Since the rubber hand illusion was first reported (Botvinick and Cohen, 1998), there has been consensus that an important factor in embodiment sensations (sense of ownership and sense of agency) is congruency between the senses or between the prediction and the feedback. Sense of ownership refers to the experience of mineness toward a body part, while sense of agency is the experience of motor control over one’s body actions. Today, this knowledge has been applied in cognitive psychology (Ehrsson et al., 2004; Tsakiris and Haggard, 2005; Ehrsson, 2007; Lopez et al., 2008; Tsakiris, 2010; Seghezzi et al., 2019; Arzy and Schacter, 2019), the medical and rehabilitation fields (Burin et al., 2015; Tieri et al., 2018) and living environments, including smartphones (Cattan et al., 2018). In the motor control theory of the feedforward model, when motor intention occurs, predicted sensory feedback is generated in the brain and integrated with actual sensory feedback (Wolpert and Ghahramani, 2000; Blakemore et al., 1998; Blakemore et al., 1999; Frith et al., 2000). When the predicted sensory and actual sensory feedback are congruent, a sense of agency occurs, but when they are incongruent, uncomfortable movement and negative feelings occur. For example, patients with musculoskeletal and central nervous system disorders experience feedback of motor trajectories and sensations that differ from the predicted outcomes with motor intention. This incongruence reduces self-attribution and often results in negative sensations such as heaviness, discomfort, and pain associated with movement (McCabe et al., 2005; Foell et al., 2013; Kuppuswamy et al., 2016). These patients experience a negative sensation of incongruency every time they move, and even moving itself becomes an uncomfortable experience.

One experimental method often used to investigate the mechanisms of feeling generation induced by incongruence is to manipulate the spatial mismatch between the senses (Katayama et al., 2016; Katayama et al., 2019). In this experiment, when spatial incongruence is induced between visual feedback of one upper extremity and somatosensory feedback of another upper extremity using a mirror, a negative sensation occurs. Other experiments use a method for manipulating temporal incongruence. For example, it is known that the temporal perception of an intention-based voluntary action and the feedback induced by the action are attracted to each other (action and feedback); this phenomenon is called intentional binding. Intentional binding is recognized as a method that can be interpreted as sense of agency decreasing as the perceived time delay between an intention-based action and its feedback increases (Haggard et al., 2002; Moore and Obhi, 2012; Haggard, 2017).

Another typical method for manipulating sensorimotor incongruence has been studied using visual feedback delay (Shimada et al., 2009; Shimada et al., 2010; Shimada et al., 2014). In these studies, various temporal incongruities between the actual movement and visual feedback were introduced using a visual feedback delay experimental setup. The delay time was systematically varied, and the participants responded to whether the predicted and actual feedback were congruent or incongruent. Studies have shown that the rate of responses indicating congruence decreases when the delay increases. A study using wrist motion also reported that body perception (sense of body heaviness) and motor data (electromyography data) can be disturbed by manipulating the incongruence with a time delay (Osumi et al., 2018). Sensory integration through this temporal congruence has also been helpful in the sense of ownership (Shimada et al., 2009).

Although these experiments induced an incongruity focused on upper-extremity movement, Kannape and Blanke (2013) reported an interesting experiment on sensory-motor incongruence of whole-body movement while walking using visually delayed feedback in healthy participants. Participants were shown delayed visual feedback while walking on a treadmill, which was instantly converted into an avatar using 3D motion analysis equipment. Similar to a recent study involving the upper extremities, the rate of responses indicating congruence decreased when the delay time increased. With reference to this recent study on the temporal incongruency method while walking, we found that the illusion of a sense of body heaviness was induced (Hayashida et al., 2025). We believe that the incongruence of delayed visual feedback over the predicted sensory feedback induces the illusion of a sense of body heaviness. Therefore, these recent studies (Osumi et al., 2018; Hayashida et al., 2025) indicated that decreasing the sense of agency can indirectly induce the sense of body heaviness.

Interestingly, in the recent study (Hayashida et al., 2025), some of our participants reported a positive sensation of an illusion of body lightness during the experience of incongruence. Why was this illusion induced? The easiest and most easily imaginable example of a light feeling in the body is the sensation of walking on the moon. If 1/6 of Earth’s gravity is fed back on the Moon, as opposed to the predicted gravity on Earth, the body should feel lighter for that amount of gravity. Let us reiterate that the illusion of a sense of body heaviness is induced by an incongruence between delayed and predicted sensory feedback. In contrast, here we propose that the illusion of a sense of body lightness might be induced by incongruence between “subjectively preceding feedback” and predicted sensory feedback.

Because walking is a periodic movement, introducing a delay close to one stride can result in visual feedback that precedes the ongoing movement. As shown in Figure 1, when a participant’s single stride time is approximately 1,000 ms, real-time feedback (delay time: 0 ms) corresponds to the terminal stance phase. In contrast, when feedback is presented during the pre-swing phase with a delay of nearly one stride (e.g., 800 ms; Figure 1, previous cycle, orange), the visual information corresponds to a slightly future phase of the gait cycle (preceding time: 200 ms; Figure 1, current cycle, orange). This configuration effectively provides what we term “subjectively preceding feedback.” Such predicted feedback could induce an illusory sense of body lightness in some participants, and the purpose of this study was to investigate this systematically.

In addition, the fact that the illusion of a sense of body lightness was induced in only some participants may have been owing to psychological characteristics. In particular, sense of agency, the subjective sensation produced by the integration of predicted and actual sensations, is affected by psychological traits. Schizophrenic personality and depression have been reported as the psychodynamic factors associated with a sense of agency. Higher schizophrenic personality tendencies are associated with reduced sensory-motor processing, which is thought to alter subjective experiences (Moore et al., 2011). It has also been reported that those with higher depressive tendencies are less able to incorporate feedback well (Obhi et al., 2013; Malik and Obhi, 2019; Nobusako et al., 2020) and have a reduced sense of agency. Therefore, it may be important for the illusion of lightness to induce a sense of agency for the visual feedback on the monitor; that is, it may be necessary to have a psychological feature that may allow immersion in it. In addition, individuals with psychological traits that are sensitive to sensory feedback may have superior sensory-motor processing, which may impact the sensory-motor integration process. This psychological trait, including sensory sensitivity, may be related to the acceptability of sensory feedback and may have influenced the elicitation of a sense of body lightness.

In summary, incongruence between predicted and actual feedback in upper limb movements causes negative effects, such as a sense of body heaviness. However, it is theoretically possible to induce positive sensations of a different nature than sense of body heaviness, such as a sense of body lightness, by using visual delay tasks while walking. Many patients with neurological and musculoskeletal disorders suffer from body perception disturbances, including a sense of body heaviness (Kuppuswamy et al., 2016; Serrada et al., 2021; Matsuda and Osumi, 2023; Dagenais et al., 2025). If it is possible to intentionally induce a sense of lightness in contrast to the negative sensation of body heaviness that many patients experience, this may be used as a future intervention tool. This new illusion has the potential to contribute not only to the medical and rehabilitation fields but also to extended reality technology and other interdisciplinary fields. Therefore, the purpose of this study was to investigate whether the application of a visual delay system while walking could induce an illusion of a sense of body lightness, and to investigate the relationship between this illusion and psychological characteristics.

This study was exploratory and preliminary in nature. Accordingly, the exploratory questions were as follows.

All participants completed the experimental task without missing their balance and responded to two questions within the 3 s.

The participants were divided into an Illusion group (n = 9) and a Non-illusion group (n = 21). All participants in the Illusion group reported that they experienced the subjectively preceding feedback when they responded to the sense of body lightness. The minimum Q2 mean values (i.e., sense of body lightness) and the maximum Q2 mean values (i.e., sense of body heaviness) for each participant used in this grouping are shown in Figure 4. In the Illusion group, the median delay was 1,120 ms (interquartile range (IQR) = 570–1,220) in the Heaviness condition and 820 ms (IQR = 820–1,020) in the Lightness condition. In the Non-illusion group, the median delay was 570 ms (IQR = 520–1,120) in the Heaviness condition and 270 ms (IQR = 233–695) in the Lightness condition. The histograms of each body perception and each delay condition are shown in Figure 5.

The one-way ANOVA of the congruence response rate showed a significant main effect (F(2, 16) = 21.051, p < 0.01, η²_p = 0.725). A post hoc power analysis using G*Power 3.1.9.2 (Faul et al., 2007) indicated that the achieved power was high (Power ≈ 1.00). Multiple comparisons revealed that the congruence response rate in the Minimum delay condition (mean = 0.752, SD = 0.153) was significantly higher than in the Heaviness (mean = 0.182, SD = 0.198) (mean difference = 0.570, 95%CI = 0.413 to 0.727, p < 0.01) and Lightness conditions (mean = 0.364, SD = 0.261) (mean difference = 0.388, 95%CI = 0.118 to 0.658, p < 0.01). The congruence response rate in the Lightness condition and the Heaviness condition was not significantly (mean difference = 0.182, 95%CI = − 0.167 to 0.532, p = 0.46) (Figure 6). The two-way ANOVA of the congruence response rate showed no interaction (F(1, 28) = 1.562, p = 0.222, η2p = 0.053), no main effect in the group (F(1, 28) = 0.829, p = 0.370, η2p = 0.029) and main effect in the conditions (F(1, 28) = 224.261, p < 0.01, η2p = 0.889). In the main effect in the conditions, a post hoc power analysis using G*Power 3.1.9.2 (Faul et al., 2007) indicated that the achieved power was high (Power ≈ 1.00). The descriptive statistics of congruence response for each condition in the Non-illusion group were as follows: the Minimum delay condition (mean = 0.752, SD = 0.198) and the Heaviness condition (mean = 0.079, SD = 0.152).

The results of the correlation analysis between the minimum and maximum Q2 mean values show that no significant correlations were found in the Illusion group (ρ = 0.315, p = 0.409) or in the Non-illusion group (ρ = −0.168, p = 0.466). These results indicate that individuals who reported higher sense of body lightness did not necessarily report a higher sense of body heaviness, indirectly suggesting that the illusion group was not a population with high demand characteristics.

For the SPQ, there was no significant difference (U = 66.000, r = −0.285, p = 0.125) between the Illusion (median [IQR] = 27 [17–31]) and Non-illusion groups (median [IQR] = 20 [15–25]). A post hoc power analysis using G*Power 3.1.9.2 (Faul et al., 2007) indicated that the achieved power was low (Power ≈ 0.307). For the BDI-II, there was no significant difference (U = 99.500, r = 0.042, p = 0.824) between the Illusion (median [IQR] = 8 [5–9]) and Non-illusion groups (median [IQR] = 6 [5–11]). A post hoc power analysis using G*Power 3.1.9.2 (Faul et al., 2007) indicated that the achieved power was low (Power ≈ 0.054). In the HSPS, there was no significant difference (U = 120.500, r = 0.215, p = 0.244) between the Illusion (median [IQR] = 81 [73–92]) and Non-illusion groups (median [IQR] = 92 [78–101]). A post hoc power analysis using G*Power 3.1.9.2 (Faul et al., 2007) indicated that the achieved power was low (Power ≈ 0.44).

The purpose of this study was to investigate whether the illusion of a sense of body lightness can be induced by applying a visual delay system. Our previous study (Hayashida et al., 2025) showed that subjectively preceding feedback can generate the illusion of a sense of body lightness; however, as it was based only on the introspective reports of a very small number of participants, it was uncertain whether the subjectively preceding feedback was actually generated and thereby caused this illusion. All the current participants in the Illusion group reported that a sense of body lightness was induced only when they were given subjectively preceding feedback. Identifying participants who could be classified into the Illusion group largely achieved the primary purpose of this study. However, the proportion was only approximately one-third, and the reasons for this and the underlying mechanisms remain largely unclear. In the following sections, we discuss these issues cautiously and aim to develop this preliminary, exploratory study into future research.

The lower congruence response rate in the Lightness condition than in the Minimum delay condition indicates that incongruence occurred when inducing the illusion of a sense of body lightness. There was no significant difference in the congruence response rate between the Lightness and Heaviness conditions. This might indicate that a sense of bodily lightness, as opposed to a sense of bodily heaviness, is not unique when considering mechanisms of body-sensation illusions based on sensorimotor incongruence. In addition, the results of the two-way ANOVA showed that there was no difference between the groups, and the Heaviness condition was similar to that of the Non-illusion group, indicating that the Illusion group was not uniquely less likely to feel a sense of body heaviness.

In studies using time delay tasks, it is common to fit a logistic curve to the Yes/No response probability for Q1 across delay times and to treat its slope (i.e., sensitivity) and the point of subjective equality (PSE) as the primary dependent variables (Franck et al., 2001; Shimada et al., 2009; Shimada et al., 2010; Kannape and Blanke, 2013; Shimada et al., 2014); however, we chose an analysis procedure based on Q2 for the following reasons. First and foremost, the main purpose of this study was to investigate whether or not there are people who feel a sense of lightness. Therefore, first of all, it was necessary to make groupings. Second, the analysis of congruence using a time delay task generally uses a method that regresses the amplification of the incongruent response rate as the delay level increases on a logistic curve (S-curve). In this case, however, a sine-like curve (rather than an S-curve) is assumed in which the discrepant response rate is amplified for the time being as the delay level increases, but the discrepant response rate decreases owing to the “periodic motion.” Therefore, the sensitivity of congruence cannot be detected by logistic curve regression analysis.

Before considering why incongruency could produce positive effects, we would like to situate the present study methodologically and discuss the future potential of this novel bodily illusion by referring to recent psychophysical literature on bodily illusions. In the present experimental paradigm, the sense of body lightness was assessed using a questionnaire. It is well known that questionnaire-based measures are influenced by introspective ability, cognitive bias, demand characteristics, and memory processes, as such questionnaires are typically administered after illusion induction. To address these limitations, recent work has introduced quantitative, model-based methods to measure the sensitivity of body ownership and embodiment to multisensory integration, including approaches based on signal detection theory and psychometric function estimation (Chancel and Ehrsson, 2020; Lanfranco et al., 2023; Lanfranco et al., 2024; Lanfranco et al., 2025). In these studies, instead of conventional subjective ratings, discriminating body ownership perception in a two-alternative forced-choice task using two rubber hands has been employed. These studies have demonstrated that the multisensory integration window underlying body ownership illusions is considerably narrower—both spatially (Kalckert and Ehrsson, 2014; Lloyd, 2007; Preston, 2013) and temporally (Shimada et al., 2009)—than previously suggested by earlier research. Applying such psychophysical approaches may further enhance the precision with which the body lightness illusion can be elicited and quantified. Although the present report focuses on the presence or absence of the illusion, we believe that this phenomenon holds substantial potential for revealing positive effects in future investigations.

In previous reports, the effects of incongruency were negative sensations, such as a sense of body heaviness or pain. On the other hand, another previous study (Ito et al., 2022) reported the subjectively preceding visual feedback made the perception of muscle fatigue weaker than normal feedback. This previous study and the present study differ substantially in several respects, including the target sensations (perception of muscle fatigue vs. bodily lightness), the context (motor control vs. bodily illusion), and the dynamism of movement (upper-limb movements vs. whole-body walking). In addition, whereas the previous study focused on the alleviation of negative sensations, the present study addresses the generation of a novel bodily illusion. Nevertheless, both studies suggest that the perception of bodily states can be positively modulated by the sign of prediction error, and our findings are therefore conceptually consistent with this framework. That is, the current study may have induced a sense of body lightness because the predicted sensation exceeded the actual sensation. Visual information from the subjectively preceding feedback was given as indicating the current state (actual sensory experience) and may have induced a sense of body lightness by exceeding the predicted body position of the self, based on somatosensory information. This positive effect of the illusion of a sense of body lightness occurred because of the subjectively preceding feedback that could be generated by the combination of the most important features of walking, namely periodicity and the visual delay system. In this sense, our results do not contradict previous reports but rather complement and extend them by showing that sensorimotor incongruence, depending on its direction, can generate not only negative but also positive subjective bodily experiences.

When given positive or negative feedback, whether the feedback is attributed to the self may also be important. In intentional binding experiments, several reports found that the sense of agency increased when the feedback was positive and decreased when feedback was negative (Takahata et al., 2012; Yoshie and Haggard, 2013). From these findings, we can infer that even with a slight incongruence of the subjectively preceding feedback, the cognitive process of attributing positive images to the self may have been executed. The timing of the self-response Q1 and Q2 after the visual feedback in our experiment was optional; thus, it is possible that the participants responded after the cognitive process of attributions or not to themselves. Although the size–weight illusion is conceptualized at the sensorimotor level and relates to vision and proprioception, studies on intentional binding have addressed it at the action level by manipulating reward (Takahata et al., 2012) and emotion (Yoshie and Haggard, 2013). Whether considering the size–weight illusion or intentional binding, although they involve different sensory modalities or computational mechanisms, they share a common feature: positive illusion effects induced by incongruences between prediction and feedback. This suggests that an integrated mechanism encompassing both the sensorimotor level and the action level operates in the background.

These attribution processes should be involved in inducing not only a sense of agency but also a sense of ownership. A previous study successfully induced a sense of ownership and agency in participants toward the avatar’s body, despite the lack of a perfect match between intent and the actual movement (Burin et al., 2020). This indicates a certain degree of flexibility in embodiment sensations. Our procedures may also induce a sense of ownership and agency for the body projected on the screen. For our participants, the “tolerable incongruence” that can induce a sense of ownership may have induced a different sensation, including that of lightness. This previous study compared the first-person perspective with a third-person perspective, and our study corresponds to the third-person perspective. Another difference from the previous study was that in our setup the body shown was not an avatar but the participant’s own projected body, and both the intended movement and the actual movement were those of the participant. Perhaps the sense of ownership could have been enhanced if our procedure had used a first-person perspective, and the effects on congruence response rate and the sense of body lightness might also have been strengthened.

One of the well-known first-person perceptual illusions is vection, the illusory perception of self-motion induced by visual motion. The term vection is defined as the illusory sense of self-motion (Palmisano et al., 2015). A classic example is the train illusion, whereby observing a forward train begin to move creates the illusory sensation that one’s own train, although stationary, is in motion. Although vection can be induced and modulated through various senses, including auditory (Väljamäe, 2009), tactile (Kooijman et al., 2022) and biomechanical stimulation (Ash et al., 2013), this phenomenon can mainly occur when misleading visual signals convey a sense of motion, overriding the vestibular signals that would otherwise indicate a lack of movement. It has been suggested that, under conditions in which vestibular signals are weak, the brain may generate an illusion by interpreting wide-field visual motion as self-motion (McManus and Harris, 2023). If the present experimental paradigm had been conducted from a first-person perspective, the sensory modality modulated by the visual stimulus might have been the vestibular sense, potentially inducing a sense of body lightness originating from vestibular signals. In contrast, in the present study, the illusion appears to have arisen from mismatches in the positions of the upper and lower limbs. That is, in the current experimental paradigm, the sensory modality modulated by the visual stimulus may have been proprioception, and the induced sense of body lightness may therefore have originated from proprioceptive signals. An integrated understanding of sense of body lightness—namely, whether it arises from a specific sensory modality or from the combined contribution of multiple modalities—would be desirable.

There were no significant differences between groups on any of the psychological measures assessed in this study. Although the mechanisms underlying the sense of agency and the sense of bodily lightness partly overlap in that both involve monitoring the congruence between sensory prediction and actual sensory feedback, the present results suggest that the psychological constructs measured in this study are not shared. In a previous study, the cutoffs for the SPQ were 41 and 12 (Raine, 1991). The means for both groups were within the cutoffs. The BDI-II classifies the lowest severity of depression at 13 or below (Beck et al., 1996). The participants in this study were within this range. As depression is associated with a sense of agency (Obhi et al., 2013; Malik and Obhi, 2019; Nobusako et al., 2020), if BDI-II values were dispersed, they might have also affected the sense of agency and sense of body lightness Although there is no clear cutoff for the HSPS, the mean score of 86.68 ± 12.70 from Takahashi (2016) is similar to that of both groups and seems to be an average value. Similar to depression, this value could be related to the sense of agency if sufficient variance were present.

This study has several limitations that should be resolved. First, individual differences in stride times were not considered. To generate the subjectively preceding feedback, a delay time close to that of one stride was required. For example, the meaning of a delay time of 1,000 ms would differ for participants with a single stride time of either 1,000 ms or 1,200 ms; if a participant with a stride of 1,000 ms was given feedback with a delay of 1,000 ms, he or she would almost feel this as being real-time feedback. Hence, the insertion of a 1,200 ms delay would be necessary for a participant with a stride time of 1,200 ms to feel this as being real-time feedback. As the 200 ms delay in our study (Hayashida et al., 2025) was noticed by most participants, if the setting was 200 ms shorter than the individual’s stride time (if the single stride time was 1,000 ms, the delay time would be 800 ms), some participants in the Non-illusion group might have experienced the subjectively preceding feedback.

To address this issue, robustness could be improved by measuring each participant’s stride time in advance and setting individualized delay durations based on that value. In addition, the question of how much temporal mismatch is tolerable remains unresolved; therefore, it would be necessary to utilize mismatch detection sensitivity assessed by Q1. For example, if a participant’s 50% threshold for mismatch detection is 200 ms, it would be necessary to provide predicted feedback that exceeds 200 ms by a small margin.

Second, the temporally incongruent visual feedback used in this study was from a third-person perspective. This is based on our previous study (Hayashida et al., 2025). However, the visual feedback originally provided was from a first-person perspective. For example, temporal incongruence between first-person perspective visual feedback, such as optical flow (Warren et al., 2001; Fajen, 2013), and walking speed may induce a strong sense of body lightness or heaviness. As the first-person perspective involves visual feedback with which we are familiar, it may be able to induce a sense of body lightness better than a third-person perspective can. This issue can be solved immediately with the use of virtual reality technology. Third, it is unclear whether it is possible to counteract this and create the illusion of body lightness in patients who complain of a sense of body heaviness. Patients may have incongruities in many aspects of motor, sensory, and cognitive functions. As the manipulations used in the current study were based on visual information only, future studies should clarify whether incongruence can be corrected across sensory modalities. Fourth, we believed that this experimental paradigm could be related to the sense of body ownership. As the participants knew that what they saw on the screen was themselves, the integration of visual and somatosensory perception might have induced a sense of ownership. To investigate the extent to which this sense of body ownership is involved in inducing the sense of body lightness, it is necessary to quantify the sense of body ownership on a Likert scale in future studies. In our experiment, we can only indicate that it is a phenomenon based on the relative timing of tactile, proprioceptive, and visual feedback.

The most important characteristic of walking, periodicity, was involved in inducing the illusion of a sense of body lightness through the interesting phenomenon of subjectively preceding feedback. We propose that this phenomenon is an important discovery that could not have been found in previous experiments using upper-extremity tasks only. This new kind of illusion could be adapted to physical movements other than walking and developed into an interdisciplinary study. Although its adaptations, limitations, and mechanisms remain unknown, we believe that, in the near future, the development of this study may help patients suffering from unpleasant subjective experiences, such as a sense of body heaviness or pain associated with movement caused by incongruence.

Future studies should be designed with adequate a priori statistical power to detect small-to-moderate group differences in psychological scale scores. Based on the observed effect sizes in the present study, larger sample sizes will be required to reliably identify between-group differences using nonparametric tests. Importantly, a priori sample size planning informed by theoretically or clinically meaningful effect sizes, rather than post hoc estimates, will be essential to improve the robustness and interpretability of future findings.

Herein, we report an experience of the illusion of a sense of bodily lightness while walking. This illusion, related to the mechanism of sensorimotor incongruence, was induced by subjectively preceding feedback through the application of a visual delay system during walking. This study was preliminary and exploratory in nature and included several limitations and unresolved issues. However, this new and simple idea may help individuals who suffer from unpleasant subjective experiences caused by incongruence in the future.