Is mirror therapy all it is cracked up to be? Current evidence and future directions

Mirror therapy has attracted a lot of attention recently, here are extracts from a paper reviewing some of the evidence.

G. Lorimer Moseley[1], Alberto Gallace[2], Charles Spence[3]

[1] Department of Physiology, Anatomy and Genetics and Pain Imaging Neuroscience Group, Le Gros Clark Building,
Oxford Centre for fMRI of the Brain, University of Oxford, South Parks Road, Oxford OXON OX1 3QX, UK
[2] Department of Psychology, University of Milano-Bicocca, Italy
[3] Crossmodal Research Laboratory, Department of Experimental Psychology, University of Oxford, Oxford, UK


That viewing oneself through a mirror can evoke peculiar experiences has intrigued researchers for more than a century [34]. The typical approach involves placing one limb behind a mirror that is situated along an observer’s midline. The observer who looks at the mirror’s surface will perceive the reflected limb to be the limb that is hidden behind the mirror. People subjectively report the experience of ‘seeing through’ the mirror’s surface, as though it were actually transparent. This approach exploits the brain’s predilection for prioritising visual feedback over somatosensory/proprioceptive feedback concerning limb position. For the amputee who ‘places’ their phantom behind the mirror, it may feel as though the phantom has ‘come alive’ [30]. This has led to a novel approach to pain reduction in this notoriously difficult to treat population [30], one that has understandably received a great deal of attention in the scientific, clinical, and popular press. We critically evaluate the current state of the evidence that mirror therapy reduces pain, summarise relevant findings concerning the other effects on the human brain of using mirrors, and suggest implications for clinical practice and research.

Does mirror therapy reduce pain?

Case studies and anecdotal data are overwhelmingly supportive of mirror therapy, or ‘virtual’ mirror therapy, in which a virtual reality environment is used instead of a mirror [29], to relieve phantom limb pain, complex regional pain syndrome (CRPS), and for post-surgical rehabilitation. Complete relief is often reported in these studies [2,12,19,23,30,33,35], but case studies are, for obvious reasons, likely to present an overly optimistic picture. More convincing are the results of a clinical trial in which 22 patients with phantom limb pain were randomly allocated to four weeks of mirror therapy, a covered-mirror control group, or to the mental imagery of movement, for 15 min daily [4]. All patients (6/6) in the mirror therapy group, 1/6 of the control group, and none of the imagery group reported a decrease in pain….

Perhaps the most robust trial of mirror therapy undertaken thus far found it to be no better for the immediate reduction of phantom limb pain than motor imagery without a mirror [3]. In that trial, which was part of a larger investigation (n = 80) of mirror therapy and phantom experiences, 15 patients with phantom limb pain were randomly allocated to mirror therapy or to a covered-mirror control group. About 50% of each group reported complete pain relief.

Such contrasting conclusions are possible when studies involve different interventions and measure the effects of those interventions in different ways. The most parsimonious conclusion of the data published to date would therefore appear to be that mirror therapy does not provide any greater immediate pain relief than motor imagery alone [3], but that a program of daily mirror therapy might [4,25], particularly if it constitutes part of a wider graded motor imagery program [28]. In short, the assertion made over a decade ago, that robust experimental trials are required to determine if the visual feedback is indeed an important part of mirror therapy [30], still holds true today…..

Implications for clinical practice and research

Although there is good evidence that programs that incorporate mirror therapy can be helpful for patients with CRPS or phantom limb pain, the current evidence concerning mirror therapy per se, is uncompelling. The obvious implication is that we need more interpretational models and additional data – a robust blinded randomised controlled trial of daily mirror therapy for an extended period and with long-term follow-up, remains to be undertaken, despite calls for such trials more than a decade ago [30]. Furthermore, we also need to critically interrogate the theoretical bases for mirror therapy so that we might also come closer to determining who will benefit and who would not…..

Why is it taking so long? One clear barrier is the difficulty of accessing appropriate patients in sufficient numbers to satisfy the statistical power analysts. Another is ensuring the patients’ participation throughout a trial in which many of them will not gain relief from what is often severe and unremitting pain. One approach that may overcome these barriers is to consider alternate experimental designs, for example, replicated case series, which, if undertaken within strict design constraints, can permit robust conclusions about the effects of treatment [15]….


Despite widespread support of mirror therapy for pain relief in the peer-reviewed [2,4,12,19,23,24, 25,30,33,35], clinical (e.g. [20] (and popular (e.g. [5]) literature, the overwhelming majority of positive data comes from anecdotal reports, which constitute weak evidence at best. Only two well described and robust trials of mirror therapy in isolation exist, on the basis of which we conclude that mirror therapy per se, is probably no better than motor imagery for immediate pain relief, although it is arguably more interesting and might be helpful if used regularly over an extended period. Three high quality trials indicate positive results for a motor imagery program that incorporates mirror therapy, but the role of mirror therapy in the overall effects is not known. Obviously, more robust clinical trials and experimental investigations are still required. In the meantime, the relative dominance of visual input over somatosensory input suggests that mirrors might have utility in pain management and rehabilitation via multisensory interactions. Indeed, mirrors may still have their place in pain practice, but we should be open-minded as to exactly how.


[1] Acerra N, Moseley GL. Dysynchiria: watching the mirror image of the unaffected limb elicits pain on the affected side. Neurology 2005;65:751–3.

[2] Altschuler EL, Hu J. Mirror therapy in a patient with a fractured wrist and no active wrist extension. Scand J Plastic Reconstr Surg Hand Surg 2008;42:110–1.

[3] Brodie EE, Whyte A, Niven CA. Analgesia through the looking glass? A randomized controlled trial investigating the effect of viewing a ‘virtual’ limb upon phantom limb pain, sensation and movement. Euro J Pain 2007;11:428–36.

[4] Chan BL, Witt R, Charrow AP, Magee A, Howard R, Pasquina PF, et al. Mirror therapy for phantom limb pain. N Engl J Med 2007;357:2206–7.

[5] Young S. For amputees, an unlikely painkiller: mirrors, 2008.

[6] Daly A, Bialocerkowski A. Does evidence support physiotherapy management of adult complex regional pain syndrome type one? A systematic review. Euro J Pain, in press.

[7] Durgin F, Evans L, Dunphy N, Klostermann S, Simmons K. Rubber hands feel the touch of light. Psychol Sci 2007;18:152–7.

[8] Flor H, Nikolajsen L, Jensen TS. Phantom limb pain: a case of maladaptive CNS plasticity? Nat Rev Neurosci 2006;7:873–81.

[9] Franz EA, Packman T. Fooling the brain into thinking it sees both hands moving enhances bimanual spatial coupling. Exp Brain Res 2004;157:174–80.

[10] Funase K, Tabira T, Higashi T, Liang N, Kasai T. Increased corticospinal excitability during direct observation of self-movement and indirect observation with a mirror box. Neurosci Lett 2007;419:108–12.

[11] Garry MI, Loftus A, Summers JJ. Mirror, mirror on the wall: viewing a mirror reflection of unilateral hand movements facilitates ipsilateral M1 excitability. Exp Brain Res 2005;163:118–22.

[12] Giraux P, Sirigu A. Illusory movements of the paralyzed limb restore motor cortex activity. NeuroImage 2003;20:S107–11.

[13] Graziano MSA. Where is my arm? The relative role of vision and proprioception in the neuronal representation of limb position. Proc Natl Acad Sci USA 1999;96:10418–21.

[14] Halligan PW, Marshall JC, Wade DT. Left on the Right – Allochiria in a case of left visuospatial neglect. J Neurol Neurosurg Psychiatry 1992;55:717–9.

[15] Hayes SC. Single case experimental-design and empirical clinicalpractice. J Consult Clin Psychol 1981;49:193–211.

[16] Holmes N, Spence C. Beyond the body schema: visual, prosthetic, and technological contributions to bodily perception and awareness. In: Knoblich G, Thornton I, Grosjean M, Shiffrar M, editors. Human body perception from the inside out. Oxford University Press; 2006.

[17] Holmes NP, Spence C. Visual bias of unseen hand position with a mirror: spatial and temporal factors. Exp Brain Res 2005;166:489–97.

[18] Hunter JP, Katz J, DavisKD. The effect of tactile and visual sensory inputs on phantom limb awareness. Brain 2003;126:579–89.

[19] Karmarkar A, Lieberman I. Mirror box therapy for complex regional pain syndrome. Anaesthesia 2006;61:412–3.

[20] Katz J, Melzack R. Phantom limb pain. In: Grafman J, Robertson I, editors. Plasticity and rehabilitation, vol. 9. Amsterdam: Elsevier; 205–29.

[21] Kramer HH, Seddigh S, Moseley GL, Birklein F. Dysynchiria is not a common feature of neuropathic pain. Euro J Pain 2008;12:128–31.

[22] Lundborg G, Rosen B. Hand function after nerve repair. Acta Physiol 2007;189:207–17.

[23] MacLachlan M, McDonald D, Waloch J. Mirror treatment of lower limb phantom pain: a case study. Disabil Rehabil 2004;26:901–4.

[24] McCabe CS, Haigh RC, Blake DR. Mirror visual feedback for the treatment of complex regional pain syndrome (Type 1). Curr Pain Headache Rep 2008;12:103–7.

[25] McCabe CS, Haigh RC, Ring EFJ, Halligan PW, Wall PD, Blake DR. A controlled pilot study of the utility of mirror visual feedback in the treatment of complex regional pain syndrome (type 1). Rheumatology 2003;42:97–101.

[26] Moseley GL. Graded motor imagery is effective for longstanding complex regional pain syndrome. Pain 2004;108:192–8.

[27] Moseley GL. Is successful rehabilitation of complex regional pain syndrome due to sustained attention to the affected limb? A randomised clinical trial. Pain 2005;114:54–61.

[28] Moseley GL. Graded motor imagery for pathologic pain – A randomized controlled trial. Neurology 2006;67:2129–34.

[29] Murray CD, Pettifer S, Howard T, Patchick EL, Caillette F, Kulkarni J, et al. The treatment of phantom limb pain using immersive virtual reality: three case studies. Disabil Rehabil 2007;29:1465–9.

[30] Ramachandran VS, Rogers Ramachandran D, Cobb S. Touching the phantom limb. Nature 1995;377:489–90.

[31] Rizzolatti G, Craighero L. The mirror–neuron system. Ann Rev Neurosci 2004;27:169–92.

[32] Ro T, Wallace R, Hagedorn J, Farne A, Pienkos E. Visual enhancing of tactile perception in the posterior parietal cortex. J Cog Neurosci 2004;16:24–30.

[33] Rosen B, Lundborg G. Training with a mirror in rehabilitation of the hand. Scand J Plastic Reconstr Surg Hand Surg 2005;39:104–8.

[34] Stratton G. Spatial harmony of touch and sight. Mind 1899;8:492–505.

[35] Tichelaar Y, Geertzen JHB, Keizer D, van Wilgen CP. Mirror box therapy added to cognitive behavioural therapy in three chronic complex regional pain syndrome type I patients: a pilot study. Int J Rehabil Res 2007;30:181–8.

See full article at Pain 138 (2008) 7–10


  1. Why can I not find any papers on type 2 CRPS? The mirror box worsens the pain for me!

    Shane Reply:

    The peripheral nerve damage associated with CRPS-2 seems to render a lot of treatments ineffective, so far most of the research seems to be concerned with CRPS-1 as the prognosis seems more positive. I can only assume that as more research is done, once effective treatments are established for CRPS-1 they will start looking at their effectiveness in the chronic CRPS-2 cases.
    The whole CRPS situation is characterized by a severe lack of research at this time, and mirror box therapy has only more recently been looked at in relation to CRPS.

  2. I started playing with a mirror box last year, and interesting things started to happen. My phantom sensations got a lot less, but intermittently, new sharp pain started to occur on occasion. So the sitution *both* got better for the most part, and worse inasmuch as phantom pain occurences are concerned. So I stopped mirror exercises. A neurophysiologist asked “can that be, is it possible”. I find that question funny but obviously some realities are constructed from papers, from articles. So it is indeed relevant to ask for the validity of such research.