Stroke and virtual reality: what the evidence shows (and what it does not)

After a stroke, around 80% of survivors are left with some degree of upper limb motor impairment. Rehabilitation works — but it is slow, it demands repetition at high volume, and it depends on something hard to guarantee in practice: that the person completes more therapy time than they can tolerate in a conventional gym.

That is precisely where virtual reality has been studied. We gathered six recent systematic reviews and meta-analyses — including the reference Cochrane review, updated in 2025 — to answer a simple question: what does the evidence show, and what does it not yet show?

1. The biggest benefit comes from adding, not replacing

This is the most consistent finding, and the easiest to overlook.

The Cochrane review (Laver et al., 2025) finds moderate- to low-certainty evidence that VR is slightly more beneficial than alternative approaches for upper limb function, balance and activity limitation. But it stresses that the benefit is clearest when VR is added to usual care — increasing total therapy time — rather than replacing it.

The correct reading is not "VR beats the physiotherapist". It is that VR makes more rehabilitation dose possible, and dose matters.

2. For gait and balance, the evidence is solid

Corbetta et al. (2015), covering 15 randomised trials and 341 participants, found that substituting some or all of conventional rehabilitation with VR-based rehabilitation produced significant improvements in walking speed, balance and mobility. Added to conventional therapy, the mobility benefit held.

3. For the upper limb, the effect is broad — and timing matters

The largest review in the set, Soleimani et al. (2024), screened 11,834 studies and included 55, with 2,142 patients. VR outperformed conventional therapy on upper limb motor function, functional independence, quality of life, spasticity and dexterity.

Two practical parameters emerge:

• interventions longer than six weeks produced better results;
• starting within six months of the stroke optimised outcomes.

Villarroel et al. (2025) adds two useful notes: gains are most pronounced in the acute and subacute phases, and combining VR with conventional therapy beat VR alone. Notably, whether the software was a commercial game or a purpose-built rehabilitation application did not change outcomes.

4. Where the evidence still disagrees: immersion

There is no consensus here, and it would be dishonest to pretend otherwise.

Soleimani concluded that fully immersive VR produced the greatest gains in gross motor function, with non-immersive approaches better for fine dexterity.

But the network meta-analysis by Zhang et al. (2025), with 34 trials and 1,704 participants, ranked the modalities differently: Microsoft Kinect was most effective for upper limb motor function, followed by Nintendo Wii and non-immersive VR — and head-mounted immersive devices showed no statistically significant difference against conventional therapy.

And Kenea et al. (2025), on immersive VR, found statistically significant improvements on the Fugl-Meyer scale and the Box and Block Test which, in the authors' own words, did not reach clinical significance thresholds.

5. And cost? These studies do not answer that

Worth stating plainly: none of these six reviews measured cost, time to discharge, or return on investment. They measure motor function, balance, dexterity, independence and quality of life.

Economic evidence on VR in healthcare does exist, but it concerns mainly pain and inpatient stays — not stroke. The question "does VR make stroke rehabilitation cheaper?" remains, honestly, open, and be wary of anyone who tells you otherwise without citing a study.

What this means in practice

Read together, the pattern is reasonably clear:

1. VR is a complement, not a substitute. The gain comes mainly from more therapy dose, under clinical supervision.
2. Start early and sustain it. First six months, programmes beyond six weeks.
3. Modality matters and is still contested. Anyone claiming otherwise is selling, not citing.
4. The economic case is not yet made for stroke specifically.

RVer is a virtual reality system designed to be used under the supervision of healthcare professionals, as a non-pharmacological complement to existing rehabilitation programmes — never as a replacement for medical care.

The six reviews cited are linked throughout the article: five have free full text on PubMed Central and the sixth has a free abstract on PubMed. You do not have to take our word for it — read the sources.

← Back to the blog