Mirror therapy for chronic pain.

How does mirror therapy work?

To produce pain, the brain has to decide that a body part is under threat. It makes that decision using every source of information available. Signals from the tissue count, but so do vision, memory, expectation, and context. Mirror therapy exploits the fact that vision is unusually persuasive.

When you watch a reflection of your healthy hand opening and closing where your painful hand should be, the visual system delivers unambiguous evidence: that limb is moving freely, and nothing bad is happening. That evidence conflicts with the brain's existing prediction, and the conflict is what creates room for the prediction to change.

What's happening in the brain

Watching a limb move activates somatosensory and premotor circuitry associated with that limb, much of the same machinery involved in moving it yourself. In chronic pain and after amputation, the cortical representation of the affected body part is often distorted. Mirror feedback appears to help renormalize it. Studies of illusion-based mirror setups during functional imaging show measurable changes in brain activation, which is why the mechanism is usually described as cortical reorganisation rather than distraction.

What conditions does mirror therapy treat?

Phantom limb pain

This is where mirror therapy began. Ramachandran's insight was that phantom pain often involves a limb that feels frozen or clenched in a painful position, and that the brain has no visual evidence to contradict it. A mirror reflecting the intact limb restores the missing feedback. The phantom is seen to move, and for many people the pain eases. See our page on phantom limb pain.

Complex regional pain syndrome

CRPS features severe pain, allodynia, and often a distorted sense of the affected limb's size or position. Mirror therapy is standardly used as the final stage of graded motor imagery, and the GMI protocol has randomized evidence in CRPS. See complex regional pain syndrome.

Stroke rehabilitation

Mirror therapy is widely used to improve motor function in a hemiparetic arm, and to address CRPS type 1 that develops after stroke. It is used alongside conventional rehabilitation rather than instead of it. This is the largest evidence base the technique has, and it gets its own section below: what the research shows for stroke.

Chronic limb and joint pain

Evidence here is younger and thinner than in the conditions above, but the rationale carries over: where movement is limited by anticipated pain rather than by structure, visual feedback offers a way to test the limit without triggering it. Karuna's mechanistic work sits in this territory.

Does mirror therapy work after a stroke?

Stroke is where mirror therapy has been studied most. The reference point is a Cochrane systematic review by Thieme and colleagues, updated in 2018, pooling 62 studies and 1,982 participants. It found evidence that mirror therapy improves movement and daily function when used as an addition to conventional rehabilitation.

Those are small-to-moderate benefits on moderate-certainty evidence, which in rehabilitation research is a respectable result. Two caveats belong right next to them, though, and are usually left out.

The two caveats worth knowing

The pain benefit is mostly a CRPS benefit. Across all trials, pain improved (SMD −0.89). Remove the CRPS studies and the remaining four trials show SMD −0.23, with a confidence interval from −0.53 to 0.08 that includes zero. So mirror therapy's analgesic effect after stroke rests largely on people who had developed complex regional pain syndrome, not on stroke pain in general.

The movement gain may sit below the threshold people notice. On the Fugl-Meyer upper extremity scale, the pooled improvement was 4.32 points (95% CI 2.46 to 6.19) across 28 trials. The review itself cites a minimum important difference of roughly 5.25 points, so the average gain did not reach the level normally considered clinically meaningful. Statistically real is not the same as noticeable in daily life.

What the guidelines say

The UK's NICE guideline on stroke rehabilitation (NG236, 2023) recommends that clinicians *consider* mirror therapy for people with muscle weakness in their upper or lower limbs after a stroke, as an adjunct to their rehabilitation programme. In NICE's grading language, “consider” is a conditional recommendation rather than a strong one.

NICE also specifies a protocol: start within the first six months after a stroke, with sessions of around 30 minutes, at least five times per week over four weeks, supervised at first. That matches what trials typically did, an average of roughly 30 minutes a session, five times a week, for four weeks.

For balance, the American Heart Association and American Stroke Association's 2016 adult stroke rehabilitation guideline does not address mirror therapy at all. Its evidence review closed in 2014, before the Cochrane update. Anyone citing an AHA class or evidence level for mirror therapy is mistaking it for the guideline's separate recommendation on bilateral training.

The evidence is not uniform. A 2025 meta-analysis in Clinical Rehabilitation of 18 trials and 633 patients found upper limb motor function improved by a mean difference of 1.79 (95% CI 0.04 to 3.54, p = 0.04), a result only marginally clear of zero, with 16 of the 18 trials rated as having some risk-of-bias concerns. A separate 2021 review found benefit in the subacute phase but non-significant results against sham comparisons.

The practical summary: worth doing, cheap, safe, best started within six months, and an addition to rehabilitation rather than a substitute for it. If you are recovering from a stroke, the person to set this up is your physiotherapist or occupational therapist.

What has Karuna's research found about mirror visual feedback?

Karuna Labs ran a small preliminary mechanistic study, approved by an institutional review board, in 17 adults with chronic shoulder pain, to test whether mirror visual feedback delivered in virtual reality could influence pain-free range of motion.

Participants wore a VR headset and moved each arm while a virtual avatar mirrored their movement. In one condition the avatar matched the arm they were actually moving; in the other, the movement of the unaffected arm was displayed as though it were the painful arm.

The second condition is what makes the mechanism visible. When people moved their healthy arm but saw the painful side moving, their pain-free range of motion decreased, significantly so for shoulder flexion and scaption. Nothing about the healthy arm itself had changed. The only difference was which limb they believed was moving.

This was single-session work in a small sample, and it was designed to probe mechanism rather than to demonstrate a treatment effect. What it establishes is that mirror visual feedback in VR reaches self-perceived movement limits, which is exactly what graded exposure work targets. It also underpins the design of our virtual reality training, where the same principle is used deliberately, in the helpful direction.

How do you do mirror therapy?

The setup is simple enough to do at home, though it's worth having a clinician guide the first sessions, particularly with CRPS, where starting too aggressively can provoke a flare.

1. Sit at a table with a mirror positioned vertically along your midline, reflective side facing your unaffected limb.
2. Place the affected limb behind the mirror, out of sight. Remove watches and jewellery from both sides.
3. Position yourself so the reflection appears where the affected limb would be.
4. Move the unaffected limb slowly and watch the reflection, allowing yourself to accept it as the affected limb.
5. Start with 5–10 minutes and simple movements. Build toward 20–30 minutes and more complex ones.
6. Stop if pain increases sharply, or if you feel disoriented or distressed. Restart shorter and slower next time.

Consistency matters more than duration. Most protocols use daily or near-daily sessions over several weeks. The point is to accumulate evidence, and that takes repetition.

What are the limits of mirror therapy, and how does VR extend it?

Mirror therapy has one structural constraint that no amount of technique can solve: it requires a body part that comes in a pair. A mirror can substitute a left hand for a right hand. It can do nothing for the lower back, the neck, or the midline, and nothing for someone whose pain affects both sides.

Since chronic low back pain is the single largest chronic pain problem there is, that limitation excludes an enormous number of people from the technique.

What virtual reality changes

In VR, the person occupies a virtual body from a first-person perspective. The system does not need to reflect anything, because it can simply render what the body appears to be doing. That removes the lateralization requirement and adds capabilities a mirror does not have:

• Non-lateralized regions. Spine, neck, and trunk can be addressed, because no mirrored pair is needed.
• Augmentation. Movement can be visually amplified or reduced relative to actual motion, so a small real movement can be seen as a larger one, and vice versa.
• Measurement. Joint angles are tracked continuously, so range of motion can be graded and paced to what a person can currently do.
• Graded difficulty. Exercises adjust to current capability, which keeps exposure within tolerance and reduces flare-ups.

This is the reasoning behind Karuna's approach: take the mechanism that makes mirror therapy work, and remove the constraint that limits where it can be applied. More on that in VR pain management and how the programme works.