulus presentation a total of 600 dots were displayed. Each trial started with a blank screen (jittered between 1000–1500 ms, in steps of 100 ms) on which a fixation cross was centrally presented. After stimulus presentation a blank was presented for 1000 ms to avoid the influence on prolonged evidence accumulation5,31.

Experiment 1

In the first experiment, we created three conditions by varying the amount of available action information at the moment a metacognitive decision had to be provided (Fig. 1a). In condition 1, the stimulus and blank screen were followed by a response cue (2000 ms), instructing participants whether the left or right button corresponded to the answer “green” or “red” (Fig. 1a). The stimulus-response mapping was randomized so that in approximately half of all trials the left response button signaled ‘red’ and in approximately the other half of the trials it signaled ‘green’. This randomized stimulus-response mapping prevented participants from preparing their response immediately after the visual stimulus had appeared and enabled us to disentangle motor preparation from motor action in both our behavioral and EEG analyses. After the presentation of the response cue, participants were asked to indicate whether the majority of the dots were green or red by pressing the corresponding button with their left or right index finger. Next, participants had to provide a metacognitive judgment about their decision by indicating their level of confidence in being correct on a labeled scale from 1–4, where 1 indicated being very uncertain and 4 being very certain that their first-order response was correct. Participants were encouraged to use the whole range of the scale. Participants verbally reported their confidence rating in order to link the manual motor response exclusively to the first-order decision (red-green decision). A microphone registered all verbal responses using speech recognition software in Presentation (Neurobehavioral Systems, version 18.1), allowing automatic recording of verbal responses. To ensure an accurate transcription of the responses, we set a threshold level of certainty (0.8). Flagged trials below 0.8 certainty were checked manually and corrected if necessary (4% of all trials).

Critically, confidence ratings were given at different points in the trial sequence depending on the condition. Typically, confidence ratings are given after the first-order task response (ACT). However, in this experiment we manipulated the amount of action information (i.e., response execution, action preparation) available for metacognitive decisions by varying the position of metacognitive judgments in a trial. In PRE_ACT, metacognitive judgments had to be provided before the first-order response (after the response cue), while in PRE_CUE metacognitive decisions had to be made prior to the first-order response and presentation of the response cue. This resulted in two conditions in which action information was minimal (response preparation) or absent at the moment the second-order (metacognitive) decision was made.

Control experiment

In the control experiment, we investigated whether observed EEG results were specific to metacognitive processes, by studying the non-specific effect of epiphenomenal/lingering motor activity from first-order responses. Therefore, we used a similar task design as used in the first experiment. Critically, in the control experiment participants were instructed to verbally report one randomly chosen letter out of four presented letters (‘E, ‘G’, ‘P’, ‘T’), instead of providing a confidence rating. Here, we focused on differences between ACT and PRE_ACT, since we did not observe behavioral and functional connectivity differences between PRE_ACT and PRE_CUE in the first experiment, see Fig. 1b.

Experiment 2

In the second experiment, the response cue was removed in order to establish reliable stimulus-response mappings throughout the experiment. The rest of the design was kept similar to that of the first experiment (Fig. 1c).

Behavioral analyses

In the present experiment, we aimed to investigate whether we could observe changes in metacognitive (second-order) performance depending on experimental condition. We therefore used a staircase procedure before starting the experiment (see above) and employed an exclusion criterion of da or meta da > 0.5 and <2.0 observed in the ACT condition (metacognitive performance is typically measured after first-order responses) in order to avoid floor and ceiling effects. As stated above, the aim of this study was to investigate fluctuations in metacognitive performance and such floor and ceiling effects would preclude the aim of the experimental design. Additionally, by filtering the data we tried to avoid potential issues with respect to the structure of the data (.e., by having little correct/incorrect trials in the data that are necessary for meta-d’ measures; for a recent discussion see72). To illustrate, nine participants that were excluded from the last experiment had a mean meta da of −0.02 in the ACT condition. This means that those participants performed the second-order task at chance level. One possible explanation could be that the way we recorded the second-order response (an English verbal report) was challenging for some of the excluded (native French-speaking) participants, thereby avoiding more difficult pronounceable answers.

For analyses, 15 participants were included in the first experiment, 18 in the control experiment and 19 in the second experiment. In order to find out whether first-order and metacognitive performance differed we calculated first-order task sensitivity (because the data was split into three conditions we calculated da35, metacognitive sensitivity (meta-da) and metacognitive efficiency (meta da – da30,32), for each condition separately. First-order task sensitivity (da) and metacognitive sensitivity (meta-da) are bias-free measures of the ability to distinguish two signals from each other and the ability to distinguish between correct and incorrect decisions, respectively (both in units of first-order da). Metacognitive efficiency reflects metacognitive sensitivity relative to different levels of first-order task performance, which is important because metacognitive sensitivity is known to be influenced by first-order task performance30.

We performed three repeated measures analyses of variance (ANOVA) on first- and second-order task performance (da, meta-da, and metacognitive efficiency) with condition as the independent variable. All behavioral analyses were performed using JASP (Version, Matlab (Matlab 12.1, The MathWorks Inc.), type 2 SDT scripts99 and SPSS (IBM SPSS Statistics, 22.0). For the Bayesian analysis in JASP a Cauchy prior distribution centered around zero was used with an interquartile range of r = 0.707.

EEG measurements and analyses

EEG was recorded and sampled at 1048 Hz using a Biosemi ActiveTwo 64-channel system, with four additional electrodes for horizontal and vertical eye-movements, each referenced to their counterpart (Biosemi – Amsterdam, The Netherlands). High-pass filtering (0.5 HZ), additional low-pass filtering (100 HZ) and a notch filter (50 HZ) were used. Next, we down-sampled to 512 Hz and corrected for eye movements on the basis of Independent Component Analysis100. The data was epoched −1.5 s to + 0.5 sec preceding confidence judgments. We removed trials containing irregularities due to EMG or other artifacts by visually inspecting all trials. To increase spatial specificity and to filter out deep sources we converted the data to spline Laplacian signals100,101. We used a sliding window Fourier transform102, window length: 400 ms, step size: 50 ms, to calculate the time-frequency representations of the EEG power (spectrograms) for each channel and each trial. We used a single Hanning taper for the frequency range 3–30 Hz (frequency resolution: 2.5 Hz, bin size: 1 Hz27). To examine the way information might be distributed during metacognitive decision-making, we assessed measures of interregional functional connectivity in the beta range. In our previous study, we specifically observed effects in prefrontal channel Fz related to metacognitive performance9. Therefore, we specifically examined consistencies of the difference of time–frequency phase values between motor channels (C3/C4, depending on the hand that responded) and central frontal electrode Fz (Intersite Phase Clustering (ISPC)25,38) in the 500 ms time period immediately preceding the confidence judgment (see Fig. 1). We used ISPC measurements to determine whether reducing the amount of motor information available at the moment of confidence judgments changed the level of functional connectivity (i.e., alpha/beta phase synchronisation) between central prefrontal9 and motor regions. In experiment 2 three participants had to be excluded from further EEG analyses due failed EEG recordings.

Power modulations were characterized as the percentage of power change at a given time and frequency bin relative to baseline power value for that frequency bin. The baseline was calculated as the mean power across the pre-stimulus interval (from −0.3 to 0 s relative to stimulus onset). All signal processing steps were performed using Brain Vision Analyzer (BrainProducts) and Matlab (Matlab 12.1, The MathWorks Inc.), X code103 and Fieldtrip104.


Monitoring and control of our decision process is a critical part of every day decision-making. When feedback is not available, metacognitive skills enable us to modify current behavior and adapt prospective decision-making. Here, we investigated what kind information is being used to compute an estimate about the quality of our decisions. Results indicate that during perceptual decision-making, information about one’s actions towards perceptual events is being used to evaluate the quality of one’s decisions. EEG results indicate that functional connectivity between motor regions and prefrontal cortex could serve as a mechanism to convey action information during metacognitive decision-making. Considered together, our results demonstrate that post-decisional information contributes to metacognition, thereby evaluating not only what one perceives (e.g., strength of perceptual evidence) but also how one responds towards perceptual events.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. The scripts and toolboxes used for analyzing the data can be downloaded at: TF analyses:, SDT: Statistics:


  1. 1.

    Morales, J., Lau, H. & Fleming, S. M. Domain-General and Domain-Specific Patterns of Activity Supporting Metacognition in Human Prefrontal Cortex. J. Neurosci. 38, 3534–3546 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. 2.

    Vaccaro, A. G. & Fleming, S. M. Thinking about thinking: A coordinate-based meta-analysis of neuroimaging studies of metacognitive judgements. Brain Neurosci. Adv. 2, 1–14 (2018).

    Article  Google Scholar 

  3. 3.

    Rouault, M., Mcwilliams, A., Allen, M. G. & Fleming, S. M. Human metacognition across domains: insights from individual differences and neuroimaging. Pers. Neurosci. 1–28 (2018).

  4. 4.

    Kiani, R. & Shadlen, M. N. Representation of Confidence Associated with a Decision by Neurons in the Parietal Cortex. Science (80-.). 324, 759–764 (2009).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  5. 5.

    Yeung, N. & Summerfield, C. Metacognition in human decision-making: confidence and error monitoring. Philos. Trans. R. Soc. B Biol. Sci. 367, 1310–1321 (2012).

    Article  Google Scholar 

  6. 6.

    Fetsch, C. R., Kiani, R., Newsome, W. & Shadlen, M. N. Effects of Cortical Microstimulation on Confidence in a Perceptual Decision. Neuron 83, 797–804 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. 7.

    Wierzchoń, M., Paulewicz, B., Asanowicz, D., Timmermans, B. & Cleeremans, A. Different subjective awareness measures demonstrate the influence of visual identification on perceptual awareness ratings. Conscious. Cogn. 27C, 109–120 (2014).

    Article  Google Scholar 

  8. 8.

    Fleming, S. M. et al. Action-Specific Disruption of Perceptual Confidence. Psychol. Sci. 26, 89–98 (2015).

    Article  PubMed  PubMed Central  Google Scholar 

  9. 9.

    Wokke, M. E., Cleeremans, A. & Ridderinkhof, K. R. Sure I’m Sure: Prefrontal Oscillations Support Metacognitive Monitoring of Decision Making. J. Neurosci. 37, 781–789 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. 10.

    Berg, R. V. D., Zylberberg, A., Kiani, R., Shadlen, M. N. & Wolpert, D. M. Confidence is the bridge between multi-stage decisions. Curr. Biol. 26, 3157–3168 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. 11.

    Palser, E. R., Fotopoulou, A. & Kilner, J. M. Altering movement parameters disrupts metacognitive accuracy. Conscious. Cogn. 57, 33–40 (2018).

    Article  CAS  Google Scholar 

  12. 12.

    Maniscalco, B. & Lau, H. The signal processing architecture underlying subjective reports of sensory awareness. Neurosci. of Consci. 1, 1–17 (2016).

    Google Scholar 

  13. 13.

    Cisek, P. & Kalaska, J. F. Neural correlates of reaching decisions in dorsal premotor cortex: Specification of multiple direction choices and final selection of action. Neuron 45, 801–814 (2005).

    Article  CAS  Google Scholar 

  14. 14.

    Maniscalco, B. et al. Tuned normalization in perceptual decision-making circuits can explain seemingly suboptimal confidence behavior. bioRxiv: 558858 (2019).

  15. 15.

    Allen, M. et al. Unexpected arousal modulates the influence of sensory noise on confidence. Elife 5 (2016).

  16. 16.

    Urai, A. E., Braun, A. & Donner, T. H. Pupil-linked arousal is driven by decision uncertainty and alters serial choice bias. Nat. Commun. 8, 14637 (2017).

    Article  ADS  PubMed  PubMed Central  Google Scholar 

  17. 17.

    Siedlecka, M., Paulewicz, B. & Wierzchoń, M. But I Was So Sure! Metacognitive Judgments Are Less Accurate Given Prospectively than Retrospectively. Front. Psychol. 7, 218 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  18. 18.

    Pasquali, A., Timmermans, B. & Cleeremans, A. Know thyself: Metacognitive networks and measures of consciousness. Cognition 117, 182–190 (2010).

    Article  Google Scholar 

  19. 19.

    Fleming, S. M. & Daw, N. D. Self-evaluation of decision-making: A general Bayesian framework for metacognitive computation. Psychol. Rev. 124, 91–114 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  20. 20.

    Pfurtscheller, G. & Lopes da Silva, F. H. Event-related EEG/MEG synchronization and desynchronization: basic principles. Clin. Neurophysiol. 110, 1842–57 (1999).

    Article  CAS  Google Scholar 

  21. 21.

    Donner, T. H., Siegel, M., Fries, P. & Engel, A. K. Buildup of Choice-Predictive Activity in Human Motor Cortex during Perceptual Decision Making. Curr. Biol. 19, 1581–1585 (2009).

    Article  CAS  Google Scholar 

  22. 22.

    Tzagarakis, C., Ince, N. F., Leuthold, A. C. & Pellizzer, G. Beta-Band Activity during Motor Planning Reflects Response Uncertainty. J. Neurosci. 30, 11270–11277 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Donner, T. H. et al. Population Activity in the Human Dorsal Pathway Predicts the Accuracy of Visual Motion Detection. J. Neurophysiol. 98, 345–359 (2007).

    Article  Google Scholar 

  24. 24.

    Haegens, S., Nácher, V., Luna, R., Romo, R. & Jensen, O. α-Oscillations in the monkey sensorimotor network influence discrimination performance by rhythmical inhibition of neuronal spiking. Proc. Natl. Acad. Sci. USA 108, 19377–82 (2011).

    Article  ADS  Google Scholar 

  25. 25.

    Siegel, M., Donner, T. H. & Engel, A. K. Spectral fingerprints of large-scale neuronal interactions. Nat. Rev. Neurosci. 13, 20–25 (2012).

    Article  CAS  Google Scholar 

  26. 26.

    Engel, A. K. & Fries, P. Beta-band oscillations — signalling the status quo? Curr. op. in neurobiol. 20, 156–165 (2010).

    Article  CAS  Google Scholar 

  27. 27.

    Kloosterman, N. A. et al. Top-down modulation in human visual cortex predicts the stability of a perceptual illusion. J. Neurophysiol. 113, 1063–76 (2015).

    Article  Google Scholar 

  28. 28.

    Spitzer, B. & Haegens, S. Beyond the Status Quo: A Role for Beta Oscillations in Endogenous Content (Re)Activation. eneuro 4, ENEURO.0170-17.2017 (2017).

  29. 29.

    Wenke, D., Fleming, S. M. & Haggard, P. Subliminal priming of actions influences sense of control over effects of action. Cognition 115, 26–38 (2010).

    Article  Google Scholar 

  30. 30.

    Fleming, S. M. & Lau, H. How to measure metacognition. Front. Hum. Neurosci. 8 (2014).

  31. 31.

    Hebart, M. N., Schriever, Y., Donner, T. H. & Haynes, J.-D. The Relationship between Perceptual Decision Variables and Confidence in the Human Brain. Cereb. Cortex (2014).

  32. 32.

    Brinkman, L., Stolk, A., Dijkerman, H. C., de Lange, F. P. & Toni, I. Distinct roles for alpha- and beta-band oscillations during mental simulation of goal-directed actions. J. Neurosci. 34, 14783–92 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. 33.

    Nieuwenhuis, S., Forstmann, B. U. & Wagenmakers, E.-J. Erroneous analyses of interactions in neuroscience: a problem of significance. Nat. Neurosci. 14, 1105–1107 (2011).

    Article  CAS  Google Scholar 

  34. 34.

    Boldt, A., de Gardelle, V. & Yeung, N. The impact of evidence reliability on sensitivity and bias in decision confidence. J. Exp. Psychol. Hum. Percept. Perform. 43, 1520–1531 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  35. 35.

    Macmillan, N. & Creelman, C. Detection Theory: A User’s Guide. (Psychology Press, 2004).

  36. 36.

    Kepecs, A., Uchida, N., Zariwala, H. A. & Mainen, Z. F. Neural correlates, computation and behavioural impact of decision confidence. Nature 455, 227–31 (2008).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  37. 37.

    Boldt, A. & Yeung, N. Shared Neural Markers of Decision Confidence and Error Detection. J. Neurosci. 35, 3478–3484 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. 38.

    Calderon, C. B., Gevers, W. & Verguts, T. The Unfolding Action Model of Initiation Times, Movement Times, and Movement Paths. Psychol. Rev. 125, 785–805 (2018).

    Article  Google Scholar 

  39. 39.

    Pleskac, T. J. & Busemeyer, J. R. Two-stage dynamic signal detection: A theory of choice, decision time, and confidence. Psychol. Rev. 117, 864–901 (2010).

    Article  Google Scholar 

  40. 40.

    Charles, L., King, J.-R. & Dehaene, S. Decoding the dynamics of action, intention, and error detection for conscious and subliminal stimuli. J. Neurosci. 34, 1158–70 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. 41.

    Zylberberg, A., Barttfeld, P., Sigman, M. & Pereira, A. The construction of confidence in a perceptual decision. Front. int. neurosci. 6, 1–10 (2012).

    Google Scholar 

  42. 42.

    Maniscalco, B., Peters, M. A. K. & Lau, H. Heuristic use of perceptual evidence leads to dissociation between performance and metacognitive sensitivity. Atten Percept Psychophys 923–937 (2016).

  43. 43.

    Simon, D. A. & Bjork, R. A. Metacognition in Motor Learning. J. Exp. Psychol. Learn. Mem. Cogn. 27, 907–912 (2001).

    Article  CAS  Google Scholar 

  44. 44.

    Kilavik, B. E., Zaepffel, M., Brovelli, A., MacKay, W. A. & Riehle, A. The ups and downs of beta oscillations in sensorimotor cortex. Exp. Neurol. 245, 15–26 (2013).

    Article  Google Scholar 

  45. 45.

    Gilbertson, T. et al. Existing Motor State Is Favored at the Expense of New Movement during 13-35 Hz Oscillatory Synchrony in the Human Corticospinal System. J. Neurosci. 25, 7771–7779 (2005).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. 46.

    Koelewijn, T., van Schie, H. T., Bekkering, H., Oostenveld, R. & Jensen, O. Motor-cortical beta oscillations are modulated by correctness of observed action. Neuroimage 40, 767–775 (2008).

    Article  Google Scholar 

  47. 47.

    Swann, N. et al. Intracranial EEG reveals a time- and frequency-specific role for the right inferior frontal gyrus and primary motor cortex in stopping initiated responses. J. Neurosci. 29, 12675–85 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. 48.

    Piantoni, G., Kline, K. A. & Eagleman, D. M. Beta oscillations correlate with the probability of perceiving rivalrous visual stimuli. J. Vis. 10, 18–18 (2010).

    Article  Google Scholar 

  49. 49.

    Bastos, A. M. et al. Visual Areas Exert Feedforward and Feedback Influences through Distinct Frequency Channels. Neuron 85, 390–401 (2015).

    Article  CAS  Google Scholar 

  50. 50.

    Siegel, M., Warden, M. R. & Miller, E. K. Phase-dependent neuronal coding of objects in short-term memory. Proc. Natl. Acad. Sci. USA 106, 21341–6 (2009).

    Article  ADS  CAS  Google Scholar 

  51. 51.

    Hanslmayr, S., Staresina, B. P. & Bowman, H. Oscillations and Episodic Memory: Addressing the Synchronization/Desynchronization Conundrum. Trends Neurosci. 39, 16–25 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. 52.

    Wyart, V., Myers, N. E. & Summerfield, C. Neural Mechanisms of Human Perceptual Choice Under Focused and Divided Attention. J. Neurosci. 35, 3485–3498 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. 53.

    Benchenane, K., Tiesinga, P. H. & Battaglia, F. P. Oscillations in the prefrontal cortex: a gateway to memory and attention. Curr. Opin. Neurobiol. 21, 475–485 (2011).

    Article  CAS  Google Scholar 

  54. 54.

    Thompson, E. & Varela, F. J. Radical embodiment: neural dynamics and consciousness. Trends Cogn. Sci. 5, 418–425 (2001).

    Article  Google Scholar 

  55. 55.

    Tsuchiya, N., Wilke, M., Frässle, S. & Lamme, V. A. F. No-Report Paradigms: Extracting the True Neural Correlates of Consciousness. Trends Cogn. Sci. 19, 757–770 (2015).

    Article  Google Scholar 

  56. 56.

    Fleming, S. M. & Dolan, R. J. The neural basis of metacognitive ability. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 367, 1338–49 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  57. 57.

    Murphy, P. R., Robertson, I. H., Harty, S. & O’Connell, R. G. Neural evidence accumulation persists after choice to inform metacognitive judgments. Elife 4 (2015).

  58. 58.

    Fleming, S. M., Huijgen, J. & Dolan, R. J. Prefrontal Contributions to Metacognition in Perceptual Decision Making. J. Neurosci. 32, 6117–6125 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. 59.

    Desender, K., Van Opstal, F. & Van den Bussche, E. Feeling the conflict: the crucial role of conflict experience in adaptation. Psychol. Sci. 25, 675–83 (2014).

    Article  Google Scholar 

  60. 60.

    Questienne, L., Opstal, F. V. & Dijck, J. V. Metacognition and cognitive control: behavioural adaptation requires conflict experience. Q. J. Exp. Psychol. 1–15 (2016).

    Google Scholar 

  61. 61.

    Susser, J. A. & Mulligan, N. W. The effect of motoric fluency on metamemory. Psychon. Bull. Rev. 22, 1014–1019 (2015).

    Article  Google Scholar 

  62. 62.

    Hagura, N., Haggard, P. & City, S. Perceptual decisions are biased by the cost to act. Elife, 1–20 (2017).

  63. 63.

    Desender, K., Calderon, C. B., Van Opstal, F. & Van den Bussche, E. Avoiding the conflict: Metacognitive awareness drives the selection of low-demand contexts. J. Exp. Psychol. Hum. Percept. Perform. 43, 1397 (2017).

    Article  Google Scholar 

  64. 64.

    Pacherie, E. The phenomenology of action: A conceptual framework. Cognition 107, 179–217 (2008).

    Article  Google Scholar 

  65. 65.

    Lange, F. P. D., Rahnev, D. A., Donner, T. H. & Lau, H. Prestimulus Oscillatory Activity over Motor Cortex Reflects Perceptual Expectations. J. Neurosci. 33, 1400–1410 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  66. 66.

    Song, J. & Nakayama, K. Hidden cognitive states revealed in choice reaching tasks. Trends Cogn. Sci. 13, 360–366 (2009).

    Article  Google Scholar 

  67. 67.

    Fleming, S. M. et al. Action-Specific Disruption of Perceptual Confidence. Psychol. Sci. 26, 89–98 (2014).

    Article  Google Scholar 

  68. 68.

    Pannu, J. K. & Kaszniak, A. W. Metamemory Experiments in Neurological Populations: A Review. Neuropsychol. Rev. 15, 105–130 (2005).

    Article  Google Scholar 

  69. 69.

    Rounis, E., Maniscalco, B., Rothwell, J. C., Passingham, R. E. & Lau, H. Theta-burst transcranial magnetic stimulation to the prefrontal cortex impairs metacognitive visual awareness. Cogn. Neurosci. 1, 165–75 (2010).

    Article  Google Scholar 

  70. 70.

    Ryals, A. J., Rogers, L. M., Gross, E. Z., Polnaszek, K. L. & Voss, J. L. Associative Recognition Memory Awareness Improved by Theta-Burst Stimulation of Frontopolar Cortex. Cereb. Cortex 26, 1200–1210 (2016).

    Article  Google Scholar 

  71. 71.

    Shekhar, M. & Rahnev, D. Distinguishing the Roles of Dorsolateral and Anterior PFC in Visual Metacognition. J. Neurosci. 38, 5078–5087 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  72. 72.

    Bor, D., Schwartzman, D. J., Barrett, A. B. & Seth, A. K. Theta-burst transcranial magnetic stimulation to the prefrontal or parietal cortex does not impair metacognitive visual awareness. PLoS One 12, e0171793 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  73. 73.

    Ruby, E., Maniscalco, B. & Peters, M. A. K. On a ‘failed’ attempt to manipulate visual metacognition with transcranial magnetic stimulation to prefrontal cortex. Conscious. Cogn. 62, 34–41 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  74. 74.

    Falkenstein, M., Hohnsbein, J., Hoormann, J. & Blanke, L. Effects of crossmodal divided attention on late ERP components. II. Error processing in choice reaction tasks. Electroencephalogr. Clin. Neurophysiol. 78, 447–455 (1991).

    Article  CAS  Google Scholar 

  75. 75.

    Bates, A. T., Kiehl, K. A., Laurens, K. R. & Liddle, P. F. Low-frequency EEG oscillations associated with information processing in schizophrenia. Schizophr. Res. 115, 222–230 (2009).

    Article  Google Scholar 

  76. 76.

    Cohen, M. X., Ridderinkhof, K. R., Haupt, S., Elger, C. E. & Fell, J. Medial frontal cortex and response conflict: Evidence from human intracranial EEG and medial frontal cortex lesion. Brain Res. 1238, 127–142 (2008).

    Article  CAS  Google Scholar 

  77. 77.

    Cohen, M. X. & Cavanagh, J. F. Single-Trial Regression Elucidates the Role of Prefrontal Theta Oscillations in Response Conflict. Front. Psychol. 2, 30 (2011).

    Article  PubMed  PubMed Central  Google Scholar 

  78. 78.

    Cavanagh, J. F., Cohen, M. X. & Allen, J. J. B. Prelude to and Resolution of an Error: EEG Phase Synchrony Reveals Cognitive Control Dynamics during Action Monitoring. J. Neurosci. 29, 98–105 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  79. 79.

    Luu, P. & Tucker, D. M. Regulating action: alternating activation of midline frontal and motor cortical networks. Clin. Neurophysiol. 112, 1295–306 (2001).

    Article  CAS  Google Scholar 

  80. 80.

    Jensen, O. & Lisman, J. E. Position Reconstruction From an Ensemble of Hippocampal Place Cells: Contribution of Theta Phase Coding. J. Neurophysiol. 83, 2602–2609 (2000).

    Article  CAS  Google Scholar 

  81. 81.

    Cavanagh, J. F. & Frank, M. J. Frontal theta as a mechanism for cognitive control. Trends Cogn. Sci. 18, 414–421 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  82. 82.

    Dragoi, G. & Buzsáki, G. Temporal Encoding of Place Sequences by Hippocampal Cell Assemblies. Neuron 50, 145–157 (2006).

    Article  CAS  Google Scholar 

  83. 83.

    Sauseng, P. et al. Relevance of EEG alpha and theta oscillations during task switching. Exp. Brain Res. 170, 295–301 (2006).

    Article  CAS  Google Scholar 

  84. 84.

    van Driel, J., Sligte, I. G., Linders, J., Elport, D. & Cohen, M. X. Frequency Band-Specific Electrical Brain Stimulation Modulates Cognitive Control Processes. PLoS One 10, e0138984 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  85. 85.

    van de Vijver, I., Ridderinkhof, K. R. & Cohen, M. X. Frontal Oscillatory Dynamics Predict Feedback Learning and Action Adjustment. J. Cogn. Neurosci. 23, 4106–4121 (2011).

    Article  Google Scholar 

  86. 86.

    Fleming, S. M. Changing our minds about changes of mind. Elife 5, 3–5 (2016).

    Article  Google Scholar 

  87. 87.

    Holroyd, C. B. & Coles, M. G. H. The neural basis of human error processing: Reinforcement learning, dopamine, and the error-related negativity. Psychol. Rev. 109, 679–709 (2002).

    Article  Google Scholar 

  88. 88.

    Cleeremans, A. The Radical Plasticity Thesis: How the Brain Learns to be Conscious. Front. Psychol. 2, 86 (2011).

    Article  PubMed  PubMed Central  Google Scholar 

  89. 89.

    Cleeremans, A., Timmermans, B. & Pasquali, A. Consciousness and metarepresentation: a computational sketch. Neural Netw. 20, 1032–9 (2007).

    Article  Google Scholar 

  90. 90.

    Buzsaki, G., Peyrache, A. & Kubie, J. Emergence of Cognition from Action. Cold Spring Harb. Symp. Quant. Biol. (2014).

  91. 91.

    Buzsaki, G. The Brain from Inside Out. (Oxford University Press, USA., 2019).

  92. 92.

    Benwell, C. S. Y., Beyer, R., Wallington, F. & Ince, R. A. A. History biases reveal novel dissociations between perceptual and metacognitive decision-making. bioRxiv Prepr (2019).

  93. 93.

    Rouault, M., Dayan, P. & Fleming, S. M. Forming global estimates of self-performance from local confidence. Nat. Commun. 1–11 (2019).

  94. 94.

    Wilson, R. C., Takahashi, Y. K., Schoenbaum, G. & Niv, Y. Orbitofrontal cortex as a cognitive map of task space. Neuron 81, 267–279 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  95. 95.

    Schuck, N. W., Wilson, R. & Niv, Y. In Goal-Directed Decision Making 259–278 (Elsevier Inc., 2018).

  96. 96.

    Schuck, N. W., Cai, M. B., Wilson, R. C., Niv, Y. & Road, W. Human Orbitofrontal Cortex Represents a Cognitive Map of State Space. Neuron 91, 1402–1412 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  97. 97.

    Wokke, M. E., Knot, S. L., Fouad, A. & Richard Ridderinkhof, K. Conflict in the kitchen: Contextual modulation of responsiveness to affordances. Conscious. Cogn. 40, 141–146 (2016).

    Article  Google Scholar 

  98. 98.

    Wokke, M. E. & Ro, T. Competitive Frontoparietal Interactions Mediate Implicit Inferences. SO – J. Neurosci. 2019 Jun 26;39(26):5183–5194. J. Neurosci (2019).

  99. 99.

    Maniscalco, B. & Lau, H. A signal detection theoretic approach for estimating metacognitive sensitivity from confidence ratings. Conscious. Cogn. 21, 422–30 (2012).

    Article  Google Scholar 

  100. 100.

    Vigário, R. N. Extraction of ocular artefacts from EEG using independent component analysis. Electroencephalogr. Clin. Neurophysiol. 103, 395–404 (1997).

    Article  Google Scholar 

  101. 101.

    Mitra, P. P. & Pesaran, B. Analysis of Dynamic Brain Imaging Data. Biophys. J. 76, 691–708 (1999).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  102. 102.

    Cohen, M. X. Comparison of different spatial transformations applied to EEG data: A case study of error processing. Int. J. Psychophysiol. (2015).

  103. 103.

    Cohen, M. X. Analyzing Neural Time Series Data: Theory and Practice. MIT Press (2014).

  104. 104.

    Oostenveld, R., Fries, P., Maris, E. & Schoffelen, J.-M. FieldTrip: Open Source Software for Advanced Analysis of MEG, EEG, and Invasive Electrophysiological Data. Comput. Intell. Neurosci. 2011, 1–9 (2011).

    Article  Google Scholar 

Download references


We thank Lisa Padding and Adeline de Cia for helping with data collection and Maries E. Vissers for her useful comments. This work was supported by the European Research Council (Advanced Grant RADICAL to A.C.), the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement (Meta_Mind - DLV-704361 to M.E.W.), and by the Belgian Science Policy Office (Interuniversity Poles of Attraction Grant P7/33 to A.C.). A.C. is a research director of the National Fund for Scientific Research and a senior fellow of the Canadian Institute for Advanced Research (CIFAR).

Author information




M.E.W. and D.A. designed the study, D.A. conducted the experiment, M.E.W. and D.A. analyzed the data, and M.E.W., D.A. and A.C. wrote the manuscript.

Corresponding author

Correspondence to Martijn E. Wokke.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Wokke, M.E., Achoui, D. & Cleeremans, A. Action information contributes to metacognitive decision-making. Sci Rep 10, 3632 (2020).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing