Editorial: Cybersickness in Virtual Reality and Augmented Reality

Early virtual reality (VR) systems introduced abnormal visual-vestibular integration and vergenceaccommodation, causing cybersickness (McCauley and Sharkey, 1992) reminiscent of simulator sickness reported bymilitary pilots, e.g., having some shared causes and overlapping (Lawson, 2014a) but distinguishable symptoms (Stanney et al., 1997). Improved processing, head tracking, and graphics were expected to overcome cybersickness (Rheingold, 1991), yet it persists in today’s muchimproved VR (Stanney et al., 2020a, 2020b). This must be resolved, because VR and Augmented Reality (AR) are proliferating for training for stressful tasks, exposure therapy for post-traumatic stress, remote assistance/control, and operational situation awareness (Hale and Stanney, 2014; Beidel et al., 2019; Stanney et al., 2020b, 2021; NATO Science and Technology Office, 2021). Experts considered the cybersickness problem recently at a 2019 Cybersickness Workshop and a 2020 Visually-Induced Motion Sensations meeting. Military aspects were discussed during 2019–2021 meetings of a Cybersickness Specialist Team (NATO Science and Technology Office, 2021). The Bárány Society’s Classification Committee just developed relevant international symptom standards for visually-induced motion sickness (VIMS; Cha et al., 2021). Finally, >40 authors produced twelve articles comprising this Frontiers Research Topic initiated by Dr. Stanney. Below, we summarize their work and provide recommendations.


INTRODUCTION
introduced into VR or AR (via a partial virtual frame) should improve balance and lessen cybersickness. They discussed two small studies of balance-impaired VR/AR users. Their VR study detected a cueing difference for two balance measures and the Simulator Sickness (SSQ) Disorientation measure 5 , while their AR study (which allowed sight of the room) detected a difference in one balance measure but no SSQ measures. Benefits were seen only with balance-impaired subjects. While the findings were mixed, an appropriately-designed Earth-referenced cue should aid orientation. Expanded studies of this type should compare similar VR-versus-AR fields of view. Finally, 3) Cao et al. provided VR users with Earth-stable granulated peripheral cues that allowed some peripheral vision, which improved visual target searching better than restricting field-of-view (FOV), a typical countermeasure. Could this approach also mitigate cybersickness better than FOV restriction? Two Articles Discussed Aspects Of Tracking Latency As A Cybersickness Contributor 4) Stauffert et al. explored cybersickness implications of latency between the movement of a tracked object and its movement on a head-worn display. They provided information to assist in assessing latency, and stressed the need for comparable assessments. 5) Palmisano et al. posited that a key (and readily quantifiable) contributor to cybersickness is a large, temporally inconsistent difference between actual and virtual head position. Their findings are relevant to , who found that varying head tracking latency was sickening. As many studies have observed that visually-moving fields elicit symptoms even when the head is still (e.g., Webb and Griffin, 2002), however, the contribution of visual field motion versus head position/motion conflict should be studied.

Three Articles Explored Additional Effects
Of Head Motion, Head Orientation, Or Head-Mounting Of Displays 6) Kim et al. posited that linear head oscillations increase sensory conflict in VR devices that only track angular motion. While they failed to detect device-related differences in perceived scene stability, spatial presence, or cybersickness, this was a creative pilot study exploring implications of different tracking devices. 7) Wang et al. confirmed that vection (the illusion of self-motion) elicited by viewing a rotating dot pattern was stronger when concordant with expected graviceptive cues. VR/AR designers should know that when vection is desired, its direction should not contradict somatosensory/vestibular cues that would be present during real motion. Also, specific motion/ orientation perceptions will tend to be altered to minimize sensory conflict (Young et al., 1975;Lackner and Teixeira, 1977;Dizio and Lackner, 1986;Howard et al., 1987;Golding, 1996;Tanahashi et al., 2012). The notion that vection can reduce sickening conflict is better supported than vection as a cause of sickness (Lawson, 2014a;Stanney et al., 2020b). Finally, 8) Hughes et al. evaluated head-worn versus tabletbased AR during tactical combat casualty training. They observed greater sickness with head-worn AR, but symptoms for both devices were mostly limited to the Oculomotor cluster of the SSQ, with little Nausea. Moreover, while subjects in the head-worn condition completed fewer training scenarios in the time allotted, they had more correct responses in completed scenarios. AR could be a less-sickening training approach, and solutions to mitigate oculomotor disturbances would make it even better.
Three Articles Explored The Role Of Active Sensorimotor Engagement Or Maintenance Of Postural Equilibrium 9) Curry et al. evaluated participants in a head-worn racing game. They did not detect main differences in cybersickness between active drivers versus passengers. The reasons for this should be explored, as a difference has been observed in other contexts (Rolnick and Lubow, 1991;Stanney and Hash, 1998;Seay et al., 2002;Sharples et al., 2008). 10) Weech et al. found a correlation between visually-influenced body sway (reflected by the center-of-pressure [COP] ratio) 6 and SSQ Disorientation and Oculomotor sub-scores in a VR. It makes sense for the Disorientation score to be related to sway; expanded studies should determine if COP ratio correlates with SSQ Total Sickness or Nausea scores, as these are likely to predict quitting a training session. Finally, 11) Jasper et al. evaluated the efficacy of different cybersickness recovery strategies. Their study elicited sufficient cybersickness (Stanney et al., 2003). Greatest recovery was observed for resting with the VR off (real natural decay), while doing a virtual hand-eye task yielded the least recovery. We agree with the authors' implication that administration of the SSQ during VR/AR should be explored further.
Three Studies Addressed The Role Of Individual Cybersickness Susceptibility (Two Of Which Were Mentioned Immediately Above) 12) Golding et al. found that sickness severity in a moving visual surround is predicted by history of susceptibility to motion sickness, migraine, and fainting. They did not detect a relationship between sickness and vection, adding to the many studies failing to find this relation (Lawson, 2014a;Stanney et al., 2021). 7 Consistent with the literature (Lawson, 2014a;Stanney et al., 2020a), the aforementioned article #11 by Jasper et al. and #9 by Curry et al. observed mixed findings concerning sex as a factor in cybersickness susceptibility. Jasper et al. observed that women reported more cybersickness, but this was confounded by women having less experience with video games. The sex difference detected in Curry et al. was solely among the subset of subjects who discontinued participation early, wherein women quit earlier when driving, but not when passengers. Future studies of individual cybersickness differences should estimate variance accounted for by experience with motion sickness, driving, video games, and head-worn displays.

CAUSAL HYPOTHESES RELEVANT TO THE 12 TOPIC ARTICLES
While the explanatory capabilities of a complete motion/ simulator/cybersickness theory have been described (Lawson, 2014a), there is no universally accepted theory. Six hypotheses were discussed by Stanney et al. (2021) and ten by Keshavarz et al. (2014). Most of these can be grouped into four established categories (Table 26.1, Keshavarz et al.), which in Table 1 are linked to the 12 articles in this Research Topic. This taxonomy may aid further literature inquiries concerning theoretical implications. 8

CONCLUDING RECOMMENDATIONS TO THE RESEARCH COMMUNITY
We thank the authors for contributing many provocative studies. As is common in research, as many questions were raised as were answered. Answering the key cybersickness questions requires controlled, labor-intensive research entailing: 1. Assessment of relevant stimulus experiences (Jasper et al.) and past susceptibility (Golding et al.): This is vital to interpretation and such measures can be used as covariates to improve analyses.
2. Larger samples (e.g.,  than have commonly been employed (e.g., Kim et al.; Shahnewaz Ferdous et al.), in order to deal with high individual variability in susceptibility (Lawson, 2014a). 3. Stimuli that elicit functionally relevant cybersickness (Stanney et al., 2014 10 ), to avoid basement effects or detection of statistical differences lacking clear functional significance (e.g., Hemmerich et al.). 4. Managing sessions and session intervals to reduce carryover effects which may confound studies with many cybersickness sessions held closely together (e.g., Hemmerich et al.; Kim et al.). Sickening VR or simulator studies should ideally limit the number of sessions to three (Lawson et al., 2009 11 ) and allow 1 week of recovery between sessions, to reduce visual-vestibular and vergence-accommodation carry-over effects due to adaptation (Dai et al., 2011) or sensitization (Dizio and Lackner, 2000), as well as learning, fatigue, classical conditioning, subject attrition, and ultradian variation (Lawson et al., 2009;Lawson, 2014a) (Bos and Lawson, 2021), and an established symptom scale is required for validation (Lawson, 2014b).

AUTHOR CONTRIBUTIONS
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.