Review protocol

The review is reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) extension for Scoping Reviews (PRISMA-ScR; Tricco et al. [34]). Prior to the initial searches and screening, the review protocol was pre-registered with the Open Science Framework (https://doi.org/10.17605/OSF.IO/8ERVH).

Research question

Scoping review: To what extent… is fully IVR cognitive training effective for individuals living with MCI or dementia, in relation to global cognition and specific cognitive domain outcomes.

Identifying relevant studies

The eligibility for our study was guided by the population, concept, and context (PCC) framework. The population included individuals living with predementia or dementia, the concept focused on fully IVR cognitive training, and the context was cognitive functioning. Eligible studies had to meet the following criteria: 1) empirical research; 2) published in English in peer-reviewed journals; and 3) included participants diagnosed with predementia or dementia (refer to Table 1). The population of interest consisted of individuals with cognitive decline, specifically those diagnosed with MCI or dementia, with clear diagnostic criteria such as Petersen’s guidelines [35]. The primary concept of interest was fully IVR, defined by the use of head-mounted displays (e.g., Oculus Quest Rift). Studies combining other interventions (e.g., physical exercise, diet) were included if at least one intervention group used only IVR. Cognitive training was defined as repeated, guided practice of standardized tasks aimed at improving specific cognitive domains [36]. The context involved fully IVR cognitive training targeting multiple or specific cognitive domains and its effect on cognitive function. Outcome measures were required to assess cognitive function, including near or far transfer effects. Neurological measures related to cognitive domains of interest [e.g., electroencephalogram (EEG)] were also included. However, studies focusing solely on non-cognitive outcomes, such as activities of daily living, were excluded, as our study focuses on the impact of cognitive training on cognitive outcomes.

Study selection

The search strategy was developed through discussions with the research team and guidance from an academic librarian. On June 20, 2024, the first author conducted searches across four databases: PubMed (Medline), Web of Science, PsycINFO, and PsycARTICLES. Additional searches were conducted using Taylor and Francis Online and Google Scholar. The search terms were designed to capture the population (MCI or dementia), concept (fully IVR cognitive training), and context (cognitive functioning), and were appropriately adapted for each database. In addition to database searches, relevant literature was identified through manual searches of reference lists, forward citation tracking of included studies, and existing reviews. Duplicates were automatically removed. The first and last authors independently screened titles and abstracts using Rayyan. They met on July 8, 2024, to review outcomes and resolve any conflicts from the initial screening. The full texts of the remaining studies were then independently reviewed by the same authors. On August 21, 2024, the authors met again to discuss their findings, resolve any conflicts by discussing their rationale, and confirm whether the studies met the eligibility criteria for final inclusion.

Following McHugh’s guidelines [37] for interpreting kappa values, the agreement between the first and last authors was appropriate. At the title and abstract screening stage, the kappa value was κ = 0.69, and at the full-text review stage, the agreement improved to κ = 0.84.

Data extraction

To address the research question, we created a data extraction table in Microsoft Word. This was designed to capture relevant information related to the study characteristics, cognitive training components, and main outcomes. The data extraction table included information on authors, publication year, county, study design, participant characteristics, head-mounted technology, cognitive training, training duration, outcome measures, main results, and comments on any additional information of interest. The first author independently extracted the data from the included articles, and this was subsequently reviewed and discussed with the last author to ensure accuracy and consistency in the data extraction process.

Literature search

The PRISMA flow diagram shows the selection process for the scoping review. The search strings identified 816 articles, 8 were included in the review (see Figure 1). All studies included participants who were diagnosed with MCI or dementia. Table 2 provides a breakdown of all the included studies.

Display full size

Figure 1. PRISMA flowchart for articles included in the scoping review. Adapted from [34]. © The Author(s) 2018. Licensed under a Creative Commons Attribution (CC BY 4.0). IVR: immersive virtual reality; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analysis. *: identified studies; **: excluded studies.

Virtual reality cognitive training intervention

Among the eight included studies, seven articles implemented multi-domain cognitive training intervention targeting memory, attention, executive functioning, and visuospatial processing. White and Moussavi [45] was the only study that implemented a single cognitive training intervention, focused on visuospatial processing, following a case study design. Across the eight studies, the duration of individual sessions ranged from 20 to 100 min, with sessions delivered either two or three times per week. The total training duration ranges from 4 to 12 weeks.

All the included studies provided inclusion criteria for individuals with MCI or dementia. Of the included studies, the samples included MCI (K = 4), amnesic MCI (K = 1), predementia combining MCI and cognitive decline sample (K = 1), mild dementia (K = 2), and onset stage Alzheimer’s disease (K = 1). Maeng et al. [43] and Zając-Lamparska et al. [40] compared the sample group diagnosed with MCI and mild dementia (respectively) with the healthy adult sample group.

Out of the included studies, changes in global cognition (K = 5), memory (K = 5), attention (K = 6), visuospatial (K = 4), executive function (K = 6), and language (K = 4) were measured pre-and post- virtual reality cognitive training intervention (see Table 3 for all the cognitive measures). There were no consistent significant within-subject results across the outcome measures. Significant within-subject effects on global cognition were found in 2 studies (K = 2/5), on language ability were found in one study (K = 1/4), on executive function were found in five studies (K = 5/6), on visuospatial were found in three studies (K = 3/4), on attention were found in three studies (K = 3/6), on memory were found in four studies (K = 4/5). Park et al. [38] was the only study that found a non-significant difference pre- and post-fully IVR intervention across all cognitive outcome measures. White and Moussavi’s case study [45] found a change in global cognition, visuospatial ability, and memory, which were assessed quantitatively through tests [e.g., Montreal Cognitive Assessment (MoCA)] and qualitatively from the participant’s wife’s feedback. Four out of the eight studies included a control/comparison group. Two studies compared the virtual reality cognitive training intervention with an active educational program control. While the other two studies used a control group with a passive “business as usual” approach.

Theoretical standing

Among the included studies, two (K = 2/8) explicitly referenced a cognitive theory or model. White and Moussavi [45] applied cognitive reserve theory and the Morris water task model to justify their focus on virtual cognitive training solely targeting spatial navigation. Their case study involving a patient with early-onset Alzheimer’s disease found improvement in the participant’s global cognition after the cognitive training, but the gains declined at a 28-week follow-up. They suggested that continued training sessions may help maintain cognitive reserve. Kang et al. [42] also highlighted cognitive reserve theory, proposing that continued cognitive practice could help restore cognitive reserve, which would improve overall cognition. Additionally, they introduced the Van Sommers model of drawing as a theoretical explanation for the observed improvements in participants’ performance on the Rey-Osterrieth Complex Figure Test copy task after the virtual cognitive training. While cognitive reserve was not explicitly measured in either study, Kang et al. [42] measured brain functional connectivity; findings showed increased connectivity within the visual network after training.

Considerations

While this review highlights the promise of fully IVR cognitive training for individuals with MCI or dementia, there are several broader considerations to highlight during any development and implementation. Beyond cognitive outcomes, there may be indirect benefits with fully IVR in terms of economic benefits. The financial burden of dementia is substantial and is expected to rise in the coming years [2]; interventions that delay the progression of severe cognitive decline and promote independence may offer significant cost savings at the individual and societal level [21, 50]. A well-designed fully IVR cognitive training game can be self-administered, dependent on the severity of the condition, without the need for direct supervision by healthcare professionals. Findings from the present review revealed variations in the development and design elements of fully IVR cognitive training intervention, such as differences in training duration (i.e., minutes and days per week), length of training (i.e., number of weeks), difficulty progression, in-game adaptivity, and integration of narrative storyline. Given these variations, establishing consensus on the optimal core elements necessary for an effective fully IVR cognitive training intervention would be valuable. Future research exploring fully IVR cognitive training from a mechanistic perspective, rooted in theoretical grounding, could provide strong foundations for developers and researchers. For instance, designing cognitive tasks that specifically align with cognitive strategy training, rather than restorative practice-based training, may be more beneficial for individuals with MCI. Moreover, the scalability of the technology can allow for delivery in home-based or residential settings, reducing the need for travel to specialist clinic visits. The present review does not suggest that fully IVR cognitive training should be implemented in isolation. Rather, it is important to recognise the potential benefits of cognitive interventions as part of a holistic approach to nonpharmaceutical support for individuals living with MCI or dementia. This is consistent with the ecological models of care, which emphasize the importance of addressing the individuals within their broader interpersonal, physical, and environmental context. In this broader perspective, dementia has been discussed as a lifestyle-related disorder by some authors [51, 52]. A relatively recent concept, “vitaction”, is described as “vitamins for the brain”, referring to lifestyle-related actions that may support brain health (e.g., socializing, learning, having personal relationships, etc.) [53]. The concept proposes that deficiencies in vitaction (e.g., lack of interest in learning) may contribute to neurodegenerative diseases (i.e., Alzheimer’s disease), and supplementing these deficiencies with relevant vitaction could support brain health and potentially reduce dementia risk [52]. In this context, engagement with novel technology (i.e., fully IVR) represents an important component of modern lifestyle and may contribute to brain health depending on how it is used and designed for people living with MCI or dementia. Fully IVR cognitive training should be researched both as a standalone cognitive intervention and in terms of how they complement established care strategies and broader lifestyle-related approaches. That said, conducting experimental research in this population presents inherent challenges, particularly due to the variability and complexity of individual care plans, which are rarely standardized (see Table 2 for some reported methodological limitations). Beyond the focus on within-subject comparison in the present review, establishing an active comparison group when assessing the effectiveness of fully IVR cognitive training would strengthen the evidence of integrating fully IVR into care plans. It is also essential to collect contextual information regarding participants’ engagement with other therapies or activities, for example, their diet or physical engagement. Among the studies included in this review, only the case study [45] collected qualitative data outside of just cognitive measures; the participant’s wife reported changes in driving performance and involvement with other activities.

The economic and contextual factors are important to consider in the future of fully IVR cognitive training. However, there are other factors that should be equally considered, such as data protection and ethics. Policymakers should begin to establish regulatory frameworks that ensure safe use of IVR technologies and support initiatives that promote digital literacy within the population. For instance, digital exclusion is still a concern, as individuals living with MCI or dementia in underserved communities may face significant challenges accessing the technology required for fully IVR cognitive interventions. Furthermore, it is essential that these cognitive training tools are co-designed with, rather than merely for, individuals living with MCI or dementia to ensure that their voices, preferences, and lived experiences are integrated into development [54–56].

Conclusions

The present scoping review provides a summary of the literature on fully IVR cognitive training for individuals living with MCI or dementia. Although the number of included studies was sparse, the findings highlight the potential of fully IVR to enhance specific cognitive domains, such as memory, executive functioning, and visuospatial abilities. Overall, these findings offer a valuable foundation for advancing research and practice in this field. Future research should aim to establish consensus on the optimal elements necessary for designing and developing fully IVR cognitive training for this population and to conduct experimental research examining the impact of specific key design elements. The present review contributes to the growing literature supporting the use of fully IVR as a promising tool for cognitive intervention. The authors aim to build upon these insights by developing a fully IVR cognitive training game adopting a co-design approach with individuals with lived experience. With continued development and evaluation, fully IVR may become a significant tool for maintaining or enhancing cognitive functioning and quality of life for individuals affected by MCI or dementia. The future of cognitive care must be guided not only by advanced technologies but also by compassion, ensuring that these novel interventions are co-designed, equitable, and grounded in the lived experience of those they intend to support.