Dissociating the External Manifestation and Internal Representation of Volition

LUO Xiaoxiao¹, ZHOU Xiaolin²,³

¹ Faculty of Education, Yunnan Normal University, Kunming 650500, China
² Shanghai Key Laboratory of Mental Health and Psychological Crisis Intervention, Shanghai 200062, China
³ School of Psychology and Cognitive Science, East China Normal University, Shanghai 200062, China

Abstract

Volition is the ability to voluntarily control oneself, a core characteristic that distinguishes humans from animals and machines, and the cornerstone of physical and mental health and social order. Its external manifestation is voluntary action, and its internal representation is control belief. The former refers to actions generated based on one's own will; the latter refers to the belief that the implementation of voluntary actions can influence external events. Most previous studies have investigated volition based on voluntary action, yet individuals simultaneously hold control beliefs when performing voluntary actions. Consequently, relevant research has conflated the external manifestation and internal representation of volition. This study proposes to employ the volition-motivated performance (VMP) paradigm to dissociate voluntary action from control belief, combined with computational modeling and multimodal neuroimaging techniques (electromyography/electroencephalography/functional magnetic resonance imaging), to systematically reveal their shared and distinct cognitive neural mechanisms. Based on this, we propose the "dual-path hypothesis of human volition processing": one is the voluntary action-related path, reflecting the action attribute of volition; the other is the control belief-related path, reflecting the motivational attribute of volition.

Keywords: volition, voluntary action, control belief, voluntary choice, volition-motivated performance paradigm
Classification Code: B842

Human inquiry into volition has a long history, spanning philosophy, psychology, neuroscience, artificial intelligence, and other fields both ancient and emerging, yet no consensus has been reached to date. Although researchers from different domains define volition differently, their discussions generally center on "the ability of individuals to choose or decide autonomously" (Kane, 1996). At the individual level, volition is the cornerstone of human physical and mental health; the inability to control one's own behavior is a typical manifestation of various physical and mental diseases, such as Parkinson's disease (Ricciardi et al., 2017) and schizophrenia (Daprati et al., 1997). At the group level, volition is even more fundamental to human society as a whole; for instance, a basic premise of lawmaking is that people should be held responsible for actions performed based on their own volition (Christensen et al., 2024; Haggard, 2019).

Volition generally refers to the ability to voluntarily control oneself, particularly in controlling goal-directed actions (Haggard, 2008, 2019): that is, the psychological and behavioral mechanism by which individuals actively initiate, maintain, or adjust actions to achieve goals (Frith, 2013; Kuhl, 1984). A complete volitional process includes both the generation and expression of volition, both of which are indispensable. On the one hand, if volition remains only in the internal generation process without being expressed through external behavior, it loses the core characteristic of volition—"teleology" (Haggard, 2019)—the notion that volition emerges to achieve specific purposes. Without external behavior, no purpose can be realized. On the other hand, if volition is reduced to external actions without an internal generation process, it also loses another core characteristic of volition—"subjectivity" (Haggard, 2019)—the subjective experience that "I can plan and control actions to achieve goals, and these actions reflect my own intentions." Without an internal processing mechanism, any subjective experience would be impossible. Thus, the generation of volition corresponds to the individual's internal "control belief"—the belief that one's own actions can influence external events and achieve goals/intentions—whereas the expression of volition corresponds to the individual's external "voluntary action"—actions initiated based on one's own will. Only through external actions that influence external events can goals/intentions be achieved. Voluntary action is the external manifestation of volition, while control belief is its internal representation.

In the volitional process, voluntary action and control belief are closely related—individuals need to possess control beliefs (i.e., believe that their actions are meaningful and can influence external events to achieve goals/intentions) to be inclined to perform voluntary actions. Conversely, when individuals lose control beliefs (i.e., believe that performing actions is meaningless and doubt that their actions can produce any effect), voluntary actions decrease or disappear, resulting in "learned helplessness" (Huys & Dayan, 2009; Maier & Seligman, 1976). Therefore, when individuals perform voluntary actions, they often simultaneously hold control beliefs (Desantis et al., 2011; Dogge et al., 2012; Huys & Dayan, 2009; Luo et al., 2022; Maier & Seligman, 1976).

Although the external manifestation and internal representation of volition are closely related, research suggests that subjective experience and actual actions in the volitional process are dissociable. For example, direct stimulation of specific brain regions can induce in individuals the subjective experience of "wanting to move" without any actual motor output (Desmurget et al., 2009; Fried et al., 1991). This indicates that the external manifestation and internal representation of volition may involve different processing mechanisms. By reviewing previous studies (see "1 Research Status" for details), we find that because the internal representation process is difficult to directly measure or manipulate, existing research has predominantly investigated volition based on its external manifestation—immediately executed voluntary actions—without distinguishing between voluntary action and control belief. Consequently, relevant findings may be caused by voluntary action alone, control belief alone, or the combined effect of both. Therefore, previous studies have failed to further dissect volition, and their findings may have confounded the roles of voluntary action and control belief (see "2 Problem Statement" for details). Thus, attempting to dissociate voluntary action and control belief through non-invasive experimental design represents an important question worthy of further investigation in the field of human volition research.

This study proposes to develop variations of a recently developed paradigm for investigating volitional effects—the volition-motivated performance (VMP) paradigm (Luo et al., 2022; see also Luo, Wang, Gu, et al., 2024; Luo, Wang, & Zhou, 2024; for a detailed introduction to the VMP paradigm, see "1.6 Volition-Motivated Performance Paradigm")—to dissociate the external manifestation (voluntary action) and internal representation (control belief) within volition. Simultaneously, we will combine behavioral computational modeling and multiple neurophysiological techniques, including electromyography (EMG), electroencephalography (EEG), and functional magnetic resonance imaging (fMRI), to reveal the underlying cognitive neural mechanisms.

This study holds both theoretical and practical significance. Theoretically, conducting a multi-faceted and multi-angle analysis of human volition based on experimental design, computational models, and neurophysiological techniques will help clarify the common and unique roles of different components of volition, open new perspectives for investigating volitional questions, and thereby deepen our understanding of human volition and even the problem of consciousness. Clinically, "control belief deficiency" is closely associated with certain mental health issues, such as depression (Huys & Dayan, 2009), learned helplessness (Maier & Seligman, 1976), and schizophrenia (Daprati et al., 1997). By isolating control belief from the volitional process, this study may provide new perspectives for research and treatment of related mental health problems. Industrially, control belief plays an important role in interactions between individuals and virtual environments or intelligent machines (Wen & Imamizu, 2022). People need to hold control beliefs about virtual environments (or intelligent machines) to complete subsequent tasks, and the feedback from these environments (or machines) in turn directly affects individuals' control beliefs. Investigating control belief and its neural mechanisms will facilitate the future development of virtual environments or intelligent machines with better control experiences and provide a foundation for artificial intelligence to simulate human control processes.

1 Research Status

Volition has no physical substance and is difficult to directly measure or manipulate. Previous research has developed various paradigms to investigate different aspects of volition (see Figure 1 [FIGURE:1] for a summary of paradigms). Among these, voluntary action, as the external manifestation of volition, is easy to observe and measure, and has been extensively studied. Researchers have primarily focused on two aspects: the generation of voluntary action and the subjective experience of voluntary action ("sense of agency"). Related paradigms mainly include self-generated action, explicit judgment of agency, and implicit measurement of agency. Control belief, as the internal representation of volition, is difficult to measure and quantify, and has received less research attention. Related paradigms mainly include explicit judgment of control, motivation from control paradigm, and volition-motivated performance paradigm. However, whether focusing on voluntary action or control belief, existing paradigms all rely on immediately executed voluntary actions. As discussed above, individuals performing voluntary actions simultaneously hold control beliefs; therefore, previous studies have failed to distinguish the respective roles of voluntary action and control belief. The following sections will briefly summarize these paradigms and further demonstrate potential problems in existing research.

1.1 Self-Generated Action Paradigm

When investigating the generation of voluntary action, researchers often require participants to self-generate actions at arbitrary time points while simultaneously recording neural activity. Sometimes these are compared with non-voluntarily generated actions (e.g., twitches induced by transcranial magnetic stimulation, actions produced in response to external stimuli). For example, researchers using EEG technology have confirmed that voluntary action generation is preceded by a readiness potential (RP) (Khalighinejad et al., 2018; Libet et al., 1983). Using fMRI technology, they have identified brain regions related to voluntary action planning and execution, such as the supplementary motor area (SMA) and premotor cortex (PMC) (Penfield, 1954; Rizzolatti & Kalaska, 2013). In addition to recording neural activity, researchers have also examined brain damage to infer the neural basis for voluntary action generation. For instance, SMA damage affects the generation of spontaneous actions but not stimulus-driven actions, whereas the opposite pattern is observed with PMC damage (Passingham, 1987), suggesting the important role of SMA in voluntary action generation.

Brain-computer interface research has found that when actions are induced using neuromuscular electrical stimulation, the peak timing of neuronal activity in the primary motor cortex (M1) coincides with the time when individuals subjectively experience the emergence of motor intention (Noel et al., 2025). Furthermore, low-frequency alpha oscillations in M1 prior to action generation can predict individuals' sense of agency (Bertoni et al., 2025). This suggests that M1 not only plays an important role in voluntary action generation but is also related to the subjective experience of voluntary action.

However, investigating volition through the direct generation of voluntary actions inevitably involves performing voluntary actions, during which individuals simultaneously hold control beliefs. Therefore, this paradigm cannot dissociate voluntary action from control belief.

1.2 Explicit Judgment of Agency Paradigm

Researchers often refer to the subjective experience of voluntary action as "sense of agency," "sense of motor control," or "sense of active control" (Haggard, 2017; Tian et al., 2018; Wu et al., 2019; Zhang et al., 2018)—that is, "the experience generated during the execution of voluntary actions of feeling that one can control one's own actions and thereby control external events" (Haggard, 2017; Haggard et al., 2002). Sense of agency accompanies the execution of voluntary actions and can be subjectively reported through introspection, constituting the explicit judgment of agency paradigm. Related studies often require participants to report the extent to which an action was performed by "myself" or was under their control. Higher subjective report scores indicate stronger sense of agency for that action and correspondingly stronger volitional experience (e.g., Georgieff & Jeannerod, 1998; Sirigu et al., 1999).

However, when an action is consistent with a participant's own actions, participants can experience a sense of agency even when they are not the agent producing the action (Tsakiris et al., 2005; Wegner & Wheatley, 1999). This suggests that individuals' subjective reports tend to overestimate their own agentic capacity, attributing actions that are actually unrelated to their own volition to themselves. Moreover, requiring individuals to introspect about the subjective experience of voluntary action also cannot be separated from the execution of voluntary action, making it impossible to dissociate voluntary action from control belief. Therefore, explicitly reported sense of agency often serves as an auxiliary rather than primary measure in volition research and is not suitable for dissociating voluntary action and control belief based on this paradigm.

1.3 Implicit Measurement of Agency Paradigm

To overcome the limitations of subjective reporting methods for sense of agency, researchers have developed paradigms for implicitly measuring agency, the most commonly used being the "intentional binding task" (Haggard, 2017; Haggard et al., 2002). In this task, participants perform an action (e.g., key press), after which an event occurs (e.g., a sound stimulus). Participants must estimate the time interval between the action and the event (or the timing of both). Results show that compared with non-voluntary actions, when a voluntary action is performed, participants' time estimates are shorter. This phenomenon is called the "intentional binding effect" or "temporal compression effect." The magnitude of this effect reflects the strength of individuals' sense of agency and has been replicated in numerous studies (e.g., Moore et al., 2009; Obhi & Hall, 2011; Schwarz et al., 2019). Related neuroscience research has also found that the degree of SMA activation can predict the strength of the intentional binding effect (Kühn et al., 2013). Using transcranial magnetic stimulation (TMS) or transcranial direct current stimulation (tDCS) to disrupt SMA activity in healthy participants weakens the intentional binding effect (Moore et al., 2010; Cavazzana et al., 2015). These findings again highlight the important association between SMA and human volitional processes.

In addition to the intentional binding effect, sensory attenuation (Blakemore et al., 1998; Schafer & Marcus, 1973) has also been used as an implicit measure of agency. Similar to the intentional binding task, in sensory attenuation tasks, participants perform an action followed by a sensory stimulus (e.g., visual, auditory, or tactile stimuli), and they must judge the intensity of the sensory stimulus. Results show that compared with non-voluntary actions, when a voluntary action is performed, participants' perception of stimulus intensity decreases, with reduced amplitude of related ERP components (e.g., N1) (e.g., Mifsud et al., 2018; Weiss et al., 2011) and decreased activation in related primary sensory cortices (e.g., auditory and somatosensory areas) (e.g., Arikan et al., 2021; Reznik et al., 2014)—that is, sensory attenuation occurs. The magnitude of this effect is also considered to reflect the strength of sense of agency. It should be noted that the relationship between sensory attenuation and intentional binding effects remains debated; although both are related to sense of agency, they may involve different cognitive processing mechanisms (Borhani et al., 2017; Lindner et al., 2025).

Whether intentional binding effects or sensory attenuation effects, their key measures—the time interval between action and subsequent events, and judgments of sensory intensity triggered by actions—are directly associated with the execution of voluntary actions. Researchers still cannot dissociate the effects of voluntary action and control belief.

1.4 Explicit Judgment of Control Paradigm

Previous research exploring control belief has also been inseparable from the execution of voluntary actions, though the approach has differed. Researchers have frequently used subjective reporting methods to investigate control belief, known as explicit judgment of control. Related studies often present specific stimuli as "outcomes" while participants perform voluntary actions and require participants to report the extent to which the "outcome" was caused by their own actions. Research consistently finds that individuals tend to judge "outcomes" that are temporally and spatially closer to voluntary actions as being caused by those actions, even when such causal relationships do not actually exist. This phenomenon is also called the "illusion of control" (Langer, 1975; Thompson et al., 1998; Chen et al., 2010). In addition to directly rating the causal relationship between specific outcomes and actions, one variation of explicit judgment of control involves having participants voluntarily generate a series of actions (e.g., moving a mouse or joystick) while judging whether and to what extent stimuli on the screen (e.g., movement of a light dot) are consistent with their own actions (e.g., judging whether the trajectory of the light dot on the screen matches their own mouse movement trajectory). Studies show that when the inconsistency between stimuli and actions increases in time or space, participants are more likely to perceive the stimulus as not under their control (Applebaum et al., 2025; Wen et al., 2023; Wen & Haggard, 2020), indicating weakened control belief.

However, similar to subjective reports of sense of agency, the presence of the control illusion means that individuals' subjective reports tend to overestimate the controllability of the external environment. Moreover, if the "outcome" is too distant from the voluntary action in time or space, individuals are not inclined to establish a connection between the outcome and the action (Matute & Blanco, 2014; Yarritu et al., 2014). Therefore, explicit judgment of control is also more suitable as an auxiliary measure, and it similarly cannot be separated from immediately executed voluntary actions, making it impossible to isolate control belief.

1.5 Motivation from Control Paradigm

The motivation from control paradigm attempts to measure control belief through implicit means. Similar to the explicit judgment of control paradigm described above, this paradigm presents specific stimuli while participants perform voluntary actions and manipulates the presentation manner of these stimuli (i.e., manipulating the "outcome" of actions), but does not require participants to provide subjective reports. Instead, it records measures such as reaction time and frequency of voluntary actions themselves. Research finds that when "outcomes" are temporally and spatially closer to voluntary actions, participants' voluntary actions become faster and more frequent (Eitam et al., 2013; Karsh et al., 2016), consistent with findings from subjective reporting methods and interpreted as effects of control belief. Further research based on this paradigm has found that the effectiveness of "outcomes" (e.g., manipulating whether an "outcome" is produced or the predictability of the "outcome") also affects participants' reaction times for voluntary actions, and this influence changes dynamically with the effectiveness of the "outcome" (Hemed et al., 2020). This suggests that individuals' control beliefs are dynamically adjusted in real-time based on external feedback, and such dynamic adjustment of control belief may further influence subsequent voluntary action execution. Similarly, individuals can dynamically adjust the bias of autonomous behavior based on feedback and belief learning to prevent autonomous behavior from being influenced by habitual responses or sequential effects. However, this adjustment is also limited, and individuals find it difficult to eliminate the influence of action-outcome dependence (Ota et al., 2024). This indicates the important role of the "action-outcome" relationship in forming control beliefs (see also Luo et al., 2022).

However, evaluating control belief using measures related to voluntary actions themselves (reaction time, frequency) still requires immediate execution of voluntary actions and cannot dissociate voluntary action from control belief.

1.6 Volition-Motivated Performance (VMP) Paradigm

The VMP paradigm employs a more purposeful type of voluntary action—voluntary choice (Leotti et al., 2010; Chen & Wu, 2019)—to investigate volition. Voluntary choice provides the freedom to make decisions among different options and is an actively generated goal-directed action (specifying options based on one's own will). However, it should be noted that many factors may drive individuals to engage in voluntary choice behavior, with volition being only one possibility. Therefore, in the VMP paradigm, voluntary choice must be compared with "forced choice"—a non-selective voluntary action where options are directly specified and participants are required to perform the corresponding key press—to reflect the role of volition. Compared with forced choice, voluntary choice is associated with a stronger sense of agency (Barlas et al., 2018; Caspar et al., 2016; Yavuz et al., 2025). More importantly, compared with forced choice, voluntary choice facilitates subsequent cognitive performance, including time estimation tasks (Murayama et al., 2015), declarative memory tasks (Murty et al., 2015; Ruiz et al., 2023), and reaction time-related tasks (e.g., visual search tasks: Luo et al., 2022; Luo, Wang, Gu, et al., 2024; conflict control tasks: Legault & Inzlicht, 2013; Luo, Wang, & Zhou, 2024). The VMP paradigm investigates the role of volition based on the differential impact of voluntary versus forced choice on subsequent cognitive performance, using the difference in cognitive performance (e.g., task reaction time) between voluntary and forced choice conditions as an index reflecting volitional effects (Luo et al., 2022; see also Luo, Wang, Gu, et al., 2024; Luo, Wang, & Zhou, 2024).

The typical trial procedure of the VMP paradigm is shown in Figure 2 [FIGURE:2]. A complete trial comprises three phases: cue phase, choice phase, and task phase. Specifically, the cue phase uses different colors to indicate whether the upcoming trial involves voluntary or forced choice. In the choice phase, two neutral images are presented. In the voluntary choice condition, participants can voluntarily press a key to select one image, whereas in the forced choice condition, one image is pre-specified (surrounded by a box), and participants can only press the key to select the designated image. In the task phase, the selected image is presented as a task-irrelevant background against which participants complete a cognitive task (e.g., visual search task). Studies based on this paradigm have found that compared with forced choice, cognitive performance following voluntary choice is facilitated (faster reaction times) (Luo et al., 2022; Luo, Wang, Gu, et al., 2024; Luo, Wang, & Zhou, 2024), and the magnitude of control belief reported subjectively by participants can predict the magnitude of the facilitative effect of voluntary choice on cognitive performance (Luo et al., 2022). This suggests that cognitive performance indices in the VMP paradigm can, to some extent, reflect individuals' control beliefs.

Figure 2. Typical trial procedure of the volition-motivated performance (VMP) paradigm. Cue phase: Circles of different colors indicate the choice type for the current trial (voluntary choice vs. forced choice). Choice phase: Voluntarily or forcibly press a key to select one of two images. Task phase: Complete a cognitive task (visual search task here) with the selected image as background. Note: There is a fixation point lasting 0.5–0.8 s between different phases (cue, choice, task phases). For simplicity, these fixation points are not shown in the flowchart. Source: Fig. 1 in Luo, Wang, Gu, et al. (2024). Color figures are available in the electronic version of this article.

However, the key index reflecting volitional effects in the current VMP paradigm (cognitive performance after voluntary choice) is still measured immediately after executing the voluntary action (choice), failing to distinguish the respective effects of voluntary action and control belief. Nevertheless, the advantage of this paradigm over those described above is that the measurement index (cognitive performance) is completely unrelated to the voluntary action (choice), and the two can be completely separated in time. In contrast, the measurement indices in the aforementioned paradigms are directly related to voluntary actions and cannot be separated. The separability between voluntary action and measurement index in the VMP paradigm provides a foundation for dissociating voluntary action and control belief through experimental design (as will be further elaborated in "2 Problem Statement" below).

2 Problem Statement

As can be seen from the summary of previous paradigms above, these approaches to investigating human volition from different angles are all directly associated with immediately executed voluntary actions. Considering that individuals simultaneously hold control beliefs when performing voluntary actions (Desantis et al., 2011; Dogge et al., 2012; Huys & Dayan, 2009; Luo et al., 2022; Maier & Seligman, 1976), research using these paradigms inevitably involves both the processing of voluntary actions and the processing of control beliefs. Therefore, previous studies have failed to dissociate voluntary action and control belief. Most studies have generally categorized related behavioral or neural results as "effects of voluntary action" or "effects of volition," when their findings may be due to voluntary action alone, control belief alone, or the combined effect of both. In other words, in several classic volition research paradigms, the dependent variable measuring volitional effects is closely related to the execution of voluntary actions, and the measurement of the index cannot be temporally separated from the execution of voluntary actions. This may be an important factor limiting researchers' ability to dissociate voluntary action and control belief. Moreover, previous volition research has primarily focused on voluntary action itself. Although some classic computational models have explored the generation of voluntary action (e.g., Blakemore et al., 2000), these models are more action-oriented, and discussions at the volitional level have rarely incorporated computational models, which has also limited researchers' ability to parse volition to some extent.

The VMP paradigm can provide insights for further dissecting volition. Compared with the classic voluntary action-related paradigms described above, the VMP paradigm based on voluntary choice has unique advantages in dissociating voluntary action and control belief: First, in the VMP paradigm, there is no association between the voluntary action behavior (the action of pressing a key to select an image during the choice phase) and the dependent variable measuring volitional effects (cognitive performance during the task phase), allowing the two to be completely separated in time. This ensures that the dependent variable index (cognitive performance) is not influenced by factors at the voluntary action execution level. In other words, based on the temporal separability between voluntary action and measurement index, the influence of voluntary action on the dependent variable index can be eliminated by extending the time interval between them (e.g., a 24-hour interval). However, it should be noted that this separability is also a limitation of using the VMP paradigm to measure volition—the index reflecting volition in the paradigm is not measured directly during the process of generating and expressing volition (i.e., during the selection process) but is measured indirectly after the volitional expression process ends (i.e., after selection is completed). This may only reflect certain aspects rather than the full picture of the volitional process. Therefore, attempting to directly analyze brain activity during the generation and expression of volition using EEG, fMRI, and other techniques is essential.

Second, in the VMP paradigm, there is no association between the outcome triggered by voluntary action (the background image appearing in the task phase) and the dependent variable measuring volitional effects (cognitive performance in the task phase). Therefore, the relationship between voluntary action and its outcome can be directly manipulated to alter control belief without affecting the cognitive task, thereby investigating how changes in control belief influence subsequent cognitive performance. Previous research has found (Luo et al., 2022) that in the VMP paradigm, when the "choice-outcome" causal relationship is confirmed (i.e., the image selected during the choice phase matches the background image presented in the task phase), voluntary choice facilitates cognitive performance. However, when the "choice-outcome" causal relationship is nullified (i.e., the image selected during the choice phase is unrelated to the background image presented in the task phase—choice cannot control the outcome, and control belief is removed), the facilitative effect of voluntary choice on subsequent cognitive performance disappears. This suggests that based on the VMP paradigm, control belief can be altered by manipulating the "choice-outcome" relationship without changing voluntary action itself, thereby eliminating the influence of control belief on the dependent variable index.

Third, the effect of voluntary choice facilitating cognitive performance based on the VMP paradigm shows consistency across tasks and effectors, representing a stable and reliable method for quantifying volitional effects. On the one hand, studies based on the VMP paradigm have found that the facilitative effect of voluntary choice on subsequent cognitive performance stably exists across various cognitive tasks, including visual search tasks (Luo et al., 2022; Luo, Wang, Gu, et al., 2024) and multiple conflict control tasks (including Flanker, Stroop, Simon, Stroop-Simon, and Flanker-Simon tasks) (Luo, Wang, & Zhou, 2024). That is, regardless of how the cognitive task varies, the facilitative effect of voluntary choice on cognitive performance (primarily facilitating response speed) can be demonstrated in these tasks. On the other hand, studies based on the VMP paradigm have also found that the facilitative effect of voluntary choice is independent of the effector used to implement the choice. Whether selecting by hand (i.e., participants press keys with their fingers), by foot (i.e., participants press pedals with their left/right foot), by mouth (i.e., participants verbally instruct which option to select, and the experimenter presses the key), or by eye (i.e., based on eye-tracking to capture participants' gaze location, participants select by fixating on the desired option), the facilitative effect of voluntary choice on cognitive performance can be observed (Luo, Wang, Gu, et al., 2024). These findings collectively suggest the "domain-general" nature of VMP effects.

Fourth, the VMP paradigm evaluates the role of volition based on performance in specific cognitive tasks, which facilitates the use of computational models to distinguish different cognitive processing components within the tasks and thereby characterize the influence of volition in detail. Specifically, many classic computational models are applicable to cognitive tasks in the VMP paradigm. For example, the drift diffusion model (DDM) (Ratcliff & McKoon, 2008) is suitable for most two-alternative-forced-choice (2AFC) cognitive tasks and has been extended into various versions for different specific tasks. In studies based on the VMP paradigm, researchers have begun using a DDM variant—the EZ-diffusion model (Wagenmakers et al., 2007)—to fit visual search tasks (Luo et al., 2022; Luo, Wang, Gu, et al., 2024). Another DDM variant—the diffusion model for conflict tasks (DMC) (Ulrich et al., 2015)—has been used to fit conflict control tasks (Luo, Wang, & Zhou, 2024). Model fitting of these cognitive tasks allows researchers to go beyond reaction times and error rates to further analyze different processing components in cognitive tasks in detail. For example, the drift rate parameter in DDM may be related to internal cognitive processing, the boundary parameter may be related to response bias, and non-decision time may be related to response execution. Such detailed model parameter analysis helps reveal how volition specifically influences different cognitive components, thereby deepening our understanding of the cognitive mechanisms of volition.

In summary, although current VMP paradigm-related research has not yet distinguished between voluntary action and control belief, dissociating the two based on this paradigm is clearly feasible. Therefore, this study proposes to dissociate the external manifestation (voluntary action) and internal representation (control belief) within volition based on the VMP paradigm, combined with behavioral computational modeling and multiple neurophysiological techniques (including EMG, EEG, and fMRI), to reveal their unique and common underlying cognitive neural mechanisms.

3 Research Framework

3.1 Research Framework and Objectives

This study aims to design variations of the VMP paradigm to dissociate the roles of voluntary action and control belief in volition's influence on cognitive performance. Furthermore, it will combine computational modeling and multiple neurophysiological techniques to reveal the cognitive neural mechanisms through which voluntary action and control belief jointly and separately influence cognitive processing. The research framework and technical roadmap are shown in Figure 3 [FIGURE:3].

Study 1 dissociates the effects of voluntary action and control belief in human volition through experimental design, separates different components of cognitive processing through computational modeling (the diffusion model for conflict tasks, DMC), and reveals the unique and common influences of voluntary action and control belief on cognitive processing at the behavioral level.

Study 2 uses simultaneous EMG and EEG recording techniques to reveal the unique and common dynamic processes of activation at the muscle level (action) and brain level (belief) for voluntary action and control belief. Furthermore, combined with DMC model fitting, it will explore the correspondence between different components of cognitive processing and electrophysiological indices, ultimately revealing the dynamic neurophysiological mechanisms through which volition influences cognitive performance.

Study 3 uses fMRI technology to reveal the unique and common neural basis, functional connectivity, and activation patterns of voluntary action and control belief influencing cognitive processing at the blood oxygenation level dependent (BOLD) signal level, thereby elucidating the brain mechanisms through which different components of volition influence cognitive performance.

Figure 3. Research framework and technical roadmap. Black boxes represent research questions, blue boxes represent research paradigms, green boxes represent research content, yellow boxes represent technical methods, and red boxes represent corresponding scientific questions. In this study, the influence of volition on cognitive performance is primarily evaluated through the effect of voluntary choice on Simon task performance. Color figures are available in the electronic version of this article.

3.2 Key Experimental Paradigm Design

This study proposes to use performance on the Simon task (Hommel, 2011; Simon & Rudell, 1967) as the index for evaluating volitional effects in the VMP paradigm. The Simon task is a classic conflict control task that requires participants to respond to stimuli presented on the left or right side of the screen while ignoring stimulus location. When the stimulus location is inconsistent with the location of the correct response effector (e.g., stimulus appears on the left but the correct response is with the right hand), cognitive performance declines (a conflict occurs). The reasons for selecting this task are as follows: First, the cognitive processes involved in the Simon task are relatively clear and can be easily separated and interpreted through neurophysiological techniques (Egner, 2008; Wang et al., 2021). Second, based on VMP paradigm findings, voluntary choice facilitates resolution of Simon conflict (Luo, Wang, & Zhou, 2024), indicating that the Simon task is sensitive to voluntary choice effects. Third, the Simon task is suitable for fitting the specialized computational model for conflict control tasks, the DMC (Ulrich et al., 2015), which can separate different cognitive processing components of conflict control tasks at the model level (e.g., Luo, Wang, & Zhou, 2024; Mittelstädt et al., 2022).

The specific experimental procedure is shown in Figure 4 [FIGURE:4]. The experiment is conducted over two days. On Day 1, participants first complete the standard VMP paradigm task to obtain the combined effect of voluntary action and control belief. After 24 hours (Day 2), participants complete two variations of the VMP paradigm—Variation ① isolates the unique effect of voluntary action while eliminating the effect of control belief; Variation ② isolates the unique effect of control belief while eliminating the effect of voluntary action (the order of completing the two VMP variations is counterbalanced across participants).

The standard VMP paradigm is shown in Figure 4A and consists of three phases: cue, choice, and task (Luo et al., 2022). The cue phase uses different colors to indicate whether the current trial involves voluntary or forced choice. In the choice phase, if it is a voluntary choice condition, participants can press a key to select one of two images; if it is a forced choice condition, participants can only press the key to select the pre-specified image. Subsequently, the task phase begins, where the selected image is presented as a task-irrelevant background against which participants complete the Simon task. In the standard VMP paradigm, participants both execute voluntary actions and hold control beliefs (i.e., believe that their selected image can appear as the background). The influence of volition on cognitive performance simultaneously includes the effect of voluntary action and the effect of control belief (combined effect).

VMP Variation ① is shown in Figure 4B and aims to eliminate control belief while retaining only the effect of voluntary action. Its key difference from the standard VMP paradigm is that the background image in the task phase is always a fixed, unchanging image unrelated to participants' selection in the choice phase. In this way, the causal "choice-outcome" relationship is broken, leading participants to recognize that "the presented image is not what I selected, not my own volition, and my choice cannot influence the outcome." In this variation, although participants execute voluntary choice actions, they do not possess control belief. The effectiveness of this manipulation in eliminating control belief has been validated in previous research (Luo et al., 2022).

VMP Variation ② is shown in Figure 4C and aims to eliminate voluntary action while retaining only the effect of control belief. Its key difference from the standard VMP paradigm is the absence of a choice phase—that is, participants are not required to perform selection actions. The cue phase uses different colors to inform participants whether the image to be presented in the current trial was voluntarily or forcibly selected by them on the previous day in the standard VMP task (i.e., directly informing participants of the image's source without requiring them to select an image). Subsequently, the task phase begins directly, presenting the image background to initiate the Simon task. In this way, participants do not perform any voluntary actions before the cognitive task begins but can establish control beliefs based on the cue information. That is, participants can recognize through the cue that "this image was selected by me yesterday and represents my own volition."

Figure 4. Volition-motivated performance (VMP) paradigm and its variations. (A) Standard VMP paradigm. Cue phase: Different colors indicate the upcoming choice type (voluntary choice vs. forced choice). Choice phase: Voluntarily or forcibly press a key to select one of two images. Task phase: Complete the Simon task with the selected image as background. In the standard VMP paradigm, the influence of volition on cognitive performance simultaneously includes the effect of voluntary action and the effect of control belief (combined effect). (B) VMP Variation ①. Compared with the standard VMP paradigm, the background image in the task phase is always a fixed, unchanging image unrelated to the selection in the choice phase. This variation aims to eliminate control belief while retaining only the effect of voluntary action. (C) VMP Variation ②. Compared with the standard VMP paradigm, there is no choice phase. Cue phase: Different colors indicate the type of image to be presented (image voluntarily selected yesterday vs. image forcibly selected yesterday). Task phase: Present the corresponding image background for the Simon task. This variation aims to eliminate voluntary action while retaining only the effect of control belief. Note: There is a fixation point lasting 0.5–0.8 s between different phases (cue, choice, task phases). For simplicity, these fixation points are not shown in the flowchart. Color figures are available in the electronic version of this article.

3.3 Research Plan

(1) Study 1: Dissociating the Effects of Voluntary Action and Control Belief on Cognitive Performance

Based on the standard VMP paradigm and its variations (see "3.2 Key Experimental Paradigm Design" for details), Study 1 will use behavioral experimental techniques to investigate the unique and common influences of voluntary action and control belief on cognitive task processing. The experiment is conducted over two days. On Day 1, participants perform the standard VMP paradigm task to obtain the combined effect of voluntary choice and control belief. After 24 hours, participants perform the Day 2 tasks, which include the two VMP variations (order counterbalanced across participants): VMP Variation ① eliminates control belief while retaining only the effect of voluntary action; VMP Variation ② eliminates voluntary action while retaining only the effect of control belief. Furthermore, the DMC model (Ulrich et al., 2015) will be used to fit Simon task performance to separate different components of this cognitive processing, ultimately revealing at the behavioral level how different components of volition influence different components of cognitive processing.

Primary Measures and Data Analysis: The primary measures for model-free analysis are Simon task reaction time and accuracy, analyzed using a 2 (Simon stimulus congruency: congruent vs. incongruent) × 2 (choice type: voluntary choice vs. forced choice) × 3 (task type: standard VMP paradigm, VMP Variation ① vs. VMP Variation ②) repeated-measures ANOVA. Further DMC model analysis will be conducted to separate task-related action activation from task-irrelevant action activation in the Simon task (Luo, Wang, & Zhou, 2024). The main parameters of the DMC are: (1) evidence accumulation rate for task-related action activation (μc), (2) peak activation of task-irrelevant actions (A), (3) time required for task-irrelevant action activation to reach peak (tpeak), (4) decision boundary (b), and (5) non-decision time (Rmean). The same ANOVA as in the model-free analysis will be performed on the optimal parameters.

Primary Predicted Results: Model-free analysis of Simon task reaction times is expected to show a significant three-way interaction. Specifically, in the standard VMP paradigm, voluntary choice overall speeds up Simon task reaction times while simultaneously reducing the Simon effect. However, in VMP Variation ①, voluntary choice only speeds up Simon task reaction times overall, while in Variation ②, it only reduces the Simon effect. DMC model analysis is expected to show primary effects concentrated on the peak activation of task-irrelevant actions (A): Variation ② should show the same pattern as the standard VMP paradigm—voluntary choice reduces A—whereas Variation ① should not show this effect.

(2) Study 2: Dynamic Processes of Voluntary Action and Control Belief Influencing Cognitive Performance

Building upon the replication of Study 1's findings, Study 2 will use high temporal resolution simultaneous EMG and EEG recording techniques to investigate the unique and common dynamic processes of activation at the muscle level (action) and brain level (belief) for voluntary action and control belief. Furthermore, combined with DMC model fitting, it will explore the correspondence between different components of cognitive processing and electrophysiological indices, ultimately revealing the dynamic neurophysiological mechanisms through which volition influences cognitive performance. The paradigm is the same as in Study 1 but with the addition of simultaneous EMG and EEG recording.

Primary Measures and Data Analysis: Model-free and DMC model analyses are the same as in Study 1. For EMG data, the analysis will focus on the magnitude of electromyographic signals within several critical time windows: after cue presentation, after option presentation, before action execution, and after Simon stimulus presentation. For EEG data, the focus will be on the contingent negative variation (CNV) during the cue phase, readiness potential (RP) during the choice phase, and posterior contralateral negativity (N2pc) and lateralized readiness potential (LRP) during the task phase. Further analysis will combine temporal change patterns of EMG and EEG signals to conduct representational similarity analysis (RSA) to explore the coupling between the two types of electrophysiological signals. Finally, individual difference analyses will be performed based on electrophysiological indices and DMC model parameters. Additionally, when investigating the effect of control belief alone, memory-related EEG components—frontal negative potential (FN400, Rugg & Curran, 2007)—will also be examined, as they may be related to the establishment of control belief.

Primary Predicted Results: Behavioral data results are expected to be the same as in Study 1. For electrophysiological signals, we expect that in the standard VMP paradigm, both CNV patterns during the cue phase and RP patterns during the choice phase can predict choice type and predict the influence of voluntary choice on cognitive performance. In VMP Variation ①, this predictive effect will be primarily in RP during the choice phase; in VMP Variation ②, it will be primarily in CNV during the cue phase. EMG and EEG signals from the same phase can predict each other, but the prediction accuracy will differ across variations: highest in the standard VMP paradigm, followed by Variation ①, and lowest in Variation ②.

(3) Study 3: Neural Basis of Voluntary Action and Control Belief Influencing Cognitive Performance

Building upon the replication of Study 1's findings, Study 3 will use high spatial resolution fMRI technology to investigate the unique and common brain neural basis of voluntary action and control belief, functional connectivity between brain regions, and differences in brain activation patterns. Furthermore, combined with DMC model fitting, it will explore the relationships between different components of cognitive processing and brain activation, functional connectivity, and whole-brain or regional activation patterns, ultimately revealing the brain mechanisms through which volition influences cognitive performance. The paradigm is essentially the same as in Study 1, with only minor adjustments suitable for fMRI research.

Primary Measures and Data Analysis: Model-free and DMC model analyses are the same as in Study 1. For fMRI data, general linear model results will primarily focus on activation of SMA and insula (INS) during the cue and choice phases, SMA and M1 activation during the task phase (e.g., Wang et al., 2019), and will also examine reward-related brain regions during different phases (e.g., striatum, Leotti & Delgado, 2011; Murayama et al., 2015; Murty et al., 2015). fMRI functional connectivity analysis will primarily use SMA and INS as seed regions to investigate their functional connectivity with other brain regions. Differences in brain activation patterns will be analyzed primarily through RSA methods to examine whole-brain activation and activation pattern differences in SMA and INS between different phases (e.g., cue phase vs. choice phase). Additionally, in VMP Variation ②, memory-related brain region hippocampus (Murty et al., 2015) will also be examined, as it may be related to the establishment of control belief.

Primary Predicted Results: Behavioral data results are expected to be the same as in Study 1. At the brain level, we expect that in the standard VMP paradigm, the degree and pattern of SMA activation during both the cue and choice phases can predict choice type and predict the influence of voluntary choice on cognitive performance, and that brain activation patterns during the cue phase can predict brain activation patterns during the choice phase. In VMP Variation ①, only brain activation during the choice phase will show similar predictive effects; in VMP Variation ②, only brain activation during the cue phase will show similar predictive effects. Brain activation patterns from the same phase will not be mutually predictable between the two variations.

4 Theoretical Construction

Although previous research has designed different paradigms from various angles to explore human volition, these studies have all relied on immediately executed voluntary actions (i.e., the external manifestation of volition), while individuals simultaneously hold control beliefs (i.e., the internal representation of volition) when performing voluntary actions. Therefore, previous studies have not dissociated the external manifestation and internal representation of volition. This study proposes to directly dissociate voluntary action and control belief in the volitional process and, combined with existing research findings, construct a "dual-path hypothesis of human volition processing" (see Figure 5 [FIGURE:5]). The rationale for this construction is detailed below.

First, in neuroscientific research related to volition, findings can be divided into two categories: "action-related" effects and "action-unrelated" effects, suggesting that the external manifestation and internal representation of volition may have different neural bases. Specifically, (1) regarding "action-related" effects, studies using EEG technology have found that voluntary actions can elicit stronger RP components (Khalighinejad et al., 2018; Libet et al., 1983), and RP may originate from SMA (Shibasaki & Hallett, 2006), indicating the important role of SMA in volitional expression. Additionally, fMRI studies have found SMA activation during voluntary action execution (Penfield, 1954; Rizzolatti & Kalaska, 2013), and the degree of SMA activation is related to sense of agency (Cavazzana et al., 2015; Kühn et al., 2013; Moore et al., 2010). These studies again emphasize the close connection between SMA and human volition, with SMA being an important brain region related to action planning. (2) Regarding "action-unrelated" effects, related research has primarily found a "reward-like effect" of expressing volition (voluntary choice): when individuals anticipate making a voluntary choice, reward-related brain regions such as the striatum are activated (Leotti & Delgado, 2011; Murayama et al., 2015; Murty et al., 2015). This suggests an association between expressing volition and reward circuits; expressing personal volition may enhance individuals' motivation, as if obtaining a reward. Thus, neuroscientific research on volition suggests that human volitional effects may involve two paths: an action-related path (RP, SMA-related) and an action-unrelated path (reward-related).

Second, in behavioral computational modeling research related to volition, altering control belief changes the pattern of volition's influence on different cognitive components, suggesting that the effect of volitional expression on cognitive performance may involve two paths: one related to control belief and another unrelated to control belief (related to voluntary action). Specifically, in Luo et al.'s (2022) study, individuals expressed volition through voluntary choice, which facilitated subsequent cognitive performance. More importantly, using the EZ-diffusion model (Wagenmakers et al., 2007), a variant of DDM, to fit and decompose the influence of voluntary choice on different cognitive components, results showed that for the model's non-decision time parameter, voluntary choice reduced non-decision time regardless of whether the choice outcome could be controlled (i.e., regardless of whether control belief was held). Therefore, in this process, control belief may be unimportant; what matters is the "action" of voluntary choice itself. Activation of the brain's action system caused by voluntary action (e.g., SMA) (Cunnington et al., 2002; Deiber et al., 1999) may generally facilitate subsequent action execution, thereby shortening non-decision time. In contrast, for the model's boundary parameter, when individuals could control the choice outcome (i.e., held control belief), voluntary choice (compared with forced choice) lowered the decision boundary, whereas when individuals could not control the choice outcome (i.e., no longer held control belief), voluntary choice no longer influenced the boundary. This suggests that control belief may make individuals' responses more liberal. Correspondingly, studies have found that providing rewards also lowers individuals' decision boundaries, making responses more liberal (Bowen et al., 2020; Luo et al., 2020), which aligns with the "association between expressing volition and reward effects" mentioned above. Conversely, depressed patients not only fail to show the effect of reward in lowering decision boundaries (Henriques & Glowacki, 1994) but also exhibit higher decision boundaries than healthy controls—that is, depressed patients' responses are more conservative (Lawlor et al., 2020)—suggesting that loss of control belief may be related to a "temporary depressive state." Additionally, reinforcement learning model-based studies have found that positive outcomes triggered by voluntary choice receive higher weight in the learning process compared with forced choice (Chambon et al., 2020). In summary, these studies again suggest that volitional effects may be divided into two aspects: voluntary action-related (influencing response execution) and control belief-related ("reward-like effect," influencing decision boundaries).

Third, based on the above previous research, this study directly dissociates the external manifestation (voluntary action) and internal representation (control belief) of volition. It further proposes to combine multimodal data including behavioral computational modeling (DMC model), electrophysiological signals (EMG and EEG data), and BOLD signals (fMRI data) to verify the similarities and differences in effects produced by voluntary action and control belief, and thereby infer the similarities and differences in their processing mechanisms. This can provide direct evidence for the "dual-path hypothesis of human volition processing." Specifically, on the one hand, we will focus on response-related indices/brain regions—for example: peak activation of task-irrelevant actions (A) and time to reach peak (tpeak) and non-decision time (Rmean) in the DMC model, EMG data, RP during the choice phase and LRP during the task phase in EEG data, and activation of SMA, PMC during the cue and choice phases and M1 during the choice phase in fMRI data—which may be related to the voluntary action path of volition. On the other hand, we need to focus on non-response-related (e.g., reward) indices/brain regions—for example: decision boundary (b) and evidence accumulation rate for task-related action activation (μc) in the DMC model, CNV during the cue phase and N2pc during the task phase in EEG data, and activation of INS and striatum during the cue and choice phases in fMRI data—which may be related to the control belief path of volition. In the design of this study (see Figure 4), the standard VMP paradigm includes both voluntary action and control belief, VMP Variation ① eliminates control belief and retains only voluntary action, and VMP Variation ② eliminates voluntary action and retains only control belief. If the volitional process can indeed be distinguished into two paths of voluntary action and control belief, then the pattern of results for response-related indices/brain regions should be mutually predictable between the standard VMP paradigm and VMP Variation ① (both include voluntary action), whereas the pattern of results for non-response-related indices/brain regions should not be mutually predictable between the standard VMP paradigm and VMP Variation ① (the former includes control belief, the latter does not). Similarly, the pattern of results for response-related indices/brain regions should not be mutually predictable between the standard VMP paradigm and VMP Variation ② (the former includes voluntary action, the latter does not), whereas the pattern of results for non-response-related indices/brain regions should be mutually predictable between the standard VMP paradigm and VMP Variation ② (both include control belief).

Fourth, in the "dual-path hypothesis of human volition processing," the voluntary action path and the control belief path are not completely independent but are closely related and mutually influential. On the one hand, only with control belief (believing that voluntary actions can influence the external world) will individuals be inclined to perform voluntary actions. This is also an important reason why individuals easily develop an "illusion of control" (overestimating the causal link between voluntary actions and external effects) (Langer, 1975; Thompson et al., 1998; Chen et al., 2010). Conversely, when individuals lose control belief, voluntary actions decrease or disappear, resulting in "learned helplessness" (Huys & Dayan, 2009; Maier & Seligman, 1976). On the other hand, the outcomes triggered by voluntary actions also regulate control belief in turn—if voluntary actions produce expected outcomes, control belief may be strengthened; if voluntary actions produce outcomes inconsistent with expectations, control belief may be weakened (see Figure 5). VMP paradigm research has found (Luo et al., 2022) that the facilitative effect of voluntary choice on subsequent cognitive performance strengthens over time. This suggests that individuals' control beliefs are reinforced through multiple "voluntary choice-outcome confirmation" processes (the selected image is presented as the task background). When voluntary choice cannot influence outcomes, the facilitative effect of voluntary choice on subsequent cognitive performance disappears. In summary, the volitional process is not static; its dynamic changes are manifested in the mutual regulation between the voluntary action path and the control belief path.

Finally, it is necessary to consider neural overlap: the neural basis of the control belief path hypothesized in this study (the reward system) lacks specificity—the reward system is also involved in many other non-volitional cognitive processes, such as risk calculation (Rolls et al., 2022), motivation regulation (Kringelbach & Berridge, 2016), and habitual behavior (Wood & Rünger, 2016). The internal representation process of volition may involve multiple reward-related cognitive processes, such as motivating behavior (Leotti et al., 2010), value calculation of action-outcome associations (Murayama et al., 2015), and reinforcement learning based on behavioral outcomes (Chambon et al., 2020). Therefore, the reward system may be a "necessary" but not an "exclusive" neural basis for the control belief path. One possible mechanism is that although multiple different cognitive processes involve the reward system, these cognitive processes have different neural representation patterns within the reward system. However, this study does not include other cognitive tasks related to the reward system and thus cannot verify whether control belief has a unique neural representation pattern. Future research should consider incorporating new variables that can activate the reward system (e.g., behavioral risk) in a dissociative design with the control belief variable, and use neural decoding analysis to verify the unique neural representation pattern of control belief.

The "dual-path hypothesis of human volition processing" expected to be constructed in this study (see Figure 5) can finely distinguish the roles of the external manifestation (voluntary action) and internal representation (control belief) of human volition. The hypothesis proposes that human volition processing occurs simultaneously through two paths: one path related to voluntary action, reflecting the action attribute of volition (action anticipation, planning, execution, etc.)—for example, expressing volition may facilitate subsequent response execution; the other path related to control belief, reflecting the motivational attribute of volition ("reward-like effect")—for example, expressing volition may influence individuals' decision boundaries (response bias), and restricted volition may be related to depression. In summary, human volition may possess dual attributes of action and motivation, forming a complete volitional process through dual-path processing.

5 References

陈煦海, 吴茜. (2019). 自主选择偏好：表现、机制与应用. 心理科学进展, 27(8), 1460–1467.

陈雪玲, 徐富明, 刘腾飞, 蒋多, 张军伟. (2010). 控制幻觉的研究方法、形成机制和影响因素. 心理科学进展, 18(5), 800–809.

田昊月, 李力红, 徐喆, 李飞, 金丹, 安灿翎. (2018). 最小自我中的施动感. 心理科学进展, 26(5), 804–810.

吴迪, 顾晶金, 李明, 张淼, 张明, 赵科, 等. (2019). 动作的主动控制感与因果关系的主动控制感：主动动作时间压缩效应的产生机制. 心理科学进展, 27(5), 804–810.

张淼, 吴迪, 李明, 凌懿白, 张明, 赵科. (2018). 主动控制感的测量及认知神经机制. 心理科学进展, 26(10),

Applebaum, A., Netzer, O., Stern, Y., Zvilichovsky, Y., Mashiah, O., & Salomon, R. (2025). The body knows better: Sensorimotor signals reveal the interplay between implicit and explicit sense of agency in the human mind. Cognition, 254, 105992. https://doi.org/10.1016/j.cognition.2024.105992

Arikan, B. E., Voudouris, D., Voudouri-Gertz, H., Sommer, J., & Fiehler, K. (2021). Reach-relevant somatosensory signals modulate activity in the tactile suppression network. NeuroImage, 236, 118000. https://doi.org/10.1016/j.neuroimage.2021.118000

Barlas, Z., Hockley, W. E., & Obhi, S. S. (2018). Effects of free choice and outcome valence on the sense of agency: Evidence from measures of intentional binding and feelings of control. Experimental Brain Research, 236, 129–139. https://doi.org/10.1007/s00221-017-5112-3

Bertoni, T., Noel, J-P., Bockbrader, M., Foglia, C., Colachis, S., Orset, B., …, Serino, A. (2025). Pre-movement sensorimotor oscillations shape the sense of agency by gating cortical connectivity. Nature Communications, 16, 3594. https://doi.org/10.1038/s41467-025-58683-9

Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (1998). Central cancellation of self-produced tickle sensation. Nature Neuroscience, 1(7), 635–640. https://doi.org/10.1038/2870

Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (2000). Why can't you tickle yourself? Neuroreport, 11(11), R11–R16. http://doi.org/10.1097/00001756-200012180-00006

Borhani, K., Beck, B., & Haggard, P. (2017). Choosing, doing, and controlling: Implicit sense of agency over somatosensory events. Psychological Science, 28(7), 882–893. https://doi.org/10.1177/0956797617697693

Bowen, H. J., Marchesi, M. L., & Kensinger, E. A. (2020). Reward motivation influences response bias on a recognition memory task. Cognition, 203, 104337. http://doi.org/10.1016/j.cognition.2020.104337

Caspar, E. A., Christensen, J. F., Cleeremans, A., & Haggard, P. (2016). Coercion changes the sense of agency in the human brain. Current Biology, 26, 585–592. https://doi.org/10.1016/j.cub.2015.12.067

Cavazzana, A., Penolazzi, B., Begliomini, C., & Bisiacchi, P. S. (2015). Neural underpinnings of the 'agent brain': New evidence from transcranial direct current stimulation. European Journal of Neuroscience, 42(3), 1889–1894. http://doi.org/10.1111/ejn.12937

Christensen, J. F., Rödiger, C., Claydon, L., & Haggard, P. (2024). Volition and control in law and in brain science: Neurolegal translation of a foundational concept. Frontiers in Human Neuroscience, 18, 1401895. https://doi.org/10.3389/fnhum.2024.1401895

Chambon, V., Théro, H., Vidal, M., Vandendriessche, H., Haggard, P., & Palminteri, S. (2020). Information about action outcomes differentially affects learning from self-determined versus imposed choices. Nature Human Behaviour, 4, 1067–1079. https://doi.org/10.1038/s41562-020-0919-5

Cunnington, R., Windischberger, C., Deecke, L., & Moser, E. (2002). The preparation and execution of self-initiated and externally-triggered movement: A study of event-related fMRI. NeuroImage, 15(2), 373–385. http://doi.org/10.1006/nimg.2001.0976

Daprati, E., Franck, N., Georgieff, N., Proust, J., Pacherie, E., & Dalery, J. et al. (1997). Looking for the agent: An investigation into consciousness of action and self-consciousness in schizophrenic patients. Cognition, 65(1), 71–86. http://doi.org/10.1016/s0010-0277(97)00039-5

Deiber, M., Honda, M., Ibañez, V., Sadato, N., & Hallett, M. (1999). Mesial motor areas in self-initiated versus externally triggered movements examined with fMRI: Effect of movement type and rate. Journal of Neurophysiology, 81(6), 3065–3077. http://doi.org/10.1152/jn.1999.81.6.3065

Desantis, A., Roussel, C., & Waszak, F. (2011). On the influence of causal beliefs on the feeling of agency. Consciousness and Cognition, 20, 1211–1220. http://doi.org/10.1016/j.concog.2011.02.012

Desmurget, M., Reilly, K. T, Richard, N., Szathmari, A., Mottolese, C., & Sirigu, A. (2009). Movement intention after parietal cortex stimulation in humans. Science, 324(5928), 811–813. http://doi.org/10.1126/science.1169896

Dogge, M., Schaap, M., Custers, R., Wegner, D. M., & Aarts, H. (2012). When moving without volition: Implied self-causation enhances binding strength between involuntary actions and effects. Consciousness and Cognition, 21(1), 501–506. http://doi.org/10.1016/j.concog.2011.10.014

Egner, T. (2008). Multiple conflict-driven control mechanisms in the human brain. Trends in Cognitive Sciences, 12(10), 380. http://doi.org/10.1016/j.tics.2008.07.001

Eitam, B., Kennedy, P. M., & Higgins, E. T. (2013). Motivation from control. Experimental Brain Research, 229(3), 475–484. http://doi.org/10.1007/s00221-012-3370-7

Fried, I., Katz, A., McCarthy, G., Sass, K. J., Williamson, P., Spencer, S. S., & Spencer, D. D. (1991). Functional organization of human supplementary motor cortex studied by electrical stimulation. Journal of Neuroscience, 11(11), 3656–3666. http://doi.org/10.1523/JNEUROSCI.11-11-03656.1991

Frith, C. (2013). The psychology of volition. Experimental Brain Research, 229, 289–299. https://doi.org/10.1007/s00221-013-3407-6

Georgieff, N., & Jeannerod, M. (1998). Beyond consciousness of external reality: A "who" system for consciousness of action and self-consciousness. Consciousness and Cognition, 7(3), 465–477. http://doi.org/10.1006/ccog.1998.0367

Haggard, P. (2008). Human volition: Towards a neuroscience of will. Nature Reviews Neuroscience, 9(12), 934–946. http://doi.org/10.1038/nrn2497

Haggard, P. (2017). Sense of agency in the human brain. Nature Reviews Neuroscience, 18(4), 196–207. http://doi.org/10.1038/nrn.2017.14

Haggard, P. (2019). The neurocognitive bases of human volition. Annual Review of Psychology, 70, 9–28. http://doi.org/10.1146/annurev-psych-010418-103348

Haggard, P., Clark, S., & Kalogeras, J. (2002). Voluntary action and conscious awareness. Nature Neuroscience, 5(4), 382–385. http://doi.org/10.1038/nn827

Hemed, E., Bakbani-Elkayam, S., Teodorescu, A. R., Yona, L., & Eitam, B. (2020). Evaluation of an action's effectiveness by the motor system in a dynamic environment. Journal of Experimental Psychology: General, 149(5), 935–948. https://doi.org/10.1037/xge0000692

Henriques, J. B., & Glowacki, J. M. (1994). Reward fails to alter response bias in depression. Journal of Abnormal Psychology, 103(3), 460–466. http://doi.org/10.1037/0021-843X.103.3.460

Hommel, B. (2011). The Simon effect as tool and heuristic. Acta Psychologica, 136(2), 189–202. http://doi.org/10.1016/j.actpsy.2010.04.011

Huys, Q. J. M., & Dayan, P. (2009). A Bayesian formulation of behavioral control. Cognition, 113(3), 314–328. http://doi.org/10.1016/j.cognition.2009.01.008

Kane, R. (1996). The significance of free will. Oxford University Press

Karsh, N., Eitam, B., Mark, I., & Higgins, E. T. (2016). Bootstrapping agency: How control-relevant information affects motivation. Journal of Experimental Psychology: General, 145(4), 1333–1350. http://doi.org/10.1037/xge0000212

Khalighinejad, N., Schurger, A., Desantis, A., Zmigrod, L., & Haggard, P. (2018). Precursor processes of human self-initiated action. NeuroImage, 165, 35–47. http://doi.org/10.1016/j.neuroimage.2017.09.057

Kringelbach, M. L. & Berridge, K. C. (2016). Neuroscience of reward, motivation, and drive. In S. Kim, J. Reeve, & M. Bong (Eds.), Recent developments in neuroscience research on human motivation (Advances in motivation and achievement, Vol. 19) (pp. 23–35). Emerald Group Publishing Limited, Leeds. https://doi.org/10.1108/S0749-742320160000019020

Kuhl, J. (1984). Volitional aspects of achievement motivation and learned helplessness: Toward a comprehensive theory of action control. In B. A. Maher (Ed.), Progress in experimental personality research (Vol. 13, pp. 99–171). Academic Press. https://doi.org/10.1016/B978-0-12-541413-5.50007-3

Kühn, S., Brass, M., & Haggard, P. (2013). Feeling in control: Neural correlates of experience of agency. Cortex, 49(7), 1935–1942. http://doi.org/10.1016/j.cortex.2012.09.002

Langer, E. J. (1975). The illusion of control. Journal of Personality & Social Psychology, 32(2), 311–328. http://doi.org/10.1037/0022-3514.32.2.311

Lawlor, V. M., Webb, C. A., Wiecki, T. V., Frank, M. J., Trivedi, M., & Pizzagalli, D. A. et al. (2020). Dissecting the impact of depression on decision-making. Psychological Medicine, 50(10), 1613–1622. http://doi.org/10.1017/S0033291719001570

Leotti, L. A., Iyengar, S. S., & Ochsner, K. N. (2010). Born to choose: The origins and value of the need for control. Trends in Cognitive Sciences, 14(10), 457–463. http://doi.org/10.1016/j.tics.2010.08.001

Leotti, L. A. & Delgado, M. R. (2011). The inherent reward of choice. Psychological Science, 22(10), 1310–1318. http://doi.org/10.1177/0956797611417005

Legault, L. & Inzlicht, M. (2013). Self-determination, self-regulation, and the brain: Autonomy improves performance by enhancing neuroaffective responsiveness to self-regulation failure. Journal of Personality & Social Psychology, 105(1), 123–138. https://doi.org/10.1037/a0030426

Libet, B., Gleason, C. A., Wright, E. W., & Pearl, D. K. (1983). Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential): The unconscious initiation of a freely voluntary act. Brain, 106(3), 623–642. http://doi.org/10.1093/brain/106.3.623

Lindner, E., Schöberl, T., Desantis, A., Gail, A. (2025). Temporal action-effect prediction does not affect perceived loudness, but the sense of agency. Consciousness and Cognition, 130, 103837. https://doi.org/10.1016/j.concog.2025.103837

Luo, X., Kang, G., Guo, Y., Yu, X., & Zhou, X. (2020). A value-driven McGurk effect: Value-associated faces enhance the influence of visual information on audiovisual speech perception and its eye movement pattern. Attention, Perception, & Psychophysics, 82, 1928–1941. http://doi.org/10.3758/s13414-019-01918-x

Luo, X., Wang, L., Gu, J., Zhang, Q., Ma, H., & Zhou, X. (2024). The benefit of making voluntary choices generalizes across multiple effectors. Psychonomic Bulletin & Review, 31, 340–352. http://doi.org/10.3758/s13423-023-02350-x

Luo, X., Wang, L., & Zhou, X. (2022). Belief in control: Voluntary choice enhances subsequent task performance under undefeated choice-outcome causation. Cognition, 255, 105108. http://doi.org/10.1016/j.cognition.2022.105108

Luo, X., Wang, L., & Zhou, X. (2024). Volition motivates cognitive performance at the response-execution level by attenuating task-irrelevant motor activations. Cognition, 245, 105738. http://doi.org/10.1016/j.cognition.2024.105738

Maier, S. F., & Seligman, M. E. (1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology: General, 105(1), 3–46. http://doi.org/10.1037/0096-3445.105.1.3

Matute, H., & Blanco, F. (2014). Reducing the illusion of control when an action is followed by an undesired outcome. Psychonomic Bulletin & Review, 21(4), 1087–1093. http://doi.org/10.3758/s13423-014-0584-7

Mifsud, N. G., Beesley, T., Watson, T. L., Elijah, R. B., Sharp, T. S., & Whitford, T. J. (2018). Attenuation of visual evoked responses to hand and saccade-initiated flashes. Cognition, 179, 14–22. https://doi.org/10.1016/j.cognition.2018.06.005

Mittelstädt, V., Miller, J., Leuthold, H., Mackenzie, I. G., & Ulrich, R. (2022). The time-course of distractor-based activation modulates effects of speed-accuracy tradeoffs in conflict tasks. Psychonomic Bulletin & Review, 29, 837–854. http://doi.org/10.3758/s13423-021-02003-x

Moore, J. W., Lagnado, D., Deal, D. C., & Haggard, P. (2009). Feelings of control: Contingency determines experience of action. Cognition, 110(2), 279–283. http://doi.org/10.1016/j.cognition.2008.11.006

Moore, J. W., Ruge, D., Wenke, D., Rothwell, J., & Haggard, P. (2010). Disrupting the experience of control in the human brain: Pre-supplementary motor area contributes to the sense of agency. Proceedings of the Royal Society B: Biological Sciences, 277(1693), 2503–2509. http://doi.org/10.1098/rspb.2010.0404

Murayama, K., Matsumoto, M., Izuma, K., Sugiura, A., Ryan, R. M., & Deci, E. L. et al. (2015). How self-determined choice facilitates performance: A key role of the ventromedial prefrontal cortex. Cerebral Cortex, 25(5), 1241–1251. http://doi.org/10.1093/cercor/bht317

Murty, V. P., DuBrow, S., & Davachi, L. (2015). The simple act of choosing influences declarative memory. Journal of Neuroscience, 35(16), 6255–6264. http://doi.org/10.1523/JNEUROSCI.4181-14.2015.

Noel, J-P., Bockbrader, M., Bertoni,T., Colachis, S., Solca, M., Orepic, P., …, Serino, A. (2025). Neuronal responses in the human primary motor cortex coincide with the subjective onset of movement intention in brain–machine interface-mediated actions. PLoS Biology, 23(4), e3003118. https://doi.org/10.1371/journal.pbio.3003118

Obhi, S. S., & Hall, P. (2011). Sense of agency and intentional binding in joint action. Experimental Brain Research, 211(3–4), 655–662. http://doi.org/10.1007/s00221-011-2675-2

Ota, K., Charles, L., Haggard, P. (2024). Autonomous behaviour and the limits of human volition. Cognition, 244, 105684. https://doi.org/10.1016/j.cognition.2023.105684

Passingham, R. E. (1987). Two cortical systems for directing movement. Ciba Foundation Symposium, 132, 151–164. http://doi.org/10.1002/9780470513545.ch10

Penfield, W. (1954). Mechanisms of voluntary movement. Brain, 77(1), 1–17. http://doi.org/10.1093/brain/77.1.1

Ratcliff, R., & McKoon, G. (2008). The diffusion decision model: Theory and data for two-choice decision tasks. Neural Computation, 20, 873–922. http://doi.org/10.1162/neco.2008.12-06-420

Reznik, D., Henkin, Y., Schadel, N., & Mukamel, R. (2014). Lateralized enhancement of auditory cortex activity and increased sensitivity to self-generated sounds. Nature Communications, 5, 4059. https://doi.org/10.1038/ncomms5059

Ricciardi, L., Haggard, P., de Boer, L., Sorbera, C., Stenner, M., & Morgante, F. et al. (2017). Acting without being in control: Exploring volition in Parkinson's disease with impulsive compulsive behaviours. Parkinsonism & Related Disorders, 40, 51–57. http://doi.org/10.1016/j.parkreldis.2017.04.011

Rizzolatti, G., & Kalaska, J. F. (2013). Voluntary movement: The parietal and premotor cortex. In E. R. Kandel, J. H. Schwartz, T. M. Jessell, S. A. Siegelbaum, A. J. Hudspeth & S. Mack (Eds.), Principles of neural science, fifth edition (pp. 865–893). McGraw Hill.

Rolls, E. T., Wan, Z., Cheng, W., & Feng, J. (2022). Risk-taking in humans and the medial orbitofrontal cortex reward system. NeuroImage, 249, 118893. https://doi.org/10.1016/j.neuroimage.2022.118893

Ruiz, N. A., DuBrow, S., & Murty, V. P. (2023). Agency as a bridge to form associative memories. Journal of Experimental Psychology: General, 152(6), 1797–1813. https://doi.org/10.1037/xge0001356

Rugg, M. D., & Curran, T. (2007). Event-related potentials and recognition memory. Trends in Cognitive Sciences, 11(6), 251–257. http://doi.org/10.1016/j.tics.2007.04.004

Schafer, E. W. P., & Marcus, M. M. (1973). Self-stimulation alters human sensory brain responses. Science, 181(4095), 175–177. https://doi.org/10.1126/science.181.4095.175

Schwarz, K. A., Weller, L., Klaffehn, A. L., & Pfister, R. (2019). The effects of action choice on temporal binding, agency ratings, and their correlation. Consciousness and Cognition, 75, 102807. http://doi.org/10.1016/j.concog.2019.102807

Shibasaki, H., & Hallett, M. (2006). What is the bereitschaftspotential? Clinical Neurophysiology, 117(11), 2341–2356. http://doi.org/10.1016/j.clinph.2006.04.025

Simon, J. R., & Rudell, A. P. (1967). Auditory S-R compatibility: The effect of an irrelevant cue on information processing. Journal of Applied Psychology, 51(3), 300–304. http://doi.org/10.1037/h0020586

Sirigu, A., Daprati, E., Pradat-Diehl, P., Franck, N., & Jeannerod, M. (1999). Perception of self-generated movement following left parietal lesion. Brain, 122(10), 1867–1874. https://doi.org/10.1093/brain/122.10.1867

Thompson, S. C., Armstrong, W., & Thomas, C. (1998). Illusions of control, underestimations, and accuracy: A control heuristic explanation. Psychological Bulletin, 123(2), 143–161. http://doi.org/10.1037/0033-2909.123.2.143

Tsakiris, M., Haggard, P., Franck, N., Mainy, N., & Sirigu, A. (2005). A specific role for efferent information in self-recognition. Cognition, 96(3), 215–231. http://doi.org/10.1016/j.cognition.2004.08.002

Ulrich, R., Schröter, H., Leuthold, H., & Birngruber, T. (2015). Automatic and controlled stimulus processing in conflict tasks: Superimposed diffusion processes and delta functions. Cognitive Psychology, 78, 148–174. http://doi.org/10.1016/j.cogpsych.2015.02.005

Wagenmakers, E. J., van der Maas, H. L. J., & Grasman, R. P. P. P. (2007). An EZ-diffusion model for response time and accuracy. Psychonomic Bulletin & Review, 14(1), 3–22. http://doi.org/10.3758/BF03194023

Wang, L., Chang, W., Krebs, R. M., Boehler, C. N., Theeuwes, J., & Zhou, X. (2019). Neural dynamics of reward-induced response activation and inhibition. Cerebral Cortex, 29(9), 3961–3976. http://doi.org/10.1093/cercor/bhy275

Wang, L., Luo, X., Yuan, T., & Zhou, X. (2021). Reward facilitates response conflict resolution via global motor inhibition: Electromyography evidence. Psychophysiology, e13896. http://doi.org/10.1111/psyp.13896

Weiss, C., Herwig, A., & Schütz-Bosbach, S. (2011). The self in action effects: Selective attenuation of self-generated sounds. Cognition, 121(2), 207–218. https://doi.org/10.1016/j.cognition.2011.06.011

Wen, W., Charles, L., & Haggard, P. (2023). Metacognition and sense of agency. Cognition, 241, 105622. https://doi.org/10.1016/j.cognition.2023.105622

Wen, W. & Imamizu, H. (2022). The sense of agency in perception, behaviour and human–machine interactions. Nature Reviews Psychology, 1, 211–222. https://doi.org/10.1038/s44159-022-00030-6

Wen, W. & Haggard, P. (2020). Prediction error and regularity detection underlie two dissociable mechanisms for computing the sense of agency. Cognition, 195, 104074. https://doi.org/10.1016/j.cognition.2019.104074

Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation. American Psychologist, 54(7), 480–492. http://doi.org/10.1037/0003-066X.54.7.480

Wood, W. & Rünger, D. (2016). Psychology of habit. Annual Review of Psychology, 67, 289–314. https://doi.org/10.1146/annurev-psych-122414-033417

Yarritu, I., Matute, H., & Vadillo, M. A. (2014). Illusion of control: The role of personal involvement. Experimental Psychology, 61(1), 38–47. http://doi.org/10.1027/1618-3169/a000225

Yavuz, M., Bonicalzi, S., Schmitz, L., Battich, L., Esmaily, J., & Deroy, O. (2025). Rational choices elicit stronger sense of agency in brain and behavior. Cognition, 257, 106062. https://doi.org/10.1016/j.cognition.2025.106062

(Corresponding author: Luo Xiaoxiao, E-mail: luoxiaoxiao@ynnu.edu.cn)

Author Contribution Statement:
Luo Xiaoxiao: Developed research ideas, designed research plan, drafted and revised the manuscript.
Zhou Xiaolin: Revised research ideas and methods, revised the manuscript.