The Relaxing Effect of Virtual Nature - Immersive Technology Provides Relief in Acute Stress Situations

Abstract

INTRODUCTION
Research on attention restoration theory brought evidence on the recreational effect of natural environments on humans (White et al., 2013). Since virtual reality (VR) creates a compelling illusion of being present in the virtual world (Heeter, 1992), it is reasonable to use this technology to expose humans to such restorative environments (Berto, 2014). This approach can provide relief from everyday stress but is especially valuable for those suffering from acute stress and emotional strain whose access to nature is limited, for instance due to health-related issues or certain working environments (Anderson et al., 2017).

The purpose of the present study is to investigate the possibility to elicit positive feelings and provide relief from distress using VR and natural virtual environments (VE) during acute stressful situations (e.g. medical treatments). Furthermore, we explore whether a high level of immersion increases the recreational effect of natural VEs.

METHOD
We developed a VR version of the commonly used Trier Social Stress Test (Kirschbaum, Pirke & Hellhammer, 1993) as a standardized instrument to evoke acute cognitive and social stress. In a preliminary study (N = 47) we compared our VR-TSST to the original real-world protocol to ascertain efficacy and validity of our method. We observed a significant increase of physiological stress (heart rate variability, HRV; cortisol) and psychological stress (anxiety, affect) compared to a baseline measurement. Thus, we consider our VR-TSST to be a reliable and efficient stress induction method. In the main study, subjects were stressed using our VR-TSST. A second stress induction was announced at the end of the VR-TSST to keep the stress level constant. Subsequently, subjects were randomly assigned to one of three conditions: participants in the VR group were exposed to a realistic simulation of an underwater environment shown on a head-mounted display (high immersion). The desktop group experienced the same VE on a non-immersive desktop screen (low immersion). Subjects in the control group were not exposed to the VE but were left waiting without medial distraction (no immersion). Objective (HRV, cortisol) and subjective (self-report) stress indicators were measured. Questionnaires were used to assess anxiety (State-Trait Anxiety Inventory), affect (Positive and Negative Affect Schedule), immersion (Game Experience Questionnaire), and presence (Igroup Presence Questionnaire).

RESULTS
A total of 62 healthy subjects aged 18 - 48 years (M = 22.6, SD = 5.36) participated in the study. HRV changed significantly in all three groups over the course of the experiment with higher values (indicating low stress) during baseline and relaxation phase and lower values (indicating high stress) during the stress induction phase. Further, in the VR condition significantly higher HRV values were measured in the relaxation phase than the desktop and the control conditions. Mean cortisol levels decreased during the post-induction phase about 27.66% in the VR group, 28.74% in the desktop group, and 18.18% in the control group. Although these differences were not significant, we found a medium effect size (.70). Subjective measures showed similar tendencies. A significant decrease of anxiety and negative affect was observed in all groups in the relaxation phase. The VR group reached the lowest scores and differed significantly from desktop and control group. The mean positive affect score was highest in the VR group. Desktop and control group did not differ significantly in neither of the observed variables. Additionally, immersion and anxiety were negatively correlated, while immersion and positive affect were positively correlated.

CONCLUSION
Natural virtual environments can be used to enhance mood and well-being. Our results highlight that deeper immersion facilitates relaxation. Thus, especially when access to nature is limited, VR is as an effective solution to reduce acute stress and anxiety.

Full text

The Relaxing Effect of Virtual Nature:
Immersive Technology Provides Relief in
Acute Stress Situations
Stefan LISZIOa,
1
, Linda GRAFa and Maic MASUCHa
a Entertainment Computing Group, University of Duisburg-Essen, Germany
Abstract. The present study investigates the possibility to provide relaxation using
virtual reality (VR) technology and natural virtual environments (VE) during acute
stress situations (e.g. medical treatments). To explore the relaxing and mood
enhancing effect of immersion, 62 participants were exposed to a realistic
underwater VE either in VR or on a desktop screen after stress induction using a
VR-TSST. Systematic changes in physiological (heart rate variability, cortisol) and
psychological (anxiety, affect) measures were observed: The VR group experienced
significantly lower stress and higher positive affect than the desktop group and a
control group. Our findings demonstrate the mood enhancing effect of immersion in
virtual nature and thus the benefit of VR in situations of acute emotional strain.
Keywords. Virtual Reality, natural virtual environments, stress, anxiety, affect,
mood induction, Trier Social Stress Test (TSST), immersion, presence
1. Introduction
Findings from environmental psychology highlight the numerous advantages of certain
natural environments in their relaxing and calming effects. In particular, research on
attention restoration theory [2] brought evidence on the recreational effect of natural
environments on humans [1]. Nature visits provide relief from everyday stress but are
especially valuable for those suffering from acute stress and emotional strain. However,
circumstances limiting access to such restorative environments [3] are manifold. For
instance, certain working environments [4], immobility or health-related isolation, or
certain medical treatments [5] may preclude from pursuing uplifting outdoor activities.
In these cases, virtual reality (VR) technology can be beneficial since it is assumed that
immersion in natural virtual environments (VE) has comparable positive effects as
exposure to nature [6]. This is due to the innate capability of VR to create an illusion of
being in another place, that is to elicit the experience of presence in the virtual world.
Thus, immersion can be utilized to distract people from situations causing acute
emotional distress [17]. It has been shown that confronting users with positive images of
bright colors and high saturation using VR technology positively affects mood,
motivation, and self-efficacy [7]. VR can also be used to elicit specific emotions by the
design of the respective VE (i.e. joy, sadness, anger, anxiety) [8,9]. Moreover, a
reciprocal interaction between presence and affect is assumed: While affective design
1
Corresponding Author: stefan.liszio@uni-due.de

elements in VEs increase the perceived level of presence, higher levels of presence again
influence affect [9,10]. However, what is perceived as relaxing is a matter of individual
preferences [4]. Thus, investigating the relaxing and mood enhancing effects of certain
VEs as well as identifying facilitating factors enables researchers and designers to build
effective VR applications for a wide range of target groups even in distress situations.
2. Method
2.1. Stimulus Material
The Trier Social Stress Test (TSST) [11] is a widely applied protocol for experimental
induction of social stress. Despite its efficacy, the TSST suffers from limited replicability
and comparability as well as a demand for considerable amount of resources [12].
Addressing these drawbacks, we developed a VR version of the original protocol (VR-
TSST). To ascertain efficacy and validity, we conducted a pilot study (N = 47) comparing
the VR-TSST with the real world TSST. We observed an increased salivary cortisol
concentration after exposition to the VR-TSST (from M = 12.7 nmol/l, SD = 7.32 nmol/l
to M = 17.1 nmol/l, SD = 11.9 nmol/l) comparable to the original TSST (from M = 13.1
nmol/l, SD = 10.9 nmol/l to M = 20.2 nmol/l, SD = 15.5 nmol/l). Other physiological
(HRV) and psychological (self-reported anxiety and affect) measures resembled these
findings. Thus, we consider the VR-TSST as a reliable and efficient stress induction
method comparable to the original TSST.
The restorative effect of green environments (e.g. parks, forests) has widely been
researched. Complementing this research, we used a realistic audiovisual VR underwater
simulation (“theBlu”, Wevr, 2016) to induce relaxation. To investigate whether the level
of immersion influences the effect of the VE, we prepared a screen recording as a less
immersive version of the original VR application and presented on a 17 inch desktop
screen. All VR content was displayed on an Oculus Rift CV1. Sounds of the VR-TSST
and the underwater simulation were played on either the built-in headphones of the HMD
or on additional desktop speaker.
2.2. Objective and Subjective Measures
Heart rate variability (HRV) was calculated with the standard deviation of successive
differences (SDSD) method from five-minute intervals. Therefore, heart rate was
recorded using a commercial heart rate monitor and chest belt. A low HRV is associated
with stress and emotional strain, a high HRV indicates relaxation and well-being
respectively. Besides HRV, an increased salivary cortisol concentration is another
indicator of physiological stress and emotional strain. Hence, saliva samples were
collected using cotton swabs and sent to an endocrinological laboratory for analysis. The
subjective experience of emotional strain was assessed with the State-Trait Anxiety
Inventory [13]. Additionally, the Positive and Negative Affect Schedule [14] was used
to identify current mood. Sense of presence was measured using the Igroup Presence
Questionnaire [15]. Moreover, immersion was determined with the corresponding
dimension of the Game Experience Questionnaire [16]. The original phrasing of the
items was slightly adapted to fit the stimulus material (Cronbach’s α = .70).

2.3. Participants and Procedure
62 healthy subjects (58% female) aged 18 to 48 (M = 22.6, SD = 5.36) participated.
Subjects were mostly students at the University of Duisburg-Essen. To avoid demand
characteristics, it was proclaimed that the study investigates the influence of VR on
concentration. Prior to the experiment, all participants were informed about the
possibility to experience emotional strain and gave written consent before filling out a
screening questionnaire to ensure physical and mental health. The experiment started
with a baseline phase (1), that is the collection of a first saliva sample and the fitting of
heart rate monitor and chest belt. At this point, HRV measurement started. Thereafter,
the participants filled out the questionnaires. In the subsequent induction phase (2), all
subjects underwent the VR-TSST procedure (20 min), with the instruction to fulfill all
presented tasks. This phase ended with the announcement of a second part of this
treatment to keep stress at a consistent level, thus the subjects anticipated more
unpleasant tasks. In the post-induction phase (3), subjective measures were again
recorded with the questionnaires. Since cortisol responses to mental stress are delayed, a
second saliva sample was collected 15 minutes after stress induction. In the subsequent
manipulation phase (4), the subjects were randomly assigned either to the VR (N = 22),
the desktop (N = 17) or the control condition (N = 23). The control group was left waiting
without any distraction. In the control group, the experimenter asked the subjects to wait
a few minutes before the experiment continued. In each condition, this phase took seven
minutes. Subjects in VR and desktop group were asked to just watch the simulation. In
the final post-manipulation phase (5), questionnaires were filled out and a third saliva
sample was collected.
3. Results
3.1. Heart Rate Variability and Salivary Cortisol
Data of four subjects was excluded from analysis due to measurement errors.
Requirements for parametric testing were checked (Kolmogorov-Smirnov and Levene’s
test) for all following analyses reported in this paper.
A repeated measures ANOVA revealed a significant difference between the mean
HRV values in the three phases baseline, induction, and manipulation (Figure 1), F(2.51,
138.1) = 3.35, p = .028, ηp2 = 0.06, a significant interaction, F(5.02, 138.1) = 3.62,
p = .004, ηp2 = 0.12, but no significant group difference, F(2, 55) = 0.72, p = .493. To
analyze the influence of the three conditions, we conducted a univariate ANOVA on the
HRV values in the manipulation phase, F(2, 55) = 5.46, p = .007, ηp2 = 0.17.
The post hoc analysis indicated significant differences between VR and desktop group
(p = .019) as well as between VR and control group (p = .020). Desktop and control
group did not differ significantly (p > .999). We calculated the difference Δ of the HRV
levels from induction to manipulation phase and compared the differences between the
three groups (Table 1). After manipulation, the VR group exhibited the lowest stress
level as indicated by HRV compared to desktop and control group.
We observed cortisol concentrations high above the usual average of 10.0 - 14.6
nmol/l for some participants. Thus, we excluded data of 10 participants with cortisol
levels higher than one standard deviation above the sample’s average (> 32.5 nmol/l). A
repeated measures ANOVA showed a significant difference between the measurements,

F(1.21, 59.4) = 10.8, p = .001, ηp2 = 0.18, no significant interaction, F(2.43, 59.4) = 0.29,
p = .789, and no group difference, F(2, 49) = 0.11, p = .900. Post hoc comparison
indicated a significant difference between the measurements baseline and induction (p
= .007) as well as between induction and manipulation (p < .001). No difference was
found between baseline and manipulation (p > .999). Mean cortisol levels were higher
than the baseline (M = 12.8 nmol/l, SD = 6.9 nmol/l) after stress induction (M = 21.3
nmol/l, SD = 18.5 nmol/l) and lower again after the manipulation (M = 14.2 nmol/l, SD
= 11.1 nmol/l), as shown in Figure 2. As for HRV, we calculated the difference Δ of the
cortisol levels from induction to manipulation (Table 1). We found a medium effect size
of ηp2 = 0.7 and descriptive statistics indicated a stronger decrease in the VR and desktop
group as compared to the control group. A univariate ANOVA did not show significant
difference between the three groups, F(2, 49) = 1.72, p = .190.
3.2. Anxiety and Affect
A repeated measures ANOVA indicated a significant difference in the anxiety levels
between the measurements, F(2.59) = 58.82, p < .001, ηp2 = 0.50, a significant interaction,
F(2,59) = 6.19, p < .001, ηp2 = 0.17, but no group difference, F(2,59) = 2.12, p = .129.
The post hoc tests showed a significant difference between baseline (M = 42.52, SD =
8.13) and post-induction (M = 49.81, SD = 10.79, p < .001), as well as between post-
induction and post-manipulation (M = 39.15, SD = 9.98, p = .004). Moreover, we found
a significant difference between baseline and post-manipulation (p < .001), indicating
that the manipulation reduced anxiety to a level lower than the baseline. Comparable
results were obtained for positive affect, F(2,59) = 9.51, p < .001, ηp2 = 0.14 and negative
affect F(2,59) = 27.68, p < .001, ηp2 = 0.32. For closer investigation of the decrease in
anxiety and negative affect after the manipulation and the increase of positive affect, we
conducted a univariate ANOVA. We calculated the difference Δ in anxiety, negative and
positive affect between post-manipulation and post-induction (Table 1). The Δ scores of
anxiety differ significantly between the groups, F(2,59) = 9.46, p < .001 ηp2 = 0.24. Post
hoc analysis revealed a significant difference between VR and desktop group (p = .012)
Table 1. Descriptive statistics M (SD) of objective and subjective data for each condition. Δ scores denote the
difference between (post-)induction and (post-)manipulation measurements.
Virtual Reality
Desktop
Control
Δ HRV (ms)
17.02 (11.82)
-9.45 (-2.34)
-2.61 (-3.08)
Δ Cortisol (nmol/l)
-8.11 (11.71)
-8.18 (9.67)
-5.24 (7.45)
Δ State Anxiety
–16.40 (9.27)
-8.76 (7.96)
-6.57 (6.19)
Δ Positive Affect
3.41 (4.79)
-2.35 (7.09)
-3.04 (5.17)
Δ Negative Affect
–5.90 (6.12)
-5.06 (3.40)
-3.35 (4.69)
Immersion
3.52 (0.67)
2.27 (0.91)
-
Presence
4.64 (0.90)
1.47 (1.66)
-
Figure 1. Mean HRV values for each phase.
■ = VR, = desktop, ◊= control group.
Figure 2. Mean cortisol levels for each phase.
■ = VR, = desktop, ◊= control group.

as well as between VR and control group (p < .001), but no significant difference between
control and desktop group (p > .999). Hence, the VR group experienced the highest
decrease of anxiety compared to desktop and control group.
We did not find significant results for the Δ values of negative affect. However, the
descriptive statistics indicate that the VR group experienced the highest decrease of
negative affect, followed by desktop, and control group. The Δ scores for positive affect,
however, differ significantly, F(2,59) = 8.56, p = .001, ηp2 = 0.23. Post hoc tests indicated
a significant difference between the VR and both the control (p = .001) and the desktop
group (p = .007). Differences between control and desktop group were non-significant
(p > .999). Hence, the VR group exhibited a higher increase of positive affect after the
manipulation than the two other groups.
3.3. Immersion and Presence
A t-test for independent samples revealed a significant difference in perceived immersion,
t(37) = 4.93, p < .001, d = -1.56 between VR group and the desktop group (Table 1).
Furthermore, a significant difference in the general feeling of being present, t(37) = 7.62,
p < .001, d = -2.37, indicates that the VR group perceived higher levels of presence than
the desktop group (Table 1). We observed significant correlations in the post-
manipulation measurement between immersion and anxiety, Pearson’s r(39) = -.53,
p = .001, as well as immersion and positive affect, Pearson’s r(39) = .49, p = .002. A
linear regression analysis indicated immersion as a significant predictor of anxiety,
R2 = .276, β = -.53, t(38) = -3.75, p = .001. Hence, high immersion results in less anxiety.
4. Discussion and Conclusion
Our results prove that the presentation of simulated natural environments is an effective
method to provide relaxation and positive mood in acute stress situations. The reception
of a computer-generated underwater scenario in VR reduces physiological stress, anxiety,
and negative feelings effectively. Significantly higher HRV levels (i.e. less stress) during
the exposition to the VE were measured in the VR group than in the desktop and control
group. Additionally, we observed a medium-sized effect of immersion in the VE on the
cortisol concentration. Since cortisol concentration is determined by a multitude of
factors (e.g. sex, age, daytime, chronic stress), we registered a high variability in our
measurements, but did not find statistically relevant connections to our results.
With respect to the psychological data, the VR group exhibited lower anxiety levels
than the desktop and the control group as well as less negative affect than the control
group. Moreover, perceived immersion impacts anxiety directly. This finding indicates
that VR can distract the user from acute distressing situations more effectively than less
immersive media. The slight decline of HRV and positive affect in the non-VR groups
supports this assumption. Hence, we follow the argumentation that immersion ties up
cognitive resources, which then are not available for negative mental processes [17].
While the majority enjoyed the underwater scenario, a small number of subjects felt
uncomfortable because of the “open water”. Others expected getting shocked or were
afraid of certain animals (e.g. jellyfish). However, we did not find relations between the
experience of unpleasantness of the VE and the observed measures. Moreover, one could
argue that the mere novelty of VR causes fascination which masks the actual effect of
the VE. Hence, we asked the participants whether they have used any VR technology

before: 37 participants reported to have used VR before, 25 had no prior experience.
With prior experience as a group variable, we did not find significant differences between
any of the measures. Thus, the novelty effect as a confounding variable can be omitted.
Our findings highlight the role of immersion and presence as facilitators of relaxing
and mood inducing effects of computer-generated natural VEs. VR technology can
provide relief from emotional strain in acute stress situations and is a viable solution to
enhance well-being of people who cannot benefit from the restorative effects of nature.
Acknowledgments
As part of the project “VR-RLX” (EFRE-0800500), this work was supported by the
European Regional Development Fund (ERDF) 2014-2020.
References
[1] M.P. White, S. Pahl, K. Ashbullby, S. Herbert and M.H. Depledge, Feelings of restoration from recent
nature visits, Journal of Environmental Psychology 35 (2013), 40-51.
[2] S. Kaplan, The restorative benefits of nature: Toward an integrative framework, Journal of Environmental
Psychology 15(3) (1995), 169-182.
[3] R. Berto, The role of nature in coping with psycho-physiological stress: A literature review on
restorativeness, Behavioral sciences (Basel, Switzerland) 4(4) (2014), 394-409.
[4] A.P. Anderson, M.D. Mayer, A.M. Fellows, D.R. Cowan, M.T. Hegel and J.C. Buckey, Relaxation with
Immersive Natural Scenes Presented Using Virtual Reality, Aerospace medicine and human performance
88(6) (2017), 520-526.
[5] K. Tanja-Dijkstra, S. Pahl, M.P. White, J. Andrade, J. May, R.J. Stone, M. Bruce, I. Mills, M. Auvray, R.
Gabe and D.R. Moles, Can virtual nature improve patient experiences and memories of dental treatment?
A study protocol for a randomized controlled trial, Trials 15 (2014).
[6] D. Valtchanov, K.R. Barton and C. Ellard, Restorative Effects of Virtual Nature Settings, Cyberpsychology,
Behavior, and Social Networking 13(5) (2010), 503-512.
[7] R. Herrero, A. García-Palacios, D. Castilla, G. Molinari and C. Botella, Virtual reality for the induction of
positive emotions in the treatment of fibromyalgia: a pilot study over acceptability, satisfaction, and the
effect of virtual reality on mood, Cyberpsychol Behav Soc Netw 17(6) (2014), 379-384.
[8] R.M. Baños, V. Liñao, C. Botella, M. Alcañiz, B. Guerrero and B. Rey, Changing Induced Moods Via
Virtual Reality, in: Persuasive Technology, W.A. IJsselsteijn, Y. de Kort, C. Midden, B. Eggen and E.
van den Hoven, Lecture Notes in Computer Science, Springer, Berlin and New York, 2006, pp. 7-15.
[9] G. Riva, F. Mantovani, C.S. Capideville, A. Preziosa, F. Morganti, D. Villani, A. Gaggioli, C. Botella and
M. Alcañiz, Affective Interactions Using Virtual Reality: The Link between Presence and Emotions,
CyberPsychology & Behavior 10(1) (2007), 45-56.
[10] D. Villani, F. Riva and G. Riva, New technologies for relaxation: The role of presence, International
Journal of Stress Management 14(3) (2007), 260-274.
[11] C. Kirschbaum, K.-M. Pirke and D.H. Hellhammer, The ‘Trier Social Stress Test’: A Tool for
Investigating Psychobiological Stress Responses in a Laboratory Setting, Neuropsychobiology 28 (1993),
76-81.
[12] O. Kelly, K. Matheson, A. Martinez, Z. Merali and H. Anisman, Psychosocial stress evoked by a virtual
audience: Relation to neuroendocrine activity, CyberPsychology & Behavior 10(5) (2007), 655-662.
[13] C.D. Spielberger, R.L. Gorsuch and R.E. Lushene, Manual for the state-trait anxiety inventory,
Consulting Psychologists Press, Palo Alto, CA, 1970.
[14] D. Watson, L.A. Clark and A. Tellegen, Development and validation of brief measures of positive and
negative affect: The PANAS scales, J Pers Soc Psychol 54(6) (1988), 1063-1070.
[15] T.W. Schubert, F. Friedmann and H.T. Regenbrecht, Decomposing the sense of presence: Factor analytic
insights, in: 2nd international workshop on presence, Vol. 1999, 1999.
[16] W.A. IJsselsteijn, Y.A.W. de Kort and K. Poels, The Game Experience Questionnaire: Development of
a self-report measure to assess the psychological impact of digital games. Manuscript., 2013.
[17] H.G. Hoffman, J.N. Doctor, D.R. Patterson, G.J. Carrougher and T.A. Furness III, Virtual reality as an
adjunctive pain control during burn wound care in adolescent patients, Pain 85(1) (2000), 305-309.