“For in and out, above, about, below,
‘Tis nothing but a Magic Shadow-show
Play’d in a Box whose Candle is the Sun,
Round which we Phantom Figures come and go.”

— The Rubâiyât of Omar Khayyâm

We have been raised to believe in a concrete reality defined by matter, which is something that occupies space, and is either perceived by our sensory organs, or measured by other tools [1]. As beings with a physical dimension, we humans tend to think of our “selves” as unified “wholes”. Of course, our everyday lives and activities pose little threat to this perception: the reality of the different parts of this “whole” is proven to us every day through our sensory experiences. We are constantly seeing, hearing, and feeling the world through our sensory organs. It is then only natural that any sensation in these organs, including pain, should be considered “normal”, if not always welcome, for we have learned that our body parts, which compose our physical substance, are entitled to sensations. The question that comes up, however, is whether the opposite is necessarily untrue.

The phantom limb phenomenon has intrigued mankind for centuries. The medical description of the condition dates back to the 1530s, when the French military surgeon Ambroise Paré described pain in the non-existent amputated limbs of soldiers [2]. The term “phantom limb” was coined by Silas Weir Mitchell, another military surgeon, in 1872. Like many others, Mitchell was intrigued by soldiers’ accounts of the vivid sensation of pain in these “ghost organs”, i.e. the limbs that were no longer present [2]. Naturally, the questions that came to mind were: what is causing this sensation of pain? And is this pain real or illusory?

The initial speculations regarding pain in a non-existent body part attributed it to some form of hallucination [3], for how could an organ that does not exist claim to sense anything? This may reflect the huge variability in the reports of the prevalence of phantom limb pain, which is said to affect anywhere between 2%-80% of all amputees [4]. Older reports tend to indicate a much lower rate in comparison with more recent studies, possibly because amputees may not have been willing to report their phantom pain for concerns over a label of insanity [5]. Even today, the view that phantom limb pain is “all in the [patient’s] head” is still prevalent, and there are those investigators who place it in the same category as “other mental disturbances” [3 and 5].

However, for those suffering from this phenomenon, the feeling of pain in an organ that no longer exists is a living reality. This pain has been described with various qualities, such as shooting, stabbing, boring, squeezing, throbbing, and burning [6]. The pain is usually experienced intermittently and is primarily localized to those parts of the missing limb which are farthest from the site of amputation (for example in the fingers of an arm amputee). Various factors may trigger the sensation of phantom pain, including certain movements, pressure, and psychological factors [7]. The onset of pain is often early after the amputation, and although some studies report that the pain diminishes over time, others report no such reduction [4 and 5]. The pain could often be excruciating, and can involve experiences of voluntary movement, paralysis, or strong involuntary spasms in the absent limb [8].

So what is the nature of this well-characterized, “real”, and often debilitating pain? How can one be made to feel a vivid sensation in a body part which no longer exists? And how can this unwelcome sensation be treated when the afflicted site has no physical reality? These are some of the questions that drove researchers to deciphering the mechanisms of phantom limb pain. In this attempt, the earlier studies turned their eyes to peripheral nerve factors. For example, it was found that the amputated nerve endings at the stump cluster together to form structures called “neuromas”. It was then postulated that neuromas (which are quite sensitive and also capable of spontaneous neural activity) generate impulses that go to the brain and are perceived as pain [2 and 4]. Support for this hypothesis came from the finding that in many patients, a surgical removal of the neuromas leads to pain relief [4]. However, not all patients benefit from this treatment. Furthermore, the peripheral theory does not explain why some amputees suffer from phantom limb pain and some don’t [2]. Therefore, although the peripheral nerve mechanism may play a role in generating phantom limb pain, it cannot be the entire story behind this ghostly sensation.

One remarkable facet of the phantom limb pain is that it seems to be related to the pre-amputation pain [2, 5, and 6]. It has been reported that if the individual has had chronic pain in the limb prior to amputation, phantom limb pain has a higher incidence. It has a lower incidence if the amputation is performed when the individual is very young [5]. Furthermore, the phantom pain seems to be qualitatively similar to the pre-amputation pain. For example, if a patient suffers a painful ulcer on her leg prior to amputation, s/he may later report phantom pain with the same characteristics and in the same location as the ulcer pain [2]. These findings have led some researchers to suggest a role for “pain memory” in the formation of phantom limb pain [5]. It is however quite remarkable that phantom limb pain can also be observed in people born without a limb. And even though this congenital phantom limb pain is reported to be rare, it may bring the importance of pain memory into question [2 and 5].

Another attempt at understanding phantom limb pain was made by studying the brain and the spinal cord. It has been postulated that after an amputation, the loss of sensory input to the spinal cord will cause changes in spinal structures called dorsal root ganglia (which previously received input from the presently missing limb). These changes include increased excitability and increased activity in N-methyl D-aspartate (NMDA) receptors, which may lead to the perception of pain [2, 6 and 7]. Another theory is that there is a genetically determined “matrix” of interconnected neurons in the brain, and phantom limb pain is caused when the matrix is deprived of normal sensory input from the limbs [9]. The deprivation causes the neurons to fire abnormally, giving rise to the perception of pain [2 and 9]. Although this theory provides a framework that is consistent with peripheral, spinal, and central mechanisms and explains some previously puzzling data, it is difficult to test and thus its accuracy remains unknown [4].

Some interesting insight into the phantom pain phenomenon has come from studies of the somatosensory cortex of the brain. It has long been known that the various parts of the body relay their somatosensory (i.e. touch) messages to different parts in the somatosensory cortex. When amputation leads to an abolition of sensory input from a specific body part, the area of the somatosensory cortex corresponding to that body part is often “invaded” by the adjacent parts [2]. For example, it has been observed that when the face of someone with a hand amputation is touched, the person feels a sensation of touch in his/her phantom hand [8]. This may be due to an “invasion” or “cortical reorganization”, because the face and hand areas are adjacent in the map of the somatosensory cortex. Remarkably, it has been found that phantom pain intensity increases with an increased degree of cortical reorganization [6]. Cortical remapping can sometimes be reversed by eliminating input from the amputation stump by inducing anaesthesia, and in one study, peripheral anaesthesia completely eliminated cortical reorganisation and phantom pain in three of six patients [5]. Yet cortical reorganization does not explain all phantom limb-related phenomena. For instance, little reorganization is observed in non-painful phantom limb sensations [10] and the extent to which this mechanism plays a role in pain sensation is still not clear [2].

Even though these neurological studies have contributed significantly to our understanding of phantom limb pain, the role of psychological explanations should not be neglected either. Strangely enough, the view that the phantom limb pain is only in the patient’s head, and is a somatic manifestation of an underlying anxiety or personality disorder is still accepted, even though empirical evidence has shown that people who suffer from phantom limb pain have normal psychological profiles [2 and 5]. It is however important to note that phantom limb pain can be elicited and exacerbated by a variety of psychological factors, in particular stress [5]. Cognitive factors may also play a role, for example those patients with a better pain appraisal and better coping strategies may be able to reduce the amount of phantom pain experienced [2 and 5].

Despite the above findings, the phantom limb pain phenomenon remains poorly understood and although various treatment approaches have been experimented with, the results have been of limited use. Medication (particularly the use of antidepressants, sodium channel blockers, NMDA-receptor antagonists, and calcitonin) remains the most “successful” form of treatment [6, 7, and 10]. Various anaesthetic blocks have been claimed to be effective, but none have proven to be effective in well-controlled trials [6]. Other forms of treatment have also been implemented, including electrical stimulation of the spinal cord and deep brain structures (with some benefit in alleviating chronic pain), and neurosurgical techniques (which may be effective at providing temporary pain relief) [6 and 11]. More recently, alternative approaches have been suggested, such as acupuncture and use of “virtual reality” (such as the mirror box experiment) [12 & 13]. Preliminary studies have indicated some benefit for these methods, but larger samples and more careful controls are needed to establish any definitive results. For example, several studies have shown that when patients with an upper arm amputation placed their intact hand inside a box with a mirror inside (so that they could see the reflection of their hand in the mirror) and asked to place their phantom hand behind the mirror (where the image of the intact hand is reflected), they can eventually project the movements of the intact hand onto the phantom limb, and in this way control the involuntary movements of the phantom [8 and 14]. However, there have been mixed reports about the effects of this procedure on phantom pain [8 and 13].

The difficulty with studying phantom limb pain stems from several factors. Most studies have small sample sizes and short follow up periods [6]. Furthermore, all descriptions of pain are reported by patients, usually after the amputation, and patients’ description of pain may be imprecise, and their memories of previous pain may not always be accurate [6]. In addition amputees are quite a diverse population and may differ in a variety of characteristics [4]. These factors also create difficulties in comparing the studies on phantom limb pain and reaching a consensus about its causes and mechanisms [4].

Thus, the mysteries of phantom limb pain live on to this day. However, studying this phenomenon has led to profound insights into the organization of the normal human brain. For instance, it was previously thought that brain remapping can only take place early in development, whereas it is now known that even the adult human brain can undergo functional reorganization [5 and 8]. Another example is how the brain integrates visual and tactile stimuli in the mirror box experiment to “feel” the non-existent limb. This shows how these “fragments of information” coming from different senses are combined to form a unified body image and an enduring sense of self [8]. Many people are fascinated by questions relating to the nature of the self and how our brain comes to construct a body image. And although these questions are not directly addressed by most neuroscience experiments, extraordinary puzzles such as the phantom limb pain, though difficult to understand, may shed some light upon the mysteries of the human mind.


[1] Cowie, A.P. (ed.) Oxford Advanced Learner’s Dictionary of Current English, 4th Edition. Oxford University Press, 1989.

[2] Woodhouse, A. Clin Exp Pharmacol Physiol. 2005, 32(1-2):132-4.

[3] MacLachlan, M., Desmond, D., Horgan, O. Journal of Rehabilitation Research & Development. 2003, 40(1): 59 – 66.

[4] Hill, A. Journal of Pain and Symptom Management. 1999, 17(2): 125-142.

[5] Flor, H. The Lancet Neurology. 2002, 1(3): 182-189.

[6] Nikolajsen, L., & Jensen, T.S. British Journal of Anaesthesia. 2001, 87(1):107–116.

[7] D’Arcy, Y. Nursing. 2005, 35(11): 17.

[8] Ramachandran, V.S. Phil. Trans. R. Soc. Lond. B. 1998, 353:1851-1859.

[9] Melzack R, Coderre TJ, Katz J, Vaccarino AL.. Ann N Y Acad Sci. 2001, 933:157-74.

[10] Flor, H. Orthopade. (2004). 33(5):553-7.

[11] Bittar, R.G., Otero, S., Carter, H. & Aziz T.Z. Journal of Clinical Neuroscience. 2005, 12(4):399-404.

[12] Bradbrook, D. Acupunct. Med. 2004, 22(2):93-7.

[13] Murray, C.D., Patchick, EL, Caillette F, Howard T, Pettifer S. Stud Health Technol Inform. 2005, 119:407-12.

[14] MacLachlan, M., McDonald, D., & Waloch, J. Disabil Rehabil. 2004, 26(14-15):901-4.