Physical inactivity has been identified as the fourth leading cause of death globally [1]. It is now well established that a sedentary lifestyle is a unique risk factor for several illnesses, such as type 2 diabetes and cardiovascular diseases, which account for about 30% of global mortality [2]. Exergames represent a promising approach that has been widely examined for various population groups (ie, children [3], young individuals [4], and older adults [5]) to promote regular exercise (defined as planned, structured, repetitive, and intentional movement intended to improve or maintain physical fitness) in unmotivated or inactive target groups [6,7].

In recent years, exergames have been proven to have the potential to improve enjoyment, motivation, and long-term engagement when compared with other conventional exercises (eg, cardiovascular exercises like biking [8,9]), and as such, they can be effective in promoting both physical and mental health [10,11]. Various nonimmersive virtual reality (VR) [12] (like using interfaces such as a flat-screen television/monitor) exergames have been designed to encourage people to be more active [5] and promote a positive lifestyle [13] and self-care [14]. Previous literature has shown that exergames could bring physical and health outcomes to players. For example, Peng et al [15] performed a meta-analysis of energy expenditure in exergames, and their main finding suggests that exergames are as effective as traditional exercises in facilitating light- and moderate-intensity physical exertion. Huang et al [16] reported that participants who were enthusiastic about exercising showed positive changes in happiness, perceived energy levels, and relaxation in a 2-week exergame intervention. Sapi et al [17] reported that participants showed improvements in balance following a 6-week exergame intervention, and the improvements were in favor of using the exergame than conventional balance training. da Silva Alves et al [18] found that participants showed improvements in functional well-being and physical well-being after 10 sessions of exergaming. In the study by Garcia et al [19], participants showed improvements in stepping, standing balance, gait speed, and mobility following a 12-week exergame intervention.

With the recent advances in immersive virtual reality (iVR) head-mounted displays (HMDs), an increasing number of iVR exergames [20-22] are being developed. They have opened the possibility of altering more radically how we engage users in performing exercises. Studies have shown that iVR has the potential to produce benefits that other types of displays (ie, a standard display like television) cannot offer. For instance, iVR exergames can offer users the illusion of more exceptional physical capabilities than they have. As such, iVR may increase motivation for exercising in general [23]. Moreover, iVR games can offer benefits such as increased perceived competence and the feeling of body movements that are more in line with how we perform exercise in the real world. Participants have described exaggerated movements to be natural, fun, and empowering [24]. Furthermore, exercising in iVR has been found to be an effective intervention to increase enjoyment and motivation than standard televisions or monitors [8,25], where enjoyment and motivation are, in turn, linked to increased adherence to physical exercises in general [26-28].

Full-body gesture-based exergames have been widely explored with people in standing positions in iVR [20,22,23]. However, they have not been adapted and explored in seated versions. Seated iVR exergames could have the following benefits: (1) suitability for users with a sedentary lifestyle (eg, university students [29]); (2) feasibility for mobility-impaired users (eg, elderly users and wheelchair users [5,30]); (3) possibility of reducing the risk of injuries due to falls or motion sickness [31,32]; and (4) avoidance of injury from hitting other objects (eg, furniture) when the space is small or surroundings are cluttered, because players are not required to walk around.

Motivation can be divided into intrinsic (enjoyment of the activity) and extrinsic (driven by external outcomes, eg, losing weight and improving fitness) [33]. Intrinsic motivation (ie, motivation derived from enjoyment and satisfaction gained from an activity) plays an essential role in long-term adherence to exercising [26,27,33], whereas extrinsic motivation, such as competitive pressure, may lead to tension and feelings of compulsion, and can diminish intrinsic motivation [34,35]. There is evidence that exergames increase enjoyment and intrinsic motivation compared with conventional exercises (eg, biking) and distract from uncomfortable bodily sensations [25,36-39].

A sedentary lifestyle is a problem for older adults and people with physical disabilities and is a serious health problem among university students [40]. Most of the research on exergames has been targeted at older adults or disabled people [5,30,41] but not university students, who are underrepresented in such studies. Research has shown that lack of time and not liking exercising are the major barriers for university students [29]. These barriers could be overcome by using full-body gesture-based exergames that can be played either standing or seated at any time and in small spaces because exergames are perceived to be more enjoyable and preferred by university students than other conventional exercises (eg, cardiovascular exercises like biking [8,9]).

The focus of this research was to evaluate the playability and user experience of a seated iVR exergame compared with a similar standing exergame among university students in terms of gameplay performance (ie, action completion time and number of gestures missed) and user experience (ie, motion sickness, intrinsic motivation, and fear of falling).

We employed a one-way within-subjects experiment design where the independent variable was game mode with two levels (standing and sitting). The order of the game mode was counterbalanced to compensate for any learning effects. The whole experiment lasted between 30 and 40 minutes for each participant depending on their tiredness level and resting heart rate (RestHR).

Participants were recruited from a local university campus through posters, social media platforms, and a mailing list. The study included university students with the ability to speak English, who were not disabled, were not pregnant (because of the physical exertion required to play the game), and had not consumed any alcohol during the day (blood alcohol level of approximately 0.07% could reduce symptoms of cybersickness [42], which might affect the results of our study).

Participants were excluded from the experiment if they (1) answered “yes” to any of the Physical Activity Readiness Questionnaire (PAR-Q) [43] questions, (2) had a resting blood pressure higher than 140/90 mmHg, and (3) had a RestHR level that was too low (ie, RestHR <62 beats/min for a 16 to 19-year-old female, RestHR <60 beats/min for a 20 to 39-year-old female, RestHR <56 beats/min for a 16 to 19-year-old male, or RestHR <55 beats/min for a 20 to 39-year-old male) or too high (ie, RestHR >94 beats/min for a 16 to 19-year-old female, RestHR >89 beats/min for a 20 to 39-year-old female, RestHR >87 beats/min for a 16 to 19-year-old male, or RestHR >84 beats/min for a 20 to 39-year-old male) [44].

All participants received drinks and snacks for their participation after they finished the experiment. The University Ethics Committee at Xi’an Jiaotong-Liverpool University approved the experiment. All participants signed informed consent forms prior to taking part in the study, and the research protocol was approved by the University Ethics Committee.

To determine the sample size for the study, a statistical power analysis was performed. This statistical power analysis was not based on data from prior studies owing to limited similar work. It was based on general considerations about the trade-offs between the ability to detect certain effects and the feasibility to acquire a sufficiently large sample. We used a sample size calculation software program (G*Power version 3.1.9.2 for Windows), with an effect size of 0.5 (Cohen d), statistical power of .90, and statistical level of significance of .05. The sample size was established at 44 participants, and we decided to recruit an additional 10 participants in the case of dropouts.

The game was implemented in Unity3D with the SteamVR plugin to enable positional tracking of the HTC VIVE trackers and controllers.

The experiment was conducted using HTC VIVE Pro Eye connected to an HP Z workstation (i7 CPU, 16 GB RAM, and Nvidia Quadro P5200 GPU). Two HTC VIVE handheld controllers, two HTC VIVE trackers, and two base stations were used to enable hand and feet motion tracking. A stable chair with two handles was used in the sitting condition. The HR was monitored by a Polar OH1 optical HR sensor, which has been proved to be able to capture good HR data when compared with the gold standard of HR measurement with an electrocardiography device [62,63]. The experiment was conducted in an indoor laboratory room that could not be seen from the outside. The room temperature was always set to be 23 to 24°C during the experiment. Figure 3 depicts the experiment setup and devices used in the experiment.

Before filling in the pre-experiment questionnaire that gathered demographic information (eg, age, gender, and experience with the VR device), we obtained participants’ consent to participate in the experiment and collected their RestHR and resting blood pressure. They were also asked to enter their age, gender, height, and weight into the Polar Beat app. Before each condition started, a researcher would help them to wear the VIVE Pro Eye headset with two VIVE handheld controllers and two VIVE trackers. Participants were then given 3 minutes to get familiar with the corresponding condition of the iVR exergame. Once their HR reached the equivalent RestHR level, they proceeded to the experiment stage, beginning with calibrating the game, training, and testing. In each condition, they needed to perform 120 gestures, requiring 5 minutes (120 gestures × 2.5 seconds). We fixed the number of gestures to allow for comparing the two games. After each condition, they were asked to fill in postexperiment questionnaires. They proceeded to the next condition when they felt rested and their HR was at the rest level.

We used the paired t test to understand the difference between the seated exergame and standing exergame regarding gameplay performance, IMI, MSAQ, fear of falling, and exertion measurements. For the percentage of missed gesture types for sitting/standing gesture type, we used one-way repeated analysis of variance (ANOVA) with the sitting/standing gesture type as the within-subjects variables, respectively. We also examined and reported if there were any significant gender differences in our measurements by using one-way between-subjects ANOVA. We set the α level at .05 in our analyses. We further reported the effect sizes using Cohen suggestion to classify the effect size, where Cohen suggested that d=0.2 represents a “small” effect size, 0.5 represents a “medium” effect size, and 0.8 represents a “large” effect size [64]. Analyses were performed using the Statistical Package for the Social Sciences (IBM Corp).

Fifty-four individuals were interested in participating in the experiment. Two were excluded owing to their high RestHR. At the end, a total of 52 participants were eligible to participate in the study. The characteristics of the study participants are shown in Table 2.

Gameplay data and analysis results are reported in Table 3. The analysis showed that game mode did not influence action completion time. However, the analysis showed that players missed more gestures in the seated exergame than in the standing exergame.

Exertion (AvgHR%, increased HR%, calories burned, and Borg 6-20) data and analysis results are presented in Table 3. The analysis yielded significantly higher AvgHR%, increased HR%, calories burned, and Borg 6-20 RPE for the seated exergame (all P<.001). Our results suggest that the seated exergame was rated as “very light” to “light” exercise and the standing exergame was rated as “very light” exercise according to the Borg 6-20 RPE scale.

To aid the visualization of the AvgHR% behavior of both exergames, Figure 5 shows the AvgHR% data from all participants during the 5 minutes of gameplay, averaged over the whole session. The seated exergame had a higher AvgHR% than the standing exergame after 0.34 minutes.

Data analysis of the MSAQ included the overall MSAQ score and its subscale scores (gastrointestinal, central, peripheral, and sopite-related). The MSAQ data and analysis results are reported in Table 4. The analysis showed significantly higher peripheral sickness (P=.02) and sopite-related sickness (P=.004) for the seated exergame. We did not find a significant difference between the seated exergame and standing exergame in terms of central (P=.81), gastrointestinal (P=.81), and overall sickness (P=.06).

Regarding IMI, there was no significant difference for interest/enjoyment (P=.96), perceived competence (P=.26), and pressure/tension (P=.42). However, the analysis yielded a significantly higher value/usefulness for the seated exergame (P=.04). IMI data and analysis results can be found in Table 4. The one-way between-subjects ANOVA showed that interest/enjoyment was rated significantly higher by females (mean 5.20, SD 1.23) than males (mean 4.38, SD 41.16), with a medium effect size (F_1,50=6.01, P=.02).

With the limited exploration of seated exergames in the literature of iVR exergames, this study is the first to explore the difference between full-body gesture-based seated exergames and standing exergames in iVR among university students regarding playability (ie, gameplay performance) and user experience (ie, intrinsic motivation and motion sickness). Our results suggest that participants perceived higher value in the seated exergame than in the standing exergame. However, the seated exergame was associated with a worse gameplay performance (ie, the number of missed gestures) and a higher rating of motion sickness than the standing exergame.

Although we observed that participants missed a significantly higher number of gestures in the seated exergame than in the standing exergame (P=.02), this rate was as low as 1.9% (2.3/120). Further analysis of the type of gestures missed in the seated exergame confirmed that these misses were in the early stages of the experiment, and as such, the main reason for the misses could be because of participant unfamiliarity with exercising in the seated position (none of them had previous experience of seated exercising; Table 2).

Regarding motion sickness, previous studies [31,32] have suggested that the seated exergame might result in a lower level of motion sickness. However, this was not supported by our findings. We observed that participants felt sicker (ie, peripheral and sopite-related motion sickness) in the seated exergame. However, the reason was beyond the scope of this study; a further investigation is required to understand why motion sickness was higher in the seated exergame than in the standing exergame. We suggest that future designers and researchers should carefully design full-body gestures for iVR seated exergames to minimize motion sickness.

As for intrinsic motivation, we did not observe any significant difference between the seated exergame and standing exergame (ie, interest/enjoyment, P=.96; perceived competence, P=.26; pressure/tension, P=.42). However, there was a gender effect on participants’ intrinsic motivation toward exergames, where we found that females had a significantly higher intrinsic motivation (ie, interest/enjoyment) than males (P=.02). This could be because the exergame involved in our study was more like a dance game, which was inspired by Just Dance, and prior research [65] has shown that females tend to be more physically active in dance exergames. Aside from this difference between male and female participants, no other differences were found in our experiment.

In most cases, standing exercises have a higher exercise intensity (in traditional exercises [66,67] and exergames [68,69]). We found that our seated exergame led to a higher exertion (ie, AvgHR%, increased HR%, calories burned, and Borg 6-20) than the standing exergame, possibly because the seated exergame involved more whole-body movements that required increased energy expenditure during gameplay [70-73].

In this section, we provide design guidelines that are based on suggestions provided by Wiemeyer et al [74] for future game designers who are interested in building iVR full-body standing or seated gesture-based exergames.

First, practice should be provided for each gesture during warm-up. A warm-up session before exercising is essential [75], and it should be included in exergames as well. One way to perform warm-up for full-body gesture-based exergames is to practice the gestures involved in the game, which will not only help players reduce the risk of injures but also make them familiar with the in-game gestures.

Second, the difficulty level of the game should be adapted to the current state of the individual. Regarding an offline approach, players might have difficulty in performing certain gestures during gameplay. Therefore, to match the difficulty of the game to the current state of the individual, it would be necessary for players to experience and select gestures they are comfortable performing before playing the game. Regarding an online approach, one of the adaptive methods that has been used and proven to be suitable in exergames is proportional–integral–derivative (PID) control [76]. Designers can use PID control to modify the transition time between gestures or select the gestures based on the player’s real-time HR and gameplay performance (ie, the number of gestures missed). PID control is also useful to avoid overly vigorous exercise, which might put the exerciser at risk of eliciting unwanted coronary issues [77].

Third, warning signs should be provided for standing exergames if users have left (or are about to leave) the calibration position and are too far to keep them protected. This is because players tend to move around during gameplay, which we encountered in our study and has been reported previously [23]. It could lead to potentially dangerous situations (eg, hitting objects that are in the environment and going out of the safe tracking area) or decrease the recognition performance of the sensors (eg, tracking may not work when they are too close to or far from the sensors).

Our study only focused on one sedentary lifestyle user group (university students). Future work could focus on investigating the two versions of the exergame with different population groups (eg, older adults and users who have physical disabilities). To minimize players’ cognitive workload, we set both exergames to include only 6 out of the 10 gesture types that we measured during the pilot study. In the future, we could add more gestures to increase the complexity of the game (as stated by participants 13, 20, 30, and 37). A further limitation is that our experiment did not measure which types of gestures caused the unwanted level of motion sickness in the seated exergame. Future experiments could be conducted to check issues related to motion sickness based on specific gestures and types of gestures.

Our contributions to the field of iVR exergaming regarding gameplay performance and user experience are as follows: (1) the iVR seated exergame could result in higher exertion and provide higher value to players than the standing exergame; (2) participants might feel sicker in the iVR seated exergame than in the standing exergame, and as such, full-body gestures for seated exergames need to be designed carefully to help minimize the feeling of motion sickness; and (3) participants might miss more gestures in the iVR seated exergame than in the standing exergame. Therefore, designers should allow more time for performing gestures in the seated exergame.