Table 1.

Study characteristics.

Author(s) and Year	Sample	Assessment	Task	Cognitive Function	Technology	Multisensory	Primary Outcome
Akinwuntan et al. 2010 [47]	EG: 33 CG: 36 Side of lesion: both Time post-stroke: not specified	UFOV test pretraining at 6 to 9 weeks, and follow-up at 3 months and 6 months post-stroke EG: 15 h of training over 5 weeks at 1 h per session, 3 sessions per week with following subtests: -divided attention -selective attention -speed of processing	The participants in the experimental group received specific training to improve their visual attention and speed of processing skills. The task was composed of 10 scenarios.	Visual attention skills	STISIM Drive System (version 1.03; Systems Technology Inc, Hawthorne, CA, USA)	Three senses: -sight -hearing -touch	Results revealed neither group effects nor significant interaction effects of group with time in the UFOV total score and the 3 subtests.
De Luca et al. 2018 [48]	EG: 6 CG: 6 Side of lesion: not specified Time post-stroke: 3–6 month	Pre–post intervention and follow-up (1 month). Each treatment session (24) lasted 45 min and was repeated three times a week for 8 weeks. Subtests: -MoCA -FIM -In addition: -FAB -Attentive Matrices -Trial Making Test (TMTA; TMTB, and TMTB-A) -TCT and MI	The device created a special setting called a sensory room. The rehabilitation was composed of exercises with audiovisual stimuli and feedback involving the perceptual-cognitive skills of patients, resulting in a motivational training.	Attention, executive, and spatial-visual functions; verbal and spatial–visual memory.	BTS-Nirvana (Garbagnate Milanese, Italy) using an interactive semi-immersive program (I-SIP)	Three senses: -sight -hearing -proprioception	MoCA and MA improved significantly only in the EG at post-treatment. Particularly in visuospatial attention domains, MI and TCT improved more in the EG than the CG did at T1.
Faria et al. 2016 [49]	EG: 9 CG: 9 Side of lesion: both Time post-stroke (months): EG, 7 (4–49); CG, 4 (3–11.5)	All patients were evaluated pre and post treatment 4 to 6 weeks; 12 interventions, of 20 min each session, distributed from 4 to 6 weeks. Subtests: -ACE -Trail Making Test A and B -Test picture arrangement to WAIS III -Stroke Impact Scale (SIS 3.0)	See Faria 2020 [50]	Global functioning, memory, attention, executive functions.	Reh@City v2.0	Three senses: -sight -hearing -touch	In EG, there were significantly greater improvements in global cognitive functioning ACE and MMSE, attention, and executive functions than in conventional therapy.
Faria et al. 2020 [50]	EG: 17 CG: 19 Side of lesion: both Time post-stroke, at least 6 months (months): EG, 45.93; CG, 21.33	Pre and post intervention at 4–6 weeks, and follow-up at 2 months; 12 sessions of treatment. Subtests: -MoCA -Trail Making Test A and B -Verbal Paired Associates to WMS-III -Digital span forward and backward -Digital symbol coding -Wechsler Adult Intelligence Scale III (WAIS)	Reh@city is a virtual reality simulation of a city where patients are required to solve cognitive tasks through familiar activities of daily living (ADL) in a variety of common places: buying food in a supermarket, picking up a package in the post office, paying the electricity at the bank ATM, etc. These places display billboards and real products of actual spaces and trademarks to help the patient relate the VR task to the real world. In addition, patients are also required to use their paretic arm to solve the tasks.	Global functioning, memory, attention, executive functions, and language.	Reh@City v2.0-based simulation of ADLs; customized handle with a tracking pattern on the surface of the table	Three senses: -sight -hearing -touch	In EG, there were significantly greater improvements in general cognitive functioning, visuospatial ability, executive functions, and attention on the MoCA.
Kang et al. 2009 [51]	EC: 8 CG: 8 Side of lesion: right Time post-stroke (days): EG, 64.3; CG, 58.1 (29.9)	Pretraining and post-training at 4 weeks; a total of 12 sessions (3 per week) for 30 min per session for 4 weeks. Subtests: -MMSE -Motor-Free Visual Perception Test -K-MBI	The program consisted of 12 tasks in four parts designed to improve visual perceptual function: -Task of visual reaction -Task of visual differential -Task of visual tracking and targeting -Task of visual–spatial cognition and moto functions	Global functioning, visual perception.	PSS CogRehab program (Indianapolis, IN, USA); motion-tracking technology based on the CAMSHIFT	Three senses: -sight -hearing -proprioception	The mean motor-free visual perception test score increased significantly in both EG and CG. The Barthel index score increased significantly in both groups, with the EG recording a higher increase. The mean interest scale score was greater in the EG than in the CG.
Kannan et al. 2019 [52]	EG: 10 CG: 10 Side of lesion: both Time post-stroke: greater than 6 months ago	At pre–post test and follow-up at 11 weeks; 6 weeks of training, 10 sessions, 90 min for each session. Subtests: -MoCA -WRAT-4 -BBS -TUG -6MWT -ABC -SPT + LNS -LOS + LNS -B + Cog-Balance and Cognitive task	Participants played Wii Fit games in conjunction with performing cognitive tasks. Each session was divided into three sub-sessions, with each sub-session comprising four Wii Fit games: Bubble; Table Tilt; Tight Rope Walking; Soccer Heading. The cognitive games were played in conjunction with each of the balance games and these have trained the semantic memory, abstract memory, working memory, attention, and verbal fluency.	Attention, executive functions, WM.	Wii Fit (Nintendo Co, Ltd. Kyoto, Japan), balance board	Three senses: -sight -hearing -proprioception	The EG showed a significant improvement in both motor and cognitive functions in particular areas: -movement velocity while performing the limits of stability (LOS) test under dual-task conditions. -COM state stability and cognitive ability during the dual-task slip-perturbation test. -cognitive function that was improved during dual-task performances under slip-perturbation test. -In the CG, only the motor functions were improved.
Kim et al. 2011 [53]	EG: 12 CG: 12 Side of lesion: right MCA infarction Time post-stroke (days): EG, 25.5; CG, 22.8	Pre-training, post-training; therapy for 30 min a day, 5 days per week for 3 weeks. Subtests: -Star cancellation test -Line bisection test -CBS -K-MBI	The VR training was composed of three programs at a time: “bird and ball”; “coconut”; “container”. In the three programs, patients were told to use nonparetic right hands against the left simulation.	Visual–spatial attention.	IREX system® (Toronto, Ont., Canada) is a VR system consisting of a monitor, a video camera, computer-recognizing gloves, and virtual objects.	Three senses: -sight -haring -proprioception	The EG showed improvements in attention, spatial awareness, and generalized cognitive functioning. No significant change was seen in the executive function and memory domain. For the CG, no significant change over time was found. The EG displayed a lower level of depression than the CG after treatment.
Maier et al. 2020 [54]	EG: 16 CG: 14 Side of lesion: not specified Time post-stroke: more than 6 months after stroke but less than 10 years	At baseline, post-training at 6 weeks and follow-up at 18 weeks; daily training for 6 weeks (30 min each session). Subtests: -ASCS -Corsi f -TMT A -WAIS -Corsi B -MoCA -MMSE -BI -FM-UE	Different VR scenarios for the training of spatial attention and working memory. The patient controls a virtual avatar on a computer screen, with conjunctive cognitive training scenarios which ACCT. The patients are seated at a table, and the three training scenarios are shown on the screen always in the same order. The complex spheroids The star constellations The quality controller	Attention, executive functions, spatial awareness, memory.	VR rehabilitation tool RGS with desktop computer, Microsoft Kinect, two wristbands with reflective markers (worn by the patient), and a Tobii Eye tracker T120.	Three senses: -sight -hearing -touch	The EG showed improvements in attention, spatial awareness, and generalized cognitive functioning. No significant change was seen in the executive function and memory domain. For the CG, no significant change over time was found. The EG displayed a lower level of depression than the CG after treatment.
Rogers et al. 2019 [55]	EG: 10 CG:11 Side of lesion: both Time post-stroke (days): EG, 22.8; CG, 30	At pre–post-treatment at 3 weeks and follow-up at 1 month. EG: 4 weeks of Elements virtual rehabilitation (three weekly 30–40 min sessions) combined with treatment as usual. EG and CG both received 3 h of daily conventional occupational and physiotherapy. Subtests: -BBT -GMLT -SST -NFI	Elements: a series of unimanual goal-directed tasks and audiovisual exploration tasks focus on speed and accuracy to promote motor learning and cognitive control. In particular, the rehabilitation programs consist of 7 tasks, increasingly structured, where the participants have to use four handheld objects (the four “elements” in the shape of a circle, pentagon, triangle, and rectangle) and are engaged with a virtual environment.	Global functioning, executive functions.	Elements The system consists of a large tabletop surface display (42 in), tangible user interfaces, and software for presenting both goal-directed and exploratory virtual environments.	Three senses: -sight -hearing -touch	Both groups showed significant training-related improvement in motor and cognitive functions and functional status. However, the magnitude of effect sizes was substantially larger for the EG compared with the CG. In particular, in pre–post difference scores, the EG showed significantly greater improvement. In moto function (BBT) in the hand most affected by their stroke and on all measures of cognitive function compared with the CG. Furthermore, the improvement shown by the EG as a function of training was maintained at the 1-month follow-up assessment; on difference scores between pre-test and follow-up, the EG performed significantly better than controls on motor, cognitive and functional outcomes.
Yip et al. 2013 [56]	EG: 19 CG: 18 Side of lesion: not specified Time post-stroke (days), at least 3 months post injury: EG, 145.13; CG, 167.53	Pre–post-intervention at 3 weeks and follow-up at 1 month. The program, 12 sessions, was run twice a week and each session lasted about 30 to 45 min. Subtests: -MMSE-CV -TONI-3 -ADI-CV -Virtual reality-based test of everyday -Prospective memory task -Behavior checklist of a prospective memory task in a real environment -CAMPROMT-CV -HKLLT -FAB -WFT-CV -CTT -CIQ-CV -Self-efficacy questionnaire in performing an everyday prospective memory task	Three training components were investigated: Prospective memory: the PM training consisted of event-based tasks (such as shopping for drinks with discounted prices or calling back home when a gift redemption counter was seen), time-based tasks (such as taking the food out of the microwave oven after 5 min), and ongoing tasks (shopping for items on a shopping list that had been given). Retrospective memory: in the RM training component, subjects were required to remember a list of four shopping items. Inhibition training component: participants were required to press the spacebar whenever they saw an item with a special price tag shown on the screen.	Perspective memory, retrospective memory.	VRPM: A virtual reality-based prospective memory training program was developed using a non-immersive form of VR. The 3D layout in the program was designed and built by Maya 8.0 Software called Virtools.	Three senses: -sight -hearing -touch	The results suggest that significantly better changes were seen in both EG and CG PM outcome measures, related to cognitive attributes such as frontal lobe functions and semantic fluency. In particular, the EG showed significant improvements in most of the test items: immediate recall of PM tasks, the performance of both event-based and time-based PM tasks performance of ongoing tasks, and number of time checks; in CG, the PM test showed significant improvement in event-based and time-based tasks but not in ongoing tasks.

EG: experimental group; CG: control group; UFOV: useful field of view; MoCA: Montreal Cognitive Assessment; FIM: functional independence measure; FAM: functional assessment measure; FAB: frontal assessment battery; VR: virtual reality; TMT-A: Trail Making Test part A; TMT-B: Trail Making Test part B; TCT: traditional cognitive training; MI: motricity index; ACE: Addenbrooke’s cognitive examination; WAIS-III: Wechsler Adult Intelligence Scale—third edition; WAIS: Wechsler Adult Intelligence Scale; WMS -III: Wechsler Memory Scale—third edition; ADL: activities of daily living; MMSE: Mini-Mental State Examination; K-MBI: Korean version of the modified Barthel index; BBS: Berg balance scale; TUG: timed up and go test; 6MWT: 6-meter walk test; ABC—activity-specific balance confidence; SPT: slip-perturbation test; LNS: letter–number sequencing; COM: center of mass; WRAT4: Wide Range Achievement Test 4; LOS: limits of stability test; WAIS: Wechsler Adult Intelligence Scale—fourth edition; Corsi B: Corsi Block Tapping Test Backward; ACCT: adaptive conjunctive cognitive training; BI: Barthel index; FM-UE: Fugl–Meyer assessment for the upper extremity; ASCS: averaged standardized composite scores; Corsi F: Corsi Block Tapping Test Forward; RGS: rehabilitation gaming system; RAVLT I: Rey Auditory Verbal Learning Test Immediate; BBT: box and blocks task; GMLT: Groton maze learning task; NFI: Neurobehavioral Functioning Inventory; SST: set shift task; TONI-3: Test of Nonverbal Intelligence—third edition; MMSE-CV: Chinese version Mini-Mental State Examination; VRPM: virtual reality-based prospective memory; CAMPROMT—CV: Cambridge Prospective Memory Test—Chinese Version; HKLLT: Hong Kong list learning test; WFT—CV: word fluency test—Chinese version; CTT: color trails test; CIQ–CV: Chinese version of the Community Integration Questionnaire.