Systematic review of memory assessment in virtual reality: evaluating convergent and divergent validity with traditional neuropsychological measures

Mancuso, Valentina; Sarcinella, Eleonora Diletta; Bruni, Francesca; Arlati, Sara; Di Santo, Simona Gabriella; Cavallo, Marco; Cipresso, Pietro; Pedroli, Elisa

doi:10.3389/fnhum.2024.1380575

REVIEW article

Front. Hum. Neurosci., 16 May 2024

Sec. Brain Health and Clinical Neuroscience

Volume 18 - 2024 | https://doi.org/10.3389/fnhum.2024.1380575

This article is part of the Research Topic Current Research and Future Development of Neuropsychology View all 7 articles

Systematic review of memory assessment in virtual reality: evaluating convergent and divergent validity with traditional neuropsychological measures

$\r\nValentina Mancuso*$ Valentina Mancuso^1*

Eleonora Diletta Sarcinella²

Francesca Bruni¹

Sara Arlati³

Simona Gabriella Di Santo⁴

Marco Cavallo¹

Pietro Cipresso⁵

Elisa Pedroli^1,6

¹Faculty of Psychology, eCampus University, Novedrate, Italy
²Department of Psychology, Research Center in Communication Psychology, Università Cattolica del Sacro Cuore, Milan, Italy
³Institute of Intelligent Industrial Technologies and Systems for Advanced Manufacturing–STIIMA National Research Council of Italy–CNR, Milan, Italy
⁴Department of Clinical and Behavioral Neurology, IRCCS Fondazione Santa Lucia, Rome, Italy
⁵Department of Psychology, University of Turin, Turin, Italy
⁶Department of Geriatrics and Cardiovascular Medicine, IRCCS Istituto Auxologico Italiano, Milan, Italy

Introduction: The evaluation of memory is a crucial aspect of both cognitive research and clinical applications, as it offers valuable insights into an individual’s cognitive wellbeing and performance. Conventional neuropsychological assessments represent the established method for assessing different aspects of memory. Recent technological advancements, specifically in the field of virtual reality (VR), have introduced novel methods for evaluating memory.

Objective: This systematic review aims to examine the current state of memory assessment using VR technologies, assessing the degree of convergence and divergence between VR-based memory assessments and conventional neuropsychological tests.

Method: A systematic review of the literature was conducted searching PubMed, PsycINFO, Web of Science databases, leading to the incorporation of 24 studies. Studies were grouped according to the examined memory domain (episodic, prospective, spatial domain). Convergence and divergence validity were examined for each, and information on software and hardware features was collected.

Results: This review demonstrates a notable alignment between VR-based memory assessments and conventional neuropsychological tests. Moreover, VR tasks have shown to exhibit associations with executive functions and overall cognitive performance. The inclusion of various ecological contexts, such as residential environments, commercial establishments, and simulated scenarios, serves to augment the ecological validity of memory evaluations conducted in VR.

Discussion: The findings indicate that VR assessments demonstrate a functional perspective by effectively capturing the dynamic relationship between memory, executive functions, and overall cognitive performance. Nevertheless, it is imperative to acknowledge and tackle certain constraints that may hinder the widespread adoption and utilization of VR tasks. These limitations encompass factors such as restricted accessibility to VR tasks and the presence of heterogeneity in VR hardware and software. The dynamic and ever-changing nature of VR technology presents a range of potential avenues for future investigation and utilization in the domain of memory evaluation.

1 Introduction

The primary objective of cognitive assessment is to obtain a thorough understanding of an individual’s cognitive functioning and detect any possible impairment (Benton, 1994). Traditional assessment tools have played a crucial role in diagnosing and assessing the extent of cognitive impairments.

Due to the complex and interdependent nature of memory, researchers have set up a range of experimental and clinical assessments to evaluate memory and unravel its components.

The assessment of memory impairments and their implications for daily functioning holds substantial importance, considering the widespread occurrence of various illnesses that can compromise memory capabilities and subsequently affect everyday activities and overall quality of life, such as stroke and neurodegenerative pathologies.

Memory tasks commonly employed in clinical practice or experimental investigations on aging are structured to allow for precise control over task parameters and testing conditions. This intentional design facilitates the examination of the complex processes at play. However, it is important to note that traditional neuropsychological tasks are designed to elicit the individual’s optimal performance in ideal conditions, which may not effectively capture impairments in memory functioning that occur in everyday life situations. In fact, one possible critique of numerous clinical assessments of memory is their limited ecological validity (Pause et al., 2013). The notion of ecological validity pertains to the degree to which the experimental conditions faithfully replicate real-world settings (Tupper and Cicerone, 1990) or the extent to which the conclusions drawn from a study can be generalized to real-life contexts (Franzen and Wilhelm, 1996). Traditional evaluations demonstrate limited resemblance to ordinary, everyday memory situations (Parsons, 2011). In real-world scenarios, the act of memorization often occurs within environments, with high levels of noises and complexity. Furthermore, this cognitive process is frequently performed simultaneously with the completion of additional tasks, such as participating in dialog, ambulating, or engaging in problem-solving endeavors. The testing conditions in experimental and clinical settings differ significantly from the conditions described earlier. In these settings, participants usually perform their tasks in a quiet environment, receive clear instructions about the tasks, primarily work with one-dimensional material, and focus their attention solely on the task at hand. Multiple studies have provided evidence indicating that suboptimal performance on these assessments does not consistently align with compromised performance in practical contexts (Wilson, 1993; Sbordone and Long, 1996; Manchester et al., 2004; Bottari et al., 2009). The previously mentioned situation poses a challenge to the practicality of using traditional assessments as a means to make a diagnosis, intervene and offer appropriate support to individuals with cognitive impairments.

Moreover, the correlation between performance on conventional memory assessment tools (such as the California Verbal Learning Test–CVLT–or Wechsler Memory Scale-Revised–WMS-R) and indicators of daily functioning (such as self and informant memory diaries, patient and informant memory questionnaires, and clinical ratings) tends to be modest at most. Various studies (Goldstein and Scheerer, 1941; Sunderland et al., 1983; Kaitaro et al., 1995; Chaytor and Schmitter-Edgecombe, 2003) have examined this correlation and have reported only modest connections between the two domains.

The application of virtual reality (VR) offers a promising solution for evaluating memory in ecologically authentic and standardized environments. This cognitive assessment technology is of great interest due to its ability to simulate naturalistic environments, while maintaining safe and reproducible experimental conditions, under the complete control of the experimenter (Rizzo et al., 2008). According to Fuchs (2006), VR offers users an immersive experience within a dynamic virtual environment, allowing them to participate in cognitive and sensorimotor activities while interacting with virtual stimuli. One notable benefit of VR resides in its capacity to generate immersive environments that faithfully reproduce the sensory elements of the real world, including visual landscapes and audible dialogs. Moreover, VR can successfully integrate the cognitive and physical challenges that individuals face in their everyday activities. Therefore, one could argue that carefully designed VR tasks have the potential to more accurately represent real-world capabilities compared to traditional neuropsychological evaluations (Rizzo et al., 2008).

The utilization of various hardware platforms in the development of VR tasks for memory assessment studies demonstrates the wide array of immersive experiences that can be attained. The range of hardware options can be classified into three modalities: non-immersive, semi-immersive, and fully immersive (Takac et al., 2021). Each modality provides unique levels of immersion and interaction. Studies have shown that non-immersive VR setups, commonly implemented through the use of computers and tablets, generally entail user engagement with a two-dimensional screen. The participants interact with the virtual environment by utilizing conventional input devices, such as a joystick or a touchscreen pad. Semi-immersive VR experiences, as observed in research utilizing video projectors, provide a moderate degree of immersion. Video projectors are utilized to project a virtual environment onto a larger screen or physical space, thereby generating a heightened sense of immersion for participants, while simultaneously enabling them to maintain awareness of their physical surroundings. The utilization of head-mounted displays serves as a prime example of fully immersive VR setups, which provide the utmost level of immersion. Head-mounted displays fully obstruct the participant’s visual perception of the physical surroundings and substitute it with the simulated digital environment.

Besides the hardware used, the virtual environments can be distinguished into computer-generated or real-life scenarios captured using 360° technologies (Mancuso et al., 2024). In the realm of computer-generated virtual environments, a notable benefit lies in their substantial capacity for customization. Academic researchers possess the capacity to strategically construct scenarios that closely correspond to the specific goals of their memory evaluations. This characteristic holds significant value in situations where the objective is to establish controlled and standardized conditions for cognitive assessment. One additional advantage of computer-generated environments is their capacity for reproducibility. The replication of virtual scenarios is a straightforward process, thereby increasing the dependability of research outcomes. The attainment of consistency in the evaluation process among diverse participants or at different time intervals is attainable, thereby establishing a solid basis for rigorous scientific investigation.

Nonetheless, computer-generated virtual environments may exhibit a deficiency in terms of realism when juxtaposed with authentic real-life settings. Although advancements in technology have undoubtedly enhanced the feeling of being present and fully engaged in virtual environments, it is important to acknowledge that there are still constraints when it comes to accurately reproduce the depth and complexity of real-life experiences. The potential absence of realism in these virtual environments gives rise to inquiries regarding the ecological validity of memory assessments conducted within them.

On the other hand, the utilization of 360° technologies enables the depiction of real-life situations with an exceptionally elevated degree of authenticity (Borghesi et al., 2022; Mancuso et al., 2024). Individuals who are fully engaged in these settings frequently experience a sense of authentic presence, thereby augmenting the ecological validity of evaluations of memory. The inherent behavior and reactions of individuals are more prone to correspond with situations and stimuli encountered in actual life circumstances.

Moreover, the presence of stimulus variation in authentic situations can pose a greater challenge and elicit more effective engagement of memory functions compared to controlled computer-generated environments. The diverse array of sensory encounters can result in more genuine reactions, especially when evaluating memory within settings that closely mirror everyday experiences (Mancuso et al., 2023).

Nevertheless, actual situations present inherent difficulties. Computer-generated environments are more controllable and replicable compared to other types of environments. This is because computer-generated environments have the potential to introduce confounding variables and impede the standardization of assessment conditions. Moreover, the process of capturing authentic situations through the utilization of 360° technologies can impose a significant demand on resources, requiring the acquisition of specialized equipment, the employment of trained personnel, and occasionally encountering challenges related to the availability of suitable locations and logistical considerations.

Many studies have provided evidence of notable correlations between performance on VR tasks and traditional neuropsychological assessments that measure the same cognitive functions (Matheis et al., 2007; Armstrong et al., 2013; Lalonde et al., 2013; Nolin et al., 2016; Pedroli et al., 2020, 2022). This evidence confirmed that VR protocols effectively demonstrated construct validity. Moreover, prior studies have provided evidence that diverse VR assessments can successfully differentiate between two separate cohorts based on their task performance levels, for example, between older adults with dementia and non-cognitively impaired seniors (Rizzo et al., 1997; Rand et al., 2007; Cushman et al., 2008; Werner et al., 2009; Banville et al., 2010; Tarnanas et al., 2013; Allain et al., 2014; Zygouris et al., 2015). More interesting for the aim of the present systematic review, the evaluation of performance in VR tasks often involves comparing it to performance in equivalent real-world tasks (Waller et al., 2001; Cushman et al., 2008; Renison et al., 2012; Allain et al., 2014; Vallejo et al., 2017). For example, previous research has established notable associations between spatial learning, as evaluated using a virtual maze, and its real-world counterpart in the spatial domain (Waller et al., 2001). Furthermore, Cushman et al. (2008) have documented the existence of correlations between spatial navigation abilities as assessed in a virtual hospital lobby and those demonstrated in the corresponding real-world environment.

To date, a multitude of studies have made noteworthy contributions to the advancement of innovative VR tools that are specifically tailored for the assessment of memory functions. However, most of them have primarily focused on the validation of measures by using samples composed mainly by young individuals. Furthermore, a noteworthy concern is the dearth of research investigations that offer substantiation for the construct validity and convergent validity of assessments based on virtual reality. These studies must incorporate meticulous correlation analyses utilizing established neuropsychological tests.

The primary objective of this systematic review is to comprehensively evaluate and systematically synthesize the current body of literature related to the evaluation of memory utilizing VR technologies, examining the degree of convergence and divergence observed between VR-based memory assessments and conventional neuropsychological memory tests.

Secondly, our review examines the hardware and software characteristics of the applications for memory assessment and protocols for their use, posing particular attention on the use of computer-generated virtual environments vs. 360° video technologies.

2 Methods

Preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines were followed (Page et al., 2021).

2.1 Search strategy

This section aims to provide a review of the studies that have assessed memory functions in virtual environments. Three high-quality databases (PubMed, PsycINFO, Web of Science) were used to perform the search on 4 August 2023. Keywords used for each separate string were (“memory”) AND (“assessment” OR “assess*” OR “evaluation” OR evalu* A) AND (“virtual reality” OR “VR” OR “virtual” OR “360-degree” OR “360°” OR “immersive” OR “panoramic”) and were searched through Title/Abstract, Abstract, Topic, respectively, for each database without time limit. Figure 1 shows the paper selection procedure and the number of selected/excluded articles.

FIGURE 1

Figure 1. PRISMA flow diagram.

2.2 Selection criteria

Studies that uniquely investigated the convergence and divergence validity of virtual reality-based memory assessment tasks with traditional neuropsychological measures have been included. In particular, we selected papers written in English (reason 1); experimental studies on human beings as reason 2 (we excluded reviews, meta-analysis, protocol articles, studies with animals, perspective); studies including participants over 60 years old (reason 3); use of virtual reality technologies (reason 4); assessment studies as reason 5 (we excluded usability, feasibility, rehabilitation and training studies); memory as cognitive domain evaluated (reason 6); studies with measures of convergency and divergency with traditional neuropsychological measures (reason 7).

3 Results

A total of 24 studies were identified and included in the present review.

Table 1 presents a comprehensive overview of the VR assessment tools, providing a summary of their descriptions. Table 2 displays the convergence and divergence validity of these tools in relation to neuropsychological tests. Figure 2 shows the risk of bias assessed using the Quadas 2 tool (Whiting et al., 2011).

TABLE 1

Table 1. Description of the Virtual assessment task, the hardware employed, and the virtual environment.

TABLE 2

Table 2. Description of the validation procedures: population employed, the VR task, neuropsychological measures administered, other measures, results of the VR task, and the correlation of this task with neuropsychological measures.

FIGURE 2

Figure 2. Assessment of risk of bias using Quadas 2 tool.

Through a thorough examination of the available literature, it emerges that VR tools that have been tested and proven to be reliable on older adults and have provided correlations with established neuropsychological assessments cover a wide array of memory areas such as episodic memory (Plancher et al., 2010; Arvind Pala et al., 2014; Kempe et al., 2015; Zygouris et al., 2015; Sauzéon et al., 2016; Ouellet et al., 2018; Corriveau Lecavalier et al., 2020; Eraslan Boz et al., 2020; Limoncu et al., 2021; Turner et al., 2021; Bruni et al., 2022; Pieri et al., 2022), verbal memory (Gottlieb et al., 2022), prospective memory (Lecouvey et al., 2019; Hogan et al., 2023), and spatial memory (Cushman et al., 2008; Tippett et al., 2009; Morganti et al., 2013; Lee et al., 2014; Serino et al., 2015; Mohammadi et al., 2018; Zhang et al., 2021; Da Costa et al., 2022; Rekers and Finke, 2023).

3.1 Research themes

3.1.1 Virtual reality assessment tools for episodic memory

The model of episodic-like memory is a simplified representation of memory for episodes, which is characterized by three key conditions: the specific details of “what” occurred during an episode, the location or “where” it took place, and the temporal aspect of “when” it happened (Allen and Fortin, 2013).

The domain of memory assessment has experienced significant growth in recent years due to the increased accessibility of VR technologies. Nevertheless, it is important to highlight that a considerable proportion of these studies predominantly focused on comparatively younger demographics, despite the expected applicability of these methodologies for older individuals. Furthermore, it is worth noting that there is a noticeable deficiency in the existing body of literature, as several of these studies have neglected to include convergent validity measures, thereby restricting the thorough validation of these memory assessment tools that are based on VR. This specific section of the discussion aims to fill the existing gap in knowledge by shifting attention toward studies that have specifically focused on validating these tools among older populations. This overview aims to provide insights into the efforts made to determine the effectiveness and practicality of VR technology in evaluating episodic memory functions among older adults.

Two studies (Arvind Pala et al., 2014; Sauzéon et al., 2016) assessed episodic memory using the HOMES test: the participants were presented with a large display screen that depicted two versions (referred to as versions A and B) of a virtual apartment. Each version contained a total of 40 objects, with 20 objects unique to each version and 20 objects shared between both versions. There were four distinct categories, identified by the names of the rooms to which objects belong to: bedroom, bathroom, kitchen, plus an unspecified fourth category. Each of these categories consisted of ten objects, with five objects being specific to each category and five objects being common to both versions. The participants were initially exposed to a specific iteration of the apartment, referred to as version A. This iteration was shown to the participants on two separate occasions, and after each presentation, the participants were asked to engage in a free recall task. Following a duration of 10 min, a conclusive yes/no recognition task was subsequently administered to all participants. The first study compared older adults with traumatic brain injury (TBI) patients. Overall, there was an observed increase in performance levels from trial 1 to trial 2. A post hoc analysis was conducted, which indicated that there were no statistically significant differences observed between the two groups. In general, participants exhibited superior performance on the recognition task in comparison to the free recall task, albeit with suboptimal performance observed in both groups. In the study conducted by Sauzéon et al. (2016), a comparison was made between older adults and patients diagnosed with AD. The findings revealed that older adults exhibited superior recall performances in comparison to AD patients. The ANOVA analysis performed on the retrieval tasks (specifically, free-recall trial 2 versus recognition hits) revealed significant main effects for both group and retrieval tasks, as indicated by the within-subject variable. The results of the post hoc follow-up test indicated that older adults exhibited lower mean performances, and these performances were further diminished in individuals with AD when compared to other groups. In general, as expected the participants exhibited superior performance in the recognition task compared to the free recall task.

Another tool used to assess episodic memory is the Virtual Shop (La boutique virtuelle) (Ouellet et al., 2018; Corriveau Lecavalier et al., 2020). In this task, participants initiated the task by positioning themselves in proximity to a cashier situated behind a countertop. They were subsequently provided with a compilation of 12 virtual images depicting commonly encountered items (e.g., belt, milk), which they were instructed to commit to memory and subsequently retrieve within the store environment. During the process of encoding, extraneous dialogs were delivered through the use of headgear with the intention of replicating a disruptive auditory setting. After the conclusion of the previous item, the program implemented a 20-s conversation between the cashier and the participant. This conversation served as a filled interference delay, wherein the cashier asked the participant about the time displayed on their right, for example. Following the conclusion of the designated delay period, the cashier issued instructions to the participant, tasking them with retrieving the items within the store that they had previously observed. Subsequently, the participant was granted unrestricted mobility within the room in order to locate and choose the items that were previously presented on the learning list. The shop contained a total of 24 items, consisting of 12 target items and 12 distractors.

Obreco is another tool developed to assess episodic memory using 360° videos and images. In the first version (Pieri et al., 2022), participants wear a head-mounted display and adopt the viewpoint of the 360-degree camera. The clinician engages in movement within the room and brings the target objects in close proximity to the camera for a duration of 5 s. During this time, the participants are required to verbally identify the object being presented. Subsequently, the tasks solely necessitate the retention of the ten objects that were previously presented during the Encoding Phase, with a time interval of ten minutes. Subsequently, participants are instructed to locate and nominate all ten objects that were previously presented during the encoding phase, amidst a set of 17 additional non-target objects. In an upgraded version (Bruni et al., 2022), participants are exposed to a domestic environment, a living room, from the experimenter’s first-person perspective. They follow the perspective of the 360° video which selectively stops at the 15 designated objects for a duration of 3 s each, while simultaneously affixing a label inscribed with the appellation “Marco” to each item. In the living room, there are an additional 15 objects present, serving as distractors. During this phase, participants are provided with instructions to verbally label all of the targets. Following a period of 15 min, participants are given instructions to recall and identify as many objects as possible from the previous encoding phase. Subsequently, the participants are provided with instructions to thoroughly examine the previously observed living room environment, identify and label the specific target objects from the previously encountered items, as well as an additional set of 15 unfamiliar distractor objects.

One common task/environment frequently used to assess episodic memory is the supermarket task which typically requires memorizing a shopping list and then buying those products in the supermarket with additional different demands like picking the products in the order they walk by them (Kempe et al., 2015) or correctly select the precise amount needed to pay for the items bought (Zygouris et al., 2015; Eraslan Boz et al., 2020; Limoncu et al., 2021).

Similarly, in the Virtual Reality Functional Capacity Assessment Tool (VRFCAT) (Turner et al., 2021), a tool developed to measure functional capacity of subjective cognitive decline, participants engage in an exploratory activity within a kitchen environment to ascertain the available culinary resources for the recipe at hand. Subsequently, individuals employ a bus timetable to locate a suitable bus route that will transport them to a grocery store. The individuals locate and acquire the requisite items from the retail establishment, subsequently employing the aforementioned timetable to identify the bus that will facilitate their return to their place of residence.

With the same purpose but a different setting, Plancher et al. (2010) developed an episodic assessment tool set in a virtual urban environment based on Paris’ photos: At the site of a vehicular collision, two automobiles collided, resulting in the activation of a horn and the emission of black smoke. The interconnectedness of buildings facilitated seamless transitions between distinct areas. Various elements comprised the town, including individuals, waste receptacles, obstacles, vegetation, advertising displays, and stationary vehicles. Each specific area is comprised of a distinct combination of these various elements. Near the town hall, a pedestrian can be observed traversing the area, accompanied by a billboard, barricades, and a cluster of trees. Following their experience of driving in the town, all participants were subjected to identical episodic memory tests. Participants were instructed to engage in free recall of verbal components by providing written responses. Specifically, they were asked to recall the “what” (i.e., the content), “where” (i.e., the location), “when” (i.e., the time), and associated details of as many elements as they could remember. Following the completion of the recall task, a recognition test was administered. The participants were required to select the item they had observed in the town from a set of three distinct images. The examination consisted of a total of ten inquiries pertaining to the various elements and their respective positions within the municipality. The performance of older adults was similar in both forms of encoding, but the performance of younger adults obviously increased in intentional encoding compared to incidental encoding. However, no effect of age and encoding was observed on the details of memories.

Gottlieb et al. (2022) developed and validated a virtual reality-based Rey Auditory Verbal Learning Test (VR-RAVLT): VR-RAVLT immerses the participant within a simulated office environment, wherein a virtual personal assistant (represented as an avatar) is positioned behind a desk. The avatar provides the participant with a compilation of 15 locations that they are required to visit within a single day and subsequently challenges them to recall as many of these locations as they can. The avatar notifies the participant of her impending early departure and states her intention to reiterate the list in order to enhance the participant’s recall of all the locations, employing a procedure akin to that used in the standard RAVLT. List B comprises a total of fifteen locations that the participant is required to visit on the following day. The research assistant documents participant responses using a format akin to that employed in the GS-RAVLT. Group effects were observed for the acquisition and the retention variables in the VR-RAVLT, respectively, with poorer scores for the older-adults group.

3.1.1.1 Convergence and divergence validity with traditional neuropsychological tests

The results of convergence and divergence of the virtual reality tests with neuropsychological tests are shown in Table 2. Overall, the HOMES (Arvind Pala et al., 2014; Sauzéon et al., 2016) scores correlated with all the neuropsychological measures; the Virtual Shop task correlated with the immediate and delayed free recall scores of the traditional verbal memory task in both younger and older adults (Corriveau Lecavalier et al., 2020) and traditional neuropsychological measure of episodic memory (Ouellet et al., 2018).

The OBRECO task (Bruni et al., 2022; Pieri et al., 2022) correlated with AM and RAVLT, i.e., executive functions and memory measures. The supermarket of Kempe et al. (2015) correlated with a verbal complex span; the supermarket of Zygouris et al. (2015) with most of the neuropsychological measures; the supermarket of Eraslan Boz et al. (2020) negatively correlated with general cognitive status, verbal and visual memory and visuo-spatial construction but positively with executive functions. The supermarket of Limoncu et al. (2021) correlated with memory, executive functions and visuo-spatial functions. The VRFCAT (Turner et al., 2021) was correlated with neuropsychological assessment battery (NAB) and delayed recall. The VR-RAVLT (Gottlieb et al., 2022) positively correlated with the standard RAVLT.

3.1.2 Virtual reality assessment tools for prospective memory

Regarding prospective memory (PM), two virtual tools have been developed: the Virtual Reality Prospective Memory Shopping Task (VRPMST) (Hogan et al., 2023), and a ride in a virtual car (Lecouvey et al., 2019). Prospective memory refers to the cognitive ability to remember and successfully execute intended actions or tasks at a future point in time. The differentiation between time-based (TB) and event-based (EB) intentions is commonly made based on the characteristics of the stimulus that initiates the retrieval process.

The Virtual Reality Prospective Memory Shopping Task (VRPMST) (Hogan et al., 2023) involves the completion of a task consisting of 12 errands within a virtual shopping center. Participants navigate through the virtual environment and enter stores in a predetermined sequence. The tasks have been designed to replicate the routine activities typically carried out within a shopping center setting. In the context of time-based project management, individuals are informed by their healthcare provider to monitor their heart rate. Consequently, they are required to assess their heart rate at regular intervals of 3 min by pressing the designated key on the iPad. Participants can verify the elapsed time on their virtual watch for a brief duration of three seconds by using the TIME button on the iPad. The VRPMST encompasses two distinct categories of event-driven project management tasks. Initially, participants are informed that they will be monitoring their food expenses. Consequently, upon the acquisition of food items, specifically four items, it is necessary for individuals to obtain a receipt. Furthermore, the participants are provided with information regarding the loss of their glasses during their previous shopping excursion. Consequently, it is necessary for them to inquire with the security personnel at the center on each occasion (a total of four times) regarding the potential retrieval of their glasses. This test has been validated in two groups: individuals with stroke and controls. Both groups scored similarly on the ongoing task, indicating that they did not find the task too difficult. Controls performed significantly better on time-based PM compared to individuals with stroke. Additionally, controls monitored the time significantly more than individuals with stroke (large effect size). Controls scored higher than individuals with stroke on event-based PM; however, no significant difference was found. VRPMST monitoring significantly correlated with VRPMST time-based PM for both groups.

In the study of Lecouvey et al. (2019), following a period of familiarization, the participants were notified that they would be immersed in a virtual city with the objective of retrieving a companion from the train station, located at the conclusion of said city. Additionally, they were tasked with fulfilling various intentions throughout their journey. The participants were presented with a set of seven intentions displayed on a laptop computer. In 50% of the cases, a significant association was observed between the anticipated cue and the subsequent retrospective component (referred to as Link-EB). Examples of Link-EB scenarios include purchasing a stamp booklet at the post-office. Conversely, in the remaining 50% of cases, a weak association was found between the anticipated cue and the retrospective component (referred to as NoLink-EB). An example of a NoLink-EB scenario is buying a pair of glasses at the fountain. The persistent intention involved the regular administration of medication for tuberculosis, with a frequency of once every two minutes. Subsequently, to verify the accuracy of the encoding process, a cued recall assessment was conducted after the presentation of the intentions. A 10-min interval was observed between the encoding and retrieval of intentions, during which participants completed questionnaires. The individuals were reminded of their obligation to collect a companion from the train station and accomplish various objectives during their journey. In order to accomplish this, participants were required to bring the vehicle to a halt at the designated moment or location (referred to as the prospective component) and subsequently inform the experimenter of the specific action they were instructed to execute (known as the retrospective component). This test has been administered in patients with mild AD and controls: cognitively normal older individuals needed a lower number of trials to encode–and recalled more–Link-EB, NoLink-EB, and TB intentions when compared to patients with AD.

3.1.2.1 Convergence and divergence validity with traditional neuropsychological tests

The results of the first study indicate that VR tasks are successful in capturing and reflecting the PM abilities that are evaluated by conventional measures. Moreover, the study unveiled significant correlations between time-based prospective memory tasks conducted in VRPMST and the corresponding time- and event-based PM tasks conducted in laboratory-based dual-task paradigms. Furthermore, a significant and strong association was observed between both time and event-based PM as evaluated using VRPMST and the outcomes obtained from the Trail Making Test (TMT), the Hopkins Verbal Learning Test-Revised (HVLT-R), and the Montreal Cognitive Assessment (MoCA).

In the second study, the NoLink-EB intentions (NoLink Event-Based) study revealed correlations between cognitive functions, specifically shifting as assessed by the Trail Making Test Part B, and semantic memory. Nevertheless, the study did not uncover any noteworthy associations with regards to Link-EB intentions (Link Event-Based).

In addition, it is worth noting that the potential aspect of PM exhibited a positive correlation with the TMT-B, a measure commonly used to assess cognitive flexibility and executive functioning, as well as with planning abilities.

3.1.3 Virtual reality assessment tools for spatial memory

Among all the studies investigating spatial memory in VR, only 9 of these compare the performance of older adults with classical neuropsychological tests:

Tippett et al., 2009, chose two routes, Path A and Path B, with predetermined starting and ending points. Path A had three learning trials and Path B was one in the VE task. Each trial began with passive observation of the city path. Participants were instructed to focus on the city’s landmarks’ spatial positioning and relative positions. After thoroughly observing the path, participants were instructed to return to the starting point and reproduce it as best they could. Path B was used as an interference condition with a different path design in the same urban setting to reduce perceptual priming effects. After these experimental trials, participants performed recall trials with 5- and 20-min delays. They navigated Path A by memory without passive viewing during these trials.

In Zhang et al., 2021, the virtual environment task was to find and maintain a fixed position for nine learning trials. Participants must arrive at the same place throughout the study. In subsequent trials, participants must use instructions and tasks to find the location while navigating the virtual environment. The trial ended immediately if participants did not reach the destination within 90 s, and the next trial began. Participants could use the egocentric strategy—remembering sequential body turns at each Y-shaped intersection—or other strategies to reach the target location during the acquisition phase. Each participant’s navigational proficiency was assessed using several metrics for the 9 learning trials. This included speed, distance error, rotation, percentage of successful individuals, and percentage of successful trials. The probe trial maze structure was similar to the learning phase, but without landmarks. All distal cues were intentionally removed.

Morganti et al. (2013) created two spatial assessment tasks with no landmarks and uniform building textures. Participants were shown one of five complex paper-and-pencil mazes (PP-MT) on a computer screen in the VR-Maze spatial task (VR-MT). The PP-MT tests allocentric spatial knowledge by having participants draw the trajectory from the starting point to the exit and find the best route. After completing PP-MT, participants were instructed to use it to find the exit point in the VR-MT. This assessed the transfer of allocentric to egocentric spatial knowledge. To begin VR-MT study, participants were shown the correspondence between the initial positions on the physical paper and the virtual representation of each maze. Participants had 10 min to reach the exit point to complete a maze.

Participants faced the computer screen with the Money’s Road Map Test (PP-RMT) at the bottom in the VR-RMT task. In order to navigate a predetermined route on a stylized urban map, the PP-RMT requires participants to indicate their left or right direction at each turn using 32 dotted steps. Since the dotted pathway has a non-linear trajectory that goes both away from and toward the person, the solutions require egocentric mental rotation. The individual cannot manipulate the map or make any head or body movements to find the correct answer. Participants were then instructed to use the PP-RMT to navigate the VR-RMT by following the dotted line on the PP-RMT. The purpose of this task is to compare the process of converting spatial knowledge from an allocentric to an egocentric frame of reference for navigating a simulated virtual reality environment to mentally simulating the same environment using a sketched map.

Cushman et al. (2008) created eight tasks. Participants saw the route on a computer screen during route learning. The second presentation video playback stopped at ten decision points. After the test trip along the designated route, participants had 1 min to verbally identify and recall as many objects or landmarks as possible. In the Self-orientation phase, participants saw 10 images of test route objects and locations. These images were carefully selected and presented sequentially in two sets of five. The subject faced an outside wall and the locations were strategically placed at 45° intervals. Participants used a computer mouse to mark the next decision point while viewing a scaled lobby on the video display during route drawing. In the Landmark experiment, participants were instructed to recall only those objects or fixtures that helped them navigate their second lobby traversal. Ten photos were displayed on the screen for photograph recognition. Five photos were chosen from the test route and five from the Medical Center. The subjects determined whether each photograph was from the test route. Participants were shown 10 more test route photos during the Photograph location task. They received a scale lobby outline with 10 lettered locations. Participants had to indicate the scene outline location for each scene, and their performance was measured by correct responses. The previous task, the video location task, showed ten brief video clips of the subject navigating the test route three times. After each visual presentation, participants marked an X on an empty map to indicate the clip’s start. They also drew an arrow from the X to show movement direction and magnitude. Correct responses were based on X placement and arrow direction.

In Serino et al. (2015) created a controlled simulated room for experimentation. Two entities—a botanical specimen and a mineral specimen—were displayed with an arrow on the ground. This north-facing arrow symbolized the start of navigation. Three trials were given to participants to memorize the plant’s positions. Two retrieval tasks were created. Participants first had to locate the object on an authentic map. They were asked to recall spatial allocentric information independent of their perspective. The second task required participants to access an empty virtual environment. In a retrieval task without spatial allocentric information, participants had to place the plant relative to another object.

Rekers and Finke, 2023 tested the feasibility of VIENNA, which comprises a single instruction trial, followed by two practice trials, and finally, 12 main trials. The trials conducted exhibit a visual representation from the viewpoint of a protagonist engaging in the exploration of virtual hallway environments. Furthermore, a representation of the environment is presented in the form of an allocentric map during every trial. Participants are instructed to engage in mental tracing of the character’s position and subsequently indicate the door that the character selected at the conclusion of the trial. Significantly, the design of this task does not depend on episodic memory and does not necessitate active exploration or navigation by the participant. As a result, the information available to solve the task is made uniform across all participants.

Da Costa et al., 2022 created the SOIVET Maze and Route Tasks. Participants navigated a virtual maze using the original MRMT map in the first one. A green indicator on the map indicates the most recent accurate direction change to reduce cognitive load. No topographical landmarks were located. Participants were told to follow the map’s route to navigate first-person. They also had to update their spatial awareness by using their body position at each maze junction. The second task, SOIVET Route task, involved participants accessing the virtual lobby reconstruction at the Central Institute of the University of São Paulo Clinics Hospital. A virtual avatar performed a predetermined sequence of five locations in the hospital lobby and surrounding area. Participants were told to track the avatar in first-person. After that, participants were told to independently travel the same route and visit the five designated locations in the SOIVET Route immediate order. After a 20-min break, participants repeated the route with an SOIVET Route delay.

The virtual reality navigation task (VRNT) by Mohammadi et al. (2018) used two environments, the virtual neighborhood and virtual maze, with three-dimensional first-person and two-dimensional overhead views. Participants first saw a two-dimensional aerial view for 60 s, then a three-dimensional first-person view. After seeing the two-dimensional aerial perspective, participants were told to find the goal (e.g., parking in the simulated neighborhood, the ball in the virtual maze). Each participant had three practice trials to learn the task, then five assessment trials.

Lee et al. (2014) created the virtual radial arm maze (VRAM), in which participants were told they were in a simulated environment with a central region and six appendages. The virtual room had colorful objects and visual cues for directions. The room was unchanged throughout each trial. Although instructed to find the three treasures quickly, no time limit was set. After finding all three treasures, the trial ended, and the participants reconvened at the maze’s center for the next trial. Five trials were run with 10-s intertrial intervals. All participants received the same rewarded arms. Working memory errors were measured by reentry into the same arm, while reference memory errors were measured by reentry into arms without rewards. The distance and time taken to find all rewards in each trial were also recorded.

3.1.3.1 Convergence and divergence validity with traditional neuropsychological tests

The research conducted by Mohammadi et al. (2018) revealed noteworthy associations between their spatial memory task and established memory assessments, such as Rey immediate recall and delayed recall. Significant correlations were observed between the VR-MT and VR-MTM tests and Corsi’s supra-span task, which is a commonly employed assessment tool for evaluating visuospatial memory and navigation abilities.

There was a notable association observed between the Vienna task and visuospatial short-term memory. Nevertheless, the present task failed to exhibit a substantial correlation with episodic memory, as assessed by the proportion of elements recalled during the delayed free recall phase of the Rey-Osterrieth Complex Figure Test (ROCFT).

In contrast, the study conducted by Mohammadi et al. (2018) found a correlation between the ROCFT and simplified Rey figure test (SRFT) with the virtual neighborhood task. Similarly, Lee et al. (2014) observed a correlation between ROCFT and SRFT with the virtual radial arm maze (VRAM). Additionally, Tippett et al. (2009) proposed a task that also exhibited a correlation with ROCFT and SRFT. These observations imply that the conventional visuospatial memory tests exhibit similar characteristics to their virtual reality counterparts, thus indicating their concurrent validity in evaluating spatial memory.

Just like tasks that evaluate episodic memory, a number of spatial memory assessments also exhibited associations with executive functions. In particular, Tippett et al. (2009) and Morganti et al. (2013) reported significant correlations between the Trail Making Test (TMT) and the variables under investigation. On the other hand, Rekers and Finke (2023) identified significant associations between visuospatial working memory, perspective-taking, and mental rotation. Furthermore, Da Costa et al. (2022) have documented correlations between the Tower of London task and virtual spatial memory assessments, suggesting that the latter may encompass elements of executive functioning.

Multiple studies have also demonstrated associations between evaluations of virtual spatial memory and overall cognitive functioning. In their study, Morganti et al. (2013) established associations between the mini-mental state examination (MMSE) and both VR-MT and VR-RMT. The study conducted by Rekers and Finke (2023) revealed a significant correlation between the VIENNA task and the mini-mental state examination (MMSE). Moreover, the study conducted by Lee et al. (2014) revealed significant associations between Virtual Reality Assessment Measure (VRAM) scores and MMSE scores. In contrast, Da Costa et al. (2022) identified associations between the Addenbrooke’s Cognitive Examination (ACE) and the SOIVET Maze task, underscoring the capacity of virtual spatial memory evaluations to serve as indicators of broader cognitive performance.

3.2 Populations

The demographic groups primarily subjected to examination through the utilization of VR tools for the purpose of memory assessment primarily consist of healthy older adults (Cushman et al., 2008; Tippett et al., 2009; Plancher et al., 2010; Morganti et al., 2013; Arvind Pala et al., 2014; Lee et al., 2014; Kempe et al., 2015; Serino et al., 2015; Zygouris et al., 2015; Sauzéon et al., 2016; Mohammadi et al., 2018; Lecouvey et al., 2019; Corriveau Lecavalier et al., 2020; Eraslan Boz et al., 2020; Limoncu et al., 2021; Zhang et al., 2021; Bruni et al., 2022; Da Costa et al., 2022; Gottlieb et al., 2022; Pieri et al., 2022; Rekers and Finke, 2023), older adults with subjective cognitive decline (Ouellet et al., 2018), individuals diagnosed with Mild Cognitive Impairment (MCI) (Cushman et al., 2008; Tippett et al., 2009; Serino et al., 2015); MCI-single domain (SD) (Lee et al., 2014; Serino et al., 2015; Zygouris et al., 2015; Mohammadi et al., 2018; Eraslan Boz et al., 2020), MCI-multiple domain (MD) (Zygouris et al., 2015; Mohammadi et al., 2018), patients with Parkinson’s disease (PD) (Turner et al., 2021), and PD-MCI (Turner et al., 2021), cognitively normal individuals with small vessel disease (SVD) (Limoncu et al., 2021), individuals with SVD with cognitive impairment (Limoncu et al., 2021); patients with stroke (Hogan et al., 2023), patients with AD (Cushman et al., 2008; Morganti et al., 2013; Lee et al., 2014; Serino et al., 2015; Sauzéon et al., 2016; Mohammadi et al., 2018; Lecouvey et al., 2019). See Table 3 for descriptions of the samples.

TABLE 3

Table 3. Descriptions of the populations.

3.3 Hardware

Among the studies under review, five employed video projectors as the primary tool for conducting virtual reality tasks (Tippett et al., 2009; Plancher et al., 2010; Arvind Pala et al., 2014; Sauzéon et al., 2016; Lecouvey et al., 2019). Six studies utilized HMDs as their selected hardware (Ouellet et al., 2018; Corriveau Lecavalier et al., 2020; Bruni et al., 2022; Da Costa et al., 2022; Gottlieb et al., 2022; Pieri et al., 2022).

Nine studies utilized computers as the primary hardware for their VR tasks (Cushman et al., 2008; Morganti et al., 2013; Lee et al., 2014; Kempe et al., 2015; Serino et al., 2015; Mohammadi et al., 2018; Zhang et al., 2021; Hogan et al., 2023; Rekers and Finke, 2023).

Four studies employed tablets (Zygouris et al., 2015; Eraslan Boz et al., 2020; Limoncu et al., 2021; Turner et al., 2021).

3.4 Computer-generated versus real scenarios

Out of the studies currently under examination, only three have utilized authentic scenarios as part of their research methodology (Cushman et al., 2008; Bruni et al., 2022; Pieri et al., 2022).

4 Discussion

The findings of this systematic review highlight a significant observation in the field of VR tools used for assessing memory. The primary focus of this review is the comprehensive collection of evidence that targeted three specific area of memory: episodic, prospective, and spatial memory.

The evaluation of episodic memory in VR settings has received significant attention and has become a subject of research interest. The emphasis on this aspect is not unexpected, given that VR technologies offer a distinctive medium for engaging participants in immersive and varied contexts, which can effectively assess their capacity to remember particular events, locations, and experiences. Scholars have utilized the immersive characteristics of VR in order to develop comprehensive episodic memory evaluations that closely resemble authentic real-world scenarios (Smith, 2019). Indeed, the studies being examined have purposefully endeavored to anchor their memory evaluation tasks within ecologically significant contexts. For example, certain studies have conducted their evaluations in familiar settings, including residential dwellings (Arvind Pala et al., 2014; Sauzéon et al., 2016), shops (Ouellet et al., 2018; Corriveau Lecavalier et al., 2020), and even simulated moving scenarios (Bruni et al., 2022). The utilization of these ecological settings has afforded participants with scenarios that closely approximate real-world circumstances. This exercise aimed to replicate a real-life memory challenge that individuals frequently face in their day-to-day experiences. The supermarket has been a commonly utilized ecological setting in numerous studies (Zygouris et al., 2015; Kempe et al., 2015; Eraslan Boz et al., 2020; Limoncu et al., 2021; Turner et al., 2021). The utilization of this particular context corresponds to the everyday encounters of numerous individuals and provides a practical framework for evaluating episodic memory in diverse shopping-related situations. Plancher et al. (2010) adopted an alternative methodology wherein participants were exposed to a simulated car accident scenario, followed by an assessment that tested their ability to recollect specific details and events associated with the high-stress situation. In addition, the researchers Gottlieb et al. (2022) conducted a virtual replication of the widely recognized Rey Auditory Verbal Learning Test (RAVLT). This study highlights the capacity of VR to imitate conventional neuropsychological memory assessments in settings that are both more regulated and closely aligned with real-world conditions.

The investigation of prospective memory, which pertains to the ability to recall and perform intended actions or tasks at a later time, has gained considerable attention in the realm of VR memory assessments. The inherent flexibility of VR environments enables the effective incorporation of prospective memory tasks within immersive settings. In such scenarios, individuals are tasked with the recollection and execution of intentions while actively engaging with and exploring the virtual realm (Lecouvey et al., 2019; Hogan et al., 2023). The emphasis on prospective memory highlights the capacity of VR tools to replicate and evaluate memory difficulties encountered in real-life scenarios.

The assessment of spatial memory has also been a significant area of focus in VR studies. The ability of virtual reality to accurately reproduce and manipulate spatial environments has facilitated the development of tasks that require individuals to navigate and remember intricate spatial layouts, which are integral to cognitive processes. The utilization of VR by researchers has enabled the immersion of individuals in various spatial contexts, thereby creating an advantageous environment for the examination of spatial memory processes.

Although the focus on these three memory domains is evident, it is crucial to acknowledge that the versatility of VR presents opportunities for evaluating a broader spectrum of memory functions. Potential future research endeavors may involve the investigation of additional subtypes of memory or the extension of the application of VR in the evaluation of less commonly examined memory domains such as semantic and autobiographical. The aforementioned findings indicate the necessity for a more extensive investigation into the various applications of VR in the assessment of memory. This exploration should take into account the potential contributions that this technology can make to the broader field of cognitive sciences and its practical implications in clinical practice.

Furthermore, another aim of this research was to evaluate the extent of convergence and divergence that can be observed when comparing memory assessments based on VR with conventional neuropsychological tests. Our analysis revealed a prominent pattern, indicating significant similarity between VR memory assessments and conventional neuropsychological tests. Specifically, VR memory tasks demonstrated notable and robust associations with memory assessments conducted using traditional methods. This implies that VR tasks have demonstrated the ability to capture memory processes related to episodic, prospective, and spatial memory, which are consistent with established neuropsychological assessments. The performance of participants in VR environments exhibited a strong correlation with their performance on established episodic memory assessments, highlighting the alignment in the memory processes being evaluated. In particular, regarding episodic memory, the results highlight the strong convergent validity of these virtual tools in the evaluation of episodic memory.

In the realm of prospective memory, the results of the study of Hogan et al. (2023) indicate that VR tasks are successful in capturing and reflecting the PM abilities that are evaluated by conventional measures. This finding provides evidence in favor of the validity of VR tasks as a means of assessing PM. Moreover, the study unveiled significant correlations between time-based prospective memory tasks conducted in VRPMST and the corresponding time- and event-based PM tasks conducted in laboratory-based dual-task paradigms. Furthermore, a significant and strong association was observed between both time and event-based PM as evaluated using VRPMST and the outcomes obtained from the Trail Making Test (TMT), the Hopkins Verbal Learning Test-Revised (HVLT-R), and the Montreal Cognitive Assessment (MoCA). The aforementioned correlations provide additional evidence supporting the notion that VRPMST encompasses both time-based and event-based aspects of prospective memory. These aspects have been found to have significant associations with established neuropsychological tests, thus highlighting the versatility and reliability of VRPMST as a tool for assessing PM.

The second study (Lecouvey et al., 2019) examined individuals diagnosed with AD and explored the correlation between prospective memory intentions and different cognitive functions. The results of this study provided valuable insights into this relationship: certain elements of intentions related to prospective memory are linked to different cognitive domains. The relevance of the correlation with TMT Part B is noteworthy due to its frequent utilization as an assessment tool for cognitive flexibility. Moreover, the observed association between TMT Part B and PM in individuals with AD underscores the interconnectedness between prospective memory and executive functions within this specific population.

The existing body of research investigating the use of virtual reality tests for spatial memory consistently demonstrates positive associations with conventional tests that evaluate memory and visuospatial capabilities. The consistent and strong correlation observed in these findings highlights the concurrent validity of the virtual reality tests in evaluating spatial memory and other cognitive functions that are interconnected.

The research conducted by Mohammadi et al. (2018) revealed noteworthy associations between their spatial memory task and established memory assessments, such as Rey immediate recall and delayed recall. This implies that the virtual reality task employed successfully captures and evaluates aspects of episodic memory, highlighting its efficacy as a dependable instrument for assessing memory functions within a realistic spatial framework.

Significant correlations were observed between the VR-MT and VR-MTM tests and Corsi’s supra-span task, which is a commonly employed assessment tool for evaluating visuospatial memory and navigation abilities. The results of this study highlight the correlation between the virtual reality assessments and well-established visuospatial tasks, thereby strengthening their concurrent validity.

There was a notable association observed between the Vienna task and visuospatial short-term memory. Nevertheless, the present task failed to exhibit a substantial correlation with episodic memory, as assessed by the proportion of elements recalled during the delayed free recall phase of the Rey-Osterrieth Complex Figure Test (ROCFT). The particularity of this task underscores its capacity to evaluate specific visuospatial memory domains, without necessarily encompassing episodic memory.

In contrast, the study conducted by Mohammadi et al. (2018) found a correlation between the ROCFT and simplified Rey figure test (SRFT) with the virtual neighborhood task. Similarly, Lee et al. (2014) observed a correlation between ROCFT and SRFT with the virtual radial arm maze (VRAM). Additionally, Tippett et al. (2009) proposed a task that also exhibited a correlation with ROCFT and SRFT. These observations imply that the conventional visuospatial memory tests exhibit similar characteristics to their virtual reality counterparts, thus indicating their concurrent validity in evaluating spatial memory.

In conclusion, the findings presented in this study underscore the strong concurrent validity of virtual reality assessments in measuring spatial memory. These assessments consistently demonstrate significant correlations with conventional measures used to evaluate memory, visuospatial skills, executive functions, and overall cognitive functioning. This highlights the potential of utilizing virtual spatial memory assessments as valuable instruments for evaluating diverse cognitive functions within authentic spatial contexts.

Nevertheless, our analysis also uncovered fascinating occurrences of disparity between VR based memory evaluations and traditional neuropsychological examinations. The divergence that stood out the most was observed in the study conducted by Plancher et al., where a virtual task was found to lack correlations with conventional memory assessments. This discovery implies that the particular virtual reality task in question may engage distinct memory components or cognitive mechanisms that are not fully assessed by traditional assessments. Additionally, it is important to highlight that certain novel VR tools demonstrated associations with functions that extend beyond memory. Several studies have demonstrated associations between VR tasks and cognitive functions such as memory and executive functions. Notable examples include the works of Arvind Pala et al. (2014), Zygouris et al. (2015), Sauzéon et al. (2016), Ouellet et al. (2018), Eraslan Boz et al. (2020), Limoncu et al. (2021), Bruni et al. (2022), and Gottlieb et al. (2022). This implies that the utilization of VR memory tasks can be advantageous in the simultaneous evaluation and differentiation of memory and executive function elements, thereby offering a more comprehensive understanding of cognitive profiles.

In addition, a number of the VR assessments that were examined displayed significant correlations with overall cognitive functioning, suggesting that they have the potential to be useful instruments for evaluating an individual’s overall cognitive status.

The findings of this study highlight the potential of VR based memory assessments as comprehensive instruments for evaluating the overall cognitive wellbeing of individuals.

The findings of our review are noteworthy as they establish a connection between VR memory assessments and executive functions, as well as general cognitive functions. These correlations serve as compelling evidence for the effectiveness of the functional approach utilized in these assessments. The aforementioned result highlights the potential of memory tasks based on VR to provide a comprehensive and integrated understanding of cognitive function, in contrast to the conventional approach driven by specific constructs.

The correlations observed concerning executive functions indicate that virtual reality memory assessments effectively engage memory processes in a manner that is inherently interconnected with other cognitive domains. This integration is consistent with the principles of functional assessment, which posits that cognitive functions are not discrete constructs but rather function in conjunction with one another within the larger cognitive system. Virtual reality has emerged as a promising method for evaluating the relationship between memory and executive functions. By utilizing VR-based memory tasks, researchers can gain a more comprehensive and realistic understanding of cognitive performance, thereby enhancing the ecological validity of their assessments.

These correlations highlight the notion that VR memory assessments do not adhere to a strict construct-driven methodology, but instead adhere to a functional approach that aims to measure cognitive processes as they occur organically within intricate, real-world situations. The functional perspective is highly appropriate for comprehensively understanding the complex interaction among memory, executive functions, and overall cognitive performance, thereby offering a more comprehensive assessment of cognitive wellbeing. Therefore, the utilization of VR as a tool for evaluating memory holds great potential for advancing cognitive research and clinical applications, as it allows for a more comprehensive and realistic understanding of cognitive abilities.

Another aspect explored in this review pertained to the discernment of whether the settings employed for memory evaluations were computer-generated or obtained through the utilization of 360-degree technologies. The objective of this investigation was to analyze the technological underpinnings of the virtual scenarios utilized in the research, providing insight into the adaptability and benefits associated with various methods of constructing immersive environments for cognitive evaluations. Out of the studies currently under examination, only three have utilized authentic scenarios as part of their research methodology (Cushman et al., 2008; Bruni et al., 2022; Pieri et al., 2022). Overall, the decision between computer-generated virtual environments and real scenarios captured with 360° technologies in memory assessment tasks necessitates a compromise between the level of control exerted over the environment and the degree of realism achieved. The selection of a particular approach should be in accordance with the research objectives, taking into account the inherent trade-offs between these crucial dimensions. In situations where ecological validity is of utmost importance, the utilization of authentic scenarios may be more favorable. However, for assessments that require strict control and standardization, computer-generated environments present distinct advantages in terms of customization and reproducibility. Researchers should carefully consider these factors in relation to their research objectives, the specific population being studied, and the resources at their disposal.

One noteworthy constraint of this systematic review is the limited accessibility of the virtual reality memory assessments utilized in the studies under review. The review sought to comprehensively synthesize the extant literature; however, it was noted that only one study made their VR tasks openly accessible and downloadable (Rekers and Finke, 2023). The restricted accessibility of these studies poses a challenge to the wider scientific community in terms of replicating, validating, and expanding upon their findings. This emphasizes the importance of promoting transparency and adopting open science practices within the realm of VR-based memory assessment. Furthermore, the studies included in the analysis demonstrated significant variability in the VR hardware and software utilized for evaluating memory performance. The range of variability encompassed various technologies, including video projectors, head-mounted displays, computer interactions, and tablets. The presence of diverse hardware and software options in VR can pose challenges when attempting to compare and integrate research findings. This is because the selection of specific hardware and software can have a substantial influence on the user’s experience and the outcomes of VR tasks. The presence of heterogeneity underscores the importance of implementing standardization within the discipline.

Author contributions

VM: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. ES: Writing – review & editing. FB: Writing – review & editing. SA: Writing – review & editing. SD: Writing – review & editing. MC: Writing – review & editing. PC: Methodology, Writing – review & editing. EP: Conceptualization, Writing – review & editing.

Funding

The authors declare that financial support was received for the research, authorship, and/or publication of this article. This work was supported by Italian Ministry of Health–PRIN 2022R8P3TZ.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Allain, P., Foloppe, D. A., Besnard, J., Yamaguchi, T., Etcharry-Bouyx, F., Le Gall, D., et al. (2014). Detecting everyday action deficits in Alzheimer’s disease using a nonimmersive virtual reality kitchen. J. Int. Neuropsychol. Soc. 20, 468–477. doi: 10.1017/S1355617714000344

PubMed Abstract | Crossref Full Text | Google Scholar

Allen, T. A., and Fortin, N. J. (2013). The evolution of episodic memory. Proc. Natl. Acad. Sci. U. S. A. 110, 10379–10386. doi: 10.1073/PNAS.1301199110

PubMed Abstract | Crossref Full Text | Google Scholar

Armstrong, C. M., Reger, G. M., Edwards, J., Rizzo, A. A., Courtney, C. G., and Parsons, T. D. (2013). Validity of the virtual reality stroop task (VRST) in active duty military. J. Clin. Exp. Neuropsychol. 35, 113–123.