Perceptual Sensitivity, but not Metacognitive Monitoring, is Dependent on Varying Levels of Control

Perceptual Sensitivity, but not Metacognitive Monitoring, is Dependent on Varying Levels of Control | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Perceptual Sensitivity, but not Metacognitive Monitoring, is Dependent on Varying Levels of Control Kazuma Takada, Wen Wen, Shunichi Kasahara, Tom Froese This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8589893/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 29 Apr, 2026 Read the published version in Experimental Brain Research → Version 1 posted You are reading this latest preprint version Abstract Sense of agency is the feeling of control over one’s actions in the environment. This is typically explained in terms of a comparator that evaluates the consistency between sensorimotor predictions and actual outcomes. Under high movement controllability, there is high predictability and hence a strong sense of agency, while a decrease in control produces salient prediction errors. Conversely, under low controllability, the sense of agency may rely more on detecting regularities between actions and outcomes until control increases. However, it remains unclear whether these distinct perceptual processes share the same metacognitive monitoring process. We addressed this question using a control change detection task, where participants moving a single dot on a screen had to detect whether their level of control had changed and report their confidence. Across two experiments, we observed that perceptual sensitivity was higher for decreases than for increases in controllability, but metacognitive process showed no directional difference. Our findings suggest that while distinct perceptual processes are involved for different levels of controllability, metacognitive monitoring shares a common underlying mechanism. sense of agency metacognition prediction error regularity detection m-ratio Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1 Introduction The sense of agency is the subjective feeling that one’s own actions cause changes in the environment and that one is in control (Gallagher, 2000 ). This subjective feeling is essential for our interactions with the environment, inducing movement and its learning (Karsh & Eitam, 2015 ; Tanaka & Imamizu, 2025 ; Wen & Haggard, 2018 ), vocalization (Ohata et al., 2020 ; Wen et al., 2022 ), and decision-making (Haggard, 2017 ; Haggard & Chambon, 2012 ; Haggard & Tsakiris, 2009 ; Wen & Imamizu, 2022 ). For example, when controlling a tool, a sense of agency arises when the actual sensory feedback matches one’s predicted (expected) outcome. Conversely, when disturbances like delays or deviations cause sensory feedback to differ from predicted outcomes, the sense of agency decays. This process of forming the sense of agency based on the prediction error is called the prediction process . In this process, the intensity of sense of agency is determined by comparing the internal model that accurately predicts movement outcomes with the actual feedback received through sensory modalities (Kawato, 1999 ; McNamee & Wolpert, 2019 ; Wolpert et al., 1995 ). For example, Zama et al. demonstrated that delayed visual feedback errors during reaching movement attenuate the sense of agency (Zama et al., 2017 ). Similarly, results from a button-pressing experiment using a virtual hand by Villa et al. showed that monitoring during movement, rather than the success of the intended goal, serves as a stable source of information for generating sense of agency (Villa et al., 2018 , 2021 ). Such prediction errors are not only detected consciously (Oancea et al., 2025 ), but also unconsciously (Pereira et al., 2023 ). In addition to the prediction error in the direction of movement, Vigh & Limanowski ( 2024 ) found that the loss of sense of agency at large delays makes temporal changes easier to detect. These results suggest they contribute to the formation of the sense of agency as a largely innate function. On the other hand, in highly uncertain environments or situations requiring novel motor learning, predictions based on the internal model may fail when a reliable internal model has not been established. Under such situations, it is necessary to detect the regularities between one’s own movements and the resulting changes in the environment (Wen & Haggard, 2020 ). The perception of those regularities can contribute to the emergence of sense of agency. Developing studies involving infants, who do not have the accurate internal model, have reported that when infants perceive causality with environmental changes, their foot movements (Rovee & Rovee, 1969 ) and eye movement are enhanced (Miyazaki et al., 2014 ; Wang et al., 2012 ). Adults can also use regularity detection to distinguish between self-caused events and externally caused events even when they have not yet established an internal model for motor control (Wen & Haggard, 2020 ). Previous studies suggest these two processes may have distinct characteristics. Wen et al. applied a metacognitive measurement approach to a reaching task using a single dot and a control detection task using two dots to examine how consciously each process functions (Wen et al., 2023 ). The results showed a negative correlation between performance on the prediction process and metacognitive sensitivity, while no correlation was observed between performance on the regularity detection process and metacognitive sensitivity. This suggests that the prediction process may be less consciously monitored in people with highly sensitive sense of agency, whereas the metacognitive monitoring of regularity detection does not appear to vary depending on the sensitivity of this process. Furthermore, a recent study suggested that two different modes - exploration and exploitation - may be activated in sense of agency depending on the prior belief of control, and that regularity detection and prediction processes may be involved differently across these modes (Wen, Mei, et al., 2024 ). In the exploitation mode, even small prediction errors can be salient and lead to a significant drop in the sense of agency, while in the exploration mode, sense of agency increases more gradually with sensorimotor input (Wen, Mei, et al., 2024 ). Another study also showed that the sensitivity in detecting a decrease and an increase in control differed even when only the two ends of the change were reversed (i.e., a decrease from 100% to 80% vs an increase from 80% to 100%), suggesting that the processes underlying the perception of a change in sense of agency may differ depending on the direction of the change (Wen, Shimazaki, et al., 2020 ). In summary, these studies suggest that different processes, such as prediction and regularity detection processes, may be involved differently in perceiving changes in sense of agency depending on the direction of change, likely because the belief of control differs across these situations. The present study aims to examine this hypothesis by examining both the perceptual sensitivity and metacognitive sensitivity when perceiving an increase or a decrease in sense of agency. If perceptual sensitivity differs between the directions of change, this would indicate that the underlying perceptual processes are likely to be different. If metacognitive sensitivity differs between the directions of change, this suggests that the conscious accessibility of these perceptual processes also differs, with one being more explicit and the other more implicit. In this study, we applied the signal detection theory and metacognitive approach to the control detection paradigm with changing levels of control designed a previous study (Wen et al., 2020 ). In the experimental task, participants manipulated a single dot displayed on a screen for 5 seconds, answered yes/no (first-order judgement) whether the dot’s motion changed during this period, and rated their confidence in this response on a 4-point scale (second-order judgement). This design separates individual judgment biases from actual perceptual sensitivity (Wen et al., 2024 ), while measuring how explicitly this movement information is recognized and contributes to the formation of sense of agency. Participants’ movements were synthesized with pre-recorded movement components according to the level of control and reflected in the dot (Wen et al., 2020 ). For example, at an 80% control, the dot contained 80% participant movement and only 20% pre-recorded movement. The detection of increase and decrease in control was conducted in different blocks. Half of the trials in each block remained at the initial control level without any change, while in the other half the control level changed at the midpoint. In Experiment 1, the level of control varied within a fixed range between 90% and 60%. In Experiment 2, to standardize the subjective difficulty of change detection for each participant, the detection task was performed using a staircase method that constantly adjusted the magnitude of change, while fixing the upper level of control to 90% (i.e., control increased from a lower level to 90% or decreased from 90% to a lower level). In Experiment 1, we compared d´ and m-ratio between the increase and decrease conditions. In Experiment 2, we used a staircase procedure to equate accuracy across the two conditions and then compared m-ratio. In addition to these metrics, we examined whether individual performance in both the first-order and second-order judgments correlated across the two conditions. Using these designs and analyses, we aimed to determine whether the perceptual processes and metacognitive monitoring differ between perceiving an increase versus a decrease in the sense of agency. 2 Experiment 1 2.1 Participants The sample size was determined based on a power analysis for a paired t-test. The effect size was set at 0.34 based on the m-ratio results from Charalampaki et al. (2023). The power calculation using G*Power 3.1 showed that a sample size of 55 participants (one-tailed test with α = 0.05 and a power of 0.80) is required. Because we planned to conduct the experiment using an online system, in which participant exclusion is likely to be higher and the effect size is likely to be smaller than in face-to-face experiments, we decided to recruit 110 participants for this experiment. We recruited 110 healthy participants (mean age = 32.58, range = 20-60, SD = 9.06, 43 females, 1 other) through Prolific 1 . Participants were required to have an approval rate of 90% or higher on experiments conducted on Prolific. All participants were English speakers and used the computer that met the experimental system requirements. Written informed consent was obtained from all participants prior to the experiment, and they provided gender (male, female, or other) and age. After providing informed consent, they worked on the experimental tasks using their own computer, and they received financial compensation after finishing all tasks. This study was approved by the local ethics committee (Okinawa Institute of Science and Technology Graduate University, ethics number: #HSR-2024-022). 2.2 Experimental design Experiment 1 employed a within-participants design consisting of two blocks that manipulated the direction of change in the level of control (two directions: increase/decrease). The block order was counterbalanced across participants. The level of control of the stimulus (the dot) was manipulated between 60% and 90%. The level of control refers to the proportion of the participant’s movement contained within the dot’s motion. The magnitude of the change in level of control was fixed at 30%. In the increase block, a single-step change from 60% to 90% was applied; in the decrease block, a single-step change from 90% to 60% was applied. This change occurred 2.5 seconds after the participant began moving the dot. Each block consisted of 60 main trials, including 30 change trials and 30 no-change trials. In addition, six practice trials (three change and three no-change trials) were presented before the main trials in each block. In total, each participant completed 132 trials. 2.3 Experimental task and apparatus Figure 1A shows the flow of a trial. Each trial consisted of a 3-second countdown, 5 seconds of controlling the dot, a response indicating whether a change in control occurred, and a level of confidence in their own response. After the control period, the question “Did the control of the dot change midway?” appeared on the screen, and participants pressed the Q key (“Yes, it changed”) or the W key (“No, it didn’t change”). They were then asked, “How likely do you feel that your answer is correct?” and rated their confidence on a four-point scale (“Not very likely”, “Not likely”, “Likely”, “Very likely”) by pressing the 1, 2, 3, or 4 key, respectively. When the trial involved a change, the level of control of the dot changed abruptly at 2.5 seconds after the first movement of participants’ mouse. During the practice trials, correctness feedback for the detection of change was provided after participants completed the confidence judgement. The dot’s movement was generated by blending participants’ real-time movements and pre-recorded movements from other individuals. At each frame, the dot’s movement direction was computed as a weighted average (e.g., 90/10 in the 90%-control condition) of the participant’s mouse-movement angle (i.e., change in position from the previous frame) and the moving angle from a continuous sequence of pre-recorded mouse movements (Wen et al., 2020). The speed of the dot’s movement always matched the speed of the participant’s mouse movement. When participants stopped moving the mouse, the dot also stopped on the screen. Participants were instructed to keep moving their mouse throughout the trial. Figure 1B shows examples of dot trajectories under the decrease in control and increase in control conditions. When the participant’s ratio is high, the dot’s movement closely aligns with their actual movement. However, as their ratio decreases, the prediction error increases. 2.4 Procedure The experiment was conducted using an online experimental platform (Kasahara & Takada, 2021). Participants were eligible for the experiment if their laptop or desktop met the following requirements: Full HD or higher resolution, browser size of 1600x940 or larger, refresh rate of 60Hz, macOS or Windows operating system, and use of Google Chrome as the browser. These requirements were checked by the platform. Participants who did not meet these requirements were declined for participation. For the experimental task, participants were asked to use a wired mouse and keyboard. To prevent the OS pointer from moving outside the experimental window in the browser, participants were first asked to click the center of the screen. This locked the OS pointer to the browser window and made it invisible. During the experimental task, mouse position at each frame was acquired to generate the dot’s movement. If a participant wished to withdraw from the experiment, they could press the ESC key to unlock the pointer and then close the browser window. A message explaining this withdrawal option was always displayed in the lower left corner of the screen. The experiment consisted of two blocks - an increase block and a decrease block. At the beginning of each block, participants were informed of the direction in which their control over the dot might change in the upcoming trials. Participants pressed the A key to start a block. This allowed them to take a short break before starting the second block. The first six trials of each block were practice trials. The level of control over the dot’s movement changed in three trials and remained unchanged in the other three trials. The trial order was randomized. During the practice trials, participants were shown whether their response of control change detection was correct or not after they had rated their confidence. After the practice trials, the main trials were conducted. The main trial consisted of 60 trials in total, with half containing a change in control over the dot’s movement and the other half containing no change. The trial order was randomized for each participant. In total, participants completed 132 trials: 6 practice trials and 60 main trials in each of the two blocks. The experiment lasted approximately 45 minutes, including the time participants spent reading the informed consent and instructions before the task. 2.5 Data analysis We first calculated d´ of the signal detection theory (Green & Swets, 1966) for each participant and condition. The d´ was computed based on the hit rate and false alarm rate. When the hit rate or false alarm rate took extreme values of 0 or 1, the z-transformation required for computing d’ would not converge. Therefore, we applied a standard log-linear correction to each rate prior to computing d´. d´ serves as an indicator of the perceptual sensitivity to a change in control independent of response bias. In the analysis of meta-d´ and m-ratio, we excluded participants based on their d´ and their use of the confidence scale. Specifically, participants with a d´ ≤ 0.5 in either condition were excluded. This exclusion criterion was used because metacognitive sensitivity would not be meaningful if the first-order judgment was around or lower than the chance level. The exclusion criterion of d´ was determined based on prior studies (Castillo et al., 2024; Chung et al., 2023; Reyes et al., 2023). In addition, we also excluded participants who used two or fewer distinct confidence ratings across trials, because limited variability in confidence ratings prevents reliable estimation of meta-d´ using maximum likelihood estimation. The m-ratio for each participant was defined as meta-d´/d´ (Fleming & Lau, 2014). A higher m-ratio indicates a stronger ability to accurately reflect, via confidence ratings, the information used in first-order judgements. After excluding all participants who met either of these criteria, we conducted paired t -tests for d´, meta-d´, and m-ratio. The metaSDT 2 library was used to compute meta-d´ and m-ratio (Maniscalco & Lau, 2012). In addition, the results for meta-d´ and m-ratio calculated using hierarchical Bayesian modelling (Fleming, 2017) are included in Supplementary Materials S1. The different methods of calculating meta-d´ did not affect our findings. Therefore, we reported the results using the HMeta-d 3 toolbox. Lastly, we calculated the correlations for all four metrics between the increase and decrease in control conditions to examine whether the perceptual and metacognitive processes involved in detecting an increase versus a decrease in control reflect similar individual differences. 2.6 Results In total, 76 participants showed low sensitivity (d´ ≤ 0.5, including negative values) in one or both conditions, particularly in the increase condition. Consequently, these participants were excluded from all subsequent analyses, resulting in a final sample of 34 participants in Experiment 1. Paired t -tests revealed significant differences for d´ ( t (33) = 5.55, p < .001, Cohen’s d = 0.95; Figure 2A left), and meta-d´ ( t (33) = 3.77, p < .001, Cohen’s d = 0.65; Figure 2A right). These results indicate that changes in level of control were detected more accurately and with higher metacognitive sensitivity in the decrease in control condition than in the increase in control condition. Furthermore, the results of paired t -test for confidence rating showed that there was a significant difference between conditions ( t (33) = 4.21, p < .001, Cohen’s d = 0.72). In other words, for the same magnitude of change, participants were more sensitive to decrease than increases in the level of control and were more confidence about their sense of agency. We also conducted correlation analysis using linear regression to examine performances across conditions. This analysis revealed weak positive associations between d´ across conditions ( r (32) = 0.30, p = .085; Figure 2B left) and between meta-d´ across conditions ( r (32) = 0.25, p = .161; Figure 2B right). However, neither association reached statistical significance. These results indicate that perceptual sensitivity (d´) and metacognitive sensitivity (meta-d´) are likely to be independent across the two conditions. Next, we calculated the m-ratio for 34 participants and analyzed their differences and correlation. Paired t -tests revealed that there was no significant difference in m-ratio between the two conditions ( t (33) = 0.83, p = .415, Cohen’s d = 0.14; Figure 3A). The results from the correlation analysis showed that m-ratios in the two conditions were not significantly correlated ( r (32) = -0.03, p = .887; Figure 3B). In addition to the absence of these differences and the lack of correlation, substantial individual difference was observed (Figure 3C). For example, participants were observed not only with similar m-ratios across both conditions, but also with extremely large m-ratios in one condition. Similar results were also demonstrated in the supplementary analysis using a hierarchical Bayesian model (Supplementary Figure S1). 3 Experiment 2 In Experiment 2, we used a staircase method for adjusting the magnitude of change in control to equate participants’ subjective difficulty levels. The experimental task, apparatus, and procedure were identical to those in Experiment 1. 3.1 Participants The sample size was determined in the same way as in Experiment 1. We recruited 110 healthy participants (mean age = 37.19, range = 20-60, SD = 10.38, 48 females, 2 others) through Prolific. Participants had an approval rate of 90% or higher on experiments conducted on Prolific, their native language was English, and the computers they used met the experimental system requirements. Written informed consent was obtained from all participants prior to the experiment, and they provided gender (male, female, or other) and age. After providing informed consent, they worked on the experimental tasks using their own computer, and they received financial compensation after finishing all tasks. This study was approved by the local ethics committee (Okinawa Institute of Science and Technology Graduate University, ethics number: #HSR-2024-022). 3.2 Experimental design Experiment 2 employed a within-participants design consisting of two blocks manipulating the direction of change in the level of control (two levels: increase and decrease), as in Experiment 1. The block order was counterbalanced across participants. In experiment 2, the range of level of control changes (the difficulty of detection) was adjusted using a staircase method. The higher end of the level of control was fixed at 90% (i.e., control either increased from a lower level to 90% or decreased from 90% to a lower level). The staircase adjusted the lower end of the level of control, which was initially set at 60% at the beginning of each block. Lowering the lower end reduced task difficulty, whereas raising it increased task difficulty. The task difficulty was adjusted using a 2-up/1-down algorithm. On trials in which the level of control changed (change trials), the difficulty increased by one step if a participant responded correctly on two in a row. After any incorrect response on a change trial, the difficulty decreased by one step immediately. Furthermore, for difficulty to increase, the correct answer rate over the previous 30 trials with changing the level of control had to exceed 71%. The step size for changes in level of control was 2.5%. These procedures were designed to achieve an approximate 71% accuracy rate for participants across two conditions, enabling evaluation of change detection in level of control and its cognitive processing. As in Experiment 1, the change in the level of control occurred 2.5 seconds after the participant began moving. Each experiment block consisted of 110 trials. The first 10 trials in each block were not adjusted based on the staircase. The next 100 trials were organized into 25 sets of four trials: within each set, two trials contained a change in the level of control and two trials did not, and the order of change and no-change trials was randomized within each set. In each block, the participant performed 6 practice trials, containing three change trials and three no-change trials before the main trials. During practice trials, feedback on the correctness of change detection was provided at the end of each trial. Each participant completed a total of 232 trials. 3.3 Data analysis In Experiment 2, we applied an exclusion criterion based on the range of the staircase. Participants who reached the minimum level of control of 2.5% in either condition (i.e., increase from 2.5% to 90% or decrease from 90% to 2.5%) were excluded from all analyses. After applying this criterion, we calculated d´ for each participant and condition and then excluded participants with d´ ≤ 0.5 and participants who used few confidence ratings, as in Experiment 1. For the remaining participants, we calculated meta-d´ and m-ratio, and then we conducted paired t -tests and correlation analyses on all four metrics (d´, meta-d´, confidence rating, and m-ratio) and mean level of control, defined as the level of control reached through the staircase procedure in each condition. In addition, we conducted correlation analyses for the mean level of control, meta-d´, and m-ratio to assess whether these measures show similar individual differences across the increase and decrease conditions. For the meta-d´ and m-ratio, we also estimated these values using hierarchical Bayesian modeling with HMeta-d as in Experiment 1. 3.4 Results First, 23 participants were excluded based on the staircase range and confidence-rating criteria. In addition, 31 participants showed low sensitivity (d´ ≤ 0.5) in at least one of the two conditions and were also excluded, resulting in a final sample of 56 participants for the subsequent analyses. Paired t -tests showed a significant difference in mean level of control between conditions ( t (55) = 6.64, p < .001, Cohen’s d = 0.89; Figure 4A and Figure 4B). As shown in Figure 4C, significant differences were also observed in d´ ( t (55) = 6.01, p < .001, Cohen’s d = 0.80; Figure 4C left). The difference in meta-d´ was relatively small but statistically significant ( t (55) = 3.30, p = .002, Cohen’s d = 0.44; Figure 4C right). In the confidence rating, there was no significant difference between conditions based on paired t -test ( t (55) = 2.60, p = .012, Cohen’s d = 0.35). To examine how consistently individuals performed across the two conditions, we conducted linear regression analyses between the decrease and increase in control conditions for the mean level of control and meta-d´. Consequently, no significant correlation was found between conditions for meta-d´ ( r (54) = 0.17, p = .21), but none of these reached statistical significance. Nevertheless, a significant positive correlation was found for the mean level of control ( r (54) = 0.65, p < .001), indicating that the relative amount of change required for detection was fairly consistent across conditions even when detection performance and metacognitive measures showed little shared variance. Figure 5A shows the distribution of m-ratio for the 56 participants. There was no significant difference between conditions ( t (55) = 0.16, p = .870, Cohen’s d = 0.02). Moreover, m-ratios across conditions were not significantly correlated ( r (54) = .187, p = .168; Figure 5B), and there were large individual differences as shown in Figure 5C. These findings suggest that, even when the task difficulty was equated among the participants, substantial individual differences remained in the metacognitive monitoring of control. Similar results were obtained when we applied a hierarchical Bayesian model (Supplementary Figure S2). 4 General Discussion The present study aimed to clarify whether perceiving an increase and a decrease in control relies on different perceptual processes, and whether these processes are monitored differently at metacognitive level. Participants controlled a dot on the screen for 5 seconds, and they first judged whether they experienced a change in control over the dot during the trial, and then rated their confidence in that response on a 4-point scale. The results of Experiment 1 showed that the perceptual sensitivity (d´) was significantly higher in the control decrease condition than the control increase condition, even though the magnitude of change was identical across conditions. Moreover, there was no significant correlation in d´ between the two conditions, suggesting that the perceptual processes involved in detecting increases and decreases in control are likely to differ. In contrast, the m-ratio did not significantly differ between the conditions, indicating that the metacognitive monitoring of these perceptual processes is similar. Both the detection of increases and decreases in the sense of agency appear to be highly explicit. Furthermore, Experiment 2 used a staircase to hold detection accuracy as close as possible to 71%. The findings replicated the main results of Experiment 1: the m-ratio again did not significantly differ between the conditions, while achieving the target accuracy required a larger magnitude of change for detecting an increase in control compared with detecting a decrease in control. Sense of agency has typically been explained within prediction framework based on internal models. In this framework, the brain is thought to generate predicted sensory feedback from efference copies of motor commands and compares these predictions with the actual sensory feedback. Prediction errors are continuously monitored, but only those exceeding a certain threshold give rise to explicit awareness that control has been lost. This error-monitoring mechanism may be efficient when a stable mapping between one’s actions and sensory feedback has already been acquired and the belief of control is strong. In this case, even small prediction errors can trigger salient drop in sense of agency (Wen, Chang, et al., 2024 ; Wen, Mei, et al., 2024 ). Our results are consistent with this view. The detection of a decrease in control was significantly more sensitive than the detection of an increase in control, even when the magnitude of change was identical (Experiment 1). Moreover, participants were able to detect relatively smaller changes in control when control decreased from 90%, compared to when it increased to 90%. Such sensitive detection reflects a reliable error-detection mechanism. In contrast, when control increases from a lower level to a higher level, people must explore the regularity between their actions and the sensory feedback. In this case, prediction errors are less reliable and informative because the mapping between actions and outcome is uncertain. The process of detecting regularities between actions and outcomes is likely to be more dominant than the error-detection process in perceiving an increase in control. A previous study reported that the detection of an increase in control shows much larger individual differences than the detection of a decrease in control (Wen et al., 2021 ). However, it was unclear whether the large individual differences arose from the criterion used to determine whether they had gained sufficient control, or from the sensitivity in perceiving such a change. Our results show a similar trend of larger individual differences in the detection of an increase than the detection of a decrease in control, indicating that a large portion of the individual differences lines in sensitivity. This is likely because regularity detection is linked to motor learning (Nobusako et al., 2022 ; Wen et al., 2021 ), and different strategies and skills can be developed during such learning. This asymmetry in sensitivity of detecting an increase and a decrease in control mirrors findings from motor learning and speech studies. For example, sensitivity to error is modulated by environmental stability; it is higher in stable environments compared to random ones. Notably, within a stable environment, sensitivity drops at the onset of perturbation and subsequently grows as adaptation progresses (Todorov et al., 2025 ). In addition, a speech study showed stronger neural responses when auditory feedback deviates further from the learned distribution than when it moves closer to it (Tang et al., 2025 ). Taken together, these findings suggest that the brain differentiates the source of variability, upregulating sensitivity when errors signal internal motor instability while downregulating it in response to external environmental noise. When control drops from a high level, the consequence can be critical in many circumstances, and a highly sensitive error-monitoring system may be important for survival. Such a system may be innate and thus show relatively small individual differences. Furthermore, the selection between error-detection processes and regularity-detection processes is thought to depend on the belief in control (Wen, Chang, et al., 2024 ; Wen, Mei, et al., 2024 ). When the belief in control is strong, a mode called exploitation is activated, and the prediction process dominates in the sense of agency (Wen, Mei, et al., 2024 ). On the other hand, when the belief in control is weak, a mode called exploration mode is activated, and the process of regularity detection dominates in the sense of agency (Wen, Mei, et al., 2024 ). The range of control levels (i.e., 60%-90%) larges lies in the category of being in control (Wen et al., 2020 ). This may explain the higher sensitivity in the detection of a decrease in control than in the detection of an increase. The mode of the sense of agency may switch depending on the belief in control, and thus the sensitivities in detecting an increase and a decrease in control may become comparable or even reverse in different circumstances. Nevertheless, our results show that the two types of processes underlying the sense of agency play very different roles in the detection of a change in control depending on the direction of the change. The present study also highlighted the metacognitive monitoring of the processes involved in detecting an increase and a decrease in control. The results of m-ratio did not differ between the two conditions, suggesting that the metacognitive monitoring is comparable between the two types of processes. The m-ratio was high in both experiments (Experiment 1: mean _inc = 0.87, SD _inc = 0.39, mean _dec = 0.96, SD _dec = 0.35; Experiment 2: mean _inc = 0.79, SD _inc = 0.65, mean _dec = 0.80, SD _dec = 0.44), indicating that the detection of a change in sense of agency is highly explicit. However, individual m-ratios did not correlate across the two conditions. This indicates that although the metacognitive efficiency was comparable at the group level, the processes underlying the metacognitive monitoring may still differ. Moreover, we observed that while the sensitivity (d´) was strongly supported by a conservative decision criterion (i.e., higher criteria is associated with higher d´), the m-ratio was independent of such decision biases (Supplementary Figure S3 and Figure S4). This suggests that the substantial between-participant variability we observed in m-ratio reflects genuine variations in the precision of the monitoring system, unconfounded by individual differences in individual strategy. For example, some participants showed high metacognitive efficiencies in both conditions, whereas others showed high metacognitive efficiency in one condition but poor metacognitive efficiency in the other condition. It remains unclear where the individual differences in metacognitive monitoring come from and how they shape the sense of agency. One possibility is that each individual may use the prediction process and the regularity-detection process differently in the sense of agency, and that metacognitive efficiency may be driven by the relative weightings of these processes. However, the present study was not designed to directly test this possibility, and it is worth further examination in future research. 5 Limitation This study has several limitations. First, all experiments were conducted online. While current web browsers allow us to obtain participants’ screen resolution and refresh rate, they cannot capture the physical display size or the control-display ratio. Consequently, the physical parameters of the experimental stimuli could not be strictly controlled across participants. To mitigate the potential noise introduced by these variations, we recruited twice the number of participants required by the power analysis for each experiment. After exclusions, data from 34 participants in Experiment 1 and 56 in Experiment 2 were retained for the final analysis. Although the final sample size of Experiment 1 fell below the target determined by the power calculation, Experiment 2 met the required sample size and successfully replicated the findings of Experiment 1. This replication supports the robustness of our results despite the variations in individual experimental environments. Furthermore, the number of trials per block in each experiment was relatively limited. Rahnev ( 2025 ) suggests that at least 100 trials are typically required to accurately estimate metacognitive indices. In the present study, Experiment 1 included 60 trials per block, and Experiment 2 included 110 trials per block. To address the issue of limited trial counts, we employed hierarchical Bayesian modeling (Fleming, 2017 ), which allows for accurate parameter estimation even with fewer trials, in addition to standard estimation methods. The results from both approaches were consistent, supporting the robustness of our findings. Future research should nevertheless aim to increase the number of trials per condition to further enhance estimation accuracy. 6 Conclusion The present study shows that the contribution of the prediction process and the regularity detection process to the sense of agency are not simply two expressions of a single mechanism but instead rely on different perceptual processes. The metacognitive efficiency of these processes is comparable, but it remains unclear whether the processes involved in their metacognitive monitoring differ. Our findings refine current models of sense of agency by highlighting the importance of considering different perceptual processes across different modes of sense of agency and shed light on the large individual differences in both perceptual sensitivity and metacognitive monitoring in sense of agency. Declarations Acknowledgements We are grateful to Tae Morrisey for her assistance with the ethical application and coordination with the ethics committee, and to Kaori Yamashiro for her support in processing the online experiments. Funding: This work was supported by JST FOREST Program (to WW; Grant Number JPMJFR2144), JST PRESTO (to SK; Grant Number JPMJPR23I4), JST Moonshot R&D Program (to WW and SK; Grant Number JPMJMS2013), and JSPS DC1 (to KT; Grant Number 25KJ2242). Conflicts of interest: The authors declare that there is no conflict of interest. Availability of data and material: The datasets in this study are available in the Open Science Framework repository, https://osf.io/u5qv7/ Code availability: The scripts in this study are available in the Open Science Framework repository, https://osf.io/u5qv7/ Ethics Approval: All study were approved by the local ethics committee (Okinawa Institute of Science and Technology Graduate University, ethics number: #HSR-2024-022). Consent to participate: Informed consent was obtained from all individual participants included in the study. Author Contribution K.T., W.W., and T.F. conceived the study and defined the project direction. K.T. and W.W. designed the experiments. K.T. implemented the experimental environment, conducted the experiments, and analyzed the data. K.T. wrote the first draft of the manuscript with figures. W.W. contributed to the writing, editing, and critical revision of the manuscript. S.K. and T.F. contributed to the editing and critical revision of the manuscript. W.W., S.K., and T.F. supervised the project. All authors read and approved the final manuscript. References Castillo, J., Sieweyumptewa, P., & Phelps, E. A. (2024). Differential effects of negative valence and memory type on accuracy, confidence, and metacognitive efficiency. Scientific Reports , 14 (1), 25685. Charalampaki, A., Peters, C., Maurer, H., Maurer, L. K., Müller, H., Verrel, J., & Filevich, E. (2023). Motor outcomes congruent with intentions may sharpen metacognitive representations. Cognition , 235 (105388), 105388. Chung, Y. H., Tam, J., Wyble, B., & Störmer, V. S. (2023). Object meaningfulness increases incidental memory of shape but not location. In PsyArXiv . https://doi.org/10.31234/osf.io/th2rm Fleming, S. M. (2017). HMeta-d: hierarchical Bayesian estimation of metacognitive efficiency from confidence ratings. Neuroscience of Consciousness , 2017 (1), nix007. Fleming, S. M., & Lau, H. C. (2014). How to measure metacognition. Frontiers in Human Neuroscience , 8 , 443. Gallagher, I., I. (2000). Philosophical conceptions of the self: implications for cognitive science. Trends in Cognitive Sciences , 4 (1), 14–21. Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics . John Wiley & Sons. Haggard, P. (2017). Sense of agency in the human brain. Nature Reviews. Neuroscience , 18 (4), 196–207. Haggard, P., & Chambon, V. (2012). Sense of agency. Current Biology: CB , 22 (10), R390-2. Haggard, P., & Tsakiris, M. (2009). The experience of agency: Feelings, judgments, and responsibility. Current Directions in Psychological Science , 18 (4), 242–246. Karsh, N., & Eitam, B. (2015). I control therefore I do: judgments of agency influence action selection. Cognition , 138 (C), 122–131. Kasahara, S., & Takada, K. (2021). Stealth Updates of Visual Information by Leveraging Change Blindness and Computational Visual Morphing. ACM Transactions on Applied Perception , 18 (4), 1–17. Kawato, M. (1999). Internal models for motor control and trajectory planning. Current Opinion in Neurobiology , 9 (6), 718–727. Maniscalco, B., & Lau, H. (2012). A signal detection theoretic approach for estimating metacognitive sensitivity from confidence ratings. Consciousness and Cognition , 21 (1), 422–430. McNamee, D., & Wolpert, D. M. (2019). Internal Models in Biological Control. Annual Review of Control, Robotics, and Autonomous Systems , 2 , 339–364. Miyazaki, M., Takahashi, H., Rolf, M., Okada, H., & Omori, T. (2014). The image-scratch paradigm: a new paradigm for evaluating infants’ motivated gaze control. Scientific Reports , 4 (1), 5498. Nobusako, S., Wen, W., Nagakura, Y., Tatsumi, M., Kataoka, S., Tsujimoto, T., Sakai, A., Yokomoto, T., Takata, E., Furukawa, E., Asano, D., Osumi, M., Nakai, A., & Morioka, S. (2022). Developmental changes in action-outcome regularity perceptual sensitivity and its relationship to hand motor function in 5-16-year-old children. Scientific Reports , 12 (1), 17606. Oancea, G., Maniscalco, B., Peters, M. A. K., & Chapman, C. S. (2025). Measuring motor awareness and metacognition at the start, middle, and end of a reaching movement. Consciousness and Cognition , 132 , 103878. Ohata, R., Asai, T., Kadota, H., Shigemasu, H., Ogawa, K., & Imamizu, H. (2020). Sense of Agency Beyond Sensorimotor Process: Decoding Self-Other Action Attribution in the Human Brain. Cerebral Cortex , 30 (7), 4076–4091. Pereira, M., Skiba, R., Cojan, Y., Vuilleumier, P., & Bègue, I. (2023). Preserved metacognition for undetected visuomotor deviations. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience , 43 (35), 6176–6184. Rahnev, D. (2025). A comprehensive assessment of current methods for measuring metacognition. Nature Communications , 16 (1), 701. Reyes, M., Morales, M. J., & Bajo, M. T. (2023). Judgments of learning in bilinguals: Does studying in a L2 hinder learning monitoring? PloS One , 18 (12), e0286516. Rovee, C. K., & Rovee, D. T. (1969). Conjugate reinforcement of infant exploratory behavior. Journal of Experimental Child Psychology , 8 (1), 33–39. Tanaka, T., & Imamizu, H. (2025). Sense of agency for a new motor skill emerges via the formation of a structural internal model. Communications Psychology , 3 (1), 70. Tang, D.-L., Parrell, B., Beach, S. D., & Niziolek, C. A. (2025). The brain’s sensitivity to sensory error can be modulated by altering perceived variability. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience , 45 (5), e0024242024. Todorov, D. I., Moulin, M., Buch, E., & Quentin, R. (2025). Modulation of error sensitivity during motor learning across time, space and environment variability. In bioRxivorg (p. 2025.05.30.656374). https://doi.org/10.1101/2025.05.30.656374 Vigh, G., & Limanowski, J. (2024). Baseline dependent differences in the perception of changes in visuomotor delay. Frontiers in Human Neuroscience , 18 , 1495592. Villa, R., Tidoni, E., Porciello, G., & Aglioti, S. M. (2018). Violation of expectations about movement and goal achievement leads to Sense of Agency reduction. Experimental Brain Research , 236 (7), 2123–2135. Villa, R., Tidoni, E., Porciello, G., & Aglioti, S. M. (2021). Freedom to act enhances the sense of agency, while movement and goal-related prediction errors reduce it. Psychological Research , 85 (3), 987–1004. Wang, Q., Bolhuis, J., Rothkopf, C. A., Kolling, T., Knopf, M., & Triesch, J. (2012). Infants in control: rapid anticipation of action outcomes in a gaze-contingent paradigm. PloS One , 7 (2), e30884. Wen, W., Chang, A. Y.-C., & Imamizu, H. (2024). The sensitivity and criterion of sense of agency. Trends in Cognitive Sciences , 28 (5), 397–399. Wen, W., Charles, L., & Haggard, P. (2023). Metacognition and sense of agency. Cognition , 241 (105622), 105622. Wen, W., & Haggard, P. (2018). Control Changes the Way We Look at the World. Journal of Cognitive Neuroscience , 30 (4), 603–619. Wen, W., & Haggard, P. (2020). Prediction error and regularity detection underlie two dissociable mechanisms for computing the sense of agency. Cognition , 195 , 104074. Wen, W., & Imamizu, H. (2022). The sense of agency in perception, behaviour and human–machine interactions. Nature Reviews Psychology , 1 (4), 211–222. Wen, W., Ishii, H., Ohata, R., Yamashita, A., Asama, H., & Imamizu, H. (2021). Perception and control: individual difference in the sense of agency is associated with learnability in sensorimotor adaptation. Scientific Reports , 11 (1), 20542. Wen, W., Mei, J., Aktas, H., Chang, A. Y.-C., Suzuishi, Y., & Kasahara, S. (2024). Control over self and others’ face: exploitation and exploration. Scientific Reports , 14 (1), 15473. Wen, W., Okon, Y., Yamashita, A., & Asama, H. (2022). The over-estimation of distance for self-voice versus other-voice. Scientific Reports , 12 (1), 420. Wen, W., Shimazaki, N., Ohata, R., Yamashita, A., Asama, H., & Imamizu, H. (2020). Categorical Perception of Control. ENeuro , 7 (5). https://doi.org/10.1523/ENEURO.0258-20.2020 Wolpert, D. M., Ghahramani, Z., & Jordan, M. I. (1995). An internal model for sensorimotor integration. Science , 269 (5232), 1880–1882. Zama, T., Takahashi, Y., & Shimada, S. (2017). The effects of trajectory and endpoint errors in a reaching movement on the sense of agency. Psychology , 08 (14), 2321–2332. Footnotes Prolific. https://www.prolific.com/ metaSDT. https://github.com/craddm/metaSDT HMeta-d. https://github.com/metacoglab/HMeta-d Additional Declarations No competing interests reported. Supplementary Files supplementaryvideo.mp4 SupplementaryMaterial.docx Cite Share Download PDF Status: Published Journal Publication published 29 Apr, 2026 Read the published version in Experimental Brain Research → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8589893","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":579556741,"identity":"a9883872-c340-4efb-8496-7fba27e33c9a","order_by":0,"name":"Kazuma Takada","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAx0lEQVRIiWNgGAWjYBACCSBmbGBgZuAHUQwMbMRoYYZokWwAUgdI0mJwAMg7QIzDJBv4D36cUWMtb3y8ufnzBwa+PIJapBmYmSU3HEs33HbmYJsE0GHFBLXIgbzxgO0w47YbiW0gvyQ2EKGF+eeDf4ftN89IbP5AlBagw9gkN7YdTtwgkdggQZQWyWZmM8uZfenJM0B+OWNAhF8kjjc+vtnzzdq2v7398YeKimOEQwzoe2RgcCyBsBY0UEO6llEwCkbBKBj2AABuwT1jXlZreAAAAABJRU5ErkJggg==","orcid":"","institution":"Okinawa Institute of Science and Technology","correspondingAuthor":true,"prefix":"","firstName":"Kazuma","middleName":"","lastName":"Takada","suffix":""},{"id":579556743,"identity":"79e1ce4d-6e39-4b7e-afcf-efec840c6160","order_by":1,"name":"Wen Wen","email":"","orcid":"","institution":"Rikkyo University","correspondingAuthor":false,"prefix":"","firstName":"Wen","middleName":"","lastName":"Wen","suffix":""},{"id":579556744,"identity":"86204b96-0a52-4a81-a16f-2a9130009033","order_by":2,"name":"Shunichi Kasahara","email":"","orcid":"","institution":"Sony Computer Science Laboratories","correspondingAuthor":false,"prefix":"","firstName":"Shunichi","middleName":"","lastName":"Kasahara","suffix":""},{"id":579556747,"identity":"48e428a4-c9de-41cd-b0b5-51c087ad6612","order_by":3,"name":"Tom Froese","email":"","orcid":"","institution":"Okinawa Institute of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Tom","middleName":"","lastName":"Froese","suffix":""}],"badges":[],"createdAt":"2026-01-13 09:23:39","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8589893/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8589893/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00221-026-07306-w","type":"published","date":"2026-04-29T15:58:36+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":101404725,"identity":"048cfab1-1df4-40d9-bd76-863e5003c345","added_by":"auto","created_at":"2026-01-29 10:37:39","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":301757,"visible":true,"origin":"","legend":"\u003cp\u003ePanel A depicts\u003cstrong\u003e \u003c/strong\u003ethe timeline of the control change detection task. Experiment 1 and Experiment 2 used the same task design, but the magnitude of the change differed between them. Each trial began with a 3-seconds countdown, after which a dot appeared at a random position on the screen. Participants then freely moved the dot using a mouse for 5 seconds. Five seconds after their first mouse movement, the dot froze in place, and participants indicated whether their control over dot’s movement had changed during the trial by pressing the Q or W key. They then rated their confidence in their judgement on a 4-point scale using the 1, 2, 3, and 4 keys. Panel B illustrates an example of blending the participant’s movement with a pre-recorded movement. The blending ratio varies depending on the condition. In the “decrease in control” condition, the dot’s movement initially tracks the participant’s actual movement. However, the weight of the pre-recorded movement increases after 2.5 seconds, causing the dot to deviate from the participant’s input. Conversely, in the “increase in control” condition, the pre-recorded movement is dominant initially, but the participant’s contribution increases after 2.5 seconds, resulting in the dot moving closer to the participant’s actual trajectory.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/c3bcf07c2a1cfd7ec16bb8bc.png"},{"id":101404807,"identity":"011df0e0-fe96-4641-a6ad-759c0c89ca18","added_by":"auto","created_at":"2026-01-29 10:37:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":227957,"visible":true,"origin":"","legend":"\u003cp\u003eResults of the d´ and meta-d´ in Experiment 1. All panel present results after excluding participant with d′ ≤ 0.5; data from 34 participants were included in the analysis. Panel A shows the participants’ distribution in d´ and meta-d´. Significant differences were observed between the two conditions for each metrics based on the paired t-tests. This result suggests that sensitivity in the two processes and its reflection have different characteristics. Panel B depicts the correlations of d´ and meta-d´ between the two conditions. Although the slopes remained positive, they were smaller and not statistically reliable.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/fe180f04cf3af312948c2585.png"},{"id":101404723,"identity":"da9d08f6-ff34-4ef5-9d0e-fe5827cb92b6","added_by":"auto","created_at":"2026-01-29 10:37:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":167842,"visible":true,"origin":"","legend":"\u003cp\u003eResults of m-ratio. The data were 34 participants after excluding the participants with d´ ≤ 0.5. Panel A shows the distribution of participants’ m-ratio under each condition, and no significant differences were observed between the two conditions. Panel B shows the scatter plot of the m-ratio with correlation in decrease condition and increase condition. Each dot represents one participant, and no correlation was observed between two conditions. Panel C depicts the difference in m-ratio between conditions. These results demonstrate that individual sensitivity depends on the direction of change in level of control. For example, some participants were more sensitive to increases in level of control, others to decreases, and some showed similar sensitivity to both directions of change.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/0aa7274af48baacdf233fe36.png"},{"id":101404724,"identity":"db1647e9-c21a-481c-9073-b8c5e0a31a73","added_by":"auto","created_at":"2026-01-29 10:37:39","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":408123,"visible":true,"origin":"","legend":"\u003cp\u003eResults in Experiment 2, after exclusion based on the minimum level of control, few confidences rating, and d´ ≤ 0.5 (N=56). Panel A shows the level of control transition in each condition and Panel B demonstrates the distribution of the mean level of control that participant reached. A significant difference was observed in the required change magnitude for detection between increases and decreases in the level of control. Panel C shows the result of d´ and meta-d´. A significant difference was also observed in d´ and this result suggests that a decrease in level of control is easier to detect, as in Experiment 1. However, there was no significant differences in meta-d´. Panel D shows the results after excluding participants who had d´ ≤ 0.5 and selected fewer types of confidence ratings. Weak positive correlations were observed between conditions for meta-d´ (b = 0.17, p = .21, R\u003csup\u003e2\u003c/sup\u003e = .03, r = .17), but none were statistically significant.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/98fdef2b79e9aa74a062e00e.png"},{"id":101404727,"identity":"5d04b393-789a-4ef9-858d-082612506851","added_by":"auto","created_at":"2026-01-29 10:37:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":168725,"visible":true,"origin":"","legend":"\u003cp\u003eResults of m-ratio (N=56). Panel A shows the distribution of participants’ m-ratio under each condition, and no significant differences were observed between the two conditions. Panel B shows the scatter plot of the m-ratio with correlation in decrease and increase in control condition. Each dot represents one participant, and no correlation was observed between two conditions. Panel C depicts the difference in m-ratio between conditions. These results indicate that even when subjective difficulty is equated across participants, individual differences in the relative advantages of each process persist.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/6e89f5be6219e52676969f76.png"},{"id":108438043,"identity":"8e6ad5a3-7306-47b3-99b0-c90839113595","added_by":"auto","created_at":"2026-05-04 16:05:51","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1584266,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/9181159b-34e5-4113-bcc2-3890c346923a.pdf"},{"id":101404748,"identity":"26be5d4c-b6e3-44e0-8ee8-94c7520ac554","added_by":"auto","created_at":"2026-01-29 10:37:42","extension":"mp4","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":93147674,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaryvideo.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/f3daaf48826f04c2fed31f65.mp4"},{"id":101404728,"identity":"2da7dc0a-5164-492f-ada7-0a45354fef7c","added_by":"auto","created_at":"2026-01-29 10:37:40","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":14783184,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-8589893/v1/d7a5450011e56fe8f1ce033d.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Perceptual Sensitivity, but not Metacognitive Monitoring, is Dependent on Varying Levels of Control","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eThe sense of agency is the subjective feeling that one\u0026rsquo;s own actions cause changes in the environment and that one is in control (Gallagher, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). This subjective feeling is essential for our interactions with the environment, inducing movement and its learning (Karsh \u0026amp; Eitam, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Tanaka \u0026amp; Imamizu, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Wen \u0026amp; Haggard, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), vocalization (Ohata et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wen et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and decision-making (Haggard, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Haggard \u0026amp; Chambon, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Haggard \u0026amp; Tsakiris, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Wen \u0026amp; Imamizu, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). For example, when controlling a tool, a sense of agency arises when the actual sensory feedback matches one\u0026rsquo;s predicted (expected) outcome. Conversely, when disturbances like delays or deviations cause sensory feedback to differ from predicted outcomes, the sense of agency decays. This process of forming the sense of agency based on the prediction error is called the \u003cem\u003eprediction process\u003c/em\u003e. In this process, the intensity of sense of agency is determined by comparing the internal model that accurately predicts movement outcomes with the actual feedback received through sensory modalities (Kawato, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; McNamee \u0026amp; Wolpert, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Wolpert et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e1995\u003c/span\u003e). For example, Zama et al. demonstrated that delayed visual feedback errors during reaching movement attenuate the sense of agency (Zama et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Similarly, results from a button-pressing experiment using a virtual hand by Villa et al. showed that monitoring during movement, rather than the success of the intended goal, serves as a stable source of information for generating sense of agency (Villa et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Such prediction errors are not only detected consciously (Oancea et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), but also unconsciously (Pereira et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In addition to the prediction error in the direction of movement, Vigh \u0026amp; Limanowski (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) found that the loss of sense of agency at large delays makes temporal changes easier to detect. These results suggest they contribute to the formation of the sense of agency as a largely innate function.\u003c/p\u003e \u003cp\u003eOn the other hand, in highly uncertain environments or situations requiring novel motor learning, predictions based on the internal model may fail when a reliable internal model has not been established. Under such situations, it is necessary to detect the regularities between one\u0026rsquo;s own movements and the resulting changes in the environment (Wen \u0026amp; Haggard, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The perception of those regularities can contribute to the emergence of sense of agency. Developing studies involving infants, who do not have the accurate internal model, have reported that when infants perceive causality with environmental changes, their foot movements (Rovee \u0026amp; Rovee, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1969\u003c/span\u003e) and eye movement are enhanced (Miyazaki et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Adults can also use regularity detection to distinguish between self-caused events and externally caused events even when they have not yet established an internal model for motor control (Wen \u0026amp; Haggard, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrevious studies suggest these two processes may have distinct characteristics. Wen et al. applied a metacognitive measurement approach to a reaching task using a single dot and a control detection task using two dots to examine how consciously each process functions (Wen et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The results showed a negative correlation between performance on the prediction process and metacognitive sensitivity, while no correlation was observed between performance on the regularity detection process and metacognitive sensitivity. This suggests that the prediction process may be less consciously monitored in people with highly sensitive sense of agency, whereas the metacognitive monitoring of regularity detection does not appear to vary depending on the sensitivity of this process.\u003c/p\u003e \u003cp\u003eFurthermore, a recent study suggested that two different modes - exploration and exploitation - may be activated in sense of agency depending on the prior belief of control, and that regularity detection and prediction processes may be involved differently across these modes (Wen, Mei, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In the exploitation mode, even small prediction errors can be salient and lead to a significant drop in the sense of agency, while in the exploration mode, sense of agency increases more gradually with sensorimotor input (Wen, Mei, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Another study also showed that the sensitivity in detecting a decrease and an increase in control differed even when only the two ends of the change were reversed (i.e., a decrease from 100% to 80% vs an increase from 80% to 100%), suggesting that the processes underlying the perception of a change in sense of agency may differ depending on the direction of the change (Wen, Shimazaki, et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In summary, these studies suggest that different processes, such as prediction and regularity detection processes, may be involved differently in perceiving changes in sense of agency depending on the direction of change, likely because the belief of control differs across these situations. The present study aims to examine this hypothesis by examining both the perceptual sensitivity and metacognitive sensitivity when perceiving an increase or a decrease in sense of agency. If perceptual sensitivity differs between the directions of change, this would indicate that the underlying perceptual processes are likely to be different. If metacognitive sensitivity differs between the directions of change, this suggests that the conscious accessibility of these perceptual processes also differs, with one being more explicit and the other more implicit.\u003c/p\u003e \u003cp\u003eIn this study, we applied the signal detection theory and metacognitive approach to the control detection paradigm with changing levels of control designed a previous study (Wen et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In the experimental task, participants manipulated a single dot displayed on a screen for 5 seconds, answered yes/no (first-order judgement) whether the dot\u0026rsquo;s motion changed during this period, and rated their confidence in this response on a 4-point scale (second-order judgement). This design separates individual judgment biases from actual perceptual sensitivity (Wen et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), while measuring how explicitly this movement information is recognized and contributes to the formation of sense of agency. Participants\u0026rsquo; movements were synthesized with pre-recorded movement components according to the level of control and reflected in the dot (Wen et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). For example, at an 80% control, the dot contained 80% participant movement and only 20% pre-recorded movement. The detection of increase and decrease in control was conducted in different blocks. Half of the trials in each block remained at the initial control level without any change, while in the other half the control level changed at the midpoint. In Experiment 1, the level of control varied within a fixed range between 90% and 60%. In Experiment 2, to standardize the subjective difficulty of change detection for each participant, the detection task was performed using a staircase method that constantly adjusted the magnitude of change, while fixing the upper level of control to 90% (i.e., control increased from a lower level to 90% or decreased from 90% to a lower level). In Experiment 1, we compared d\u0026acute; and m-ratio between the increase and decrease conditions. In Experiment 2, we used a staircase procedure to equate accuracy across the two conditions and then compared m-ratio. In addition to these metrics, we examined whether individual performance in both the first-order and second-order judgments correlated across the two conditions. Using these designs and analyses, we aimed to determine whether the perceptual processes and metacognitive monitoring differ between perceiving an increase versus a decrease in the sense of agency.\u003c/p\u003e"},{"header":"2 Experiment 1","content":"\u003ch2\u003e2.1 Participants\u003c/h2\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; The sample size was determined based on a power analysis for a paired t-test. The effect size was set at 0.34 based on the m-ratio results from Charalampaki et al. (2023). The power calculation using G*Power 3.1 showed that a sample size of 55 participants (one-tailed test with \u0026alpha; = 0.05 and a power of 0.80) is required. Because we planned to conduct the experiment using an online system, in which participant exclusion is likely to be higher and the effect size is likely to be smaller than in face-to-face experiments, we decided to recruit 110 participants for this experiment.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe recruited 110 healthy participants (mean age = 32.58, range = 20-60, \u003cem\u003eSD\u003c/em\u003e = 9.06, 43 females, 1 other) through Prolific\u003ca href=\"#_ftn1\" name=\"_ftnref1\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e1\u003c/sup\u003e. Participants were required to have an approval rate of 90% or higher on experiments conducted on Prolific. All participants were English speakers and used the computer that met the experimental system requirements. Written informed consent was obtained from all participants prior to the experiment, and they provided gender (male, female, or other) and age. After providing informed consent, they worked on the experimental tasks using their own computer, and they received financial compensation after finishing all tasks. This study was approved by the local ethics committee (Okinawa Institute of Science and Technology Graduate University, ethics number: #HSR-2024-022).\u003c/p\u003e\n\u003ch2\u003e2.2 Experimental design\u003c/h2\u003e\n\u003cp\u003eExperiment 1 employed a within-participants design consisting of two blocks that manipulated the direction of change in the level of control (two directions: increase/decrease). The block order was counterbalanced across participants. The level of control of the stimulus (the dot) was manipulated between 60% and 90%. The level of control refers to the proportion of the participant\u0026rsquo;s movement contained within the dot\u0026rsquo;s motion. The magnitude of the change in level of control was fixed at 30%. In the increase block, a single-step change from 60% to 90% was applied; in the decrease block, a single-step change from 90% to 60% was applied. This change occurred 2.5 seconds after the participant began moving the dot. Each block consisted of 60 main trials, including 30 change trials and 30 no-change trials. In addition, six practice trials (three change and three no-change trials) were presented before the main trials in each block. In total, each participant completed 132 trials.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003e2.3 Experimental task and apparatus\u003c/h2\u003e\n\u003cp\u003eFigure 1A shows the flow of a trial.\u0026nbsp;Each trial consisted of a 3-second countdown, 5 seconds of controlling the dot, a response indicating whether a change in control occurred, and a level of confidence in their own response. After the control period, the question \u0026ldquo;Did the control of the dot change midway?\u0026rdquo; appeared on the screen, and participants pressed the Q key (\u0026ldquo;Yes, it changed\u0026rdquo;) or the W key (\u0026ldquo;No, it didn\u0026rsquo;t change\u0026rdquo;). They were then asked, \u0026ldquo;How likely do you feel that your answer is correct?\u0026rdquo; and rated their confidence on a four-point scale (\u0026ldquo;Not very likely\u0026rdquo;, \u0026ldquo;Not likely\u0026rdquo;, \u0026ldquo;Likely\u0026rdquo;, \u0026ldquo;Very likely\u0026rdquo;) by pressing the 1, 2, 3, or 4 key, respectively. When the trial involved a change, the level of control of the dot changed abruptly at 2.5 seconds after the first movement of participants\u0026rsquo; mouse. During the practice trials, correctness feedback for the detection of change was provided after participants completed the confidence judgement.\u003c/p\u003e\n\u003cp\u003eThe dot\u0026rsquo;s movement was generated by blending participants\u0026rsquo; real-time movements and pre-recorded movements from other individuals. At each frame, the dot\u0026rsquo;s movement direction was computed as a weighted average (e.g., 90/10 in the 90%-control condition) of the participant\u0026rsquo;s mouse-movement angle (i.e., change in position from the previous frame) and the moving angle from a continuous sequence of pre-recorded mouse movements (Wen et al., 2020). The speed of the dot\u0026rsquo;s movement always matched the speed of the participant\u0026rsquo;s mouse movement. When participants stopped moving the mouse, the dot also stopped on the screen. Participants were instructed to keep moving their mouse throughout the trial. Figure 1B shows examples of dot trajectories under the decrease in control and increase in control conditions. When the participant\u0026rsquo;s ratio is high, the dot\u0026rsquo;s movement closely aligns with their actual movement. However, as their ratio decreases, the prediction error increases.\u003c/p\u003e\n\u003ch2\u003e2.4 Procedure\u003c/h2\u003e\n\u003cp\u003eThe experiment was conducted using an online experimental platform (Kasahara \u0026amp; Takada, 2021). Participants were eligible for the experiment if their laptop or desktop met the following requirements: Full HD or higher resolution, browser size of 1600x940 or larger, refresh rate of 60Hz, macOS or Windows operating system, and use of Google Chrome as the browser. These requirements were checked by the platform. Participants who did not meet these requirements were declined for participation. For the experimental task, participants were asked to use a wired mouse and keyboard. To prevent the OS pointer from moving outside the experimental window in the browser, participants were first asked to click the center of the screen. This locked the OS pointer to the browser window and made it invisible. During the experimental task, mouse position at each frame was acquired to generate the dot\u0026rsquo;s movement. If a participant wished to withdraw from the experiment, they could press the ESC key to unlock the pointer and then close the browser window. A message explaining this withdrawal option was always displayed in the lower left corner of the screen.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe experiment consisted of two blocks - an increase block and a decrease block. At the beginning of each block, participants were informed of the direction in which their control over the dot might change in the upcoming trials. Participants pressed the A key to start a block. This allowed them to take a short break before starting the second block. The first six trials of each block were practice trials. The level of control over the dot\u0026rsquo;s movement changed in three trials and remained unchanged in the other three trials. The trial order was randomized. During the practice trials, participants were shown whether their response of control change detection was correct or not after they had rated their confidence. After the practice trials, the main trials were conducted. The main trial consisted of 60 trials in total, with half containing a change in control over the dot\u0026rsquo;s movement and the other half containing no change. The trial order was randomized for each participant. In total, participants completed 132 trials: 6 practice trials and 60 main trials in each of the two blocks. The experiment lasted approximately 45 minutes, including the time participants spent reading the informed consent and instructions before the task.\u003c/p\u003e\n\u003ch2\u003e2.5 Data analysis\u003c/h2\u003e\n\u003cp\u003eWe first calculated d\u0026acute; of the signal detection theory (Green \u0026amp; Swets, 1966) for each participant and condition. The d\u0026acute; was computed based on the hit rate and false alarm rate. When the hit rate or false alarm rate took extreme values of 0 or 1, the z-transformation required for computing d\u0026rsquo; would not converge. Therefore, we applied a standard log-linear correction to each rate prior to computing d\u0026acute;. d\u0026acute; serves as an indicator of the perceptual sensitivity to a change in control independent of response bias.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the analysis of meta-d\u0026acute; and m-ratio, we excluded participants based on their d\u0026acute; and their use of the confidence scale. Specifically, participants with a d\u0026acute; \u0026le; 0.5 in either condition were excluded. This exclusion criterion was used because metacognitive sensitivity would not be meaningful if the first-order judgment was around or lower than the chance level. The exclusion criterion of d\u0026acute; was determined based on prior studies (Castillo et al., 2024; Chung et al., 2023; Reyes et al., 2023). In addition, we also excluded participants who used two or fewer distinct confidence ratings across trials, because limited variability in confidence ratings prevents reliable estimation of meta-d\u0026acute; using maximum likelihood estimation. The m-ratio for each participant was defined as meta-d\u0026acute;/d\u0026acute; (Fleming \u0026amp; Lau, 2014). A higher m-ratio indicates a stronger ability to accurately reflect, via confidence ratings, the information used in first-order judgements. After excluding all participants who met either of these criteria, we conducted paired \u003cem\u003et\u003c/em\u003e-tests for d\u0026acute;, meta-d\u0026acute;,\u0026nbsp;and m-ratio.\u0026nbsp;The\u0026nbsp;metaSDT\u003ca href=\"#_ftn2\" name=\"_ftnref2\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e2\u003c/sup\u003e library was used to compute meta-d\u0026acute; and m-ratio (Maniscalco \u0026amp; Lau, 2012). In addition, the results for meta-d\u0026acute; and m-ratio calculated using hierarchical Bayesian modelling (Fleming, 2017) are included in Supplementary Materials S1. The different methods of calculating meta-d\u0026acute; did not affect our findings. Therefore, we reported the results using the HMeta-d\u003ca href=\"#_ftn3\" name=\"_ftnref3\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e3\u003c/sup\u003e toolbox. Lastly, we calculated the correlations for all four metrics between the increase and decrease in control conditions to examine whether the perceptual and metacognitive processes involved in detecting an increase versus a decrease in control reflect similar individual differences.\u003c/p\u003e\n\u003ch2\u003e2.6 Results\u003c/h2\u003e\n\u003cp\u003eIn total, 76 participants showed low sensitivity (d\u0026acute; \u0026le; 0.5, including negative values) in one or both conditions, particularly in the increase condition. Consequently, these participants were excluded from all subsequent analyses, resulting in a final sample of 34 participants in Experiment 1. Paired \u003cem\u003et\u003c/em\u003e-tests revealed significant differences for\u0026nbsp;d\u0026acute; (\u003cem\u003et\u003c/em\u003e(33) = 5.55, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, Cohen\u0026rsquo;s d = 0.95; Figure 2A left), and meta-d\u0026acute; (\u003cem\u003et\u003c/em\u003e(33) = 3.77, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, Cohen\u0026rsquo;s d = 0.65; Figure 2A right). These results indicate that changes in level of control were detected more accurately and with higher metacognitive sensitivity in the decrease in control condition than in the increase in control condition. Furthermore, the results of paired \u003cem\u003et\u003c/em\u003e-test for confidence rating showed that there was a significant difference between conditions (\u003cem\u003et\u003c/em\u003e(33) = 4.21, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, Cohen\u0026rsquo;s d = 0.72). In other words, for the same magnitude of change, participants were more sensitive to decrease than increases in the level of control and were more confidence about their sense of agency. We also conducted correlation analysis using linear regression to examine performances across conditions. This analysis revealed weak positive associations between d\u0026acute; across conditions (\u003cem\u003er\u003c/em\u003e(32) = 0.30, \u003cem\u003ep\u003c/em\u003e = .085; Figure 2B left) and between meta-d\u0026acute; across conditions (\u003cem\u003er\u003c/em\u003e(32) = 0.25, \u003cem\u003ep\u003c/em\u003e = .161; Figure 2B right). However, neither association reached statistical significance. These results indicate that perceptual sensitivity (d\u0026acute;) and metacognitive sensitivity (meta-d\u0026acute;) are likely to be independent across the two conditions.\u003c/p\u003e\n\u003cp\u003eNext, we calculated the m-ratio for 34 participants and analyzed their differences and correlation. Paired \u003cem\u003et\u003c/em\u003e-tests revealed that\u0026nbsp;there was no significant difference in m-ratio between the two conditions (\u003cem\u003et\u003c/em\u003e(33) = 0.83, \u003cem\u003ep\u003c/em\u003e = .415, Cohen\u0026rsquo;s d = 0.14; Figure 3A). The results from the correlation analysis showed that m-ratios in the two conditions were not significantly correlated (\u003cem\u003er\u003c/em\u003e(32) = -0.03, \u003cem\u003ep\u003c/em\u003e = .887; Figure 3B). In addition to the absence of these differences and the lack of correlation, substantial individual difference was observed (Figure 3C). For example, participants were observed not only with similar m-ratios across both conditions, but also with extremely large m-ratios in one condition. Similar results were also demonstrated in the supplementary analysis using a hierarchical Bayesian model (Supplementary Figure S1).\u003c/p\u003e"},{"header":"3 Experiment 2","content":"\u003cp\u003eIn Experiment 2, we used a staircase method for adjusting the magnitude of change in control to equate participants\u0026rsquo; subjective difficulty levels. The experimental task, apparatus, and procedure were identical to those in Experiment 1.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003e3.1 Participants\u003c/h2\u003e\n\u003cp\u003eThe sample size was determined in the same way as in Experiment 1. We recruited 110 healthy participants (mean age = 37.19, range = 20-60, \u003cem\u003eSD\u003c/em\u003e = 10.38, 48 females, 2 others) through Prolific. Participants had an approval rate of 90% or higher on experiments conducted on Prolific, their native language was English, and the computers they used met the experimental system requirements. Written informed consent was obtained from all participants prior to the experiment, and they provided gender (male, female, or other) and age. After providing informed consent, they worked on the experimental tasks using their own computer, and they received financial compensation after finishing all tasks. This study was approved by the local ethics committee (Okinawa Institute of Science and Technology Graduate University, ethics number: #HSR-2024-022).\u003c/p\u003e\n\u003ch2\u003e3.2 Experimental design\u003c/h2\u003e\n\u003cp\u003eExperiment 2 employed a within-participants design consisting of two blocks manipulating the direction of change in the level of control (two levels: increase and decrease), as in Experiment 1. The block order was counterbalanced across participants. In experiment 2, the range of level of control changes (the difficulty of detection) was adjusted using a staircase method.\u0026nbsp;The higher end of the level of control was fixed at 90% (i.e., control either increased from a lower level to 90% or decreased from 90% to a lower level). The staircase adjusted the lower end of the level of control, which was initially set at 60% at the beginning of each block. Lowering the lower end reduced task difficulty, whereas raising it increased task difficulty. The task difficulty\u0026nbsp;was adjusted using a 2-up/1-down algorithm. On trials in which the level of control changed (change trials), the difficulty increased by one step if a participant responded correctly on two in a row. After any incorrect response on a change trial, the difficulty decreased by one step immediately. Furthermore, for difficulty to increase, the correct answer rate over the previous 30 trials with changing the level of control\u0026nbsp;had to\u0026nbsp;exceed 71%. The step size for changes in level of control was 2.5%. These procedures were designed to achieve an approximate 71% accuracy rate for participants across two conditions, enabling evaluation of change detection in level of control and its cognitive processing.\u003c/p\u003e\n\u003cp\u003eAs in Experiment 1, the change in the level of control occurred 2.5 seconds after the participant began moving. Each experiment block consisted of 110 trials. The first 10 trials in each block were not adjusted based on the staircase. The next 100 trials were organized into 25 sets of four trials: within each set, two trials contained a change in the level of control and two trials did not, and the order of change and no-change trials was randomized within each set. In each block, the participant performed 6 practice trials, containing three change trials and three no-change trials before the main trials. During practice trials, feedback on the correctness of change detection was provided at the end of each trial. Each participant completed a total of 232 trials.\u003c/p\u003e\n\u003ch2\u003e3.3 Data analysis\u003c/h2\u003e\n\u003cp\u003eIn Experiment 2, we applied an exclusion criterion based on the range of the staircase. Participants who reached the minimum level of control of 2.5% in either condition (i.e., increase from 2.5% to 90% or decrease from 90% to 2.5%) were excluded from all analyses. After applying this criterion, we calculated d\u0026acute; for each participant and condition and then excluded participants with d\u0026acute; \u0026le; 0.5 and participants who used few confidence ratings, as in Experiment 1. For the remaining participants, we calculated meta-d\u0026acute; and m-ratio, and then we conducted paired \u003cem\u003et\u003c/em\u003e-tests and correlation analyses on all four metrics (d\u0026acute;, meta-d\u0026acute;, confidence rating, and m-ratio) and mean level of control, defined as the level of control reached through the staircase procedure in each condition. In addition, we conducted correlation analyses for the mean level of control, meta-d\u0026acute;, and m-ratio to assess whether these measures show similar individual differences across the increase and decrease conditions. For the meta-d\u0026acute; and m-ratio, we also estimated these values using hierarchical Bayesian modeling with HMeta-d as in Experiment 1.\u003c/p\u003e\n\u003ch2\u003e3.4 Results\u003c/h2\u003e\n\u003cp\u003eFirst, 23 participants were excluded based on the staircase range and confidence-rating criteria. In addition, 31 participants showed low sensitivity (d\u0026acute; \u0026le; 0.5) in at least one of the two conditions and were also excluded, resulting in a final sample of 56 participants for the subsequent analyses. Paired \u003cem\u003et\u003c/em\u003e-tests showed a significant difference in mean level of control between conditions (\u003cem\u003et\u003c/em\u003e(55) = 6.64, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, Cohen\u0026rsquo;s d = 0.89; Figure 4A and Figure 4B). As shown in Figure 4C, significant differences were also observed in d\u0026acute; (\u003cem\u003et\u003c/em\u003e(55) = 6.01, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, Cohen\u0026rsquo;s d = 0.80; Figure 4C left). The difference in meta-d\u0026acute; was relatively small but statistically significant (\u003cem\u003et\u003c/em\u003e(55) = 3.30, \u003cem\u003ep\u003c/em\u003e = .002, Cohen\u0026rsquo;s d = 0.44; Figure 4C right). In the confidence rating, there was no significant difference between conditions based on paired \u003cem\u003et\u003c/em\u003e-test (\u003cem\u003et\u003c/em\u003e(55) = 2.60, \u003cem\u003ep\u003c/em\u003e = .012, Cohen\u0026rsquo;s d = 0.35). \u003cstrong\u003eTo examine how consistently individuals performed across the two conditions, we conducted linear regression analyses between the decrease and increase in control conditions for the mean level of control and meta-d\u0026acute;.\u0026nbsp;\u003c/strong\u003eConsequently,\u0026nbsp;no significant correlation was found between conditions for meta-d\u0026acute; (\u003cem\u003er\u003c/em\u003e(54) = 0.17, \u003cem\u003ep\u003c/em\u003e = .21), but none of these reached statistical significance. Nevertheless, a significant positive correlation was found for the mean level of control (\u003cem\u003er\u003c/em\u003e(54) = 0.65, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001), indicating that the relative amount of change required for detection was fairly consistent across conditions even when detection performance and metacognitive measures showed little shared variance.\u003c/p\u003e\n\u003cp\u003eFigure 5A shows the distribution of m-ratio for the 56 participants. There was no significant difference between conditions (\u003cem\u003et\u003c/em\u003e(55) = 0.16, \u003cem\u003ep\u003c/em\u003e = .870, Cohen\u0026rsquo;s d = 0.02). Moreover, m-ratios across conditions were not significantly correlated (\u003cem\u003er\u003c/em\u003e(54) = .187, \u003cem\u003ep\u003c/em\u003e = .168; Figure 5B), and there were large individual differences as shown in Figure 5C. These findings suggest that, even when the task difficulty was equated among the participants, substantial individual differences remained in the metacognitive monitoring of control. Similar results were obtained when we applied a hierarchical Bayesian model (Supplementary Figure S2).\u003c/p\u003e"},{"header":"4 General Discussion","content":"\u003cp\u003eThe present study aimed to clarify whether perceiving an increase and a decrease in control relies on different perceptual processes, and whether these processes are monitored differently at metacognitive level. Participants controlled a dot on the screen for 5 seconds, and they first judged whether they experienced a change in control over the dot during the trial, and then rated their confidence in that response on a 4-point scale. The results of Experiment 1 showed that the perceptual sensitivity (d\u0026acute;) was significantly higher in the control decrease condition than the control increase condition, even though the magnitude of change was identical across conditions. Moreover, there was no significant correlation in d\u0026acute; between the two conditions, suggesting that the perceptual processes involved in detecting increases and decreases in control are likely to differ. In contrast, the m-ratio did not significantly differ between the conditions, indicating that the metacognitive monitoring of these perceptual processes is similar. Both the detection of increases and decreases in the sense of agency appear to be highly explicit. Furthermore, Experiment 2 used a staircase to hold detection accuracy as close as possible to 71%. The findings replicated the main results of Experiment 1: the m-ratio again did not significantly differ between the conditions, while achieving the target accuracy required a larger magnitude of change for detecting an increase in control compared with detecting a decrease in control.\u003c/p\u003e \u003cp\u003eSense of agency has typically been explained within prediction framework based on internal models. In this framework, the brain is thought to generate predicted sensory feedback from efference copies of motor commands and compares these predictions with the actual sensory feedback. Prediction errors are continuously monitored, but only those exceeding a certain threshold give rise to explicit awareness that control has been lost. This error-monitoring mechanism may be efficient when a stable mapping between one\u0026rsquo;s actions and sensory feedback has already been acquired and the belief of control is strong. In this case, even small prediction errors can trigger salient drop in sense of agency (Wen, Chang, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Wen, Mei, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Our results are consistent with this view. The detection of a decrease in control was significantly more sensitive than the detection of an increase in control, even when the magnitude of change was identical (Experiment 1). Moreover, participants were able to detect relatively smaller changes in control when control decreased from 90%, compared to when it increased to 90%. Such sensitive detection reflects a reliable error-detection mechanism.\u003c/p\u003e \u003cp\u003eIn contrast, when control increases from a lower level to a higher level, people must explore the regularity between their actions and the sensory feedback. In this case, prediction errors are less reliable and informative because the mapping between actions and outcome is uncertain. The process of detecting regularities between actions and outcomes is likely to be more dominant than the error-detection process in perceiving an increase in control. A previous study reported that the detection of an increase in control shows much larger individual differences than the detection of a decrease in control (Wen et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, it was unclear whether the large individual differences arose from the criterion used to determine whether they had gained sufficient control, or from the sensitivity in perceiving such a change. Our results show a similar trend of larger individual differences in the detection of an increase than the detection of a decrease in control, indicating that a large portion of the individual differences lines in sensitivity. This is likely because regularity detection is linked to motor learning (Nobusako et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Wen et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and different strategies and skills can be developed during such learning.\u003c/p\u003e \u003cp\u003eThis asymmetry in sensitivity of detecting an increase and a decrease in control mirrors findings from motor learning and speech studies. For example, sensitivity to error is modulated by environmental stability; it is higher in stable environments compared to random ones. Notably, within a stable environment, sensitivity drops at the onset of perturbation and subsequently grows as adaptation progresses (Todorov et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). In addition, a speech study showed stronger neural responses when auditory feedback deviates further from the learned distribution than when it moves closer to it (Tang et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Taken together, these findings suggest that the brain differentiates the source of variability, upregulating sensitivity when errors signal internal motor instability while downregulating it in response to external environmental noise. When control drops from a high level, the consequence can be critical in many circumstances, and a highly sensitive error-monitoring system may be important for survival. Such a system may be innate and thus show relatively small individual differences. Furthermore, the selection between error-detection processes and regularity-detection processes is thought to depend on the belief in control (Wen, Chang, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Wen, Mei, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). When the belief in control is strong, a mode called \u003cem\u003eexploitation\u003c/em\u003e is activated, and the prediction process dominates in the sense of agency (Wen, Mei, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). On the other hand, when the belief in control is weak, a mode called \u003cem\u003eexploration\u003c/em\u003e mode is activated, and the process of regularity detection dominates in the sense of agency (Wen, Mei, et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The range of control levels (i.e., 60%-90%) larges lies in the category of being in control (Wen et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This may explain the higher sensitivity in the detection of a decrease in control than in the detection of an increase. The mode of the sense of agency may switch depending on the belief in control, and thus the sensitivities in detecting an increase and a decrease in control may become comparable or even reverse in different circumstances. Nevertheless, our results show that the two types of processes underlying the sense of agency play very different roles in the detection of a change in control depending on the direction of the change.\u003c/p\u003e \u003cp\u003eThe present study also highlighted the metacognitive monitoring of the processes involved in detecting an increase and a decrease in control. The results of m-ratio did not differ between the two conditions, suggesting that the metacognitive monitoring is comparable between the two types of processes. The m-ratio was high in both experiments (Experiment 1: mean\u003csub\u003e\u003cem\u003e_inc\u003c/em\u003e\u003c/sub\u003e = 0.87, SD\u003csub\u003e\u003cem\u003e_inc\u003c/em\u003e\u003c/sub\u003e = 0.39, mean\u003csub\u003e\u003cem\u003e_dec\u003c/em\u003e\u003c/sub\u003e = 0.96, SD\u003csub\u003e\u003cem\u003e_dec\u003c/em\u003e\u003c/sub\u003e = 0.35; Experiment 2: mean\u003csub\u003e\u003cem\u003e_inc\u003c/em\u003e\u003c/sub\u003e = 0.79, SD\u003csub\u003e\u003cem\u003e_inc\u003c/em\u003e\u003c/sub\u003e = 0.65, mean\u003csub\u003e\u003cem\u003e_dec\u003c/em\u003e\u003c/sub\u003e = 0.80, SD\u003csub\u003e\u003cem\u003e_dec\u003c/em\u003e\u003c/sub\u003e = 0.44), indicating that the detection of a change in sense of agency is highly explicit. However, individual m-ratios did not correlate across the two conditions. This indicates that although the metacognitive efficiency was comparable at the group level, the processes underlying the metacognitive monitoring may still differ. Moreover, we observed that while the sensitivity (d\u0026acute;) was strongly supported by a conservative decision criterion (i.e., higher criteria is associated with higher d\u0026acute;), the m-ratio was independent of such decision biases (Supplementary Figure S3 and Figure S4). This suggests that the substantial between-participant variability we observed in m-ratio reflects genuine variations in the precision of the monitoring system, unconfounded by individual differences in individual strategy. For example, some participants showed high metacognitive efficiencies in both conditions, whereas others showed high metacognitive efficiency in one condition but poor metacognitive efficiency in the other condition. It remains unclear where the individual differences in metacognitive monitoring come from and how they shape the sense of agency. One possibility is that each individual may use the prediction process and the regularity-detection process differently in the sense of agency, and that metacognitive efficiency may be driven by the relative weightings of these processes. However, the present study was not designed to directly test this possibility, and it is worth further examination in future research.\u003c/p\u003e"},{"header":"5 Limitation","content":"\u003cp\u003eThis study has several limitations. First, all experiments were conducted online. While current web browsers allow us to obtain participants\u0026rsquo; screen resolution and refresh rate, they cannot capture the physical display size or the control-display ratio. Consequently, the physical parameters of the experimental stimuli could not be strictly controlled across participants. To mitigate the potential noise introduced by these variations, we recruited twice the number of participants required by the power analysis for each experiment. After exclusions, data from 34 participants in Experiment 1 and 56 in Experiment 2 were retained for the final analysis. Although the final sample size of Experiment 1 fell below the target determined by the power calculation, Experiment 2 met the required sample size and successfully replicated the findings of Experiment 1. This replication supports the robustness of our results despite the variations in individual experimental environments.\u003c/p\u003e \u003cp\u003eFurthermore, the number of trials per block in each experiment was relatively limited. Rahnev (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) suggests that at least 100 trials are typically required to accurately estimate metacognitive indices. In the present study, Experiment 1 included 60 trials per block, and Experiment 2 included 110 trials per block. To address the issue of limited trial counts, we employed hierarchical Bayesian modeling (Fleming, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), which allows for accurate parameter estimation even with fewer trials, in addition to standard estimation methods. The results from both approaches were consistent, supporting the robustness of our findings. Future research should nevertheless aim to increase the number of trials per condition to further enhance estimation accuracy.\u003c/p\u003e"},{"header":"6 Conclusion","content":"\u003cp\u003eThe present study shows that the contribution of the prediction process and the regularity detection process to the sense of agency are not simply two expressions of a single mechanism but instead rely on different perceptual processes. The metacognitive efficiency of these processes is comparable, but it remains unclear whether the processes involved in their metacognitive monitoring differ. Our findings refine current models of sense of agency by highlighting the importance of considering different perceptual processes across different modes of sense of agency and shed light on the large individual differences in both perceptual sensitivity and metacognitive monitoring in sense of agency.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe are grateful to Tae Morrisey for her assistance with the ethical application and coordination with the ethics committee, and to Kaori Yamashiro for her support in processing the online experiments.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eThis work was supported by JST FOREST Program (to WW; Grant Number JPMJFR2144), JST PRESTO (to SK; Grant Number JPMJPR23I4), JST Moonshot R\u0026amp;D Program (to WW and SK; Grant Number JPMJMS2013), and JSPS DC1 (to KT; Grant Number 25KJ2242).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest:\u0026nbsp;\u003c/strong\u003eThe authors declare that there is no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material:\u0026nbsp;\u003c/strong\u003eThe datasets in this study are available in the Open Science Framework repository, https://osf.io/u5qv7/\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability:\u0026nbsp;\u003c/strong\u003eThe scripts in this study are available in the Open Science Framework repository, https://osf.io/u5qv7/\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Approval:\u0026nbsp;\u003c/strong\u003eAll study were approved by the local ethics committee (Okinawa Institute of Science and Technology Graduate University, ethics number: #HSR-2024-022).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate:\u0026nbsp;\u003c/strong\u003eInformed consent was obtained from all individual participants included in the study.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eK.T., W.W., and T.F. conceived the study and defined the project direction. K.T. and W.W. designed the experiments. K.T. implemented the experimental environment, conducted the experiments, and analyzed the data. K.T. wrote the first draft of the manuscript with figures. W.W. contributed to the writing, editing, and critical revision of the manuscript. S.K. and T.F. contributed to the editing and critical revision of the manuscript. W.W., S.K., and T.F. supervised the project. All authors read and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eCastillo, J., Sieweyumptewa, P., \u0026amp; Phelps, E. A. (2024). Differential effects of negative valence and memory type on accuracy, confidence, and metacognitive efficiency. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1), 25685.\u003c/li\u003e\n \u003cli\u003eCharalampaki, A., Peters, C., Maurer, H., Maurer, L. K., M\u0026uuml;ller, H., Verrel, J., \u0026amp; Filevich, E. (2023). Motor outcomes congruent with intentions may sharpen metacognitive representations. \u003cem\u003eCognition\u003c/em\u003e, \u003cem\u003e235\u003c/em\u003e(105388), 105388.\u003c/li\u003e\n \u003cli\u003eChung, Y. H., Tam, J., Wyble, B., \u0026amp; St\u0026ouml;rmer, V. S. (2023). Object meaningfulness increases incidental memory of shape but not location. In \u003cem\u003ePsyArXiv\u003c/em\u003e. https://doi.org/10.31234/osf.io/th2rm\u003c/li\u003e\n \u003cli\u003eFleming, S. M. (2017). HMeta-d: hierarchical Bayesian estimation of metacognitive efficiency from confidence ratings. \u003cem\u003eNeuroscience of Consciousness\u003c/em\u003e, \u003cem\u003e2017\u003c/em\u003e(1), nix007.\u003c/li\u003e\n \u003cli\u003eFleming, S. M., \u0026amp; Lau, H. C. (2014). How to measure metacognition. \u003cem\u003eFrontiers in Human Neuroscience\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e, 443.\u003c/li\u003e\n \u003cli\u003eGallagher, I., I. (2000). Philosophical conceptions of the self: implications for cognitive science. \u003cem\u003eTrends in Cognitive Sciences\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(1), 14\u0026ndash;21.\u003c/li\u003e\n \u003cli\u003eGreen, D. M., \u0026amp; Swets, J. A. (1966). \u003cem\u003eSignal detection theory and psychophysics\u003c/em\u003e. John Wiley \u0026amp; Sons.\u003c/li\u003e\n \u003cli\u003eHaggard, P. (2017). Sense of agency in the human brain. \u003cem\u003eNature Reviews. Neuroscience\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(4), 196\u0026ndash;207.\u003c/li\u003e\n \u003cli\u003eHaggard, P., \u0026amp; Chambon, V. (2012). Sense of agency. \u003cem\u003eCurrent Biology: CB\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(10), R390-2.\u003c/li\u003e\n \u003cli\u003eHaggard, P., \u0026amp; Tsakiris, M. (2009). The experience of agency: Feelings, judgments, and responsibility. \u003cem\u003eCurrent Directions in Psychological Science\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(4), 242\u0026ndash;246.\u003c/li\u003e\n \u003cli\u003eKarsh, N., \u0026amp; Eitam, B. (2015). I control therefore I do: judgments of agency influence action selection. \u003cem\u003eCognition\u003c/em\u003e, \u003cem\u003e138\u003c/em\u003e(C), 122\u0026ndash;131.\u003c/li\u003e\n \u003cli\u003eKasahara, S., \u0026amp; Takada, K. (2021). Stealth Updates of Visual Information by Leveraging Change Blindness and Computational Visual Morphing. \u003cem\u003eACM Transactions on Applied Perception\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(4), 1\u0026ndash;17.\u003c/li\u003e\n \u003cli\u003eKawato, M. (1999). Internal models for motor control and trajectory planning. \u003cem\u003eCurrent Opinion in Neurobiology\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(6), 718\u0026ndash;727.\u003c/li\u003e\n \u003cli\u003eManiscalco, B., \u0026amp; Lau, H. (2012). A signal detection theoretic approach for estimating metacognitive sensitivity from confidence ratings. \u003cem\u003eConsciousness and Cognition\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(1), 422\u0026ndash;430.\u003c/li\u003e\n \u003cli\u003eMcNamee, D., \u0026amp; Wolpert, D. M. (2019). Internal Models in Biological Control. \u003cem\u003eAnnual Review of Control, Robotics, and Autonomous Systems\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e, 339\u0026ndash;364.\u003c/li\u003e\n \u003cli\u003eMiyazaki, M., Takahashi, H., Rolf, M., Okada, H., \u0026amp; Omori, T. (2014). The image-scratch paradigm: a new paradigm for evaluating infants\u0026rsquo; motivated gaze control. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(1), 5498.\u003c/li\u003e\n \u003cli\u003eNobusako, S., Wen, W., Nagakura, Y., Tatsumi, M., Kataoka, S., Tsujimoto, T., Sakai, A., Yokomoto, T., Takata, E., Furukawa, E., Asano, D., Osumi, M., Nakai, A., \u0026amp; Morioka, S. (2022). Developmental changes in action-outcome regularity perceptual sensitivity and its relationship to hand motor function in 5-16-year-old children. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(1), 17606.\u003c/li\u003e\n \u003cli\u003eOancea, G., Maniscalco, B., Peters, M. A. K., \u0026amp; Chapman, C. S. (2025). Measuring motor awareness and metacognition at the start, middle, and end of a reaching movement. \u003cem\u003eConsciousness and Cognition\u003c/em\u003e, \u003cem\u003e132\u003c/em\u003e, 103878.\u003c/li\u003e\n \u003cli\u003eOhata, R., Asai, T., Kadota, H., Shigemasu, H., Ogawa, K., \u0026amp; Imamizu, H. (2020). Sense of Agency Beyond Sensorimotor Process: Decoding Self-Other Action Attribution in the Human Brain. \u003cem\u003eCerebral Cortex\u0026nbsp;\u003c/em\u003e, \u003cem\u003e30\u003c/em\u003e(7), 4076\u0026ndash;4091.\u003c/li\u003e\n \u003cli\u003ePereira, M., Skiba, R., Cojan, Y., Vuilleumier, P., \u0026amp; B\u0026egrave;gue, I. (2023). Preserved metacognition for undetected visuomotor deviations. \u003cem\u003eThe Journal of Neuroscience: The Official Journal of the Society for Neuroscience\u003c/em\u003e, \u003cem\u003e43\u003c/em\u003e(35), 6176\u0026ndash;6184.\u003c/li\u003e\n \u003cli\u003eRahnev, D. (2025). A comprehensive assessment of current methods for measuring metacognition. \u003cem\u003eNature Communications\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(1), 701.\u003c/li\u003e\n \u003cli\u003eReyes, M., Morales, M. J., \u0026amp; Bajo, M. T. (2023). Judgments of learning in bilinguals: Does studying in a L2 hinder learning monitoring? \u003cem\u003ePloS One\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(12), e0286516.\u003c/li\u003e\n \u003cli\u003eRovee, C. K., \u0026amp; Rovee, D. T. (1969). Conjugate reinforcement of infant exploratory behavior. \u003cem\u003eJournal of Experimental Child Psychology\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(1), 33\u0026ndash;39.\u003c/li\u003e\n \u003cli\u003eTanaka, T., \u0026amp; Imamizu, H. (2025). Sense of agency for a new motor skill emerges via the formation of a structural internal model. \u003cem\u003eCommunications Psychology\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(1), 70.\u003c/li\u003e\n \u003cli\u003eTang, D.-L., Parrell, B., Beach, S. D., \u0026amp; Niziolek, C. A. (2025). The brain\u0026rsquo;s sensitivity to sensory error can be modulated by altering perceived variability. \u003cem\u003eThe Journal of Neuroscience: The Official Journal of the Society for Neuroscience\u003c/em\u003e, \u003cem\u003e45\u003c/em\u003e(5), e0024242024.\u003c/li\u003e\n \u003cli\u003eTodorov, D. I., Moulin, M., Buch, E., \u0026amp; Quentin, R. (2025). Modulation of error sensitivity during motor learning across time, space and environment variability. In \u003cem\u003ebioRxivorg\u003c/em\u003e (p. 2025.05.30.656374). https://doi.org/10.1101/2025.05.30.656374\u003c/li\u003e\n \u003cli\u003eVigh, G., \u0026amp; Limanowski, J. (2024). Baseline dependent differences in the perception of changes in visuomotor delay. \u003cem\u003eFrontiers in Human Neuroscience\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e, 1495592.\u003c/li\u003e\n \u003cli\u003eVilla, R., Tidoni, E., Porciello, G., \u0026amp; Aglioti, S. M. (2018). Violation of expectations about movement and goal achievement leads to Sense of Agency reduction. \u003cem\u003eExperimental Brain Research\u003c/em\u003e, \u003cem\u003e236\u003c/em\u003e(7), 2123\u0026ndash;2135.\u003c/li\u003e\n \u003cli\u003eVilla, R., Tidoni, E., Porciello, G., \u0026amp; Aglioti, S. M. (2021). Freedom to act enhances the sense of agency, while movement and goal-related prediction errors reduce it. \u003cem\u003ePsychological Research\u003c/em\u003e, \u003cem\u003e85\u003c/em\u003e(3), 987\u0026ndash;1004.\u003c/li\u003e\n \u003cli\u003eWang, Q., Bolhuis, J., Rothkopf, C. A., Kolling, T., Knopf, M., \u0026amp; Triesch, J. (2012). Infants in control: rapid anticipation of action outcomes in a gaze-contingent paradigm. \u003cem\u003ePloS One\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(2), e30884.\u003c/li\u003e\n \u003cli\u003eWen, W., Chang, A. Y.-C., \u0026amp; Imamizu, H. (2024). The sensitivity and criterion of sense of agency. \u003cem\u003eTrends in Cognitive Sciences\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e(5), 397\u0026ndash;399.\u003c/li\u003e\n \u003cli\u003eWen, W., Charles, L., \u0026amp; Haggard, P. (2023). Metacognition and sense of agency. \u003cem\u003eCognition\u003c/em\u003e, \u003cem\u003e241\u003c/em\u003e(105622), 105622.\u003c/li\u003e\n \u003cli\u003eWen, W., \u0026amp; Haggard, P. (2018). Control Changes the Way We Look at the World. \u003cem\u003eJournal of Cognitive Neuroscience\u003c/em\u003e, \u003cem\u003e30\u003c/em\u003e(4), 603\u0026ndash;619.\u003c/li\u003e\n \u003cli\u003eWen, W., \u0026amp; Haggard, P. (2020). Prediction error and regularity detection underlie two dissociable mechanisms for computing the sense of agency. \u003cem\u003eCognition\u003c/em\u003e, \u003cem\u003e195\u003c/em\u003e, 104074.\u003c/li\u003e\n \u003cli\u003eWen, W., \u0026amp; Imamizu, H. (2022). The sense of agency in perception, behaviour and human\u0026ndash;machine interactions. \u003cem\u003eNature Reviews Psychology\u003c/em\u003e, \u003cem\u003e1\u003c/em\u003e(4), 211\u0026ndash;222.\u003c/li\u003e\n \u003cli\u003eWen, W., Ishii, H., Ohata, R., Yamashita, A., Asama, H., \u0026amp; Imamizu, H. (2021). Perception and control: individual difference in the sense of agency is associated with learnability in sensorimotor adaptation. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1), 20542.\u003c/li\u003e\n \u003cli\u003eWen, W., Mei, J., Aktas, H., Chang, A. Y.-C., Suzuishi, Y., \u0026amp; Kasahara, S. (2024). Control over self and others\u0026rsquo; face: exploitation and exploration. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1), 15473.\u003c/li\u003e\n \u003cli\u003eWen, W., Okon, Y., Yamashita, A., \u0026amp; Asama, H. (2022). The over-estimation of distance for self-voice versus other-voice. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(1), 420.\u003c/li\u003e\n \u003cli\u003eWen, W., Shimazaki, N., Ohata, R., Yamashita, A., Asama, H., \u0026amp; Imamizu, H. (2020). Categorical Perception of Control. \u003cem\u003eENeuro\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(5). https://doi.org/10.1523/ENEURO.0258-20.2020\u003c/li\u003e\n \u003cli\u003eWolpert, D. M., Ghahramani, Z., \u0026amp; Jordan, M. I. (1995). An internal model for sensorimotor integration. \u003cem\u003eScience\u003c/em\u003e, \u003cem\u003e269\u003c/em\u003e(5232), 1880\u0026ndash;1882.\u003c/li\u003e\n \u003cli\u003eZama, T., Takahashi, Y., \u0026amp; Shimada, S. (2017). The effects of trajectory and endpoint errors in a reaching movement on the sense of agency. \u003cem\u003ePsychology\u0026nbsp;\u003c/em\u003e, \u003cem\u003e08\u003c/em\u003e(14), 2321\u0026ndash;2332.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e Prolific. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.prolific.com/\u003c/span\u003e\u003cspan address=\"https://www.prolific.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e metaSDT. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/craddm/metaSDT\u003c/span\u003e\u003cspan address=\"https://github.com/craddm/metaSDT\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e HMeta-d. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/metacoglab/HMeta-d\u003c/span\u003e\u003cspan address=\"https://github.com/metacoglab/HMeta-d\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":true,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"sense of agency, metacognition, prediction error, regularity detection, m-ratio","lastPublishedDoi":"10.21203/rs.3.rs-8589893/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8589893/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSense of agency is the feeling of control over one\u0026rsquo;s actions in the environment. This is typically explained in terms of a comparator that evaluates the consistency between sensorimotor predictions and actual outcomes. Under high movement controllability, there is high predictability and hence a strong sense of agency, while a decrease in control produces salient prediction errors. Conversely, under low controllability, the sense of agency may rely more on detecting regularities between actions and outcomes until control increases. However, it remains unclear whether these distinct perceptual processes share the same metacognitive monitoring process. We addressed this question using a control change detection task, where participants moving a single dot on a screen had to detect whether their level of control had changed and report their confidence. Across two experiments, we observed that perceptual sensitivity was higher for decreases than for increases in controllability, but metacognitive process showed no directional difference. Our findings suggest that while distinct perceptual processes are involved for different levels of controllability, metacognitive monitoring shares a common underlying mechanism.\u003c/p\u003e","manuscriptTitle":"Perceptual Sensitivity, but not Metacognitive Monitoring, is Dependent on Varying Levels of Control","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-29 10:37:30","doi":"10.21203/rs.3.rs-8589893/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c753df58-ccd3-4b19-accc-ae8fa50428e3","owner":[],"postedDate":"January 29th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-05-04T16:05:44+00:00","versionOfRecord":{"articleIdentity":"rs-8589893","link":"https://doi.org/10.1007/s00221-026-07306-w","journal":{"identity":"experimental-brain-research","isVorOnly":false,"title":"Experimental Brain Research"},"publishedOn":"2026-04-29 15:58:36","publishedOnDateReadable":"April 29th, 2026"},"versionCreatedAt":"2026-01-29 10:37:30","video":"","vorDoi":"10.1007/s00221-026-07306-w","vorDoiUrl":"https://doi.org/10.1007/s00221-026-07306-w","workflowStages":[]},"version":"v1","identity":"rs-8589893","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8589893","identity":"rs-8589893","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}