Virtual reality-based sensory stimulation preferences at Amundsen-Scott South Pole station in Antarctica

Abstract

Virtual Reality (VR)-based sensory stimulation can provide relaxation and psychological restoration in isolated, confined and extreme conditions such as long-duration spaceflight, but it remains unclear which aspects of VR would be most beneficial. To investigate individual preferences for various VR attributes, 25 overwintering crew members at the Amundsen-Scott South Pole Station in Antarctica underwent 16 variations of VR stimulation at the end of their mission, with manipulations in delivery mode (VR vs. laptop), content (nature vs. city environments), duration (4 vs. 10 min), and sensory augmentation (with or without temperature cues). Data collection included pre- and post-intervention surveys on perceived quality (value, immersiveness, restorativeness) and mood, as well open-ended qualitative feedback. We found that VR was viable and restorative in a high-fidelity spaceflight analog. Although longer-lasting nature scenes were preferred overall, interindividual variation in preferences for sensory stimulation emphasizes the need for a personalized approach.

Introduction

“I didn’t realize how much I missed the sound of wind in trees until this study. I think being reminded of the outside world could help bring perspective into these environments, and help contextualize some of the issues we deal with here that seem bigger than they are.” [participant #25].

Prolonged exposure to isolated, confined, and extreme (ICE) conditions, such as long-duration spaceflight, increases the risk of adverse cognitive or behavioral conditions^1,2. Mood disorders, sleeping difficulties, psychosomatic symptoms, impaired cognition, and interpersonal conflict have all been described during missions on space stations, as well as in space analog environments such as Antarctic research stations^2–4. With ongoing plans to set foot on the Moon and Mars, humans will extend greater distances into our solar system for longer durations. Crew members will be exposed to new or elevated threats with reduced rescue opportunities, and will be required to perform complex tasks with unprecedented autonomy and teamwork. Mission success and safety will depend on the crew’s mental composure and ability to maintain high levels of operational performance. Reducing the impact of the psychological stressors encountered in these environments has become one of the highest priorities in space exploration⁵.

Previous work suggests a link between the reduced sensory stimulation and sensory monotony experienced in ICE environments with a loss of pleasure, satisfaction, and engagement⁶. Crew members are challenged with living and working in a limited habitable volume for prolonged periods, devoid of new smells, sights, sounds, or surfaces other than the enclosure itself, with limited opportunities for external interactions, persistently high levels of danger, and a lack of ability to ‘escape’. In such situations, augmented sensory stimulation can offer novelty, prevent boredom, reduce stress, and restore cognitive resources – thereby providing opportunities for targeted countermeasure implementation to promote individual behavioral health⁶.

Virtual Reality (VR) is particularly promising for use in ICE environments, with many new applications in the areas of psychology and healthcare^7,8. It offers a portable, configurable platform to transport crewmembers beyond the limited sensory exposure, autonomously, privately, and confidentially. Previous studies have already shown the feasibility and acceptability of exposure to VR-based nature scenes in different spaceflight analog missions, and provided an opportunistic, preliminary analysis of the impact of specific scene content on perceived immersiveness, restoration, and mood^9,10. To effectively implement VR as a behavioral health countermeasure, further optimization of such sensory stimulation experiences will be needed. While the industry is rapidly developing features to improve interaction and presence in the virtual world^11,12, it remains unclear, however, which aspects of VR-based sensory stimulation would be most beneficial for relaxation and restoration in ICE environments.

Here, we conducted a study on overwintering crew members at a high-fidelity analog for long-duration spaceflight— the Antarctic Amundsen-Scott South Pole station¹³. Our aim was to manipulate various aspects of VR-based scene presentation, including scene content, duration, and temperature-based sensory augmentation, to understand how these attributes contribute to the VR experience, and help to maintain behavioral health and performance in ICE. We focused on outcomes of perceived quality, mood impacts, and individual preferences using questionnaires and qualitative feedback.

Methods

Amundsen-Scott South Pole Station

Amundsen-Scott Station is a research base located in Antarctica at the geographical South Pole (90°S), on the nearly featureless Antarctic ice sheet at an altitude of 2835 meters (9306 ft). Temperatures have varied from −13.6° C to −82.8° C, with an annual mean of -49° C, and in the austral, winter the sun does not rise for six months. Each year during the winter, some 50 crew members remain confined to the station and completely isolated from the rest of the world for about nine months, as no outside help can reach the base over land or air. During this period, the crew consists of scientists and support personnel (e.g. electricians, mechanics, cooks) to maintain the station and perform scientific activities. No supplies can be provided and evacuation of the crew in case of emergencies is highly unlikely. The station offers an exceptional opportunity for researchers to study the effects of prolonged isolation and confinement on human health and performance, and Amundsen-Scott Station stands as one of the few places on Earth that closely simulates the extreme environment astronauts may encounter during long-duration space missions. It has been used by the National Aeronautics and Space Administration (NASA), in collaboration with the United States Antarctic Program and National Science Foundation, as an analog environment for testing technologies and behavioral health countermeasures that could be employed during space missions, using a larger sample than spaceflights could provide.

Participants

Participants were recruited from the 2022 winter-over crew assigned to a one-year stay at the station. Exclusion criteria included a history of neurological or psychiatric disorders; uncorrected visual, auditory, or other perceptual impairments that would interfere with the ability to view VR and computer videos; and a history of discomfort with VR. Twenty-five out of 44 crew members volunteered to undergo repeated sensory stimulation sessions during a five-week period in September and October 2022, at the end of nine months of isolation during the Antarctic winter. Eight were female (32%), and the mean age was 34.5 ± 9.0 years. Seven (28%) had a Master’s degree or higher, and 11 (44%) had prior winter-over experience. Four (16%) had previous experience with VR. All participants underwent physical and psychological screening as part of the crew selection process and were deemed fit for winter-over service before deployment. The study was ethically approved by the Institutional Research Boards of NASA and Massachusetts General Hospital, and has been in accordance with the Declaration of Helsinki. All participants provided informed consent.

Study design and procedures

This study followed a repeated measures design in which participants were exposed to 16 variations of a sensory stimulation protocol in a pre-scheduled, randomized order. Over the course of the five-week study period, participants performed the protocol approximately three times per week and not more than one time per day at their preferred times, depending on availability of the research equipment. Before the first sensory stimulation exposure, participants were familiarized with the equipment and data collection methods. All subsequent testing sessions were self-administered by participants. Study coordination and data management was overseen by the station’s physician (JMW). Data collections were performed in a private room at room temperature, in the seated position. Participants were asked not to use the equipment outside of study procedures.

The protocol consisted of watching a video scene on an HTC Vive Focus 3 VR headset (HTC, New Taipei City, Taiwan), or on a regular laptop as a control condition (Alienware, Miami, FL, USA). Participants watched eight different videos that included filmed urban or nature scenarios with natural audio, without the possibility of manipulating or interacting with the environment other than through observation. Some of the videos involved motion (walking, driving), while other scenes were stationary with transitions to new vantage points at a regular pacing. Video length ranged from 10 to 26 minutes, but participants were instructed to watch a self-chosen fragment for either 4 or 10 minutes, and each of the eight available videos was watched twice to cover both durations (resulting in 16 separately performed experimental conditions). For some of the VR-based videos an Embr Wave 2 wristband (Embr Labs, Boston, MA, USA) was worn on the right palmar side of the wrist, to provide additional haptic presentation via temperature cues complementary to the visual content shown (e.g. hot for a sunny day in Portugal; cold for an Autumn night in New York City). Participants activated the Embr at the start of a video, which then provided either heating or cooling waves using a closed-loop temperature control system throughout the VR experience. An overview of the different videos and corresponding temperature cue settings is shown in Table 1. Representative still images from each video are provided in Supplementary Notes 1, and the full videos can be made available on request to the corresponding author.

Table 1.

Sensory stimulation videos and temperature cue settings*

Scene	Video length (mm:ss)	Modality	Content	Motion	Temperature cue
New York City	25:44	VR	City scenes with urban skyscapes, people, and traffic	Walking	Cold
Chicago	10:04	VR	City scenes with urban skyscapes, people, and traffic	Driving	None
Boston	15:25	VR	City park scenes with urban skyscapes and people in view	Stationary	None
Houston	15:03	VR	City park scenes with urban skyscapes and people in view	Stationary; riding a streetcar	Warm
Portugal	13:42	VR	Nature scenes including vistas, forest and beach scenes	Stationary	Warm
Ireland	13:22	VR	Nature scenes including vistas, forests, and beach scenes	Stationary	None
New York City	11:53	Laptop	City scenes with urban skyscapes, people, and traffic	Walking	None
Ireland	13:19	Laptop	Narrated travel video withmostly nature, some city scenes	Stationary	None

^*Participants watched a self-chosen fragment of each setting twice—for a 4-min and a 10-min duration—resulting in 16 pre-scheduled, randomized experimental conditions.

Each of the 16 experimental conditions was labelled in four categories to assess different attributes of the sensory stimulation, including modality (VR vs. a laptop control condition), scene content (nature vs. city environments), duration (4 vs. 10 min), and the use of sensory augmentation (with or without temperature cues). Motion features were inconsistent across the videos and did not enable clear categorization for pairwise comparison, but were included for analysis of subjective feedback as an exploratory attribute that may inform future exploration of this feature.

For each data collection, participants completed a questionnaire battery immediately before and after engaging with the scenes to assess acceptability and perceived quality of the experience, impact on mood, and individual preferences. An overview of collected questionnaires is shown in Fig. 1. After completing all 16 conditions, participants completed a debrief survey with open-ended questions for additional qualitative feedback.

Fig. 1. Overview of the experimental protocol.

Including collected questionnaires.

Measures of perceived quality

The Value of Virtual Reality Questionnaire (VVRQ) was completed post-session as a measure of perceived value of the sensory stimulation experience⁹. The questionnaire consists of six statements about the sensory stimulation system that were rated on a 7-point Likert scale from 1 (Strongly disagree) to 7 (Strongly agree). Scores were calculated as the sum of the six items, ranging from 6 to 42. A response on at least five items (83%) was required for inclusion in analysis, and missing responses were scored as the mean of the other items.

The Modified Reality Judgement and Presence Questionnaire (MRJPQ) was completed post-session as a measure of scene immersiveness^9,14,15. It consists of 10 questions on the extent the person felt they were present (e.g. “To what extent did you feel you were physically in the virtual world?”), experienced the situation as real (e.g. “To what extent was what you saw in the virtual world similar to reality?”), and felt an emotional impact (e.g. “To what extent did the virtual world make you feel emotions (anxiety, sadness, happiness, etc.)?”). The items were rated on a scale from 0 (Not at all) to 10 (Absolutely), and a total score was calculated as the sum of the 10 items, ranging from 0 to 100. A response on at least eight items (80%) was required for inclusion in analysis, and missing responses were scored as the mean of all other items.

The Perceived Restorativeness Scale (PRS) was completed post-session as a measure of the subjective sense of emotional restoration experienced as a result of the sensory stimulation^16,17. It is made up of 26 items and measures the perception of five established restorative factors (Being-Away, Fascination, Coherence, Scope, Compatibility). The items were rated on a 7-point scale from 0 (Not at all) to 6 (Completely), and a total score was calculated as the sum of all items, collapsing across subscales, ranging from 0-156. A response on at least 20 items (77%) was required for inclusion in analysis, and missing responses were scored as the mean of the other items within the same subscale.

Measures of impact on mood

The abbreviated Profile of Mood States (POMS) was completed both pre- and post-session to assess the impact of the sensory stimulation on reported mood^18,19. It consists of 40 items describing how a person feels at a particular moment in time, which are rated on a scale from 0 (Not at all) to 4 (Extremely). Items are grouped into five negative (Tension-anxiety, Depression, Anger-hostility, Confusion-bewilderment, Fatigue) and two positive (Vigor, Esteem-related affect) subscales, and items are averaged for each subscale. The Total Mood Disturbance score was calculated as the sum of scores from negative subscales subtracted by the sum of positive subscales, ranging from -8 (no mood disturbance) to 20 (high mood disturbance). If more than two items were missing within one subscale, the Total Mood Disturbance score was considered invalid. For each experimental condition, Total Mood Disturbance was analyzed as change from baseline scores (pre-session scores were subtracted from post-session scores), so that positive values reflect an increase in Total Mood Disturbance, while negative values point to a decrease.

The State Anxiety Inventory for adults (STAI-STATE) was completed both pre- and post-session to evaluate the impact of the sensory stimulation on anxiety levels. It consists of 20 statements that ask a person how they feel at that moment, and each item is rated on a 4-point scale from 1 (Not at all) to 4 (Very much so), and a total score was calculated as the sum of all items, ranging from 20 (no anxiety) to 80 (high anxiety). A response on at least 15 items (75%) was required for inclusion in analysis, and missing responses were scored as the mean of all other items. For each experimental condition, anxiety scores were analyzed as change from baseline (pre-session scores were subtracted from post-session scores), so that positive values reflect an increase in anxiety, while negative values point to a decrease.

Qualitative Feedback

Participants also provided qualitative feedback after each session and after completing all 16 conditions to assess acceptability and individual preferences for the sensory stimulation protocol. Post-session, the surveys included open-ended questions asking about their experience of the sensory stimulation in general, and, if applicable, whether they liked the Embr feature or the inclusion of motion vs. fixed locations in the scenes. Participants were also asked whether they experienced motion sickness. The post-study debrief included open-ended questions on participants’ experiences during the study, use of the equipment, preferred sensory stimulation conditions, how useful they thought the sensory stimulation was for promoting relaxation or behavioral health, and recommendations on protocol and equipment improvements for overwintering crew members and astronauts on missions to Mars.

For each participant, responses were coded on the inclusion of various aspects of the sensory stimulation and organized into various themes (modality, scene content, motion, duration, the inclusion of temperature cues, equipment and video quality, and recommendations). For each theme, counts were made of mentioning positive and negative experiences (e.g. whether a participant had a positive experience with city scenes, or felt that the duration was too short), suggestions (e.g. a desire for a higher variety of experiences), as well as mentioned contributing factors (e.g. seeing new faces contributed to a positive experience, or motion sickness contributed to a negative experience). Responses that overlapped between participants were collapsed into categories. For each participant, after all their post-session and post-study responses were assessed, their individual preference within each theme was determined (e.g. for the duration theme, whether they preferred 4 vs. 10 minutes, or had no preference). Exemplar quotes were identified.

Statistical Analysis

Statistical analysis and generation of graphs were performed using GraphPad Prism version 9.5.0 for MacOS (GraphPad Software, Boston, MA, USA). To assess how different attributes contribute to the sensory stimulation experience, outcome measures from all experimental conditions were averaged within-subjects into two groups for each of the four categories for pairwise comparisons (4 Laptop conditions vs. 12 VR conditions; 10 City conditions vs. 6 Nature conditions; 8 4-min conditions vs. 8 10-min conditions; 6 conditions with Embr vs. 6 conditions without Embr). As none of the laptop-based conditions included the Embr, these conditions were excluded from this category. For each category, normality of the difference between the two groups was assessed using the Shapiro-Wilk test, and visual inspection of histograms and Q-Q plots. Outliers were removed. Pairwise comparisons were performed using the two-tailed paired t-test for normal distributions, and the two-tailed Wilcoxon signed-rank test for non-normal distributions. Bonferroni corrections were done to correct for multiple testing within each outcome measure, so that differences were considered statistically significant if p < 0.0125. Data is expressed as the mean and standard deviation unless stated otherwise.

Results

All participants completed the study. A total of 401 sensory stimulation sessions were recorded, with one subject (#11) accidentally performing one experimental condition twice. One subject (#20) missed 1 of the 16 different experimental conditions, and this condition was therefore not included in analysis. For 2 subjects (#18 and #21), one of the experimental conditions was excluded from analysis due to invalid sensory stimulation exposure (no sound, incorrect video settings). VVRQ, MRJPQ, and PRS responses were missing in one experimental condition in one subject (#16), and STAI-STATE responses were entered incorrectly in one condition in another subject (#20). These conditions were therefore not included in the analysis of these questionnaires. One subject (#9) did not complete the VVRQ, MRJPQ, PRS, and qualitative feedback in 7 out of 16 experimental conditions. This subject was completely excluded from the analysis of the VVRQ, MRJPQ, and PRS, but remained included for analysis of his remaining qualitative feedback.

In the remaining questionnaires for which scores were calculated, 3 out of 2,280 items (0.1%) were missing for VVRQ; 7 out of 3800 items (0.2%) for MRJPQ; 6 out of 9880 items (0.1%) for PRS; 37 out of 15,880 items (0.2%) for POMS; and 22 out of 7920 items (0.3%) for STAI-STATE. None of the questionnaires had more than 1 missing item. Three questionnaires missed an answer on one of the post-session open-ended questions, and no answers were missing from the post-study qualitative debrief.

Perceived sensory stimulation quality

Figures 2 to 4 show rankings of the 16 test conditions for perceived VR quality (value, immersiveness, restorativeness). Figure 5 shows within-subject comparisons of perceived VR quality for four different aspects of sensory stimulation: modality (Laptop vs. VR), content (City vs. Nature), duration (4 vs. 10 min), and temperature-based sensory augmentation (No Embr vs. Embr).

Fig. 3. Perceived immersiveness of the sensory stimulation per test condition.

Ranked from highest to lowest mean score. Box and whisker plots show range and interquartile intervals.

Fig. 2. Perceived value of the sensory stimulation per test condition.

Ranked from highest to lowest mean score. Box and whisker plots show range and interquartile intervals.

Fig. 4. Perceived restorativeness of the sensory stimulation per test condition.

Ranked from highest to lowest mean score. Box and whisker plots show range and interquartile intervals.

Fig. 5. Within-subject comparisons of perceived value (VVRQ), immersiveness (MRJPQ), and restorativeness (PRS) of the sensory stimulations in 4 different aspects of sensory stimulation.

Including modality (Laptop vs. VR), content (City vs. Nature), duration (4 vs. 10 min), and temperature-based sensory augmentation (No Embr vs. Embr). * reflects a p-value < 0.125 (two-tailed paired t-test or Wilcoxon signed-rank test) to indicate significance after Bonferroni correction.

Value of Virtual Reality (VVRQ) scores were significantly higher in VR vs. laptop-based sensory stimulation (34.0 ± 6.0 vs. 26.1 ± 8.0; W = 298, p < 0.001), in nature vs. city environments (33.7 ± 5.2 vs. 31.1 ± 6.9; t = 2.984, p = 0.007), and after 4 vs. 10 minutes of sensory stimulation (32.7 ± 5.9 vs. 31.4 ± 6.2; t = 2.956, p = 0.007). Perceived value of VR was not significantly different in sessions that included the Embr temperature cues vs. without the Embr (33.7 ± 6.2 vs. 34.3 ± 6.0; t = 1.353, p = 0.189).

Scene Immersiveness (MRJPQ) scores were significantly higher in VR vs. laptop-based sensory stimulation (66.6 ± 15.8 vs. 35.8 ± 17.3; t = 8.727, p < 0.001), and in nature vs. city environments (63.3 ± 14.0 vs. 56.3 ± 16.0; t = 3.015, p = 0.006). Scene immersiveness was not significantly different after 4 vs. 10 minutes of sensory stimulation (59.1 ± 13.2 vs. 58.8 ± 15.7; t = 0.278, p = 0.784), or with the addition of Embr temperature cues vs. no Embr (66.5 ± 16.8 vs. 66.7 ± 15.8; t = 0.105, p = 0.917).

Perceived Restorativeness (PRS) scores were significantly higher in VR vs. laptop-based sensory stimulation (106.2 ± 21.0 vs. 79.2 ± 27.4; W = 300, p < 0.001), and in nature vs. city environments (114.5 ± 20.2 vs. 90.4 ± 27.3; t = 4.255, p < 0.001). Perceived restorativeness was not significantly different after 4 vs. 10 minutes of sensory stimulation (99.6 ± 20.5 vs. 99.3 ± 22.2; t = 0.181, p = 0.858), or with the addition of Embr temperature cues vs. no Embr (103.3 ± 22.4 vs. 109.1 ± 21.9; t = 2.000, p = 0.059).

Impact on mood

Figures 6 and 7 show rankings of changes in mood (Total Mood Disturbance, state anxiety) for each of the 16 test conditions. Figure 8 shows within-subject comparisons of mood changes for four different aspects of sensory stimulation: modality (Laptop vs. VR), content (City vs. Nature), duration (4 vs. 10 min), and temperature-based sensory augmentation (No Embr vs. Embr).

Fig. 6. Change in Total Mood Disturbance after the sensory stimulation per test condition.

Ranked from lowest to highest mean score. Box and whisker plots show range and interquartile intervals.

Fig. 7. Change in anxiety scores after the sensory stimulation per test condition.

Ranked from lowest to highest mean score. Box and whisker plots show range and interquartile intervals.

Fig. 8. Within-subject comparisons of changes in Total Mood Disturbance (POMS) and anxiety levels (STAI-STATE) from pre- to post-session in 4 different aspects of sensory stimulation.

Including modality (Laptop vs. VR), content (City vs. Nature), duration (4 vs. 10 min), and temperature-based sensory augmentation (No Embr vs. Embr). * reflects a p-value < 0.125 (two-tailed paired t-test or Wilcoxon signed-rank test) to indicate significance after Bonferroni correction.

Overall change in POMS Total Mood Disturbance between pre- and post-session data collection across all subjects and sessions was -0.2 ± 1.3 on a scale ranging from -8 to 20. There were no significant differences in the change in Total Mood Disturbance changed in VR vs. laptop-based stimulation (-0.2 ± 0.7 vs. 0.0 ± 0.6; t = 1.648, p = 0.112), nature vs. city environments (-0.4 ± 0.6 vs. -0.0 ± 0.7; t = 2.691, p = 0.013), 4 vs. 10 minutes of sensory stimulation (-0.1 ± 0.6 vs. -0.2 ± 0.6; t = 0.498, p = 0.623), or with the addition of Embr temperature cues vs. no Embr (-0.1 ± 0.8 vs. -0.3 ± 0.7; t = 2.346, p = 0.028).

Overall change in STAI-STATE anxiety scores between pre- and post-session data collection across all subjects and sessions was -0.1 ± 5.8 on a scale ranging from 20 to 80. Changes in STAI-STATE scores reflected significantly less anxiety in VR vs. laptop-based sensory stimulation (-0.6 ± 2.7 vs. 0.6 ± 1.8; t = 2.903, p = 0.008), and in nature vs. city environments (-1.7 ± 2.5 vs. 1.0 ± 3.3; t = 4.805, p < 0.001). There were no significant differences in changes in STAI-STATE scores after 4 vs. 10 minutes of sensory stimulation (0.3 ± 3.0 vs. -0.4 ± 2.9; t = 1.493, p = 0.149), or with the addition of Embr temperature cues vs. no Embr (0.4 ± 4.1 vs. -0.9 ± 3.2; t = 1.891, p = 0.071).

Individual preferences in qualitative feedback

Based on the assessment of all qualitative feedback at the end of the study, all participants (100%) preferred VR-based sensory stimulation over using a laptop. Fifteen out of 25 participants (60%) preferred watching nature scenarios, while 3 participants (12%) preferred city scenes, and 7 (28%) did not seem to have a preference. Below are various representative quotes that underscore individual variation in content preference:

“Overall I preferred the nature scenes from Ireland and Portugal over anything else; they were the easiest to get lost in and left me feeling the most relaxed. The city scenes were very exciting though and definitely did a good job of holding my attention.” [participant #11].

“I prefer cities to natural, rural environments. This made me feel at peace, like I was on a nice drive into a city.” [participant #7].

“Cityscapes are a bit intense for me anyway because I grew up in a more rural area. I love to be in them in person and see all the action but I feel very sensitive to stimulation right now because I haven’t left one square mile for almost a year.” [participant #6].

“More nature, way more nature. Nobody wants to see car when we winter-over…” [participant #20].

Seven out of 24 participants (29%) that had a positive experience with city scenarios mentioned that they enjoyed being in the urban environment. At the same time, 9/19 participants (47%) mentioned that the same environment contributed to a negative experience of one or more of the city scenarios:

“I personally like city scapes and crowds from time to time. The traffic and people and all the noises of the city can be very relaxing.” [participant #3].

“The whole city scape experience was overstimulating and I was anxious for it to end.” [participant #6].

“It is infuriating watching cars not following traffic laws.” [participant #4].

Similarly, 12/24 participants (50%) mentioned that being around people contributed to a positive sensory stimulation experience, while 10/19 participants (53%) mentioned that being around people contributed to a negative experience:

“Loved the people watching, city noises and familiar usual things we miss down here.” [participant #19].

“It was strange being around an immense amount of people. After a year at the South Pole you see the same 44 people every day and it takes time to reintegrate with society once you leave this place. I love exploring cities, but seeing the dark and hundreds of people around me at that moment is not what I was expecting.” [participant #17].

“I really hate places with large crowds that I can not get away from. I can honestly say I never wanted to visit New York before this, and this [experience] cements that I do not want to ever do it.” [participant #4].

Overall, 7 participants (28%) explicitly expressed a desire for more nature scenarios, while only 1 (4%) also recommended more urban environments. Interestingly, various elements of scene content were commonly mentioned to contribute to a positive experience in both nature and city scenarios, and were not mentioned as a contributor to any of the negative experiences across all participants. For instance, 16 participants (64%) mentioned the sound and movement of flowing water, 9 (36%) mentioned watching bodies of water such as lakes or oceans, and 8 (32%) mentioned the virtual sunshine. For urban scenarios, 10 participants (40%) also mentioned parks as a positive element. On a particular note, 20 participants (80%) mentioned animals as a positive element, with 8 participants (32%) explicitly expressing a desire for more animals:

“Mmh, have I already mentioned more nature and animals? If I didn’t, please add more nature and more animals. If I did already, then just add them, thanks.” [participant #9].

“Also: more dogs. Just…set up a 360 cam in different parts of a doggy day care. Or a cat cafe.” [participant #2].

With regards to preferred duration of the sensory stimulation, 18 participants (72%) preferred the duration of 10 minutes, while only 1 participant (4%) preferred the duration of 4 minutes, and 6 (24%) did not express a clear preference:

“I found with the shorter session I was prepping myself for the timer to go off, and disappointed every time. With the long sessions I could more easily forget that I’m on a timer and could more easily relax into the session.” [participant #4].

“The shorter videos with the VR and Embr were my favorites, especially the ones near running water and away from the city. I found those the most calming. The longer ones did not match well with me limited attention span.” [participant #1].

“Why did it only have to be 4 min? Can I have a little more?” [participant #2].

Using the Embr for temporary-based sensory augmentation was preferred in 8 participants (32%), while 13 (52%) preferred sensory stimulation without the Embr, and 4 (16%) did not express a preference. In all Embr scenarios combined, 58 ± 11% of the participants responded that they liked the sensory augmentation during a session, with 64 ± 1% for the “warm” Embr scenarios alone, and 46 ± 14% for the “cold” Embr scenarios:

“The Embr is so nice. It made the sun in the videos feel warm” [participant #1].

“I barely noticed it, so it had a minimal impact on me” [participant #24].

“It is just cooking my wrist.” [participant #20].

Reasons mentioned for not liking the Embr included a lack of feeling of immersion, not being able to notice the Embr enough, the Embr sensations not corresponding well to the scenes, or feelings of discomfort (e.g. too hot or too cold). Nevertheless, 15 participants (60%) expressed that they saw potential in using Embr or other temperature-based sensory augmentation to improve the sensory stimulation experience. Recommendations included a larger area of impact (e.g. whole room, heater, fan), a stronger temperature sensation, and better synchronization with the scenes:

“I think the Embr wasn’t enough temperature gradient over a large enough amount of my body surface area for me to really notice a difference. I think something like a desk fan or a space heater might make the experience more immersive. For astronauts, maybe a circulating water jacket where you control the water temperature with a chiller or heater?” [participant #24].

With regards to motion, 6 participants (24%) expressed a preference for scenes that incorporated motion. Four (16%) rather engaged with stationary scenes, and 15 (60%) did not express a clear preference:

“Movement is nice because you can see more of your surroundings and be exposed to more stimuli. However I do wish I could control the journey instead of just being a spectator.” [participant #3].

“I liked the static images, so that I could look around freely rather then get “pulled along”, which would remind me that I am in fact watching a video” [participant #16].

From the 18 participants (72%) that mentioned at least one negative experience with motion, contributing factors included that it didn’t feel natural or real (44%), that it felt disorienting or that motion had an impact on visuals (61%), physical discomfort such as motion sickness or headache (50%), or that it wasn’t relaxing (17%). Thirteen participants (52%) reported motion sickness in one or more of the scenes that contained motion. Three participants (12%) reported motion sickness in scenes that were stationary. Motion sickness was reported in 37 out of 399 sessions (9%).

In accordance with the above findings, multiple participants further highlighted the desire for individualized sensory stimulation. Seventeen participants (68%) expressed a desire for more variation in the scenarios to choose from, and 10 participants (40%) expressed a desire for more autonomy, including autonomy in scheduling the sensory stimulation sessions (6 participants, 24%), scenario selection (5 participants, 20%), and control within the VR experience, such as the ability to change environments themselves instead of making such transitions without control (4 participants, 16%):

“The enjoyment of each content type is highly subjective. If there was a vast library of content that users could choose from they would be able to find something that resonated with them. As an anecdote, the movie the Martian detailed each person’s unique music collection; I think VR content would naturally be similar to this.” [participant #10].

“If I could have spent an hour with nature scenes any time I wanted, I would have enjoyed that.” [participant #21].

“I didn’t like the lack of control of the motion. I would have like to look at the greenery in the park longer but the camera man turned toward the street. Now, who wants to look at a busy street on a camera? Anyone? Anywhere on earth? Definitely not here.” [participant #6].

In addition, 7 participants (28%) expressed a desire for more interaction such as being able to move around or the inclusion of game elements.

“Sometimes, being taken on a tour of a virtual landscape can be really enjoyable… However, being able to explore the environment on my own and find my very own “Zen Spot” that I like to visit again and again would be very rewarding.” [participant #3].

Finally, several other relevant mentions of personalized sensory stimulation were made:

“[I recommend] tailored scents for individual astronauts that evoke feelings of home, favorite places, or childhood. [Participant #5].

“Give the people what they want. Let the crew members direct their own sensory experience because they are coming to each session with their own unique mental state. Maybe one day they want to explore a new video they haven’t seen yet but another day they just want to return to that one peaceful waterfall and just be in that place.” [participant #6].

“I have to believe that for the highest measure of individual impact, the conditions would have to be laid out alongside each participants respective personalities.” [participant #8].

Discussion

In this study, we collected questionnaires and qualitative feedback in an Antarctic winter-over crew to understand how VR-based sensory stimulation is best implemented as a behavioral health intervention for isolated, confined, and extreme (ICE) conditions such as long-duration spaceflight. Previous studies have already demonstrated the feasibility and acceptability of using VR in various spaceflight analog missions^9,10, with an opportunistic and preliminary analysis of the impact of specific scene content on perceived immersiveness, restoration, and mood. Here, we have expanded on this prior work by implementing VR in the operational and high-fidelity analog environment of Antarctica, and by systematically investigating preferences for various attributes of VR, including scene content, duration, and the use of multisensory stimuli.

Our results showed clear overall preferences for using VR over watching videos on a laptop, for viewing nature scenarios over urban scenarios, and for longer VR exposure times. VR-based nature scenarios received overall high scores of perceived value, immersion and restoration, and had the most beneficial impact on mood, while other scenarios showed mixed results. Nevertheless, several participants also enjoyed or even preferred city scenarios over nature, with higher scores of perceived quality and mood. Similarly, the use of temperature-based sensory augmentation and the incorporation of motion in the scenarios was experienced with mixed positive and negative results. Recognizing these individual differences, many participants expressed their desire for a more tailored approach, recommending more variation in scene content, and autonomy in scheduling and selecting the scenarios.

The POMS was the only outcome measure that did not show a statistical effect of VR or nature scenes, likely due to a floor effect. Both pre- and post-session, Total Mood Disturbance scores were generally low, making the detection of any beneficial effects of sensory stimulation difficult. An exploratory descriptive analysis of POMS subscales showed similar small changes between pre- and post-session data collections across all participants and sessions, including -0.0 ± 0.3 for Tension-Anxiety, -0.1 ± 0.2 for Depression-Dejection, -0.0 ± 0.2 for Anger-Hostility, -0.0 ± 0.2 for Confusion-Bewilderment, -0.1 ± 0.4 for Fatigue-Inertia, -0.1 ± 0.4 for Vigor-Activity, and -0.0 ± 0.3 for Esteem-related Affect. No hypothesis testing was performed on these subscales considering the limited value in a small sample size and to reduce the type 1 error rate. Still, when ranking mean scores for each scenario, Total Mood Disturbance was lowest for VR nature scenarios.

The observed preference for nature in our study corresponds with findings from other researchers^9,14,20–22, and fits within the interpretative framework of Attention Restoration Theory^23,24. According to this theory, exposure to nature can be restorative—with reduced stress, improved mood, lower fatigue, and renewed productivity—as a result of the psychological distance from routine mental concerns (i.e., a sense of ‘being away’), combined with effortless, interest-driven attention (‘fascination’) grounded in an environment of substantial scope (‘extent’). Each of these contributing mechanisms is typically found in natural settings. While people living and working in ICE environments cannot freely seek out such settings, our results confirm previous findings that virtual exposure to nature can be similarly restorative^14,25. When implementing VR nature scenes in various other spaceflight analog settings, Anderson et al.⁹ showed comparable preliminary results on the VVRQ, MRJPQ, PRS, and POMS questionnaires.

Nevertheless, our data showed a clear variability in individual preferences and outcomes. In Attention Restoration Theory, Kaplan puts forth that the environment and the person’s desired intent in the environment must match for it to be restorative (“compatibility”)²⁴. Prior work has suggested that, in general, people subjectively prefer natural settings to urban settings, perhaps because they find it mentally restorative subconsciously²⁶. Similarly, scenes with water tend to be rated more highly than scenes without water²⁷. Beyond this general and perhaps subconscious framework, however, there is still a great degree of personal preference that may influence the effectiveness of sensory stimulation for a given individual¹⁴. In ICE crews, that variation in preference may increase in response to the lack of environmental elements and experiences that are typically present in their lives back home. Where urban scenes have not had restorative effects in average populations^22,28, these kinds of environments may present a sense of normalcy and an escape from their current setting. Our participants repeatedly mentioned elements of greenery, water, animals, and sunshine—all elements the Antarctic winter-over environment is devoid of—, but also in varying degrees people, children, “the hustle and bustle” of the city, and in general the familiarity of environments. What elements or scenarios provide such an experience of familiarity, however, is highly dependent on the individual. By adding variety in available content, crew members are more likely to have access to compatible experiences:

“A typical year back home isn’t spent at the same location. There’s variety. Sometimes people take a trip to somewhere exotic. Sometimes they want to go for a walk in the forest. Sometimes they stop in for a coffee. Nothing about life is static, so the scenarios should reflect that.” [participant #7].

Sensory stimulation in general can fulfill multiple needs, each relating differently to stress and coping⁶. Vessel and Russo identified two needs that are primarily relevant for VR-based applications, including 1) a need to recover from stress and restoration of attentional resources (“blowing off steam”), and 2) a need to engage the human information foraging drive (“learning and exploration”) to prevent boredom and subsequent stress. Our findings of both a general desire for natural scenarios—often associated with restorative properties—and an increased variety of experiences—including those that are highly arousing and present novel stimuli—seem to reflect these different needs. In ICE environments, given the unique circumstances of reduced sensory stimulation and monotony, the potential for an unmet need for information foraging should drive the selection of VR experiences beyond those that provide relaxation and restoration⁶:

“A boat on the water, maybe, for people who would be into that. Dogs. Lots of dog videos. Roller coasters for people who could handle it. Rock climbing, nature walks. And for those who don’t like motion, maybe a nice forest setting, or a beach, an art gallery, a coffee house. There will be crew members from every walk of life on these missions, more variety would be well-received.” [participant #7].

The inclusion of specific types of video gaming or cognitive training programs could be particularly valuable, where clever game design can be used to foster learning and exploration through rewards, hidden surprises, and supporting perceptions of autonomy and competence^6,29.

To leverage the benefits of VR as a behavioral health countermeasure, immersion and achieving presence is key^6,12. In this study, we attempted to improve the immersiveness of the virtual experience by integrating scenario-driven temperature cues. While our data showed mixed results, 23 participants (92%) mentioned that Embr did not increase the feeling of immersion in one or more of the sessions, and about half of the participants seemed to prefer sensory stimulation without the Embr. In these cases, the temperature stimuli of the Embr did not seem to be sufficiently coherent with real-world scenarios, which is in line with previous findings that a lack of credibility of multisensory feedback can negatively affect the user¹². Nevertheless, more than half of the participants also saw potential in temperature-based sensory augmentation. Multisensory VR systems have generally shown positive outcomes in the literature¹², and new technologies may soon allow for the implementation of more realistic thermal stimuli³⁰. Other types of stimuli should be considered as well, such as tactile feedback, taste, or smell. A recent study by Abbott and Diaz-Artiles for instance demonstrated that the addition of digital scents had a positive impact on mood, providing a promising multisensory VR experience to support behavioral health for long-duration spaceflight³¹. Further research is needed to verify the acceptability and added value of such technologies in ICE settings. Moreover, given the emphasis on personalizing the VR experience, it would be prudent to leave the final decision of applying additional stimuli to the end-user.

Immersion and presence can be achieved further by the timing and amount of VR exposure. Except for the VVRQ ratings, our results suggest that 4 minutes is too short for most users. Participants mentioned that it prevented them from getting immersed, or that they wanted to explore more, up to a point of irritation from being teased by having to stop the session. Ten minutes seemed more appropriate, and is more in line with findings from previous studies^9,14,22,31. Several participants expressed that on some occasions, also 10 minutes was not long enough to get settled into the experience, or notice an impact on stress or mood. While operational scheduling constraints in ICE environments may preclude prolonged use of VR, a certain degree of autonomy is likely going to improve the effects of sensory stimulation as a countermeasure. Restorative effects on attention and stress may have different timescales, with physiological effects occurring relatively quickly (4-5 minutes), followed by emotional effects ( ~ 10 minutes), and effects on attention and performance taking longest (e.g. 20+ minutes)^6,21,32. In addition, as expressed by our participants, having the option to experience VR at-will would allow people to make use of opportunities when sensory stimulation is likely to meet their personal needs the most, with full engagement, and lower distraction. Too much autonomy, on the other hand, may be counterproductive, as some of our participants shared a concern for potential overuse:

“I think I would really really enjoy VR. I wonder, though, if this could become a dependency in a place where there are no external stimuli. People here spend a lot of their free time in their rooms watching movies or browsing the internet when available, or even when together with other people some just focus on their smartphones. What would be the effects of having something even more effective in isolating people?” [Participant #9].

Overall, our study was positively received by the Antarctic crew, and 22 (88%) participants expressed that VR-based sensory stimulation has potential as a countermeasure for winter-over or spaceflight missions. For VR to be effective in these environments however, one other factor that was frequently brought forward by participants—and for many perhaps the most important one—was image quality. The scenarios used in this study had varying resolutions and frame rates, and some were below the recommended resolution of 4 K. Twenty (80%) participants mentioned that the quality of the visuals or sounds diminished at least one of the experiences. Reasons included feeling disoriented from a lack of stability during motion, not being able to clearly see objects at greater distances, or pixelation or glitches taking away from the realism. Higher quality scenarios would likely have revealed a higher restorative impact, and this may have included content that is deemed less restorative in traditional VR research. The differences in video quality may also have influenced our comparisons of questionnaire outcomes between scenarios, although the narrative feedback confirms that our results are at least partially driven by factors beyond video quality.

The logistical constraints of performing research in a highly remote, operational environment created some limitations. We were not able to conduct our study until the very end of the winter-over, which coincides with the return of sunlight after an extended period of darkness. This period is often described (anecdotally) to correlate to changes in mood and behavior, resulting from excitement of returning daylight, but also increased levels of activity in preparation of the arrival of summer personnel, and feelings of anticipation that the mission is coming to an end. Even though we randomized our experimental conditions to account for temporal changes, it would be worth monitoring people during a mission period more reflective of the long-term exposure to space-like living conditions, when sensory deprivation is higher, and distractions are lower. Finally, even though this study was designed to investigate a behavioral countermeasure for spaceflight, differences between ICE environments and their unique characteristics could still yield different outcomes³³. While the current results offer valuable insight into the acceptability and restorative effects of VR in an Antarctic winter-over mission, they may not be directly transferable to the spaceflight environment due to the absence of gravity, the presence of space radiation, variations in mission duration, different crew compositions, different personal space and amenities, and other factors. Nevertheless, South Pole station is regarded as one of the highest-fidelity analogs on Earth to help understand future life on missions to the Moon or Mars, and our study provides promising results to be leveraged for the implementation of VR in space.

In conclusion, this study demonstrates the viability and restorative potential of VR-based sensory stimulation during an Antarctic winter-over, and provides insight into preferences for implementation as a behavioral health countermeasure against monotony, sensory deprivation, and stress. Beyond the restorative effects of virtual nature, VR has the ability to provide access to emotions and experiences that are otherwise unavailable in an isolated, confined habitat¹⁰. Interindividual variation in the needs and preferences for sensory stimulation will necessitate a personalized approach. With a wider range of content, state-of-the-art technology and image quality, minimum exposure times, and more autonomy, we believe that VR will present a valuable addition to the behavioral countermeasure suite for long-duration space missions.

Future work could include the characterization of a more personalized and autonomous VR paradigm, and investigating aspects of VR that are relevant for operational implementation. Gaining a better understanding on how autonomy in selecting and scheduling VR experiences compared to pre-assigned trials may benefit restoration, or how sensory stimulation affects the time course of not just subjective, but also physiological and cognitive restoration, will help to tailor VR-based sensory stimulation for use in space. Choosing scenarios that are most compatible with the microgravity environment (e.g. underwater settings, where the discrepancy between walking and floating is reduced, and the perceived mis-match between the visual scene and the true physical environment is minimized) will require further testing in parabolic flights or in spaceflight. Leveraging the rapidly evolving industry will enable access to the latest immersive multisensory technologies, and scenarios that are specifically aimed at restoration could be further optimized. For instance, interactive experiences could be developed that leverage VR-guided relaxation techniques such as breathing or meditation³⁴, and could include the use of biofeedback. Specific content, such as the inclusion of animals or people, seems to be particularly impactful after isolation and confinement, and their restorative benefit could also be further explored. All of the above will accelerate our ongoing efforts to develop optimized VR countermeasures to support crew members, so that they can succeed on missions to the Moon and Mars.

“I did have a realization when I saw a group of people cross the street. All at once I remembered that the entirety of the human race (besides other research bases, ISS, etc.) have been living their regular scheduled life the entire time I’ve been here. The world has just kept going and none of the people in the simulation probably thought about the South Pole or the ISS once in that entire time. With one sunset and one sunrise in a year, flat white, and everything you see is man-made it’s easy for time to stop, slow and linger.” [participant #6].

Author contributions

Study concept and design (S.T., A.F., P.B., G.S., J.B., A.S.). Acquisition of data (S.T., J.M.W.). Data Analysis (S.T.). Drafting of manuscript (S.T.). Critical revision of the manuscript (J.M.W., G.S., J.B., A.S.). Approval of final manuscript (S.T., J.M.W., A.F., P.B., G.S., J.B., A.S.).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. Moreover, the data can be accessed upon request to the NASA Life Sciences Data Archives.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41526-025-00471-2.