VR in Rehabilitation: a Therapist's Guide
What VR rehabilitation is, how immersion levels differ, where the evidence is strongest, and how certain it is. A plain-language guide for therapists.
The short answer
Stroke is the best-studied use, with a modest benefit. Compared with the same amount of another therapy, VR slightly improved upper limb function (low certainty), balance (low certainty) and activity limitation (moderate certainty). Added to usual care, it improved upper limb function by a moderate amount (moderate certainty), although that design also gives patients more therapy time [4].
Most trials used screens or consoles, not headsets [4], and results for one format shouldn’t be assumed for another, or for digital rehabilitation in general. Outside stroke, the evidence is smaller. VR is one way to deliver graded, repeated practice with feedback, and its value for a given patient depends on the goal, the evidence for that goal and how well it fits the programme.
What is VR in rehabilitation?
In rehabilitation, VR means computer-generated environments in which a patient moves and interacts while the system responds in real time, so tasks can be practised with feedback, graded and repeated. The label covers both a screen game with a camera tracking the body and reaching for virtual objects in a headset, which is one reason results are hard to compare.
Immersion describes what the technology delivers: how far it shuts out the real world and tracks the user’s body. Presence is the user’s subjective sense of “being there”. Slater and Wilbur’s framework treats immersion as a property of the system and presence as the user’s response to it [1]. Clinical benefit has not been shown to follow from either.
Types of immersion
| Level | Typical setup | Therapist’s note |
|---|---|---|
| Non-immersive | Screen or monitor, with a camera or controller tracking movement | Easiest to set up; the patient stays aware of the room |
| Semi-immersive | Large projection or wide display | Shared viewing; more space needed |
| Fully immersive | Head-mounted display (HMD, “headset”) | Strongest sense of presence; possible cybersickness; needs supervision |
Classifications vary: one review counts the Wii and Kinect as semi-immersive [12], while most papers call them non-immersive.
A 2025 network meta-analysis of 34 RCTs (1,704 participants) compared formats for upper limb recovery after stroke on the Fugl-Meyer scale. Kinect, Wii and other screen-based VR improved scores more than conventional therapy; headsets did not reach statistical significance (MD 4.16 points, 95% CI −0.02 to 8.38). That is not evidence that headsets don’t work. Comparisons between formats were indirect and certainty was not graded [2].
Evidence at a glance
“Certainty” is the review authors’ GRADE rating, not our own.
| Condition and outcome | Source | What was found | Certainty |
|---|---|---|---|
| Stroke, upper limb function: VR vs alternative therapy | Cochrane, 190 RCTs, 7,188 participants [4] | SMD 0.20 (95% CI 0.12 to 0.28); 67 studies, 2,830 participants | Low |
| Stroke, upper limb function: VR added to usual care | Same review [4] | SMD 0.42 (0.26 to 0.58); 21 studies, 689 participants | Moderate |
| Stroke, balance: VR vs alternative | Same review [4] | SMD 0.26 (0.12 to 0.40); 24 studies, 871 participants | Low |
| Stroke, activity limitation: VR vs alternative | Same review [4] | SMD 0.21 (0.11 to 0.32); 33 studies, 1,495 participants | Moderate |
| Stroke, participation/quality of life: VR vs alternative | Same review [4] | SMD 0.11 (−0.02 to 0.24); 16 studies, 963 participants. Not statistically significant | Low |
| Stroke, gait speed: VR vs alternative | Same review [4] | MD 0.05 m/s (−0.02 to 0.13); 10 studies, 304 participants. Not statistically significant | Very low |
| Parkinson’s disease, balance: screen-based VR vs conventional therapy | Meta-analysis, 28 RCTs [15] | SMD 0.42 (0.19 to 0.65); 11 trials, 630 participants | Low |
| Mild cognitive impairment (older adults) | Meta-analysis, 18 RCTs, 722 participants [6] | Memory SMD 0.20, attention 0.25, executive function 0.22; global cognition not significant | Moderate |
What this table does not show: that any specific product works. The reviews pool many different systems.
Where the evidence is strongest: stroke
The 2025 Cochrane review (Laver et al.) pooled 190 randomised trials with 7,188 participants, with searches to September 2023 [4]. Three details matter when reading its results.
- Comparator. Against alternative therapy, controls usually received the same amount of a different therapy [4]. That is the cleanest test of VR itself, and the upper limb effect was small (SMD 0.20).
- Dose. Added to usual care, the effect was larger (SMD 0.42), but the VR group also had more therapy time, and the authors note greater improvement with more therapy [4]. The review can’t separate VR from extra practice.
- Robustness. When the authors excluded studies at unclear or high risk of bias in two or more domains, 17 studies remained in the main upper limb analysis and the effect disappeared (SMD 0.04, 95% CI −0.09 to 0.16) [4].
Gait speed and participation or quality of life showed no statistically significant difference: the trials did not show a benefit, which is not the same as showing equivalence. (SMD is the standardised mean difference; by Cohen’s convention, about 0.2 is small and 0.5 moderate.)
Beyond stroke
Parkinson’s disease. The 2016 Cochrane review (8 trials, 263 participants) found low-quality evidence that VR improved step and stride length compared with physiotherapy; gait, balance and quality of life did not differ significantly [5]. A 2025 meta-analysis found screen-based exergames improved balance more than conventional therapy in mild to moderate disease (low certainty), with no significant difference for mobility (moderate certainty) [15]. We cover it in Does VR improve balance in Parkinson’s disease?
Mild cognitive impairment. A 2025 meta-analysis reported small improvements in memory, attention and executive function compared with conventional treatment or no intervention (moderate certainty), and no significant effect on global cognition [6]. Samples were small and blinding was lacking [6].
Hand dexterity after stroke. A 2025 meta-analysis of immersive VR found a gain of 2.85 blocks on the Box and Block Test over conventional rehabilitation (95% CI 0.70 to 4.99; I² = 93%), and no significant difference on the Action Research Arm Test [13]. That average is below the smallest real difference for an individual, about 5.5 blocks [14], though that threshold applies to one person, not a group mean. See VR and hand dexterity after stroke: is the gain clinically meaningful? and our video Does VR Improve Hand Dexterity After Stroke?
Open questions
- Dose. The right amount of VR is not settled [4].
- Long-term effects. Most outcomes were measured at the end of treatment.
- Cost. A 2024 review found only five economic studies of VR in neurological rehabilitation, all console-based. It suggests possible savings, mainly from low system prices and less travel in telerehabilitation, but the studies varied widely [12]. Robust cost-effectiveness evidence is lacking.
Safety and side effects
In the Cochrane review, adverse events were few and relatively mild, mostly transient dizziness, headache and pain, but only 59 of the 190 studies reported monitoring them [4].
Cybersickness (nausea and disorientation with headsets) deserves attention. A meta-analysis of 55 studies (3,016 participants) found relatively high symptom scores, highest with gaming content and influenced by locomotion and exposure time. It excluded clinical samples and included few older users, so the findings may not transfer directly to patients [8].
Implementation: separate from the outcome evidence
Choosing and running a system raises questions the outcome trials don’t answer: how far it adapts to the patient, how much supervision it needs, what it records and how it fits the service. These are reasonable selection criteria, but no study cited here shows that they change outcomes.
Regulation. In the EU, software intended for a medical purpose may qualify as a medical device under Regulation (EU) 2017/745 (MDR) [9], and MDCG 2019-11 explains qualification and classification [10]. A wellness game is different from a product claiming a therapeutic effect; ask which one you are looking at.
Measuring outcomes. Consensus recommendations for stroke recovery trials set out core sensorimotor measures and make a strong case for adding kinematic and kinetic measures [11]. If a system records movement data, check how it relates to the scales your service uses.
Choosing a system. Our guide to choosing a VR rehabilitation system covers adaptation, supervision, monitoring and day-to-day use. For format differences, see Immersive vs non-immersive VR.
How to interpret and apply the evidence
What the evidence supports: the stroke benefits summarised above, with few, mild adverse events where they were monitored [4]. Evidence in other conditions is smaller [5][6][15].
Our suggestions for integrating VR well. These are editorial guidance based on general rehabilitation practice. The studies did not test them.
- Start from the therapeutic goal, such as reaching, weight shifting or attention, and choose tasks that train it.
- Set difficulty and progression from your own assessment, and adjust them as the patient changes.
- Use VR within the programme rather than as a separate activity.
- With headsets, start with short seated sessions and ask about dizziness and nausea.
- Reassess with the scales you already use, and change the plan if the patient isn’t improving.
Limitations to keep in mind. Only 36 of the 190 stroke trials had more than 50 participants, and risk of bias was often unclear because of poor reporting [4]. An accompanying editorial argues that the review’s 2001 definition of VR groups very different interventions together [3]; a consensus paper proposes staged VR trial designs (VR1 to VR3) [7]. Two Cochrane authors wrote included trials and did not assess them; one is a named inventor on VR rehabilitation patents, and another has industry links in VR rehabilitation [4].
Bottom line for clinicians
- Supported benefit: after stroke, VR slightly improves upper limb function, balance and activity compared with other therapy, and adds a moderate upper limb benefit on top of usual care.
- Integration opportunity: VR can add graded, repeated practice with feedback to upper limb and balance work, as extra therapy time or within existing sessions.
- Your role: choosing patients and goals, progressing the tasks, supervising tolerance and checking the response with your usual measures. That is what decides whether VR is the right tool for a given patient.
References
- Slater M, Wilbur S. A Framework for Immersive Virtual Environments (FIVE). Presence 1997;6(6):603-616. DOI: 10.1162/pres.1997.6.6.603
- Zhang J, et al. Effects of virtual reality with different modalities on upper limb recovery: systematic review and network meta-analysis. Front Neurol 2025. DOI: 10.3389/fneur.2025.1544135
- Wade DT. Virtual reality for stroke rehabilitation: time to rethink? Lessons from a Cochrane review. Cochrane Database Syst Rev 2025, ED000174 (editorial). DOI: 10.1002/14651858.ED000174
- Laver KE, Lange B, George S, Deutsch JE, Saposnik G, Chapman M, Crotty M. Virtual reality for stroke rehabilitation. Cochrane Database Syst Rev 2025;6:CD008349. DOI: 10.1002/14651858.CD008349.pub5
- Dockx K, et al. Virtual reality for rehabilitation in Parkinson’s disease. Cochrane Database Syst Rev 2016. DOI: 10.1002/14651858.CD010760.pub2
- Yang Q, et al. Virtual reality interventions for older adults with mild cognitive impairment: systematic review and meta-analysis of RCTs. J Med Internet Res 2025;27:e59195. DOI: 10.2196/59195
- Birckhead B, et al. Recommendations for methodology of virtual reality clinical trials in health care. JMIR Ment Health 2019;6(1):e11973. DOI: 10.2196/11973
- Saredakis D, et al. Factors associated with virtual reality sickness in head-mounted displays: systematic review and meta-analysis. Front Hum Neurosci 2020. DOI: 10.3389/fnhum.2020.00096
- Regulation (EU) 2017/745 on medical devices. OJ L 117, 5 May 2017. eur-lex.europa.eu/eli/reg/2017/745/oj
- MDCG 2019-11 rev.1. Qualification and classification of software. European Commission, June 2025.
- Kwakkel G, et al. Standardized measurement of sensorimotor recovery in stroke trials. Int J Stroke 2017;12(5):451-461. DOI: 10.1177/1747493017711813
- Cano-de-la-Cuerda R, et al. Economic cost of rehabilitation with robotic and virtual reality systems in people with neurological disorders: a systematic review. J Clin Med 2024;13(6):1531. DOI: 10.3390/jcm13061531
- Kenea CD, Abessa TG, Lamba D, Bonnechère B. Immersive Virtual Reality in Stroke Rehabilitation: A Systematic Review and Meta-Analysis of Its Efficacy in Upper Limb Recovery. J Clin Med 2025;14(6):1783. DOI: 10.3390/jcm14061783
- Chen HM, et al. Test-retest reproducibility and smallest real difference of 5 hand function tests in patients with stroke. Neurorehabil Neural Repair 2009;23(5):435-40. PubMed 19261767
- Lheureux A, Lejeune T, Simons A, Gillis A, Everard G. Virtual reality effects on balance and mobility in people with Parkinson’s disease: a systematic review with meta-analysis. Ann Phys Rehabil Med 2025;68(6):101967. DOI: 10.1016/j.rehab.2025.101967
Frequently asked questions
Is VR rehabilitation effective after stroke?
It may help modestly. The 2025 Cochrane review found a small upper limb benefit versus alternative therapy (low certainty) and a larger one when VR was added to usual care (moderate certainty) [4]. It does not show that any specific product works.
What is the difference between immersive and non-immersive VR?
Non-immersive VR uses a screen and the patient sees the room around them. Immersive VR uses a headset that blocks the real environment. Immersion describes the technology; presence is the user's feeling of being in the virtual space [1].
Is VR rehabilitation safe?
In stroke trials that monitored it, adverse events were few and relatively mild: transient dizziness, headache and pain [4]. Only 59 of 190 studies reported monitoring for them [4], and the main meta-analysis of headset cybersickness excluded clinical samples [8].
Does VR work for conditions other than stroke?
Evidence exists but is thinner. In Parkinson's disease, a 2025 meta-analysis found screen-based VR improved balance more than conventional therapy, with a small effect and low certainty [15]. For mild cognitive impairment, a meta-analysis found small gains in memory, attention and executive function, with moderate-certainty evidence [6].
Is VR rehabilitation a medical device in the EU?
It depends on the intended purpose. Software with a medical purpose can fall under the MDR [9]; MDCG 2019-11 provides guidance on qualification and classification [10].