Effects of Window Green View Index on Stress Recovery of College Students from Psychological and Physiological Aspects

Abstract

Students often experience high levels of daily academic pressure, spending extended periods within indoor classroom environments. Windows, as a medium of proximity to nature, play an important role in relieving stress. However, the broader implications of the Window Green View Index (WGVI) on individual well-being remain underexplored. This study aims to assess the effects of WGVI on stress recovery in college students by utilizing virtual reality technology to create five classroom environments with varying WGVI levels: 0%, 25%, 50%, 75%, and 100%. Twenty-four participants were subjected to the Trier Social Stress Test before engaging with the different WGVI scenarios for stress recovery. Both subjective assessments and objective physiological indicators were evaluated. Results indicated that participants exhibited the lowest Profile of Mood States (POMS) score (−4.50) and significantly improved systolic blood pressure recovery at a 25% WGVI level. The examination of EEG data revealed that the O2 channel in the occipital region exhibited the highest level of activity in the alpha frequency range during the experiment. Additionally, a significant association was observed between the EEG measurements and the subjective rating of stress. This study underscores the significance of incorporating WGVI into the design and planning of college buildings to promote mental health and well-being among students.

1. Introduction

The mental health of college students has been increasingly recognized as a critical public health concern [1]. During the pivotal developmental stage of transitioning from adolescence to adulthood, students face numerous academic challenges, interpersonal conflicts, and career-related decisions, all of which contribute to heightened stress levels [2]. This substantial stress can adversely impact both their physical and mental well-being. Stress reduction theory posits that being exposed to nature improves cognitive function and reduces stress levels [3]. Additionally, research indicates that interaction with nature fosters a greater sense of security. As a result, integrating natural elements into campus planning is regarded as an effective strategy for promoting the psychological well-being of university students and mitigating the prevalence of mental health disorders [4].

Recent research has extensively demonstrated the positive effects of incorporating natural spaces on individual well-being [5,6]. However, college students spend the majority of their time in classroom settings, limiting their direct access to nature for stress alleviation. This restriction underscores the need for alternative strategies to reconnect students with natural environments, with one such approach being the optimization of window views. Windows have emerged as a prevalent means through which college students engage with the natural environment [7,8]. A positive correlation has been observed between having a view of nature outside windows and individuals’ mental health and happiness [9]. Du et al. [10] investigated the effect of natural window views on the perceived quality of indoor environments, concluding that such views can significantly enhance overall comfort. Similarly, Jiang et al. [11] examined the health benefits of daylight and natural landscape views through windows in experimental settings. They found that exposure to natural landscapes through windows could alleviate stress and reduce fatigue. Zhang et al. [12] studied the effects of green window landscapes on sleep quality among older adults, revealing that a higher ratio of green views through windows was associated with improved sleep quality in this population. Gilchrist et al. [13] also suggested that viewing trees, grasses, and ornamental plants through windows can enhance mental health. These findings underscore the strong correlation between natural elements visible in the windowscape and their perceived restorative effects.

The World Green Building Council framework [14] for promoting a healthy and sustainable indoor environment identifies biophilia and views of nature as key components of a healthy space. Researchers have shown considerable interest in integrating Biophilic Design (BD) strategies into indoor environments, with natural window views serving as one aspect of this approach. Numerous studies have explored the effects of outdoor nature visible through windows on individuals’ recovery. However, definitive conclusions are still lacking regarding quantitative research on the visibility of greenery and the varying effects of different levels of nature exposure. In response to this gap, scholars have introduced the concept of the green view index (GVI) to quantify the visibility of nature [15]. The GVI is the proportion of green in an image that is visible to the human eye. Subsequent studies have explored the restorative benefits of GVI through subjective evaluation [9,16]. For instance, Jin et al. [16] used the “Profile of Mood State” and “Semantic Analysis Scale” as psychological assessment indicators to study the GVI of urban parks ranging from 10% to 66%. They found that the higher the GVI, the greater the physical and mental recovery effect on individuals. Xu et al. [17] examined the effect of street GVI on people by constructing 15 virtual street models. Their findings revealed that 25% of street GVIs were more attractive to people. This is because too much green visibility can result in obstructed views and poor spatial extension, which reduces fascination. Similar conclusions were found in a study by Zhang et al. [12]. They found that the greenness visible through windows was objectively assessed using the green window ratio and found that a 25% green visibility ratio was better for the health of the elderly. Huang et al. [18] compared scenarios with varying GVI levels through skin conductance measurements and questionnaire surveys, demonstrating that higher GVI ratios were more effective at relieving stress. Similarly, Jin et al. [16] found that when GVI was higher than 35%, the scene had a promoting effect on the subjects’ cutaneous electrical activity. The GVI was positively correlated with cutaneous electrical conductivity and physiological recovery. To sum up, while these studies offer valuable insights into the effects of GVI, they primarily focus on subjective evaluations and basic physiological indicators, without considering the influence of more complex physiological metrics, such as brain activity.

It is important to note that vision is the dominant sense for receiving information from the environment [19]. Visual receptors capture information and transmit it to the brain through neurons, where it is interpreted and processed. Therefore, some scholars began to study the electrical response of people under different GVI. Nie et al. [20] constructed a panoramic scene using VR and discovered significant changes in the α values of the frontal and occipital lobes. As the GVI increased from 0% to 60%, subjects reported greater pleasure, with the brain’s topographic map shifting from dark blue to brown-yellow. Olszewska et al. [21] investigated the effect of various levels of greening on brain activity. They demonstrated that higher levels of visible greenery led to increased α-wave activity and heightened relaxation. GVI has various effects on people in different environments, resulting in varying research findings. For instance, Elsadek et al. [19] investigated the effects of high-rise window views on well-being by measuring brain activity. They found that window views featuring green space with water or green space alone produced the most beneficial outcomes, including increased alpha brain wave activity and enhanced parasympathetic nervous system activation. Overall, inconsistencies in GVI research findings can be attributed to variations in research focus, such as examining the street green view index [17] or the floor green view index [21,22]. These different contexts yield distinct effects of GVI on individuals, leading to divergent conclusions. Future studies should therefore aim to clarify research objectives and design more targeted experiments to better understand these dynamics.

In summary, some scholars have studied the GVI from various perspectives, including the street and floor levels. However, there is a noticeable lack of quantitative research on GVI from the window view angle, particularly in relation to its effects on stress recovery as measured through EEG. The key findings from these studies are summarized in

Table 1. Summary of key findings.

Topic	Key Findings	Advantages	Limitations	Ref.
The green view index	The green view index can be used to quantify nature.	The green view index allows nature to be quantitatively analyzed and a clearer understanding of its visual extent.	Details are not clearly defined and lack the nuanced categorization of GVIs.	[15,17,22]
Subjective restoration benefits of GVI	The higher the GVI, the better the psychological recovery.	The subjective evaluation method has effectively studied the effect of GVI on human psychological stress.	The subjective evaluation method is greatly affected by personal emotional factors, and the research results are not objective.	[9,12,16,17,18]
Physiological restoration benefits of GVI	The higher the GVI, the lower the SCL. The more WGVI, the higher the brain’s α activity.	From the physiological point of view of feedback, GVI can relieve stress.	Most of the studies have applied simple physiologic indicators, and fewer studies have been conducted by EEG.	[18,19,20,21]

.

To address the limitations of previous studies, this study introduces the concept of the Window Green View Index (WGVI), which quantifies the proportion of natural scenery visible through a window. Three research objectives were proposed as follows:

• Quantitatively study the impact of WGVI on students’ subjective stress and physiological indicators, thereby addressing the gap in physiological evaluations in earlier studies.
• Investigate the EEG characteristics of individuals during stress recovery under different WGVI conditions to explore the underlying physiological mechanisms of stress recovery.
• Propose an optimal WGVI range for promoting stress recovery.

Using VR technology, five classroom environments were created with WGVI levels of 0%, 25%, 50%, 75%, and 100%. Participants underwent the Trier Social Stress Test (TSST) and wore VR glasses to view a natural scene outside the window to alleviate their stress. Both subjective evaluations and physiological indicators were recorded. A one-way ANOVA was applied to the data to examine the influence of WGVI on stress recovery among college students. The significance of the research is to provide new design ideas for university architecture design and campus landscape planning.

2. Methodologies

The focus of this study is the Window Green View Index (WGVI), intending to investigate its effects on human stress recovery. To achieve this, virtual reality (VR) technology was employed to create immersive environments with varying levels of WGVI. The subjects were asked to perform the Trier Social Stress Test (TSST) in a virtual space to induce psychological and physical stress, and their blood pressure was recorded post-stress. During the recovery phase, participants viewed scenes with different WGVI levels while their psychological and physiological indicators were monitored. This study analyzed changes in blood pressure before and after recovery, along with heart rate variability (HRV) and EEG data, to explore the effects of different WGVI levels on stress recovery.

2.1. Subjects

The sample size was determined using G. Power 3.0 software. To minimize errors and ensure the practical significance of the research findings, an effect size of 0.5 (medium) was set, with a significance level (α) of 0.05 and a power of 0.8. This calculation indicated that a minimum of 24 participants was required. Consequently, 24 college students were selected as subjects for the experiment. All subjects had no color blindness, hypertension, 3D vertigo, or any other conditions unsuitable for wearing VR devices. Additionally, their corrected vision was within the normal limits. Before the experiment, subjects were instructed to obtain adequate rest, avoid caffeine and alcohol, and sign an informed consent form. This study adhered to the Declaration of Helsinki and was approved by the Ethics Committee of Qingdao University of Technology. The traits of subjects are displayed in

Table 2. The traits of subjects.

Traits of Subjects		Number/People
Gender	Male	8
	Female	16
Corrected vision	Left vision ≥ 1.5	24
	Right vision ≥ 1.5	24
Educational background	Undergraduate student	6
	Master’s student	18

.

2.2. Experimental Process

The experiment was completed in the VR lab of Qingdao Technological University. Figure 1 shows the experimental site. The experiment lasted approximately 120 min per subject (including rest time) and consisted of five 20 min scenarios, on average. The aim was to fully immerse the subjects in the scene while avoiding fatigue and simulation sickness. As shown in Figure 2, the experiment was divided into two stages: preparation and repetition. During the preparation phase, subjects wore VR glasses and physiological monitoring devices and sat quietly for 10 min to eliminate pre-experiment activity distractions. In the formal phase, the subjects performed the Trier Social Stress Test (TSST) while facing a virtual blackboard. The TSST, a widely used standardized stress-inducing protocol [23], consisted of a one-minute Stroop stress test followed by three minutes of basic addition and subtraction tasks with fewer than 1000 calculations. The Stroop test required participants to quickly identify the meanings and colors of words. Blood pressure and pulse were measured at the last minute of the stress test and recorded as BP1 and HR1, respectively. Following the stress phase, subjects rotated 90 degrees to view a window with varying levels of greening for five minutes to recover from stress. EEG and HRV data were recorded during the recovery phase, with blood pressure and pulse measured in the final minute and recorded as BP2 and HR2. Finally, subjects completed the POMS questionnaire. The experiment was then repeated with different WGVI scenes in random order. The subjects were asked to rest 5–8 min after each scene.

2.3. VR Scenes and Experimental Environment

This study employed virtual reality (VR) technology to explore the effects of different WGVI levels on stress recovery. A classroom space model was developed using SketchUp Pro 2021 software, where WGVI was manipulated by altering the density of natural elements outside the window. Subjects wore VR glasses for the experiment to make the subjects immersed. Studies have found no significant differences between the mental and bodily reactions of subjects in virtual space and real physical space [24]. Conducting experiments in VR allows for better control over window illumination, [25] as well as modifications to the view’s elements and the depth of field [26]. The ability to adjust WGVI through VR simulations—including scenarios with extremely high WGVI levels rarely encountered in real environments—overcomes the limitations of studying WGVI in physical spaces [20]. Therefore, varying architectural parameters within VR is a much more feasible method than varying the physical parameters of the actual building [27,28].

As shown in Figure 3a, the classroom model created in SketchUp measured 10.00 m × 9.00 m with a ceiling height of 3.80 m. The windows were located on the southern side of the classroom, featuring an aspect ratio of 1:1.05 and a window-to-wall ratio of 1:3. Trees were placed 8 m from the windows, and the outdoor environment simulated sunny conditions. Subjects were seated in the middle of the classroom with a window view containing sky and natural elements (content showing trees, plants, and other sources of greenery) [29]. Figure 3b is the subject’s perspective from the virtual scene.

As shown in Figure 4, the grid method was used to segment the window view, and the pixel calculation method was used to determine the WGVI. WGVI was calculated as shown in Formula (1):

$$WGVI={\displaystyle \frac{S_g}{Sa}}\times 100\%$$

(1)

where S_g represents the visible area of the green part of the plant irrespective of its distance from the window view. Nevertheless, the distance for the nature seen from the window view was 8 m in this experiment. Sa refers to the total size of the window. Pixel computing may identify the area occupied by the branches of a tree as vegetation (i.e., the pores between the branches and leaves are also counted as vegetation). Therefore, the WGVI measured by the experiment is about 3% to 5% higher than the manual calculation, and the scene with 100% WGVI is only close to 100%.

Figure 5 shows five kinds of scenes outside the window. The viewing angle in Figure 3 simulates the horizontal viewing angle of human eyes when sitting in the middle of the classroom. Figure 5a without nature outside the window (0% WGVI and 100% sky) is the control group of the experiment. Figure 5b has 25% WGVI and 75% sky. Figure 5c,d have 50% WGVI and 55% sky, 75% WGVI and 25% sky, and 100% WGVI without sky, respectively. The plant in the scene is elm, a common plant in China. Except for the difference in WGVI, the visual characteristics of the elements are identical.

2.4. Experimental Equipment

The equipment utilized in the experiment is shown in

Table 3. Experiment equipment.

Usage	Model	Test Program	Range	Accuracy/ Sampling Rate	Figure
Showing experimental scenes.	HTC Vive Pro	-	110°	2880 × 1660
Measuring physiological indicators.	Emotiv EPOC (Flex Gel Sensor Kit)	EEG	0.16–43 Hz	128 Hz
	Healink-R211B	HRV	0.6–40 Hz	1000 Hz
	Yuwell YE660	BP	0–299 mmHg	±4 mmHg
	Yuwell YE660	HR	Times/min	1

. The subjects wore HTC Vive Pro to immerse themselves in virtual scenes. EEG, HRV, and BP were measured by Emotiv EPOC (Flex Gel Sensor Kit, San Francisco, CA, USA); Healink-R211B, Bengbu, China; and Yuwell YE660, Shanghai, China, devices. Equipment monitoring EEG, HRV, and BP were worn throughout the entire experiment.

Moreover, the Enscape 3.1 software was used to control lighting parameters in virtual scenes to ensure that each experimental environment had consistent window lighting.

2.5. Psychological and Physiological Indicators

2.5.1. Psychological Indicators

The questionnaire of Profile of Mood States (POMS) was employed to assess the subjects’ emotional state following the visual stimulus [30,31]. To reduce the burden on the participants, the Chinese version of the POMS was utilized, with a short form containing 28 questions (POMS citation in

Table A1. POMS.

	Not At All	A Little	Moderately	Quite a Lot	Extremely
Tense	0	1	2	3	4
Worn out	0	1	2	3	4
Unhappy	0	1	2	3	4
Lively	0	1	2	3	4
Uneasy	0	1	2	3	4
Confused	0	1	2	3	4
Fatigued	0	1	2	3	4
Sad	0	1	2	3	4
Energetic	0	1	2	3	4
Unable to concentrate	0	1	2	3	4
Restless	0	1	2	3	4
Exhausted	0	1	2	3	4
Discouraged	0	1	2	3	4
Active	0	1	2	3	4
Nervous	0	1	2	3	4
On-edge	0	1	2	3	4
Weary	0	1	2	3	4
Grouchy	0	1	2	3	4
Full of pep	0	1	2	3	4
Forgetful	0	1	2	3	4
Bitter	0	1	2	3	4
Bushed	0	1	2	3	4
Worthless	0	1	2	3	4
Vigorous	0	1	2	3	4
Uncertain about things	0	1	2	3	4
Anxious	0	1	2	3	4
Hopeless	0	1	2	3	4
Competent	0	1	2	3	4

). The subjects were asked to score their emotional intensity on a 5-point Likert scale, with 0 indicating “not at all” and 4 indicating “very”. As shown in

Table 4. Five subscales of POMS.

Five Subscales of POMS	Relationship with Emotional States
Tension and Anxiety (T-A)	A lower score on T-A, C, D, and F indicates a better mood [32]
Confusion (C)
Depression (D)
Fatigue (F)
Vitality (V)	A higher score on V indicates a better mood [32]

, five subscales were selected as the psychological indicators for this experiment. TMD was calculated by summing the scores of the four negative emotions (T-A, C, D, and F) and subtracting the score for the positive emotion (V). The lower the TMD score, the less mood disorders and stress [32].

2.5.2. Physiological Indicators

Blood pressure is the most widely studied stress-related physiological indicator. Studies have shown that stress responses can raise blood pressure [33]. However, cardiovascular recovery values were utilized to assess the subjects’ stress recovery status. This approach was chosen due to the substantial influence of individual differences on cardiovascular indicators [34] The recovery value was calculated by subtracting the mean physiological value post-recovery from the mean value post-stress [35]. A positive value indicates successful recovery, while a negative or zero value suggests inadequate recovery.

Heart rate variability (HRV) is a measure of variations in the time interval between heartbeats. Stress activates the sympathetic nervous system (SNS), while relaxation engages the parasympathetic nervous system (PNS) [36]. HRV in the low-frequency (LF) and high-frequency (HF) ranges are correlated with stress exposure. Under stress, HF decreases and the HRV-nLF/nHF increases. Therefore, HRV-nLF/nHF is inversely related to mental stress [37]. In this study, Kubios 4.1.1 software was employed to analyze HRV in the frequency domain [38,39]. HRV-nLF/nHF was adopted as the index for evaluating stress levels, with lower HRV-nLF/nHF values indicating lower stress levels [40].

Through electrodes positioned on the scalp’s surface, the EEG equipment detects and logs electrical impulses produced by the brain. The voltage of the cell membranes varies as a result of brain neuronal activity. The electrodes record the activity of the central nervous system and provide an accurate depiction of the subject’s mental state by detecting these voltage changes happening in tens of thousands of neurons [41]. In addition, EEG signals show unwanted artifacts (such as eye flickering, muscle movements, etc.) that need to be eliminated.

2.6. EEG Data Processing

The EEGLAB of the MATLAB toolkit is used to preprocess EEG signals. As shown in Formulas (2) and (3), the procedure divides the time series data into fragments that may overlap and then performs the FFT. According to the frequency, they can be divided into different frequency bands. After that, the size of each segment is calculated, and the spectrum is averaged. The PSD characteristics of artifact-free EEG signals were obtained [42].

$$x_d\left(n\right)=x\left(n+dM-M\right),0\le n\le M,1\le d\le L$$

(2)

$$P_d\left(f\right)={\displaystyle \frac{1}{MU}}|{\sum }_{n=0}^{M-1}{x}_d\left(n\right)w\left(n\right)e^{-j2\pi fn}{|}^2$$

(3)

In Formulas (2) and (3), $x_d\left(n\right)$ is a sequence, where $d=1,2,3,\dots$; L is the signal interval; M is the interval length; and $w\left(n\right)$ is the windowed data. In the window function, U is the normalization factor of the power and the calculation formula is shown in Formula (4).

$$U={\displaystyle \frac{1}{M}}{\sum }_{n=0}^{M-1}{\left|w\left(n\right)\right|}^2$$

(4)

The Welch method was used to make corrections to improve the resolution and smoothness of the power spectral density curve, as in Formula (5).

$$PWelch(f)={\displaystyle \frac{1}{L}}{\sum }_{i=0}^{L-1}{P}_d\left(f\right)$$

(5)

Since the absolute intensity of each person’s brain wave is different, the α frequency related to the relaxation degree is selected for the average study. This study also used the relative frequency of the brain wave power electrical degree of relaxation [43] and mental stress [44], as Formulas (6) and (7), respectively. Relaxation is the ratio of the alpha power value to beta power value of EEG. Mental stress is mainly determined by the frontal and parietal lobes, which is the ratio of the frontal theta power to the parietal alpha power.

$$\mathrm{R}\mathrm{e}\mathrm{l}\mathrm{a}\mathrm{x}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{e}\mathrm{e}={\displaystyle \frac{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{a}\mathrm{l}\mathrm{p}\mathrm{h}\mathrm{a}}{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{b}\mathrm{e}\mathrm{t}\mathrm{a}}}$$

(6)

$$\mathrm{M}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{s}\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{s}={\displaystyle \frac{{\mathrm{R}\mathrm{T}}_{\mathrm{F}\mathrm{r}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{l}\mathrm{o}\mathrm{b}\mathrm{e}}}{{\mathrm{R}\mathrm{A}}_{\mathrm{P}\mathrm{a}\mathrm{r}\mathrm{i}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{l}\mathrm{o}\mathrm{b}\mathrm{e}}}}$$

(7)

2.7. Data Analysis

In this study, a one-way analysis of variance was used to compare the subjective evaluation and physiological indicators of different WGVI. The results were verified by a t-test to ensure that the data were statistically significant at the level of 0.05. The correlation between subjective evaluation and physiological indexes was compared by a Pearson correlation analysis [45] to prove that the selected physiological indicators can be used as evaluation indexes of stress recovery. In addition, a multi-parameter fitting analysis was used to further determine the optimal range of WGVI Settings that can improve stress recovery benefits.

3. Results

To study the effect of WGVI on the relaxation degree of college students, the POMS scale was used as a subjective evaluation indicator, and three objective indicators, including blood pressure, HRV, and EEG, were analyzed and discussed.

3.1. Subjective Questionnaire Evaluation

Figure 6 shows the results of the average POMS scores at different WGVI. The results showed that WGVI of 25% was significantly different from that of the control group (WGVI = 0%). As shown in Figure 6, negative emotions exhibited a trend of initially decreasing, followed by an increase as WGVI rose from 0% to 100%. At 25% WGVI, negative emotion scores were at their lowest, with Tension–Anxiety (T-A) at 3.00, Depression (D) at 2.27, Fatigue (F) at 3.09, and Confusion (C) at 2.75. These values represented decreases of 5.25, 5.28, 6.36, and 4.58, respectively, compared to the control group with 0% WGVI. This suggests that introducing moderate levels of greenery can significantly enhance mood by reducing negative emotions. However, the benefits appear to plateau, as further increases in WGVI beyond this threshold did not yield noticeable improvements in stress recovery. The Vigor (V) score reached its peak at 25% WGVI, with an average score of 13.00—approximately 1.7 times higher than the control group (4.83). Other groups with varying WGVI levels also showed significantly higher Vigor scores than the 0% WGVI group. These results demonstrate that window greening can positively influence the mood of college students, with different WGVI levels having varying effects on emotional states. Specifically, the WGVI of 25% produced the greatest improvement, alleviating negative emotions such as tension, anxiety, and confusion, while significantly enhancing vitality. These findings agree with the research conducted by Xu et al. on street GVI, where they also observed that a GVI of approximately 25% is particularly appealing to individuals [17]. Although the WGVI and street GVI differ in scope, their conclusions share notable similarities. Therefore, the design of the WGVI may be informed, to some extent, by the design principles applied to street greening.

Figure 7 shows the average TMD scores of subjects under various WGVI. At the level of 0.05, the overall mean showed a significant difference (F = 9.73, p < 0.05). As demonstrated in Figure 7, the average TMD score fluctuated significantly across varying WGVI levels. Initially, the average TMD score decreased, reaching its lowest value of −4.50 at a WGVI of 25%, after which it began to increase as WGVI exceeded 25%. The TMD scores at 50%, 75%, and 100% WGVI were 4.25, 6.58, and 16.42, respectively, all substantially lower than the control group’s (TMD = 28.58). These results suggest that natural landscapes visible through windows can effectively mitigate mood disorders, with a WGVI of 25% being particularly beneficial. The effectiveness of 25% WGVI in enhancing emotional states may stem from its balance between a green view and spatial openness, which is typically conducive to alleviating anxiety.

3.2. Physiological Data Analysis

3.2.1. Blood Pressure Analysis

Figure 8a shows the systolic blood pressure (SBP) recovery values of subjects with different WGVI. As shown in Figure 8a, the SBP recovery value of the subjects at 25% WGVI was up to 4.60 mmHg, representing an increase of 84.0% compared to the group at 0% WGVI (p < 0.05). At WGVIs of 50%, 75%, and 100%, the SBP recovery values were 1.70 mmHg, 1.20 mmHg, and 1.80 mmHg, respectively, none of which differed significantly from the control group. Nevertheless, these values were still lower than those of the control group, indicating that a WGVI of 25% is the most effective in improving SBP recovery efficiency and providing quicker relief from stress-induced SBP elevation. Figure 8b shows the diastolic blood pressure (DBP) recovery values of subjects with different WGVIs. As shown in the figure, when WGVI increased from 0% to 100%, the DBP recovery values of the subjects were 1.50 mmHg, 1.90 mmHg, 0.90 mmHg, 0.40 mmHg, and −0.60 mmHg, respectively, with no significant difference. The results showed that WGVI had little effect on DBP. This is because the systolic blood pressure depends on the elasticity of the arteries and the diastolic blood pressure depends on the tension of the arteries, and therefore the systolic blood pressure fluctuates more in acute stress experiments. However, when WGVI was 100%, the DBP recovery value of the subjects was negative. This suggests that high WGVI is not beneficial to stress recovery but may produce additional physiological stress.

3.2.2. HRV Analysis

Figure 9 shows the subjects’ HRV-nLF/nHF under different WGVI. At 0% WGVI, the HRV-nLF/nHF was 0.95. As WGVI increased, the HRV-nLF/nHF first decreased and then increased. When WGVI was at 25%, the HRV-nLF/nHF was the lowest (0.81), which was 14.74% lower than the group at 0% WGVI (p < 0.05). This suggests that subjects had lower stress at 25% WGVI. The HRV-nLF/nHF was 0.88, 0.96, and 0.97 at 50%, 75%, and 100% WGVI, respectively, and did not significantly differ from the group at 0% WGVI. This finding suggests that the level of mental stress experienced by students is the lowest when the WGVI is at 25%, which aligns with the observed SBP recovery as depicted in Figure 8a. Therefore, consideration can be given to controlling the green view index of the window view in a design to increase the comfort of people inside. The findings also indicate that the augmentation of WGVI has minimal impact on HRV. Thus, a single HRV metric alone is insufficient to determine the effect of window greenness on the human body; more complex feedback from physiological metrics is required.

3.2.3. EEG Analysis

There is a demonstrable positive association between the alpha frequency of EEG and relaxation. To more effectively assess the impact of different WGVI on the recovery of physiological indicators in subjects, this study tracked the cortical regions and channels exhibiting the highest levels of alpha frequency activity. EEG electrodes were positioned according to the 10–20 international standard [46]. In this experiment, the energy thermograms of the EEG signals from 24 subjects were analyzed at α frequencies (8 Hz, 9 Hz, 10 Hz, 11 Hz, 12 Hz) while they were exposed to five different WGVI scenes. The channel with the highest energy value in each heat map was recorded as the Ep points, yielding a total of 600 Ep points for the analysis.

Table 5. Statistics for the 24 participants’ active point (Ep) frequency. (With the red squares marking the channels with the highest frequencies and the number of frequencies.).

Frontal		Central		Occipital		Parietal		Left Temporal		Right Temporal
Chan.	Times	Chan.	Times	Chan.	Times	Chan.	Times	Chan.	Times	Chan.	Times
Fp1	76	CP2	8	O1	14	P7	16	TP9	8	FT10	38
Fp2	80	CP6	4	O2	96	P3	10	T7	2	T8	10
F3	10	C3	2	Oz	60	Pz	22			TP10	16
F4	6	Cz	8			P4	28
F8	26	C4	8			P8	20
Fz	2	FC5	8
		FC6	22

presents the distribution frequency of active channels across all subjects. The O2 channel located in the occipital lobe exhibited an active frequency of 96 times, rendering it the most active channel among all experimental channels in terms of alpha wave activity. The occipital region is predominantly linked to the processing of visual information, including color identification. It exhibits a heightened sensitivity towards the perception of green plant color. In addition to the O2 channel, the Fp1 channel and Fp2 channel in the frontal lobe and the Oz channel in the occipital lobe also maintained a high frequency of activity, which were 76, 80, and 60 times, respectively. This may be attributed to the frontal lobe’s association with positive emotions [47], while the occipital lobe is linked to arousal—both regions being critical in studying stress recovery.

The energy value of the alpha frequency domain can be obtained by preprocessing and the Fourier conversion of the original electrical signal collected by EEG equipment. Figure 10 shows the power spectral density curve and topographic map of the EEG O2 channel for all subjects under different WGVIs. Studies have shown that adults have noticeably less alpha event-related activity in the parietal and occipital areas of the scalp during periods of acute stress [48]. As shown in Figure 10, the power spectral density curve demonstrates that within the alpha frequency range (8–12 Hz), alpha waves with varying WGVI values reach their maximum power spectrum at approximately 10 Hz. Additionally, 25% WGVI has the highest alpha peak energy, followed by 50% and 75%, and 100% WGVI has the lowest alpha peak energy. The topographic map in Figure 10 further shows the alpha wave energy distribution at 10 Hz. When WGVI was 0%, the majority of the energy was concentrated in the frontal lobe, and the O2 channel in the occipital region was inactive. At 25% and 50% WGVI, the energy distribution shifted primarily to the occipital lobe, and the O2 channel became significantly more active. Conversely, 100% WGVI led to a negative energy value. These findings suggest that a moderate amount of natural landscape visible from a window can aid in stress recovery, with an optimal WGVI range between approximately 25% and 50%.

Figure 11 shows the average power spectral density (PSD) energy values of all subjects in the alpha frequency domain (8–12 Hz) of the O2 channel under different WGVI. At a significance level of 0.05, the overall mean revealed a significant difference (F = 20.18, p < 0.05). As shown in Figure 10, PSD-α initially decreased and then increased as WGVI increased. At 25% WGVI, the highest PSD-α value was 4.13 dB, representing a significant increase of 50.18% compared to the control group with 0% WGVI (2.75 dB). The second-highest PSD-α value was recorded at 50% WGVI (3.74 dB), indicating a 36.00% increase compared to the control. In contrast, when the WGVI reached 75% and 100%, the energy decreased by 14.18% and 30.18%, respectively, compared to the control group. This indicates a higher activity of alpha waves in the O2 channel between 25% and 50% WGVI, suggesting a more relaxed brain state in the subjects. In this interval, there exists a WGVI that causes PSD-α to reach its highest value. This may be because a high WGVI gives a sense of visual density. On the contrary, when the WGVI is 25–50%, the natural window scenery brings a kind of space extension and makes people feel comfortable.

Figure 12a illustrates the relaxation degree of all subjects under different WGVI. At the level of 0.05, the overall mean showed a significant difference (F = 6.00, p < 0.05). Compared to the control group with 0% WGVI (0.90), the relaxation degree increased by 0.19–1.25 when WGVI ranged from 25% to 100%. When WGVI was 25%, the relaxation was the highest; when WGVI was 50% and 75%, the relaxation was close. Figure 12b shows the mental stress of all subjects under different WGVIs, which is opposite to the relaxation degree in Figure 12a. At the level of 0.05, the overall mean showed a significant difference (F = 10.81, p < 0.05). Compared to the WGVI of 0% (2.08), WGVI of 25–100% reduced mental stress by 0.09–1.33, respectively. The lowest mental stress was observed when WGVI was 25%. According to the results of Figure 12, among the five kinds of WGVI set up in the experiment, 25% WGVI is most conducive to improving relaxation and reducing mental stress. In contrast, 100% WGVI has no significant effect on relieving stress. Similar results were found in the study of Nie et al. [20], where subjects experienced a significant decrease in EEG pleasure in an environment with a green vision rate of 90%. And the subjects subjectively showed stress. Nie et al. speculated that excessive greening may lead to negative emotions among viewers and affect stress recovery.

Table 5 shows the Pearson correlation analysis of alpha energy values in the O2 channel, relaxation degree, and mental stress of EEG and TMD of subjective evaluation, respectively. As shown in the table, the PSD-α in the O2 channel and relaxation degree were significantly negatively correlated with the TMD average score, while mental stress was significantly positively correlated with the TMD average score. Therefore, EEG can be used as a physiological indicator to judge the stress recovery of students.

The above results indicate that among the five scenarios tested, a WGVI of 25% yielded the best stress recovery outcomes, followed by 50%. In order to determine the best WGVI range for improving stress recovery, this study used multi-objective optimization for a further analysis. As shown in

Table 6. Analysis of correlation between the 24 subjects’ TMD and EEG rhythm. (*, p < 0.05; **, p < 0.01).

		PSD-α in O2 Channel	Relaxation Degree	Mental Stress
TMD	Pearson correlation	−0.237 *	−0.410 **	0.569 **
	Sig. (2-tailed)	0.035	0.001	0.009

, the relaxation degree and mental stress were selected as the representative databases of the pressure objective function. A polynomial regression analysis was performed to estimate the curve, and a functional model was used to evaluate the relationship between the relaxation degree, mental stress, and WGVI. The functional model is shown in Formulas (8) and (9), where x is WGVI and ranges from 0 to 100%. Figure 13 draws the curve of the function model.

$$\mathrm{Relaxation} \mathrm{degree}: y_1=0.92+2.59x_1-5.42x_1^2+3.02x_1^3; R^2=0.20$$

(8)

$$\mathrm{Mental} \mathrm{stress}: y_2=2.05-7.49x_2+12.91x_2^2-5.49x_2^3; R^2=0.38$$

(9)

As shown in Figure 13, when WGVI was 14.92–65.77%, the relaxation degree surpasses fatigue, making this range conducive to stress recovery. The highest degree of relaxation occurs between 29.13% and 37.04%, while the lowest level of mental stress is observed between 37.04% and 39.94%. Since the optimal range of WGVI for relaxation is different from the optimal range of WGVI for mental stress, the results are taken as an ensemble of thresholds for the relaxation degree and mental stress for accuracy. Consequently, from the perspective of balancing relaxation and stress, the optimal range for WGVI is between 29.13% and 39.94%.

4. Discussion

4.1. Contributions

GVI is an intuitive quantitative indicator of visual greening. Previously, many studies judged the urban natural level by the GVI [49,50], which was conducive to the construction of urban natural landscapes. However, the sole focus on the numerical value of GVI overlooks its impact on human well-being. This paper contends that the study of the green view index should be based on human welfare. Therefore, this study further proposed WGVI to study the effects of different degrees of natural views of windows on stress recovery of college students. The results confirmed previous findings that increasing visible nature in one’s view positively influences well-being. Moreover, this study identified the optimal WGVI range for restorative benefits.

Increasing visible nature in a view is conducive to stress recovery, and different WGVI results in different stress recovery benefits. Subjectively, subjects perceived themselves as experiencing less emotional disturbance when the WGVI was 25%. The analysis of BP and HRV indicated that stress recovery was the most pronounced at 25% WGVI, with significant SBP recovery and a lower HRV-nLF/nHF ratio, both indicating better stress recovery. These results, when compared with a control group with 0% WGVI, suggest that some exposure to nature is beneficial, but the effect saturates quickly. Beyond a certain point, increasing WGVI further does not significantly enhance stress recovery. In the EEG analysis, the alpha activity of the O2 channel was significantly increased at 25–50% WGVI. This increase indicates enhanced activity in the occipital lobe, higher alpha power, greater relaxation, and reduced mental stress. And the findings indicated that WGVI involved more relaxation in the range of 25% to 50%.

Stress reduction theory (SRT) provides an explanation for the impact of nature experience on affect. This theory posits that natural environments have a restorative advantage over artificial environments due to the role that they played in our evolution as a species. According to the stress reduction theory [51], viewing natural scenes can reduce stress. Moreover, according to the EEG interpretation of human emotions, the occipital lobe controls the emotional reaction and green visual processing [52], and alpha represents positive emotions [53]. A well-designed landscape is often linked to lower anxiety levels [54]. Previous studies have also found that people feel better physically and mentally if more than 25% of their view is green. This is because the year with the most green space in the previous study was 2002, and only 16.49% of the study area had more than 25% GVI. Therefore, we need to explore the optimal range of WGVIs that are beneficial to people’s health, and it is also critical to provide potential strategies for improving green landscapes for those who are currently unable to enjoy them. Polynomial fitting further determined that the WGVI range of 29.13% to 39.94% could better improve stress recovery, considering the balance between the relaxation degree and mental stress. This result was in agreement with the research of some scholars. For example, earlier research in Japan suggested that when the GVI exceeds 30%, people will have a good impression of greening, and when the GVI exceeds 50%, people will feel that the degree of greening is better [55]. Jiang et al. [56] found that the fastest human stress recovery occurred when GVI reached 24%~34%, with 40% being considered relatively reasonable.

In contrast to previous studies, this research refined the classification of the green view index by specifically examining the impact of WGVI on human stress recovery. This offers a new perspective for architecture and landscape design aimed at enhancing well-being. Based on these findings, we recommend that architects integrate WGVI considerations into their building designs. An appropriate natural landscape layout can be tailored to various settings, optimizing stress recovery and mental health benefits. For instance, techniques like “borrowed scenery” could be used to incorporate green views into window designs, enhancing the visual connection between interior spaces and nature. Achieving symmetry and balance, as well as carefully managing proportion, scale, contrast, and harmony, is essential in this design process. By fostering a sense of unity between natural landscapes and indoor spaces, architects can ensure that nature and architecture form a cohesive, complementary environment that enhances well-being. In addition, the WGVI is usually higher for low-rise buildings because high-rise and high-density buildings shade urban green spaces, making the WGVI lower. From this perspective, integrating green walls and green roofs in limited urban spaces can effectively improve the green landscape. Green walls are usually systems that are green vertical surfaces, while green roofs are urban green spaces that are separate from the ground. In highly urbanized areas, the provision of green walls and green roofs is necessary to increase green space and enhance the WGVI of high floors. This study provides a novel human-centered perspective on landscape design from a practical standpoint. In addition, it encourages architects to consider the relationship between window size and outdoor greening when designing building facades.

4.2. Limitations and Future Research

In this paper, VR technology and EEG are combined to explore the effects of WGVI on stress recovery, which contributes to landscape construction and the formation of multidisciplinary research. The utilization of VR technology helps eliminate potential confounding factors in the experiments, such as external illumination, which could influence the experimental outcomes. This study demonstrated that WGVI could change the state of the brain waves of the subjects. Between 29.13% and 39.94% of WGVI can induce relaxation-related brain wave patterns. Nevertheless, this study has some limitations. Firstly, this study had no control cases and lacked resolution across WGVI. As we all know, window views can vary significantly in content. In EN17037, it is considered that the window quality should include at least three layers, including the ground, city or landscape, and sky [29,57]. To simplify variable control, this study focused on just two layers—natural greenery and the sky—omitting a deeper exploration of the relationship between architecture and nature. Future research will need to incorporate this relationship. In addition, to verify the accuracy of the current study, in the future, we will apply the results of this study to projects and conduct experiments through the actual planting and planning of green visibility to further verify the accuracy of the current results. Secondly, this study did not consider the relationship between different floors and WGVI. In addition, we have not taken into account factors such as the height of the building, its proximity to other buildings, and the effect of light, all of which will affect the green visibility of windows. In fact, the WGVI differed across different floors of a building. Future work will also consider the influence of other factors to supplement the research model and refine the findings. To further optimize the scope of WGVI research, the distinctions in stress recovery of WGVI between various floors should be studied in the future. Thirdly, the subjects consisted solely of college students, chosen for feasibility and exploratory purposes; thus, the results may only be representative of this particular age group. Moreover, these data were measured during the COVID-19 pandemic, a period during which students’ physical and mental anxiety increased and the WGVI’s alleviation of stress resilience was more pronounced [58]. Future research should include a more diverse sample, considering variations in age and gender.

5. Conclusions

Increasing visible nature in a view positively impacts the stress recovery of college students, and different WGVIs yield varying effects on stress recovery. This study determined the optimal setting range of WGVI conducive to stress recovery to guide the design of urban natural landscapes. The specific conclusions are as follows:

(1)

After the stress test, the TMD of the subjects who watched the natural scenery outside the window was significantly lower than those without any natural scenery outside their window. When WGVI was 25%, the subjects’ TMD was the lowest (−4.5). This indicated that the natural scenery outside the window effectively reduced the negative emotions and enhanced the positive emotion in the subjects. The optimal effect occurred when the WGVI was 25%.

(2)

The SBP recovery value of the subjects was the maximum, 4.60 mmHg, when WGVI was 25%. However, the DBP recovery value had no significant difference under different WGVI. In addition, the subjects’ SNS activity was weaker when the WGVI was 25%, while the PNS activity was enhanced. The HRV-nLF/nHF was 0.07–0.16 lower than that of the other four groups. BP and HRV consistently showed that stress recovery was best when WGVI was at 25%.

(3)

The O2 channel in the occipital region was found to be the most active channel with an α frequency by the EEG test. All the α-waves in the five scenes reach the peak power spectrum at approximately 10 Hz. The PSD-α values of 25% and 50% WGVI were 4.13 μv² and 3.74 μv², respectively, significantly higher than those of other groups. Compared to non-natural scenery outside a window, the subjects with the natural scenery experienced an increase in relaxation levels ranging from 0.19 to 1.25, and a decrease in mental stress ranging from 0.09 to 1.33. The Pearson correlation analysis demonstrated a substantial negative correlation between the PSD-α and relaxation degree, as well as TMD. Additionally, there was a strong positive link between mental stress and TMD. Therefore, EEG indicators can be used as physiological markers to assess students’ stress levels, providing a state-of-the-art approach for future research.

(4)

The polynomial fitting analysis of the relaxation degree and mental stress showed that the optimal interval of WGVI was 29.13–39.94%. Within this range, the greatest stress relief and improved stress recovery were observed, resulting in enhanced physical and mental well-being.