Electroencephalography-based psychological and physiological effects of winter virtual forest trail landscapes on youth

Abstract

In the context of escalating urbanization and modernization, urban residents are facing a progressive rise in stress levels, particularly during winter when severe cold and limited daylight hours intensify psychological strain and physical fatigue. Despite the acknowledged health benefits of brief exposure to natural settings, research on the restorative effects of virtual winter forest settings remains limited. This study undertook the creation of immersive winter forest trail landscapes using virtual technology, generating six distinct audiovisual configurations by manipulating variables such as Evergreen Tree Density (ETD) and Event Ambient Sound (EAS). A cohort of 132 participants (N = 22) engaged in a stress-induction experiment involving a 5-minute virtual landscape exploration within indoor settings. Psychological metrics were assessed through the Profile of Mood States (POMS) and Perceived Restored Soundscape Scale (PRSS), while changes in brain alpha and beta waves and neuroemotional indicators were monitored via Emotiv EPOC X during the participants’ virtual experience. The impact of the virtual winter forest trail landscape on participants’ psychological and physiological perceptions was analyzed. Findings revealed that: (1) the virtual winter forest trail contributed to heightened positive emotions (p = 0.001); (2) diverse audiovisual configurations positively influenced audiovisual nerve relaxation, as evidenced by EEG data, albeit with varying degrees of efficacy; (3) winter forest trail environments characterized by high green visibility significantly facilitate the physical and mental recovery of visitors; (4) multi-person activity sounds outperformed single-player audio in terms of restorative benefits, while companionship enhances the healing process; and (5) interactions involving multiple people and sound production significantly enhanced recovery benefits (p = 0.000) in a forested trail environment dense with evergreens during winter. In contrast, the lowest recovery benefits were observed when individuals strolled alone. This research offers a theoretical foundation for the advancement and implementation of winter forest landscape therapy and serves as a scholarly reference for leveraging snow and ice tourism resources in forest environments.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-91345-w.

Introduction

The impact of winter exposure to urban forests on physical and mental health

With the continuous advancement of urbanization and modernization, the accelerated pace of life and increased work pressure have significantly reduced the time and opportunities for urban residents to contact the natural environment, which has led to increasingly serious psychological and physiological health problems^1,2. Currently, there is a growing body of research on forest therapy within the international scholarly community^3–5. It has been established that short-term exposure to forest environments can promote physical and mental health of young people^5,6. Cold weather and shorter daylight hours in winter make people more prone to depression and fatigue^7–9. Consequently, winter-related mental health issues have increasingly emerged as a significant area of research. Studies have shown that exposure to natural environments in winter can significantly improve people’s psychological state and enhance well-being^10,11, and that urban forests can help to heal mental illnesses, enhance physical movement, as well as promote physiological relaxation^12–14, and markedly enhance concentration and cognitive function¹⁵. Although winter forest landscapes appear sluggish due to lower temperatures and withered trees, in contrast to the lush vegetation of summer¹⁶, the silver lining of winter provides a unique experience of serenity and helps people to withdraw from the busyness of daily life to achieve deep relaxation and restoration^17,18.

Perceived restorative benefits of virtual forest trail environments

As winter snow and ice tourism programs have gained increasing popularity among the public^19,20, there has been a significant rise in participation in snow and ice activities, leading to the burgeoning appeal of winter forest tours. However, due to geographical constraints, seasonal variations, and physical limitations, not all individuals have the opportunity to access natural forests adorned with snow and ice for physical and psychological rejuvenation. For recreationists who are unable to personally experience snow and ice activities, as well as for those who seek healing through winter landscapes^21,22, anti-seasonal healing based on virtual snow and ice landscapes has become an effective way to fulfill that inherent need. As a result, virtual reality (VR) technology has gained increasing popularity due to its immersive, interactive, and experiential characteristics^23,24. Immersive virtual environments (IVE) offer participants a heightened sense of realism, immersion, and presence compared to traditional 2D images and videos²⁵. Several studies have demonstrated that IVEs can elicit perceptual restoration effects comparable to those experienced in real environments^26–28. Through virtual technology, individuals can engage with natural landscapes and the therapeutic benefits of forests at any time and in any location. This is particularly advantageous during winter months when outdoor activities are restricted; virtual forest trail landscapes serve as an alternative means of healing²⁹.

The impact of audiovisual interaction on physiological and psychological well-being in forest trails

Furthermore, the interplay among various sensory modalities may significantly affect the experience of forest environments. The robust connection between visual and auditory perception is regarded as a critical foundation for landscape assessment^30,31. Research indicates that audiovisual interactions within forest environments can capture a broader range of participants’ attention compared to silent settings³².

Currently, investigations into virtual forest perception predominantly emphasize visual and auditory dimensions, with researchers examining how visual and auditory preferences affect psychophysiological responses within virtual forest landscapes³³. In the realm of visual perception, studies have concentrated on specific landscape attributes, such as growth density, plant species, and vegetation height^34–36. Among these characteristics, forest density is particularly crucial, as it is a key component of the forest environment and significantly affects its restorative effects³⁷. Despite this, research on how evergreen tree density influences human restorative perception remains relatively scarce. Therefore, we have chosen evergreen tree density as one of the variables in this study. Regarding soundscape perception, existing studies primarily focus on natural and artificial sounds, revealing that both categories positively influence human emotions³⁸. Furthermore, human voices within forest settings can enhance visitors’ feelings of security³⁹, while the companionship of multiple individuals can mitigate feelings of loneliness⁴⁰. Most current research on anthropogenic sounds in forests centers on footsteps, conversations, and children’s play⁴¹. Research on natural sounds includes bird songs, flowing water, and sounds of vegetation caused by the wind⁴². In forest environments, ambient sound is composed of both natural and human-made sounds, with different types of activities generating specific ambient sound effects. However, existing studies lack in-depth exploration of the ambient sounds produced by specific activities. A common approach for assessing the impact of audiovisual interaction on landscape perception and restorative benefits is to combine visual and auditory elements in a between-group design⁴³. Consequently, it is essential to investigate the impact of auditory interactions between evergreen tree density and various activity modes on individuals’ physical and mental well-being across diverse forest trail environments.

Assessing the perceived physical and mental restorative benefits of forest landscapes

Currently, research on the perceived restorative benefits of forest landscapes predominantly emphasizes psychological and physiological dimensions. Psychological indicators are typically evaluated using standardized scales^44,45, with the Profile of Mood States (POMS) being one of the most widely employed instruments in related studies. Physiological indicators are often derived from the assessment of neurological and cardiovascular systems. Brainwave technology, characterized by its millisecond temporal resolution, effectively captures rapid fluctuations in brain activity in real time, making it increasingly favored by researchers for neurological evaluations³¹. Electroencephalography (EEG) provides a comprehensive representation of the electrophysiological activities of cortical and scalp nerve cells, revealing the brain’s activity state by recording electrical impulses from the cerebral cortex⁴⁶. Monitoring brain activity serves as the most direct and effective method for assessing subjects’ environmental perceptions and physiological responses. Specifically, alpha (α) and beta (β) waves in EEG are recognized as the most pertinent indicators of environmental reflection variability⁴⁷. In summary, this study investigates the perceived restorative benefits of forest landscapes from both psychological and physiological perspectives. The psychological assessment employs the POMS scale to quantify psychological states, while the physiological evaluation reflects subjects’ relaxation and tension through the analysis of α-wave, β-wave, and β/α ratios. Neuroemotional indicators are also utilized for supplementary validation.

The objective of this study is to examine the effects of the interaction between the density of evergreen trees in various forest trail environments and the sound produced by different activity modes on human physical and mental health under simulated winter snow-covered conditions. To effectively control experimental variables and mitigate the influence of confounding factors, we collected live images and audio data, utilizing virtual technology to generate immersive videos of winter forest trail environments as experimental stimuli. We assessed perceptual recovery through various audio-visual combinations, employing psychological indicators (changes in POMS scores pre- and post-test, PRSS scores) and physiological measures (EEG α-wave and β-wave power spectral density values, arousal index β/α, and neuro-emotional indicators). This research aims to provide scientific guidance for the design and implementation of forest therapy, optimizing its therapeutic effects and enabling broader access to this emerging form of natural healing. Building upon previous studies, the following research questions are proposed:

(RQ1): How do different virtual forest trail environments influence stress recovery under snow-covered conditions in winter?

(RQ2): What is the impact of evergreen tree density on the perceived restorative benefits of virtual snow-covered forest trail environments?

(RQ3): How do ambient sounds from different activities after snowfall affect perceived restorative benefits?

Materials and methods

Study sites

Jingyuetan National Forest Park, a national 5 A-level tourist attraction, is located in the Jingyue Economic and Technological Development Zone, Changchun City, Jilin Province. The park covers an area of 96.38 square kilometers and has a forest coverage rate exceeding 96%⁴⁸. It is situated in a temperate semi-moist monsoon climate zone, with maximum and minimum temperatures of 28.32 °C and − 22.48 °C, respectively. The average winter temperature ranges from − 10 °C to − 18 °C. The snowy season extends for more than 180 days, with an average snow depth of 21.6 cm⁴⁹. Figure 1c illustrates the snow depth and the number of snowfall days in Changchun City from 2019 to 2023. The substantial average snowfall and high frequency of snowfall days make Jingyuetan National Forest Park an ideal site for studying winter landscapes.

Fig. 1.

Site status. (a) Site location and location of sampling points, (b) schematic of plant density, (c) snow depth and number of snowfall days in Changchun, 2019–2023.

Source: Plotted by the authors using ArcMap 10.8.1 (https://www.esri.com/) and Adobe Photoshop 2024 (https://www.adobe.com/products/photoshop.html).

Experimental materials

View data acquisition

In January 2024, a 360° panoramic video (FHD: 1920 × 1080, 60 fps) was captured in the field using a GoPro Max camera. The specific locations of the sampling sites are depicted in Fig. 1a. Three experimental sample sites were identified, each exhibiting varying densities of evergreen trees (Picea koraiensis Nakai) within larch (Larix olgensis A. Henry) forests after thorough investigation. To accurately determine the tree density at the three aforementioned locations, aerial video footage was recorded using a DJI Air 2 S drone. This footage, in MP4 format with a resolution of 3840 × 2160 and a frame rate of 30 fps, was utilized to analyze and calculate the tree densities. All photos were taken on clear days during the same time period (between 10 a.m. and 2 p.m.) to minimize lighting differences due to time variations. Factors such as photorealism, composition, orientation, and perspective were also considered during the shooting. The planting schematic and calculated density results are presented in Fig. 1b.

Soundscape data acquisition

The sound environment was recorded in the field using the TASCAM Portacapture X6 recorder in outdoor mode, ensuring recordings were made during the same time period. The recordings took place in early January 2024, with a format of 44.1 kHz 16-bit WAV (stereo), and all equipment was calibrated prior to recording. To replicate the live sound as perceived by the human ear during human activity, the recording equipment was mounted on a tripod fixed to the wooden trestle at the three sample sites. The height above the ground was 1.60 m, and the recording duration at each site was 5 min. The recorded environmental sounds primarily included bird chirping, wind, and traffic noise, while the activity ambient sounds comprised conversations and footsteps. To minimize ambient noise interference and ensure the purity and authenticity of the recorded human activity sounds, we selected the quieter of the three occasions and SITE ③, which is farther from the main road, for recording. We recorded the wind speed during each audio capture and paused the recording whenever the wind speed exceeded a set threshold, ensuring that the collected sound data had a high signal-to-noise ratio.

Once the audio was recorded, it was imported into Adobe Audition 2022 for noise reduction. The ambient sound was reduced using the intelligent noise reduction function, to ensure the clarity and consistency of the audio data. The noise reduction level was set to 55%. Due to the complexity of human activity noise, the noise sampling noise reduction function was employed. After capturing the noise samples, the noise reduction extent was adjusted to 90%, the noise reduction amplitude was set to 16dB, the spectral attenuation rate was 60%, and the smoothing extent was 20%. The spectral comparison of the audio before and after noise reduction is shown in Fig. 2a.

Fig. 2.

(a) Roaming video production process, spectrum comparison before and after audio noise reduction and comparison between virtual and real maps, (b) Landscape components extraction and calculation method.

Model rendering and video production

SketchUp Pro 2022 was utilized for virtual modeling, based on the aerial video and the 360° panoramic video, combined with the field density data of larch and evergreen trees. According to field measurements, the width of the wooden trestle was set to 2 m, and the average walking speed was set to 1.06 m per second. Research indicates that, compared to shorter or longer exposure times, participants experience better stress recovery after being exposed to virtual natural landscapes for 5 min. Therefore, the final rendered video was set to 5 min⁵⁰, and the total length of the modeled wooden trestle was 320 m.

The model was imported into Lumion 12.5, and to maximize the realism of the visual stimulation of the video samples, the sun altitude was set to the environment at 10:00 a.m. on January 10. Rendering parameters such as precipitation, reflection, shadow, floodlight, and fog were subsequently adjusted to create a winter environment. The video rendering quality was set to 16x antialiasing, and the video format was MP4, with a resolution of FHD: 1920 × 1080 and a frame rate of 60fps. A frame from the rendered video was compared with a panoramic image from the field shooting, as shown in Fig. 2a.

The noise-canceled audio and the rendered video were imported into Adobe Premiere Pro 2022 for audio-video integration. The exported video was encoded in HEVC (H.265) format with a resolution of 1920 × 1080, a frame rate of 60 fps, and a duration of 5 min, and saved in MP4 format.

Landscape characterization descriptive statistics

To quantify the landscape characteristics of different forested timber trail environments and compare the differences between them, we employed a computational approach similar to that of H. Nordh⁵¹, as illustrated in Fig. 2b. Key node frames were intercepted from the three produced virtual trail landscape roaming videos to ensure consistent image viewpoints and rendering parameters. The landscape components were categorized into five variables: sky visibility, ground snow cover visibility, visibility of evergreen, percentage of man-made facilities, and other (including long white larch and other dead wood weeds, etc.).

The intercepted images were all of 1080 × 1920 resolution, and each type of landscape element in the image was selected and painted with different colors to distinguish each element using the grid tool of Adobe Photoshop 2023. A pattern with a grid spacing of 20 pixels was laid on each photo, totaling 5184 (54 × 96) grid cells. Each grid cell for a specific variable was outlined and counted. A comparative analysis of the landscape component occupancy calculations is shown in Table 1 below.

Table 1.

Comparison of occupancy rates of different landscape components.

Landscape constituents	Percentage of man-made facilities	Ground snow cover visibility	Sky visibility	Visibility of evergreen	Other
Real photo	7.35 ± 0.51%	20.37 ± 0.70%	8.24 ± 1.21%	16.28 ± 0.78%	47.76 ± 1.78%
Non-evergreen tree	6.66 ± 0.00%	22.26 ± 0.82%	6.59 ± 0.23%	0.00 ± 0.00%	64.49 ± 0.82%
Low-density evergreens	6.66 ± 0.00%	18.73 ± 0.41%	5.29 ± 0.23%	8.14 ± 0.47%	61.18 ± 0.55%
High-density evergreens	6.66 ± 0.00%	17.96 ± 0.57%	5.05 ± 0.59%	17.89 ± 0.89%	52.45 ± 1.10%

EEG data acquisition and processing

Alpha waves (8–13 Hz) typically manifest during states of relaxation, such as sitting with closed eyes or mild attentiveness, indicating a wakeful yet relaxed brain state⁵². These waves diminish or vanish in response to unfamiliar sounds or heightened tension⁵³. Conversely, beta waves (13–30 Hz) signify an active brain state associated with logical and analytical thought processes, emerging when individuals concentrate or attempt to solve problems, often leading to increased tension and anxiety⁵⁴. The arousal index, represented by the β/α ratio, is commonly utilized to assess stress levels, with the ratio increasing under stress and decreasing during relaxation⁵⁵. Brainwave data were collected using the Emotiv EPOC X-14 channel, and its neural emotions were measured, a common indicator in neuroassessment research⁵⁶. The schematic diagram of the EPOC X, along with the corresponding brain regions and electrode positions, is shown in Fig. 7 below. The 14 electrodes were distributed across the headset as follows: frontal lobe regions (F3, F4, F7, F8, AF3, AF4, FC5, FC6), parietal lobe regions (P7, P8), temporal lobe regions (T7, T8), and occipital lobe regions (O1, O2). These electrodes monitored EEG activities and output performance index data, which included six neuroemotional indices (excitation, engagement, relaxation, interest, stress, and focus) evaluated by Emotiv’s performance index algorithm. Several studies have validated the scientific reliability of these neuroemotional performance metrics^57,58. The purpose of this experiment is to investigate the effect of virtual roaming video on human visual and auditory stimuli. Consequently, the PSD values of α-waves and β-waves, along with the β/α-wave ratio in the temporal and occipital lobe regions, were selected as the research indices corresponding to these functions.

Fig. 7.

(a) Schematic of EEG topography, (b) Schematic of brain regions, electrode locations and corresponding functional maps.

Topological features of EEG demonstrate robustness to re-referencing and preprocessing⁵⁹. The analysis of EEG data is divided into four stages: preprocessing, feature extraction, post-processing, and result analysis⁶⁰. The raw EDF data exported by Emotiv PRO were preprocessed using the EEGLAB toolbox (version 14.0.0) in MATLAB R2021b. This process involved localizing electrode positions, resampling, re-referencing, and filtering EEG waves within the frequency range of 8–30 Hz using the filter tool. Subsequently, recorded events were segmented, and common artifacts were manually removed. Independent component analysis (ICA) was then applied to eliminate artifacts such as eye movements, muscle contractions, and noise components, thereby enhancing the reliability of the preprocessed EEG data.

Power spectral density (PSD) quantifies the contribution of each frequency component to the overall intensity of a signal segment, serving as a measure of brain activity intensity across various frequencies^61,62. Average power spectral density (APSD) represents the mean signal power over frequency, characterizing the energy distribution of brain activity across different frequency ranges within a specified time period, thereby providing critical insights into the level of brain activity⁶³. The preprocessed EEG data underwent a fast Fourier transform (FFT) to convert the time-domain signal into a frequency-domain representation⁶⁴. The average power spectral density within the frequency ranges of α- and β-waves was calculated and normalized to facilitate comparability of EEG signals across different conditions.

Measurement and calculation of psychological indicators

The Profile of Mood States (POMS) is a widely utilized psychological assessment tool for measuring emotional states. In this study, we employed the POMS model developed by Grove and Prapavessis⁶⁵, which was subsequently simplified by Zhu for the Chinese population⁶⁶. The POMS comprises 40 distinct feelings, with the frequency of their occurrence rated on a 5-point Likert scale (0 = almost none, 1 = somewhat, 2 = moderately, 3 = quite a bit, 4 = very strongly). The scores for these 40 feelings were utilized to calculate seven mood scores: tension (T), anger (A), fatigue (F), depression (D), panic (P), energy (E), and self-esteem (S). The Total Mood Disturbance (TMD) = Negative Mood (T + A + F + D + P) − Positive Mood (E + S) + 100. Higher scores for each mood indicate more pronounced mood changes⁶⁷.

The Perceived Restored Soundscape Scale (PRSS) was utilized to evaluate the restorative effects of soundscapes on psychological well-being⁶⁸. This scale includes questions pertaining to four dimensions: fascination, being away, compatibility, and extent, with responses measured on a 5-point Likert scale to assess the level of agreement with each statement.

Participants

A total of 132 participants were recruited from Jilin Agricultural University, the gender ratio was nearly equal, aged between 18 and 42 years (M = 22.55, SD = 1.93). The demographic statistics are shown in Table 2. The sample size was determined using G*Power 3.1 (α = 0.05, power = 0.8), and the number of participants satisfied the required sample size criteria. All participants adhered to EEG experimental protocols, reported no history of smoking or psychiatric disorders, exhibited normal auditory and visual capabilities, and were free from color vision deficiencies.

Table 2.

Basic demographic characteristics statistics.

Categorization	Attributes	Numeric
Gender	Male	63
	Female	69
Age	18–20 years	21
	20–30 years old	97
	30–40 years old	8
	40–44 years old	6
Educational background	Undergraduate	78
	Graduate students and above	54

Experimental sites

To minimize external interference, the experiment was conducted in the multimedia laboratory of Jilin Agricultural University (7 m long, 5 m wide, and 4 m high). The laboratory was equipped with multimedia equipment featuring a resolution of 2048 × 1080 and a refresh rate of 120 Hz, along with blackout curtains on the north-facing window. The window remained open to simulate outdoor temperature and ventilation, while the curtains were drawn to reduce visual distractions, ensuring the room was free from extraneous noise. Participants wore headphones, selected a relaxed and comfortable seating position, and viewed the virtual roaming video under low-brightness LED diffuse lighting. The average daytime temperature in January 2024 was − 11 °C, with temperatures ranging from − 5 to 2 ℃ and humidity levels between 52% and 69%.

Experimental group design

Based on field investigations, three primary types of trail landscape environments were identified in Jingyuetan National Forest Park. The three target sites, SITE ①, SITE ②, and SITE ③, depicted in Fig. 1b, serve as typical representatives of these environments, characterized by varying densities of evergreen trees and a consistent planting density of larch across all sites. Evergreen Tree Density (ETD) is the first variable, categorized into three levels: no density (N), low density (L), and high density (H). Additionally, another variable, Event Ambient Sound (EAS), refers to the sounds generated by activities on the trail landscape, which can be classified into two types: single-person travel (S) and multiple-people travel (M). Multiple individuals engage in conversation while traveling together, in contrast to a solitary person walking alone. The combinations of these two variables resulted in six experimental groups: N-S, N-M, L-S, L-M, H-S, and H-M. The design of the experimental groups is illustrated in Fig. 3.

Fig. 3.

Schematic diagram of the experimental group design.

Experimental flow design

The experiment was conducted during the second half of January 2024, from 8:00 to 11:30 a.m. and 2:00 to 6:30 p.m. each day for 1 week. Participants were required to arrive 10 min in advance. Prior to the experiment, all participants were briefed on the procedure and precautions, and information was collected while informed consent forms were signed. They were assured that their personal information would be utilized exclusively for academic research and would remain confidential. The study received approval from the Ethics Committee of the College of Forestry and Grassland Science, Jilin Agricultural University, confirming that all experiments were conducted in accordance with relevant guidelines and regulations. All participants were randomly assigned to one of six groups.

Prior to the experiment, participants remained still for 10 min, then donned the Emotiv EPOC X wireless EEG headset and acclimatized for 1 min. Research indicates that watching stress-inducing videos and virtual reality (VR) technology effectively elicit stress and anxiety responses^69–72. After confirming proper equipment function, VR technology was used to immerse participants in a 5-min scenario designed to evoke fear and anxiety, featuring a disaster movie, a realistic emergency video, and a horror film in that order. To minimize content influence on stress responses, all participants viewed the same material. Baseline EEG data were recorded for 1 min post-stress induction. Subsequently, the POMS (pre-test) and PRSS (pre-test) questionnaires were administered to assess pre-test psychological baseline status.

Participants wore headphones, adjusted to the appropriate volume, and engaged in a 5-min virtual winter forest trail landscape experience while real-time EEG data were recorded. After the video playback, the headset and EEG device were removed, and POMS (post-test) and PRSS (post-test) questionnaires were administered to assess psychological changes. At the conclusion of the experiment, the collected EEG and questionnaire data were promptly recorded, stored, and backed up to ensure security and integrity. The experimental flowchart is presented in Fig. 4 below.

Fig. 4.

Experiment flowchart (a) Experiment in progress, (b) Interface schematic, (c) Experiment site, (d) Staff division, (e) Experiment flow.

Statistical analysis

Physiological and psychological data from 132 participants were statistically analyzed in this study. ANOVA was employed to explore the interaction between two experimental variables, while t-tests were used to analyze changes in emotional states on the POMS scale, EEG activity before and after the experiment (during the last minute), and emotional responses measured before and after the experiment. All statistical analyses were performed using IBM SPSS Statistics 27.0 (IBM Corporation, Armonk, NY, USA). The significance level was set at α = 0.05, with p < 0.05 considered statistically significant. The results were visualized using OriginPro2024 (OriginLab Corporation, Northampton, MA, USA). Additionally, Cohen’s d (standardized mean difference) was used as an effect size metric to quantify the mean differences between pre- and post-test measurements.

Analysis of results

Analysis of the results of psychological perception of landscape and soundscape

Psychologically restorative outcomes

The differences in mood changes, as measured by the POMS scale, were analyzed using paired samples t-tests. The POMS results are presented in Fig. 5. The findings indicated that, following exposure to various audiovisual combinations while maintaining a constant soundscape for each data group (N-S vs. L-S, H-S vs. N-M, and L-M vs. H-M), significant alterations were observed across all seven mood dimensions. This suggests that, with the soundscape held constant, distinct landscape features exert a measurable influence on mood states. Furthermore, comparisons among the groups N-S vs. N-M, L-S vs. L-M, and H-S vs. H-M, where the soundscape parameter remained unchanged, revealed varying extents of change in pre- and post-test scores across the seven emotional dimensions. These significant differences underscore the impact of diverse pedestrian activity modes on emotional states, even when the soundscape is held constant.

Fig. 5.

POMS results. (a) Tension, (b) Anger, (c) Fatigue, (d) Depression, (e) Panic, (f) Energy, (g) Self-esteem, (h) TMD.

One-way ANOVA was used to compare the differences in perceived restorative changes across different groups under audiovisual interaction conditions. The results are shown in Table 3. indicated that the changes in “anger” were not significant for most groups, though the L-S group exhibited the greatest decrease in “anger” (p = 0.059), which was not statistically significant. The “energy” (p = 0.001) showed the most significant improvement. The N-M group significantly reduced “tension” (p = 0.008) and alleviated “panic” (p = 0.272), though the latter result was not significant. The N-S group showed a significant increase in the “self-esteem” (p = 0.001) emotional indicator. The H-S group showed a more significant reduction in “fatigue” (p = 0.011). The H-M group was notably more effective in alleviating “depression” (p = 0.012) compared to the other groups. Comparisons of TMD values revealed that the H-M group had significantly higher TMD values (p = 0.000) than the other groups, indicating the most substantial improvement in overall emotional state before and after the experiment. In other words, the interaction of high-density evergreen trees with multiple people and activity sounds had the strongest positive impact on emotional regulation in participants.

Table 3.

One-Way ANOVA of POMS.

Indicator	Group	Mean	SD	ANOVA		Post-hoc comparison
				F	Sig
ΔTension	N-S	− 3.455	4.469	0.357	0.008**	H-S > H-M > L-S > N-S > L-M > N-M
	N-M	− 4.136	2.569
	L-S	− 3.046	3.015
	L-M	− 3.500	3.635
	H-S	− 2.818	5.491
	H-M	− 2.955	2.836
ΔFatigue	N-S	− 2.591	4.148	1.829	0.011**	L-S > H-M > N-S > L-M > N-M > H-S
	N-M	− 3.455	2.154
	L-S	− 1.364	3.048
	L-M	− 3.227	2.927
	H-S	− 3.818	3.375
	H-M	− 1.909	3.753
ΔAnger	N-S	− 1.500	3.635	0.742	0.059	H-M > N-S > L-M > H-S > N-M > L-S
	N-M	− 2.546	2.857
	L-S	− 3.000	2.563
	L-M	− 1.546	3.555
	H-S	− 1.727	4.548
	H-M	− 1.455	3.751
ΔDepression	N-S	− 3.000	2.182	3.072	0.012*	L-S > N-S > N-M > L-M > H-S > H-M
	N-M	− 3.273	3.058
	L-S	− 2.318	2.147
	L-M	− 3.773	3.054
	H-S	− 4.364	5.803
	H-M	− 6.227	4.287
ΔEnergy	N-S	9.636	3.947	0.575	0.001**	L-S > N-S > H-M > N-M > L-M > H-S
	N-M	8.682	4.561
	L-S	9.773	5.163
	L-M	8.682	4.874
	H-S	7.500	5.180
	H-M	8.818	6.329
ΔPanic	N-S	− 2.000	2.673	1.290	0.272	N-S > L-M > L-S > H-S > H-M > N-M
	N-M	− 4.227	2.506
	L-S	− 2.864	2.416
	L-M	− 2.818	2.575
	H-S	− 3.273	4.289
	H-M	− 3.546	3.725
ΔSelf-esteem	N-S	6.773	3.408	0.912	0.001**	N-S > H-M > L-S > L-M > H-S > N-M
	N-M	4.909	3.544
	L-S	5.864	3.167
	L-M	5.727	2.931
	H-S	5.727	2.931
	H-M	6.500	3.349
ΔTMD	N-S	− 28.955	14.660	0.215	0.000***	H-S > L-S > N-S > L-M > N-M > H-M
	N-M	− 31.227	10.076
	L-S	− 28.227	12.701
	L-M	− 29.273	16.216
	H-S	− 27.909	20.853
	H-M	− 31.409	13.710

a: * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001.

b: N-S, single travel in non-evergreen environments; L-S, single travel in low-density evergreen environments; H-S, single travel in high-density evergreen environments; N-M, multi-person in non-evergreen environments; L-M, multi-person travel in low-density evergreen environments; H-M, multi-person travel in high-density evergreen environments.

Soundscape restorability results

A one-way ANOVA was used to compare the differences across the four dimensions of the PRSS scale in different groups. The results are shown in Table 4. In the fascination dimension, the soundscape environment of the H-M group significantly captured participants’ interest and attention, exhibiting the strongest appeal and novelty. The N-S group scored the highest in the being-away dimension, with its soundscape environment displaying good order, coherence, and comprehensibility, allowing participants to easily extract information. In the compatibility dimension, the soundscape environment of the H-M group aligned with participants’ emotions, needs, and expectations, matching their psychological state and personality traits, though the result was not significant. In the extent dimension, the soundscape environment of the H-M group conveyed information with good breadth and depth, allowing individuals to perceive and deeply engage with the environmental sounds. In terms of the overall score across all dimensions, the ranking was: H-M group > L-M group > L-S group > H-S group > N-M group > N-S group. The H-M group performed the best, indicating that its soundscape environment was the most ideal virtual roaming experience in terms of attractiveness, information acquisition, emotional compatibility, and experiential depth. Additionally, within environments with the same density of evergreen trees, soundscapes involving multiple people always scored higher than those involving a single participant.

Table 4.

One-Way ANOVA of PRSS.

Indicator	Group	Mean	SD	ANOVA		Post-hoc comparison
				F	Sig
Fascination	N-S	1.803	0.801	2.45	0.037*	H-M> L-S> L-M> H-S> N-S> N-M
	N-M	1.773	0.670
	L-S	2.258	0.803
	L-M	2.242	0.868
	H-S	1.955	0.602
	H-M	2.394	0.918
Being-away	N-S	2.382	0.580	4.616	0.001**	N-S> N-M> L-S> L-M> H-S> H-M
	N-M	2.262	0.916
	L-S	2.185	0.514
	L-M	2.059	0.679
	H-S	1.993	0.736
	H-M	1.637	0.754
Compatibility	N-S	2.091	0.714	2.072	0.073	H-M> N-M> N-S> H-S> L-M> L-S
	N-M	2.288	0.730
	L-S	1.894	0.538
	L-M	2.015	0.604
	H-S	2.076	0.642
	H-M	2.485	0.859
Extent	N-S	1.788	0.655	3.130	0.011*	H-M> L-M> L-S> N-M> H-S> N-S
	N-M	2.121	0.702
	L-S	2.152	0.512
	L-M	2.273	0.588
	H-S	1.955	0.415
	H-M	2.470	0.834
Overall score	N-S	7.239	1.954	3.344	0.007**	H-M> L-M> L-S> H-S> N-M> N-S
	N-M	8.238	1.870
	L-S	8.428	1.392
	L-M	8.894	2.138
	H-S	8.326	1.024
	H-M	9.485	2.682

a: * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001.

b: N-S, single travel in non-evergreen environments; L-S, single travel in low-density evergreen environments; H-S, single travel in high-density evergreen environments; N-M, multi-person in non-evergreen environments; L-M, multi-person travel in low-density evergreen environments; H-M, multi-person travel in high-density evergreen environments.

Analysis of results of physiological changes in landscape and soundscape

Results of EEG analysis

A paired-sample t-test was used to compare EEG changes and the arousal index in terms of the β/α ratio between the pre-test (baseline) and post-test (last 1 min). The results are shown in Fig. 6. The findings indicated that after viewing the six different audiovisual combinations of virtual roaming videos, all groups exhibited an increase in alpha waves (α) and a decrease in beta waves (β), though the extent of change varied across groups. The β/α ratio decreased for all groups, as this ratio is commonly used to assess stress levels⁷³, suggesting that different forest boardwalk environments were effective in reducing stress.

Fig. 6.

EEG results. (a) αPSD in occipital region, (b) βPSD in occipital region, (c) αPSD in temporal region, (d) βPSD in temporal region, (e) β/α in occipital region, (f) β/α in temporal region.

A one-way ANOVA was used to compare the differences in EEG changes after viewing the six different audiovisual combinations of virtual roaming videos. The results are shown in Tables 5 and 6. In the temporal lobe, Δα: H-M group > L-M group > N-M group > L-S group > H-S group > N-S group, Δβ: N-S group > H-M group > L-M group > H-S group > N-M group > L-S group; in the occipital lobe, Δα: H-M group > L-M group > N-M group > L-S group > N-S group > H-S group, Δβ: N-S group > H-S group > L-M group > H-M group > N-M group > L-S group. The H-M group showed the greatest impact on alpha waves in both the temporal and occipital lobes, while the N-S group had the greatest impact on beta waves in both regions, and these differences were statistically significant. In other words, under the interaction conditions of high-density evergreen trees and activity sounds from multiple people, alpha wave activation in the temporal and occipital lobes of the brain was significantly enhanced. Conversely, in a forest environment without evergreen trees, solitary walking significantly increased beta wave activation in these two regions.

Table 5.

Statistical results of one-way ANOVA for EEG in the Temporal lobe.

Indicator	Group	Mean	S D	ANOVA		Post-hoc comparison
				F	Sig
Δα	N-S	0.00247	0.00519	2.38	0.042*	H-M > L-M > N-M > L-S > H-S > N-S
	N-M	0.00585	0.00492
	L-S	0.00466	0.00667
	L-M	0.00714	0.01300
	H-S	0.00264	0.00309
	H-M	0.00973	0.01245
Δβ	N-S	− 0.00105	0.00173	2.852	0.018*	N-S > H-M > L-M > H-S > N-M > L-S
	N-M	− 0.00022	0.00027
	L-S	− 0.00011	0.00035
	L-M	− 0.00039	0.00043
	H-S	− 0.00034	0.00048
	H-M	− 0.00057	0.00123
Δβ/α	N-S	− 0.00858	0.01388	1.979	0.001**	H-M > L-M > N-M > N-S > L-S > H-S
	N-M	− 0.00903	0.00650
	L-S	− 0.00684	0.00883
	L-M	− 0.01101	0.01537
	H-S	− 0.00535	0.00532
	H-M	− 0.01529	0.01539

a: *indicates p < 0.05, **indicates p < 0.01, ***indicates p < 0.001.

Note b: N-S, single travel in non-evergreen environments; L-S, single travel in low-density evergreen environments; H-S, single travel in high-density evergreen environments; N-M, multi-person in non-evergreen environments; L-M, multi-person travel in low-density evergreen environments; H-M, multi-person travel in high-density evergreen environments.

Table 6.

Statistical results of one-way ANOVA for EEG in the occipital lobe.

Indicator	Group	Mean	SD	ANOVA		Post-hoc comparison
				F	Sig
Δα	N-S	0.00369	0.00662	1.094	0.037*	H-M > L-M > N-M > L-S > N-S > H-S
	N-M	0.00554	0.01447
	L-S	0.00511	0.00679
	L-M	0.00647	0.01474
	H-S	0.00274	0.00383
	H-M	0.00982	0.01437
Δβ	N-S	− 0.00095	0.00132	4.32	0.001**	N-S > H-S > L-M > H-M > N-M > L-S
	N-M	− 0.00025	0.00024
	L-S	− 0.00021	0.00020
	L-M	− 0.00026	0.00032
	H-S	− 0.00029	0.00031
	H-M	− 0.00026	0.00065
Δβ/α	N-S	− 0.00973	0.01368	1.054	0.004**	H-M > N-S > L-M > N-M > L-S > H-S
	N-M	− 0.00769	0.01467
	L-S	− 0.00755	0.00725
	L-M	− 0.00909	0.01631
	H-S	− 0.00524	0.00507
	H-M	− 0.01378	0.01688

a: * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001.

b: N-S, single travel in non-evergreen environments; L-S, single travel in low-density evergreen environments; H-S, single travel in high-density evergreen environments; N-M, multi-person in non-evergreen environments; L-M, multi-person travel in low-density evergreen environments; H-M, multi-person travel in high-density evergreen environments.

Regarding the arousal index, the changes in Δβ/α values showed: In the temporal lobe, H-M group > L-M group > N-M group > N-S group > L-S group > H-S group; in the occipital lobe, H-M group > N-S group > L-M group > N-M group > L-S group > H-S group. Overall, the H-M group exhibited the greatest changes in both the temporal and occipital lobes, while the H-S group showed the least change. This suggests that in forest boardwalk environments with high-density evergreen trees, the interaction with sounds from multiple people results in the maximum activation of the temporal and occipital lobes, effectively reducing the stress on the visual and auditory nervous systems. On the other hand, solitary walking produced the least restorative effect.

The changes in brain activity exhibited by participants after viewing the six different audiovisual combinations of virtual roaming videos are shown in Fig. 7.

Neuroemotional indices results

We used the emotional indicators exhibited by participants to assess their emotional responses after viewing the six different audiovisual combinations of virtual roaming videos. The results, shown in Table 7, reveal that all groups experienced an increase in engagement, relaxation, and interest, while excitement, stress, and concentration decreased. These effects showed significant differences in engagement, stress, and relaxation indicators. Among the groups, the N-M group exhibited the most significant increase in engagement (p = 0.035) and the largest decrease in excitement (p = 0.051), though the statistical difference was not significant. In terms of concentration, the N-S group showed the greatest decrease (p = 0.299), but the result was not significant. The H-S group’s video most effectively increased participants’ interest (p = 0.442), although the difference was not significant. The H-M group’s video was the most effective in reducing participants’ stress (p = 0.047) and enhancing relaxation (p = 0.043), both showing significant differences. Generally, physiological arousal levels are positively correlated with measured scores. In other words, in environments without evergreen trees, engaging in group activities led to higher levels of engagement and relatively lower excitement, while solitary walking was more likely to decrease concentration. In high-density evergreen tree environments, solitary walking stimulated more interest, but in the presence of multiple people and their accompanying sound interactions, participants experienced greater reductions in stress and increased feelings of relaxation.

Table 7.

Neuroemotional indicator one-way ANOVA results.

Indicator	Group	Mean	SD	ANOVA		Post-hoc comparison
				F	Sig
ΔEngagement	N-S	0.097	0.178	2.491	0.035*	N-M > L-M > H-S > H-M > N-S > L-S
	N-M	0.241	0.125
	L-S	0.051	0.215
	L-M	0.114	0.188
	H-S	0.104	0.223
	H-M	0.103	0.193
Δexcitement	N-S	− 0.170	0.182	0.861	0.051	H-M > H-S > L-M > N-S > L-S > N-M
	N-M	− 0.179	0.194
	L-S	− 0.175	0.132
	L-M	− 0.126	0.259
	H-S	− 0.105	0.182
	H-M	− 0.092	0.196
Δstress	N-S	− 0.132	0.194	0.925	0.047*	L-S > N-S > L-M > N-M > H-S > H-M
	N-M	− 0.151	0.183
	L-S	− 0.111	0.232
	L-M	− 0.149	0.226
	H-S	− 0.167	0.134
	H-M	− 0.223	0.124
Δrelaxation	N-S	0.106	0.216	0.675	0.043*	H-M > N-M > L-M > N-S > L-S > H-S
	N-M	0.160	0.181
	L-S	0.103	0.210
	L-M	0.155	0.225
	H-S	0.093	0.206
	H-M	0.175	0.162
Δinterest	N-S	0.137	0.179	0.966	0.442	H-S > H-M > L-M > N-S > N-M > L-S
	N-M	0.130	0.193
	L-S	0.104	0.213
	L-M	0.167	0.204
	H-S	0.218	0.177
	H-M	0.170	0.167
Δfocus	N-S	− 0.198	0.247	1.229	0.299	H-M > N-M > L-M > H-S > L-S > N-S
	N-M	− 0.076	0.141
	L-S	− 0.136	0.180
	L-M	− 0.111	0.192
	H-S	− 0.135	0.215
	H-M	− 0.060	0.253

a: * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001.

b: N-S, single travel in non-evergreen environments; L-S, single travel in low-density evergreen environments; H-S, single travel in high-density evergreen environments; N-M, multi-person in non-evergreen environments; L-M, multi-person travel in low-density evergreen environments; H-M, multi-person travel in high-density evergreen environments.

Discussion

Virtual winter forest trail environments facilitate the induction of positive emotional states

Immersive virtual environments (IVEs) can partially replicate the recovery experience found in real-world settings and have demonstrated a perceptual recovery effect²⁶. The winter forest timber trail, characterized by its distinctive landscape features and spatial qualities, significantly enhances the overall emotional state of transient visitors^6,74,75. Liu, QH’s research further indicates that coniferous forests are particularly effective in eliciting positive emotions and promoting mental well-being^5,76. We further demonstrated through virtualization that walking in a snow-covered winter urban forest trail environment alleviates negative emotions and provides significant perceived restorative benefits^14,77. This phenomenon may be attributed to the diminished presence of flora and fauna during winter⁷⁸, which generally creates a quieter atmosphere than in other seasons, thereby alleviating psychological stress and anxiety⁴⁴. The blanket of snow covering the ground and vegetation simplifies the terrain and plant communities, presenting a smooth and continuous visual surface⁷⁹. This creates a predominantly white landscape that is both monotonous and pristine, fostering a serene and pure visual experience⁸⁰. Such an environment effectively stimulates positive emotional responses⁸¹. Additionally, the snowy landscape serves as a restorative element, enabling individuals to engage with environments, activities, and experiences that diverge from their daily routines¹⁴. Although previous studies have shown that virtual technology can largely replicate the landscape perception outcomes found in real-world environments⁸², there are still limitations in its ability to reproduce physical environments⁸³. This gap may be even more pronounced in winter settings, as the sounds and tactile feedback of snow in real environments are also unique therapeutic elements, which virtual environments are not yet able to fully replicate in terms of these multisensory experiences⁸⁴.

EEG validation that different winter landscape forest wood trail environments promote relaxation of the audiovisual nerves

During the transfer of information within the brain, neurons convey data through electrochemical signals⁸⁵. EEG technology amplifies these electrical signals and converts them into digital formats, enabling real-time monitoring of cerebral activity⁸⁶. The temporal lobe of the cerebral cortex is primarily responsible for auditory processing, while the occipital lobe is chiefly involved in visual information processing^73,87. In this study, we observed that after experiencing different audiovisual restorative environments, all participants showed a significant increase in α-wave activity in the temporal and occipital lobes, while β-wave activity decreased. The increase in α-waves is typically associated with relaxation and a reduced state of arousal, while the decrease in β-wave activity may reflect a reduction in cognitive tasks and a lowered brain response to environmental stimuli⁸⁸. In conjunction with neuroemotional assessments, we discovered that the landscape features and spatial experiences of various winter-view forest trail environments exert a significant therapeutic effect in alleviating stress. Furthermore, these environments facilitate the relaxation of audiovisual pathways. Notably, perceptual recovery levels differ among various demographic groups, corroborating findings from previous studies^30,89. Additionally, the increase in α-wave activity in participants who experienced restorative environments with natural sounds suggests relaxation and attention recovery. This is consistent with the Attention Restoration Theory (ART), which posits that environments can promote relaxation and reduce cognitive load⁹⁰. Sound propagates as a pressure wave, vibrating the eardrum and prompting hair cells on the basilar membrane to transduce mechanical energy into electrical signals⁹¹. These signals are subsequently transmitted via nerve fibers to the temporal lobe, where auditory information is processed⁹². Visual information acquisition commences with the absorption of photons by photoreceptors in the retina, which convert the light stimulus into electrical signals⁹³. These signals are conveyed through optic nerve fibers to the greater occipital lobe for processing, culminating in visual perception⁹⁴. Distinct audiovisual combinations of light signals and pressure waves produced by virtual roaming videos elicit varying visual and auditory stimuli, which, in turn, provoke diverse emotional responses and physiological effects, accounting for the discrepancies in perceptual recovery levels.

Winter forest trail environments characterized by high green visibility significantly facilitate the physical and mental recovery of visitors

Research indicates that medium-density forest environments provide superior perceived restoration benefits compared to low- and high-density forests^95,96. These studies are primarily based on summer landscapes. Seasonal differences significantly affect landscape perception experiences, with preferences for summer landscapes generally being higher than for winter landscapes⁹⁷. Compared to summer landscapes, experiencing more green spaces in winter is more beneficial for perception recovery, especially in single-layer forest landscapes⁹⁸. While this investigation focuses on winter forests, revealing that forest trail environments with high green visibility significantly enhance the physical and mental recovery of recreationists. From an environmental psychology perspective, evergreens retain their foliage during winter, and higher densities of evergreens correspond to increased green visibility, thereby providing continuous visual comfort⁹⁹. Additionally, the rich sensory experiences and opportunities for social interaction within natural environments significantly promote physical and mental well-being^100,101. Ecologically, high-density forests in winter regulate microclimates¹⁰², offering warm, wind-sheltered outdoor spaces¹⁰³. Additionally, they effectively attenuate noise, fostering a tranquil and relaxing atmosphere conducive to alleviating stress and fatigue¹⁰⁴. According to Attention Restoration Theory (ART), a high density of evergreens cultivates a rich natural landscape in winter forests, which can mitigate stress and fatigue through multisensory stimulation and gentle attraction¹⁰⁵, thereby aiding in the restoration of fatigued cognitive function¹⁰⁶.

Better restoration benefits from the interaction of winter high-density evergreen tree forest trail environments with multi-person activity ambient sound

Research indicates a significant negative correlation between anthropogenic sound and environmental assessment^41,107,108. Zeng, C.C. argued that quieter anthropogenic sounds in forests can be more physiologically stressful³¹. However, our findings demonstrate that the restorative benefits of multi-person activity sounds consistently surpass those of single-person activity sounds in winter forest trail environments with equivalent densities of evergreen trees. This may be attributed to the fact that crowd-gathering scenarios are less likely to occur in expansive winter forest settings, leading to increased feelings of loneliness and anxiety when alone¹⁰⁹. When sounds exhibit significant temporal variation, calm sound stimuli can induce stress⁷⁷. In quiet winter environments, higher loudness levels are associated with greater perceptual sensitivity¹¹⁰. Engaging in common activities with multiple individuals fosters increased engagement¹¹¹, allowing individuals to divert their attention from negative emotions or stressors^112,113 and enhancing overall well-being¹¹⁴, thereby effectively reducing anxiety and restlessness. Furthermore, multi-person activity sounds are richer and more dynamic than those of single-person activities, providing a sense of social interaction and emotional support¹¹⁵. This increased sense of affinity and belonging promotes mental health and alleviates stress.

Research has demonstrated that both quiet sounds in high-density forest environments and loud natural sounds in low-density forests can evoke positive emotions³¹. However, our findings show that in the high-density evergreen tree forest boardwalk environment, the interaction with sounds generated by group activities has a significant positive effect on both physiological and psychological recovery, while solitary walking shows the lowest recovery benefits. While anthropogenic sound is generally acknowledged to negatively impact perceived recovery^77,116, sound transmission is partially obstructed in high-density evergreen forests during winter, resulting in footsteps that possess a gentle and rhythmic quality. Additionally, the sounds of multiple individuals conversing can partially mask these footsteps, minimizing excessive interference with auditory perception. Conversely, the background sounds of multiple activities enhance the effects of attention recovery¹¹⁷. The combination of a dense evergreen forest in winter and the sounds of various activities provides a rich array of visual and auditory stimuli, which can alleviate feelings of bleakness and seasonal depression¹¹⁸. This interaction also fosters a sense of belonging and security, promoting both physical and mental health. In contrast, walking alone in a winter high-density evergreen tree forest environment, due to the lack of diverse auditory stimuli, may lead to increased alertness¹¹⁹ and create a sense of silence and psychological feelings of loneliness or helplessness¹²⁰, which in turn diminishes the restorative effects.

Limitations and future directions

Due to the limitations of objective experimental conditions, this study has several constraints. (1) Uniqueness of Visual and Auditory Factors: In this study, evergreen trees were the only visual factor considered. However, factors such as boardwalk design, light and shadow variations, and the vertical design of the landscape also have an impact on physical and mental recovery. Additionally, variations in decibel levels caused by activities and the complexity of footstep sounds can influence auditory experiences, which require further exploration in future studies. (2) Sample Size Bias: The participants in this study were predominantly young adults with an average age of 22.55 years, and the sample size was relatively small. Although the results are statistically significant to some extent, there is a lack of discussion on demographic characteristics such as BMI, age distribution, gender, and cultural background. In the future, we plan to invite a larger and more diverse group of participants, including those with varying BMI, age, and cultural backgrounds. (3) Limitations of the Virtual Environment: While this study used a virtual environment to simulate real physical settings, the replication of actual physical environments still has limitations, and there is room for further research in this area.

Conclusion

This study investigates the physiological and psychological responses of participants engaged in simulated walks within a winter forest park, utilizing electroencephalography (EEG) and virtualization techniques. The research focuses on the effects of two variables—evergreen tree density (ETD) and event ambient sound (EAS)—on participants traversing a forested wooden trail. The interaction between these variables elucidates distinct mechanisms for stress alleviation and the enhancement of positive emotions. Our findings indicate that: (1) virtual winter forest trail environments significantly promote positive emotional states; (2) EEG data confirm that various winter forest trail environments facilitate the relaxation of audiovisual nerves, albeit with differing levels of perceptual recovery; (3) winter forest trail environments characterized by high green visibility significantly facilitate the physical and mental recovery of visitors; (4) multi-person activity sounds outperformed single-player audio in terms of restorative benefits, while companionship enhances the healing process; and (5) in a winter high-density evergreen tree forest boardwalk environment, the interaction with sounds generated by group activities offers better restorative benefits, while solitary walking provides the least restorative effects.

The findings of this study provide scientific evidence for designing more effective nature-based therapies and virtual rehabilitation environments, as well as offering important practical guidance for the future planning of forest park trails and plant landscapes. In the future construction of forest wellness areas, increasing the density of evergreen trees, carefully planning activity programs, and encouraging group walks can promote communication. Additionally, incorporating social interaction features to enhance the sense of social support can more effectively alleviate negative emotions and induce positive emotional expression. Furthermore, for individuals with mobility limitations, the promotion of virtual forest therapy technology, especially under extreme weather conditions like winter, offers convenient health intervention options.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, Y.C. and W.N.; methodology, Y.C.; software, Q.W., M.S. and W.B.; validation, Q.W.; formal analysis, Q.W.; investigation, Y.C., W.N., Q.W. and M.S.; resources, Y.C., W.N. and J.Y.; data curation, Q.W.; writing—original draft preparation, Q.W; writing—review and editing, Q.W; visualization, Q.W.; supervision, W.B.; project administration, Y.C. and W.N.; funding acquisition, Y.C. All authors have read and agreed to the published version of the manuscript.

Data availability

The data support the findings of the present study and are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Institutional review board statement

This research experiment was approved by the Ethics Committee of the College of Forestry and Grassland Science of Jilin Agricultural University (No. 2024-CFGSJLAU-1-1)).

Informed consent

Informed consent was obtained from all subjects involved in this study.