Virtual nature, psychological and psychophysiological outcomes: A systematic review
虛擬自然、心理及心理生理結果：系統性回顧

Handling Editor: L. McCunn
處理編輯：L. McCunn

Virtual reality 虛擬現實
Psychological outcome 心理結果
Psychophysiological outcome
心理生理結果
Affect 影響
Stress 壓力
Restorativeness 恢復性

The psychological benefits that nature offers to humans are undisputed. Several decades of research attest to how exposure to outdoor natural environments can improve mental and physical health (e.g., Bratman, Daily, Levy, & Gross, 2015; Carrus et al., 2015; Coventrye et al., 2021; Giannico et al., 2021; McMahan & Estes, 2015; Yao, Zhang, & Gong, 2021). According to the biophilia hypothesis, people have an innate need for affiliation with nature since it is related to our evolution (Wilson, 1984), while for both Stress Reduction Theory (SRT; Ulrich, 1983) and Attention Restoration Theory (ART; Kaplan & Kaplan, 1989), nature is restorative, in terms of affect and cognitive recovery from stress and mental fatigue.
自然對人類所提供的心理益處是無可爭辯的。數十年的研究證實，接觸戶外自然環境可以改善心理和身體健康（例如，Bratman、Daily、Levy 和 Gross，2015；Carrus 等，2015；Coventrye 等，2021；Giannico 等，2021；McMahan 和 Estes，2015；Yao、Zhang 和 Gong，2021）。根據生物親和假說，人類對自然的依附有著與我們進化相關的內在需求（Wilson，1984），而根據壓力減少理論（SRT；Ulrich，1983）和注意力恢復理論（ART；Kaplan 和 Kaplan，1989），自然在情感和認知上對於從壓力和心理疲勞中恢復具有修復作用。

However, under some circumstances, it can be difficult to access nature. For instance, several studies point to the importance of exposure to nature for populations with special needs, such as hospitalized patients (Nejati, Rodiek, & Shepley, 2016) and prisoners (Li, Zhang, et al., 2021). Some evidence already shows that even non-direct experience of nature, such as exposure to nature views, images, or videos can have temporary positive health and cognitive performance outcomes (Bratman, Hamilton, & Daily, 2012; Mcsweeney, Rainham, Johnson, Sherry, & Singleton, 2014; Pasca et al., 2021). However, as compared to brief changes in people's mental states, what is desirable for psychological intervention are long-term changes.
然而，在某些情況下，接觸自然可能會很困難。例如，幾項研究指出，對於有特殊需求的人群，如住院病人（Nejati, Rodiek, & Shepley, 2016）和囚犯（Li, Zhang, et al., 2021），接觸自然的重要性。一些證據已經顯示，即使是非直接的自然體驗，例如接觸自然景觀、圖像或視頻，也能帶來暫時的健康和認知表現的正面效果（Bratman, Hamilton, & Daily, 2012；Mcsweeney, Rainham, Johnson, Sherry, & Singleton, 2014；Pasca et al., 2021）。然而，與人們心理狀態的短暫變化相比，心理干預所期望的是長期的變化。

Virtual reality (VR) is defined as the simulation of environments in 3D, accessible often through special equipment such as head-mounted displays but also 2D screens, that provide the user with a sense of reality (Jayaram, Connacher, & Lyons, 1997). The immersive experience
虛擬實境（VR）被定義為三維環境的模擬，通常通過特殊設備如頭戴顯示器以及二維螢幕來訪問，為用戶提供現實感（Jayaram, Connacher, & Lyons, 1997）。沉浸式體驗

may be the most important characteristic for the associated psychological benefits, such as improved mood and perceived restorativeness ( Li , Dong, et al., 2021; Liszio, Graf, & Masuch, 2018). A greater sense of experienced presence is what differentiates virtual reality from 2D photos and videos (Yeo et al., 2020). The fruition of virtual nature (i.e., exposure to natural environments via VR) can occur through computer-generated environments or panoramic photos and videos of real natural environments. While the former allows for interactivity, the latter is also associated with benefits through the sense of reality and vividness that may evoke (Yeo et al., 2020). Natural environments used are more often green (e.g., forests and urban parks) and blue (e.g., beaches and rivers' shores). Sometimes the VR experience may be combined with auditory and olfactory stimuli (e.g., Hedblom et al., 2019).
可能是與心理益處相關的最重要特徵，例如改善情緒和感知的恢復性（Li, Dong 等，2021；Liszio, Graf, & Masuch，2018）。更強烈的存在感是虛擬現實與 2D 照片和視頻之間的區別所在（Yeo 等，2020）。虛擬自然的實現（即通過虛擬現實接觸自然環境）可以通過計算機生成的環境或真實自然環境的全景照片和視頻來實現。雖然前者允許互動，但後者也因其可能喚起的現實感和生動性而與益處相關（Yeo 等，2020）。使用的自然環境通常是綠色的（例如，森林和城市公園）和藍色的（例如，海灘和河岸）。有時虛擬現實體驗可能會結合聽覺和嗅覺刺激（例如，Hedblom 等，2019）。

In recent years, empirical evidence revolving around the impact of exposure to VR is growing. Especially since the outbreak of the Covid-19 pandemic, the uses of VR for health (Mantovani et al., 2020; Singh et al., 2020) seems to have drawn even greater attention (Ball, Huang, & Francis, 2021). In addition, people's relationship with nature was shaped by the pandemic and the lockdowns. Indeed, some authors noted that spending prolonged time indoors due to containment measures can cause psychological distance from natural environments, especially for the youngest (Honey-Rosés et al., 2020). Indoor exposure to nature through VR may mitigate the impact of home confinement if it becomes necessary again due to future pandemics or other causes (Spano et al., 2021; Theodorou et al., 2021). For instance, influential companies such as Meta are investing in these technologies, facilitating a decrease in the prices of VR devices, and indicating that there could be significant dissemination of this medium in the near future (Kraus, Kanbach, Krysta, Steinhoff, & Tomini, 2022).
近年來，關於虛擬實境（VR）接觸影響的實證證據不斷增加。特別是自新冠疫情爆發以來，VR 在健康領域的應用（Mantovani et al., 2020; Singh et al., 2020）似乎引起了更大的關注（Ball, Huang, & Francis, 2021）。此外，人們與自然的關係也受到疫情和封鎖措施的影響。事實上，一些作者指出，由於防控措施而長時間待在室內可能會導致與自然環境的心理距離，尤其是對於年輕人（Honey-Rosés et al., 2020）。透過 VR 在室內接觸自然可能會減輕因未來疫情或其他原因再次需要居家隔離的影響（Spano et al., 2021; Theodorou et al., 2021）。例如，像 Meta 這樣的影響力公司正在投資這些技術，促進 VR 設備價格的下降，並表明在不久的將來這一媒介可能會有顯著的普及（Kraus, Kanbach, Krysta, Steinhoff, & Tomini, 2022）。

Despite the growing interest in the application of virtual nature, there has been as yet limited effort to gather and systematically summarize the available studies (Browning, Shipley, et al., 2020; Li, Zhang, et al., 2021; Nukarinen et al., 2022; White et al., 2018), especially regarding psychological outcomes not strictly related to well-being (Frost et al., 2022; Riches, Azevedo, Bird, Pisani, & Valmaggia, 2021).
儘管對虛擬自然應用的興趣日益增長，但迄今為止，對可用研究的收集和系統性總結的努力仍然有限（Browning, Shipley 等，2020；Li, Zhang 等，2021；Nukarinen 等，2022；White 等，2018），特別是在與心理結果相關的方面，這些結果並不僅僅與福祉有關（Frost 等，2022；Riches, Azevedo, Bird, Pisani, & Valmaggia，2021）。

The aim of the current study was to systematically review the available studies on the psychological and psychophysiological benefits of virtual nature, including but not limited to well-being. In particular, first, we critically reviewed studies that used virtual exposure to different kinds of natural environments and that measured effects on different psychological and psychophysiological variables, focusing on instruments and methods (e.g., type of exposure, type of natural environment). The basic idea was to synthesize the current state of work and provide a broad picture of the available findings that could be inclusive as much as possible in terms of outcomes and methods. We believe that this is an added value as compared to recently published systematic reviews (e.g., Frost et al., 2022; Riches et al., 2021). Second, we identified potential intervening variables in the relationship between virtual nature experience and psychological and psychophysiological outcomes, namely relevant moderators and mediators. Moderators can inform about specific populations or conditions that may hinder or strengthen the observed effects, while mediators may suggest relevant mechanisms and help develop valuable theories.
本研究的目的是系統性地回顧現有的有關虛擬自然的心理和心理生理益處的研究，包括但不限於幸福感。具體而言，我們首先批判性地回顧了使用虛擬暴露於不同類型自然環境的研究，並測量對不同心理和心理生理變數的影響，重點關注工具和方法（例如，暴露類型、自然環境類型）。基本思想是綜合當前的研究狀態，提供一個盡可能包羅萬象的可用發現的廣泛圖景，無論是在結果還是方法上。我們相信，這相較於最近發表的系統性回顧（例如，Frost et al., 2022; Riches et al., 2021）具有附加價值。其次，我們確定了虛擬自然體驗與心理和心理生理結果之間關係中的潛在干預變數，即相關的調節變數和中介變數。 版主可以告知特定的人群或情況，這些可能會妨礙或加強觀察到的效果，而中介則可以提出相關的機制並幫助發展有價值的理論。

In accordance with the EQUATOR Reporting Guideline Decision Tree (Simera et al., 2010), the structure of this paper was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Moher, Liberati, Tetzlaff, Altman, & Prisma Group., 2009). The PRISMA statement ensures evidence-based reporting of systematic reviews according to a checklist and a flow diagram describing the process of identification of the final set of selected studies.
根據 EQUATOR 報告指導方針決策樹（Simera 等，2010），本文的結構受到系統評價和 Meta 分析的首選報告項目（PRISMA）指導方針（Moher, Liberati, Tetzlaff, Altman, & Prisma Group，2009）的指導。PRISMA 聲明確保根據檢查清單和描述最終選定研究識別過程的流程圖，對系統評價進行基於證據的報告。

The principal database used for bibliographic searches was Scopus. Since our intent was to ensure complete coverage of psychological and psychophysiological outcomes, we started our search by indicating only the exposure, thus not including outcomes and measures in the search query. A search query was created based on general, basic keywords on the tool of interest (i.e., "virtual reality" OR "immersive virtual reality" OR oculus OR viewer) and types of exposure to nature ("virtual nature" OR "greenspace" OR "green space" OR "blue space" OR "bluespace" OR "white space" OR "whitespace" OR "natural environment" OR "natural virtual environment" OR "exposure to nature" OR "exposure with nature" OR "contact with nature" OR "contact to nature" OR "proximity to nature" OR "urban nature"). No limits regarding the publication year were set. After the first step of selection on Scopus, we checked for possible further publication records on Web of Science and PsycINFO. Since only duplicates of records already retrieved on Scopus were found, we continued the selection process for eligibility via this platform. This phase of study selection was completed on March 16, 2023. In Fig. 1, the number of studies considered and their distribution per year is graphically represented. As can be seen, there is evidence of growing interest over the years, from 1996 to 2022.
主要用於文獻檢索的數據庫是 Scopus。由於我們的目的是確保心理和心理生理結果的全面覆蓋，我們的搜索從僅指明暴露開始，因此在搜索查詢中不包括結果和測量。根據感興趣的工具（即「虛擬現實」或「沉浸式虛擬現實」或 oculus 或觀眾）和接觸自然的類型（「虛擬自然」或「綠地」或「綠色空間」或「藍色空間」或「藍色空間」或「白色空間」或「白色空間」或「自然環境」或「自然虛擬環境」或「接觸自然」或「與自然接觸」或「接觸自然」或「接觸自然」或「接近自然」或「城市自然」）創建了一個搜索查詢。未設置有關出版年份的限制。在 Scopus 的第一步選擇後，我們檢查了 Web of Science 和 PsycINFO 上可能的進一步出版記錄。由於僅發現了在 Scopus 上已檢索到的記錄的重複項，因此我們繼續通過此平台進行合格的選擇過程。這一階段的研究選擇於 2023 年 3 月 16 日完成。 1，所考慮的研究數量及其每年的分佈以圖形方式表示。如圖所示，從 1996 年到 2022 年，顯示出對此的興趣逐年增長的證據。

English peer-reviewed articles published in scientific journals were considered eligible; thus, gray literature was not considered. We excluded (a) conference papers, since they were not available in their full-text versions and/or the reported contents had also been published in the form of an article, (b) reviews, (c) conference reviews, (d) book chapters, since full texts were not available, (e) notes, (f) books, and (g) retracted papers. A further selection step was based on the subject area. Since we were interested in psychological and psychophysiological outcomes, we excluded (h) articles published in journals classified in the areas of mathematics, physics and astronomy, chemistry, and material science, i.e., areas in which these outcomes are typically not contemplated. Table S1 shows the complete syntax of the search query used in the Scopus database.
英語的同行評審文章發表在科學期刊中被視為合格；因此，灰色文獻不被考慮。我們排除了 (a) 會議論文，因為它們的全文版本不可用和/或報告的內容也已以文章形式發表，(b) 評論，(c) 會議評論，(d) 書籍章節，因為全文不可用，(e) 註釋，(f) 書籍，以及 (g) 撤回的論文。進一步的選擇步驟是基於主題領域。由於我們對心理和心理生理結果感興趣，我們排除了 (h) 發表在數學、物理和天文學、化學和材料科學等領域的期刊中的文章，即這些結果通常不被考慮的領域。表 S1 顯示了在 Scopus 數據庫中使用的搜索查詢的完整語法。

Records selected were screened for eligibility in two further steps. First, they were assessed by title, abstract, and keywords, in order to make sure that they met all the exclusion and inclusion criteria, i.e., focus on psychological and/or psychophysiological outcomes, human subjects, and exposure to nature via VR. Psychological and psychophysiological outcomes were defined as outcomes of interest to the psychological field, in that their investigation: 1) would be helpful to describe how people behave; 2) help understand why people behave in a certain way; 3) help predict how people will behave in certain situations; 4) help develop ways to intervene to alter potentially nocive behaviors and situations (see Plotnik & Kouyoumdjian, 2013, p. 4). Identified eligible records were extracted and their full texts were downloaded. As a final check, the last step for eligibility consisted in reading the full texts and checking for the aforementioned inclusion/exclusion criteria. The process of article selection was conducted by two independent reviewers (GS and AT). On five occasions, a third independent rater (AP) was employed to solve disagreements since the majority rate was considered. A total of 59 articles were included in this systematic review. The entire process of the search strategy and article selection is displayed in Fig. 2.
所選記錄經過兩個進一步步驟篩選以確定其資格。首先，通過標題、摘要和關鍵詞進行評估，以確保它們符合所有排除和納入標準，即專注於心理和/或心理生理結果、人類受試者，以及通過虛擬現實接觸自然。心理和心理生理結果被定義為對心理學領域感興趣的結果，其研究：1）有助於描述人們的行為；2）有助於理解人們為何以某種方式行為；3）有助於預測人們在特定情況下的行為；4）有助於開發干預方法以改變潛在有害的行為和情況（見 Plotnik & Kouyoumdjian, 2013, p. 4）。已識別的合格記錄被提取並下載其全文。作為最後檢查，資格的最後一步是閱讀全文並檢查上述納入/排除標準。文章選擇過程由兩位獨立評審（GS 和 AT）進行。 在五次情況下，聘請了第三位獨立評估者（AP）來解決分歧，因為考慮了大多數評分。這次系統性回顧共納入了 59 篇文章。搜索策略和文章選擇的整個過程顯示在圖 2 中。

The set of selected articles includes mainly studies with experimental
所選文章的集合主要包括實驗研究

Fig. 1. Number of scientific articles on virtual nature per year (source: Scopus).
圖 1. 每年有關虛擬自然的科學文章數量（來源：Scopus）。
Note. The year 2023 has not been considered since it is not yet finished at the time of the search.
注意。由於在搜尋時 2023 年尚未結束，因此未考慮該年。
and quasi-experimental research designs, but also randomized controlled trials, and interventions. To our knowledge, no available checklist for simultaneously assessing each of the aforementioned research designs was available, and using three different checklists to evaluate studies among the same set might not lead to consistent selections. To overcome this issue, we created a new checklist for quality assessment based on two available checklists of The Joanna Briggs Institute's critical appraisal tools for use in JBI Systematic Reviews (Tufanaru, Munn, Aromataris, Campbell, & Hopp, 2020), i.e., the Checklist for Quasi-Experimental Studies (Non-Randomized Experimental Studies) and the Checklist for Randomized Controlled Trials. Our checklist consisted of eight items scored on (1) research design, (2) control group, (3) comparability between samples, (4) VR task, (5) measures, (6) statistical analysis, (7) appropriateness of results, and (8) presence of power analysis for sample size calculation (Table S2). Each item was scored with 1 (Yes/addressed/clear), 0.5 (Partially addressed/partially clear), or 0 (No/not addressed/not clear). Quality assessment was conducted by two independent reviewers (GS and AT). The interrater reliability as assessed with the Cohen's Kappa coefficient for the 8 items of the quality assessment ranged from 1 ( ; i.e., perfect agreement, in case of items , and 8 ) to 0.667 ( .001) for item 1. In case of disagreement, the decision of another reviewer (AP) was applied. The maximum score for quality assessment was 8.
和準實驗研究設計，但也包括隨機對照試驗和干預措施。據我們所知，尚未有可用的檢查表來同時評估上述每一種研究設計，使用三個不同的檢查表來評估同一組研究可能不會導致一致的選擇。為了解決這個問題，我們根據兩個可用的檢查表創建了一個新的質量評估檢查表，這些檢查表來自喬安娜·布里格斯研究所的關鍵評估工具，用於 JBI 系統評價（Tufanaru, Munn, Aromataris, Campbell, & Hopp, 2020），即準實驗研究檢查表（非隨機實驗研究）和隨機對照試驗檢查表。我們的檢查表由八個項目組成，評分標準為（1）研究設計，（2）對照組，（3）樣本之間的可比性，（4）虛擬現實任務，（5）測量，（6）統計分析，（7）結果的適當性，以及（8）樣本大小計算的效能分析存在性（表 S2）。每個項目的評分為 1（是/已解決/清楚）、0.5（部分解決/部分清楚）或 0（否/未解決/不清楚）。 質量評估由兩位獨立評審（GS 和 AT）進行。使用 Cohen's Kappa 係數評估的質量評估 8 項目的評分者間可靠性範圍從 1（ ；即完美一致，對於項目 和 8）到 0.667（ .001）對於項目 1。在意見不一致的情況下，採用另一位評審（AP）的決定。質量評估的最高分為 8。

Fifty-nine studies were included in this systematic review (Table S3). The studies taken into consideration were distributed among Europe, Asia, and North America. Specifically, twenty-four were conducted in Europe (10 in Germany, 2 in Norway, 2 in Sweden, 3 in The Netherlands, 1 in Finland, 1 in France, 1 in Ireland, 1 in Italy, and 3 in the United Kingdom), eighteen in Asia (12 in China, 4 in Taiwan, and 2 in Singapore), twelve in North America (10 in the United States and 2 in Canada), and four in Australia. One study was conducted on a mixed sample composed of people living in the United States, Taiwan, or Thailand for at least 5 years prior to the study.
本系統性回顧納入了五十九項研究（表 S3）。考慮的研究分佈於歐洲、亞洲和北美。具體而言，二十四項在歐洲進行（德國 10 項、挪威 2 項、瑞典 2 項、荷蘭 3 項、芬蘭 1 項、法國 1 項、愛爾蘭 1 項、義大利 1 項和英國 3 項），十八項在亞洲（中國 12 項、台灣 4 項和新加坡 2 項），十二項在北美（美國 10 項和加拿大 2 項），以及四項在澳大利亞。有一項研究針對一個混合樣本進行，該樣本由在美國、台灣或泰國居住至少 5 年的人士組成。

The majority of studies (forty-nine) were conducted with healthy adult samples, of which thirteen involved university students and three older adults. Four studies selected participants based on their anxiety levels, either selecting only those with high levels of anxiety and/or depression (Li, Dong, et al., 2021; Shu, Wu, & Zhai, 2022; Wang, Sit, Tang, & Tsai, 2020) or just comparing VR benefits in participants with high vs. low levels of anxiety (O'meara, Cassarino, Bolger, & Setti, 2020; Tanja-Dijkstra et al., 2014). Another study focused on individuals with a low level of connectedness to nature (Leung, Hazan, & Chan, 2022). Two studies focused on individuals with acquired physical disabilities (Je & Lee, 2020; Lakhani et al., 2020), one on esophageal and gastrointestinal cancer patients (Song et al., 2022), one on a sample of pregnant women (Sun et al., 2023), one on patients undergoing hemodialysis (Hsieh & Li, 2022), one on dental patients (Tanja-Dijkstra et al., 2018). Special samples considered were remote workers (Ch et al., 2023). Taking into consideration all articles, the minimum sample size of participants per study is 14 while the maximum was 1280 . Sample size varied based on the number of conditions in the study design.
大多數研究（四十九項）是針對健康成人樣本進行的，其中十三項涉及大學生，三項涉及老年人。四項研究根據參與者的焦慮水平選擇參與者，或僅選擇那些具有高水平焦慮和/或抑鬱的人（Li, Dong, et al., 2021; Shu, Wu, & Zhai, 2022; Wang, Sit, Tang, & Tsai, 2020），或僅比較高焦慮與低焦慮參與者的虛擬現實效益（O'meara, Cassarino, Bolger, & Setti, 2020; Tanja-Dijkstra et al., 2014）。另一項研究專注於與自然連結感較低的個體（Leung, Hazan, & Chan, 2022）。兩項研究專注於獲得性身體殘疾的個體（Je & Lee, 2020; Lakhani et al., 2020），一項針對食道和胃腸癌患者（Song et al., 2022），一項針對孕婦樣本（Sun et al., 2023），一項針對接受血液透析的患者（Hsieh & Li, 2022），一項針對牙科患者（Tanja-Dijkstra et al., 2018）。特別考慮的樣本包括遠程工作者（Ch et al., 2023）。考慮到所有文章，每項研究的參與者最小樣本量為 14，最大樣本量為 1280。 樣本大小根據研究設計中的條件數量而有所不同。

The total score of the quality assessment checklist was obtained by summing the scores of each item of the checklist. We categorized the studies into four categories: "very high quality" (score ), "high quality" (score from 5 to 6.5)"medium quality" (score from 3 to 4.5), and "low quality" (score ) (Table S4). Sixteen studies were classified as having a very high quality, a further thirty-nine studies as having "high quality". Three studies were classified as having "medium quality" and only one as having "low quality". The item on which we found the lowest average score is that of power analysis. Only thirteen studies reported having carried out an a priori power analysis. On the other hand, all of the included studies described clear and complete study results. The VR tasks used and the statistical analyses performed were also found to be highly satisfactory (quality criteria numbers 4 and 6 in Table S2). Table S4 shows the scores and judgment for the quality assessment for each study considered.
質量評估檢查表的總分是通過將檢查表中每個項目的分數相加而得出的。我們將研究分為四個類別：“非常高質量”（分數 ）、“高質量”（分數從 5 到 6.5）、“中等質量”（分數從 3 到 4.5）和“低質量”（分數 ）（表 S4）。十六項研究被分類為非常高質量，另外三十九項研究被分類為高質量。三項研究被分類為中等質量，只有一項被分類為低質量。我們發現平均分數最低的項目是功效分析。只有十三項研究報告進行了事前功效分析。另一方面，所有納入的研究都描述了清晰且完整的研究結果。所使用的虛擬現實任務和進行的統計分析也被認為是非常滿意的（表 S2 中的質量標準編號 4 和 6）。表 S4 顯示了每項研究的質量評估分數和判斷。

Findings are reported in a detailed way and organized by type of virtual exposure and outcome investigated: Each section can be read independently by readers interested in specific results. The current section is structured as follows. We first report findings compared by type of virtual exposure (i.e., virtual nature vs. on-site exposure, virtual nature vs. virtual urban environment, and virtual nature vs. virtual indoor experience). We then present results divided by outcomes investigated. We found that psychophysiological outcomes were only used to investigate physiological stress, thus they are reported in this section. Outcomes are listed by number of evidence available (i.e., mood, affect, and emotion, perceived stress and relaxation, physiological stress, restorativeness, environment preference and pleasantness, cognitive performance, nature connectedness, behavior and behavioral intentions, creativity, perceived safety, subjective vitality, and dental experience). Next, we focus on findings on the intervening variables and, lastly, we report results on the variables related to the VR experience.
研究結果以詳細的方式報告，並根據虛擬暴露類型和調查結果進行組織：每個部分都可以獨立閱讀，供對特定結果感興趣的讀者參考。本節的結構如下。我們首先報告按虛擬暴露類型比較的結果（即虛擬自然與現場暴露、虛擬自然與虛擬城市環境、以及虛擬自然與虛擬室內體驗）。然後，我們呈現按調查結果劃分的結果。我們發現心理生理結果僅用於調查生理壓力，因此在本節中報告。結果按可用證據的數量列出（即情緒、情感和情緒、感知壓力和放鬆、生理壓力、恢復性、環境偏好和愉悅感、認知表現、自然連結、行為和行為意圖、創造力、感知安全性、主觀活力和牙科體驗）。接下來，我們專注於干預變數的發現，最後，我們報告與虛擬現實體驗相關的變數結果。

Fig. 2. Systematic review flow chart detailing the literature search, number of abstracts screened, and full texts retrieved.
圖 2. 系統性回顧流程圖，詳細說明文獻搜尋、篩選的摘要數量及檢索的全文。

All studies featured in this review had at least one virtual nature exposure condition. Types and conditions of exposure varied significantly among the selected studies. Almost all studies included green environments e.g., forests, urban green spaces, parks, courtyards, and gardens. Blue environments were beaches, rivers, lakes, waterfalls, and underwater environments such as a tropical coral reef. Most of them combined virtual nature with the use of auditory and even olfactory stimuli such as bird and water sounds and garden smell. Among others, only one study (Lakhani et al., 2020) exposed participants to a white (polar) environment, and one other (Yin, Bratman, Browning, Spengler,
所有在本次回顧中提到的研究至少有一種虛擬自然暴露條件。所選研究中的暴露類型和條件差異顯著。幾乎所有研究都包括綠色環境，例如森林、城市綠地、公園、庭院和花園。藍色環境則包括海灘、河流、湖泊、瀑布以及熱帶珊瑚礁等水下環境。大多數研究將虛擬自然與聽覺甚至嗅覺刺激結合，例如鳥鳴聲、水聲和花園的氣味。在其他研究中，只有一項研究（Lakhani et al., 2020）讓參與者接觸到白色（極地）環境，另一項研究（Yin, Bratman, Browning, Spengler）則未提及。
& Olvera-Alvarez, 2022) to a brown (desert) environment. The length of a single exposure was characterized by a wide variability, which lasted from a minimum of 26 s to a maximum of 45 min . The majority of studies used footage of real environments (32, 54.24%), while another consistent part used computer-generated images (23, 38.98%). Only one study (1.69%) digitally manipulated shootings of real environments and one study (1.69%) had both real and computer-generated images in their study design. Lastly, two studies (3.40%) did not report this information.
& Olvera-Alvarez, 2022) 到一個棕色（沙漠）環境。單次曝光的長度特徵顯示出廣泛的變異性，持續時間從最少 26 秒到最多 45 分鐘。大多數研究使用了真實環境的影像（32，54.24%），而另一部分則使用了電腦生成的影像（23，38.98%）。只有一項研究（1.69%）對真實環境的拍攝進行了數位操作，還有一項研究（1.69%）在其研究設計中同時包含了真實和電腦生成的影像。最後，兩項研究（3.40%）未報告此信息。

Eighteen studies (Blum, Rockstroh, & Göritz, 2019; Evensen et al., 2021; Gao et al., 2019; Huang, Yang, Jane, Li, & Bauer, 2020; Je & Lee, 2020; Lakhani et al., 2020; Lindquist, Maxim, Proctor, & Dolins, 2020;
十八項研究（Blum, Rockstroh, & Göritz, 2019；Evensen et al., 2021；Gao et al., 2019；Huang, Yang, Jane, Li, & Bauer, 2020；Je & Lee, 2020；Lakhani et al., 2020；Lindquist, Maxim, Proctor, & Dolins, 2020；

Liszio et al., 2018; Liszio & Masuch, 2019; Mattila et al., 2020; Pals, Steg, Dontje, Siero, & van Der Zee, 2014; Reese, Kohler, & Menzel, 2021; Rockstroh, Blum, & Göritz, 2019; Shu et al., 2022; Song et al., 2022; van Vliet et al., 2021; Wang, Shi, Zhang, & Chiang, 2019; Yeo et al., 2020) explored the effect of a number of natural environments reproducing different types of stimuli. Such stimuli might consist of images, 2D videos, immersive VR through headsets, shutter glasses, or Cave VE system, or through the manipulation of certain features of the same condition in virtual nature (e.g., interactive vs. noninteractive virtual game, as in Liszio & Masuch, 2019). Overall, the immersive virtual experience was found to be more effective than other media (e.g., Knaust et al., 2022); however, the wide variability of exposure (visual, olfactory, and auditory stimuli, and a variety of features embedded into natural environments), media, and outcomes considered, led to a remarkable variability of differentiated findings (see Table S3). As an example, Gao, Zhang, Zhu, Gao, and Qiu (2019) exposed participants to six different virtual environments, namely gray space (also defined as "built environment"), blue space, open green space, partly open green space, partly closed green space, and closed green space through panoramic photos with a head-mounted display. Authors reported a general positive effect of the six landscapes on attentional fatigue and negative mood, nevertheless, more specifically, partly open space seemed to provide a higher positive effect on mood, while blue and partly closed spaces were identified as the most preferred for recreational purposes. At the same time, a courtyard with grass and trees was demonstrated to induce a lower level of stress in previously stressed participants compared to a courtyard without vegetation (Huang et al., 2020). Shu et al. (2022), comparing different landscape structures, found that different permeability levels and viewing distance have a crucial role both in landscape preference and in anxiety and depression levels.
Liszio 等人（2018）；Liszio & Masuch（2019）；Mattila 等人（2020）；Pals、Steg、Dontje、Siero 和 van Der Zee（2014）；Reese、Kohler 和 Menzel（2021）；Rockstroh、Blum 和 Göritz（2019）；Shu 等人（2022）；Song 等人（2022）；van Vliet 等人（2021）；Wang、Shi、Zhang 和 Chiang（2019）；Yeo 等人（2020）探討了多種自然環境對不同類型刺激的影響。這些刺激可能包括圖像、2D 影片、透過頭戴式顯示器的沉浸式虛擬實境、快門眼鏡或洞穴虛擬環境系統，或通過操控虛擬自然中相同條件的某些特徵（例如，互動式與非互動式虛擬遊戲，如 Liszio & Masuch，2019）。總體而言，沉浸式虛擬體驗被發現比其他媒介更有效（例如，Knaust 等人，2022）；然而，考慮到曝光的廣泛變異性（視覺、嗅覺和聽覺刺激，以及嵌入自然環境中的各種特徵）、媒介和結果，導致了顯著的差異化發現變異性（見表 S3）。 作為一個例子，高、張、朱、高和邱（2019）讓參與者接觸六種不同的虛擬環境，即灰色空間（也定義為「建成環境」）、藍色空間、開放綠地、部分開放綠地、部分封閉綠地和封閉綠地，通過 全景照片和頭戴顯示器進行展示。作者報告了這六種景觀對注意力疲勞和負面情緒的一般正面影響，然而，更具體地說，部分開放空間似乎對情緒提供了更高的正面影響，而藍色和部分封閉空間被認為是最受歡迎的休閒用途。同時，與沒有植被的庭院相比，帶有草和樹木的庭院被證明能在先前感到壓力的參與者中引發較低的壓力水平（黃等，2020）。舒等（2022）比較不同的景觀結構，發現不同的通透性水平和觀賞距離在景觀偏好以及焦慮和抑鬱水平中起著關鍵作用。

Noteworthy are two studies (Blum et al., 2019; Rockstroh et al., 2019) that employed Virtual-Based Heart Rate Variability (VR-HRV) biofeedback (as a treatment for a wide range of stress-related conditions through the use of computer-generated virtual environments) in comparison with traditional HRV biofeedback techniques on relaxation, mind wandering, attentional resources, and heart variability, and on physiological stress, mood, motivational aspects and attentional focus, respectively. The first study reported no significant differences in physiological responses and relaxation between the two conditions; nevertheless, the VR-HRV biofeedback was found to be more effective in reducing mind wandering and preserving attentional resources. Consistently, the other study reported no significant differences in heart-rate variability between the two techniques. Simultaneously, VR-based biofeedback was associated with higher motivation and higher attentional focus compared to the traditional one. Other attempts to combine traditional interventions and virtual nature exposure have been done. In a recent study (Ch et al., 2023) a sample of remote workers, after baseline measurements, was exposed to an intervention composed of a session of three weeks of virtual nature experience followed by three weeks of virtual nature experience combined with mindfulness practice. Findings showed that both interventions were useful in reducing stress and increasing focus; however, the combined intervention was the most effective for convergent thinking.
值得注意的是兩項研究（Blum et al., 2019; Rockstroh et al., 2019）使用虛擬基礎心率變異性（VR-HRV）生物反饋（作為治療各種與壓力相關的狀況，通過使用計算機生成的虛擬環境）與傳統的 HRV 生物反饋技術進行比較，研究了放鬆、心智漫遊、注意資源和心率變異性，以及生理壓力、情緒、動機方面和注意力焦點。第一項研究報告在兩種情況下的生理反應和放鬆之間沒有顯著差異；然而，發現 VR-HRV 生物反饋在減少心智漫遊和保留注意資源方面更為有效。另一項研究一致報告在兩種技術之間的心率變異性沒有顯著差異。同時，與傳統生物反饋相比，基於虛擬的生物反饋與更高的動機和更高的注意力焦點相關。還有其他嘗試將傳統干預和虛擬自然暴露結合起來。在最近的一項研究中（Ch et al.）。在 2023 年，對一組遠程工作者進行了基線測量後，進行了一項干預，該干預包括為期三週的虛擬自然體驗，隨後是為期三週的虛擬自然體驗結合正念練習。研究結果顯示，這兩種干預對減少壓力和提高專注力均有幫助；然而，結合的干預對於聚合思維最為有效。

In general, repeated exposure to virtual nature is demonstrated to be effective in reducing heart rate in patients undergoing hemodialysis (Hsieh & Li, 2022), in improving mood states in individuals with physical disabilities (Kalantari et al., 2022), and in reducing worry and panic in university students (Browining et al., 2023).
一般來說，反覆接觸虛擬自然被證明對於接受血液透析的患者能有效降低心率（Hsieh & Li, 2022），對於身心障礙者改善情緒狀態（Kalantari et al., 2022），以及對於大學生減少擔憂和恐慌（Browining et al., 2023）都有幫助。
3.3.1.1. Virtual nature vs. on-site exposure. Alongside a virtual nature condition, twelve studies (Browning, Shipley, et al., 2020; Calogiuri et al., 2018; Chirico & Gaggioli, 2019; Deringer & Hanley, 2021; Gao, Liang, Chen, & Qiu, 2019; Hofman, Hughes, & Walters, 2021; Léger & Mekari, 2022; Luo, Tao, Lu, Lu, & He, 2023; Reese, Stahlberg et al.,
3.3.1.1. 虛擬自然與現場接觸。除了虛擬自然條件外，還有十二項研究（Browning, Shipley 等，2020；Calogiuri 等，2018；Chirico & Gaggioli，2019；Deringer & Hanley，2021；Gao, Liang, Chen & Qiu，2019；Hofman, Hughes & Walters，2021；Léger & Mekari，2022；Luo, Tao, Lu, Lu & He，2023；Reese, Stahlberg 等，
2022; Sneed, Deringer, & Hanley, 2021; Xiang et al., 2021; Unal, Pals, Steg, Siero, & van der Zee, 2022) paired an in vivo exposure of a naturalistic setting. Browning, Saeidi-Rizi, McAnirlin, Yoon, and Pei (2020) exposed participants to the same type of natural environment in two ways. Participants either watched a 6 -min 360-degree VR nature video or spent 6 min in an outdoor forest setting. Positive mood levels increased in the outdoor condition, decreased in the control condition (i. e., an indoor setting with no visual or auditory access to nature), but remained stable in the VR condition. However, participants in both nature-based conditions reported significantly greater perceived restorativeness than participants in the control condition. Those findings were consistent with Calogiuri and colleagues (Calogiuri et al., 2018) who claimed that outdoor exposure, even in brief sessions, enhances mood states. In their study, green exercise was performed outdoors or in two VR conditions, namely a sedentary exposition of a nature walk or through a virtual nature walk while on a manually activated treadmill. Interestingly, Luo et al. (2023) reported an improvement in children's auditory and visual attention test performance when a session of physical activity was followed by virtual nature exposure, but not in the opposite order. Sneed et al. (2021) compared 12-min VR videos of stationary clips using headsets and a short hike consisting in walking along a trail. They found that connectedness to nature was significantly higher in participants who experienced real-life exposure to nature. Regarding cognitive outcomes in university students, scores in tests on memory and executive functions were higher both after a virtual and a real walking session with no significant difference between the two conditions (Léger & Mekari, 2022). Evidence for similar beneficial effects of virtual experiences is provided by Chirico and Gaggioli (2019). They exposed participants to either a panoramic view of a lake or to an immersive 360 -footage of the same landscape. In this experiment, positive and negative emotions did not differ between the two conditions. In agreement with these results, Gao, Liang, and colleagues (2019) found that participants most strongly preferred a VR environment, compared to an on-site exposure or mere photo presentation. Landscape preference was also investigated by Xiang et al. (2021), showing that virtual nature (i.e., open, semi-open, and closed green space across the four seasons) elicited more similar effects with on-site exposure than photo-elicitation. As expected, differentiated results were obtained across different seasons, media, and types of green space. Reese, Stahlberg et al. (2022) did not report differences between the two conditions (VR and in-vivo) on affect, restorativeness, subjective vitality, and stress, in healthy adults. Results on restorativeness were confirmed by Ünal et al. (2022) on a sample of university students. In line with this, Deringer and Hanley (2021) found no difference in the improvement in ecological behavior intentions after the two conditions. Regarding blue space, no difference in conservation behavior between VR blue environment (a 6-min video on the importance of the Great Barrier Reef) and real-life experience of the Great Barrier Reef was detected (Hofman et al., 2021).
2022; Sneed, Deringer, & Hanley, 2021; Xiang et al., 2021; Unal, Pals, Steg, Siero, & van der Zee, 2022）將自然環境的現場暴露配對。Browning, Saeidi-Rizi, McAnirlin, Yoon, 和 Pei（2020）以兩種方式讓參與者接觸相同類型的自然環境。參與者要麼觀看一段 6 分鐘的 360 度虛擬現實自然視頻，要麼在戶外森林環境中待 6 分鐘。在戶外條件下，正面情緒水平上升，在控制條件下（即一個沒有視覺或聽覺接觸自然的室內環境）下降，但在虛擬現實條件下保持穩定。然而，兩個基於自然的條件下的參與者報告的感知恢復感顯著高於控制條件的參與者。這些發現與 Calogiuri 及其同事（Calogiuri et al., 2018）的研究一致，他們聲稱戶外暴露，即使在短暫的會議中，也能提升情緒狀態。在他們的研究中，綠色運動是在戶外或在兩個虛擬現實條件下進行的，即靜態的自然散步展示或在手動啟動的跑步機上進行虛擬自然散步。有趣的是，Luo 等人。 （2023）報告指出，當一段體育活動後接著虛擬自然體驗時，兒童在聽覺和視覺注意力測試中的表現有所改善，但反之則沒有。Sneed 等人（2021）比較了使用頭戴式顯示器觀看的 12 分鐘靜態 VR 影片和沿著小徑短途徒步旅行的效果。他們發現，與自然的連結在經歷真實自然體驗的參與者中顯著更高。關於大學生的認知結果，記憶和執行功能測試的分數在虛擬和真實步行後均較高，兩種情況之間沒有顯著差異（Léger & Mekari，2022）。Chirico 和 Gaggioli（2019）提供了虛擬體驗類似有益效果的證據。他們讓參與者接觸到湖泊的全景視圖或同一景觀的沉浸式 360 度影片。在這個實驗中，正面和負面情緒在兩種情況下並無差異。 根據這些結果，Gao、Liang 及其同事（2019）發現參與者對虛擬現實環境的偏好最強烈，相較於現場曝光或僅僅是照片展示。Xiang 等人（2021）也調查了景觀偏好，顯示虛擬自然（即四季中的開放、半開放和封閉綠地）在效果上與現場曝光更為相似，而非照片引發的效果。正如預期的那樣，不同季節、媒介和綠地類型之間獲得了不同的結果。Reese、Stahlberg 等人（2022）在健康成年人中未報告虛擬現實和現場條件在情感、恢復性、主觀活力和壓力方面的差異。Ünal 等人（2022）在一組大學生樣本中確認了恢復性的結果。與此一致，Deringer 和 Hanley（2021）發現兩種條件後生態行為意圖的改善沒有差異。關於藍色空間，未發現虛擬現實藍色環境（關於大堡礁重要性的 6 分鐘視頻）與大堡礁的現實體驗之間在保護行為上的差異（Hofman 等人）。, 2021).
3.3.1.2. Virtual nature vs. virtual urban environment. Seventeen studies offered virtual urban environments (e.g., squares and city districts), as a control condition, compared to virtual natural environments (Chan et al., 2021; Hedblom et al., 2019; Leung et al., 2022; Li, Dong, et al., 2021; Mostajeran, Krzikawski, Steinicke, & Kühn, 2021; O'meara et al., 2020; Palanica, Lyons, Cooper, Lee, & Fossat, 2019; Reese, Mehner, Nelke, Stahlberg, & Menzel, 2022; Schebella, Weber, Schultz, & Weinstein, 2020; Schutte, Bhullar, Stilinović, & Richardson, 2017; Song et al., 2022; Suppakittpaisarn et al., 2023; Sun et al., 2023; Tabrizian, Baran, Smith, & Meentemeyer, 2018; Tanja-Dijkstra et al., 2018, Study 2; Yu, Lee, & Luo, 2018; 2020). Hedblom et al. (2019) compared three multisensory environments, i.e., an urban area, an urban park, and an urban forest. Each visual stimulus was combined with appropriate olfactory stimuli, such as diesel for the urban area, grass and forest odors for the urban park and urban forest. Similarly, matching auditory stimuli were administered, such as traffic noise, bird sounds, and the sound of a
3.3.1.2. 虛擬自然與虛擬城市環境。十七項研究提供了虛擬城市環境（例如，廣場和城市區域）作為對照條件，與虛擬自然環境進行比較（Chan et al., 2021; Hedblom et al., 2019; Leung et al., 2022; Li, Dong, et al., 2021; Mostajeran, Krzikawski, Steinicke, & Kühn, 2021; O'meara et al., 2020; Palanica, Lyons, Cooper, Lee, & Fossat, 2019; Reese, Mehner, Nelke, Stahlberg, & Menzel, 2022; Schebella, Weber, Schultz, & Weinstein, 2020; Schutte, Bhullar, Stilinović, & Richardson, 2017; Song et al., 2022; Suppakittpaisarn et al., 2023; Sun et al., 2023; Tabrizian, Baran, Smith, & Meentemeyer, 2018; Tanja-Dijkstra et al., 2018, Study 2; Yu, Lee, & Luo, 2018; 2020）。Hedblom et al. (2019) 比較了三種多感官環境，即城市區域、城市公園和城市森林。每個視覺刺激都與適當的嗅覺刺激相結合，例如城市區域的柴油氣味、城市公園和城市森林的草和森林氣味。同樣，還提供了匹配的聽覺刺激，例如交通噪音、鳥鳴聲和某種聲音。
slight breeze for the respective environments. As expected, the two virtual nature settings resulted in stronger stress reduction than the urban setting. Interestingly, perceived pleasantness was associated with olfactory and auditory stimuli, rather than with visual stimuli. The study highlights the relevance of multisensory approaches for VR research. This was subsequently supported by Schebella et al. (2020) who found a significant effect of perceived multisensory biodiversity on stress recovery over a control condition characterized by the absence of biodiversity (i.e., urban environment). In Li, Dong, et al. (2021), each group of participants with mild-to-moderate anxiety and depression was randomly assigned to one over four conditions on virtual natural scenes, or to a VR urban environment, i.e., a classic city with buildings, streets, and a couple of trees. Results confirmed a significant improvement in positive emotions and self-efficacy for individuals exposed to VR natural environments than to the urban scene. However, a good recovery effect was also found for VR urban scenes. According to the authors (Li, Dong, et al., 2021), it may be the case that providing a quiet urban environment without people produces a healing effect. Thus, one could hypothesize that a positive effect of the urban scene might be induced also by the way the stimulus is presented. In this regard, Mostajeran et al. (2021) recently compared the effects of a forest and urban environment presented in two different ways, namely photo slideshows and immersive videos. Although the type of environment influenced mood disturbance, fatigue, and cognitive performance (higher for urban exposure and lower for nature exposure), exposure to any condition reduced the heart rate. These findings are in line with Yu et al. (2018; 2020). In both studies, participants assigned to experimental and control conditions were immersed in scenes of blue and green environments or in urban scenes, such as a subway station and a shopping plaza. Results showed that participants' systolic blood pressure and heart rate decreased regardless of the condition. Nevertheless, in both studies, psychological benefits were greater after virtual nature exposure, except for attention performance which, in Yu, Lee, Lu, Huang, and Browning (2020), did not differ between the two conditions. Sun et al. (2023) observed a difference in psychophysiological outcomes only by comparing certain levels of green space exposure (high) and an urban condition. The same difference is not detectable when moderate levels of green space were displayed. Mixed evidence is available on mood and affect. Song et al. (2022) found higher positive emotions in the virtual nature conditions compared to the urban condition; however, no differences were found for negative affect and depression levels. Leung et al. (2022) reported an increase in restorativeness, nature connectedness, and enjoyment but the results on affects were inconsistent. Reese, Mehner, et al. (2022) found no differences in subjective vitality and restorativeness between the two conditions. Conversely, Palanica et al. (2019) observed that virtual nature increased perceived restorativeness and enhanced creative thinking more than viewing an urban environment, regardless of the medium (i.e., a 2D mobile tablet and a 3D VR headset). In the study by O'meara et al. (2020), participants with low or high anxiety were immersed in an alleyway of an urban town or in a lush green forest. Results showed that exposure to virtual nature significantly impacted high-anxiety university students in terms of reduction in negative affect. A town was also used in a previous study by Schutte and collaborators (2017) in which the VR condition resulted in a higher positive affect and restorativeness. Furthermore, restorativeness was found to be a mediator in the relationship between VR experience (both conditions) and positive affect, while connectedness to nature may act as a mediator between VR experience and reduced negative affect in young adults and positive affect in older adults (Chan et al., 2021). Tabrizian et al. (2018) identified perception of safety as a mediator in the relationship between urban green enclosure and perceived restorativeness with a stronger effect in the virtual park compared to the virtual plaza condition. Lastly, a virtual blue environment was observed to have a stronger effect on reduced experienced and recalled pain on a sample of dental patients, than standard dental care, while the same difference was not found comparing the VR urban environment and standard care
輕微的微風對於各自的環境來說。正如預期，這兩個虛擬自然環境的壓力減輕效果比城市環境更強。有趣的是，感知的愉悅感與嗅覺和聽覺刺激有關，而不是與視覺刺激有關。這項研究突顯了多感官方法在虛擬實境研究中的重要性。這一點隨後得到了 Schebella 等人（2020）的支持，他們發現感知的多感官生物多樣性對壓力恢復的顯著影響，與缺乏生物多樣性的對照條件（即城市環境）相比。在 Li、Dong 等人（2021）的研究中，每組輕度至中度焦慮和抑鬱的參與者被隨機分配到四種虛擬自然場景中的一種，或一個虛擬城市環境，即一個經典的城市，有建築、街道和幾棵樹。結果確認，接觸虛擬自然環境的個體在正面情緒和自我效能方面有顯著改善，與城市場景相比。然而，虛擬城市場景也發現了良好的恢復效果。根據作者（Li、Dong 等人）。，2021 年），提供一個沒有人的安靜城市環境可能會產生療癒效果。因此，可以假設城市場景的正面效果也可能是由刺激呈現的方式所引起的。在這方面，Mostajeran 等人（2021 年）最近比較了以兩種不同方式呈現的森林和城市環境的效果，即照片幻燈片和 沉浸式視頻。儘管環境類型影響了情緒困擾、疲勞和認知表現（城市暴露較高，自然暴露較低），但接觸任何條件都降低了心率。這些發現與 Yu 等人（2018 年；2020 年）的研究一致。在這兩項研究中，分配到實驗組和對照組的參與者沉浸在藍色和綠色環境的場景中，或在城市場景中，如地鐵站和購物廣場。結果顯示，無論條件如何，參與者的收縮壓和心率均有所下降。 儘管如此，在這兩項研究中，心理益處在虛擬自然暴露後更為明顯，除了注意力表現，在 Yu、Lee、Lu、Huang 和 Browning（2020）的研究中，兩種條件之間並無差異。Sun 等人（2023）僅通過比較某些高水平的綠地暴露與城市條件，觀察到心理生理結果的差異。當顯示中等水平的綠地時，則無法檢測到相同的差異。關於情緒和情感的證據混雜。Song 等人（2022）發現虛擬自然條件下的正面情緒高於城市條件；然而，對於負面情感和抑鬱水平則未發現差異。Leung 等人（2022）報告了恢復性、與自然的連結感和享受感的增加，但情感的結果不一致。Reese、Mehner 等人（2022）發現兩種條件之間在主觀活力和恢復性方面沒有差異。相反，Palanica 等人（2019）觀察到虛擬自然比觀看城市環境更能提高感知的恢復性和增強創造性思維，無論媒介為何（即。在 O'meara 等人（2020）的研究中，低焦慮和高焦慮的參與者被置身於城市小巷或郁郁蔥蔥的森林中。結果顯示，接觸虛擬自然對高焦慮的大學生在減少負面情緒方面有顯著影響。Schutte 及其合作者（2017）在之前的研究中也使用了城鎮，結果顯示虛擬現實條件下的正面情緒和恢復感更高。此外，恢復感被發現是虛擬現實體驗（兩種條件）與正面情緒之間關係的中介，而與自然的連結可能在虛擬現實體驗與年輕人減少負面情緒及老年人正面情緒之間起到中介作用（Chan 等人，2021）。Tabrizian 等人（2018）確定安全感的感知是城市綠色圍合與感知恢復感之間關係的中介，並且在虛擬公園中相比虛擬廣場條件下有更強的效果。最後，觀察到虛擬藍色環境對一組牙科病患所經歷和回憶的疼痛減輕效果比標準牙科護理更強，而在比較虛擬城市環境和標準護理時則未發現相同的差異
(Tanja-Dijkstra et al., 2018).
3.3.1.3. Virtual Nature vs. Virtual Indoor Experience. Among the selected articles, in six studies (Anderson et al., 2017; Annerstedt et al., 2013; Burmeister, Moskaliuk, & Cress, 2018; Fleury, Blanchard, & Richir, 2021; Yin et al., 2022; Zhang, Wu, & Yang, 2023) the control group was exposed to an indoor environment, such as a virtual office room. In Anderson et al.'s study (2017), each participant viewed three 15-min scenes of rural Ireland, remote beaches, and empty indoor classrooms. Results showed that both VR natural scenes provided relaxation and, interestingly, the preferred one elicited a higher reduction of negative affect and increased mood. Another study (Annerstedt et al., 2013) compared two conditions of virtual nature (i.e., with and without the sound of bird songs and water) and a control condition in which a test for inducing social stress in laboratory settings was virtually simulated. All conditions took place in a Cave Automatic Virtual Environment (CAVE) system, an immersive VR environment where projectors are directed onto the walls of a room-sized cube. After being exposed to a stressful situation, participants immersed in a natural environment with sounds through VR showed a better stress recovery, assessed through cardiovascular data and saliva cortisol, than participants of the other groups (i.e., soundless and control conditions). In line with this, Yin et al. (2022) found a reduction in physiological stress in participants exposed to a virtual desert environment compared to others exposed to a virtual office. Noteworthy was the sample composed of young males' resident in the desert. Burmeister et al. (2018), focusing on people's concentration, manipulated two virtual environments (i.e., a typical room office with furniture and a leisurely natural environment with trees, a garden, and a view of mountains, rivers, and a cottage). In this research, objective concentration (i.e., efficiency and accuracy) when performing a cognitive task was found to be higher in the control group (i.e., VR indoor condition) than in the group exposed to the VR leisure natural environment. However, no differences were found in accuracy, calculation speed, and counting tasks, nor in self-rated concentration level. On the other hand, a recent study (Zhang et al., 2023) suggested an increase in brain activities (e.g., attentional readiness and cognitive flexibility) after the exposition to virtual nature compared to the control condition. Fleury et al. (2021) investigated the environmental influences on creativity, through a sample of engineering school students, by asking them to perform a creative design activity in three different types of virtual environments (i.e., a forest, an office, and an empty dark room). This research supports the hypothesis that the natural condition, compared to two other conditions, boosts creativity score.
3.3.1.3. 虛擬自然與虛擬室內體驗。在所選的文章中，有六項研究（Anderson et al., 2017; Annerstedt et al., 2013; Burmeister, Moskaliuk, & Cress, 2018; Fleury, Blanchard, & Richir, 2021; Yin et al., 2022; Zhang, Wu, & Yang, 2023）控制組暴露於室內環境，例如虛擬辦公室。根據 Anderson et al.的研究（2017），每位參與者觀看了三個 15 分鐘的場景，分別是愛爾蘭鄉村、偏遠海灘和空曠的室內教室。結果顯示，兩個虛擬自然場景都提供了放鬆，並且有趣的是，偏好的場景引發了更高的負面情緒減少和情緒提升。另一項研究（Annerstedt et al., 2013）比較了兩種虛擬自然條件（即有和沒有鳥鳴和水聲）以及一個控制條件，在該條件下，實驗室環境中誘發社交壓力的測試被虛擬模擬。所有條件均在洞穴自動虛擬環境（CAVE）系統中進行，這是一種沉浸式虛擬現實環境，投影儀被指向一個房間大小的立方體的牆壁。 在經歷壓力情境後，沉浸在自然環境中並透過虛擬實境（VR）聆聽聲音的參與者，與其他組別（即無聲和控制條件）相比，顯示出更好的壓力恢復，這是通過心血管數據和唾液皮質醇來評估的。與此一致，Yin 等人（2022）發現，與暴露於虛擬辦公室的參與者相比，暴露於虛擬沙漠環境的參與者生理壓力有所減少。值得注意的是，樣本由居住在沙漠的年輕男性組成。Burmeister 等人（2018）專注於人們的專注力，操控了兩個虛擬環境（即一個典型的辦公室房間，配有家具，和一個悠閒的自然環境，裡面有樹木、花園，以及山脈、河流和小屋的景觀）。在這項研究中，執行認知任務時，控制組（即 VR 室內條件）的客觀專注力（即效率和準確性）被發現高於暴露於 VR 休閒自然環境的組別。然而，在準確性、計算速度和計數任務方面，並未發現差異，也沒有在自評的專注水平上發現差異。 另一方面，最近的一項研究（張等，2023）表明，與對照條件相比，接觸虛擬自然後大腦活動（例如注意準備和認知靈活性）有所增加。Fleury 等（2021）通過一組工程學校的學生，調查了環境對創造力的影響，要求他們在三種不同類型的虛擬環境中（即森林、辦公室和空曠黑暗的房間）進行創意設計活動。這項研究支持了自然條件相比其他兩種條件能提升創造力得分的假設。

Lastly, we note an interesting study (Wang et al., 2020) focusing on individuals with Generalized Anxiety Disorder. In this study, authors design the experimental condition through VR using slideshows with green and blue environments (forests, parks, woods, and rivers), paired with a control one in which a number of Virtual Abstract Paintings (VAP) have been shown to participants. The experiment took place in a CAVE VE system, and each participant was asked to exercise (cycling activity) at a moderate intensity. Results showed that both groups had higher brain alpha activity, (i.e., relaxation), but the group exposed to virtual nature exhibited higher levels of perceived stress relief than those exposed to VAP.
最後，我們注意到一項有趣的研究（Wang et al., 2020），專注於廣泛性焦慮症的個體。在這項研究中，作者通過虛擬實境設計實驗條件，使用綠色和藍色環境（森林、公園、林地和河流）的幻燈片，並與一個控制組進行配對，該控制組向參與者展示了一些虛擬抽象畫（VAP）。實驗在 CAVE 虛擬環境系統中進行，每位參與者被要求以中等強度進行運動（騎自行車活動）。結果顯示，兩組的腦部α波活動均較高（即放鬆），但接觸虛擬自然的組別所感受到的壓力緩解程度高於接觸 VAP 的組別。

3.3.2.1. Mood, affect, and emotion. In total, 32 studies measured changes in mood due to exposure to virtual nature (see Table 1 for classification of the articles by outcome investigated). In the majority of cases (i.e., seventeen studies; see Table S3), mood was measured with the Positive Affect and Negative Affect Scale (PANAS), six studies used the Profile of Mood States (POMS), four studies assessed affect using the State-Trait Anxiety Inventory (STAI-S), while other scales were used by single studies (e.g., Summary of Positive and Negative Experiences
3.3.2.1. 心情、情感與情緒。總共有 32 項研究測量了因接觸虛擬自然而引起的心情變化（見表 1 以分類研究文章的結果）。在大多數情況下（即十七項研究；見表 S3），心情是使用正向情感與負向情感量表（PANAS）來測量的，六項研究使用了心情狀態輪廓（POMS），四項研究使用狀態-特質焦慮量表（STAI-S）來評估情感，而其他量表則由單一研究使用（例如，正向與負向經驗摘要）。

Table 1 表格 1
Articles per outcome investigated.
每個結果調查的文章數量。

Table 1 (continued) 表 1（續）

Table 1 (continued) 表 1（續）

SPANE scale). Single-item measures were also used across different studies. In eight studies (Anderson et al., 2017; Huang et al., 2020; Liszio et al., 2018; Liszio & Masuch, 2019; Mostajeran et al., 2021; Schebella et al., 2020; Shu et al., 2022; Sun et al., 2023), mood was measured before and after stress induction (recovery) and in one after a boredom induction (Yeo et al., 2020), while in all the other cases it was measured after exposure to virtual and other types of nature. Four studies selected participants based on their anxiety levels, either selecting only those with mild/high levels of anxiety and/or depression (Li, Dong, et al., 2021; Shu et al., 2022; Wang et al., 2020) or just comparing participants on their level of anxiety (O'meara et al., 2020).
SPANE 量表）。在不同的研究中也使用了單項測量。在八項研究中（Anderson et al., 2017; Huang et al., 2020; Liszio et al., 2018; Liszio & Masuch, 2019; Mostajeran et al., 2021; Schebella et al., 2020; Shu et al., 2022; Sun et al., 2023），情緒在壓力誘導前後（恢復）進行測量，而在一項研究中則是在無聊誘導後進行測量（Yeo et al., 2020），而在其他所有情況下則是在接觸虛擬和其他類型的自然後進行測量。四項研究根據參與者的焦慮水平進行選擇，或僅選擇那些具有輕度/高度焦慮和/或抑鬱的參與者（Li, Dong, et al., 2021; Shu et al., 2022; Wang et al., 2020），或僅比較參與者的焦慮水平（O'meara et al., 2020）。

Positive Emotions. The effect of virtual nature on positive emotions is inconsistent. A series of studies found that positive emotions did not increase after exposure to virtual nature (Anderson et al., 2017; Gao, Zhang, et al., 2019; O'meara et al., 2020), not even during a long-term intervention (Hsieh & Li, 2022; Leung et al., 2022). Other studies, however, report significant effects suggesting that exposure to virtual nature was at least no different from exposure to other conditions, such as outdoor nature and biofeedback without VR or no biofeedback (Chirico & Gaggioli, 2019; Reese, Stahlberg et al. 2022; Rockstroh et al., 2019). In one study, VR conditions even decreased positive affect. Calogiuri et al. 's (2018) results report more enjoyment after an outdoor walk than the two VR conditions (i.e., sitting and treadmill). Positive affect and levels of tranquility were significantly reduced after the two VR conditions, with the sitting condition with significantly more reduced positive emotions than the treadmill condition.
正面情緒。虛擬自然對正面情緒的影響並不一致。一系列研究發現，接觸虛擬自然後，正面情緒並未增加（Anderson et al., 2017; Gao, Zhang, et al., 2019; O'meara et al., 2020），即使在長期介入期間也是如此（Hsieh & Li, 2022; Leung et al., 2022）。然而，其他研究報告顯示出顯著的效果，表明接觸虛擬自然至少與接觸其他條件（如戶外自然和無虛擬實境的生物反饋或無生物反饋）沒有區別（Chirico & Gaggioli, 2019; Reese, Stahlberg et al. 2022; Rockstroh et al., 2019）。在一項研究中，虛擬實境條件甚至降低了正面情感。Calogiuri et al.（2018）的結果報告顯示，戶外散步後的享受程度高於兩個虛擬實境條件（即坐著和跑步機）。在兩個虛擬實境條件後，正面情感和寧靜感顯著降低，其中坐著的條件比跑步機條件的正面情緒減少得更為明顯。

Another set of studies found an increase in positive affect (Browning, Mimnaugh, et al., 2020; Huang et al., 2020; Lakhani et al., 2020; Leung et al., 2022; Mattila et al., 2020; Schutte et al., 2017; Song et al., 2022; Yu et al., 2018, Yu et al., 2020). For instance, in Mattila et al. (2020), positive affect increased after exposure to a virtual forest. In Browning, Shipley, et al. (2020), exposure to virtual nature as well as outdoor nature was associated with higher positive affect as compared to an indoor condition. In three studies, higher positive affect and vigor were found in the VR natural (vs. urban) environment (Leung et al., 2022; Schutte et al., 2017; Song et al., 2022; Sun et al., 2023; Yu et al., 2018). In two studies (Chan et al., 2021; Yu et al., 2020), after being exposed to a virtual urban environment, positive affect decreased. In Huang et al. (2020), exposure to a courtyard with grass (as compared to one with trees and one without vegetation) after stress induction had the highest effect on positive affect.
另一組研究發現正向情感的增加（Browning, Mimnaugh 等，2020；Huang 等，2020；Lakhani 等，2020；Leung 等，2022；Mattila 等，2020；Schutte 等，2017；Song 等，2022；Yu 等，2018，Yu 等，2020）。例如，在 Mattila 等（2020）中，接觸虛擬森林後正向情感增加。在 Browning, Shipley 等（2020）中，接觸虛擬自然以及戶外自然與室內環境相比，正向情感更高。在三項研究中，VR 自然環境（與城市環境相比）中發現更高的正向情感和活力（Leung 等，2022；Schutte 等，2017；Song 等，2022；Sun 等，2023；Yu 等，2018）。在兩項研究中（Chan 等，2021；Yu 等，2020），在接觸虛擬城市環境後，正向情感減少。在 Huang 等（2020）中，經過壓力誘導後，接觸有草的庭院（與有樹的庭院和沒有植被的庭院相比）對正向情感的影響最大。

Two studies found an increment in positive affect after virtual nature exposure as compared to other delivery modes. In Liszio et al. (2018), a higher increment of positive affect was found in the virtual condition as compared to the desktop and control conditions. In another study (Yeo et al., 2020), positive affect was associated with the delivery mode (i.e.,
兩項研究發現，與其他傳遞模式相比，虛擬自然體驗後的正向情感有所增加。在 Liszio 等人（2018）的研究中，虛擬條件下的正向情感增幅高於桌面和控制條件。在另一項研究（Yeo 等人，2020）中，正向情感與傳遞模式相關聯。
2D screen, videos of real nature, and computer-generated virtual nature) in a way that the major increment in positive affect was observed in the most immersive mode, i.e., virtual nature rather than the less immersive ones, i.e., 2D screen and videos.
2D 螢幕、 個真實自然的影片，以及電腦生成的虛擬自然，以一種方式顯示出在最具沉浸感的模式中，即虛擬自然，觀察到正面情感的主要增長，而不是在較不具沉浸感的模式中，即 2D 螢幕和 影片。

Interesting to note is the effect on positive affect in those studies that compared the effect of VR with or without interaction. In Li, Dong, et al. (2021), positive affect decreased just in the natural interactive environment but not in an urban and a natural non-interactive environment in participants with mild to moderate anxiety and depression. In the study by Liszio and Masuch (2019), no significant differences were found after mere exposure to all conditions (interactive, noninteractive, and control conditions). Lastly, control was irrelevant in determining positive affect. Similarly, Reese et al. (2021) found that the positive affect was higher after VR regardless of whether the participant had or had not active control in the navigation.
有趣的是，在比較有互動與無互動的虛擬實境（VR）效果的研究中，對於正向情感的影響。根據李、董等人（2021）的研究，在自然互動環境中，輕度至中度焦慮和抑鬱的參與者的正向情感下降，但在城市和自然非互動環境中則沒有下降。在 Liszio 和 Masuch（2019）的研究中，對所有條件（互動、非互動和控制條件）的單純暴露後，並未發現顯著差異。最後，控制在決定正向情感方面並不相關。同樣，Reese 等人（2021）發現，無論參與者是否在導航中有主動控制，VR 後的正向情感均較高。

Few studies investigated the differences in positive affect as compared to exposure to different natural environments. For instance, in Schebella et al.'s (2020) study, when compared to the virtual urban environment, the low biodiversity virtual environment showed greater happiness. All the other differences between higher levels of biodiversity were nonsignificant. In Wang et al. (2019), it emerged that three environments were successful in increasing vigor, namely waterfall with trees, wood with bench, and the most artificial natural environment. On the contrary, the other three environments, namely wood with a platform and bench, a platform with trees, and a pool with plants decreased it.
很少有研究探討正向情感與不同自然環境的接觸之間的差異。例如，在 Schebella 等人（2020）的研究中，與虛擬城市環境相比，低生物多樣性的虛擬環境顯示出更大的快樂。所有其他較高生物多樣性之間的差異均不顯著。在 Wang 等人（2019）的研究中，發現三種環境成功地提高了活力，即有樹的瀑布、帶長椅的林地，以及最人工的自然環境。相反，其他三種環境，即有平台和長椅的林地、有樹的平台，以及有植物的池塘則降低了活力。

Negative Emotions. In this section, we review studies showing how exposure to virtual nature decreases negative affect. As compared to a virtual urban environment, the virtual natural environment significantly decreased negative affect, confusion, anger-hostility, tension, depression, and fatigue (Chan et al., 2021; Leung et al., 2022; Mattila et al., 2020; Yu et al., 2018, 2020). Overtime exposure to virtual natural environments decreased worry and panic in Browning et al. (2023), with no similar significant changes found for a condition with no exposure. Nervousness decreased in Kalantari et al. (2022) after three modules of exposure. Mostajeran et al. (2021), after stress induction, found that the virtual urban environment (vs. virtual forest) worsened the mood while, on the contrary, fatigue decreased after the virtual forest (vs. urban environment) exposure. Interestingly, in three studies, fatigue even increased after the virtual urban environment (Mostajeran et al., 2021; Yu et al., 2018, 2020). O'meara et al. (2020) found a three-way interaction between time, condition, and anxiety levels. In other words, they found a significant reduction (between pre- and post-levels) in negative affect in students with high (but not with low) anxiety exposed to a virtual natural environment and the increase was higher than in the urban condition. Chirico and Gaggioli (2019) found negative emotions such as anger, disgust, and sadness significantly decreased after exposure to virtual nature, but not after outdoor nature.
負面情緒。在本節中，我們回顧了研究顯示接觸虛擬自然環境如何減少負面情感。與虛擬城市環境相比，虛擬自然環境顯著減少了負面情感、困惑、憤怒-敵意、緊張、抑鬱和疲勞（Chan et al., 2021; Leung et al., 2022; Mattila et al., 2020; Yu et al., 2018, 2020）。Browning et al.（2023）發現，隨著時間的推移，接觸虛擬自然環境減少了擔憂和恐慌，而在沒有接觸的情況下則未發現類似的顯著變化。Kalantari et al.（2022）在三個模塊的接觸後，緊張感減少。Mostajeran et al.（2021）在壓力誘導後發現，虛擬城市環境（與虛擬森林相比）使情緒惡化，而在接觸虛擬森林（與城市環境相比）後，疲勞則減少。有趣的是，在三項研究中，接觸虛擬城市環境後，疲勞甚至增加（Mostajeran et al., 2021; Yu et al., 2018, 2020）。O'meara et al.（2020）發現時間、條件和焦慮水平之間存在三向交互作用。 換句話說，他們發現高焦慮（但低焦慮則不然）學生在接觸虛擬自然環境後，負面情緒的顯著減少（在前後水平之間），而且這種增加的幅度高於城市環境。Chirico 和 Gaggioli（2019）發現，接觸虛擬自然後，憤怒、厭惡和悲傷等負面情緒顯著減少，但在接觸戶外自然後則沒有。

Different scholars focused on the type of environments that could result in a higher decrease of negative emotions. For instance, in Schebella et al. (2020) a low biodiversity environment showed lower anxiety and stress as compared to the urban environment and a higher biodiversity environment. Anxiety was also lower in a multisensorial environment than in a visual-only environment, demonstrating that auditory and olfactory stimuli can make a difference in effectiveness. Wang et al. (2019) compared different natural environments and found that the highest recovery in mood and anxiety was in the natural environment with the most presence of artificial elements and in blue spaces (vs. forest environments with different elements such as waterfalls and artificial elements).
不同學者專注於可能導致負面情緒顯著減少的環境類型。例如，在 Schebella 等人（2020）的研究中，低生物多樣性的環境相比於城市環境和高生物多樣性的環境顯示出較低的焦慮和壓力。在多感官環境中，焦慮也低於僅有視覺的環境，這表明聽覺和嗅覺刺激在效果上可以產生差異。Wang 等人（2019）比較了不同的自然環境，發現情緒和焦慮的恢復在人工元素最為豐富的自然環境和藍色空間中達到最高（相比於擁有不同元素如瀑布和人工元素的森林環境）。

In Gao, Zhang, et al. (2019) the most effective natural environments in reducing negative emotions were the partly open green space, and then the open green space, the partly closed green space, and the blue space, while the gray space and the closed green space were the least effective. Shu et al. (2022) investigated edge permeability and viewing distance of green spaces and found that the higher anxiety and
在高、張等人（2019）的研究中，最有效減少負面情緒的自然環境是部分開放的綠地，其次是開放的綠地、部分封閉的綠地和藍色空間，而灰色空間和封閉的綠地則是最不有效的。舒等人（2022）研究了綠地的邊緣通透性和觀賞距離，發現焦慮程度越高，
depression scores were related to landscapes at a viewing distance of 20 m (vs. 100 m and 200 m ) and characterized by high edge permeability. Interestingly, it seems that a role is played by preference. Indeed, in Anderson et al. (2017), the preferred natural environment shown (green or blue) decreased negative emotions. Despite these positive findings, two studies did not find any differences in the reduction of negative affect due to the type of natural environment shown (Huang et al., 2020; Li, Dong, et al., 2021).
抑鬱分數與在 20 米的觀賞距離（相較於 100 米和 200 米）下的景觀相關，並以高邊緣通透性為特徵。有趣的是，偏好似乎在其中扮演了一個角色。事實上，在 Anderson 等人（2017）的研究中，所展示的偏好自然環境（綠色或藍色）減少了負面情緒。儘管有這些正面的發現，兩項研究並未發現由於所展示的自然環境類型而導致負面情感減少的差異（Huang 等人，2020；Li, Dong 等人，2021）。

Some studies investigated additional characteristics that could enhance the efficacy of VR. In the study by Reese et al. (2021), negative emotions were lower after the VR experience regardless of the level of activity in the navigation exerted by participants. In Liszio and Masuch (2019) exposure to interactive VR was associated with less anxiety than in the control group after exposure to the environments. Mostajeran et al. (2021), after stress induction, found no differences in depressive and anxiety symptoms, or stress for the type of medium used (VR or photos). One study (Liszio et al., 2018) suggests the added value of virtual environments on traditional media. Here, exposure to a natural environment reduced anxiety and the reduction was significantly different for VR condition vs. desktop and control. Lastly, two studies investigated VR as a potential tool to lower anxiety in combination with other important clinical instruments such as biofeedback. Findings by Blum et al. (2019) on anxiety attested that exposure to a natural VR environment combined with biofeedback was more successful than biofeedback alone in sustaining the levels of anxiety. However, Rockstroh et al. (2019) did not find any specific effects of virtual nature combined with biofeedback on mood.
一些研究調查了可以增強虛擬實境（VR）效能的額外特徵。在 Reese 等人（2021）的研究中，無論參與者在導航中所施加的活動水平如何，VR 體驗後的負面情緒均較低。在 Liszio 和 Masuch（2019）的研究中，與對照組相比，接觸互動式 VR 後的焦慮感較少。Mostajeran 等人（2021）在壓力誘導後發現，使用的媒介類型（VR 或照片）對抑鬱和焦慮症狀或壓力沒有差異。一項研究（Liszio 等人，2018）表明虛擬環境對傳統媒介的附加價值。在這裡，接觸自然環境減少了焦慮，且 VR 條件下的減少與桌面和對照組相比顯著不同。最後，兩項研究調查了 VR 作為降低焦慮的潛在工具，並結合其他重要的臨床工具，如生物反饋。Blum 等人（2019）對焦慮的研究表明，接觸自然 VR 環境結合生物反饋在維持焦慮水平方面比單獨使用生物反饋更成功。然而，Rockstroh 等人。 （2019）未發現虛擬自然結合生物反饋對情緒的任何特定影響。

Lastly, a small set of studies did not find any specific effect of exposure to virtual nature on negative mood. For instance, tiredness did not decrease after one-time exposure in Kalantari et al. (2022), while depression and rumination did not change after repeated exposure in Browning et al. (2023). In Schutte et al. (2017) there was no effect of condition (VR natural vs. VR urban) on negative affect. Three studies found a decrease in negative affect and boredom across all conditions, with no specific effects for virtual natural environments (vs. outdoor nature or a no exposure condition; Browning, Mimnaugh, et al. (2020); Reese, Stahlberg et al. (2022) and vs. virtual urban environments; Song et al., 2022) and the virtual medium (Yeo et al., 2020). In the study by Calogiuri et al. (2018), in the sitting VR condition, there was even a significant increase in negative emotions such as fatigue, which was significantly higher in than the other two conditions (i.e., in the outdoor condition and in the VR treadmill condition). Interestingly, negative emotions were positively associated with motion sickness.
最後，一小組研究並未發現虛擬自然環境對負面情緒的具體影響。例如，Kalantari 等人（2022）的研究中，經過一次接觸後疲勞感並未減少，而 Browning 等人（2023）的研究中，重複接觸後抑鬱和反芻情緒也沒有變化。在 Schutte 等人（2017）的研究中，條件（虛擬自然與虛擬城市）對負面情緒沒有影響。三項研究發現，在所有條件下，負面情緒和無聊感均有所減少，但對虛擬自然環境（與戶外自然或無接觸條件相比；Browning, Mimnaugh 等人（2020）；Reese, Stahlberg 等人（2022）以及與虛擬城市環境相比；Song 等人，2022）和虛擬媒介（Yeo 等人，2020）並無具體影響。在 Calogiuri 等人（2018）的研究中，在坐著的虛擬現實條件下，負面情緒如疲勞感甚至顯著增加，且在這一條件下的疲勞感顯著高於其他兩個條件（即戶外條件和虛擬跑步機條件）。有趣的是，負面情緒與暈動病呈正相關。
3.3.2.2. Perceived stress and relaxation. Eleven studies measure perceived stress as an outcome of exposure to virtual nature. Six studies confirmed the benefit of exposure to virtual nature in lowering stress ( Ch et al., 2023; Chan et al., 2021; Reese et al., 2021; Schebella et al., 2020; Wang et al., 2020; Zhang et al., 2023), while two did not (Mostajeran et al., 2021; Reese, Stahlberg et al., 2022). Interestingly, the study by Suppakittpaisarn et al. (2023) found that higher stress recovery was found with videos of 5 min (vs. 1 min and 15 min ) and it was greater for women in natural (vs. urban) VR environments. In the study by Reese et al. (2021), perceived stress was lower after exposure to virtual nature, and this decrease was significantly stronger in the condition of no control (experience controlled by the experimenter) as opposed to active control by the participant. In the study by Wang et al. (2020) as well, in adults with generalized anxiety disorder, perceived stress was reduced after physical exercise combined with exposure to virtual videos and this decrease was greater when the contents of the virtual videos were natural environments vs. abstract painting. In Schebella et al. (2020), perceived stress was significantly lower in the low biodiversity than in the urban environment and in the multisensory rather than the visual-only natural environment. In two studies, relaxation was higher when traditional biofeedback was associated with VR nature (vs. traditional biofeedback alone; Blum et al., 2019) and in a VR condition (vs. 2D and no video condition; Knaust et al., 2022). In Ch et al. (2023), stress was lower in the conditions of VR nature and VR nature with mindfulness as compared to a condition with no intervention, with no added value of mindfulness. Conflicting are the findings by Mostajeran et al. (2021), who did not find any significant effect in perceived stress measure with two different instruments, neither regarding the type of environment (natural vs. urban) nor for the level of immersion (VR and 2D photos). Following the same line, in Reese, Stahlberg et al. (2022) no differences in stress were found between virtual and outdoor nature exposure.
3.3.2.2. 感知壓力與放鬆。十一項研究測量了感知壓力作為接觸虛擬自然的結果。六項研究確認了接觸虛擬自然在降低壓力方面的好處（Ch et al., 2023; Chan et al., 2021; Reese et al., 2021; Schebella et al., 2020; Wang et al., 2020; Zhang et al., 2023），而兩項則未能確認（Mostajeran et al., 2021; Reese, Stahlberg et al., 2022）。有趣的是，Suppakittpaisarn et al. (2023) 的研究發現，觀看 5 分鐘的影片（與 1 分鐘和 15 分鐘相比）能夠更好地促進壓力恢復，且在自然虛擬環境中女性的恢復效果更佳（與城市環境相比）。在 Reese et al. (2021) 的研究中，接觸虛擬自然後的感知壓力較低，且在無控制條件下（由實驗者控制的體驗）這一減少的效果顯著強於參與者主動控制的情況。在 Wang et al. (2020) 的研究中，對於患有廣泛性焦慮症的成年人，結合身體運動與虛擬影片的接觸後，感知壓力減少，且當虛擬影片的內容為自然環境而非抽象畫時，這一減少的效果更為明顯。 在 Schebella 等人（2020）的研究中，低生物多樣性的環境中感知壓力顯著低於城市環境，而多感官環境的壓力則低於僅視覺的自然環境。在兩項研究中，當傳統生物反饋與虛擬現實自然環境結合時，放鬆感較高（與僅傳統生物反饋相比；Blum 等人，2019），而在虛擬現實條件下（與 2D 和無視頻條件相比；Knaust 等人，2022）。在 Ch 等人（2023）的研究中，虛擬現實自然環境和結合正念的虛擬現實自然環境的壓力較低，與無干預的條件相比，正念並未帶來額外的價值。Mostajeran 等人（2021）的研究結果則相互矛盾，他們在使用兩種不同工具測量感知壓力時，無論是環境類型（自然與城市）還是沉浸程度（虛擬現實與 2D 照片）均未發現顯著效果。沿著同樣的思路，Reese、Stahlberg 等人（2022）在虛擬與戶外自然暴露之間未發現壓力的差異。
3.3.2.3. Physiological stress. In total, twenty-six articles measured physiological stress regarding exposure to virtual nature (see Table 1). Eleven studies measured physiological stress after stress induction (Anderson et al., 2017; Annerstedt et al., 2013; Hedblom et al., 2019; Huang et al., 2020; Knaust et al., 2022; Liszio et al., 2018; Liszio & Masuch, 2019; Mostajeran et al., 2021; Schebella et al., 2020; Sun et al., 2023; Yin et al., 2022) and one after fatigue induction (Zhang et al., 2023), while in all the other cases stress was measured after exposure to virtual and other types of nature.
3.3.2.3. 生理壓力。總共有二十六篇文章測量了與接觸虛擬自然相關的生理壓力（見表 1）。十一項研究在壓力誘導後測量了生理壓力（Anderson et al., 2017; Annerstedt et al., 2013; Hedblom et al., 2019; Huang et al., 2020; Knaust et al., 2022; Liszio et al., 2018; Liszio & Masuch, 2019; Mostajeran et al., 2021; Schebella et al., 2020; Sun et al., 2023; Yin et al., 2022），而一項在疲勞誘導後測量（Zhang et al., 2023），其餘情況下則是在接觸虛擬和其他類型的自然後測量壓力。

Heart Rate. Regarding heart rate, most studies found that heart rate and heart rate variability decreased over time suggesting relaxation, with no differences between conditions, namely urban vs. natural (Mostajeran et al., 2021; Song et al., 2022; Yu et al., 2018, 2020), natural vs. indoor environment (Anderson et al., 2017) and immersive VR videos vs. 2D photo slideshow (Knaust et al., 2022; Mostajeran et al., 2021). One study reported decreases in heart rate after VR natural videos, with the first exposure being more effective than the second and third over a three-week period (Hsieh & Li, 2022). The same result did not apply to heart rate variability indices. Moreover, two studies found that standard and VR biofeedback were both effective in inducing relaxation with no differences between them in heart rate and heart rate variability (Blum et al., 2019; Rockstroh et al., 2019). Only one study found that heart rate variability indicated significantly higher relaxation in the nature VR condition than in desktop and control conditions (Liszio et al., 2018). In just one study, heart rate variability suggested higher relaxation in the VR nature vs. VR urban condition (Chan et al., 2021). Recovery from stress induction in terms of heart rate and heart rate variability was not found in Liszio and Masuch (2019) per level of biodiversity of the environment and in Shu et al. (2022) between different viewing distances and edge permeability levels of green spaces. No differences in recovery in terms of heart rate and heart rate variability were also found in Schebella et al. (2020) between natural and urban environments and in Sun et al. (2023) per level (high, moderate, none) of green space exposure.
心率。關於心率，大多數研究發現心率和心率變異性隨時間減少，顯示出放鬆的趨勢，且在不同條件之間沒有差異，即城市與自然環境（Mostajeran et al., 2021；Song et al., 2022；Yu et al., 2018, 2020）、自然與室內環境（Anderson et al., 2017）以及沉浸式虛擬實境視頻與 2D 照片幻燈片（Knaust et al., 2022；Mostajeran et al., 2021）。有一項研究報告指出，在觀看虛擬實境自然視頻後心率下降，第一次接觸的效果比第二次和第三次在三週期間更為明顯（Hsieh & Li, 2022）。相同的結果並不適用於心率變異性指標。此外，兩項研究發現標準生物反饋和虛擬實境生物反饋在誘導放鬆方面均有效，且在心率和心率變異性上沒有差異（Blum et al., 2019；Rockstroh et al., 2019）。只有一項研究發現，心率變異性在自然虛擬實境條件下顯示出顯著更高的放鬆程度，相較於桌面和控制條件（Liszio et al., 2018）。在僅有的一項研究中，心率變異性顯示在虛擬實境自然環境與虛擬實境城市環境之間的放鬆程度更高（Chan et al.）。，2021）。Liszio 和 Masuch（2019）在環境的生物多樣性水平上未發現壓力誘導後的心率和心率變異性的恢復，而 Shu 等人（2022）在不同的觀賞距離和綠地的邊緣通透性水平之間也未發現差異。Schebella 等人（2020）在自然和城市環境之間以及 Sun 等人（2023）在綠地暴露水平（高、中、無）上也未發現心率和心率變異性恢復的差異。

Regarding the virtual experience, two studies found that recovery from stress induction in terms of reduced heart rate and heart rate variability was higher in the condition of virtual forest with nature sound (vs. without sound; Annerstedt et al., 2013) and multisensory environment (vs. visual-only environment; Schebella et al., 2020). Heart rate variability indicated higher relaxation in the condition with active interaction with the virtual natural environment (vs. no interaction and control groups; Liszio & Masuch, 2019). Regarding the types of environments, higher recovery from stress induction was found in a low (vs. moderate) biodiversity natural environment (Schebella et al., 2020) and in natural environments such as green space and blue spaces, while a natural context with more artificial elements was the only one with an increase in heart rate (Wang et al., 2019).
關於虛擬體驗，兩項研究發現，在虛擬森林中伴隨自然聲音的情況下（與無聲音相比；Annerstedt et al., 2013）和多感官環境（與僅視覺環境相比；Schebella et al., 2020），在減少心率和心率變異性方面，從壓力誘發中恢復的程度更高。心率變異性顯示，在與虛擬自然環境進行主動互動的情況下（與無互動和控制組相比；Liszio & Masuch, 2019），放鬆程度更高。關於環境類型，在低（與中等）生物多樣性的自然環境中，從壓力誘發中恢復的程度更高（Schebella et al., 2020），以及在綠地和藍色空間等自然環境中，而含有更多人工元素的自然背景則是唯一一個心率增加的情況（Wang et al., 2019）。

Electrodermal Activity (EDA). EDA gives indications regarding physiological arousal related to emotions and stress. EDA decreased over time after VR exposure and the decrease was greater for natural vs. urban environments in three studies (Anderson et al., 2017; Hedblom et al., 2019; Huang et al., 2020), while in one study the pre-post exposure changes did not differ between high and moderate green level vs.
電皮膚活動（EDA）。EDA 提供有關與情緒和壓力相關的生理喚起的指示。在三項研究中（Anderson et al., 2017; Hedblom et al., 2019; Huang et al., 2020），VR 暴露後 EDA 隨時間減少，且自然環境的減少幅度大於城市環境，而在一項研究中，高綠化水平與中等綠化水平的前後暴露變化並無差異。
urban environment (Sun et al., 2023). One study did not find differences between virtual nature vs. outdoor nature vs. indoor environment; Browning, Shipley, et al., 2020), and no differences were found also for viewing distance and edge permeability (Shu et al., 2022). One study showed how a photo slideshow was more effective than the VR condition in lowering the EDA scores between the stress induction and the exposure phases (Mostajeran et al., 2021). On the contrary, in another study, the reduction in EDA was higher in VR and PC conditions vs. no video, but the scores in the VR condition did not differ from the PC condition (Knaust et al., 2022).
城市環境（Sun et al., 2023）。一項研究未發現虛擬自然、戶外自然與室內環境之間的差異（Browning, Shipley, et al., 2020），在觀看距離和邊緣通透性方面也未發現差異（Shu et al., 2022）。一項研究顯示，照片幻燈片在降低壓力誘導與暴露階段之間的皮膚電反應（EDA）分數方面，比虛擬現實（VR）條件更有效（Mostajeran et al., 2021）。相反，在另一項研究中，虛擬現實和個人電腦（PC）條件下的 EDA 減少程度高於無視頻條件，但 VR 條件的分數與 PC 條件並無差異（Knaust et al., 2022）。

Blood pressure. Regarding blood pressure, only seven studies measured systolic and diastolic blood pressure and mean arterial pressure (see Table 1). The findings by Wang et al. (2019) indicated relaxation after VR exposure, with significant differences in pre- and post-measurements for green natural environments and blue space, with the latter one showing a greater decrease. Sun et al. (2023) found that the observed decrease in systolic blood pressure after exposure was higher in a green (vs. urban) condition, but no differences in diastolic blood pressure were found. Recovery after stress induction as suggested by the decrease of mean arterial pressure after VR desert exposure was significantly greater than the VR urban exposure condition in desert residents (Yin et al., 2022). Yu et al. (2018) found that blood pressure was decreasing over time, regardless of condition (natural vs. urban). Conversely, Yu et al. (2020) and Song et al. (2022) found no differences between pre- and post-exposure measurements for blood pressure, nor any differences between the two conditions (natural and urban). Two other blood-related measures, namely blood volume amplitude and blood oxygen value, did not differ significantly per viewing distance and edge permeability in Shu et al. (2022).
血壓。關於血壓，只有七項研究測量了收縮壓、舒張壓和平均動脈壓（見表 1）。王等人（2019）的研究結果顯示，在虛擬現實（VR）暴露後放鬆，綠色自然環境和藍色空間的前後測量有顯著差異，後者顯示出更大的下降。孫等人（2023）發現，暴露後收縮壓的下降在綠色（與城市相比）條件下更高，但舒張壓沒有發現差異。根據虛擬現實沙漠暴露後平均動脈壓的下降，壓力誘導後的恢復在沙漠居民中顯著高於虛擬現實城市暴露條件（尹等人，2022）。余等人（2018）發現，無論條件（自然與城市）如何，血壓隨時間下降。相反，余等人（2020）和宋等人（2022）發現血壓的前後暴露測量之間沒有差異，也沒有自然和城市兩種條件之間的差異。 在 Shu 等人（2022）的研究中，另外兩項與血液相關的指標，即血容量振幅和血氧值，並未因觀賞距離和邊緣滲透性而顯著不同。

Electroencephalography (EEG). Studies measuring EEG focused primarily on alpha waves, which suggest relaxation. In participants with generalized anxiety disorder, the increase of the alpha waves between pre- and post-exposure measurements was higher in the condition in which aerobic exercise was combined with virtual nature vs. abstract painting (Wang et al., 2020). In subjects with acquired physical disabilities, the increase in alpha waves and alpha/beta ratio indicated that subjects after exposure to virtual nature had a significantly greater change and post-exposure alpha waves than those exposed to a 2 D video of a garden, suggesting higher relaxation (Je & Lee, 2020). In another study on gastrointestinal cancer patients (Song et al., 2022), alpha waves were higher than baseline in the VR natural condition, while in the VR urban condition they were higher during baseline. One study did not find differences in alpha waves between the baseline (resting state) and during visual VR and between different natural environments (Gao, Zhang, et al., 2019). Li, Dong, et al. (2021) collected data on participants with mild to moderate anxiety and depression and computed three EEG indices namely engagement, calmness, and alertness. Indicators of engagement and calmness did not change significantly across conditions while alertness was significantly higher after virtual restorative natural environments, suggesting cognitive recovery and improved cognitive functions. Following the same line, results by Zhang et al. (2023) suggested higher attentional states and cognitive flexibility, and lower cognitive processing load in a VR natural vs. indoor environment condition.
腦電圖（EEG）。測量 EEG 的研究主要集中在α波上，這表明放鬆。在患有廣泛性焦慮症的參與者中，結合有氧運動與虛擬自然的情況下，前後測量之間α波的增加高於與抽象畫相比的情況（Wang et al., 2020）。在有獲得性身體殘疾的受試者中，α波和α/β比率的增加表明，接觸虛擬自然後的受試者相比於接觸 2D 花園視頻的受試者，顯示出顯著更大的變化和更高的後接觸α波，這表明更高的放鬆程度（Je & Lee, 2020）。在另一項針對腸胃道癌症患者的研究中（Song et al., 2022），在虛擬自然條件下的α波高於基線，而在虛擬城市條件下則在基線期間較高。一項研究未發現基線（靜息狀態）與視覺虛擬現實期間以及不同自然環境之間的α波差異（Gao, Zhang, et al., 2019）。李、董等。 （2021）收集了輕度至中度焦慮和抑鬱參與者的數據，並計算了三個腦電圖指標，即參與度、平靜度和警覺性。參與度和平靜度在不同條件下沒有顯著變化，而警覺性在虛擬恢復性自然環境後顯著提高，這表明認知恢復和認知功能改善。沿著同樣的思路，張等人（2023）的研究表明，在虛擬自然環境與室內環境條件下，注意力狀態更高，認知靈活性更強，認知處理負荷更低。

Salivary Cortisol. Only four studies measured cortisol levels before and after exposure (recovery from stress induction). Three found no statistical change per condition: high and moderate levels of green vs. no green (urban) environment (Sun et al., 2023), natural with vs. without sound vs. indoor condition (Annerstedt et al., 2013), and VR vs. desktop vs. control condition (Liszio et al., 2018). In Yin et al. (2022), in desert residents, salivary cortisol decreased in all conditions (desert, green, and office), but this reduction was significantly higher in the desert vs. office condition.
唾液皮質醇。只有四項研究在暴露前後測量了皮質醇水平（從壓力誘導中恢復）。三項研究在各條件下未發現統計學上的變化：高和中等水平的綠色環境與無綠色（城市）環境（Sun et al., 2023）、自然環境有聲與無聲與室內條件（Annerstedt et al., 2013），以及虛擬實境與桌面與控制條件（Liszio et al., 2018）。在 Yin et al.（2022）中，沙漠居民的唾液皮質醇在所有條件下均有所下降（沙漠、綠色和辦公室），但在沙漠與辦公室條件下的減少顯著更高。

Salivary Amylase. Three studies measured salivary amylase. While one did not find any significant decreases per condition of VR nature vs. VR urban (Yu et al., 2018), another study (Sun et al., 2023) found that salivary amylase decreased more in a VR natural vs. VR urban condition. The last study found that among seven natural environments, in the one with wood with a platform and bench (a moderately artificial environment) salivary amylase was higher, indicating more stress (Wang et al., 2019).
唾液澱粉酶。三項研究測量了唾液澱粉酶。雖然有一項研究未發現虛擬現實自然環境與虛擬現實城市環境之間有任何顯著的減少（Yu et al., 2018），但另一項研究（Sun et al., 2023）發現虛擬現實自然環境下的唾液澱粉酶減少得更多。最後一項研究發現，在七個自然環境中，具有平台和長椅的木質環境（中度人工環境）中的唾液澱粉酶較高，顯示出更多的壓力（Wang et al., 2019）。
3.3.2.4. Restorativeness. In total, fifteen studies measured the restorative quality of the natural environments displayed through VR and the restorative process reported by the individuals exposed to those environments (see Table 1). Of these works, seven used the Perceived Restorativeness Scale (PRS; Browning, Mimnaugh, et al., 2020; Calogiuri et al., 2018; Leung et al., 2022; Reese, Stahlberg et al., 2022; Schutte et al., 2017; Song et al., 2022; Tabrizian et al., 2018), four the Restoration Outcome Scale (ROS; Mattila et al., 2020; Reese et al., 2021; Reese, Mehner et al., 2022; Reese, Stahlberg et al., 2022) and others used other instruments, such as the Restorative Environmental Scale (RES), the Perceived Restorative Characteristics Questionnaire (PRCQ), and the Restorative Components Scale (RCS).
3.3.2.4. 恢復性。總共有十五項研究測量了通過虛擬現實展示的自然環境的恢復質量，以及暴露於這些環境中的個體所報告的恢復過程（見表 1）。在這些研究中，有七項使用了感知恢復性量表（PRS；Browning, Mimnaugh 等，2020；Calogiuri 等，2018；Leung 等，2022；Reese, Stahlberg 等，2022；Schutte 等，2017；Song 等，2022；Tabrizian 等，2018），四項使用了恢復結果量表（ROS；Mattila 等，2020；Reese 等，2021；Reese, Mehner 等，2022；Reese, Stahlberg 等，2022），其他則使用了其他工具，如恢復環境量表（RES）、感知恢復特徵問卷（PRCQ）和恢復組件量表（RCS）。

As assessing the quality of the environments shown, across studies, restorativeness is measured just once after exposure to the environments. Most of the studies showed how virtual nature has a higher restorative quality than a virtual urban environment (Leung et al., 2022; Mattila et al., 2020; Schutte et al., 2017; Tabrizian et al., 2018; Yu et al., 2020; Ünal et al., 2022). In line with these results, an intervention study, with 3-5 exposures over the course of one week, found that restorativeness decreased after one week only in the urban (vs. natural) condition (Song et al., 2022). In Browning, Mimnaugh, et al. (2020) exposure to VR natural environment and outdoor natural environment also predicted greater restorativeness than an indoor setting. Moreover, studies point to a comparable effect between virtual nature and outdoor nature (Reese, Stahlberg et al., 2022; Ünal et al., 2022). Following this line, Calogiuri et al. (2018) found no differences between VR and outdoor conditions for fascination and being away. Both dimensions showed positive and significant correlations with the sense of presence. In Mattila et al. (2020), the VR forest environment was perceived as more restorative than the real urban forest environment and the real semi-urban forest in terms of all the sub-scales. The VR forest was even perceived as more coherent and fascinating than the three real forest environments and more compatible and more likely to enhance the 'being away' experience than the real semi-urban and urban forests. The virtual natural environment was perceived as more restorative during exercise than virtual abstract painting (Wang et al., 2020). Song et al. (2022) found that the blue environment was the most restorative followed by an open green space, a semi-open green space, a closed green space, and lastly, a gray space. Lastly, one study found no significant differences in the restorative quality of five different virtual environments (urban, natural, and three interactive natural environments; Li , Dong, et al., 2021).
在評估所展示環境的質量時，跨研究中，恢復性僅在接觸環境後測量一次。大多數研究顯示，虛擬自然的恢復性質量高於虛擬城市環境（Leung et al., 2022; Mattila et al., 2020; Schutte et al., 2017; Tabrizian et al., 2018; Yu et al., 2020; Ünal et al., 2022）。與這些結果一致的一項干預研究，在一週內進行 3-5 次接觸，發現恢復性在城市（與自然相比）條件下僅在一週後下降（Song et al., 2022）。在 Browning、Mimnaugh 等人（2020）的研究中，接觸虛擬自然環境和戶外自然環境也預測了比室內環境更高的恢復性。此外，研究指出虛擬自然和戶外自然之間的效果相當（Reese, Stahlberg et al., 2022; Ünal et al., 2022）。沿著這一思路，Calogiuri 等人（2018）發現虛擬現實和戶外條件在著迷和遠離感方面沒有差異。這兩個維度與存在感顯示出正向且顯著的相關性。在 Mattila 等人。 在 2020 年，虛擬現實森林環境在所有子尺度上被認為比真實的城市森林環境和真實的半城市森林環境更具恢復性。虛擬森林甚至被認為比三個真實森林環境更具一致性和吸引力，並且比真實的半城市和城市森林更兼容，更有可能增強“遠離”的體驗。虛擬自然環境在運動過程中被認為比虛擬抽象畫更具恢復性（王等，2020）。宋等（2022）發現藍色環境是最具恢復性的，其次是開放綠地、半開放綠地、封閉綠地，最後是灰色空間。最後，一項研究發現五種不同虛擬環境（城市、自然和三種互動自然環境）的恢復質量沒有顯著差異（李、董等，2021）。

Regarding the characteristics of the natural environment perceived as more restorative, Tabrizian et al. (2018) showed that in the virtual park, higher ratings on restorativeness were found for the vegetation set on one side (as compared to equally distributed) and medium or high permeability vegetation, namely the extent to which spatial arrangement of vegetation affords seeing and moving (as compared to low). In Pals et al. (2014), the most restorative natural environment was that without furniture, then the one with wooden furniture, and the least restorative was the one with metal furniture. By contrast, in Reese, Mehner et al. (2022), there were no differences in restorativeness per level of wildness and presence of human structures. As for the characteristics of the VR experience, the study by Reese et al. (2021) showed that there were no statistical differences in perceived restorativeness attributed to a virtual coastal environment with or without participants' active control in the VR experience.
關於被認為更具恢復性的自然環境特徵，Tabrizian 等人（2018）顯示，在虛擬公園中，位於一側的植被組合（與均勻分佈相比）和中等或高通透性的植被在恢復性評分上較高，即植被的空間排列在視覺和移動上所提供的程度（與低通透性相比）。在 Pals 等人（2014）的研究中，最具恢復性的自然環境是沒有家具的環境，其次是有木製家具的環境，而金屬家具的環境則是最不具恢復性的。相對而言，在 Reese、Mehner 等人（2022）的研究中，野性程度和人類結構的存在對恢復性沒有影響。至於虛擬現實體驗的特徵，Reese 等人（2021）的研究顯示，參與者在虛擬現實體驗中是否有主動控制，對於虛擬海岸環境的感知恢復性並沒有統計上的差異。
3.3.2.5. Environment preference and pleasantness. Twelve articles
3.3.2.5. 環境偏好與愉悅感。十二篇文章
focused on the type of environments and the characteristics of the material shown on perceived preference and pleasantness (see Table 1). First, VR natural environments obtained higher scores of preference and pleasure as compared to photos and on-site exposure in one study (Gao, Liang, et al., 2019) and to VR urban exposure in two studies (Song et al., 2022; Ünal et al., 2022), while it obtained comparable ratings to on-site exposure in two studies (Xiang et al., 2021; Ünal et al., 2022), while in another study there were no differences between screen and VR exposure (Lindquist et al., 2020). As for the type of natural environment, an environment of relative complexity as the semi-open green space (vs. open and closed green spaces) and blue spaces obtained the highest preference score through VR (Gao, Liang, et al., 2019; Gao, Zhang, et al., 2019; Song et al., 2022; Xiang et al., 2021). In one study, preference was comparable to on-site for semi-open spaces in all seasons, while in the other two green spaces (open and closed), VR was comparable to on-site only during winter (Xiang et al., 2021). Presumably, this occurs for the potential of VR and the winter setting, to induce participants' focus on their feelings (Xiang et al., 2021). Higher preference for a green vs. desert VR environment in desert residents suggests that familiarity may play a role in preference (Yin et al., 2022).
專注於環境類型及材料特徵對感知偏好和愉悅感的影響（見表 1）。首先，虛擬現實自然環境在一項研究中相比於照片和現場暴露獲得了更高的偏好和愉悅分數（Gao, Liang, et al., 2019），並且在兩項研究中相比於虛擬現實城市暴露也獲得了更高的分數（Song et al., 2022; Ünal et al., 2022），而在另外兩項研究中，虛擬現實與現場暴露的評分相當（Xiang et al., 2021; Ünal et al., 2022），而在另一項研究中，螢幕和虛擬現實暴露之間沒有差異（Lindquist et al., 2020）。至於自然環境的類型，相對複雜的半開放綠地（與開放和封閉綠地相比）和藍色空間通過虛擬現實獲得了最高的偏好分數（Gao, Liang, et al., 2019; Gao, Zhang, et al., 2019; Song et al., 2022; Xiang et al., 2021）。在一項研究中，半開放空間在所有季節的偏好與現場相當，而在另外兩項綠地（開放和封閉）中，虛擬現實僅在冬季與現場相當（Xiang et al., 2021）。 推測這是因為虛擬實境的潛力和冬季環境，促使參與者專注於他們的感受（Xiang et al., 2021）。沙漠居民對綠色虛擬實境環境的偏好高於沙漠環境，這表明熟悉度可能在偏好中扮演了角色（Yin et al., 2022）。

Physical features in the natural environment can influence preference ratings. For instance, Pals et al. (2014) found that a natural environment with metal furniture was less preferred and pleasurable than an environment with wooden furniture or with no furniture at all. Similarly, van Vliet et al. (2021) focused on urban parks and found that the number of trees and the presence of flowerbeds with a diversity of flowers was related to preference more than the presence of services and the amount of litter. Their analysis shows how there are individual differences in park appreciation, some people giving importance mainly to natural elements and others evaluating the park as a whole. Lastly, in a study, the least preferred viewing distance of a green space was 20 m (as compared to 100 m and 200 m ), which was also associated with higher anxiety and depression (Shu et al., 2022). Here the edge permeability had no effect on preference.
自然環境中的物理特徵可以影響偏好評價。例如，Pals 等人（2014）發現，擁有金屬家具的自然環境比擁有木製家具或完全沒有家具的環境更不受喜愛和愉悅。同樣，van Vliet 等人（2021）專注於城市公園，發現樹木的數量和花壇中多樣化花卉的存在與偏好有關，這比服務的存在和垃圾的數量更為重要。他們的分析顯示，對公園的欣賞存在個體差異，有些人主要重視自然元素，而另一些人則將公園作為整體來評估。最後，在一項研究中，對綠地的最不喜歡的觀賞距離為 20 米（相比之下，100 米和 200 米），這也與較高的焦慮和抑鬱有關（Shu 等人，2022）。在這裡，邊緣通透性對偏好沒有影響。

Regarding multisensory experience, Hedblom et al. (2019) found that among the urban forest, urban park, and city, visual pleasantness was higher for the urban forest and lower for the urban environment; olfactory pleasantness was highest for the urban park and lowest for the city, and auditory pleasantness was lowest for the city, but no differences were found between the urban park and the urban forest. General pleasantness was higher for the urban park than for the forest and the urban area. Among different environments, Lindquist et al. (2020) found that a vacant lot was the least preferred environment and the garden the most preferred one, especially in VR (vs. screen). Moreover, the highest preference ratings were found for the realistic congruent sound conditions, then for the city sound, and, lastly, for the no sound conditions (Lindquist et al., 2020).
關於多感官體驗，Hedblom 等人（2019）發現，在城市森林、城市公園和城市環境中，城市森林的視覺愉悅度較高，而城市環境則較低；城市公園的嗅覺愉悅度最高，城市則最低，聽覺愉悅度在城市中最低，但城市公園和城市森林之間沒有發現差異。一般愉悅度在城市公園中高於森林和城市區域。在不同環境中，Lindquist 等人（2020）發現，空地是最不受歡迎的環境，而花園是最受歡迎的環境，特別是在虛擬實境（與螢幕相比）。此外，對於真實一致的聲音條件的偏好評分最高，其次是城市聲音，最後是無聲條件（Lindquist 等人，2020）。

The preference for an environment is related to higher positive affect (Gao, Zhang, et al., 2019), while the preferred scene (between two natural environments) reduced negative affect more than the second chosen scene and was related to a higher sense of presence in one study (Anderson et al., 2017). In one study (Ünal et al., 2022), restorativeness explained of the variance in preference and pleasantness for the VR natural environment. By contrast, Gao, Zhang, et al. (2019) found no correlation between preference for an environment and the restoration of attention, negative mood, and physiological stress recovery. Moreover, a multi-sensorial experience can impact stress; for instance, odor pleasantness reduced stress as measured with electrodermal activity, while auditory pleasantness was marginally significant, and visual pleasantness had no effect (Hedblom et al., 2019).
對環境的偏好與較高的正向情感有關（Gao, Zhang 等，2019），而在一項研究中，偏好的場景（兩個自然環境之間）比第二個選擇的場景更能減少負向情感，並與較高的存在感相關（Anderson 等，2017）。在另一項研究中（Ünal 等，2022），恢復性解釋了虛擬現實自然環境的偏好和愉悅感的變異性 。相對而言，Gao, Zhang 等（2019）發現環境偏好與注意力恢復、負向情緒和生理壓力恢復之間沒有相關性。此外，多感官體驗可以影響壓力；例如，氣味的愉悅感減少了通過皮膚電反應測量的壓力，而聽覺的愉悅感邊際上顯著，視覺的愉悅感則沒有影響（Hedblom 等，2019）。
3.3.2.6. Cognitive performance. Eleven studies focused on cognitive performance after exposure to a natural environment through VR. In general, it seems that the mere experience of a virtual natural environment does not affect participants' attentional resources in the majority of studies. Yeo et al. (2020) found that there were no differences in attention as measured through an objective task administered on the computer in the score pre- and post-exposure to either natural or urban virtual environments. Mostajeran et al. (2021) found that cognitive performance was higher after exposure to a natural (vs. urban) virtual environment assessed in terms of both mistakes and correct answers. No differences were found, however, for the immersion level (i.e., VR or photos). No differences were found between a virtual (vs. outdoor) nature walk in memory (Léger & Mekari, 2022).
3.3.2.6. 認知表現。十一項研究專注於在虛擬現實中接觸自然環境後的認知表現。一般來說，似乎僅僅體驗虛擬自然環境並不影響參與者的注意力資源，在大多數研究中都是如此。Yeo 等人（2020）發現，在接觸自然或城市虛擬環境前後，通過計算機進行的客觀任務測量的注意力沒有差異。Mostajeran 等人（2021）發現，在接觸自然（與城市相比）虛擬環境後，認知表現更高，無論是從錯誤數量還是正確答案來評估。然而，沉浸程度（即虛擬現實或照片）之間沒有發現差異。Léger 和 Mekari（2022）發現，虛擬（與戶外）自然散步在記憶方面沒有差異。

In Burmeister et al. (2018), objective concentration in terms of efficiency and accuracy was even higher in the virtual indoor setting (office) than in the virtual green environment showed, whereas no differences were found in accuracy and speed for calculating and counting, the speed increase in the overall test, and in the reports of subjective concentration. O'meara et al. (2020) in students with low and high anxiety test exposure to virtual natural or urban environments did not affect the change in the scores at a nonverbal reasoning test for condition, time, or level of anxiety. No differences were found in attention as measured by the Stroop task per different types of virtual environments (i.e., gray, blue, and four green spaces; Gao, Zhang, et al., 2019) and between a VR forest and a VR indoor environment (Zhang et al., 2023). In one study, auditory attention improved after a VR video in a sample of children (Luo et al., 2023).
在 Burmeister 等人（2018）的研究中，虛擬室內環境（辦公室）在效率和準確性方面的客觀集中度甚至高於虛擬綠色環境，而在計算和計數的準確性和速度上則沒有發現差異，整體測試的速度有所提高，主觀集中度的報告也顯示出相同的趨勢。O'meara 等人（2020）在低焦慮和高焦慮的學生中，對虛擬自然或城市環境的測試暴露並未影響非語言推理測試的分數變化，無論是條件、時間還是焦慮水平。在不同類型的虛擬環境（即灰色、藍色和四個綠色空間；Gao, Zhang 等，2019）中，根據 Stroop 任務測量的注意力沒有發現差異，虛擬森林和虛擬室內環境之間（Zhang 等，2023）也沒有差異。在一項研究中，兒童樣本在觀看虛擬現實視頻後，聽覺注意力有所改善（Luo 等，2023）。

Different effects were found for attentional resources measured after biofeedback combined with natural computer-generated virtual environments. In the study by Rockstroh et al. (2019), reports on concentration and distraction during exposure to stimuli were higher in the condition in which the biofeedback technique was combined with VR nature as compared to standard biofeedback and a non-treated control. Blum et al. (2019) observed that in the VR (vs. standard) biofeedback condition, participants reported higher focus on the present moment both in terms of the mind and the body, less task-relevant, and irrelevant mind wandering. Moreover, attentional resources were better conserved after the VR (vs. standard) biofeedback condition, since there was a significantly higher reduction in reaction times objectively measured with the Stroop task. Lastly, in Ch et al. (2023), the focus was higher in both VR and VR and mindfulness intervention conditions (daily exposure over a two-week period) as compared to a no-intervention condition.
在結合生物反饋與自然電腦生成虛擬環境後測量的注意資源中發現了不同的效果。在 Rockstroh 等人（2019）的研究中，報告指出在生物反饋技術與虛擬現實自然環境結合的條件下，參與者在接觸刺激時的專注和分心程度高於標準生物反饋和未處理的對照組。Blum 等人（2019）觀察到，在虛擬現實（相較於標準）生物反饋條件下，參與者在心智和身體上都報告了對當下的更高專注，並且任務相關和無關的心智漫遊較少。此外，在虛擬現實（相較於標準）生物反饋條件下，注意資源的保存更好，因為在客觀測量的 Stroop 任務中反應時間顯著減少。最後，在 Ch 等人（2023）的研究中，與無介入條件相比，虛擬現實和虛擬現實與正念介入條件（為期兩週的每日接觸）中的專注度更高。
3.3.2.7. Nature connectedness. Six articles investigated the relationship between exposure to virtual nature and connectedness with nature (see Table 1). Schutte et al. (2017) investigated if the benefits of virtual nature could be different for different individuals, testing connectedness to nature as a moderator between exposure to virtual natural (vs. urban) environment and positive affect. The results showed a marginally significant effect. In particular, when connectedness to nature was low there were no differences in the positive affect for the natural (vs. urban) environment. When connectedness to nature was high, there was a tendency for positive affect to be lower for the urban (vs. natural) environment.
3.3.2.7. 自然連結性。六篇文章研究了接觸虛擬自然與自然連結性之間的關係（見表 1）。Schutte 等人（2017）調查了虛擬自然的好處是否因個體而異，測試了自然連結性作為接觸虛擬自然（與城市環境相比）和正向情感之間的調節變量。結果顯示出邊際顯著的效果。特別是，當自然連結性較低時，自然（與城市）環境的正向情感沒有差異。當自然連結性較高時，城市（與自然）環境的正向情感傾向較低。

Five studies were interested in the change in connectedness to nature after exposure to virtual nature. Yeo et al. (2020) investigated connectedness to nature in relation to three different delivery modes (i. e., 2D screen, videos of real nature, and computer-generated virtual nature). Connectedness to nature was found to increase as the delivery mode became more immersive: thus, it was higher in the virtual nature than in 2D screen and videos. In particular, it seems that the effect of virtual nature on the increment in nature connectedness was mediated by the sense of presence given by the medium. One study (Sneed et al., 2021) measured connectedness to nature and interdependence with nature before and after exposure to virtual nature, a virtual indoor environment, and outdoor nature. Changes for connectedness to nature were higher in the outdoor condition than in both VR conditions and interdependence with nature was higher in the outdoor nature and
五項研究關注於在接觸虛擬自然後，與自然的連結感的變化。Yeo 等人（2020）調查了與三種不同傳遞模式（即 2D 螢幕、真實自然的 影片和電腦生成的虛擬自然）相關的自然連結感。研究發現，隨著傳遞模式變得更具沉浸感，自然連結感有所增加：因此，虛擬自然的連結感高於 2D 螢幕和 影片。特別是，虛擬自然對自然連結感增強的影響似乎是通過媒介所帶來的存在感來中介的。一項研究（Sneed 等人，2021）在接觸虛擬自然、虛擬室內環境和戶外自然之前和之後測量了與自然的連結感和與自然的相互依賴感。與自然的連結感在戶外條件下的變化高於兩種虛擬現實條件，而與自然的相互依賴感在戶外自然中更高。
virtual nature conditions than the indoor environment condition. No differences were found between outdoor and virtual nature conditions. In Leung et al. (2022), connectedness increased after two VR natural (but not after two VR urban) sessions. Interestingly, in a study, connectedness to nature increased after a VR natural (but not after a VR urban) condition, and in the VR natural (vs. urban) condition, both the increase in positive affect and the decrease in negative affect were explained through state connectedness to nature Chan et al., 2021).
虛擬自然條件比室內環境條件更佳。戶外和虛擬自然條件之間沒有發現差異。在 Leung 等人（2022）的研究中，經過兩次虛擬自然（但不是兩次虛擬城市）會議後，連結感有所增加。有趣的是，在一項研究中，對自然的連結感在虛擬自然（但不是虛擬城市）條件下增加，而在虛擬自然（與城市相比）條件下，正向情感的增加和負向情感的減少都通過對自然的狀態連結感來解釋（Chan 等人，2021）。

An emotional connection to the marine environment to which participants were exposed (virtually or in-situ) was measured by Hofman et al. (2021). In the virtual natural environment condition (as opposed to a real marine environment in which the effect was not found), this latter was positively and significantly associated with a specific set of future behavioral intentions to engage in conservative behaviors.
參與者對海洋環境的情感連結（無論是虛擬還是實地接觸）由 Hofman 等人（2021）進行測量。在虛擬自然環境條件下（與未發現效果的真實海洋環境相對），這種情感連結與一組特定的未來行為意圖正相關且顯著，這些行為意圖涉及參與保護行為。
3.3.2.8. Behavior and behavioral intentions. Four studies investigated behavior and behavioral intentions after being exposed to virtual natural environments. In particular, Hofman et al. (2021) showed how both being immersed in a marine environment and a real snorkel experience increased intentions to engage in conservative behaviors (i.e., educational and political actions, waste reduction, and purchasing choices), with no statistical differences between the two conditions. Leung et al. (2022) found that a VR nature (vs. urban) condition was associated with greater motivation for future nature engagement and, interestingly, this behavioral intention was associated with state connectedness to nature. In Tanja-Dijkstra et al. (2014), participants with high (vs. low) dental anxiety undergoing simulated dental treatment with an audiotape reproducing sounds of a real dental intervention, reported higher intention to use VR during a real dental visit. Lastly, only one study by Deringer and Hanley (2021) investigated real behavior, namely the willingness to sign a letter to a senator for supporting ecological actions and found that participants in both VR and outdoor nature conditions (vs. a no exposure condition) exhibited greater ecological behavior.
3.3.2.8. 行為與行為意圖。四項研究探討了在接觸虛擬自然環境後的行為和行為意圖。特別是，Hofman 等人（2021）顯示，沉浸在海洋環境中和真實的潛水體驗都增加了從事保守行為（即教育和政治行動、減少浪費和購買選擇）的意圖，兩種情況之間沒有統計上的差異。Leung 等人（2022）發現，虛擬現實自然（與城市相比）條件與未來參與自然活動的動機更大相關，並且有趣的是，這種行為意圖與對自然的連結感有關。在 Tanja-Dijkstra 等人（2014）的研究中，經歷高（與低）牙科焦慮的參與者在接受模擬牙科治療時，聆聽重現真實牙科介入聲音的錄音帶，報告了在真實牙科就診時使用虛擬現實的意圖更高。 最後，只有一項由 Deringer 和 Hanley（2021）進行的研究調查了真實行為，即願意簽署一封信給參議員以支持生態行動，並發現無論是在虛擬實境還是戶外自然環境中（與無接觸條件相比），參與者表現出更大的生態行為。
3.3.2.9. Creativity. Three studies hypothesized that exposure to virtual nature could enhance creativity. In the study by Fleury et al. (2021), participants were asked to sketch an innovative item as a work post for people in a wheelchair while immersed in a virtual natural, indoor, and neutral environment. Overall creativity was higher in the virtual nature condition as compared to the other two, especially regarding the sub-criterion of novelty. In Palanica et al. (2019), participants were exposed to a natural or an urban environment in 2D or 3D mode and, in the meantime, they were asked to report possible alternative uses of a brick (creative task). Results attested that was the type of environment (natural or urban) that was related to increased creativity, but not the medium used (2D or 3D). In Ch et al. (2023), creativity was assessed daily over two weeks (per condition) and results showed that convergent thinking was higher in the condition of VR nature combined with mindfulness than in VR nature alone and control (no intervention), while divergent thinking worsened in the intervention conditions vs. control.
3.3.2.9. 創造力。三項研究假設接觸虛擬自然可以增強創造力。在 Fleury 等人（2021）的研究中，參與者被要求在沉浸於虛擬自然、室內和中立環境中時，為輪椅使用者草擬一個創新項目作為工作帖子。與其他兩種情況相比，虛擬自然條件下的整體創造力更高，特別是在新穎性這一子標準上。在 Palanica 等人（2019）的研究中，參與者在 2D 或 3D 模式下接觸自然或城市環境，同時被要求報告磚塊的可能替代用途（創意任務）。結果證明，與創造力增強相關的是環境類型（自然或城市），而不是使用的媒介（2D 或 3D）。在 Ch 等人（2023）的研究中，創造力在兩週內每天進行評估（每種條件），結果顯示，結合正念的虛擬自然條件下的聚合思維高於僅有虛擬自然和控制組（無干預），而在干預條件下的發散思維則較控制組惡化。
3.3.2.10. Perceived safety. Three studies investigated perceived safety with VR. Evensen et al. (2021) manipulated a hedge height in the evening in a virtual urban park but found no differences in perceived safety, which differed only by gender with females feeling less safe. On the same line, Shu et al. (2022) found no differences in perceived safety per viewing distance and edge permeability of a green space. By contrast, in Tabrizian et al. (2018), being surrounded by trees in an urban park was related to less safety and this effect was not present in the urban condition (plaza). Also, the park with low (vs. high) permeability was perceived as less safe.
3.3.2.10. 感知安全性。三項研究調查了虛擬實境中的感知安全性。Evensen 等人（2021）在虛擬城市公園的晚上操控了樹籬的高度，但發現感知安全性沒有差異，僅因性別而異，女性感到不太安全。同樣，Shu 等人（2022）發現根據觀看距離和綠地的邊緣通透性，感知安全性沒有差異。相反，在 Tabrizian 等人（2018）的研究中，被樹木包圍的城市公園與較低的安全感相關，而這種效應在城市環境（廣場）中並不存在。此外，低通透性（與高通透性相比）的公園被認為不太安全。
3.3.2.11. Subjective vitality. Three studies investigated subjective vitality as a consequence of exposure to virtual nature. In Mattila et al.
3.3.2.11. 主觀活力。三項研究探討了虛擬自然暴露對主觀活力的影響。在 Mattila 等人。
(2020), the authors found that exposure to a VR forest significantly increased the subjective vitality experienced by participants. In Reese, Mehner et al. (2022) subjective vitality was higher after VR, with no differences between environments per level of wildness and human structure presence. Subjective vitality did not change after exposure to VR and outdoor nature and no differences between conditions were found in Reese, Stahlberg et al. (2022).
（2020）作者發現，接觸虛擬現實森林顯著提高了參與者的主觀活力。在 Reese、Mehner 等人（2022）的研究中，虛擬現實後的主觀活力較高，且在野性和人類結構存在的環境層級之間沒有差異。在 Reese、Stahlberg 等人（2022）的研究中，接觸虛擬現實和戶外自然後，主觀活力沒有變化，且在條件之間沒有發現差異。
3.3.2.12. Dental experience. Three studies (Tanja-Dijkstra et al., 2014; 2018, Study 1 and 2) investigated how dental experience could be ameliorated in healthy adults and dental patients with exposure to VR nature. In Tanja-Dijkstra et al. (2014), healthy adults underwent a simulated dental treatment with an audiotape while exposed to a VR coastal environment (with active or passive control) or a black screen (control condition). They were additionally split at the median of their scores in dental anxiety. Results showed how participants with high (vs. low) dental anxiety in both VR conditions (active and passive control) had less vivid memories one week later than control (black screen). No such differences were found for immediate dental experience and intrusive thoughts one week later. A similar experimental apparatus was used for the subsequent studies (Tanja-Dijkstra et al., 2018). In Study 1, pain experience was elicited in healthy adults using a laboratory cold pressor task. Results showed that VR conditions (both active and passive) resulted in a reported less experienced and recollected pain after a week. In Study 2, patients undergoing a real dental treatment were exposed to a VR condition (coastal or urban environment) or to a control condition in which they simply received the healthcare treatment. Results supported the effect of VR nature on experienced and recollected pain, while the same effect was not found for the VR urban condition.
3.3.2.12. 牙科經驗。三項研究（Tanja-Dijkstra 等，2014；2018，研究 1 和 2）探討了如何在健康成人和牙科病人中通過接觸虛擬現實自然環境來改善牙科經驗。在 Tanja-Dijkstra 等（2014）中，健康成人在虛擬現實海岸環境（有主動或被動控制）或黑屏（控制條件）下，聆聽錄音帶進行模擬牙科治療。他們的牙科焦慮得分還被分為中位數以上和以下。結果顯示，在兩種虛擬現實條件下（主動和被動控制），高牙科焦慮的參與者在一週後的回憶中比控制組（黑屏）有較少的生動記憶。對於即時牙科經驗和一週後的侵入性思維，未發現此類差異。隨後的研究（Tanja-Dijkstra 等，2018）使用了類似的實驗設備。在研究 1 中，使用實驗室冷壓任務引發健康成人的疼痛經驗。結果顯示，虛擬現實條件（主動和被動）在一週後報告的疼痛經歷和回憶較少。 在研究 2 中，接受真實牙科治療的患者被暴露於虛擬實境條件（海岸或城市環境）或控制條件，在控制條件下他們僅接受醫療治療。結果支持了虛擬實境自然環境對經歷和回憶疼痛的影響，而對虛擬實境城市環境則未發現相同的效果。

Few studies focused on the mechanisms underlying the benefits of virtual nature on individuals. Regarding restorativeness, Tabrizian et al. (2018), based on the characteristics of the natural and urban environments shown, found that spatial arrangement of natural objects negatively predicted perceived personal safety (one-sided rather than two- or four-sided objects predicted a higher sense of safety) which, in turn, positively predicted perceived restorativeness. This model was true for the natural environment but not for the urban environment. Vegetation permeability (low, medium, high) negatively predicted perceived safety which, in turn, positively predicted restorativeness, again for the natural environment but not for the urban environment.
很少有研究專注於虛擬自然對個體益處的機制。關於恢復性，Tabrizian 等人（2018）根據所展示的自然和城市環境的特徵發現，自然物體的空間排列對感知的個人安全有負面預測（單面物體而非雙面或四面物體預測出更高的安全感），而這反過來又正面預測了感知的恢復性。這一模型在自然環境中成立，但在城市環境中則不成立。植被通透性（低、中、高）對感知安全有負面預測，而這又正面預測了恢復性，同樣適用於自然環境，但不適用於城市環境。

Regarding emotions, Yeo et al. (2020) found that the interactive computer-generated VR (vs. the 2D video) condition was associated with greater improvements in positive affect, and this effect was sequentially mediated by both a higher sense of presence and a higher connectedness to nature. In other words, exposure to the VR condition provoked a higher sense of presence that, in turn, enhanced connectedness to nature, which ultimately predicted greater positive affect. Following the same line, Li, Dong, et al. (2021) showed how the effects of restorativeness on positive and negative emotions and general self-efficacy were mediated by the sense of presence. In Schutte et al. (2017), restorativeness was a mediator in the relationship between VR exposure (both natural and urban conditions) and positive affect. In Chan et al. (2021), in the relationship between VR exposure and positive and negative affect, connectedness to nature was found as a significant mediator. Also, in the study by Pals et al. (2014), the coherence dimension of restorativeness fully mediated the negative effect of metal furniture in a natural environment (compared with a condition without furniture) on preference, pleasure, and restoration and partially mediated the contrast effects of metal vs. wooden furniture on preference and pleasure.
關於情感，Yeo 等人（2020）發現互動式電腦生成的虛擬實境（與 2D 影片相比）條件與正向情感的改善有更大的關聯，這一效果依次受到更高的存在感和更高的自然連結感的中介。換句話說，接觸虛擬實境條件引發了更高的存在感，進而增強了與自然的連結，最終預測了更大的正向情感。沿著這一思路，Li、Dong 等人（2021）展示了恢復性對正向和負向情感以及一般自我效能的影響是如何受到存在感的中介。在 Schutte 等人（2017）的研究中，恢復性是虛擬實境暴露（包括自然和城市條件）與正向情感之間關係的中介。在 Chan 等人（2021）的研究中，在虛擬實境暴露與正向和負向情感之間的關係中，自然連結被發現是一個重要的中介。此外，在 Pals 等人的研究中。 （2014），恢復性的一致性維度完全中介了金屬家具在自然環境中（與無家具的情況相比）對偏好、愉悅和恢復的負面影響，並部分中介了金屬家具與木質家具在偏好和愉悅上的對比效應。

Other findings suggest the same mediation paths since they found a positive correlation between the sense of presence and both negative and positive affect, in the virtual but not in the natural condition
其他研究結果顯示相同的中介路徑，因為他們發現虛擬環境中的存在感與負面和正面情感之間存在正相關，但在自然環境中則沒有
(Chirico & Gaggioli, 2019). The sense of presence was higher in the interactive (vs. non-interactive) VR condition and VR (vs. desktop) condition, and negatively predicted anxiety and positively predicted positive affect (Liszio et al., 2018; Liszio & Masuch, 2019). Only one study used electromyography (EMG), in terms of contraction of the brachioradialis muscle of the arm, as an indicator of presence and physical participation. Results indicated that pre- and post-exposure score changes were significantly higher for all conditions (urban, natural interactive, and noninteractive environments) providing differences in presence as indicated also by the results using scale measurements (Li, Dong, et al., 2021).
（Chirico & Gaggioli, 2019）。在互動（相較於非互動）虛擬實境條件下，存在感較高，且在虛擬實境（相較於桌面）條件下亦然，並且對焦慮有負向預測，對正向情感有正向預測（Liszio et al., 2018；Liszio & Masuch, 2019）。只有一項研究使用了肌電圖（EMG），以手臂的肱橈肌收縮作為存在感和身體參與的指標。結果顯示，所有條件（城市、自然互動和非互動環境）的前後暴露分數變化顯著較高，提供了存在感的差異，這一點也得到了使用量表測量的結果的支持（Li, Dong, et al., 2021）。

Lastly, several studies tested the effect of possible moderators in the relationship between VR nature exposure and psychological and psychophysiological outcomes. Browning, Shipley, et al. (2020) found no interaction (moderating) effects of previous VR experience with condition (outdoor nature vs. virtual nature vs. indoor setting) on both positive and negative affect and restorativeness, indicating that previous VR experience did not affect the relationship between the condition, restorativeness, and mood. Knaust et al. (2022) found no interaction effects of gender, age, technology anxiety, and previous VR experience with the type of medium and time (pre-post-exposure) on electrodermal activity and perceived relaxation. Yin et al. (2022) found no interaction effects between preference for a green or brown landscape and exposure to those landscapes and a control condition on cortisol levels, mean arterial pressure, and serum interleukin in desert residents. By contrast, Browning et al. (2023) found significant findings in the condition of daily VR nature exposure: Females (vs. males), those with more (vs. less) VR experience, with more (vs. less) outdoor nature experience, and those with lower (vs. higher) pre- and post-exposure engagement with beauty showed greater decreases in worry. Lastly, in two studies, participants with high (vs. low) anxiety exhibited more benefits from exposure to VR nature (O'meara et al., 2020; Tanja-Dijkstra et al., 2014).
最後，幾項研究測試了可能的調節變數在虛擬現實自然暴露與心理及心理生理結果之間的關係中的影響。Browning、Shipley 等人（2020）發現，先前的虛擬現實經驗與條件（戶外自然 vs. 虛擬自然 vs. 室內環境）之間對正面和負面情感及恢復力沒有交互（調節）效應，這表明先前的虛擬現實經驗並未影響條件、恢復力和情緒之間的關係。Knaust 等人（2022）發現，性別、年齡、技術焦慮和先前的虛擬現實經驗與媒介類型和時間（前後暴露）之間對皮膚電活動和感知放鬆沒有交互效應。Yin 等人（2022）發現，對綠色或棕色景觀的偏好與暴露於這些景觀及控制條件之間對沙漠居民的皮質醇水平、平均動脈壓和血清白介素沒有交互效應。相反，Browning 等人（2023）在每日虛擬現實自然暴露的條件下發現了顯著的結果：女性（vs. 男性）、擁有更多（vs. 更少）虛擬現實經驗的人，擁有更多（vs. 較少的戶外自然體驗，以及那些在接觸美的前後參與度較低（相較於較高）的人，顯示出更大的擔憂減少。最後，在兩項研究中，焦慮程度高（相較於低）的參與者從虛擬現實自然體驗中獲得了更多的益處（O'meara et al., 2020; Tanja-Dijkstra et al., 2014）。

Immersive VR video tends to induce a better sense of presence and immersion compared to a VR photo slideshow (Liszio et al., 2018; Mostajeran et al., 2021) especially in interactive virtual environments (Yeo et al., 2020) or in conditions in which the participants are in control and can explore the environment freely (Tanja-Dijkstra et al., 2014, 2018). Sense of presence has also been shown to be associated with lower anxiety and higher positive affect (Liszio & Masuch, 2019). On the contrary, motion (or cyber or simulator) sickness may cause a state of physical discomfort characterized by symptoms including dizziness, cold sweat nausea, or even vomiting. The aforementioned sensations are due to a mismatch of information among the brain, body, eyes, and ears, but also psychological components may play a role, such as memory of past discomfort in a similar situation (Dobie, 2019, pp. 113-127). A number of studies presented in this systematic review addressed this issue by administering appropriate items or scales and reported high ratings of motion sickness among participants exposed to VR, as reported by O'meara et al. (2020). Despite the exclusion of participants with discomfort as an essential precondition for each study, as observed, negative VR experience impacted the pleasantness of the experience itself in many cases, both in a sitting and in a treadmill condition (as in Calogiuri et al., 2018) and regardless of whether or not the experimenter helped participants during the virtual navigation (as in Reese et al., 2021). In summary, a satisfactory compromise between a high sense of reality and immersion in VR and the risk of discomfort is yet to be discovered.
沉浸式虛擬實境（VR）視頻相比於虛擬實境照片幻燈片（Liszio et al., 2018; Mostajeran et al., 2021），往往能引發更好的存在感和沉浸感，尤其是在互動虛擬環境中（Yeo et al., 2020）或參與者能夠自由探索環境的情況下（Tanja-Dijkstra et al., 2014, 2018）。存在感也被證明與較低的焦慮和較高的正向情感相關（Liszio & Masuch, 2019）。相反，運動（或網路或模擬器）暈動症可能會導致身體不適，特徵包括頭暈、冷汗、噁心甚至嘔吐。上述感覺是由於大腦、身體、眼睛和耳朵之間信息的不匹配，但心理因素也可能起作用，例如對於類似情況下過去不適的記憶（Dobie, 2019, pp. 113-127）。本系統評價中呈現的多項研究通過施用適當的項目或量表來解決這一問題，並報告了在接觸虛擬實境的參與者中運動暈動症的高評分，正如 O'meara et al.（2020）所報告的。 儘管在每項研究中排除了感到不適的參與者作為基本前提，但如所觀察到的，負面的虛擬實境體驗在許多情況下影響了體驗本身的愉悅感，無論是在坐著的情況下還是在跑步機上（如 Calogiuri 等，2018 年），也無論實驗者是否在虛擬導航過程中協助參與者（如 Reese 等，2021 年）。總之，尚未發現高現實感和沉浸感與不適風險之間的滿意妥協。

A further experience-related issue is the likely psychological effect of repeated exposure through VR technology. It has been speculated that positive psychophysiological effects of VR may be also determined by perception evoked by previous virtual experience, however, findings were inconsistent. In participants with previous VR experience a higher, albeit slightly, level of ecological validity in a natural over a virtual condition was detected (Chirico & Gaggioli, 2019), while no significant interaction effect of the novelty of the medium was found between exposure condition and psychological outcomes, such as affects and restorativeness (Browning, Mimnaugh, et al., 2020). The potential intervening effect of previous VR experience is certainly something to be monitored. A possible strategy to overcome this issue is to select participants with the same VR experience, e.g., with no or little previous experience (as in Rockstroh et al., 2019).
另一個與經驗相關的問題是通過虛擬現實技術重複接觸可能產生的心理影響。有人推測，虛擬現實的正面心理生理效應可能也受到先前虛擬經驗所引發的感知影響，然而，研究結果並不一致。在具有先前虛擬現實經驗的參與者中，發現自然條件下的生態有效性水平略高於虛擬條件（Chirico & Gaggioli, 2019），而在接觸條件與心理結果（如情感和恢復性）之間，並未發現媒介新穎性有顯著的交互作用效應（Browning, Mimnaugh, et al., 2020）。先前虛擬現實經驗的潛在干預效應無疑是需要監測的問題。一種克服此問題的可能策略是選擇具有相同虛擬現實經驗的參與者，例如，沒有或只有少量先前經驗的參與者（如 Rockstroh et al., 2019）。

Our review complements and expands available knowledge on this research topic, both from a chronological and a content point of view (i. e., findings on environment preference and pleasantness, cognitive performance, nature connectedness, behavior and behavioral intentions, creativity, perceived safety, subjective vitality, and dental experience were novel as compared to previous systematic reviews available, i.e. Frost et al., 2022; Riches et al., 2021). To our knowledge, to date, it constitutes the most updated and complete review on the effectiveness of virtual nature on psychological and psychophysiological outcomes. Efforts were made in summarizing existing evidence on the effectiveness of virtual nature on psychological outcomes, however, findings are available for specific outcomes or populations. Two recent systematic reviews (Frost et al., 2022; Riches et al., 2021) focused on the effect of virtual reality on well-being and relaxation and of virtual nature on well-being in general, respectively. A meta-analysis (Browning, Shipley, et al. (2020) is also available on the effect of virtual nature on the positive and negative affect in studies that report a comparison between in-vivo and virtual exposure. Moreover, a narrative review (White et al., 2018) is available on the effect of virtual nature in therapeutic settings.
我們的評論補充並擴展了在這一研究主題上的可用知識，無論是從時間順序還是內容的角度（即，關於環境偏好和愉悅感、認知表現、與自然的聯結、行為和行為意圖、創造力、感知安全性、主觀活力和牙科經驗的發現，相較於以往的系統性評論（如 Frost 等，2022；Riches 等，2021）都是新穎的。據我們所知，迄今為止，這是關於虛擬自然對心理和心理生理結果的有效性最更新和最完整的評論。雖然在總結虛擬自然對心理結果有效性的現有證據方面做出了努力，但發現僅針對特定結果或人群。兩篇最近的系統性評論（Frost 等，2022；Riches 等，2021）分別集中於虛擬現實對幸福感和放鬆的影響，以及虛擬自然對一般幸福感的影響。一項元分析（Browning, Shipley 等）。 （2020）也有關於虛擬自然對正面和負面情感影響的研究，這些研究報告了實地與虛擬暴露之間的比較。此外，還有一篇敘述性回顧（White et al., 2018）探討了虛擬自然在治療環境中的效果。

All studies reviewed included virtual nature exposures, featuring various types of environments such as green and blue spaces. The length and media of exposure varied, with immersive virtual experiences being more effective. Different natural environments had diverse effects on attention, mood, stress levels, and preference. Both virtual nature and urban exposures reduced heart rate, but psychological benefits were greater with virtual nature exposure. No substantial differences were found comparing virtual nature with in-vivo exposure to a naturalistic setting. Lastly, the experience of virtual nature exposure has consistently proven to be preferable over exposure to an indoor virtual environment, such as an office.
所有回顧的研究都包括虛擬自然暴露，涵蓋各種環境類型，如綠地和藍地。暴露的長度和媒介各不相同，沉浸式虛擬體驗更為有效。不同的自然環境對注意力、情緒、壓力水平和偏好有不同的影響。虛擬自然和城市暴露均降低了心率，但虛擬自然暴露的心理益處更大。比較虛擬自然與自然環境的實地暴露，未發現顯著差異。最後，虛擬自然暴露的體驗始終被證明比室內虛擬環境（如辦公室）的暴露更受歡迎。

The outcomes considered in the present review were all those identifiable as psychological and psychophysiological, therefore no further selection based on psychological domains (e.g., cognitive functioning, affect and emotion, psychosocial abilities, personality traits) was carried out. Overall, the results of the available studies attest to how VR is effective in producing effects among its users. The main findings drawn from this extensive review are summarized in Table 1.
本次回顧考慮的結果均可識別為心理和心理生理，因此未根據心理領域（例如，認知功能、情感與情緒、心理社會能力、人格特質）進行進一步的選擇。總體而言，現有研究的結果證實了虛擬實境在其使用者中產生效果的有效性。本次廣泛回顧的主要發現總結於表 1 中。

First of all, the benefits of VR exposure were found in the domain of human emotions. In most studies, VR exposure decreased negative affect and some studies suggest an increase in positive affect as well. Virtual nature exposure also seems beneficial for stress reduction, with psychophysiological measurements pointing to significant relaxation, and restorativeness. Regarding preference, it seems generally higher for virtual semi-open green spaces and personal predilection for a natural environment can be crucial in determining the benefits associated with exposure. Cognitive performance seems not to be stimulated by virtual nature alone, while attention is amplified when in combination with biofeedback. A promising result that still needs further support is that state connectedness to nature seems to grow with the level of immersion, it is comparable to outdoor fruition and explains changes in affect while more studies are needed to investigate the role of dispositional connectedness to nature. Lastly, initial results were found for behavior and behavioral intentions, creativity, perceived safety, subjective vitality, and dental experience.
首先，虛擬實境（VR）暴露的好處主要體現在人類情感領域。在大多數研究中，VR 暴露減少了負面情感，並且一些研究還表明正面情感有所增加。虛擬自然暴露似乎對減壓也有益，心理生理測量顯示出顯著的放鬆和恢復效果。關於偏好，對虛擬半開放綠地的偏好普遍較高，而對自然環境的個人喜好在決定暴露所帶來的好處方面可能至關重要。僅僅依賴虛擬自然似乎無法刺激認知表現，而當與生物反饋結合時，注意力則會增強。一個仍需進一步支持的有希望的結果是，與自然的連結感似乎隨著沉浸程度的提高而增長，這與戶外享受相當，並解釋了情感的變化，但仍需更多研究來探討與自然的性格連結的角色。最後，初步結果顯示在行為和行為意圖、創造力、感知安全性、主觀活力和牙科經驗方面的發現。

Noteworthy is the investigation of previous or current VR experience. The present review pointed out the crucial role of immersive VR video in providing a sense of presence; however, the other side of the coin is represented by the risk of incurring sensations of discomfort due to the virtual experience. Added to this, previous negative VR experiences may prevent users from benefiting from virtual nature exposure. The strategies implemented so far, such as selecting participants with the same VR experience, or excluding those experiencing discomfort, to overcome these issues need to be further developed.
值得注意的是對於過去或當前虛擬實境（VR）體驗的調查。本次回顧指出沉浸式 VR 視頻在提供存在感方面的關鍵作用；然而，另一方面則是由於虛擬體驗而產生不適感的風險。此外，以前的負面 VR 體驗可能會阻礙用戶從虛擬自然暴露中獲益。迄今為止實施的策略，如選擇具有相同 VR 體驗的參與者或排除那些感到不適的人，為了解決這些問題需要進一步發展。

The present review is not without limitations. First, we were unable to use a meta-analytic approach to the available data due to high heterogeneity in terms of the exposure methods, VR experience and instruments used, and the psychological outcomes considered. Moreover, investigating the effect of precise exposure types on specific psychological outcomes is highly warranted. As a second limitation, although our efforts in merging two standardized, commonly used checklists for the creation of a suitable instrument for the quality assessment of all the included studies, we acknowledge the use of a non-standardized instrument.
本次回顧並非沒有局限性。首先，由於暴露方法、虛擬實境體驗和所使用的工具以及考慮的心理結果之間存在高度異質性，我們無法對可用數據採用元分析方法。此外，調查精確暴露類型對特定心理結果的影響是非常必要的。作為第二個局限性，儘管我們努力合併兩個標準化的、常用的檢查表，以創建一個適合所有納入研究的質量評估工具，但我們承認使用了一個非標準化的工具。

Lastly, the reliability of the findings of our systematic review, as for any other review, can be undermined by the so-called publication bias. Therefore, further efforts should focus on determining the entity of such bias and the implications for any attempt to summarize the research findings, considering also the gray literature. It is also worth noting that, although we acknowledge that offering the reader(s) a compartmentalized synthesis of available findings could be seen as a study limitation for the risk of incurring a superficial investigation of the topic, we believe this is a strong point of our review since we allow them to specifically focus on the variable(s) of interest.
最後，我們系統性回顧的發現可靠性，與其他任何回顧一樣，可能會受到所謂的出版偏差的影響。因此，進一步的努力應該集中在確定這種偏差的實體及其對任何試圖總結研究結果的影響，同時也要考慮灰色文獻。值得注意的是，儘管我們承認為讀者提供可用發現的分門別類的綜合可能被視為研究的局限性，因為這可能導致對主題的表面性調查，但我們認為這是我們回顧的一個強項，因為我們允許他們專注於感興趣的變數。

In general, findings provide an incomplete picture. For example, we criticize the lack of studies in younger populations, such as school-age children and adolescents. Similarly, although present, the results on the elderly and populations with special needs are still insufficient. Studies on different contexts are also missing, such as schools, prisons, or retirement homes. Moreover, from the quality assessment we carried out, it appears that very few studies considered the power of the effects observed, while from an analytic overview of the studies' characteristics, it emerges that several studies employed very low sample sizes. This latter consideration warns the reader to examine the findings carefully, especially regarding those outcomes on which studies are few. However, as Reese, Mehner et al. (2022), argue, VR studies are always subject to feasibility considerations, given that VR studies require a relatively high investment of resources, given their relatively complicated setups (i.e., compared to questionnaire studies or online panels). Consequently, accumulating evidence through reviews and meta-analyses should help in interpreting the findings with more confidence.
一般來說，研究結果提供了一個不完整的畫面。例如，我們批評在年輕人群體中缺乏研究，例如學齡兒童和青少年。同樣，雖然有關於老年人和特殊需求人群的結果存在，但仍然不足。對於不同情境的研究也缺乏，例如學校、監獄或養老院。此外，根據我們進行的質量評估，似乎很少有研究考慮到觀察到的效果的效能，而從對研究特徵的分析概述中可以看出，幾項研究使用了非常小的樣本量。這一點提醒讀者仔細檢查研究結果，特別是對於那些研究較少的結果。然而，正如 Reese、Mehner 等（2022）所指出的，虛擬現實研究總是受到可行性考量的影響，因為虛擬現實研究需要相對較高的資源投入，因其設置相對複雜（即，相較於問卷研究或在線小組）。因此，通過綜述和元分析積累證據應有助於更有信心地解釋研究結果。

Most of the studies focused on three main outcomes, namely mood, stress, and restorativeness. Other outcomes, such as environmental preference and pleasantness, cognitive performance, and nature connectedness were less investigated, while there is a notable lack of studies for behavior and behavioral intentions, creativity, perceived safety, subjective vitality, and dental experience. Thus, there is a need for future studies to investigate further the underestimated outcomes, through different instruments and experimental designs. For instance, since the relationship between outdoor nature exposure and cognitive processes is well-known (Bratman et al., 2015), it would be crucial to better understand whether, and under what conditions, virtual nature can have comparable effects. Here, the available evidence points to the sense of presence as an important catalyst of virtual nature benefits. It would also appear to be interesting to test the role of multisensorial virtual experiences and other variables. Moreover, further evidence should be also collected on the effect of virtual nature on restoration through understudied biosignals, such as salivary cortisol and salivary amylase.
大多數研究集中於三個主要結果，即情緒、壓力和恢復性。其他結果，如環境偏好和愉悅感、認知表現以及與自然的聯結，則較少受到研究，而行為和行為意圖、創造力、感知安全性、主觀活力以及牙科經驗的研究則明顯不足。因此，未來的研究需要進一步探討這些被低估的結果，通過不同的工具和實驗設計。例如，既然戶外自然暴露與認知過程之間的關係已經廣為人知（Bratman et al., 2015），那麼更好地理解虛擬自然是否以及在什麼條件下能夠產生可比擬的效果將是至關重要的。在這裡，現有的證據指出存在感是虛擬自然益處的重要催化劑。測試多感官虛擬體驗和其他變量的作用似乎也很有趣。此外，還應收集更多關於虛擬自然對恢復影響的證據，通過尚未充分研究的生物信號，如唾液皮質醇和唾液澱粉酶。

Additional outcomes could be investigated in the future. In this regard, it seems that studies are missing the role of personality in the appreciation of virtual nature. In the studies considered, very few focused on personal dispositions such as engagement with beauty and disgust sensitivity (Browning, Shipley, et al., 2020), disposition towards an ecological worldview (Hofman et al., 2021), and nature connectedness (Schutte et al., 2017). It is possible that, for instance, specific individual characteristics can more strongly predispose some people to the actual recognition of virtual nature benefits. Available evidence has already shown how VR appreciation can vary among individuals as well as how the benefits of outdoor nature can be recognized by some individuals more than others (Feng et al., 2021; Kober & Neuper, 2013; Panno et al., 2020). Similarly, as recently highlighted, some individuals, more than others, could experience fewer benefits associated with virtual nature fruition; thus, future studies on individual differences could understand how to extend these benefits to more people (Theodorou, Spano et al., 2023). Moreover, studies are missing on the potential influence of the medium on behavioral change. For instance, it would be interesting to extend some results from the research on on-site nature exposure to VR. Furthermore, it would be interesting to see if virtual nature can impact pro-environmental behavior, prosocial behavior, cooperation, and aggression (Zelenski, Dopko, & Capaldi, 2015).
未來可以研究其他結果。在這方面，似乎研究忽略了個性在欣賞虛擬自然中的角色。在考慮的研究中，只有少數專注於個人特質，例如對美的投入和厭惡敏感性（Browning, Shipley 等，2020）、對生態世界觀的傾向（Hofman 等，2021）以及與自然的連結感（Schutte 等，2017）。例如，特定的個人特徵可能會更強烈地使某些人傾向於實際認識虛擬自然的好處。現有證據已經顯示，虛擬現實的欣賞在個體之間可能有所不同，以及某些個體比其他人更能認識戶外自然的好處（Feng 等，2021；Kober & Neuper，2013；Panno 等，2020）。同樣，最近強調的，有些個體可能比其他人經歷的虛擬自然享受的好處較少；因此，未來對個體差異的研究可以了解如何將這些好處擴展到更多人（Theodorou, Spano 等，2023）。 此外，尚缺乏關於媒介對行為改變潛在影響的研究。例如，將現場自然接觸的研究結果擴展到虛擬實境中將會很有趣。此外，了解虛擬自然是否能影響環保行為、親社會行為、合作和攻擊性也將是有趣的（Zelenski, Dopko, & Capaldi, 2015）。

Further advancements in the understanding of the psychological mechanisms underlying the observable benefits are desirable. In general, the studies that focused on the intervening variables (mediators) extended from outdoor nature exposure to the VR medium the mediating effect of some variables (i.e., perceived safety, restorativeness, and connectedness to nature) well-known for translating the characteristics of the natural environments in positive outcomes for the individuals (i. e., mood, restorativeness, and preference). Nevertheless, what stands out as the most important mediator specifically related to the potential of the VR medium is the sense of presence. This means that future studies, as well as interventions aimed at using this tool, should investigate further the methodological characteristics of the research design or the intervention that could enhance the sense of presence, such as multisensory stimuli. Moreover, further focus is needed on the investigation of the direction of the relationships observed. Except for exposure, which in this kind of study is manipulated, other research designs may also manipulate the proposed mediators (e.g., restorativeness) to understand if they truly are antecedents of the outcomes observed (e.g., affect). This will help establish new theories. Lastly, few studies focused on moderators, and further studies are needed to identify potential populations and conditions that can enhance or diminish the effects observed.
進一步了解可觀察益處背後的心理機制是可取的。一般而言，專注於中介變數的研究從戶外自然接觸擴展到虛擬實境媒介，某些變數（即感知安全性、恢復性和與自然的連結）的中介效應在將自然環境的特徵轉化為個體的正面結果（即情緒、恢復性和偏好）方面是眾所周知的。然而，與虛擬實境媒介潛力特別相關的最重要的中介變數是存在感。這意味著未來的研究以及旨在使用這一工具的干預措施應進一步探討研究設計或干預的 методологические 特徵，以增強存在感，例如多感官刺激。此外，還需要進一步關注觀察到的關係方向的研究。除了在這類研究中被操控的接觸外，其他研究設計也可能操控所提出的中介變數（例如）。，恢復性）以了解它們是否真的是觀察到的結果（例如，情感）的前因。這將有助於建立新的理論。最後，少數研究專注於調節變數，還需要進一步的研究來識別可能增強或減少觀察到的效果的人群和條件。

Identifying mediators and moderators could also help the development of new theories, specifically conceptualized to predict the effects of virtual nature. What we can observe so far from the articles selected is that, mostly, the theories used in formulating the hypothesis and interpreting the findings are the same used in the context of exposure to outdoor nature (e.g., the ART; Kaplan & Kaplan, 1989). This is not surprising given that the area of research is still early in its development. Nevertheless, we can speculate that more predictive theories in the future will consider the specificities of the VR medium, such as the sense of presence.
識別中介變數和調節變數也有助於新理論的發展，特別是專門用來預測虛擬自然效果的理論。從所選文章中我們目前可以觀察到的是，主要用於形成假設和解釋研究結果的理論，與在戶外自然環境中接觸時所使用的理論相同（例如，ART；Kaplan & Kaplan，1989）。考慮到這一研究領域仍處於早期發展階段，這並不令人驚訝。然而，我們可以推測，未來會有更多的預測理論考慮到虛擬現實媒介的特性，例如存在感。

Almost the totality of studies is experimental in nature, and they were conducted with considerable heterogeneity of methods and instruments. This point is not necessarily a limitation since this line of research is relatively new and in expansion. Nevertheless, what we observe is often the use of unvalidated scales and the use of single-item measures. Thus, more sound experimental paradigms are recommendable for future studies. Moreover, future research should focus on the specificities of different methods used and their efficacy (e.g.,
幾乎所有的研究都是實驗性質，並且它們的研究方法和工具具有相當大的異質性。這一點不一定是一種限制，因為這一研究領域相對較新且正在擴展。然而，我們觀察到的往往是使用未經驗證的量表和單項測量。因此，未來的研究建議採用更為可靠的實驗範式。此外，未來的研究應該專注於不同方法的特性及其有效性（例如，
differences in the use of computer-generated or real environments). Lastly, future studies could consider different study designs. First, longitudinal studies could monitor changes over time due to repeated exposure to virtual nature (e.g., every day for a certain amount of time). Second, studies could test the effectiveness of clinical protocols based on virtual nature in the general population (e.g., stress reduction) and clinical samples (e.g., reduction of specific symptomatology).
在使用電腦生成或真實環境的差異方面。最後，未來的研究可以考慮不同的研究設計。首先，縱向研究可以監測由於重複接觸虛擬自然而隨時間變化的情況（例如，每天在一定時間內）。其次，研究可以測試基於虛擬自然的臨床方案在一般人群（例如，減壓）和臨床樣本（例如，特定症狀的減少）中的有效性。

We report a wide variety of methodological approaches in terms of research design, the medium used, and exposure lengths and typology. As recently pointed out (Browning, Saeidi-Rizi et al., 2021), methodological choices may affect findings overestimating the beneficial effect of simulated nature. In fact, such a variety of experimental studies prevents scholars from performing a meta-analysis to statistically synthesize the available evidence, thus making it difficult to provide correct information on the use of this medium for application purposes (e.g., therapies). In addition, further consideration concerns the exposure type. It emerges that exposure to a random natural environment is sufficient and therefore offering "nature", whatever it may be, is certainly a beneficial factor for a variety of psychological and psychophysiological outcomes. Nevertheless, a comparison of the effects of different types of natural environments is lacking (see also Theodorou, Romano, et al., 2023).
我們報告了各種各樣的方法論方法，包括研究設計、使用的媒介以及暴露的長度和類型。正如最近指出的（Browning, Saeidi-Rizi 等，2021），方法論選擇可能會影響研究結果，過高估計模擬自然的有益效果。事實上，這種多樣的實驗研究使學者無法進行元分析來統計綜合可用證據，因此難以提供有關此媒介在應用目的（例如，治療）中使用的正確信息。此外，進一步考慮的是暴露類型。研究表明，接觸隨機的自然環境是足夠的，因此提供“自然”，無論它是什麼，無疑是對各種心理和心理生理結果的有益因素。然而，對不同類型自然環境效果的比較仍然缺乏（另見 Theodorou, Romano 等，2023）。

Similarly, the recommended duration of in-vivo exposure to nature for health and well-being purposes is 120 min a week (White et al., 2019). On the contrary, the duration of virtual nature exposure for different desired outcomes is little explored, hence there is no agreement between the selected studies. The only study (Suppakittpaisarn et al., 2023) which compared the effect of three different duration of exposure , or 15 min pointed out that the optimal duration for inducing stress recovery is equal to 5 min . Despite the appreciable effort in investigating this relevant topic, the effectiveness of only three durations has been compared and for just one outcome. Further investigations are needed in order to estimate a trade-off between the minimum duration to obtain a benefit and the maximum duration to avoid collateral issues from virtual exposure (e.g., motion sickness) in order to recommend an "optimal" duration of exposure. Concerning the virtual stimuli, among those included, only in one case (Lakhani et al., 2020) a white, icy environment was offered. White, icy environments are much less investigated than other types of natural environments, thus evidence of their effects is scarce. The duration of the effect also remains unanswered. Although the overall evidence is suggestive of the beneficial effect of virtual nature on psychological and psychophysiological outcomes, it is not yet possible to state whether this effect persists over time. It is therefore desirable that future studies arrange follow-ups to verify potential changes in the medium and long term.
同樣，為了健康和福祉目的，建議的自然實地接觸時間為每週 120 分鐘（White et al., 2019）。相反，對於不同期望結果的虛擬自然接觸時間的研究則較少，因此所選研究之間並無共識。唯一一項比較三種不同接觸時間的研究（Suppakittpaisarn et al., 2023）指出，誘導壓力恢復的最佳接觸時間為 5 分鐘。儘管在研究這一相關主題上付出了可觀的努力，但僅比較了三種接觸時間的有效性，且僅針對一個結果。需要進一步的研究來估算獲得益處的最小接觸時間與避免虛擬接觸所帶來的副作用（例如，暈動病）的最大接觸時間之間的權衡，以便推薦一個“最佳”接觸時間。關於虛擬刺激，在所包含的案例中，只有一個案例（Lakhani et al., 2020）提供了白色冰冷的環境。 白色、冰冷的環境比其他類型的自然環境研究得少得多，因此有關其影響的證據稀少。影響的持續時間也仍然沒有答案。儘管整體證據暗示虛擬自然對心理和心理生理結果有益，但目前尚無法確定這種影響是否會隨時間持續。因此，未來的研究應安排後續跟進，以驗證中長期內可能的變化。

Inconsistent findings were identified on the use of a multisensory approach for virtual nature exposure. Including sounds or noise in virtual exposure is anything but unusual and, frequently, the use of sounds in association with the visual stimulus is not even specified as it is part of the 2 D or 3 D video used as an exposure stimulus. It has been found that a combination of visual, olfactory, and auditory stimuli improves the participants' sense of reality during a virtual task and a better stress recovery compared to visual-only virtual stimuli (Schebella et al., 2020). More specifically, Hedblom et al. (2019) reported higher perceived pleasantness and lower physiological responses associated with stress in the case of olfactory stimuli (odors) rather than visual stimuli. Furthermore, we can hypothesize that not only the combination but also the specific type of visual, auditory, or olfactory stimuli may have differentiated effects for each individual. For example, sounds made by other people may distract the perceived peacefulness in an urban park (Jo & Jeon, 2020). Overall, findings on the effect of a multisensory approach on well-being and psychological outcomes are still scarce; thus, there is plenty of room for further investigation.
在虛擬自然暴露中使用多感官方法的研究結果不一致。在虛擬暴露中包含聲音或噪音並不罕見，且通常與視覺刺激相關的聲音使用甚至未被明確說明，因為它是作為暴露刺激所使用的 2D 或 3D 視頻的一部分。研究發現，視覺、嗅覺和聽覺刺激的組合能改善參與者在虛擬任務中的現實感，並且相比僅有視覺的虛擬刺激，能更好地恢復壓力（Schebella et al., 2020）。更具體地說，Hedblom 等人（2019）報告指出，在嗅覺刺激（氣味）而非視覺刺激的情況下，感知的愉悅感更高，與壓力相關的生理反應更低。此外，我們可以假設，不僅是組合，特定類型的視覺、聽覺或嗅覺刺激對每個個體可能有不同的影響。例如，其他人發出的聲音可能會分散在城市公園中感知到的寧靜感（Jo & Jeon, 2020）。 整體而言，關於多感官方法對幸福感和心理結果影響的研究仍然稀少；因此，還有很多空間進行進一步的調查。

Lastly, particular attention should be also paid to the control conditions and their comparisons. As emerged from our findings, a virtual urban environment may produce some sort of healing effect, e.g., relaxation. In light of this, it is crucial to understand how much of the benefits originate from the virtual medium and from the content of the medium, such as natural, urban, or other environments.
最後，還應特別注意控制條件及其比較。根據我們的研究結果，虛擬城市環境可能會產生某種療癒效果，例如放鬆。因此，了解這些好處有多少來自虛擬媒介以及媒介內容（如自然、城市或其他環境）是至關重要的。

Human health benefits deriving from the interaction with natural environments are now well-established (e.g., Spano et al., 2020). The main purpose for using virtual nature lies in the need for assuring a share of nature engagement to individuals unable to directly enjoy the benefits of that exposition. This may be true for urban residents, but more importantly for people with mobility constraints, physical or mental frailty, or in confinement, e.g., as in the case of Covid-19 infected patients, people under arrest or in custody or in care facilities, and other special conditions. A review on the use of virtual nature in health and care settings (White et al., 2018) suggested VR as a useful and effective substitute for real contact with nature, which therefore should be preferred, for a number of physical and psychological outcomes, including the treatment of chronic pain, relaxation for cancer patients, eating disorders, post-traumatic disorder, anxiety, and depression. Particularly noteworthy is the use of VR as a treatment for cognitive impairment and dementia. Reynolds, Rodiek, Lininger, and McCulley (2018) reported the presence of a relaxation effect, detected through heart rate reduction, and more positive emotions, after exposure to a non-immersive VR nature compared to other VR contents (e.g., a movie) in patients with dementia. VR-based therapy proved to be effective also in combination with other, conventional psychotherapy approaches, such as the case of Cognitive Behavior Therapy (CBT). As early as 2009, Kim and collaborators (2009) demonstrated a higher effectiveness of a CBT-based therapy performed in a forest environment compared with the same therapy in hospital or outpatient management, in achieving depression remission in patients with major depressive disorders. More recently, nature-based mindfulness was shown to provide a positive effect on a number of psychological, physical, and social conditions (Djernis et al., 2019). Nevertheless, to our knowledge, excluding two studies included in this systematic review on the combination of virtual nature and biofeedback (Blum et al., 2019; Rockstroh et al., 2019), VR-based therapy in combination with well-established therapeutic approaches is yet to be explored, thus, VR-based systematic protocols are lacking. More importantly, we acknowledge the lack of medium- and long-term efficacy studies. In general, it is unlikely that virtual nature will be able to replace conventional therapies; however, according to a framework recently proposed by Litleskare, E MacIntyre, and Calogiuri (2020), virtual nature may be an effective supplement to real nature, may help people in reconnecting with real nature, and enhance the benefits deriving from the human-nature interaction.
人類健康從與自然環境的互動中獲益的事實現在已經得到充分證實（例如，Spano 等，2020）。使用虛擬自然的主要目的是為了確保無法直接享受自然接觸好處的個體能夠獲得一部分自然參與。這對於城市居民來說可能是正確的，但對於行動受限、身心虛弱或被限制的人來說更為重要，例如，Covid-19 感染患者、被逮捕或拘留的人、以及在護理機構中的人和其他特殊情況。關於在健康和護理環境中使用虛擬自然的綜述（White 等，2018）建議虛擬現實作為與自然實際接觸的有用和有效的替代品，因此應該優先考慮，因為它對多種身體和心理結果有益，包括慢性疼痛的治療、癌症患者的放鬆、飲食失調、創傷後壓力症候群、焦慮和抑鬱。特別值得注意的是，虛擬現實被用作認知障礙和癡呆症的治療。 雷諾茲、羅迪克、林寧格和麥卡利（2018）報告了放鬆效應的存在，這一效應通過心率降低和更積極的情緒來檢測，這是在接觸非沉浸式虛擬現實自然環境後，相較於其他虛擬現實內容（例如電影）在癡呆症患者中的表現。虛擬現實基礎的療法在與其他傳統心理治療方法結合時也證明是有效的，例如認知行為療法（CBT）的案例。早在 2009 年，金及其合作者（2009）就展示了在森林環境中進行的 CBT 基礎療法相比於在醫院或門診管理中進行的相同療法，在實現重度抑鬱症患者的抑鬱緩解方面具有更高的有效性。最近，自然基礎的正念被證明對多種心理、身體和社會狀況產生了積極影響（Djernis 等，2019）。然而，據我們所知，除了這項系統評價中包含的兩項關於虛擬自然和生物反饋結合的研究（Blum 等，2019；Rockstroh 等）。，2019），基於虛擬實境的療法結合已建立的治療方法尚待探索，因此，基於虛擬實境的系統性方案仍然缺乏。更重要的是，我們承認缺乏中期和長期的療效研究。一般來說，虛擬自然不太可能取代傳統療法；然而，根據 Litleskare、E MacIntyre 和 Calogiuri（2020）最近提出的框架，虛擬自然可能是現實自然的有效補充，可能幫助人們重新與現實自然建立聯繫，並增強人與自然互動所帶來的好處。

The present work was aimed at systematizing the available studies on the benefits of exposure to nature through VR, focusing on psychological and psychophysiological outcomes and related intervening variables. We identified current gaps (lack of specific samples, studies of unknown power, underestimated outcomes, lack of investigation of mediators and moderators, lack of theories that could consider the specificities of the VR medium, lack of best practices) and future research directions (first of all, meta-analyses and investigations of publication bias), with the hope of challenging researchers interested in this rapidly expanding field of study.
本研究旨在系統化現有關於透過虛擬實境接觸自然的好處的研究，重點關注心理和心理生理結果及相關的介入變數。我們識別了當前的缺口（缺乏特定樣本、未知效能的研究、低估的結果、缺乏對中介和調節變數的研究、缺乏能考慮虛擬實境媒介特性的理論、缺乏最佳實踐）以及未來的研究方向（首先是進行元分析和出版偏差的調查），希望能挑戰對這一快速擴展的研究領域感興趣的研究者。

We have no conflicts of interest to disclosure. This work was carried out under the project "Establishing Urban FORest based solutions In Changing Cities" (EUFORICC) and financially supported by the Ministry of Education, University and Research (MIUR) of Italy (PRIN 20173RRN2S).
我們沒有需要披露的利益衝突。本項工作是在「在變化中的城市中建立城市森林解決方案」(EUFORICC) 項目下進行的，並得到了意大利教育、大学和研究部（MIUR）的財政支持（PRIN 20173RRN2S）。