1 Introduction

The application of computer-based technology to sport is an area of intense interest. Such technologies include computerised modelling, data acquisition and analysis, mobile computers, and information technology networks (Baca et al. 2009). Virtual reality (VR) is another technology, and it was first applied to sport research in the 1990s, although there has been a resurgence of interest in recent years. VR refers to a computer-simulated environment that aims to induce a sense of being mentally or physically present in another place (Baños et al. 2000; Sherman and Craig 2002). An important feature of VR is that the individual can interact with the environment. In the context of sport, interaction might occur through an exertion interface (Mueller et al. 2007). For example, physical effort on a machine such as an ergometer can be related to the speed of movement through a virtual race course. Motion capture video systems, infrared beams, and wearable sensors are other approaches that can be used to translate physical actions into virtual sport performance.

The key elements that define VR applications to sport are the use of computer-generated sport-relevant content and a means for the athlete to interact with the virtual environment. When defined in this way, the application of VR to sport has a number of strengths. As noted by Hoffman et al. (2014), the VR environment can be controlled and manipulated in specific and reproducible ways. Hoffman et al. used these characteristics to train participants to use a rowing race pacing strategy. VR can also be used for assessment, to gain feedback on performance, and to practice specific skills. The VR environment does not need to be limited to a single person. Other individuals may be present such as a coach, teammate, or competitor even if they are physically located in another place. The ability to connect with individuals via the Internet allows for interaction without the need for travel. Finally, the increasing availability of commercially produced software or full VR systems avoids the need for specialised technical expertise and allows VR to be used in local gyms and at home.

The present study aimed to provide a systematic review of research on VR applications to sport. The PsycINFO, SPORTDiscus, Scopus, Google Scholar, and Cochrane Library databases were first searched for the existence of similar reviews. The search yielded systematic reviews on VR in physical rehabilitation (e.g. Laver et al. 2015), VR in psychological interventions (e.g. Meyerbröker and Emmelkamp 2010), and the use of exergames or active videogames (e.g. Guy et al. 2011; Larsen et al. 2013; Peng et al. 2013). The search helped to minimise overlap with existing reviews. Accordingly, the present review focused on VR applications to sport and sport-related exercise with healthy individuals. Studies were included if they were based on recognised sports even if those sports are used as a component of physical conditioning or fitness programs (e.g. cycling, running, rowing). As a result, this review focused on sport-based tasks as distinct from research with interactive videogame systems that promote physical activity through gameplay (i.e. exergames).

The broad question examined in the present review was: What is known about the application of VR to sport? In particular, the review aimed to provide a definition of VR when used for sports. A further aim was to document the aims, methods, and the broad findings from the research conducted to date. Past research may be interpreted within the context of existing theories in sport and exercise, but of particular focus in the present review were those factors that are unique to VR applications to sport. The review also aimed to identify the gaps in the research to date and develop recommended reporting standards for researchers who apply VR to sport.

2 Literature review method

The literature search and selection method followed the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines (Liberati et al. 2009) and the use of inclusion and exclusion rules described by Meline (2006). Initially, the SPORTDiscus and PsycINFO databases were searched. The PsycINFO database includes sport and exercise psychology journals, in addition to the ACM Transactions on Applied Perception, the ACM Transactions on Computer–Human Interaction, and the IEEE Transactions on Professional Communication. The search was conducted using the terms: (sport* OR exercis* OR fitness OR physical train* or physical activit*) AND (virtual realit* OR virtual environment* OR virtual world* OR virtual system* OR virtual partner*). The search was limited to articles published from 1990 and up to the date of the search (February, 2016) and included articles that were in press. In addition, to identify any missed articles due to the inconsistent use of terms (e.g. virtual reality versus virtual competitor) the reference lists of the articles selected for final inclusion from the database search were examined. An examination was also made of the citations of these articles, as collated from the Scopus database.

The database search yielded 263 articles from the PsychINFO database and 377 articles from the SPORTDiscus database for a total of 640 articles. This reduced to 620 articles following removal of duplicates. A search of the reference lists and citations yielded a further 66 unique articles. Articles were screened for exclusion or inclusion by two individuals in a two-step process: title and abstract (Step 1) and the full article (Step 2).Footnote 1 The following exclusion criteria were used: date (published before 1990), language (not published in English language), source (a dissertation, thesis, abstract only, magazine article, or not a peer-reviewed source), study type (a review, meta-analysis, commentary, letter to the editor, editorial report, or other non-empirical article), no VR was used (a computer-generated environment was not used or there was no interactivity with the environment), population (the sample did not include healthy human participants), task (the methods did not include participation in a sport or a physical exercise that used equipment related to a sport or sports training), game (the task was based wholly on an exergame/active videogame), rehabilitation (the purpose of the task was to rehabilitate those with physical injury), and measure (performance, physiological, or psychological outcomes were not the primary measures).

Following the screening and selection process, 20 articles were included for full review. Of these articles, 18 were published in journals with journal citation metrics reported by the Web of Science database. The mean impact factor (based on the most recent year) was 2.21 (range 0.06–4.47, SD = 1.21) indicating that the journals were largely of good quality although with some exceptions. Consistent with this interpretation, the journal rankings varied evenly across the full spectrum of Q1 (n = 5), Q2 (n = 5), Q3 (n = 4), and Q4 (n = 4). The articles were coded by four authors and coding decisions were cross-checked. Articles were coded for characteristics related to the study (aims, type, location, conditions/groups, outcome measures, key findings), participants (sample size, age, experience with sport), virtual reality technology (task type, system, display features, point of view, others in the environment, immersion/presence measures), and sport task (type).

3 Defining virtual reality in sport

VR when applied to sport may be defined as instances when individuals are engaged in a sport that is represented in a computer-simulated environment which aims to induce a sense of being mentally or physically present and enables interactivity with the environment. This definition highlights the computer-simulated nature and interactivity of the virtual environment, which are key element of more general definitions of VR (e.g. Baños et al. 2000; Sherman and Craig 2002). It also aims to highlight the application of VR to sport from the perspective of the user (athlete). Realistic responses to virtual environments are suggested to occur when the system induces a sense of presence and the perception that the events are actually occurring (Slater 2009). In this respect, it is important that VR uses a computer-generated environment because this is a key feature that allows for interactivity and the perception of presence (Baños et al. 2000; Sherman and Craig 2002). In other words, the virtual environment or elements within it will move or change in response to the actions of the athlete. However, the method by which the virtual environment is presented to the athlete should not be specified in the definition because it might impose technological limitations to the application of VR to sport (see Steuer 1992).

In many applications outside of sport, the virtual environment is displayed using a computer automatic virtual environment (CAVE) or head-mounted display (HMD). The CAVE is composed of a large cube made up of display screens that the user physically enters to become surrounded by the virtual environment. A HMD is a wearable device that covers the eyes and thus removes vision of the outside world. It has one or more small screens on which the virtual world is viewed in stereovision with a wide field of view. The HMD is combined with head tracking to allow the user to view areas of the virtual environment that are outside of the immediate field of view by turning their head. Being a smaller, more portable, and a more affordable system, the HMD is more popular than the CAVE, although both may be regarded as sharing the same key features of an immersive system (Slater 2009).

However, the potential applications for using CAVE and HMD systems can be limited for some types of sports. A HMD may be impractical or potentially dangerous for some sports. For example, running a race on a treadmill using a HMD can be hazardous because vision of the moving treadmill is removed. The head movements and sweating of the athlete can also make the HMD uncomfortable to wear. Indeed, in no studies identified in this review was a HMD system used despite researchers consistently using the term virtual reality to describe their approach. The most common approach was a two-dimensional depiction of the virtual environment using a computer screen or a projector. A computer screen or projector has the advantages of ease of use and practicality with sport but may induce less presence than a HMD or CAVE system. Further research is required to determine whether there is significant difference in presence when a computer screen or projector is used.

Several instances can be identified in which researchers used methodology that approximated the proposed definition of VR applications to sport. For example, some researchers have used a visual display that shows a video of a real environment (e.g. Plante et al. 2006). Feltz et al. (2011) conducted a series of studies that investigated the Köhler motivation gain effect with a plank exercise task. These studies showed the participant via a video (i.e. not a computer-generated avatar) and included a second individual shown on a second visual display without any interaction. Videos of real environments and people may have potential for VR applications to sport, but they must include elements of interactivity to fulfil the proposed definition of VR. Similarly, other researchers have used computer-generated environments to examine baseball batting (Ranganathan and Carlton 2007), handball goalkeeping (Vignais et al. 2015), and soccer goalkeeping (Stinson and Bowman 2014), but these did not allow for any interactivity with the environment and were not included in the review. Thus, the present review was focussed more specifically on interactive VR applications to sport. In some cases, it was also found that researchers used a non-animated avatar against a blank screen (e.g. Briki et al. 2013), but these do not meet the proposed definition because the methods did not simulate a real environment.

Another important consideration for interactive VR applications to sport is the distinction between sport, exercise, and exergaming. Sport may be defined as an activity that requires motor skill and/or hand-eye coordination combined with physical exertion and includes rules and elements of competition (Australian Bureau of Statistics 2008). Exercise, used synonymously with physical exercise, is a structured activity that may include repetitive elements that is performed to maintain or improve physical fitness (Australian Bureau of Statistics 2008). Exergame/active videogame is a videogame played on commercial game console systems (e.g. Xbox, Wii, PlayStation) that combines gameplay with physical movements that are more than sedentary behaviour (Kim et al. 2014). Exercise and exergames together represent a more general case of enhancing physical activity and may not necessarily be based on a sport.

Exercises or exergames that are not based on a sport clearly do not represent instances of VR applications to sport even if they incorporate a virtual environment. However, investigators have used sport-related computer games, particularly those that run on a games console, in research. Console games based on sports have been used to examine skill acquisition and transfer in children (Reynolds et al. 2014) and adults (Tirp et al. 2015). However, these applications lacked an appropriate exertion interface (e.g. participants ran on the spot to simulate running in the game) or essential sporting equipment (e.g. no darts were used in a dart game), and these aspects can make the task substantially different to perform the sport in real life. VR has also been applied to exercise and improving physical fitness. In several studies, researchers have used sport-related tasks such as cycling, running, and rowing (e.g. Murray et al. 2016). These applications have relevance to sport performance particularly because many of these studies have introduced elements of competition or pressure to meet team goals.

4 A conceptual framework for the application of virtual reality to sport

The application of VR to sport has taken many forms, with various types of sport tasks, VR technologies, and types of athletes used in the research. Some researchers have examined questions relating to the use of VR technology itself, such as comparing outcomes when using VR and not using VR (e.g. Annesi and Mazas 1997; Legrand et al. 2011; Mestre et al. 2011; Plante et al. 2003a), the effects of immersion in the virtual environment (Ijsselsteijn et al. 2004; Vogt et al. 2015), and differences between computer-controlled and real virtual competitors (Snyder et al. 2012). In contrast, other researchers have used VR technology as part of a methodology to answer more general questions about factors related to sport performance. For example, Oliveira et al. (2015) used a virtual partner as a means to compare the effects of self-selected and externally imposed exercise intensity.

We developed a broad conceptual model that summarises and provides a framework to interpret the research conducted to date. As shown in Fig. 1, the VR system results in outcomes that occur concurrently or following engagement in the VR sport task. The VR system is composed of four components. These are the VR environment, the sport task, the athlete, and the non-VR environment. Research on VR applications to sport have largely focussed on only the first three of these components. The VR environment is the unique component for VR applications to sport and is the focus of most research. The second component, the sport task used, will differ according to the application and can vary between endurance-type sports or skill-based sports. The third component relates to characteristics of the athlete, such as skill level and competitiveness. The characteristics of the athlete may act independently or they may interact with other elements of the VR system to influence outcomes. The fourth component encompasses those aspects of the real-world environment in which the athlete completes the task. Ambient temperature, humidity, and time of day are among the relevant factors that can be present and influence outcomes. Finally, all four elements of the VR system will produce outcomes that emerge on an ongoing basis when performing the sport task (concurrent outcomes) or they may emerge at a later time (posttask outcomes). The posttask outcomes may be short term or long term.

Fig. 1
figure 1

A model of interactive virtual reality (VR) in sport and sport-related exercise showing the relationship between components of the VR system, current outcomes, and posttask outcomes

Full size image

The four components of the VR system share elements in common with other models applied to sport and exercise psychology. For example, Tenenbaum and Hutchinson (2007) proposed that perceived effort and effort tolerance are determined by the individual (e.g. dispositions, task familiarity, demographic characteristics), the task (e.g. intensity, duration), and the environmental conditions (e.g. social, physical features) that are present in a given situation. These conditions are analogous to the three non-VR components of the VR system as presented in Fig. 1. Such a similarity is to be expected because VR aims to simulate a real environment. However, research on VR applications to sport have not yet examined the effects of the real (non-VR) environment on performance. Instead, attention has been directed towards variables related to the virtual environment, such as immersion, presence, and interactivity with virtual others. Research supporting the conceptual framework depicted in Fig. 1 is presented below.

5 The virtual reality system

5.1 Virtual reality environment and task factors

The first two components, the VR environment and sport task, may be considered together because they can be closely linked. For example, a rower may complete a time trial using a rowing ergometer. However, the ergometer is merely the exertion interface. It is transformed into a virtual boat such that pulls on the ergometer handle are depicted as movements of the virtual oars through the water. Increasing exertion on the task (e.g. rowing at a higher intensity) will be reflected in changes in the virtual environment (e.g. faster movement through the water and passing scenery). Thus, performance and other factors related to the task will influence the virtual environment and this relationship can be reciprocal.

Research has shown that several characteristics of the VR environment and the task influence outcomes. A summary of the methodological approaches used to create the VR environment and task is provided in Table 1. As can be seen, the sport tasks used most often have been cycling and running, but rowing, weightlifting, and golf have also been examined. Cycling, running, and rowing are sports that contain elements of endurance and persistence. These sports are also relatively easy to translate into a virtual environment. The exertion interface of the treadmill or ergometer can readily monitor information with regard to the speed and other performance elements (e.g. cadence) and translate this information into virtual movements. Interactivity is further enhanced by including directional controls although few VR systems have been used which have this capability.

Table 1 Characteristics of the task and virtual reality system of studies investigating virtual reality in sport
Full size table

The VR software and display equipment used in research has varied from commercially available products to those that are custom made. The virtual environment is typically displayed on computer screens or projected against a wall. A larger display or the inclusion of more multimodal elements of the environment will increase the sense of immersion in the virtual world (Vogt et al. 2015) and this can influence performance. Using a more immersive virtual environment during a cycling task (i.e. showing the track from the point of view of the rider versus from a birds eye view) has increased motivation and the speed of cycling in participants (Ijsselsteijn et al. 2004). Using a virtual running task, over a third of participants have reported that the immersion induced by the VR environment is an important motivating feature (Nunes et al. 2014). There might be a dose-dependent relationship between the level of immersion induced by the VR system and the magnitude of the resulting outcomes.

The presence of others in the virtual environment has also emerged as an important feature of the VR environment. Indeed, the presence of others may be even more important than the capability of the VR system to induce feelings of immersion or the presence. In a survey study examining golf play in a virtual environment, Lee et al. (2012) distinguished between two types of presence: telepresence or the feeling of being physical immersed in the virtual environment and social presence or the feeling of being with and communicating with others in the virtual environment. Social presence was shown to play a more important role in perceived enjoyment, perceived value, and behavioural intentions than telepresence. Further, unlike social presence, telepresence did not significantly predict any of these outcomes.

The presence of others has also influenced motivation and performance for aerobic sport tasks. Using a running task, Nunes et al. (2014) reported that participants preferred to run in the presence of virtual others than to run on the virtual course alone (Nunes et al. 2014). Irwin et al. (2012) examined the Köhler motivation gain effect while participants cycled in a virtual environment. Participants cycled at an intensity of 65% of heart rate reserve for as long as they felt comfortable. Different groups of participants completed trials while cycling in the virtual environment alone or at the same time as another person (a confederate) who the participant was informed had performed moderately better than they did in a baseline trial. Cycling with the other person was either in a conjunctive situation (a “team score” would be based on the rider who quit the task first) or a coactive situation (no team partnership). Task persistence was higher in the coactive situation than when cycling alone. Moreover, a further enhancement of persistence was observed in the conjunctive situation, suggesting motivational gains when performing a VR-based sport in a team situation.

In the study by Irwin et al. (2012), the confederate was shown via a video loop on another screen and not in the VR environment. Murray et al. (2016) also examined the Köhler motivation gain effect in which the teammate was present as a virtual partner in the virtual environment. Female participants novice to rowing completed a rowing trial in the presence of a virtual teammate in a conjunctive situation (the shortest distance rowed over a 9-min trial would count as the team score) or in the VR environment alone. Prior to the trial, participants were informed that the teammate had rowed 40% longer than them in an initial baseline row. A Köhler motivation gain effect was found in that participants rowed further and had a higher heart rate in the presence of a teammate than when rowing in the VR environment alone. Moreover, the conditions did not differ in felt arousal, positive feelings, or ratings of perceived exertion. The latter finding suggests that performance improvements can be induced by a virtual partner in the absence of negative psychological costs.

The presence of others in a virtual environment can be used to more directly induce a pressure to perform in a competitive situation. Using a sample of older adults, Anderson-Hanley et al. (2011) compared cycling through a virtual course either alone or in the presence of on-screen rider avatars. In the latter condition, participants were explicitly asked to outpace the avatars. The introduction of the on-screen avatars increased cycling power output when compared to solo cycling condition. However, this effect was observed only in participants who were classified as high in competitiveness based on a self-report questionnaire. A limitation of this study was that all participants completed the solo cycling condition first and the competitive situation second. Nevertheless, the findings suggest that competitiveness is an important moderating factor in responses to VR.

Similar outcomes to Anderson-Hanley et al. (2011) were reached in a study by Snyder et al. (2012) who compared two competitive situations while participants cycled in a VR environment. In the virtual condition, the participants were informed that the avatar of the other rider was controlled by the computer. In a live rider condition, the participants were introduced to a confederate and were informed that the avatar speed was controlled by the cycling speed of the confederate. Cycling performance, measured as watts generated, was higher for the live rider condition than in the virtual rider condition. Again, this difference emerged only in participants high in competitiveness. No differences between the rider conditions emerged for participants low in competitiveness.

Competitive situations can be constructed within a virtual environment in various ways. Nunes et al. (2014) devised three competitive modes based on whether participants competed against themselves (i.e. a prior performance), against an individual chosen for them who is superior, or against any individual chosen by the participant. Using a VR running task, all types of competitive situations enhanced physical exertion (as measured by heart rate) and self-reported motivation when compared to running on the virtual course alone. Evidence was also found that participants who were not initially competitive still felt pressure to outperform the on-screen avatars. However, similar to the conclusions reached by Anderson-Hanley et al. (2011) and Snyder et al. (2012), participants who had a stronger preference for competitive situations showed the highest task performance.

A different approach to the use of another individual in the virtual environment was reported by Oliveira et al. (2015). Participants completed two conditions of a VR cycling task. In one condition, the participant self-selected the intensity of the cycling trial. In the other condition, participants were asked to follow a virtual cyclist. The virtual cyclist was set to a speed that matched the self-selected intensity condition. No significant differences were found between conditions on physiological effort or affective responses. Typically, an externally imposed intensity results in an increase in negative affect. The findings thus suggest that this affective “cost” is mitigated when participants match the imposed pace of a virtual partner. However, further research is required to confirm these findings. For example, order effects may have been a factor because all participants completed the self-selected condition first and followed by the externally imposed intensity condition.

5.2 User (Athlete) factors

The third component of the VR system is the athlete who is engaging in the virtual sport. The characteristics of the athlete user have the potential to mediate or moderate the effects of VR on performance and psychological outcomes. Athlete user factors may include physical characteristics, expertise and experience, and psychological characteristics. As shown in Table 2, the participants recruited in research to date have been relatively homogenous. The typical participant has been a young adult sampled from Western countries who are novice to the sport. It has been suggested that using novices is advantageous because it results in a sample that is more physiologically equivalent and their performance is less likely to be influenced by prior learning (Hoffmann et al. 2014). However, it reduces the generalisation of findings to participants that are younger or older or who compete at the elite level.

Table 2 Sample size and participant characteristics of studies investigating virtual reality in sport
Full size table

It is surprising that most studies have not reported comparisons between males and females given the documented gender differences in not only sport performance but also in experience with computerised environments (e.g. computer games). Plante et al. (2003a) included gender as a factor when examining the effects of VR on mood during cycling. Females showed a larger difference in reported relaxation between the VR alone (no cycling) and both the cycling alone and cycling with VR conditions when compared to males. Plante et al. (2003b) also used a cycling task and reported gender differences in ratings of energy. Males reported higher energy when cycling alone, cycling with VR, or experiencing VR alone than in a baseline control condition that did not involve VR or cycling. In contrast, females reported more energy in cycling alone or cycling with VR than when VR was used without cycling or in the baseline condition. While preliminary, there is some suggestion that females may be influenced more by the VR environment than males.

The preferences of the individual user may be an important psychological factor that moderates outcomes. Legrand et al. (2011) assigned participants to either a cycling task alone (no VR input), a self-selected VR task (either jogging or cycling), or an externally imposed VR task (either jogging or cycling). All conditions improved positive affect and reduced negative affect when assessed by pre- and posttask subjective measures. The in-task subjective measures showed that participants in the self-selected VR task reported higher pleasure than the cycling alone or the externally imposed VR task, which themselves did not differ. Autonomy or the appropriate matching of an individual to a preferred sport may thus be important for mood benefits when using VR. As noted above, individual preferences for task intensity may be another factor in that using VR technology may reduce the negative impact of performing at an externally imposed intensity (Oliveira et al. 2015).

5.3 Non-VR environment factors

The final component of the VR system, the real-world environment, has received no attention in research conducted to date. Researchers have used a controlled indoor environment and have kept key variables like temperature, humidity, and time of day constant or allowed them to vary at random. Tenenbaum and Hutchinson (2007) noted that the environment can be divided into physical and social components and a similar distinction can be made here. In particular, based on research showing that the presence of others in the virtual environment can influence performance and psychological states, it would be expected that the presence of others in the real environment will also have an influence. Further research is required to examine the effects of environmental factors and to determine the relative strength of these factors when present virtually versus when present in reality.

5.4 Concurrent and posttask outcomes

A summary of the key research aims and outcomes is shown in Table 3. The majority of the outcomes have been observed concurrently with the task, but some have been observed posttask (i.e. short-term and long-term effects; see Fig. 1). Concurrent outcomes are those that influence ongoing behaviour (e.g. performance, persistence, affective states, perceived exertion). For example, VR tasks that induce competitiveness may induce short-term increases in performance if the individual is running at a pace slower than a virtual competitor (Nunes et al. 2014). Posttask outcomes will influence behaviour at a later time and are thus independent of the ongoing interaction with the VR system (e.g. meeting performance goals, competition outcomes). For instance, Annesi and Mazas (1997) showed that an exercise program that used a VR cycling task increased adherence to the exercise program relative to cycling alone.

Table 3 Characteristics of the design, aims, conditions, measures, and key findings of studies investigating virtual reality in sport
Full size table

Outcomes may also be divided into those related to task performance, physiological effects, and psychological processes. As shown in Table 3, performance outcomes in past research include adherence (Anderson-Hanley et al. 2014; Annesi and Mazas 1997; Irwin et al. 2012), distance travelled or speed in the virtual environment (Hoffmann et al. 2014; Ijsselsteijn et al. 2004; Murray et al. 2016; Nunes et al. 2014; Snyder et al. 2012), physical intensity exerted (Anderson-Hanley et al. 2011; Chen et al. 2015; Snyder et al. 2012), in-task persistence (Irwin et al. 2012), and strategy (Hoffmann et al. 2014). Physiological outcomes have included heart rate (Nunes et al. 2014; Snyder et al. 2012), oxygen consumption and blood lactate level (Oliveira et al. 2015), muscle fatigue (Chen et al. 2015), and electroencephalogram (EEG) amplitude and frequency (Vogt et al. 2015). Psychological outcomes may relate to behavioural intentions (Lee et al. 2012), cognitive functions (Anderson-Hanley et al. 2012), motivation (Ijsselsteijn et al. 2004; Nunes et al. 2014), perceived pressure (Ijsselsteijn et al. 2004), attentional focus (Baños et al. 2016; Mestre et al. 2011), and various positive and negative feeling states.

The application of VR to sport has resulted in several beneficial outcomes. When compared to control conditions, tasks that incorporate VR have shown improved adherence (Annesi and Mazas 1997), better race strategy performance (Hoffmann et al. 2014), higher cognitive functioning (Anderson-Hanley et al. 2012), improved mood and reduced tiredness (Plante et al. 2003b), increased workload (Chen et al. 2015), and higher enjoyment (Mestre et al. 2011; Murray et al. 2016). However, the control condition used in most research has involved performance of the sport on its own. This approach may be questioned because it does not control for the presence of an external stimulus during the task. It is possible that the VR environment may produce its effects because it distracts and diverts attention away from the task (Baños et al. 2016; Mestre et al. 2011), rather than because it induces a sense of the presence or includes elements of interactivity, which are the key features of a VR environment.

It is also noteworthy that better performance or psychological outcomes have not always resulted when VR is used (e.g. Lee et al. 2012; Legrand et al. 2011) suggesting that other factors may moderate its effectiveness. As noted above and shown in Table 3, these factors may relate to the VR system or user, such as level of immersion (Ijsselsteijn et al. 2004), competitiveness (Anderson-Hanley et al. 2011; Nunes et al. 2014; Snyder et al. 2012), social presence (Irwin et al. 2012; Lee et al. 2012; Murray et al. 2016), self-selection of tasks (Legrand et al. 2011), attentional focus (Mestre et al. 2011), and the mood altering effects of the task itself (Plante et al. 2003b).

Performance and psychological outcomes may result from the additive or interactive effects of the VR system. For example, a high level of immersion will enhance motivation and performance (Ijsselsteijn et al. 2004). However, immersion may be increased in different ways. It can be enhanced by using a more realistic VR environment as done by Ijsselsteijn et al. (2004). It can also be enhanced if the individual has a high trait level to feel a greater sense of the presence. Interactions between external and individual factors may also influence outcomes. For instance, the introduction of a virtual competitor (VR environment factor) can increase performance (Nunes et al. 2014), although the increase may only be observed if the individual is competitive (athlete factor) as demonstrated in research (Anderson-Hanley et al. 2011; Snyder et al. 2012). Further research is required to examine other interactive effects.

6 Future research directions and recommendations

The present review has highlighted issues that warrant further investigation. Most research to date has focussed on VR tasks that involve aerobic sports (cycling, running, and rowing). More research is required on the effectiveness of a VR environment for learning or improving the mechanics of skill acquisition and performance in skill-based sports (see Sigrist et al. 2015 for an example). The capacity for VR environments to be created in specific and reproducible ways can allow for the training and assessment of skills and decision-making processes. Some of the factors identified as important with aerobic sports (e.g. attentional focus, competitiveness) may also be important in skill-based sports when VR is used.

Research is required to examine the generality of effects with VR. Studies should include more diverse populations, particularly experienced and elite athletes, children, and the elderly. In addition, research has also not examined relationships between performance in VR and real-world environments. Identifying how the two situations differ and how they are the same could inform how VR influences performance and psychological states. The transfer of performance from the virtual environment to the real world has also not been tested, yet it seems an essential requirement if VR is to be used as a training approach for sport.

Further research is required that aims to directly manipulate psychological processes. For example, it has been suggested that VR environments induce a dissociative attentional focus and that this may be related to affective responses (Mestre et al. 2011). Baños et al. (2016) applied this concept by asking overweight and normal weight children to walk on a treadmill while focussing their attention on their physical feelings or while focusing their attention on a virtual environment. Ratings of enjoyment were higher for the VR condition than the self-focused condition, although there were no differences in perceived exertion or feeling states. The findings are promising but are in need of replication and extension. Past research with non-VR tasks has also found that an external associative focus enhances sport and exercise outcomes (e.g. Neumann and Heng 2011; Neumann and Piercy 2013). An external associative focus involves focussing on the effects of movements on the environment and the achievement of task goals (Neumann and Brown 2013; Stevinson and Biddle 1999). Future research could thus use VR to induce an external associative focus and examine its effectiveness in enhancing performance.

Further research is required to elucidate what factors are relevant to performance and affective outcomes. Research using multiple measures or manipulations may be particularly useful to determine the relative amounts of variance in performance attributed to different aspects of the VR environment. In addition, different features of the sport task should be varied. For example, intensity may be a particularly salient factor for aerobic sports. A higher intensity level may switch attentional focus towards internal physiological states (Stevinson and Biddle 1999) and result in individuals focusing attention away from the VR environment. It may be possible to enhance attentional focus on the virtual environment by requiring participants to follow a virtual partner as done by Oliveira et al. (2015).

Finally, the nature of computer-based interactions is becoming more diverse and with a greater amount of overlap between the different forms of technology and their applications. The present review applied a definition of VR that required interactivity with the virtual environment. However, it is acknowledged that researchers are developing and testing systems that employ a virtual environment that the athlete responds to, even though the behaviour of the athlete does not affect any feature of the environment. For example, goalkeeping skills in penalty shots have been examined in both handball (Vignais et al. 2015), and soccer/football (Stinson and Bowman 2014). In these applications, the goalkeeper viewed a virtual environment depicting an individual shooting a penalty and was required to move their body in the predicted direction of the ball. Their movements did not influence the action of the virtual penalty kick taker (e.g. moving too early had no effect). Another instance that resembles VR is the use of augmented reality. In such applications, a user has an indirect view of a physical, real-world environment in which computer-generated input is added to. The input may be visual, auditory, or other senses. This blending of real and virtual environmental elements has yet to be extensively examined in sporting applications.

Based on the present review, recommendations can also be made to ensure appropriate methodology and report in studies. It is recommended that researchers:

  1. 1.

    Use the term virtual reality accurately and consistently in reference to studies that have employed VR technologies according to accepted definitions such as the one proposed here. The term should not be confused with exergames, which refers to the more general case of enhancing physical activity via interactive computer game play. If interactivity with the virtual environment is a particular feature that is to be highlighted, such as in the present review, the term interactive VR may be used.

  2. 2.

    Report participants prior experience with VR in general and with the specific VR system because experience level may be an important factor that influences outcomes.

  3. 3.

    Use a measure of immersion or the presence as a standard part of the protocol because these aspects are a core feature of VR, and the level of immersion has emerged as an important factor that influences outcomes (Ijsselsteijn et al. 2004; Vogt et al. 2015). Such measures include the Reality Judgement and Presence Questionnaire (Baños et al. 2000) and the Presence Questionnaire (Witmer and Singer 1998).

  4. 4.

    Provide full details of the VR system that is used. These details include the name of the system or software used, the participant point of view, the presence of others in the VR environment, the presence of sounds in the VR environment, and the mechanisms through which the participant interacts with the VR environment.

  5. 5.

    Report on relevant procedures that are important psychologically, such as whether participants had choice over the type of VR task or discrete elements within the task.

7 Conclusions

This review identified research studies that have investigated the application of VR to sport. The research findings to date indicate that VR can be a promising adjunct to existing real-world training and participation in sport. A VR-based system for training and participation has several advantages such as enabling athletes to train regardless of weather conditions, providing a means to compete with others in a different geographic location, and allowing precise and replicable control over features of the virtual environment. Future research would benefit from a theoretical framework of VR application to sport (see Fig. 1). The present review has shown that the characteristics of the individual user and system are important factors that can influence a range of performance, physiological, and psychological outcomes. By understanding the experience of when individuals are engaged in sport within a VR environment, researchers, coaches, and athletes will able to use the technology for the benefit of athletes and society in general.