The effect of brief exposure to virtual nature on mental wellbeing in adolescents

Adolescence (approximately ages 10–24) is marked by cognitive, emotional, physiological and social transitions and is a period of heightened vulnerability to mental health difficulties, with around half of mental disorders appearing in the teenage years and three-quarters by age 24. Depression is common, recurrent, and increasingly prevalent globally, with substantial disability and economic costs. Despite a treatment-oriented model of care, many in need do not receive treatment, and the efficacy of psychotherapies and antidepressants is often limited, with modest benefits and no clear reductions in population burden over time. This context, compounded by strained services and increased stress in adolescents, motivates a focus on prevention and alternative interventions. Nature-based interventions (NBIs) are a promising, under-researched approach; exposure and connection to nature are associated with better health and wellbeing, and early-life access to green space predicts lower psychiatric risk. Urbanization and indoor lifestyles reduce routine nature contact, suggesting potential utility for accessible, brief, virtual nature exposure. This proof-of-principle study tested whether a brief immersive virtual woodland walk (green and blue space) compared with an urban busy underground train journey would reduce stress and improve wellbeing indices, increase nature connection, and enhance nature-related spirituality in adolescents.

Prior work associates nature exposure with stress reduction and restoration of cognitive and emotional functioning (psycho-evolutionary and biophilia perspectives). Mechanisms include stress recovery, attention restoration, affect regulation, and reductions in rumination; spirituality in nature has been proposed as an additional contributor to wellbeing but is seldom examined. NBIs range from forest bathing and therapies (CBT, mindfulness, compassionate mind training) to brief self-help interventions. Virtual nature can benefit health and wellbeing, though effects may be smaller than real-world nature. Barriers to accessing real nature include urban form, socioeconomic disadvantage, mobility limitations, and symptoms like fatigue and low motivation, creating inequities. Adolescents may be time-poor due to social influences and academic pressures. Evidence reviews indicate mixed but often positive effects of virtual nature on mood, with real nature generally outperforming virtual for mood improvement while negative affect differences are less consistent. Emerging work highlights potential roles of natural sounds (water, birdsong) and specific environments (green, blue, brown spaces) in wellbeing. This study addresses gaps in adolescent-focused virtual nature exposure and explores nature spirituality alongside standard wellbeing outcomes.

Design: Experimental, proof-of-principle, remote online randomized study with 1:1 allocation to virtual nature versus urban comparison condition using Gorilla Experiment Builder. Recruitment: Opportunity sampling via social media, campus posters, and word of mouth; consent obtained via QR-linked materials. Procedure: Baseline measures completed, then automatic randomization to a 6-minute video condition (nature or urban), followed by an attention check and post-experiment self-report battery; debrief and signposting provided; prize draw entries offered. Participants: N=76 adolescents and young adults (mean age 20.42, SD 1.15, range 18–25), 61 female (85.53%), 11 male (14.47%); 65 white (85.53%), 11 other ethnicities. Conditions: Nature video (6 min) POV walk combining woodland path (green space) and riverside (blue space) with natural sounds (birdsong, leaves rustling, water). Urban video (6 min) London Underground rush-hour experience featuring crowded platform and train, with urban noise (engines, announcements, chatter), designed to emulate typical urban stressors. Measures: SWEMWBS (wellbeing, 7 items), I-PANAS-SF (positive/negative affect, 10 items), BSRI (state rumination, 8-item VAS), Stress VAS (single item), PSS-4 (perceived stress past month, 4 items), Nature Connection Index (NCI, 6 items), Nature spirituality (6-item state measure adapted from Ecospirituality Scale plus bespoke items), and bespoke single-item VAS for relaxation, positive mood, and attention. Feasibility/acceptability: Realism, perceived length, and recommendation questions; attention check after video. Randomisation check: Baseline differences tested; none observed. Sample size and power: Planned N=106 for 90% power (partial η²=0.11); achieved N=76 reduced power to 77%. Statistical analysis: Conducted in STATA 17. Baseline comparisons via ANOVA and chi-square. Primary analyses via ANCOVA with post scores as outcomes and baseline scores as covariates to assess condition effects (nature=1, urban=2). Bootstrapped marginal mean differences (1000 resamples; 95% CIs) reported. Significant condition effects followed up with within-condition repeated measures ANOVA. Visualization via grouped violin/box plots in R.

Randomization produced no baseline differences (Table 1). Stress and relaxation: Significant Condition effect on Stress VAS over time (F(1,73)=4.81, p=0.03, η²=0.06; marginal mean difference=8.63; 95% CI 1.19 to 16.07). Stress decreased significantly in the nature condition (F(1,36)=31.02, p<0.001, η²=0.87) but not in urban (F(1,36)=2.96, p=0.09, η²=0.07). Relaxation post-exposure was higher in nature (F(1,74)=35.10, p<0.001, η²=0.32; marginal mean difference=−28.39; 95% CI −37.60 to −19.17). Affect and mood: Negative affect showed a Condition effect (F(1,73)=11.28, p<0.05, η²=0.13; marginal mean difference=1.95; 95% CI 0.85 to 3.06); it decreased in nature (F(1,36)=26.69, p<0.001, η²=0.43) and did not change in urban (F(1,38)=0.06, p=0.81). Positive affect Condition effect (F(1,73)=16.20, p<0.001, η²=0.18; marginal mean difference=−3.02; 95% CI −4.44 to −1.60) reflected a significant decrease in urban (F(1,38)=36.36, p<0.001, η²=0.49) and no change in nature (F(1,36)=0.51, p=0.48). Positive mood VAS was higher post in nature (F(1,74)=25.08, p<0.001, η²=0.25; marginal mean difference=−23.41; 95% CI −32.31 to −14.50). Nature connection and spirituality: NCI Condition effect (F(1,73)=5.37, p<0.05, η²=0.07; marginal mean difference=−1.41; 95% CI −2.57 to −0.24); within-condition changes were nonsignificant (nature F(1,36)=3.09, p=0.09; urban F(1,38)=2.65, p=0.11). Nature spirituality Condition effect (F(1,73)=8.43, p<0.01, η²=0.10; marginal mean difference=−1.82; 95% CI −3.05 to −0.50); it increased in nature (F(1,36)=15.76, p<0.001, η²=0.30) and did not change in urban (F(1,38)=0.16, p=0.69). Rumination: No significant difference between conditions on BSRI (F(1,73)=1.96, p=0.17, η²=0.03). Attention: Post-exposure self-reported attention was higher in nature (F(1,74)=7.03, p<0.05, η²=0.09; marginal mean difference=−12.66; 95% CI −21.62 to −3.70). Acceptability/feasibility: Attention check was passed by most participants (nature 89.19%, urban 84.62%; χ²=0.35, p=0.56). Realism was rated higher in urban (χ²=3.95, p=0.05). Video length was often perceived as a bit long or too long with no between-condition difference (χ²=1.52, p=0.47). Recommendation: 72.22% in nature vs 43.59% in urban would recommend to someone with anxiety/depression (χ²=6.27, p=0.012). Descriptive pre/post means are provided (Table 2).

Findings support the hypotheses that brief immersive virtual nature exposure reduces stress and improves indices of wellbeing (relaxation, mood, negative affect) and increases nature connection relative to an urban comparison; the novel hypothesis of increasing nature-related spirituality was also supported. Positive affect decreased in the urban condition but was buffered in nature, indicating a potential protective effect of virtual nature against urban-induced declines in positive affect. The null rumination result may reflect reduced potency of virtual exposure compared to real-world NBIs and reduced statistical power due to smaller sample size. Attention self-reports improved after nature exposure, though objective attention findings in the literature are mixed. The intervention proved largely acceptable and feasible, with most participants passing attention checks and many recommending the nature video. Virtual nature should be viewed as an adjunct rather than a replacement for real-world nature, offering accessible support for adolescents who are time-poor or face barriers to outdoor exposure, and potential use as self-help, in therapy, and preparatory steps toward more intensive NBIs. The nature spirituality increase highlights a potentially important but underrepresented dimension of wellbeing; integrating values and spirituality considerations may enhance NBI suitability and therapeutic alignment.

A single brief (6-minute) immersive virtual nature exposure produced beneficial short-term effects in adolescents on stress, relaxation, mood, negative affect, attention, nature connection, and nature-related spirituality, and was generally acceptable and feasible. Virtual nature may serve as a scalable, accessible tool for universal or targeted prevention, adjunctive therapy, and preparatory exposure where access to real nature is limited or challenging. Future research should: assess dose-response and repeated exposures; augment virtual interventions with active components (mindfulness, breathing, reflective recall); allow preference-based selection of nature types (green/blue/brown spaces); use more immersive modalities (e.g., VR); include objective biomarkers and cognitive measures; conduct follow-ups to evaluate sustained and preventive effects; diversify samples; and directly compare virtual with real-world NBIs and test virtual nature within multi-component intervention packages.

Planned sample size was not reached (76 vs 106), reducing statistical power. Rumination effects may have been underpowered and virtual exposure may be less potent than real-world nature. No objective physiological (e.g., cortisol, HRV) or cognitive measures were collected; outcomes were self-reported. No follow-up assessments were conducted, preventing evaluation of medium-term symptom change or prevention effects. The sample was predominantly female and white, limiting generalizability. The intervention cannot identify specific beneficial environmental features or sensory modalities; real-world nature is multisensory and may include non-sensory pathways (e.g., phytoncides). Perceived realism was lower for the nature video than urban, and many participants felt the video was a bit long or too long. Virtual nature is not intended as a replacement for real nature and findings from this proof-of-principle study should be extended with more comprehensive designs.