Table 2.

Main characteristics of the included studies.

Author	Patient	Tools	Inclusion Criteria	Training	Intervention	Control Group	Assessment	Outcome
Upper limb
Mekbib et al., 2021 [67]	N = 23 EG 12; CG 11	Mirroring neuron VR Rehabilitation (MNVR-Rehab): HMD, two HTC Vive tracking stations, Leap Motion and ALIENWARE laptop	(1) first ischemic or hemorrhagic stroke with moderate to severe upper limb dysfunction (2) stroke within 3 months (3) age > 18 years (4) neither hearing nor vision deficits (5) MMSE > 16	2 h per day, 4 days a week for 2 weeks	60 min of virtual training (reach, grasp and release colored ball into a basket through MNVR-Rehab) plus 60 min of occupational training	Occupational therapy based on daily living activities, balance control, gait training, weight shift and upper limb functional training	Assessment at baseline and post-intervention (2 weeks). FM-UE; BI	MNVR-Rehab is an encouraging rehabilitation apparatus that may increase upper limb function in subacute stroke subjects compared to occupational therapy
Lin et al., 2021 [69]	N = 18 EG 9; CG 9	Virtual reality mirror therapy (Oculus Rift and Leap Motion Controller and dedicated Software)	(1) 6 months post-unilateral infarction or hemorrhage stroke (2) FMUE between 23 and 60 (3) MMSE >24	Sessions of 50 min, two days per week for 9 weeks	30 min of virtual reality mirror therapy plus 20 min of traditional motor task specific exercises	30 min of conventional mirror therapy plus 20 min of traditional motor task specific exercises	Assessment at baseline and post-intervention (9 weeks). FM-UE assessment	Adding virtual reality to mirror therapy can increase upper limb function in chronic stroke subjects.
Huang et al., 2020 [70]	N = 18 EG 20; CG 20	HTC Vive	(1) first unilateral stroke (after 3 and 24 months) (2) hemiparesis with upper limb dysfunction after stroke (3) upper limb rehabilitation to convalescents levels of Brunnstrom stages III to V (4) able to sit and stand without help ford 2 min (BBS ≥ 3)	20 sessions of 30 min, 3 times a week, over 8 weeks plus 1 h of upper limb conventional training	Upper limb conventional training plus immersive virtual reality gaming (shoot balloon, electric current stick and shooting game)	Upper limb conventional training plus physical training with climbing bar, ball bearing and pulley	Assessment at baseline and post-intervention (8 weeks). FM-UE; BBT; FIM self-care score	In stroke rehabilitation the use of an immersive virtual reality system improves upper limb function.
Ögün et al., 2019 [71]	N = 65 EG 33; CG 32	HMD + Leap Motion	(1) MMSE score ≥ 25 (2) stroke between six and 24 months (3) Modified Ashworth Scale score < 3 (4) upper extremity and hand Brunnstrom score ≥ 4	Sessions of 60 min of therapy, three days per week, for six weeks	immersive VR plus Leap motion training consists of four games (play task-oriented games that aimed at gripping and handling of objects with arm and at forearm motion and stability	45 min of conventional upper extremity active exercises including the same used in the VR group plus 15 min of sham VR training	Assessment at baseline and post-intervention (six weeks). FM-UE; ARAT; FIM, PASS	Applying immersive VR with leap motion in stroke patients has a statistical significance on upper extremity function and daily life activities but not on independence.
Subramanian et al., 2015 [74]	N = 24 EG 12; CG 12	Stereoscopic glasses + projector + screen + 3D virtual environment (CAREN)	(1) between 40 and 80 years (2) single ischemic or hemorrhagic stroke 6 to 60 months previously (3) scored 3 to 6/7 on the Chedoke-McMaster Stroke Assessment arm sub-scale (4) no other neurologic or neuromuscular/orthopedic problems affecting the upper limb and trunk	12 sessions: 3 times per week over 4 weeks	A 3D virtual environment (CAREN system) simulated a supermarket scene. Subjects had to point 6 objects placed just beyond arm’s length, without physically touching them.	Participants had to point at targets in a physical environment	Arm and trunk kinematic assessment at baseline, after 4 weeks, and 3 months following intervention. Neuropsychological assessment only at baseline Neurocognitive assessment (Stroop test, RAVLT, TOL, ROCF copy)	An increase in kinematic data for upper limb motor recovery was related to milder neurocognitive deficits.
Subramanian et al., 2013 [76]	N = 32 EG 16; CG 16	Stereoscopic glasses + projector + screen + 3D virtual environment (CAREN)	(1) between 40 and 80 years (2) single stroke 6 to 60 months previously (3) scored 3 to 6/7 on the Chedoke-McMaster Stroke Assessment arm sub-scale (4) no other neurologic or neuromuscular/orthopedic problems affecting the upper limb and trunk	12 sessions of 45 min over 4 weeks	A 3D virtual environment (CAREN system) simulated a supermarket scene. Subjects had to point to 6 objects placed just beyond arm’s length, without physically touching them.	Participants had to point at targets in a physical environment	Assessment at baseline, after 4 weeks, and 3 months following intervention. FMA, RPSS, WMFT-FAS, MAL-AS mean scores, Motivation Task Evaluation Questionnaire, and kinematic analysis	Both groups improved arm motor impairment measures, clinical impairment scores and activity levels. Improvements can be attributed to practice intensity. VE training led to better results in arm motor recovery, especially in the moderate-to-severe group.
Crosbie et al., 2012 [77]	N = 18 EG 9; CG 9	HMD + desktop computer + motion tracking system + sensors	(1) medically stable (2) 18–85 years (3) 6–24 months following a first stroke (4) able to follow a two-step command	Nine sessions of 30–45 min over three weeks	Virtual reality through an HMD for upper limb training. The virtual tasks simulated a range of upper limb tasks related to reach to target, reach and grasp and game tasks.	Conventional therapy based on muscle facilitation, stretching exercises, strengthening activities and functional tasks	Assessment at baseline, post-intervention, after three weeks, and 6 weeks following the intervention. ARAT and an exit questionnaire.	This pilot study has demonstrated the feasibility of a RCT in chronic stroke patients, if careful consideration is given to the recruitment methods and outcome measures. Larger trials are needed to offer high-quality evidence for the specific effect of virtual reality-mediated therapy in upper limb stroke rehabilitation.
Lower limb
De Rooij et al., 2021 [68]	N = 52 EG 28; CG 24	GRAIL	(1) WHO diagnosis of stroke (2) between 2 weeks and 6 months post-stroke (3) walking without help for balance and coordination (FAC > 3) (4) walking in daily life feeling self-limitation (5) community living (6) between 18 and 80 years	Sessions of 30 min, 2 times a week for 6 weeks	Training on GRAIL several virtual environment with different purpose and variable degree of complexity	10–15 min of treadmill training and 15 min of functional gait training	Assessment at baseline, 6 weeks, and 3 months post-intervention. USER-P; TUG test; 6MWT Walking activity (total number of steps a day, duration of walking activity per day and step frequency); Mini-BESTest; FES-I; SIS-16; FSS, HADS anxiety and depression; SS-QOL	The effect of VRT was not statistically different from the effect of non-VRT on analyzed outcomes in community-living people after stroke, but virtual treadmill training was safe and well-tolerated by patients and therefore could be a useful supplement to stroke rehabilitation
Kim et al., 2015 [72]	N = 27 VRCA-G 10; CA-G 11; CG 7	Treadmill + projector + screen + VR program	(1) 6 months post-stroke with hemiplegia (2) gait speed < 0.8 m/s (3) autonomous walking without device for more than 6 min (4) MMSE-K >24	Sessions of 30 min, 3 times a week for 4 weeks (VRCA-G and CA-G) plus sessions of 30 min, 2 times a day for 4 weeks (all groups)	VRCA G trained with a treadmill in 4 different VR environments (sidewalk walking, overground walking, uphill walking and stepping over obstacles) and a progressive speed increase based on the patient’s condition. CA-G trained in the real world in 4 environments: overground walking, stair walking, slope walking and unstable surface walking. Both groups received also general physical therapy.	General physical training	Assessment at baseline and 4 weeks post-intervention. TUG test; ABC; 6MWT; GAITrite walking	In post-stroke subjects, VR treadmill training-based community ambulation and community ambulation training help enhancing dynamic balance ability, activities-specific balance confidence and temporal/spatial gait or gait endurance. These methods are both useful to increase functional skills but, due to VR effects on patient’s physical and psychological fields, it would be more beneficial using this therapy before the community ambulation training.
Cho et al., 2015 [73]	N = 22 EG 11; CG 11	Treadmill + projector + screen + VR program	(1) single stroke, (2) 6 months post-stroke, (3) able to walk 10 m with and without the use of an assistive device, (4) able to understand and follow simple verbal instructions (Korean version of the MMSE score > 24) (5) no severe heart disease or uncontrolled hypertension.	Sessions of 30 min a day, 5 times a week for 4 weeks	VR training (treadmill training) with four cognitive load tasks (memory, arithmetic and two verbal tasks)	VR training (treadmill training)	Assessment at baseline, 3 days after the last experimental training. GAITRite walkway system for spatiotemporal gait parameters under single and dual task conditions.	Beneficial effect of VRTCL on walking function under single and dual task conditions in chronic stroke patients. In the VRTCL group there was a greater increase in walking function during the dual task condition than in the control group.
Lee et al., 2014 [75]	N = 21 EG 10; CG 11	HMD	(1) chronic stroke (2) no medications influencing balance (3) MMSE score < 24 (4) no pain or disability associated with acute musculoskeletal diseases (5) able to sit for over 10 s without help (6) able to stand without help for 1 min	20 sessions of 30 min over 4 weeks for all participants + 20 sessions of 30 min over 4 weeks for EG	16 exercises to train postural control organized on three degrees of difficulty (lying, sitting and standing) plus standard rehabilitation	Standard rehabilitation	Assessment at baseline and after 4 weeks post-intervention. TUG test; BBS; GAITrite walkway	Adding VR training to conventional rehabilitation leads an increase in a number of gait parameters (gait velocity, step length and stride length) compared with traditional therapy alone.
Kang et al. [78]	N = 30 TOF: 10; Treadmill G: 10; CG: 10	HMD + treadmill	(1) hemiparetic stroke 6 months after diagnosis, (2) able to walk without help for more than 15 min, (3) no visual deficits or hemianopsia, (4) MMSE ≥ 21 (5) Brunnstrum stage > 4	Sessions of 30 min, 3 times a week for 4 weeks plus conventional therapy 5 times a week for four weeks	Treadmill training with an HMD showing walking on a street with a progressive increase in speed plus conventional therapy. Treadmill group experienced treadmill without extra device and with a progressive rise of speed plus conventional therapy.	Stretching added range of motion training plus conventional therapy	Assessment at baseline and 4 weeks post-intervention. TUG test; FRT; 10 MWT; 6MWT	Treadmill training with optic flow can help improve balance and gait function in chronic post-stroke patients especially taking advantage of optic flow speed modulation.