1 Introduction

Virtual Reality (VR) headsets are garnering increasing popularity among consumers and continuously attracting new users. The use of VR is particularly appealing as a means of entertainment when users are forced to stay in one place and need to pass the time, such as during commutes or long journeys. Extended periods of travel have been found to have a negative impact on the relaxation and overall well-being of passengers. (Stutzer and Frey 2008). Especially in these situations, VR could offer advantages to its users. Due to its high level of immersion in comparison to regular entertainment systems, virtual reality technology has the potential to be an attractive addition to public transportation as it allows passengers to escape the confined spaces of a car or airplane cabin and enter a virtual environment (McGill et al. 2020, 2022a; Williamson et al. 2019). The convenience of having a personal screen shielded from prying eyes provides additional privacy (McGill et al. 2020; Medeiros et al. 2022; Williamson et al. 2019).

One transportation mode commonly associated with extended travel times, in which passengers could potentially benefit from the use of VR technology, is air travel. In-flight entertainment serves as a crucial component of a flight’s overall experience, being of high importance to both passengers and airline expenditure (Hugon-Duprat and O’Connell 2015). Passengers are willing to pay extra for additional or improved entertainment (Balcombe et al. 2009) and 65% of respondents in a survey showed moderate to strong interest in utilizing VR technology during their flights (Bajorunaite et al. 2021). In addition to providing enhanced entertainment due to high immersion and improved privacy, VR offers more advantages. Lewis (2014) reports that VR applications on airplanes can provide a distraction from physical and auditory stressors, like limited legroom or a crying baby. As a result, VR could potentially provide relief from the stressors typically encountered during economy airline flights, improving the overall flight experience and reducing stress levels.

However, the unique aspects of VR, when compared to conventional in-flight entertainment, also have drawbacks. Greater privacy can lead to decreased awareness of surroundings, which poses challenges for responsiveness and preparedness during emergencies (McGill and Brewster 2019). Additionally, some individuals may have low acceptance of using VR technology on airplanes. A recent survey indicates that passengers express concerns about the decrease in environmental awareness. The survey also found that respondents are concerned about unwanted interaction with others and being judged by the passengers sitting around them (Bajorunaite et al. 2021). In this line of thought, incorporating actual passengers into a virtual environment through avatars results in decreased security concerns and higher social acceptance (Bajorunaite et al. 2022). Therefore, the lack of situational awareness is one of the main problems for the public use of VR application (Williamson et al. 2019).

Another major cause of concern when introducing VR to a large audience is the occurrence of cybersickness, i.e. sickness as a consequence of immersion into virtual environments or more specifically in VR. This has also been cited as a reservation against using VR in public transportation (Bajorunaite et al. 2021; Mukhaimar et al. 2023; Qiu et al. 2023). In this regard, motion sickness, i.e. sickness as the result of an adverse physical motion stimulus, has historically been a significant issue in aviation. In the 1950s, during the rise of modern-day civil aviation, approximately 75% of individuals reported symptoms of motion sickness during flights (Lederer and Kidera 1954). This percentage has significantly decreased as the general public has become more accustomed to flying and technology has advanced (Turner et al. 2000). Nevertheless, the three-dimensional movements of an airplane and the potential conflicting interactions with visual motion information displayed in virtual reality present significant challenges. Thus, the use of VR in this setting requires special consideration. In this study, we aim to address this major challenge for the widespread use of VR in transportation and investigate the interaction between the presented VR content and the perceived flight motion.

1.1 Cybersickness with VR in moving vehicles

The sensory conflict theory is currently the most widely accepted theory for the development of motion sickness (Reason and Brand 1975). This theory posits that motion sickness arises due to a conflict between two senses in the perception of motion (usually the visual and vestibular senses). Unlike typical motion sickness which is triggered by a provocative motion stimulus, the presentation of moving visual cues alone can also induce similar symptoms even in the absence of actual motion. This is known as visually induced motion sickness (VIMS; Golding et al. 2021) and is consistent with sensory conflict theory because in these cases the visually presented motion is not congruent and thus in conflict with vestibularly perceived motion or absence of motion. In VR, a conflict occurs when a movement is displayed visually through the VR headset, but is not physically experienced and lacks any support from vestibular sensation (McCauley and Sharkey 1992). Cybersickness resulting from VR share some similarities with simulator sickness, a distinct form of sickness syndrome arising from motion simulators such as flight simulators (Kennedy et al. 1993). For example, simulators typically feature a virtual visual environment. However, unlike traditional VR applications, they also incorporate the simulation of real, physical movements. Importantly, these movements should be synchronized with the corresponding visual representation of the environment. Although cybersickness in virtual environments and simulator sickness share similarities, they exhibit distinct symptom patterns, with cybersickness emphasizing disorientation and generally causing stronger symptoms (Stanney et al. 1997). In addition, VIMS often shows more pronounced oculomotor symptoms such as headache compared to traditional motion sickness (Bijveld et al. 2008).

When using VR in real vehicles, there is an additional element of physical movement. Unlike in simulators, where visual representation should be closely synchronized with physical movement, VR applications in vehicles are potentially entirely independent. Using VR in moving vehicles can thus potentially lead to motion sickness due to both the impact of the visual representation of motion (VIMS) and the actual motion of the vehicle itself. This means that sensory conflicts might intensify if instead of incongruent movement (visual movement while being physically stationary), there are even contradictory physical and visually presented movements. Several studies have explored the implementation of VR technology in cars, both for co-drivers and as a means of passenger relaxation in self-driving cars (e.g. Cho and Kim 2020; Hock et al. 2017; McGill et al. 2017). As expected, many passengers develop significant cybersickness, when using VR in a car (Cho and Kim 2020). In general, the type of content has a large impact on the provocativeness of a VR application. Contents with high optical flow induce cybersickness more strongly than static contents (Chang et al. 2020). Furthermore, elevated levels of sickness were found in a moving car solely when dynamic VR movies, featuring locomotion, were shown. Static movie clips failed to induce significant cybersickness (Wienrich et al. 2017).

The significant role of VR content was supported through research indicating that the detrimental effects of car movements can be negated if the physical movements of the car are synced with the visual movements of the VR applications (Cho and Kim 2020). Moreover, VR applications synchronized with the vehicle appear to cause even lower levels of motion sickness than when using VR while inside a parked car, which again would be consistent with sensory conflict theory (Hock et al. 2017). Similarly, efforts have been made to alleviate motion sickness while riding in a car by providing motion flow information (Cho and Kim 2022). Both the possibility to see the physical motion through a transparent wall in the background of the virtual scenery and the presentation of radially flowing particles that reflected the physical motion significantly reduced the extent of cybersickness compared to the usage without any mitigation strategies. However, none of the aforementioned strategies singularly prevent the occurrence of cybersickness. This may be attributed to several technical constraints such as latency still occurring between motion of the car and rendering in VR, drift of the headset’s gyroscope, or disruptions of the optical flow (Cho and Kim 2020, 2022).

In general, there are differences between the feasibility of using VR technology in airplanes and cars. Cars experience more angular movements due to frequent turns and accelerations as a result of increased stops (e.g., at traffic lights). On the other hand, movements in civilian aircrafts are generally more sluggish. Turning flight is composed of slower acceleration and rolling at a constant speed, which is not comparable to the tight turns of a car ride through a downtown area. Williamson et al. (2019) therefore argue that employing VR technology in aircrafts as opposed to cars is viable even with current hardware.

General use of VR during regular cruise flight may prove unproblematic, but it could pose a risk to the health and well-being of passengers during extreme conditions. While aircraft movement is generally slower, turbulence may still occur and is characterized by a high-frequency acceleration motion profile. Additionally, due to the three dimensions of flying, passengers experience vertical and tilting motions that are not present in a car. Therefore, it is crucial to explore the interaction between various common aircraft movements and the tolerability of VR content across distinct typical airplane motion profiles. This investigation shall determine the feasibility of VR entertainment applications in the aircraft cabin. Prior research suggests that cybersickness among VR users was only minimally affected by simulated turbulence (Soyka et al. 2015). We intend to expand this study and systematically examine the interaction with different VR content and typical motion profiles in an ecologically valid environment.

1.2 Individual differences in susceptibility to cybersickness

Numerous factors have been examined to elucidate interindividual variability in susceptibility to motion sickness, including demographic attributes (e.g., gender or age), physiological parameters (e.g., semicircular canal function or migraine history), and psychological factors (e.g., anxiety or field dependency) (Mittelstaedt 2020). To improve user experience, VR providers should be aware of risk factors for cybersickness development and communicate these with first-time users. Since cybersickness may cause severe symptoms in rare cases, these measures can prevent emergencies during airplane travel.

Cybersickness resulting from VR and other virtual applications is often conflated with motion sickness. However, it is unclear to what degree susceptibility to motion sickness when using different modes of transportation can be applied to virtual environments. In general, the motion sickness history, collected by the Motion Sickness Susceptibility Questionnaire (Golding 2006), is considered a good predictor for individual susceptibility to cybersickness as well (Jasper et al. 2023). However, it solely considers the frequency of motion sickness occurring from physical movement, such as cars, airplanes, and amusement park rides and not from virtual environments and it shows inconsistent findings in predicting sickness as a result of VR exposure (Pöhlmann et al. 2023). This may be important as there are generally some differences between susceptibility to cybersickness and motion sickness. Traditional motion sickness is often negatively correlated with age, meaning that older individuals are less susceptible (Golding et al. 2017). In contrast, younger individuals tend to experience less cybersickness than older individuals (Arns and Cerney 2005). This could be a consequence of greater familiarity and habituation to a particular motion stimulus. Older individuals usually have more experience with a diverse range of vehicles, whereas younger generations often consume virtual content, particularly VR, making them more accustomed to the virtual motion information of such environments. While experience with VR has not proven to be a good predictor, experience with video games has been found to be a suitable alternative for developing familiarity with virtual environments and predicting cybersickness in VR (Mittelstaedt et al. 2018). One potential explanation for this is that VR technology has yet to gain widespread adoption among consumers, resulting in a lack of familiarity among users. Additionally, individuals who play more video games may also have a generally more positive basic attitude toward technification and an affinity for new technologies. It is conceivable that attitudes toward technologies are themselves predictors of tolerability and coping with negative consequences of these technologies (i.e. cybersickness).

Another aspect of cybersickness susceptibility that is frequently discussed is gender. While gender differences are not always present, studies consistently show that women are more susceptible to cybersickness than men (Kelly et al. 2023; Munafo et al. 2017).

Besides these demographic characteristics and varying degrees of experience and habituation, the individual personality of the user must be taken into account. In this context, anxiety has been shown to be a reliable predictor for the degree of motion sickness (Fox and Arnon 1988) and cybersickness experienced by passengers (Stelling et al. 2021). Anxious individuals might tend to perceive unpleasant body perceptions more strongly and amplify symptoms through catastrophizing (Mittelstaedt et al. 2019). Anxiety may also contribute to the commonly observed gender differences (Pot-Kolder et al. 2018).

1.3 Present study

In the present study, we aim to investigate the applicability of VR in the aircraft cabin in an ecologically valid environment. Our investigation focuses on the effects of flight phases with varying motion profiles (cruise, turning, and turbulence) and VR content (static and dynamic) previously identified in the literature. Specifically, we explore the general provocativeness of VR to cause cybersickness in the aircraft cabin, as well as how flight phase and VR content impact cybersickness and whether there is an interaction between these factors.

Since there are often strong interindividual differences in cybersickness among individuals, we will also investigate which individual factors are most predictive of cybersickness in the aircraft cabin.

2 Method

2.1 Design

Each participant completed two 27-minute simulator flights. The first flight served as a familiarization flight without the use VR. In the second flight, participants were immersed in a virtual environment by wearing VR headsets. The study had one between-subject factor: VR CONTENT. Participants watched VR clips during the second flight – these clips were either static with fixed POV or dynamic with linearly moving POV. Additionally, there was one within-subject factor: FLIGHT SEGMENT. Within each flight, the simulator performed three different motion profiles in randomized sequence (Cruise, Turning, Turbulence). This allowed an investigation of the interaction between motion profiles and VR content. Moreover, a Takeoff segment was carried out before all other segments, serving as a baseline measurement. During this segment, no VR headset was worn in either flight. In-flight cybersickness ratings were conducted after each segment. Detailed symptom assessment was performed before and following the second flight.

Prior to and in between flights, participants completed questionnaires to identify individual predictors of susceptibility to cybersickness in this environment.

2.2 Participants

A total of 130 participants enrolled in the study. The analyses are based on 129 candidates (64 female and 65 male) with an average age of 28.00 (SD = 8.33; range 19–60) as one participant dropped out after the familiarization flight. Participants were recruited either through a recruitment agency or university postings and were compensated with 100€ for their participation.

All participants were tested for COVID-19 prior to the study and were mandated to wear either a surgical or FFP-2 mask during the procedure.

2.3 Procedure

2.3.1 General procedure

Upon arrival, the participants received information regarding the experiment procedure and attended a mandatory safety briefing for the simulator. Following the briefing, the participants were instructed on the use of tablets to complete their questionnaires. Afterwards, they performed a familiarization flight to acclimate to the simulator’s movements and interior without the use of VR headsets. Instead, passengers were provided with an artificial external view that matched the movement of the flight, displayed on screens mounted on the walls of the cabin, similar to cabin windows. The duration of the flight was 27 min. After the familiarization flight, participants completed questionnaires on their personality and attitude towards technology, which took around 30 min. Then, they engaged in a VR flight that replicated the movements and flight profile of the familiarization flight. Participants were separated into two groups and wore VR headsets while watching either the static or dynamic 360° clips throughout the simulated flight. The entire procedure lasted between two and a half and three hours.

For ethical reasons, participants were informed in advance in the advertisement and again on-site about the possibility of experiencing motion sickness as a result of the experiment. The procedure was considered ethically acceptable by the Ethics Committee of the German Psychological Society (DGPs; 2021-09-20VA).

2.3.2 Flight protocol

Both flights (familiarization and VR) started stationary on the runway. Each flight started with the Takeoff and a climb phase that lasted for a total of 3 min. After Takeoff, the three four-minute experimental segments followed, which consisted of the three segments Cruise, Turbulence and Turning in randomized sequence. In the VR flight, participants put on the VR headsets only during these segments and watched 360° clips (static or dynamic). A flight ended after the third segment without a landing. After Takeoff, between the segments and at the end participants were asked to remove the VR headsets and answer an in-flight questionnaire on sickness symptoms within 4 min. During these phases, the motion profile of the simulator was the same as that of the Cruise segment. Prior to the start and after completion of all segments, participants were asked to fill in the Simulator Sickness Questionnaire (SSQ; Kennedy et al. 1993). The experimental protocol of the VR flight can be seen in Fig. 1.

Fig. 1
figure 1

Experimental protocol of the VR flight

Full size image

2.4 Apparatus

2.4.1 Flight simulator

The Air Vehicle Simulator (AVES; Fig. 2) was used as a motion platform for a passenger cabin (length: 4.2 m; width: 3.3 m; height: 2.4 m) module with 16 seats in four rows. The first row was not used because it has less legroom compared to the other rows. Thus, each session accommodated up to 12 participants if there were no no-shows.

Fig. 2
figure 2

The outside view of the Air Vehicle Simulator (AVES)

Full size image

Four screens (27”LCD/NEC MultiSync® EA275UHD) on each side of the inner cabin wall were mounted to present a virtual outside view of the surroundings. The visual environment matched the movements of the AVES and was generated using Microsoft Flight Simulator 2020Footnote 1.

In the cabin, air conditioning similar to that in a real airplane is implemented and held a constant temperature of 21 °C, as well as a sound system through which engine and other sounds were presented to give passengers the feeling of a real flight. An interior view of the cabin is presented in Fig. 3.

Fig. 3
figure 3

Interior view of aircraft cabin simulator with participants wearing VR headsets

Full size image

The used motion profiles consisted of four different types of segments: Takeoff (including climb), Cruise, Turbulence and Turning. Cruise and Turbulence segments consisted of straight and level flight with varying turbulence strengths. Turbulence was generated using the von Kármán turbulence model (Gage 2003) and then scaled using a gain of 0.1 for the Cruise and 1.25 for the Turbulence segment. The gains were subjectively tuned and yielded accelerations of a root mean square level of 0.014 m/s2 during Cruise and 0.177 m/s2 during Turbulence. The Turning segment consisted of alternating left and right turns using a maximum bank angle of 25°. The gain used for turbulence scaling during the Turning segment was the same as for the Cruise segment (0.1). Turn direction was changed every 30 s.

2.4.2 Virtual reality

Oculus Quest with two PenTile OLED 1440 × 1600 72 Hz displays (one for each eye) were used for the presentation of the VR. Compared to other VR headsets, this model sports the notable advantage of a wireless operation, allowing for easy use with passengers in an airplane cabin.

Our VR flights included three 360° VR clips showing scenes from the port of Hamburg, each supplemented with auditory information about the surroundings. Figure 4 showcases a typical view from one of the scenes. Each clip was split into three short scenes recorded using an INSTA360 ONE X 360° camera at various locations.

Fig. 4
figure 4

Typical view of the VR clips, in this case a clip from the static condition

Full size image

Participants were divided into two separate groups. One group (static condition) received only footage shot from a static POV with no camera movement (69 participants). From this perspective, participants could view the entire environment in 360° from a fixed vantage point. In the other condition (dynamic condition), the camera perspective moves forward in a straight line, akin to a leisurely walk (60 participants) at about 4 km/h resulting in a constant optical flow. All clips permit the user to move in three degrees of freedom, that is to look up and down, left and right and to tilt. Identical clips were shown to all participants in one flight, without knowledge of other clips or conditions.

Content on the VR headsets was managed by one of the experimenters with Showtime VR®. The clips were started and finished simultaneously for all participants of the same flight.

2.5 Questionnaire

2.5.1 Personality questionnaire

Three questionnaires were used to assess the personality of the participants. The NEO-FFI-3 (McCrae and Costa 2007) uses 60 items to assess personality structure across the Big Five dimensions: Neuroticism, Extraversion, Openness to Experience, Agreeableness, and Conscientiousness. Each item presented a statement about the participant’s behavior or personality, to be rated on a five-point Likert scale ranging from “strongly disagree” to “strongly agree”.

The Pain Catastrophizing Scale (PCS; Sullivan et al. 1995) assesses coping with pain and in particular the tendency to catastrophize pain and its effects. Thirteen statements on the coping with and state during pain have to be answered on a five-point Likert scale. While the PCS focuses on pain, it appears to measure a general tendency to catastrophize unpleasant states. Individuals who tend to catastrophize pain symptoms are also more likely to report experiencing cybersickness (Mittelstaedt et al. 2019).

The State-Trait Anxiety Inventory (STAI; Spielberger et al. 1983) uses twenty items to assess both state and trait anxiety in individuals. In this study, only trait anxiety (situation-independent) was analyzed. Participants responded to twenty statements on their personal feelings on a four-point Likert scale from “not at all” to “very much”.

2.5.2 Affinity for technology

The affinity for technology was assessed with a questionnaire for technology affinity with a focus on electronic devices (TA-EG; Karrer et al. 2009). With this questionnaire, participants indicated their attitudes towards new technological devices and their impact on everyday life. Items like “I like to own new electronic devices” or “Electronic devices cause distress” are answered on a four-point Likert scale. The responses were aggregated to four different dimensions: Enthusiasm, Positive Attitude, Negative Attitude (all 5 items) and Competency (4 items).

In addition, we asked several questions about previous experience with VR, the use of video games and the use of mobile games. Participants also had to indicate their frequency of flying.

2.5.3 Sickness

The Motion Sickness Susceptibility Questionnaire Short (MSSQ; Golding 2006) was used to determine the general tendency to develop motion sickness. It assesses the frequency of occurrence of sickness symptoms across nine different means of transportation and amusement rides, separately for child- and adulthood. Scores for motion sickness susceptibility were derived using the formula given by Golding (2006). The mean MSSQ raw score (sum of MSSQ child and MSSQ adult) of 29.46 in the present sample corresponds to a percentile score of 59.69 after percentile conversion and is indicating a slightly above average susceptible group.

To evaluate the severity of sickness in the study, the Simulator Sickness Questionnaire (SSQ; Kennedy et al. 1993) was used in addition to the in-flight questionnaire (discussed below). It was administered before and after a flight. In this questionnaire, people rate the intensity of 16 symptoms they are currently experiencing. These comprise of symptoms from the three symptom clusters disorientation, oculomotor and nausea. They were summed into a Total Score of sickness (SSQ-TS) according to the formula provided by Kennedy et al. (1993).

Throughout the flight, in between segments, participants were prompted to complete the short in-flight questionnaire. The questionnaire was automatically displayed on a tablet computer mounted at the headrest in front of them after takeoff and each segment. It included three items asking for nausea (“Do you feel nausea?”), disorientation (“Do you feel vertigo or dizziness?”) and oculomotor symptoms / headache (“Do you feel eyestrain or headache?”). The original items in German language can be seen in the supplementary material. All responses were entered using a slider on the tablet computer. The position of the slider was then translated to a rating from 0 to 100.

2.6 Analysis

For the analysis of the influence of flight phase, VR content, and their interaction on symptom severity in the SSQ, we computed a linear mixed model for repeated measures. The model included the factors measurement time (pre and post) and VR content (dynamic and static) as fixed effects and the participant as random effects.

To further investigate the temporal progression and the influence of different flight phases on sickness, linear mixed models for repeated measurements were performed with the in-flight ratings as dependent variables. Here, the factors VR content (dynamic and static) and flight phase (Takeoff without VR, Cruise, Turning and Turbulence) were fixed effects and the participants a random effect. A total of three models were computed, one for each item from the in-flight questionnaire on sickness (nausea, disorientation and oculomotor).

Correlations between predictors and SSQ after the VR flight were computed to investigate the relationship of different predictor variables with the sickness level. Since some of the predictors were expected to be correlated, a partial least squares (PLS) regression was then performed. As the participants did not view the same VR content, we included the condition as control variable to the model. To conduct a relative weighting of the predictors, variable importance in projection (VIP) was computed which indicates the relative importance of each predictor.

3 Results

3.1 VR content and SSQ

The average scores on the SSQ before the VR flight were M = 7.37 (SD = 10.40) in the static condition and M = 7.29 (SD = 8.78) in the dynamic condition. Whereas, the SSQ scores after the VR flight were at M = 13.39 (SD = 15.16) in the static and at M = 14.27 (SD = 13.56) in the dynamic condition.

Table 1 presents the linear mixed model results, indicating no significant influence for the VR content and the interaction with measurement time (pre/post). However, a significant result was found for the main effect of measurement time. As illustrated in Fig. 5, SSQ values increased significantly after the VR flight compared to before, with similar elevations observed in both VR content conditions.

Table 1 Results of the linear mixed model for the SSQ scores depending on VR content and time of measurement
Full size table
Fig. 5
figure 5

Results of the linear mixed model on SSQ scores. Error bars indicate standard errors

Full size image

3.2 Flight phases

We computed linear mixed models for the three items on sickness (nausea, dizziness, and oculomotor). These were collected a total of four times as part of the in-flight questionnaire. The results of all three linear mixed models are summarized in Table 2.

Table 2 Results of the linear mixed models for the three sickness items in the in-flight questionnaire
Full size table

For nausea, the model yielded a significant effect of flight phase. VR content and the interaction of VR content and flight phase remained insignificant. Although the model suggested an effect for both Turbulence and Turning flight phases, Tukey-adjusted post hoc tests revealed only a significant difference for Turbulence and Takeoff in the sense that nausea was significantly higher after Turbulence (Table 3). Figure 6a) shows that although the interaction effect did not become significant, increased sickness levels were reported within the dynamic VR content condition for both Turbulence and Turning flight phases.

Table 3 Post hoc pairwise comparisons of flight phases with Tukey’s HSD p-value correction in all three symptom clusters
Full size table
Fig. 6
figure 6

Results of the in-flight questionnaire by flight phase and VR content. Note that the sequence of flight phases Cruise, Turbulence and Turning was randomized. Possible scores ranged from 0 to 100

Full size image

The linear mixed model for disorientation also indicated a significant effect of flight phase, while VR content and all interactions were not significant. Tukey-adjusted post hoc comparisons showed that the only significant difference was that disorientation levels were significantly higher after Turbulence in comparison to the Takeoff (Table 3). Similar to nausea, Fig. 6b) shows that increased levels of disorientation were observed in the dynamic condition, specifically after Turbulence, despite a non-significant interaction effect.

For oculomotor symptoms, the linear mixed model also yielded a significant effect of flight phase with no significant effect for VR content or the interaction of both. Tukey-adjusted post hoc tests revealed only a significant difference for the Turning phase in contrast to the Takeoff while all other comparisons had an adjusted p-value higher than 0.05 (Table 3). As displayed in Fig. 6c), oculomotor symptoms (including headache) were rated highest compared to the other sickness symptoms. Interestingly, the ratings for oculomotor symptoms were descriptively higher in the static condition and the highest in this condition during Turbulence.

It is important to note that not a single mean of a combination of VR content and flight phase scored above 10 out of 100 in any of the three symptoms of nausea, disorientation, or oculomotor symptoms.

3.3 Predictors of cybersickness

First, the correlations of each predictor with the level of SSQ scores were computed. The results can be found in Table 4. Twelve of the nineteen variables were significantly correlated with the SSQ scores. The highest correlation was with trait anxiety from the STAI (r = .33).

Table 4 Pearson correlations (point-biserial for gender) with SSQ scores (controlled for condition) after VR flight and VIP of PLS regression
Full size table

To determine which individuals were most likely to develop sickness symptoms during VR in the aircraft cabin, hence which characteristics were most highly related to the sickness levels, we performed a partial least square regression analysis (PLS) with all possible predictors. The advantage of PLS over standard linear regression analysis is that it is less susceptible to correlated predictors (multicollinearity), which is to be expected with the present selection of factors.

The k-fold cross validation with 10 random segments of the PLS model resulted in a test root mean square error (test RMSE) of 14.06 for all 20 components (including the VR content condition as control variable) and explain a total variance of R2unadjusted = 0.343.

The variable importance in projection (VIP) values are shown in Table 4. Similar to the correlation results, the greatest importance was attributed to variables related to anxiety (trait anxiety, neuroticism, pain catastrophizing).

Great importance also resulted for the frequency of video game use, as well as enthusiasm and competence with new technologies. In this context, a greater influence of negative attitudes (concerns) towards new technologies on the levels of reported sickness was found than for positive attitudes.

Finally, a younger age, female gender and a high motion sickness history in adulthood were identified as important predictors to the susceptibility to cybersickness as a result of VR in the aircraft cabin.

4 Discussion

In the present study, we investigated the impact of using VR on cybersickness in the aircraft cabin. Special focus was on the effects of different VR content (static vs. dynamic 360° videos), varying flight profiles and the interaction between them. A major advantage of the study is the high ecological validity of the environment, as we used a realistic replica of an aircraft cabin (Stelling et al. 2021), a full motion simulator and participants were tested in groups of passengers, i.e. a realistic scenario of a real flight was created. This included a cabin with sixteen seats arranged in a realistic layout, realistic legroom, real cabin seats, exterior walls, overhead bins, and realistic air conditioning. Furthermore, a flight with realistic movements was conducted utilizing the simulator, primarily designed for use as a cockpit simulator, along with a simulated external view.

In our analyses, no significant difference was found between the effects of the two VR contents, neither in the post-flight SSQ, nor in the in-flight ratings. This was despite previous research suggesting a higher provocativeness of content in which the user’s point of view is moving (Wienrich et al. 2017). Because this was the first experiment in this setting and the effects of VR on human well-being in airplanes were not extensively researched with sizeable groups yet, we opted for content with relatively mild visual motion (only slow, straight movements and no angular movements). This could be one reason for why the dynamic condition did not result in significantly more sickness than the static condition due to the speed of navigation, hence the optical flow being directly linked to the degree of cybersickness (So et al. 2001). However, descriptive analysis suggest that dynamic content has the potential to cause problems, particularly with regard to symptoms of nausea and disorientation which primarily accounted for the increased scores after Turbulence (see below). When using more provocative content such as faster navigation speed, the effects and interactions with vigorous aircraft movements may be intensified.

For the flight phase, we compared Takeoff (without VR) to Cruise, Turbulence, and Turning. Significant differences between those flight phases were observed for each of the three symptom clusters (nausea, disorientation, and oculomotor). For nausea and disorientation, higher values were recorded after Turbulence while for oculomotor symptoms, higher values were observed after Turning flight. It is in line with the expectations that turbulence and turning flight could be an additional problem for triggering sickness in an airplane (Wienrich et al. 2017). In our flight scenarios - as in real life - these flight profiles were characterized by higher frequency and amplitudes of physical movements. However, the fact that the Turbulence was not significantly different from the Cruise phase suggests that there may have been an additive effect with VR on one side and unpleasant simulator motion on the other. It is noteworthy that oculomotor symptoms have increased during the static condition, especially during cruise flight and turbulence. This may be attributed to the static nature of the environment, with participants experiencing more ocular problems because they had to constantly look at the same scenery. Furthermore, the monotony of the stationary surroundings could have caused fatigue, which is listed as one of the oculomotor symptoms in the SSQ.

Overall, only low to moderate expressions of cybersickness occurred throughout the study. The highest score was found in the dynamic condition at 14.27. The SSQ authors suggest a threshold of 20 to classify a simulator as “bad” (Kennedy et al. 1993). However, since they used a military sample with likely lower susceptibility, a score of 20 is commonly found with various virtual applications. Therefore, the scores observed in this study seem rather low (Bimberg et al. 2020). Additionally, none of the participants had to terminate the VR experience, including removing the VR headset, vomiting or exhibiting noticeable severe sickness symptoms. Thus, our results indicate that short-term use of VR with content similar to ours is not problematic on airplanes in terms of cybersickness. This aligns with previous research by Soyka et al. (2015), which found no effect of (turbulent) simulator motion on the tolerability of VR either. However, the present study cannot fully support their assertion that “on average no one felt worse after the experiment than before” (Soyka et al. 2015; p. 39). It should be noted that both studies used similarly short exposure durations. In real-world use, VR headsets are likely to be used for extended periods. Since it is well-established that sickness levels tend to increase with exposure duration (Murata 2004), we would expect higher levels of sickness with longer exposure in an airplane as well. It would be intriguing to explore whether such duration-dependent increase of sickness is further influenced by adverse motions like turbulence. It should also be mentioned that even on the ground, some people get severe cybersickness from VR without any additional physical motion involved. This makes it all the more important to understand which users may be more susceptible to cybersickness (see below). Another source of physical discomfort might have been the mandatory face mask, participants had to wear throughout the experimental procedure due to the health and safety guidelines at the time. Although anecdotal evidence suggests that none of the participants noted a negative impact of mask wearing, the wearing of a medical mask may have influenced the results. Nevertheless, since all participants wore the masks throughout the procedure, the actual influence on the validity of relationships and the conclusions might be negligible.

When using VR in a moving vehicle, three possible sources of sickness need to be considered. First, the VR application itself, which could induce cybersickness just like on a stationary ground. Second, the motion of the moving vehicle can lead to motion sickness even without VR use. Finally, when both are combined, an additional source of sickness may occur. As a result, the symptoms experienced are neither pure cybersickness nor pure motion sickness, but rather a combination of both. When both visual and vestibular information work together to create a coherent sense of motion, any disruptions to these expected inputs can cause a sensory conflict, resulting in symptoms of sickness (Bos et al. 2008; Reason and Brand 1975). In cases where both visual and vestibular information are incongruent, the sensation of sickness may be individual and depend on whether the expected input is violated in the visual system or the vestibular system. Some participants may have experienced symptoms due to the movements of the simulator while others may have experienced cybersickness induced by the visual motion of the VR content. However, rather than examining the two systems independently, the interaction between them may be crucial, as one movement stimulus could be triggering solely because of the presence of the other. For example, the inclusion of incongruent motion within the simulator could result in sensory conflict, even for individuals who are well accustomed to virtual environments and have “learned” that virtual motion does not involve vestibular cues. In this sense, the greatest difference between visual and physical motion poses the highest risk for developing sickness symptoms. Descriptively, the combination of dynamic VR content and turbulent motion resulted in the most severe nausea and disorientation symptoms. However, it is possible that the lack of significant difference between VR contents may be attributed to the chosen content not being provocative enough, as discussed previously.

One method for mitigating the sensory conflict experienced in moving vehicles involves integrating physical motion into the visual environment. Recently published frameworks provide the means to create virtual applications that integrate vehicle movements within extended reality (XR) applications, indicating the potential for large-scale implementation in the near future (McGill et al. 2022b). Autonomous cars offer a wealth of movement data via data streams. However, acquiring comparable information for non-autonomous vehicles may prove more challenging. To ensure meaningful activities in these cases, virtual applications would require great flexibility. Applications that rely on predetermined movements or require user attention to particular locations might be difficult to realize with these settings. As an example, racing games in which the track veers left cannot simply make a right turn. Additionally, cinematic 360° movies must remain entertaining and comprehensible even with sudden shifts in focus. In general, the degrees of freedom when designing VR content are severely restricted due to being bound to the physical movements of an unpredictably moving vehicle. In addition, an infrastructure must be provided to transfer physical movements to the VR headset while integrating them into the application with minimal latency. This is particularly important when accommodating multiple users aboard an airplane, as VR content may be utilized that is not synchronized with the movements of the aircraft in the near future.

4.1 Personal characteristics

In this study, we gathered a range of personal traits to ascertain the participants most susceptible to VR-induced sickness symptoms while onboard an aircraft. The findings support previous findings that indicate a higher susceptibility to such symptoms among individuals with greater levels of anxiety (Solimini et al. 2012). Furthermore, these results are consistent with the level of correlation expected from the synopsis of a large number of studies (Mittelstaedt 2020). It has been suggested that the relationship originates from anxious individuals being more attentive and alert to physical conditions, thus enabling the emergence of sickness (Fox and Arnon 1988). An earlier finding, which is supported here, proposes that at least some of the mechanism is rooted in the catastrophizing of unpleasant bodily sensations (Mittelstaedt et al. 2019). In the present study, anxiety – represented by trait anxiety, neuroticism, and catastrophizing – was identified as the primary predictor of sickness symptoms. As the anxiety levels of female participants was higher than male participants like in many previous studies, it is also possible that differences in anxiety contributed to the gender effect observed in this study (Pot-Kolder et al. 2018). The fact that some participants had little or no experience with VR may have exacerbated their discomfort and cybersickness due to anxiety induced by unfamiliarity with the technology.

The impact of general technology use on sickness is an important yet often overlooked factor. Our study identified frequency of video game use as the second most important predictor of susceptibility to sickness, highlighting its potential as a predictor (Mittelstaedt et al. 2018; Rosa et al. 2016). One possible explanation for this is that video gaming leads to greater habituation to virtual environments, which can help prevent the occurrence of cybersickness in VR (Mittelstaedt 2020). It was therefore unexpected that the frequency of VR use did not show a significant correlation with the observed cybersickness. This could be because VR is not yet widely used in the general population. 52% of participants had never used VR before and another 25% had tried it only once. This may provide insufficient exposure to habituate to the effects of VR. By comparison, 56% of participants reported playing video games for at least one hour per week. It is conceivable that the issue of cybersickness will disappear entirely with the increased distribution of VR headsets or even video games in the population.

Interestingly, attitudes towards new technologies in general also have a strong predictive power for the development of cybersickness symptoms. Whereas positive attitudes towards new technologies such as “Electronic devices make my everyday life easier” or “Electronic devices help to get information” showed little predictive ability, the negative attitudes such as “Electronic devices reduce the personal contact between people” or “Electronic devices cause distress” exhibited substantially greater predictive power. Perhaps, the focus on concerns regarding unfamiliar technology is more important than acknowledging the beneficial opportunities of these technologies with regards to the emergence of cybersickness. As described above, the anxiety about negative consequences of technology use may contribute to the observed relationship.

One intriguing additional aspect is that we found a significant correlation with age, which, contrary to previous findings (Arns and Cerney 2005), indicated that older people were less susceptible to cybersickness. It is possible that the mixture of real movements and the presentation of VR causes a slightly different type of sickness, similar to simulator sickness. Due to their greater life experience with flights, older participants may have a higher resistance to sickness that is triggered or influenced by flight movements.

Overall, we were able to account for roughly 34% of the variation in the ratings, prompting further discussion of other potential factors contributing to cybersickness. Along with physiological factors in the processing of visual and vestibular stimuli (Mittelstaedt 2020), which are difficult to assess, situational factors like night sleep (Ng et al. 2022), diet (Rahimzadeh et al. 2023), general health, and individual response tendencies are possible explanations for the differences found.

4.2 Limitations and future research

As stated earlier, the study exclusively employed mild dynamic VR content with slow and straight locomotion. With confirmed good tolerability, we can now move on to testing more extreme content featuring fast movements and rapid rotations. The inclusion of such content would allow content creators more development flexibility and users more freedom of choice. Testing more provocative content with aircraft movement such as turbulence could explore the limits of acceptability for the general population in the aircraft cabin. Similarly, the VR headsets were only worn for a total of twelve minutes with interruptions. Although the interruptions were used for the longitudinal survey of in-flight questionnaire, allowing for a clearer separation of the different flight segments, it is unlikely that interruptions of this frequency would occur in real operations. Furthermore, on long-haul flights, an extended use would be expected. As previously indicated, long-term effects of VR should be examined, particularly in association with prolonged turbulence. It is conceivable that the sickness effects accumulate gradually over time, as has often been reported (Duzmanska et al. 2018; Murata 2004), and only pose problems with prolonged use.

Responses to the questionnaires assessing sickness levels and experiences may have been affected by a tendency towards social desirability. In fact, it has been suggested that the gender difference is a result of men feeling greater pressure to appear stronger and more resilient, leading to socially desirable responding (Klosterhalfen et al. 2006). Although response tendencies can generally never be ruled out in subjective questioning of subjects, the question answering in a group, which was preferred for reasons of ecological validity, might have slightly amplified this effect.

Finally, while we used a realistic simulation of an actual flight, it is necessary to validate the outcome in an actual aircraft.

4.3 Practical implications

This study contributes to clarifying the feasibility of VR usage in aircraft cabins. As VR continues to grow as a technology and push into the consumer market, airlines are considering VR for passenger entertainment on a larger scale (e.g. Onag 2021). The findings suggest that brief exposure to VR with mild content has no major impact on tolerability and cybersickness, even during aversive flight phases that experience high acceleration frequencies, such as turbulence. However, despite the observed good tolerability, the study conditions only partly reflect what may be used in real aircraft in terms of content and exposure durations. Turbulence presents the greatest risk of causing cybersickness in passengers, particularly with dynamic content, likely increasing with faster optical flow of displayed content. In addition to these environmental factors, individual user characteristics also play an important role. Our findings indicate that an individual’s attitudes towards technology may be an important aspect in the development of cybersickness in VR. Specifically, negative attitudes towards technology, rather than positive ones, should be considered. To ensure wider accessibility of VR applications and prevent negative experiences, providers should mitigate negative perceptions and alleviate concerns regarding VR technology.

In addition to the risk of cybersickness, there are further safety considerations to account for when using VR equipment in the aircraft cabin. Users may potentially collide their headset with the front seat or the outside wall during turbulence, putting them at risk of headset breakage or excessive pressure on facial bones. Thus, precautions must be taken to prevent collisions between VR headsets (McGill et al. 2020). Other safety considerations also play a role in the introduction of VR in aviation, like the continued responsiveness of users (McGill and Brewster 2019). Anecdotally, neither issue caused any problems in the present study. Nonetheless, the combined impact of these factors and the slightly elevated levels of cybersickness observed during turbulence may render the use of VR inadvisable during these flight phases.

4.4 Conclusion

This study investigated the use of VR in the aircraft cabin in an ecologically valid environment. Overall, there is little evidence of cybersickness in the setting used in this study. Dynamic VR content did not generate significantly higher sickness values than static content. Nonetheless, passengers could experience cybersickness problems while using VR for an extended period during flight and when encountering turbulence. This study presents a first attempt at simultaneous exploration of different VR content and motion profiles in an aviation context and points to possible interactions of turbulence with dynamic content.