Characteristics of included studies.
| Author(s), publication year; country | Sample | Participants (n); mean age and SD | Study design | HMD technology | Cognitive training | Training duration (days × weeks × minutes) | Outcome measure | Main results | Notes |
|---|---|---|---|---|---|---|---|---|---|
| Park et al. [38], 2020; Republic of Korea | Amnestic MCI | Fully IVR (n = 10); 71.80 ± 6.61Control group (n = 11); 69.45 ± 7.45 | RCT | HTC Vive, VR controllers | Memory, executive function, processing speed, attention | 2 d × 12 w × 30 min | K-MMSEKorean Colour-Word Stroop TestSGDS-K | Within-group analysis did not indicate any significant change after 12 weeks. | Control group participants continued their normal activities.No clear theoretical framework or model was reported as the rationale for the intervention. The authors proposed some explanations for the non-significant results: 1) The relatively short intervention duration compared to studies with longer training periods (e.g., 6 months), 2) the intervention did not include physical exercise components, which the authors suggest may be critical for dementia prevention.Reported methodological limitation: small sample size. |
| Zhu et al. [39], 2022; China | MCI & MD | MCI (n = 18); 82.94 ± 6.44MD (n = 13); 85.76 ± 4.67 | Pretest-posttest study design | HTC Vive Pro Eye, VR controllers | Memory, attention, executive function, calculation ability | 3 d × 5 w × 20–30 min | MMSE & MoCAAVLTSTTSDMT | Within-group analysis indicated significant change after 5 weeks for both MCI and MD groups. | No control group was included in the study.No specific theory was used to explain the reported cognitive outcomes.However, the neural mismatch model was referenced to explain participants’ discomfort experienced in the fully IVR environment. The authors suggested that increasing exposure time during the session may help reduce discomfort and improve user adaptation.Reported methodological limitation: small sample size, lack of active comparison groups. |
| Zając-Lamparska et al. [40], 2019; Poland | MD | MD (n = 27); 72.04 ± 7.43Healthy adults (n = 72); 67.86 ± 5.83 | Pretest-posttest study design | Oculus Rift DK2, Xbox 6DOF control pad | Attention, memory, language, and visuospatial processing | 2 d × 4 w × 45 min | AttentionMemory, languageVisuospatial processing | Within-group analysis indicated some significant change after 4 weeks of fully IVR intervention for the MD group. There were some non-significant cognitive measures. | No control group was included for this fully IVR intervention. The cognitive training games used in the study were not described in full detail; however, a brief description of GRADYS training software is provided in Zając-Lamparska et al. [41]. No specific theoretical framework was used to explain the reported results.Reported methodological limitation: lack of an active comparison group. |
| Kang et al. [42], 2021; Republic of Korea | Predementia {MCI or Subject cognitive decline}* | Total (n = 41); 74.51 ± 5.81 | RCT | Oculus Rift CV1 display,Oculus Touch controllers | Attention, executive function, and memory | 2 d × 4 w × 20–30 min | MMSEAttentionLanguage abilityMemoryFrontal executive functionMRI | Within-group analysis indicated a significant change in verbal memory, language, and related functions, and TMT-B for the fully IVR group after 4 weeks of intervention. | A control group was included with participants continuing “business as usual”.The cognitive reserve model was used to provide the rationale for the study. The Van Sommers model was applied to explain improvements observed in the RCFT copy test, based on proposed links between visuospatial construction demands in the fully IVR environments and frontal-occipital functional connectivity.Reported methodological limitation: small sample size, lack of an active comparison group. |
| Maeng et al. [43], 2021; Republic of Korea | MCI | MCI (n = 31); 73.2 ± 7.3Healthy adults (n = 25); 71.6 ± 4.4 | Not reported | Samsung Odyssey HMD, motion controllers | Memory, attention, executive function | 8 sessions × 4 w × 50–60 min | CERAD-K (each cognitive domain) | Within-group analysis indicated significant change in some CERAD-K outcomes (Word list memory, word list recall, word list recognition, constructional recall, TMT-A, Stroop-C, Stroop-CW) after 4 weeks of intervention for the MCI group. | No control group was included in this study.There was no specific theoretical framework or model that was used. However, the authors examined the sense of presence, hypothesising that it would differ from that of cognitively normal individuals; however, no significant difference was found.Reported methodological limitation: small sample size, no follow-up tests. |
| Yang et al. [44], 2022; Republic of Korea | MCI* | Fully IVR (n = 33); 72.5 ± 5Exercise (n = 33); 67.9 ± 3.6Control (n = 33); 72.6 ± 5.6 | RCT | Oculus Quest headset, VR controllers | Attention, working memorySourced from Thapa et al. [30] (2020) | 3 d × 8 w × 100 min | MMSE TMT-ASDSTNCGG-FAT | Within-group analysis indicated a significant change in global cognitive function (MMSE, TMT-A, and SDST) after 8 weeks of training intervention. | Three comparison groups were included: exercise intervention group, a control group receiving 30-min educational seminars, and a fully IVR group.No specific theory was used to provide a rationale for the study or to explain the cognitive outcomes. However, the authors’ reference to the sense of presence to interpret observed neural-level changes in the brain.Reported methodological limitation: inconsistent intervention frequency, female-dominant sample. |
| Thapa et al. [30], 2020; Republic of Korea | MCI | Fully IVR (n = 34); 72.6 ± 5.4Control (n = 34); 72.7 ± 5.6 | RCT | Oculus Quest headset, VR controllers | Memory | 3 d × 8 w × 100 min | MMSE-DSNCGG-FATTMT-ATMT-BSDSTEEG recording | Within-group analysis indicated a significant change in global cognitive function (TMT-A, TMT-B, SDST) after 8 weeks of training intervention. | A control group was included; participants completed eight sessions of an educational program or general healthcare, delivered once per week, with each session lasting 30–50 min. The intervention group also received educational content. No specific theoretical framework or model was provided for the rationale or to explain the significant findings.Reported methodological limitation: no follow-up test during the intervention. |
| White and Moussavi [45], 2016; Canada | Individual with the onset stage of AD | 74-year-old male | Case study | Oculus Rift DK VRN, wheelchair motion sensors | Spatial navigation | 3 d × 7 w × 45 min | MoCA | The participant’s MoCA score increased from 24 to 26 after 7 weeks of training intervention. | The cognitive reserve model was used as the theoretical rationale for the study. The Morris water task was used as a model for IVR-based spatial navigation assessment. The authors suggested that including periodic “booster” sessions to maintain and reinforce any potential improvements in cognitive reserve. |
*: Studies that conducted a power analysis. AVLT: Auditory Verbal Learning Test; CERAD: Consortium to Establish a Registry for Alzheimer’s Disease; EEG: electroencephalogram; HMD: head-mounted display; IVR: immersive virtual reality; MCI: mild cognitive impairment; MD: mild dementia; MMSE: Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; NCGG-FAT: National Center for Geriatrics and Gerontology Functional Assessment Tool; RCFT: Rey-Osterrieth Complex Figure Test; RCT: randomized controlled trial; SD: standard deviation; SDMT: Symbol Digit Modalities Test; SDST: Symbol Digit Substitution Test; SGDS-K: Short Form Geriatric Depression Scale-Korean version; STT: Shape Trial Test; TMT: Trial-Making Test.