Giving up the ghost: a review of phantom limb phenomena.

The literature concerning painful and non-painful phantom limb phenomena was reviewed to establish our present state of understanding of the phenomena and explore avenues for future research. Although advances have been great in recent years, several misconceptions remain. Points of focus in the review include past definitional deficiencies, conflicting incidence figures, separation of painful and non-painful phantom limb phenomena, precipitating and relieving factors as enumerated by amputees, psychological versus physiological explanations, and treatment attempts. Strengths and weaknesses of past research are discussed and implications for future research are drawn.

From the time, in 1551, that Pard first described the phenomenon in which an individual experiences the presence of a missing limb (cited in Shukla, Sahu, Tripathi, & Gupta, 1982), and Mitchell first coined the term phantom limb" [1892], science has struggled to explain it. In spite of centuries of study, a universally accepted explanation has not yet been established. The purpose of this discussion is to explore some of the possible reasons for our limited progress in this area by examining some of the research conducted to date. Definition

Probably the greatest flaw in past research has been the failure of the studies to adhere to universal definitions for the various phantom limb phenomena. Because of this lacking, readers cannot be certain that researchers are studying the same phenomenon.

In the past, all phantom limb has been lumped under the all-inclusive heading of phantom pain. This term, however, does not fully describe the experience. Jensen, Krebs, Nielsen and Rasmussen [1983] have shown, for example, that phantom limb can be divided into two categories. They term "phantom limb", referring to any non-painful sensations occurring in the missing limb, and "phantom pain", referring to painful sensations. They note, however, that painful and non-painful phantom limb usually occur within the same individual. In addition, they are careful to separate these perceptions from stump pain. Others [Dawson & Arnold, 1981; Dernham, 1986; Malin & Winkelmuller, 1985; Sherman & Sherinan, 1985; Sherman, Sherman, & Parker, 1984; Shukla et al., 1982] have also recognized more than one type of phantom limb. However, the results of these studies cannot be effectively compared because they do not use the same form of categorization when dividing painful from non-painful perceptions. Nor do they use a similar scale when collecting the subjective data.

For example, Jensen et al. [1983] included exteroceptive sensations (itching, tingling, etc.), temperature sensations, kinaesthesia and movement as non-painful phenomena. Kinaesthesia included the sense of limb position in various forms of length and volume; movement could be spontaneous or willed. They reported the following qualities of painful phantom limb: knife-like, sticking, shooting, pricking, burning, squeezing, throbbing, pressing, cramp-like, sawing, dull, freezing, and crushing. In contrast, Shukla et al. [1982] included only telescopy (shortening of the phantom limb) or movement as non-painful phantom limb. Their types of painful phantom limb were: burning, electric current, throbbing, cramps, itching, cutting, jactitation (twitching), stabbing, tearing, and crushing. It becomes clear by examining the categorization of just these two studies that there has been no reliable division between the painful and non-painful phantom limb. It is very possible that the electric current described as painful phantom limb in the Shukla et al. [1982] study is the same experience that is described as tingling non-painful phantom limb in the Jensen et al. [1983] study since there was no elaboration of descriptors by the respondents.

Sherman and Sherman [1985] provide the best example of the confusion between non-painful and painful phantom limb in that their descriptions are the same for both. The descriptors that they list include sharp shock/shooting, squeeze, hot, cramp, tingle, warm, unusual position and broken (in order of frequency). Part of the reason for the use of the same adjectives for both painful and non-painful limb could be the way in which the amputees were asked to report their perceptions. In the questionnaire the amputees were first asked to describe their non-painful phantom limb perceptions - "What do they feel like? (For instance: warm, squeezing, etc.)" [Sherman et al., 1984, pp. 85-861 - and then were asked the same question about the painful phantom limb, but were not given any examples. It is possible that the presence of prompting by example in the first question created a mental set for the respondents which led them to use the same descriptors for the second question. This is an essential point when considering that the authors concluded from the similarity between painful and non-painful descriptors that painful phantom limb may be an intensified form of non-painful limb. Refer to Table 1 for a comparative listing of definitional descriptors. incidence

In addition to, and partly as a result of the definitional problems, studies have been inconsistent in their reporting of incidence statistics. Only recently have researchers attempted large-scale phenomenological studies of painful and non-painful phantom limb, but these investigations have added tremendously to the body of knowledge on phantom limb phenomena. In the past it was estimated that only about 5 to 10% of individuals experienced painful phantom limb [Rosenzweig and Leiman, 1982]. However, more recent studies report that between 69 and 85% of individuals experience painful phantom limb, and the incidence of non-painful phantom limb is even higher, at between 84 and 98% [Jensen et al., 1983; Sherman & Sherman, 1985; Sherman et al., 1984; Shukla et al., 1982). The length of time between loss of limb and onset of the phantom limb phenomena was shown to vary from hours and days [Jensen et al., 1983; Sherman, Gall, & Gonnly, 1979; Shukla et al., 1982] to weeks and months [Sherman et al., 1979]. It is possible that different physiological mechanisms are responsible for immediate phantom limb as opposed to that which occurs later on, explaining why some individuals do not experience an immediate phantom limb. These mechanisms will be discussed in a later section.

Persistence of the phantom limb. In the past it was believed that phantom limb eventually faded away in most if not all cases [Dworetzky, 1985], but research does not support this. The percentages of individuals reporting the disappearance of phantom limb ranged between 14 and 22% [Jensen et al., 1983; Sherman & Sherman, 1985; Sherman et al., 1984; Shukla et al., 1982]. In fact, the percentages of individuals reporting an increase in phantom limb ranged between 8 and 44% [Jensen et al., 1983; Sherman & Sherman, 1985; Sherman et al., 1984; Shukla et al., 19821. However, 50% of amputees reported that the sensations decreased in intensity over time [Sherman et al., 1984]. Possible physiological mechanisms underlying these findings will be discussed in a later section.

Duration and frequency of painful phantom limb episodes. Other highly variable aspects of painful phantom limb include the duration and frequency of painful episodes. Duration was found to range from seconds or minutes to hours or days [Sherman & Sherman, 1985; Sherman et al., 1984], with the worst cases reporting continuous pain. Frequency was found to range from twenty days out of a month to less than once per month [Sherman & Sherman, 1985; Sherman et al., 1984], again with the worst cases reporting constant pain. Precipitating and relieving factors

One aspect of phantom limb phenomena that was unaddressed prior to the phenomenological studies is a listing of the antecedents to a phantom limb episode and the actions that facilitate relief, as reported by amputees. Precipitating factors were as varied as the descriptors, but there were some common threads. Refer to Table 2 for a comparative listing of precipitating and relieving factors.

It is important to note, that in studies reporting relieving factors, over 50% of the subjects reported that they had found no factors that could relieve the painful phantom limb. In addition, it is clear that there is some overlap in that some of the same factors are able to both increase and decrease phantom limb. A possible physiological reason for this inconsistency will be discussed in the next section. Causes

As with any issue that is as little understood as phantom limb phenomena, there has been much disagreement about the causes. Basically, the possible causes have been divided into two categories- psychological and physical. Because the possibility of a psychological cause exists for some cases of phantom limb, a brief listing of the psychological explanations for phantom limb will be included. However, the primary focus of this discussion will be on studies of organic causes.

Psychological. Dernham [1986] outlined six psychological explanations for the occurrence of painful phantom limb. These include beliefs that the individual (1) experiences a radical change in life-style; (2) may possess a rigid personality; (3) may possess a compulsive, self-reliant personality; (4) may be experiencing anxiety over the change in body image; (5) may be experiencing a distortion of the grieving process; or (6) may experience the phantom limb as a symptom of depression. Dawson and Arnold [1981] asserted that the failure of surgery to alleviate painful phantom limb has led to the assumption that the phantom limb may fulfill psychological needs. It has also been proposed that for some individuals, persisting painful phantom limb may be evidence of an unstable personality [Gillis, 1969: cited in Dawson & Arnold, 1981].

Additional theories have been postulated as explanations for the experience of painful phantom limb. The first of these was that painful phantom limb was a result of wish-fulfillment, stemming from denial of the loss of the limb. The second was presented in 1978 by Solomon and Schmidt, explaining it as the attempt to deny the emotional response to the amputation. The third theory, asserted by Kolbe [1954], stated that the experience of painful phantom limb represents the emotional response to the amputation of a body part possessing significance in the amputee's relationships with others. The fourth theory, embraced at different times by different researchers [Frazier & Kolbe, 1970; Parkes, 1973; Solomon & Schmidt, 1978] claimed that painful phantom limb may be the result of unresolved mourning for the limb, by the amputee's fantasies about the limb, and over-valuation of the amputated appendage. For a review see Shukla et al. (1982).

All of these psychological explanations take a maladaptive approach in that phantom limb phenomena are seen to be the result of the amputee's inability to successfully adjust to the amputation. It is easy to see why this maladaptive approach might provide an acceptable explanation if the incidence figures are believed to be as low as 5- 1 0% of all amputee populations as was previously thought. However, such explanations are less feasible given current incidence figures as high as 85%. It is conceivable that these maladaptive factors may be involved in some cases of phantom limb phenomena, but certainly not all.

PhysiologicaL Although several researchers have acknowledged the existence of both painful and non-painful phantom limb [Dernham, 1986; Jensen et al., 1983; Sherman & Sherman, 1985; Sherman et al., 1984; Sherman et al., 1979; Shukla et al., 1982], only recently have attempts been made to separate the causes of each [Sherman, 1989]. There have been six major avenues of research into the possible physiological explanations for the varying phantom limb phenomena. The first to be addressed in this discussion will be the peripheral theory, followed respectively by the contribution of spinal mechanisms, the supraspinal mechanisms, the central theory, the vascular mechanisms, and the vascular-related mechanisms.

Peripheral theory proposes that impulses are generated in the stump which, when delivered to the central nervous system (CNS), are perceived as originating in the absent limb. Support for this theory was found in studies where electrical stimulation of the stump increased phantom limb [Carlen, Wall, Hadvoma & Steinbach, 1978]. In addition, local anesthesia of the stump stops phantom limb for the duration of the anesthetic's effectiveness [Carlen et al., 1978]. Another reason for the acceptance of this theory is that painful phantom limb has been associated in some cases with stump pain and problems in healing in the stump [Carlen et al., 1978; Dougherty, 1979; Gross, 1982; Parkes, 1973; Sherman & Sherman, 1985; Shernan et al., 1984; Sherman et al., 1979; Sunderland, 1978]. Another mechanism involving the stump will be discussed in the section on vascular-related mechanisms.

It has also been found that when major nerves are cut, tumors of nerve tissue - something akin to scarring - form at the cut ends [Carlen et al., 1978; Parkes, 1973]. These tumors, called neuromas, emit constant nerve impulses which may be responsible for the experience of phantom limb. In addition, decreases in the thickness of skin on the stump increase the aggravation to the neuromas [Sherman, 19891, possibly explaining why some amputees experience increases in painful phantom limb over time. The formation of neuromas takes time, so this cannot explain the immediate appearance of phantom limb as noted by Jensen et al. [1983], Sherman et al. [1979] and Shukla et al. [1982]. However, Carlen et al. [1978] also reported evidence of the quick development of spontaneous nerve impulses in the portions of newly cut neurons that are closest to the center of the body. This neural activity was also found to be either increased or decreased by pressure to the stump. This finding could explain why some amputees report that stump massage decreases painful phantom limb, while others report that such massages increase the sensations. It also may explain why some individuals report that the use of a prosthesis increases phantom limb while others report that their prostheses decrease the sensations. For a review see Jensen et al. [1983] and Shukla et al. [1982]

There has been less research concerning the spinal mechanisms involved in phantom limb. There are two basic avenues to this line of research: one focuses on the neurochemical changes that take place following the severing of neurons and the other focuses on the structural changes that result. One of the neurochemicals that has been found to show progressive depletion in animals as well as in humans is Substance P [Hunt, Rossop, Emson & Clements-Jones, 1982], which is believed to be associated with the transmission of pain-related information [Bloom, Lazerson & Hofstadter, 1985]. Decreased levels of this neurotransmitter could explain why there is often a reduction in the perception of painful phantom limb over time. Among the structural changes noted, it was found that cut neurons from the dorsal horns of the spinal cord exhibit a reduction of the primary cell depolarization [Wall & Devor, 1981]. This reduction would decrease the likelihood that the nerve would "fire", offering another explanation of why the phantom limb decreases in intensity over time. There is also evidence of the degeneration of the smaller neurons and spinal cord cells [Sherman, Sherman & Gall, 1980]. This could be another explanation of why the phantom limb begins to fade with the passage of time. In addition, studies have noted changes in the receptive fields of input deprived spinal cord neurons carrying information to the CNS [Carlen, et al., 1978]. The implications of this finding are more in keeping with the next approach: supraspinal mechanisms.

The contribution of supraspinal mechanisms is related to the spinal mechanisms in that changes in the receptive fields of severed neurons may indicate that they are receiving input from pathways that did not previously excite this area of cells [Dostrovsky, Millar & Wall, 1976]. It is quite possible that this change in the receptive fields could be the mechanism responsible for the eventual telescopy experienced over time by nearly all amputees [Jensen et al., 1983; Sherman & Shernan, 1985; Sherman et al., 1979; Sherman et al., 1984; Shukla et al., 1982].

Central theory was first proposed by Melzack in 1971 as a physiological explanation of painful phantom limb, but has been interpreted as a psychological explanation by Dawson and Arnold [1981]. For the purposes of this discussion the theory will be accepted as being physiological in nature, as was intended by the author. Shukla et al. [1982] state that central theory credits conscious processes with the perception of phantom limb, and that these processes are based on schematic representations, built over time. Although these conscious processes operate independently of the peripheral sensory mechanisms, the schematic representations have been built from their sensory input throughout life. Support for this schematic explanation has been provided in studies that show that locally anesthetizing peripheral nerves of an intact limb will result in the experience of phantom limb [Melzack & Bromage, 1973]. That is to say that when the brain is no longer receiving input from the anesthetized periphery, similar to an amputee's experience, phantom limb occurs. Additional support for the schematic explanation has been provided by subjects who report that "scratching" the phantom limb can relieve phantom itching [Jacome, 1978].

As proposed by Melzack [1971] the central theory asserts that within the brainstem is a mechanism which inhibits neural responses originating in the peripheral nervous system. Cessation of input would decrease inhibition, and in so doing, increase neural activity. Neurons at this level also have the ability to "recruit adjacent neurons", presumably for the purpose of duplicating function. This ability occurs at several synapses along the pathway, and is responsible for what Melzack called the "spread of trigger zones" on the cortex.

This explanation conflicts with the interpretation of central theory [Dawson & Arnold, 1981; Shukla et al. 1982] that was referred to above as the schematic explanation in that the central theory does not insist that the central processes involved act independently of the peripheral mechanisms. Melzack's central theory allows for the effect of peripheral impact by virtue of its absence. Support for this has been found in studies which showed the presence of lesions in the CNS of some patients has resulted in the disappearance of phantom limb [Apenzellar & Bicknell, 1969]. It is clear from these and other studies that there is CNS integration involved in the phantom limb, but the question remains as to whether or not the sensory pathways play a role in this integration.

Aligned with central theory is the belief that phantom limb phenomena, especially painful phantom limb, are a result of the cerebral structures forming a memory engram or schema of the pain receptors' experience of the damage that resulted in the loss of the limb [Jensen et al., 1983]. This belief is supported by studies that show a similarity in the nature and intensity of pain in the pre-surgical limb to the post-surgical painful phantom limb [Jensen et al., 1983; Shernan et al., 1984; Shukla et al., 1982]. These findings provide support for the idea that intensity is the main difference between painful and non-painful phantom limb if it is conceivable that the memory engram could vary in intensity as a function of stress or anxiety, a theory proposed by Sherman et al. [1984], as will be discussed under the heading of treatments.

More recent studies [Arena, Sherman, Bruno, & Smith, 1989; Sherman, 1989; Sherman, Arena, Sherman, & Ernst, 1989; Sherman & Barja, 1989; Sherman, Bruno, Sherman, Grana, & Arena, 1989] have shown that certain types of phantom pain - burning, throbbing, tingling, and, to a lesser degree, cramping - increased with decreased blood flow in the stump. This vaso-constriction can occur as a result of changes in skin thickness on the stump over time or as a result of tissue swelling in response to decreases in barometric pressure [Sherman, 1989]. Moreover, studies have shown that changes in weather affect phantom limb [Arena et al., 1989], as well as does ambient temperature (see Table 2).

In addition to vaso-constriction, another vascular-related mechanism apparently operates. Several studies reported cramping types of painful phantom limb preceded by "spike-like" muscle spasms in the stump [Sherman, 1989; Sherman, Arena, et al., 1989; Sherman, Bruno, et al., 1989]. Such muscle spasms act to decrease the blood flow, but since burning phantom limb perceptions occur in the presence of the vaso-constriction alone, it is probable that vaso-constriction is not solely responsible for the cramping phantom limb. Therefore more research is necessary in order to assess the contribution of the muscle spasms, themselves. These last two explanations mark the beginning of an attempt to isolate physiological mechanisms responsible for specific types of painful phantom limb. For a review see Sherman [1989], Sherman, Arena, et al. [1989], and Sherman and Barja [1989].

Psychological Versus Physical Causes. The emphasis of this discussion on the physiological causes is not without basis in that the implications of the maladaptive approach of the psychological explanations of painful phantom limb are similar to those of somatoform disorders as described by the diagnostic manual of the American Psychological Association. The somatoform disorders involve pain experiences in the absence of a stimulus. In order for a somatoform disorder diagnosis to be made, there must not be any discernible organic cause for the experienced pain [Rosenhan & Seligman, 1984]. By this most conservative criterion, the existence of studies supporting physiological mechanisms for phantom limb phenomena would dictate that psychological explanations be set aside when referring to phantom limb in general. This suggestion receives further support from studies [Sherman, 1989; Sherman & Bajar, 1989; Sherman, Sherman, & Bruno, 1987] which have found no evidence that amputees reporting painful phantom limb differed in the incidence of major personality disorders from amputees not reporting painful phantom limb as would be expected from the maladaptive explanations. Treatments

The one thing about which all of the studies agree is the lack of success that has been found with any of the attempted methods of treating painful phantom limb on a wide scale basis. The percentage of amputees reporting effectively treated painful phantom limb ranges from I to 30% [Sherman & Sherman, 1985; Sherman et al., 1979]. In 1984 Sherman et al. asked amputees to report treatments they had tried, and the results. The list of their responses can be found in Table 3.

Out of the approximately 3,400 amputees responding to the survey only 133 individuals reported any kind of permanent change, and only 6 individuals reported having found a cure for their painful phantom limb.

Far worse than the failure of the treatments offered these patients are the numbers of individuals who were offered no treatment at all. Only 20% were offered treatments. Forty percent were told that their pain was imaginary. The other 40% were either told that nothing could be done or that the pain would go away. Some who were told that the pain would disappear were painful phantom limb sufferers of twenty years or more! Perhaps the most disturbing finding as a result of this type of treatment from medical professionals is that many amputees reported that they had stopped talking to their doctors about their phantom limb for fear that they would be judged insane, or would lose the medical benefits needed for stump care, etc. [Sherman et al., 19841. Given this, it is less difficult to understand the low incidence statistics of the past.

In a more recent review, Sherman [1989] reported 30 types of physical treatments, 6 types of chemical treatments, and 11 types of psychological or other treatments, but none cited were effective for more than a few months at a time. The author also reported that surgery was not effective in reducing painful phantom limb, but that it was effective in the treatment of stump pain.

Some studies have reported partial reduction and even complete alleviation of painful phantom limb using small, specialized groups of patients, and they will be discussed in greater detail. One study [Urban, France, Steinberger, Scott & Maltbie, 1986] found that the combination of narcotics and antidepressants allowed the dosage to be kept low enough to prevent problems associated with addiction and tolerance and remain an effective treatment for painful phantom limb. One study [Gross, 1982] found that locally anesthetizing the contralateral limb was effective in producing long-term relief from painful phantom limb. In this study, the researcher used patient report to locate points of deviation in the skin's normal electrical charge on the limb contralateral to the amputation. Examination showed that these points of hyperalgesia exhibited a lower resistance to mild electrical current. It was therefore proposed that the painful phantom limb was a result of the disturbance of a functional symmetry present in a healthy body, and that the contralateral anesthesia worked to restore the symmetry.

Two of the studies were able to show success with relaxation training. The first, a case study using biofeedback-assisted relaxation of the stump muscles [Dougherty, 1980], resulted in pain-free periods of four to five hours and the ability to quickly stop episodes of painful phantom limb that awakened the patient nightly. The researcher suggested that there were two possible explanations for the success of this treatment; one was that the relaxation accompanied a decrease in sympathetic tone, thought to have a role in peripheral pain, and the second is that the relaxation served as a distractor. However, given more recent findings [Sherman, 1989; Sherman, Arena, et al., 1989], the more reasonable explanation would be that vasodilation would accompany relaxation, leading to increased blood flow, and a reduction of pain. The second study examining the effects of relaxation on painful phantom limb in chronic sufferers used a combination of progressive muscle relaxation, muscle tension feedback and reassurance [Sherman, Gall & Gormly, 1979]. After this combination of training a 1-3 year follow-up showed that 10 of the 16 subjects were pain-free, and of the remainder, all but two reported considerable reduction in the pain experienced. The conclusion reached by the researchers was not that painful phantom limb was caused by anxiety, but that anxiety exacerbated sensations that otherwise might have been viewed as non-painful. Although they asserted that the effectiveness of this treatment rests on its ability to break the pain-tension-anxiety cycle, the vascular constriction that would accompany anxiety-produced muscle tension would also explain the treatment's effectiveness when relaxation occurred.

Conclusion In conclusion it appears that after centuries of work in this area we are just beginning to scratch the surface of understanding phantom limb phenomena. From the preceding review it is clear that the success of future research depends on the pursuit of several avenues: * It is imperative that universally acceptable definitions of painful and non-painful phantom limb be developed in order for systematic study to be possible. The production of list after list of descriptive adjectives does not contribute to our understanding, therefore the next logical step in the process is to systematically reduce the present lists to more comprehensive components in order to develop useful measures of phantom limb phenomena. * Once this defining process has taken place a comparison between painful and non-painful phantom limb is necessary. If Sherman et al. [1984] are correct in their assertion that the sensations are the same for painful and non-painful phantom limb, differing only in intensity, then more research is needed to examine the factors that affect the intensity of these perceptions. If, on the other hand, painful and non-painful phantom limb are two completely different phenomena, the defining process should reveal this as well. * The third necessary avenue of research involves the careful investigation of the precipitating and relieving factors reported by the amputees. Not only could a better understanding of these help direct attention to the physiological mechanism underlying particular phantom limb phenomena, but it could also lead to relief for painful phantom limb sufferers. The process of linking descriptors, precipitating factors, and physiological mechanisms has begun in some recent studies [Arena, et al., 1989; Sherman, 1989; Sherman, Arena, et al., 1989; Sherman & Barja, 1989; Sherman, Bruno, et al, 1989] but much more refining is necessary. * In order for the tremendous inroads in the discovery of physiological causes for phantom limb phenomena to continue to be made we must reject- for the majority of cases-the notion of the maladaptive approach taken by the psychological theories of the past. High incidence statistics and review of the literature do not support such an approach, and failure to reject it at this time can only provide a stumbling block to the progress of phantom limb research. * Perhaps the most important point to be made in the focus of future research in this area is that health care professionals must be made aware of the prevalence of phantom limb in order to dispel many of the widely held misconceptions mentioned throughout this discussion. In the United States alone one out of every 300 individuals has had a major amputation, and the incidence of new cases is 30,000 each year [Dernham, 1986]. Painful phantom limb is not a small problem, and as the evidence presented in this discussion attests, it is not a problem that will go away. Amputees must stop being told that their pain is imagined if we are to expect the reliable reports of these sensations that are all too necessary for the proper examination of them. Until this can be accomplished, little progress in the study of phantom limb phenomena can be expected.

Acknowledgments

The author is grateful to Dr. Wayne Harrison, Dr. Jorge Rodrigues-Sierra, and Dr. James Thomas for their assistance and advice in preparing this review. hi addition, appreciation is owed to Dr. Richard. A. Sherman, and several anonymous reviewers for their comments on an earlier version of the paper.

References Apenzellar, O. & Bicknell, J. M. (1969). Effects of nervous

system lesions on phantom experience in amputees. Neurology,

19, 141-146. Arena, J. G., Sherman, R. A., Bruno, G. M., & Smith, J.D. (1989).

The relationship between humidity level, temperature, and

phantom limb pain: Preliminary analysis. Proceedings of the

annual meeting of the Association for Applied Psychophysiology. Bloom, F. E., Lazerson, A., & Hofstadter, L. (1985). Brain, mind

and behavior. New York: W. H. Freeman & Company. Carlen, P. L., Wall, P. D., Hadvoma, M. D. & Steinbach, T.

(1978). Phantom limb and related phenomena in recent traumatic

amputations. Neurology, 28, 211-217. Dawson, L. & Arnold, P. (1981). Persistent phantom limb pain.

Perceptual and Motor Skills, 53, 135-138. Dernham, P. (1986). Phantom limb pain. Geriatric Nursing, I,

34-37. Dostrovsky, J. O., Millar, J. & Wall, P. D. (1976). The immediate

shift of afferent drive of dorsal column nuclear cells following

deafferentiation: A comparison of acute and chronic deafferentiation

in gracile nucleus and spinal cord. Experimental Neurology,

52, 480-495. Dougherty, J. (1979). Relief of phantom limb pain after EMG 

biofeedback-assisted relaxation: A case report. Behavior Research

and Therapy, 18, 355-357. Dworetzky, J. P. (1985). Psychology. St. Paul: West Publishing

Company. Frazier, S. H. & Kolbe, L. C. (1970). Psychiatric aspects of

phantom limb. Orthopedic Clinics of North America, 1, 48 1 - 495. Gross, D. (1982). Contralateral local anesthesia in the treatment

of phantom limb and stump pain. Pain, 13, 313-320. Hunt, S. P., Rossop, M. N., Emson, P. C. & Clements-Jones, V.

(1982). Substance P and enkaphalins in the spinal cord after

limb amputation. Lancet, I, 1023. Jacome, D. (1978). Phantom itching relieved by scratching phantom

feet. Journal of the American Medical Association, 40,

2432. Jensen, T. S., Krebs, B., Nielsen, J., & Rasmussen, P. (1983).

Phantom limb, phantom pain and stump pain in amputees

during the first six months following limb amputation. Pain,

17, 243-256. Kolbe, L. C. (1954). The painful phantom: Psychology, physiology

and treatment. Springfield, IL.: Charles C. Thomas. Malin, J.P. & Winkelmuller, W. (1985). Phantom phenomena

(phantom arm) following cervical root avulsion: Effect of

dorsal root entry zone thermocoagualtion. European Archives

of Psychiatry and Neurological Sciences, 235, 53-56. Melzack, R. (1971). Phantom limb pain: Implications for treatment

of pathologic pain. Anesthesiology, 35, 409-419. Melzack R. & Bromage, R. R. (1973). Experimental phantom

limb. Experimental Neurology, 39, 261-269. Mitchell, S. W. (1872). Injuries of nerves and their consequences.

Philadelphia: J. B. Lippincot. Parkes, C. M. (1973). Factors determining the persistence of

phantom pain in the amputee. Journal of Psychosomatic Research,

17, 97-108. Rosenhan, D. L., & Seligman, M. P. E. (1984). Abnormal psychology.

New York: W.W. Norton and Company. Rosenzweig, M. R. & Leiman, A. L. (1982). Physiological

Psychology. Lexington, Massachussetts: D.C. Heath & Company. Sherman, R. A. (1989). Stump and phantom limb pain. In (R.

Portenoy, Ed.) Neurologic Clinics of North America, 7, 249 - 263. Sherman, R. A., Arena, J. G., Sherman, C. J. , & Ernst, J. L.

(1989). The mystery of phantom pain: Growing evidence for

psychophysiological mechanisms, Biofeedback and Self-regulation,

(in press). Sherman, R. A. & Bala, R. H. (1989). Postamputation and

phantom-limb pain. In (K. Foley and R. Payne, Eds.) Current

therapy of pain. Ontario, Can.: B. C. Decker, Pub. Sherman, R. A., Bruno, G. M., Sherman, C. J., Grana, A., &

Arena, J. G. (1989). Relationships between muscle contraction 

patterns in the residual limb and phantom limb pain: An initial

evaluation. Manuscript submitted for publication. Sherman, R. A., Gall, N., & Gormly, J. (1979). Treatment of

phantom limb pain with muscular relaxation training to disrupt

the pain-anxiety-tension cycle. Pain, 6, 47-55. Sherman, R. A., & Sherman, C. J. (1985). A comparison of

phantom sensations among amputees whose amputations were

of civilian and military origins. Pain, 21, 91-97. Sherman, R. A., Sherman, C. J., & Bruno, G. M. (1987). Psychological

factors influencing chronic phantom limb pain: an

analysis of the literature. Pain, 28, 285-295. Sherman, R. A., Sherman, C. J. & Gall, N. G. (1980). A survey

of current phantom limb treatment in the United States. Pain,

8,85-99. Sherman, R. A., Sherman, C. J., & Parker, L. (1984). Chronic

phantom and stump pain among American veterans: Results of

a survey. Pain, 18, 83-95. Shukla, G. D., Sahu, S. C., Tripathi, R. P., & Gupta, D. K. (1982).

Phantom limb: A phenomenological study. British Journal of

Psychiatry, 141, 54-58. Solomon, G. F. & Schmidt, K. M. (1978). A burning issue:

Phantom limb pain and psychological preparation of the patient

for amputation. Archives of Surgery, 113, 185-186. Sunderland, S. (1978). Nerves and nerve injuries. Edinburgh:

Churchill-Livingstone. Urban, B. J., France, R. D., Steinberger, E. K., Scott, D. L., &

Maltbie, A. A. (1986). Long-term use of narcotic/antidepressant

medication in the management of phantom limb pain.

Pain, 24, 191-196. Wall, P. D. & Devor, M. D. (1981). The effect of peripheral nerve

injury on dorsal root potentials and on transmission of afferent

signals into the spinal cord. Brain Research, 209, 95-1 1 1. Received: October 1989 Revised: February 1990 Accepted: March 1990 MARGARET S. BOWSER, University of Nebraska at Omaha, Department of Psychology, 60th and Dodge, Omaha, Nebraska 68182. TABULAR DATA OMITTED