To investigate Visually-Induced Motion Sickness (VIMS) and visual effects associated with the Head-Mounted Displays (HMDs) most commonly used in Virtual Reality (VR) systems.
A comprehensive search query was performed on the Medline/PubMed, EMBASE, CENTRAL, ACM Digital Library, and IEEE Xplore databases. We identified population-based studies that evaluated HMDs as an independent factor for visual discomfort. Potential variables relevant to HMD discomfort, including system features (e.g. optical characteristics), subject characteristics (e.g. gender), and task characteristics (e.g. duration, vection, and task content) were reviewed. Total severity scores of Simulator Sickness Questionnaires (SSQT), oculomotor scores of Simulator Sickness Questionnaires (SSQO), and Visual Strain Questionnaires scores (VSQ) were used to measure HMD discomfort impact.
We analyzed data for 1040 participants from a total of seventeen studies, all published between 1998 and 2015. Our review demonstrated that exposure to HMDs resulted in higher SSQT and SSQO mean change scores, compared with exposure to traditional displays such as TV and desktop computer displays. Furthermore, HMD exposure duration had a significant impact on the mean change scores of SSQT, SSQO, and VSQ. Our analysis also showed that HMD discomfort was affected by all three of the variables we evaluated.
This meta-analysis qualifies the risk factors causing discomfort after exposure to HMDs. We recommend that HMD manufacturers increase their awareness of, and address, these visual discomfort issues in their products.
Head-mounted display, Virtual reality, Virtual environment, Visual discomfort
HMD: Head-Mounted Display; NED: Near-to-Eye Display; VE: Virtual Environment; VR: Virtual Reality; 2D: Two-Dimensional; S3D: Stereoscopic Three-Dimensional; IPD: Inter-Pupillary Distance; FOV: Field of View; VIMS: Visually-Induced Motion Sickness; D: Disorientation; O: Oculomotor disturbance; N: Nausea; SSQ: Simulator Sickness Questionnaire; VSQ: Visual Strain Questionnaire; DK2: Developmental Kit 2; CI: Confidence Interval; SS: Simulator Sickness
In recent years, Virtual Reality (VR) has become more commonplace, with rapid adoption into daily life. VR is a non-invasive simulation technology that provides an immersive, realistic, three-dimensional (3D) computer-simulated environment in which people perform tasks and experience activities as if they were in the real world. The most direct experience of VR is provided by fully immersive VR systems. The most widely adopted VR systems display is a simulated environment through special wearable head-mounted visual displays (HMDs). HMDs have evolved over the past five years from tethered systems comprising of screens and lenses fitted into a helmet, to relatively inexpensive systems that utilize mobile smart devices and fit into a light weight lens system. The optics within the HMDs vary from monocular (one eye view), binocular (both eyes view screen) and dichoptic (both eyes view different screen/image or image can be stereoscopic, adding depth cues). Recent advancements in hardware have included eye tracking and the use of multifocal optics.
Although HMDs have recently been introduced to the general public, they are not a new phenomenon (Table 1). As early as the 1960s, computer graphics pioneer Ivan Sutherland developed the first HMD, which made it possible to overlay virtual images on the real world [1,2]. HMD technology gradually evolved through the 1970s with use across military, industry, scientific research and entertainment domains. The early commercially available HMDs had limited applications due to their narrow Field-Of-View (FOV) and inherent cumbersomeness in weight, physical restrictions, and system parameters. Recent advancements have been directed toward making HMDs more comfortable for longer duration of use. Recent HMD products including Samsung Gear, HTC Vive, Oculus Rift, FOVE, and Google Daydream have become commercially available and increasingly commonplace as a result of technical advancements. For example, the latest version of the Oculus Rift at this time, the Development Kit 2 (DK2), has a higher resolution, higher refresh rate (i.e., the frequency with which a display's image is updated), lower persistence (which aids in removing motion blur) and more advanced positional tracking allowing for precise movement, when compared to its predecessor. FOVE has introduced eye tracking with real time foveal rendering to improve user experience. HMD technology advancement and cost reduction has increased its potential for widespread use.
Table 1: Optical characteristics of representative head mounted displays in the systematically reviewed articles. View Table 1
Visually induced motion sickness (VIMS) or simulation sickness, remains an obstacle to the widespread adoption and commercial development of technologies associated with VR based HMDs [3,4]. With occlusive HMD systems, which by definition, is the distinguishing factor of virtual reality vs. augmented and mixed reality systems, a user is dependent on the VR system for sensory input. This dependency involves synchrony in sensory input, and the lack of this synchrony lends to visual-vestibular mismatch. The symptoms of visual-vestibular mismatch include nausea, stomach discomfort, disorientation, postural instability and visual discomfort.
It is commonly accepted that the symptoms of nausea and instability result from various sensory input conflicts, including conflicting position and movement cues, leading to a disharmonious effect on the visual and vestibular systems [5,6]. In addition, specific types of HMDs might have mismatch problems with the user's visual system due to improper optical design, resulting in convergence-accommodation conflict and visual discomfort or fatigue [7-13].
Early evaluation of the side effects of HMDs showed variable and inconsistent results. Notably, Peli reported no objective functional visual differences between HMDs and conventional desktop computer displays [14].
Conversely, other early studies reported high incidence of visual discomfort including eyestrain, dry eye, tearing, foreign body sensation, feeling of pressure in the eyes, aching around the eyes, headache, blurred vision, and difficulty in focusing. For example, Mon-Williams, et al. found that following a 10-minute exposure to a stereoscopic VR display, 60% of study participants reported symptoms of eyestrain, headache, and nausea [15]. This finding has been confirmed in a number of more recent studies [16-26].
Other visual problems such as myopia, heterophoria, fixation disparity, accommodation-vergence disorders, and abnormal Tear Break-Up Time (TBUT) also have been reported [15,18,20,21,24,27,28]. Using HMDs may cause accommodative spasm that in turn may lead to a transient myopia [20]. Continued conflict between convergence-accommodation, the user's Inter-Pupillary Distance (IPD), and/or the systems' Inter-Optical Distance (IOD) may lead to heterophoria and fixation disparity changes [15,20,29-31]. Moreover, visual symptoms are not necessarily limited to the time of actual Virtual Environment (VE) immersion; rather, visual changes including visual fatigue, reduced visual acuity and heterophoria may continue after terminating exposure to HMD-based VE [25,32-34].
As a result of the recent advancements in the industry of virtual technology, the growing side effects associated with it require thorough documentation and characterization. To our knowledge, there has been no review article on the role of HMDs in visual discomfort. The current existing body of literature shows mixed results and different roles for different influential variables. While some HMDs studies have found significant negative impact on visual comfort, others have not. Biocca suggested that the cause of VR-induced sickness could be a technical problem, which would disappear as the technology advanced [35]. Unfortunately, this has not been the experience so far as technological advancements have not significantly reduced visual problems [12,13,36-47] (Table 2). Therefore, the extent to which HMD design impacts visual discomfort is unclear.
Table 2: Summary of empirical data from head mounted display meta-analysis studies. View Table 2
We conducted a meta-analytic review of publications related to the visual effects of HMDs. By compiling data from multiple independent studies, this study allowed for a broad review and large population sample size in assessing the role of HMDs in visual discomfort. Furthermore, we systematically reviewed factors that might potentially affect the role of HMDs in visual discomfort. These factors included the characteristics of the systems, the participants (i.e., gender), and the tasks (i.e., duration, vection, and content).
We performed an extensive and systematic review of scientific literature. The database search included Medline/PubMed (US National Institutes of Health/National Library of Medicine), Embase (Elsevier), ProQuest Central, ACM Digital Library (Association for Computing Machinery), and IEEE Xplore Digital Library (Institute of Electrical and Electronics Engineers). These databases were searched using the following keywords: VR, virtual environment (VE), HMDs, VIMS, SS, SSQ, VSQ, visual fatigue, and visual problems.
Only papers from peer-reviewed journals and large national conference proceedings were selected for inclusion in our study. For the relevant trials lacking data, we also attempted to contact the corresponding author by email for further unpublished but potentially relevant data; none of these contacts resulted in receipt of additional data. Unpublished data and abstracts were not included. No language restrictions were imposed.
The inclusion criteria consisted of any of the following: (1) Human studies evaluating visual system-related problems in HMD-based VE, (2) Studies involving self-reported questionnaires (SSQ and VSQ). Studies for visual problems without HMD-based VE, case studies with fewer than three participants, or studies without enough data to evaluate the impact of HMDs on visual discomfort were excluded.
Visually Induced Motion Sickness (VIMS) can be measured by psychological and physiological methods. The Simulator Sickness Questionnaire (SSQ), a self-reported measurement, is a well-known psychological method and the gold standard for measuring the extent of VIMS [4,55]. The questionnaire consists of three components: nausea (SSQN), disorientation (SSQD), and oculomotor symptoms (SSQO). The total score of SSQ (SSQT) is an aggregate score of the three components.
The SSQ contains 16 items (i.e general disomfot, fatigue, eye strain, nausea), and each item is scored on a 4-point scale in which 0 = none, 1 = slight, 2 = moderate and 3 = severe [56]. Given that the discomfort caused by HMDs is often connected to visual symptoms, the Visual Strain Questionnaire (VSQ), a more detailed visual strain-related questionnaire, was used to measure the severity of eyestrain symptoms (e.g. tired, sore or aching, irritated, watering or runny, dry, hot and burning eyes; blurred or double vision; and general visual discomfort) [57]. 57 These symptoms of eyestrain are also often connected to computer vision syndrome [58-61]. In this study, the SSQ and VSQ were used to verify the occurrence of VIMS. Figure 1, Figure 2, Figure 3, and Figure 4 compare the mean score of SSQT, SSQO, and VSQ respectively between the various studies analyzed.
Figure 1: A forest plot diagram of mean change of the total severity score of simulator sickness questionnaire (SSQT) in pre- and post-HMDs exposure (Chi2 = chi-square statistic; CI = confidence interval; df = degrees of freedom; I2 = I-square heterogeneity statistic; IV = inverse variance; SE = standard error; P = P value; Z = Z-statistic).
View Figure 1
Figure 2: A forest plot diagram of mean change of the oculomotor scores of a simulator sickness questionnaire (SSQO) in pre- and post-HMDs exposure (abbreviations are the same as those in Figure 1).
View Figure 2
Figure 3: A forest plot diagram of mean change of the scores of visual strains questionnaire (VSQ) in pre- and post-HMDs exposure (abbreviations are the same as those in Figure 1).
View Figure 3
Figure 4: A forest plot diagram of sensitivity analyses of the mean change of the total severity score of SSQT in pre- and post- HMDs exposure (abbreviations are the same as those in Figure 1).
View Figure 4
The following variables listed below were used for primary outcomes. Mean change refers to the difference in SSQT scores between difference HMD devices (i.e. oculus rift versus FOVE).
(1) Mean change in symptom scores between different HMD-based VE.
(2) Comparison of mean change in symptom scores between HMDs and other traditional displays.
The visual impact of HMDs was used as the effect size in the meta-analysis. For multiple session VE exposure studies, we used only the data from the first VE exposure session in order to perform a between-group analysis. For studies with multiple measures of visual discomfort, we included only the SSQ and VSQ outcomes in the meta-analysis.
All statistical analyses were performed with Review Manager Version 5.3 (The Cochrane Collaboration, Oxford, England), using two-tailed p values and a 95% Confidence Interval (CI). For generic inverse variance outcomes, the mean difference was analyzed. Meta regression was performed, and heterogeneity was explored using the Q test with calculating I2, indicating the percentage of variability due to heterogeneity rather than to chance. I2 values of 50% or more were considered substantial heterogeneity. The fixed-effects model was used to pool the data. A random-effects model was applied when P < 0.1 in the test for heterogeneity, and the fixed-effects model was used for other cases [62].
Figure 5 shows the flow chart of the selection process for the reviewed studies. Of 69 potentially relevant studies identified through the electronic search, 34 publications met all the inclusion criteria. After excluding 17 articles because specific data for SSQ or VSQ were not provided, 17 studies with 1040 participants remained for inclusion in the meta-analysis. The smallest sample size in an included article was twelve [47], and the largest was 232 [39]. All 17 studies were published between 1998 and 2015. Eight were performed in the USA; three in Finland; two in Germany; and one each in the United Kingdom, Japan, Spain, and the Netherlands.
Figure 5: The flow chart of the process of article selection.
View Figure 5
Table 2 provides a description of the pre- versus post-exposure mean score changes of SSQT, SSQO and VSQ. These were analyzed in seventeen, seven, and three studies, respectively. In three studies, the mean SSQT and SSQO score changes in exposure to HMDs and traditional displays also were included in the meta-analysis.
The forest plots show that the visual discomfort was significantly different in pre- versus post-VR exposure (Figure 1, Figure 2, and Figure 3). The results show mean difference of 11.54 (95% CI 7.44 to 15.64; P < 0.00001), 7.67 (95% CI 3.76 to 11.58; P = 0.0001) and 1.07 (95% CI 0.8 to 1.35; P < 0.00001) for the SSQT, SSQO and VSQ, respectively.
In order to investigate the possible role of publication year, we undertook a sensitivity analysis by excluding the four studies published before 2000. Older publications assessed older HMDs; consequently, their results may have been skewed in favor of greater VIMS. This analysis showed a mean difference of 8.29 (95% CI 4.80 to 11.77; P < 0.00001) for SSQT in pre- versus post- VR exposure (Figure 4).
Only four studies provided adequate information for meta-analysis of visual discomfort in HMDs compared to traditional displays (i.e., TV, desktop computer displays, or other 2D viewing condition) [37,39,43,54]. The forest plots show that the mean differences of SSQT and SSQO in HMDs versus traditional displays were 3.62 (95% CI 1.47 to 5.78; P = 0.001), and 4.78 (95% CI 1.51 to 8.05; P = 0.004) respectively; these results were statistically significant (Figure 6 and Figure 7).
Figure 6: A forest plot diagram of HMDs versus the traditional displays (TV, desktop, or other 2D viewing condition) for the mean change of the total severity score of SSQT (abbreviations are the same as those in Figure 1).
View Figure 6
Figure 7: A forest plot diagram of HMDs versus the traditional displays (TV/desktop) for the mean change of the oculomotor scores of SSQO (abbreviations are the same as those in Figure 1).
View Figure 7
The previous studies highlighted a number of factors that have the potential to cause stress to the visual system in HMD-based VR. It seems that the stress on the visual system is multifactorial [63] (Table 3). We identified several studies that reported the effects of HMDs from different influence factors. Among these studies, HMDs' optical characteristics (system features), participants' gender (individual characteristics), duration, vection, and task content (task characteristics) were systematically reviewed respectively (Table 4, Table 5, Figure 8, Figure 9, and Figure10).
Table 3: Experimental studies investigating causative factors of HMD-induced motion sicknes. View Table 3
Table 4: Eye symptoms related to head mounted display exposure. View Table 4
Table 5: Eye symptoms related to vection in a head mounted display environment. View Table 5
Figure 8: Mean SSQ scores by gender in different studies (higher scores = higher symptom severity). A: SSQT scores; B: SSQO scores. Error bars represent standard error of mean.
View Figure 8
Figure 9: Mean SSQT scores by duration of exposure in different studies. A: SSQT score and exposure duration; B: SSQT score and post exposure duration.
View Figure 9
Figure 10: SSQ and VSQ Change (Post - Pre) are demonstrated after performing different tasks in two studies (higher scores = higher symptom severity). A: SSQT change; B: SSQO change; C: VSQ change. Vertical lines represent standard error of mean.
View Figure 10
Advances in HMD technology have provided the potential for its widespread use in VR. However, VIMS, as an inherent problem, still remains an obstacle to public adoption and commercial development of this technology. HMD devices, such as Oculus Rift, HTC Vive, Samsung Gear, FOVE and Google DayDream have already entered the market, thus highlighting the importance of further research about VIMS. In our meta-analysis and systematic review, we have demonstrated through our data analysis the presence of significant visual discomfort after exposure to HMDs, when compared to traditional displays, and identified the potential moderating factors for this visual discomfort. To our knowledge, this is the first comprehensive summary and meta-analysis to address this issue.
Although the evaluation methods of HMD-induced visual problems varied between studies and in some studies the details were not provided, the results of this meta-analysis showed that regardless of the evaluation methods, the exposure to HMDs has been associated with significant visual discomfort. This meta-analysis of pre- and post-exposure questionnaires demonstrated significant associations between visual impact and the mean change scores of total severities of SS, oculomotor score of simulator sickness questionnaires scores and visual strain questionnaires scores. Furthermore, HMD exposure was associated with higher scores of mean changes of SSQT and SSQO compared to other traditional displays, such as TV and desktop computer displays.
Theoretically, HMDs with sub-optimal designs were used in the earlier articles, i.e., those published before 2000. The visual discomfort reported in those papers might have been a result of their poor design, rather than valid observations. Therefore, the meta-analysis from those papers published before 2000 may have substantially overestimated the incidence of HMD-induced visual discomfort. For example, the Virtual Research V6 HMD used for Stanney's 1998 study was a mid-range display made by the US-based Virtual Research System Inc. and had a lower brightness and poorer color presentation compared to the updated Virtual Research V8 HMD used for Sharples' 2008 study [32,54]. These advancements in the manufacturer's HMD technology may have produced a higher mean change of the questionnaire score when compared to scores from participants who had used the older Virtual Research V6.
In order to rule out any significant influence from the papers published prior to 2000, (i.e., to assess the reliability of our meta-analysis), our sensitivity analysis excluded the four studies published before 2000. This exclusion did not change the outcome of our meta-analysis; i.e. the result after eliminating the old papers was statistically consistent with the result of original meta-analysis. Even though these questionnaires are self-reported subjective evaluation methods, they provide converging evidence that HMD-based VR causes visual discomfort. These findings are consistent with a cross-sectional survey of 953 questionnaires related to VIMS, in which almost 35% of the respondents reported tired eyes during 3D movies [98]. The results suggest a need for raising public awareness about the visual discomfort that individuals may suffer after exposure to HMDs. We recommend that the HMD industry and manufacturers address the visual discomfort issue before their products become commonly used.
To date, no controlled studies have evaluated the extent to which user subjective responses are determined by characteristics resulting from pre- and post-test measures. Young gave subjects SSQ's either pre- and post-VR immersion, or only post immersion. Participant reports of sickness after immersion in VR showed higher scores when both pre- and post-test questionnaires were given to the participants than when only a post-test questionnaire was used [23]. These results are notable because measurements of sickness by both pre- and post-self-report questionnaires are significantly biased due to demand characteristics, and may substantially overestimate the incidence of HMD-induced visual discomfort. We suggest that comparative studies of the visual effects of HMD-based VEs employ experimental designs that are not subject to such biases, or at least take measures to balance these biases. Alternatively, more objective measures could be used systematically in order to evaluate visual effects after HMD exposure.
Our systematic review also shows that many factors impact HMD-induced VIMS. These factors include the characteristics of the device system, the participants in the studies we included, and the tasks they were asked to perform. Device system variables included viewing mode (e.g. monocular, binocular or dichoptic), headset design (e.g. fit, weight), optics (e.g. misalignment in the optics; contrast, luminance), Field Of View (FOV), and time lag (i.e. transport delay). HMD weight has been associated with the experience of visual discomfort and injury [99]. A key consideration for HMD design must be that weight is within the level of human tolerance to minimize head and neck fatigue. With the decrease in cost of components, HMD design has moved to more ergonomic HMD systems, which has been reflected in the adoption of mobile systems, such as Samsung Gear and Google DayDream.
Symptoms of eyestrain and blurry vision were significantly higher in monocular mode than in other modes [18,67,72]. The use of HMDs in stereoscopic mode is less comfortable than in non-stereoscopic mode [14,17,69-71]. Spatial properties of the display, i.e., Field of View (FOV), may be implicated in producing visual discomfort symptoms [34,51]. The FOV studies show that narrow FOV (< 50 degrees) reduces the perception of self-motion and wide FOV (> 100 degrees) may increase the presence and level of simulator sickness. Patterson, et al. recommend a minimum 60° FOV to achieve a full sense of immersion [100].
Resolution contributes to overall image quality but also directly affects the users' experience of VIMS. It is often uncomfortable to view low-quality images that are noisy or blurry. Anatomically, the central retinal fovea has the highest number of photoreceptors and the highest capacity for resolving an image [101]. While the limit of human visual resolution is 1 minute of arc at the central fovea, few HMDs can achieve this, primarily due to current technological limitations. The result can be a pixelated experience [101]. It is important to provide the highest possible resolution in the central field of view of the virtual environment to truly simulate a real-life experience and mimic the viewing characteristics of human vision. Some devices attempt to achieve this by creating a gradient of resolution with the highest resolution in the central field of view and the lowest resolution at the periphery [101,102]. The potential trade off of higher resolution is the overexposure of energy from the display, given the proximal distance of the display to the eyes.
Time lag between an individual's action and the system's reaction potentially could influence a user's experience of VIMS symptoms, as it affects human perception of visual and vestibular cues [33,51,64,65]. Therefore, reducing the sensor error of HMD systems may minimize the VIMS experience. HMD optical characteristics, such as eye relief (a fixed distance from the eyepiece lens to its exit pupil), convergence demand, horizontal disparity, vertical misalignment of displays, inter-ocular rotation difference, vertical-horizontal magnification differences, luminance, focus differences, temporal asynchrony, focal distance, field curvature difference and Inter-Pupillary Distance (IPD), are all potential factors that can induce visual discomfort and headache when they are poorly aligned or adjusted [36-39, 66,68,73-75,100,103-106]. Another consideration is the type of optics used in HMD systems, single focus lens systems vs. multifocal lens systems, with the latter lending to more natural adaptive zones for multiple focal lengths typically encountered in VR environments, especially interactive experiences.
High rates of ocular symptoms may be associated with certain device characteristics (Table 1). For example, the VIMS experienced with the Vuzix may be attributable to the focal distance of the display (5 m/2.4 m), luminance level (24 cd/m2), and vertical misalignment (0.8°). Conversely, the low ocular symptom rate in the MyVu may be explained by its luminance level (97 cd/m2), lack of vertical misalignment (0°), physical design and assigned tasks. It is essential to note that technical advances have reduced obvious problems such as physical ergonomic issues including HMD weight, system time delay, and luminance.
It is also important to note that the conflict between visual and vestibular input remains a significant problem. In other words, the VIMS from HMD exposure is not simply a "technical problem" that will be resolved as the technology advances. The discomfort from the vestibular-visual mismatch will not resolve unless the mismatch itself resolves, which may involve a multifactorial process of synchronizing sensory inputs dependent on hardware specifications and software/content design.
Individuals differ in their susceptibility to VIMS [107]. Age has been shown to have a significant relationship with HMD-related eyestrain symptoms [70]. Children 2-12 years of age have immature visual systems and binocular function that is worse than that of adults; this makes children more susceptible to both visual discomfort caused by HMDs and oculomotor side effects including reduced visual acuity, amblyopia, or strabismus [21,31,108,109]. Adults with limited fusional ranges experienced more visual discomfort, specifically with convergent eye movement in response to stimuli in VEs (Karpicka E, unpublished data). Therefore, age effect on HMDs needs to be further studied and taken into account in the design of future HMDs. In regard to gender, females reported more simulator sickness and more often withdrew from HMD-based VEs when compared to male participants [16,19,33,48,49,70]. This difference may be due to under-reporting of susceptibility on self-reports by males (so-called "macho effect") or hormonal effects [110]. Other possible explanation for this gender difference is that females generally have a wider FOV than males, which increases the likelihood of flicker perception and sickness susceptibility [111].
People with visual deficits may have an increased susceptibility to oculomotor side effects compared to those without such deficits, although this has yet to be verified experimentally. A past history of motion sickness has also been found to predict susceptibility to sickness in HMD-based VEs [49]. Individuals also differ in their ability to habituate or adapt to HMD-based VEs (i.e. plasticity), with some individuals adapting much more readily than others after repeated exposures to stimuli [50,81-84,87].
It has been suggested that those with greater plasticity may be less susceptible to VIMS, although the time course to adapt may vary. Greater plasticity is associated with faster symptom reduction on repeated exposures rather than with reduction of initial symptoms [81-84]. Thus far, the characteristics of individuals with greater levels of plasticity have not been identified, and this will require further study.
An individual's posture may also contribute to VIMS. The postural instability theory states that motion sickness occurs with a loss of postural control [112]. In a virtual environment setting, there is a sensory conflict between the virtual image and the real-world posture that increases the body's risk for motion sickness [70,77-80,112]. Postural stability relies on input from the visual, somatosensory, and vestibular systems. This input is processed and then controls two major reflexes, including the Vestibular Ocular Reflex (VOR) that maintains stability of visual objects on the retina as well as the vestibular spinal reflex that maintains body postural stability while an individual is in motion. Conflict between the visual and vestibular sensory inputs can give rise to postural instability (ataxia) as well as to VIMS [79]. Postural instability, which has been reported as a symptom of HMDs exposure, may last for several hours after exposure [70,77-80,85,88]. Special consideration for HMD user safety, as related to the risk of postural instability, must be kept in mind. For instance, HMD users should allow for adaptation and recovery time before engaging in potentially dangerous activities such as driving, or sports may be in order.
Task characteristics have been also identified as potentially affecting VIMS. The most important of these is the duration of exposure to VE. As shown in Figure 9, longer exposure to VE increases the incidence of VIMS. These symptoms may persist up to 60 minutes after exposure [17,21,24-26,32,34,39,48,70,84,95,96]. Another important factor shown to influence VIMS is vection (i.e. an illusion of self-motion; Table 5), with faster vection resulting in greater sickness symptoms [17,77,91,93] Viewing HMD-based VR in a sitting position may reduce symptoms, as sitting reduces the demands on postural control [22,53,78,113]. More complicated tasks, such as reading, may induce total symptom severity scores and oculomotor-related symptom scores that are significantly higher than those observed with movies or games (Figure 10) [39,52,64]. These findings imply that more demanding tasks probably will create some degree of eyestrain. Increased reading sensitivity, when compared to watching a movie or playing a game, might be due to activation of different areas of the brain, which may make reading more complex than other tasks. Alternatively, reading can affect attention and blink rate, which may also contribute to an increase in VIMS. Moreover, inappropriate vertical gaze angle may cause increased oculomotor changes and visual discomfort [30, 97].
Our meta-analysis and systematic review confirms that visual discomfort occurs after exposure to current HMDs significantly more than after exposure to traditional displays. The visual discomfort induced by HMDs is influenced by the three categories of moderator factors, which indicates that the discomfort is multi-factorial and poly-symptomatic. It is conceivable that the visual discomfort induced by HMDs will diminish gradually as the quality of design of HMDs improves and the technology of the components increases; however, the discomfort may not resolve completely until the visual-vestibular mismatch is resolved. VIMS and visual discomfort continue to be obstacles for widespread acceptance of HMDs; this increases the importance of further research into VIMS. More research is needed to resolve visual-vestibular mismatch, and to develop objective methods of evaluating and quantifying VIMS symptoms such as visual/ocular changes (e.g. ocular movements), physiological changes (e.g. changes in heart rate, blink rate, EEG [electroencephalography]), and vestibular changes (e.g. perceived spatial velocity). More research focusing on the user experience is necessary, with recommended expansion of subjective assessment methods such as questionnaires (e.g. the SSQ and VSQ).
Furthermore, as the VR market expands from early consumer adoption, a burgeoning environment for VR-related software has developed. While the emphasis of our paper relates to the current hardware limitations of HMDs, the authors recommend future research also focus on the relationship of software implementation to simulation/VR sickness. For example, perceived motion in virtual environments is affected by how head motion or controls are mapped into the graphical representation of the virtual environment. To limit some of the effects of software-related VIMS, developers may limit movements in certain directions or provide a frame of reference. Recent advancements include the usage of eye tracking with foveal rendering to simulate real-world object focus in virtual environments [101].
We have proposed recommended guidelines below, in part for both hardware and software developers, to design accordingly to minimize VIMS. Although there are still hurdles related to creating seamless virtual environments, there is a lot of promise. Continued research and development of both hardware features and software implementation will continue to improve the VR experience.
Our meta-analysis has led us to propose a few key observations and recommendations. Our observations are as follows:
1. Lighter HMDs are associated with a decrease in discomfort;
2. Monocular presentations should be avoided, as they are associated with more discomfort compared to binocular and dichoptic presentations;
3. Exposure to VR in sitting position may decrease VIMS;
4. Complex visual tasks and reading may increase VIMS severity;
5. Rapid vection results in an increase in VIMS symptoms.
Our recommendations are as follows:
1. Manufacturers need to be attentive to system characteristics of the devices they develop and market;
2. Users should be advised that children, women, users with visual field defects, postural instability, or past history of motion sickness may be especially prone to VIMS;
3. Inexperienced users are especially susceptible to developing VIMS, and users are different in their adaptation to HMDs;
4. Users should be warned to not use HMDs for a long period of time, and to take frequent breaks;
5. Users should avoid driving or operating heavy machinery after exposure to VR until VIMS and postural instability resolve.
This study was supported in part by an unrestricted grant from Research to Prevent Blindness, Inc., New York, New York, USA, to the Department of Ophthalmology and Visual Sciences, University of Utah, Salt Lake City, Utah, USA.
We are grateful for the consultation support we received from: Michael Zyda, DSc, GamePipe Laboratory, Department of Computer Science, Viterbi School of Engineering, University of Southern California; Los Angeles, CA, USA; Kushagra Shrivastava, Marketing Technologist; Advisory Board, Vizzario; and Jacob Mederos, Lead Software Engineer, Vizzario, Davis, CA, USA. Susan Schulman, University of Utah School of Medicine, provided editorial and manuscript preparation assistance.
Dr. Khaderi is an Advisory Board Member for Magic Leap Inc., an advisor for Medella Health, a consultant and speaker for Merz Pharma, and the founder and CEO of Vizzario. The other authors have declared that they have no financial disclosures.