Exposure to natural environments has been widely recognized for its potential to affect emotional and affective processes positively. However, in the research literature, the concept of a natural environment is multifaceted and not uniformly defined. An environment can be considered natural if it includes certain natural elements (Korpela, 2003; Kuo and Sullivan, 2001). Various types of natural environments exist, primarily distinguished as blue spaces and green spaces. Green spaces refer to natural areas such as parks, gardens, and forests in both urban and wilderness settings (van Dillen et al., 2012; Lee and Maheswaran, 2011; Maas et al., 2006). Blue spaces encompass aquatic environments, including oceans, lakes, rivers, and smaller water features like fountains and streams (Grellier et al., 2017; White et al., 2010).

However, this binary categorization of nature—into green and blue spaces—fails to capture the diversity of natural landscapes that might offer therapeutic benefits. The health advantages of many natural environments remain underexplored in the burgeoning literature on nature exposure. For example, “white space”, such as snow-covered landscapes, has been identified as potentially beneficial (Brooke and Williams, 2021; Finlay, 2018; Korpela et al., 2014; Olwig, 2005; Yli-Panula et al., 2022). Similarly, “brown space”, referring to deserts, and “red nature”, which includes volcanoes, are emerging concepts in the scholarly discussion of nature’s therapeutic effects (Nazif-Munoz et al., 2020; Olvera-Alvarez et al., 2021; Yin et al., 2022; Kotera et al., 2021).

Furthermore, various characteristics of natural environments can potentially influence the benefits they provide to humans, including dimensions, composition, location, and configuration of natural landscapes (Wheeler et al., 2015), tree canopy density (Jiang et al., 2016), and biodiversity (Cameron et al., 2020). Specific features of natural spaces might enhance the effects of certain interactions and offer diverse opportunities for physical activities, relaxation, and engagement (van Dillen et al., 2012). In this context, the concept of affordance is particularly relevant. Gibson (1979) highlighted that environments have properties that suggest their possible uses and action opportunities for people. The notion of affordance extends beyond physical actions to include opportunities for a variety of human behaviours, encompassing social, cognitive, and emotional dimensions (Brymer and Davids, 2014). In this regard, emotional affordance refers to the likelihood that a situation facilitates certain emotions (Schutte et al., 2008) along the dimensions of pleasure/displeasure and activation/relaxation (Roe and Aspinall, 2017).

Research also indicates that the effects of nature exposure vary depending on the type of human-nature interaction (Kahn Jr et al., 2018) and the characteristics of the exposure, such as the duration and frequency of visits to natural environments (Shanahan et al., 2015). Therefore, an important issue is the conceptualization of nature experience, that encompasses individuals’ perceptions and interactions with natural stimuli through various sensory modalities (Hartig et al., 2011).

Nature is widely acknowledged as a restorative environment, providing numerous benefits for human health, performance, and wellbeing. Interacting with nature offers a range of advantages: physiological, physical, affective, cognitive, and social. Physiologically, it helps stabilise heart rate and blood pressure, reduces muscle tension, and lowers cortisol levels (Hartig et al., 2014; Ulrich, 1983). Additionally, it is linked to reduced mortality from natural causes, respiratory diseases, cardiovascular diseases, and type 2 diabetes (Rojas-Rueda et al., 2019; Twohig-Bennett and Jones, 2018; Astell-Burt et al., 2014). Cognitive benefits include improved attention, executive function, and memory (Berman et al., 2008; Berto, 2014). Socially, nature reduces isolation and loneliness, and enhances perceptions of social support (Maas et al., 2009; Astell-Burt and Feng, 2021). Affectively, nature generally reduces stress, depression, anxiety, rumination, and negative emotions, while promoting positive emotions (Bratman et al., 2015; Ohly et al., 2016) and wellbeing (McMahan and Estes, 2015; Roberts et al., 2020), with long-term effects as well (e.g., Li D. et al., 2021; Vitale et al., 2022). Furthermore, research has shown how exposure to nature also has an effect on nature connectedness itself and on nature engagement (Leung et al., 2022), which in turn can have an effect on pro-environmental behaviour (Anderson and Krettenauer, 2021), and thus on social wellbeing.

Experimental studies have shown that exposure to natural environments can directly impact emotional processes by reducing negative emotions and enhancing positive ones (Berman et al., 2008; Hartig et al., 2003; 2011; van den Berg et al., 2003). Nature also benefits emotional processes indirectly by restoring attentional resources, thereby improving executive functioning and self-regulation (Berman et al., 2008; Kaplan and Berman, 2010; Berto, 2005). Hence, exposure to natural stimuli effectively influences both affective states (McMahan and Estes, 2015; Takayama et al., 2022) and emotion regulation processes (Vitale and Bonaiuto, 2024; Ríos-Rodríguez et al., 2024; Bratman et al., 2021).

Several theories explain nature’s restorative effects. From a philosophical and evolutionary standpoint, Fromm (1964) was the first to introduce the concept of Biophilia, which he described as a fundamental human inclination to seek connection with nature. This intrinsic tendency reflects an innate preference for natural or nature-associated environments over artificial or constructed settings. Fromm’s insight laid the groundwork for later theoretical developments, such as those of Wilson (1984), who extended the concept to evolutionary biology—suggesting that this affinity is not merely psychological, but rooted in adaptive survival mechanisms that favored individuals attuned to natural cues.

These initial theories have been expanded and experimentally tested and explained over the years by various authors. Ulrich (1983) proposed the Stress Reduction Theory (SRT) suggesting that humans are evolutionarily inclined to prefer natural environments, which helps reduce stress by lowering physiological arousal and negative emotions. Moreover, according to the Attention Restoration Theory (ART), by Kaplan and Kaplan (1989) nature enhances mental health by improving attention through effortless engagement with natural stimuli, offering a break from daily demands and providing captivating, non-demanding elements (Kaplan and Berman, 2010).

Although spending time in real natural environments may provide greater benefits for wellbeing compared to indirect experiences (Browning et al., 2020), a significant portion of evidence supporting nature’s therapeutic potential comes from studies using digital surrogates (McMahan and Estes, 2015). For instance, nature-based videos have shown potential in aiding recovery from physiological and psychological stress markers over the past 30 years (e.g., Meuwese et al., 2021; Ulrich et al., 1991). Even viewing static images of natural scenes can significantly increase positive emotions (Hartig et al., 1996), improve mood (van den Berg et al., 2003), restore attentional capacity (Berto, 2005), enhance executive attention (Berman et al., 2008), and reduce impulsivity (Berry et al., 2015).

The regenerative potential of digital nature representations through images, videos, and projections on various screen sizes has been extensively studied (Browning et al., 2020). However, these simulated forms of nature often fall short of fully replicating the qualities of real natural environments, potentially limiting their effectiveness (White et al., 2018). They may lack the extent and scale of natural settings, the ability to transport individuals away from their physical surroundings, offer personalised experiences, or engage multiple sensory modalities (Hedblom et al., 2019; Kjellgren and Buhrkall, 2010).

Advancements in Virtual Reality (VR) technology offer a way to overcome some of these limitations by allowing people to immerse themselves more completely in simulated natural environments. VR can be described as a computer-generated, three-dimensional simulation that allows navigation, interaction, and immersion in another environment (Briggs, 1996). This immersion is achieved by either sitting in a room with nature images projected on all visible surfaces (e.g., Annerstedt et al., 2013) or by using head-mounted displays (HMDs) that block out or reduce external sensory stimuli, enabling complete immersion in a panoramic view of an alternative environment (Furman et al., 2009; Mosso et al., 2009).

The benefits of nature contact appear to extend to these immersive digital environments provided through VR technology (Nukarinen et al., 2022; White et al., 2018). VR nature experiences have been linked to improvements in negative emotions, restorativeness, self-reported stress, anxiety, happiness, and creativity (e.g., Palanica et al., 2019; Schutte et al., 2017; Schebella et al., 2020; Yu et al., 2018; Grossi and Marocco, 2025; Marocco et al., 2025b). However, evidence is still mixed (e.g., Li H. et al., 2021), and further research is needed to evaluate the potential of using VR nature scenarios and their impact on emotional states and other relevant psychological variables.

However, it is crucial to recognize that, within the field of technological innovation, there are also constraints that impact users’ willingness to adopt new technologies (Marocco et al., 2024c). These limitations primarily arise from dimensions of User Experience that are closely linked to emotional perception, such as technical usability, user acceptance and domain suitability (Stickdorn and Schneider, 2011; Marocco et al., 2024a; b, c; d). Therefore, designing technology with a human-centered perspective (Talamo et al., 2021; Marocco and Talamo, 2022; Talamo et al., 2011) is essential for ensuring the effectiveness of this kind of VR interventions. Nevertheless, due to space constraints, this review will not primarily focus on the design perspective of VR technology or the impact of user acceptability on the emotional response. In the field of User Experience, indeed, emotions play a crucial role. The experience itself can be considered an intuitive and immediate perception of a feeling or an emotion, often preceding rational evaluation. Therefore, investigating how virtual nature elicits emotional responses is a foundational step toward understanding the overall experience of interaction with these technologies.

Hence, given the growing interest in the therapeutic potential of virtual natural environments, we decided to focus our narrative review on the role of virtual natural environments in emotion elicitation. Most of the existing literature has focused on the general emotional effects of VR, while relatively few studies and reviews have systematically examined the specific impact of virtual nature. A deeper analysis of the main reviews in the field (Somarathna et al., 2023, Liang et al., 2024; Bolouki, 2024; Spano et al., 2023; Li et al., 2023; Abdullah et al., 2021) reveals that many tend to overlook key dimensions that are essential to fully understanding the emotional impact of virtual nature. In particular, three recurring limitations can be identified: a lack of differentiation between types of natural environments (e.g., green or blue spaces); insufficient integration of explicit and implicit emotional measures; a lack of comparative analysis of different exposure modalities. In contrast to the studies discussed above and addressing these limitations, the present review provides a more differentiated perspective on the effects of virtual nature, which has been largely missing from prior reviews. Hence, this review will consolidate the current knowledge, identify gaps, and provide direction for future research, addressing the types of interventions used in past studies, the variety of natural environments utilised, and their differential impacts on emotional responses, measured both by explicit and implicit measures.

This review aims to address several key research questions to better understand the impact of natural virtual environments on emotion elicitation, focusing specifically on two crucial components of emotions: valence and arousal. Indeed, emotions can be mapped within a two-dimensional framework encompassing emotional valence and arousal. Valence refers to the positivity or negativity of an emotion, while arousal indicates its intensity (Russell and Barrett, 1999; Lang et al., 1998; Russell, 2003).

The following questions were formulated to guide the analysis:

1. What types of interventions have been adopted by previous studies and which have been the most effective? This question seeks to identify the various methods used in past research to expose participants to natural virtual environments, including the length of exposure and the specific techniques employed, and evaluate their relative effectiveness in eliciting emotional responses.

By addressing these research questions, the review aims to synthesize current knowledge and identify gaps in the literature, ultimately providing a foundation for future studies and practical applications in the field of virtual reality and emotion elicitation.

With these questions in mind, a literature search was conducted in June 2024 to find empirical studies on the topic of interest. To ensure the inclusion of good-quality studies that have undergone peer review, the analysis focused exclusively on articles published in scientific journals, prioritising those that have been evaluated by experts in the field. The following databases were used to identify potentially relevant articles: Scopus, Google Scholar, and ScienceDirect.

For all databases, the same set of search keywords was used. About the VR-related literature, the keywords “virtual reality” and “VR” were used as search terms. For natural spaces, the keywords included “nature”, “natural environments”, and “natural scenarios”. Regarding the emotion-related literature, the adopted keywords were “emotion”, “mood”, “emotional states”, “affect”, and “affective states”. These search terms were used in combination to conduct the research across all databases, focusing on titles and abstracts, as follows: (“virtual reality” OR “VR”) AND (“nature” OR “natural environments” OR “natural scenarios”) AND (“emotion” OR “emotional states” OR “affect” OR “affective states”).

The search was limited to peer-reviewed journal articles published in English. Given the rapid evolution of VR technology over the past decade, we decided to focus only on research published in the last 10 years to exclude studies using outdated technology (LaValle, 2023). Characteristics such as sample type (e.g., clinical vs nonclinical), age, gender, and research design were not considered exclusion criteria for this review. The full list of inclusion and exclusion criteria is reported in Table 1.

After the selection of the articles, the data extraction was carried out for all the included studies, using a pre-established form. The main components of the data extraction form were: the year of publication and country of study; study design and methods; study sample; characteristics of the virtual nature environments and the intervention (duration and modality); assessment of emotions; and results.

Furthermore, a synthesis of the evidence is provided. First, a preliminary descriptive summary of the included studies is presented, detailing their characteristics in terms of year of publication, location, sample demographics, study design and methods, features of the virtual natural environments, interventions used, and methods for assessing emotions.

Then, a narrative synthesis was adopted to further synthesise the results. Based on the guidelines for discussing narrative reviews (Baumeister and Leary, 1997), which recommend section critiques rather than critiquing each individual study, the findings have been organised around the two key components of emotion elicitation: valence and arousal. Each section includes a summary of the methods and results of relevant studies, along with a brief outline of the major flaws and limitations in the evidence. This approach allows for a more cohesive and comprehensive discussion of the overall trends and insights within the literature on the potential of virtual nature in emotion elicitation.

The studies obtained from the search were characterised in terms of participant description (including sample size and age range), country, and research design (refer to Table 2 for an overview).

Furthermore, an overview of publication years of the reviewed studies is shown in Figure 1.

From the analysis of the literature, it emerged that most studies exploring the effect of virtual natural scenarios on emotions used green spaces (i.e., forests, meadows, parks), compared to other types of environments such as blue spaces (i.e., beaches, sea, waterfalls) or brown (i.e., deserts) and red environments (i.e., savannahs and canyons).

More specifically, a significant number of studies used forest environments (i.e., nine). Browning et al. (2020) used an outdoor forest setting, while Hong et al. (2019) recorded video at Minjuji Mountain using a 360-degree camera. Mattila et al. (2020) and Reese et al. (2022) focused on a general forest environment. Wang et al. (2019) examined seven different types of forest resting environments. Yu et al. (2018) used the natural forest of The Aowanda National Forest Recreation Area, and Yu et al. (2020) showed the Neidong National Forest Recreation Area in Taiwan. Some other studies focused on different kinds of environments or mixed environments. Specifically, Gao et al. (2019) examined both blue spaces, such as an open pond with some greenery, and green spaces with various levels of canopy cover, and Yu et al. (2018) used a waterfall environment in The Aowanda National Forest Recreation Area.

Meadow environments were used in three studies (Chan et al., 2023; Voigt-Antons et al., 2021; Jin et al., 2023), providing open, grassy landscapes with natural elements. Parks, particularly urban and peri-urban parks, were featured in two studies. Ho et al. (2023) used nature-based VR videos showing different areas such as parks, hiking trails, forest paths, and bikeways. Meng et al. (2024) focused on urban parks, peri-urban parks, and nature conservation areas.

Among the blue spaces, coastal environments were featured in two studies. Anderson et al. (2017) used locations over Ireland with large expansive natural vistas with views of water and beach locations on the Australian coast, while Reese et al. (2021) investigated a general coastal nature environment. Table 3 provides a description of all the included studies, categorized by the different types of virtual natural environments used in the research.

The virtual environment interventions explored were designed with varying exposure durations and modalities.

The durations can be broadly categorised into short (5–10 min), moderate (15–20 min), and long exposures (30 min). This categorization aims to synthesize and standardize the comparison of studies that employed heterogeneous exposure times. Although these intervals are not universally standardized, they were defined based on the most frequently recurring exposure durations found across the included studies.

The modalities of exposure in virtual natural environments included seated viewing, walking activities, or free exploration, with various levels of interaction.

For short exposure sessions, typically lasting between 5 and 10 min, participants were usually seated. In the studies by Gao et al. (2019), Hong et al. (2019), Wang et al. (2019), and Reese et al. (2021) participants were seated while they watched images/videos or navigated through the virtual scenery for 5 min. Chan et al. (2023) and Mattila et al. (2020) used a swivel chair, allowing participants to turn 360° while remaining seated for 5 min. In the study by Jin et al. (2023), participants interacted within the natural virtual scene for 5 min under varying conditions (no interaction, interaction with a handle, interaction with gestures). Reese et al. (2022) conducted a study where participants were asked to remain seated on an office chair for a duration ranging from 5.5 to 9.5 min. During this time, they had the freedom to move their body, arms, and head at their discretion. Meng et al. (2024) provided a 10-min VR immersion while seated. In the studies by Yu et al. (2018), (2020), participants watched videos in any comfortable seated position for 10 min. Masters et al. (2022) allowed participants to roam within the VR scenes or explore the virtual environment for 10 min each, respectively.

Moderate exposure studies, ranging from 15 to 20 min, often involved seated participants. Anderson et al. (2017) had participants seated for 15 min, with some in a supine position on a lounge chair for specific scenes. Voigt-Antons et al. (2021) involved participants in playing games while seated for 20 min. Browning et al. (2020) combined seated and walking experiences, alternating between 6 min of sitting and 6 min of walking, totalling 18 min, instructing participants to relax and enjoy the setting during the sitting component.

Long exposure sessions lasted around 30 min. Graf et al. (2020) enabled participants to explore the “VR Forest Walk” for 30 min, guiding them on how to navigate and interact in the virtual world. Ho et al. (2023) allowed participants to watch VR videos in a 360-degree format for 30 min (once a week for 12 weeks), encouraging them to move their heads freely to view from various angles.

An overview of the interventions’ characteristics, focusing on the duration and the modality of exposure to VR, is shown in Table 4.

For the evaluation of the emotional experience, explicit or implicit measures are used. Explicit measures rely on the conscious and intentional assessment of participants. In this type of measurement, subjects are presented with direct questions or asked for subjective evaluations regarding a specific topic. On the other hand, implicit measures aim to assess individuals’ automatic and unconscious reactions. This type of measurement relies on indirect indicators of responses, such as reaction times or physiological measurements. In the included studies, it emerges that both implicit and explicit measurements were used, sometimes in combination. Specifically, implicit measures were used to assess arousal, while self-reports questionnaires were mostly adopted to obtain information about the valence of the emotional states.

This section summarizes the key findings from the reviewed studies regarding the impact of virtual nature interventions on emotional experiences, highlighting the major trends and insights. We will first examine the evidence related to the valence component of emotion elicitation, followed by an analysis of the arousal component.

The studies reviewed employed various methods to expose participants to virtual natural environments, with differences in the length of exposure and specific techniques used. The duration of exposure ranged from short sessions of 5 minutes (Chan et al., 2023; Gao et al., 2019; Hong et al., 2019; Mattila et al., 2020) to longer sessions of 30 min (Graf et al., 2020; Ho et al., 2023). The modalities included seated viewing with or without the possibility to move freely (Ho et al., 2023; Yu et al., 2018; Yu et al., 2020), walking in place (Browning et al., 2020), and interacting with the environment in different ways (Jin et al., 2023). The most effective interventions were those that combined longer durations with higher levels of interaction, allowing participants, through hand gestures, to immerse themselves deeply in the virtual natural environments. This comprehensive engagement appears crucial for maximizing the emotional and physiological benefits of VR nature experiences.

The studies included in the present review employed a broad spectrum of virtual natural environments, categorized into green, blue, red, and brown environments. Green spaces were the most frequently utilized, with a strong focus on virtual forests (Browning et al., 2020; Hong et al., 2019; Mattila et al., 2020; Reese et al., 2022; Yu et al., 2018; Yu et al., 2020). Parks, hiking trails, forest paths, and bikeways were examined by Ho et al. (2023), green meadows were investigated by Chan et al. (2023) and Jin et al. (2023), and urban parks, peri-urban parks, and nature conservation areas were included by Meng et al. (2024).

Blue spaces were represented in two studies: expansive vistas of water and beach locations were featured by Anderson et al. (2017), while coastal nature environments were explored by Reese et al. (2021). Moreover, some studies included different kinds of environments or mixed environments: open ponds and greenery, combining elements of blue and green spaces, were included by Gao et al. (2019), while mixed environments featuring forests with views of oceans or mountains were utilized by Graf et al. (2020).

Red and brown environments were less common, with two studies focusing on these settings: a red savannah was examined by Chan et al. (2023), and canyons along with forest environments were studied by Masters et al. (2022). No studies considering white environments were included in the review due to our specified inclusion criteria.

This review reveals that virtual green meadows and forest environments are the most effective in eliciting positive emotions. Specifically, green meadows consistently demonstrated an ability to enhance positive affect across multiple studies (Chan et al., 2023; Voigt-Antons et al., 2021; Jin et al., 2023). This trend indicates that green meadows are particularly effective in generating positive emotional responses in virtual environments.

Similarly, virtual forests also showed a strong capacity to elicit positive emotions. Indeed, studies by Browning et al. (2020), Hong et al. (2019), Mattila et al. (2020), Masters et al. (2022), Reese et al. (2022), Wang et al. (2019), Yu et al. (2018), and Yu et al. (2020) found that forest environments increased positive emotions, although in some cases, these increases were not statistically significant. Despite the variability in significance, the consistent trend across multiple studies underscores the effectiveness of virtual forests in enhancing positive emotional experiences.

In contrast, other types of environments such as blue spaces (e.g., beaches, coastal environments) and red or brown environments (e.g., red savannahs, canyons) showed less consistent effects on positive emotions. For example, Anderson et al. (2017) and Reese et al. (2021) reported no significant change in positive affect from blue spaces, while Chan et al. (2023) and Masters et al. (2022) found mixed results for red and brown environments. Further exploration is needed to identify potential factors contributing to the variability in their effects.

The analysis of the included studies reveals that virtual forests as green spaces are most effective in reducing arousal. Specifically, studies by Hong et al. (2019), Wang et al. (2019), and Yu et al. (2018) reported decreases in arousal associated with forest environments, indicating their effectiveness in promoting relaxation. Additionally, various types of green spaces, including open green space and partly closed green space, were discovered to be effective in decreasing arousal levels (Ho et al., 2023; Gao et al., 2019).

In contrast, green meadows demonstrated mixed results. While Jin et al. (2023) found a decrease in arousal associated with green meadows, Voigt-Antons et al. (2021) reported no significant change in arousal after exposure to a green meadow. This variability requires further exploration, given the limited number of studies conducted and analysed on these specific environments.

Blue spaces also showed a notable reduction in arousal, as evidenced by Anderson et al. (2017) and Gao et al. (2019), where expansive blue vistas and open ponds with greenery contributed to decreased arousal, validating the calming effect of water elements. However, it is important to note that the effect of blue spaces was less frequently studied compared to green environments.

In general, although the included studies revealed a positive impact of virtual natural environments on emotion elicitation, it is noteworthy that some evidence is still mixed and need for further exploration. Particularly, even if most of the studies reported a decrease in negative emotions, suggesting a positive impact of VR experience on emotional states, not all the research revealed significant effects in enhancing positive affect after the VR experience. For example, Gao et al. (2019), assessed no significant changes in emotions with positive valence but reported a reduction in negative valence emotions. Similarly, Meng et al. (2024) and Anderson et al. (2017) found positive emotions unaltered but negative ones decreased. Moreover, regarding reduction in emotional arousal, most of the studies reported a decreased arousal. However, some studies reported an increased arousal (Browning et al., 2020) or a not significant change (Voigt-Antons et al., 2021; Yu et al., 2020). These unexpected results may be attributed to methodological factors. In Browning et al.’s (2020) study, participants alternated between sitting and walking, a condition that may have physiologically stimulated them. Similarly, Voigt-Antons et al. (2021) involved participants in gameplay during the VR experience, potentially engaging mechanisms that heightened arousal rather than reducing it. Differently, in Yu et al. (2020), participants were seated in a comfortable position but the VR experience only lasted 10 min, a very short time to detect differences in physiological states. Further exploration of these dimensions is needed in future research.