Mapping the Phantom Limb; S1 and M1 Plasticity Implications in Phantom Limb Pain

Reports of phantom limb phenomena have been recorded for centuries. Amputees sensed a missing limb as though it were still a part of their bodies, executing controlled movements and experiencing normal skin sensation. Early theorists attributed this to residual neural memory. Recent studies of patients with congenitally absent limbs challenge this notion. Indeed, congenital aplastics experienced phantom limb sensation despite lacking neural memory. Thus, investigators have turned to neuroimaging techniques to discern the neural correlates of phantom limb sensation. Both monkey and human studies show reorganization of primary motor (M1) and somatosensory (S1) cortex following limb loss. Often, muscles near the amputation site respond to deefferented M1 stimulation. Additionally, cutaneous stump stimulation often activates deafferented S1. Experiments have also indicated that face activity invades silent M1/S1 regions. M1/S1 plasticity has exhibited similar pathways; thus, experimenters

Doctors and scientists have long held the view that the adult brain is relatively stable in its functional organization. Since patients with discrete brain lesions exhibit particular cognitive or functional losses, neuroscientists have concluded that each section of the brain has a localized function and, thus, any focal injury results in an irreparable deficit (Kaas, 2001). In recent years, newer neuroimaging techniques, as well as other modern technologies of neural observation, have challenged the belief in absolute functional stability within the adult brain (Brugger, Kollias, Müri, Creher, Hepp-Reymond, & Regard, 2000; Lotze, Flor, Grodd, Larbig, & Birbaumer, 2001). Indeed, the brain's ability to assimilate knowledge is indicative of a certain degree of malleability in neural pathways. This plasticity, however, is not limited to higher cognitive functions. Studies of both brain damaged and amputee patients have shown that, even at the somatosensory and motor levels, brain structures may adjust their functions over time. This paper examines the shift in primary motor and somatosensory cortex function affected by limb loss in amputee and congenitally aplastic subjects.

Patients who are missing extremities (i.e. fingers, toes, hands, feet, arms and legs), whether from birth or traumatic injury, often express the experience of phantom limb phenomena (Brugger et al., 2000; Lotze et al., 2001). This includes, but is not limited to, sensation within the limb, the ability to control phantom limb movement, and uncontrollable phantom limb pain (PLP). The phenomena have been reported for centuries, but only recently has modern technology allowed cognitive neuroscientists to delve into the causes and possible treatments for such sensation.

Structure and organization of the normal human M1 and S1

This paper will focus on the primary motor and somatosensory cortices, located on the precentral and postcentral gyri of the human brain. The primary motor cortex lies just anterior to the central sulcus and extends across the precentral gyrus. This posterior strip of frontal cortex extends its axons into the spinal cord and synapses with motor neurons that generate movement. Posterior to the central sulcus, along the frontal gyrus of the parietal lobe, is the postcentral gyrus or somatosensory cortex. The primary somatosensory cortex (S1) receives cutaneous stimulation information about touch, pain, temperature and limb position (Gazzaniga, Ivry, & Mangun, 2002). These gyri express pivotal functional reassignment following limb amputation.

Neuroimaging and surgical procedures have afforded neuroscientists a functional topographical map of both M1 and S1. In normal brains, studies have regularly observed cortical regions corresponding to specific areas on the body. The general organization of both M1 and S1 begins most medially with the lower digit representational area and extends more laterally with lower limbs, trunk, arms, hands, fingers, face and tongue. In both gyri, the face responsive region has a disproportionately large mouth area and terminates with the tongue representation most laterally. The size of the activated S1 region for each body part is directed by the sensitivity to stimulation of that body part. Thus, the disproportionately large lip and hand representations indicate that more cortex and neurons are devoted to lip and hand sensation when compared to arm or leg regions. Adjacent trunk, arm and face representations are critical areas of reorganization patterns following upper limb amputation. Similar effects are seen with neighboring lower limb areas in leg amputees.

Primary motor and somatosensory reorganization in patient A.Z.

An early theory of phantom limb sensation posited that perceptuomotor memory of the lost extremity resulted in the illusion of limb existence (Brugger et al., 2000). If this residual motor and somatosensory memory were indeed the cause of such phenomena, one would then expect that a patient congenitally missing certain extremities would not experience phantom limb sensation. The opposite is observed in patient A.Z., a 44-year-old woman born without forearms and legs (Brugger et al., 2000). Despite the fact that she never actually experienced sensation of these limbs, A.Z. perceived mental images of her missing limbs since early childhood. She experienced these sensations as though the limbs were actually a part of her body. Thus, the existence of congenitally aplastic phantoms supports the hypothesis that innate body schema, or a genetically derived sense of one's limbs, is involved and embedded in brain organization.

Using neuroimaging and neurophysiological techniques, scientists mapped A.Z.'s phantom limb sensation to uncover the neural correlates of her cortical reorganization (Brugger et al., 2000). They recorded functional magnetic resonance imaging (fMRI) while A.Z. imagined moving her phantom limbs and actually moved intact body parts (i.e., stumps, cheeks, tongue and eyelids). Experimenters expected to observe activation in the cortical areas for hand and finger movement when A.Z. moved body parts adjacent to deefferented M1 receptive areas. Additionally, transcranial magnetic stimulation (TMS) mapped A.Z. sensorimotor cortex. Motor evoked potentials (MEPs) were recorded from her deltoid muscles while she reported any sensation in stumps, phantom limbs or both.

fMRI analysis showed that phantom finger movements produced bilateral activation of the dorsal premotor cortex at the junction of the superior frontal and precentral sulci. Also activated were mesial premotor regions and bilateral superior posterior parietal cortex along the intraparietal sulcus. During imagined hand and finger movement, there was no observed activation of the region of primary motor cortex (M1) dedicated to hand representation in normals. Additionally, facial movement fMRI data recorded M1 activation only in regions associated with face movement (Brugger et al., 2000). Since the cortical hand representation region of M1 remained silent when neighboring regions were stimulated, this case study failed to confirm the expected theory that neurons controlling face regions invade nearby deefferented motor cortex. Experimenters concluded that this discrepancy between patient A.Z. and findings from traumatic amputee studies indicates a possible moderate influence of sensorimotor memory in phantom limb phenomena. TMS activation of M1 extending into the deefferented hand and foot regions did, however, result in stump movement. Left stump movement resulted in contralateral hemispheric activation, while right stump movement caused bilateral stimulation. Similarly, during the TMS phase, A.Z. felt phantom limb movement in the limb contralateral to the stimulated S1 cortex.

Experimenters found certain basic differences between A.Z. cortical reorganization and that observed in traumatic amputee patients. While repetitive imagined finger movement in traumatic amputee patients resulted in contralateral M1 finger area activation (Lotze et al., 2001), this effect was non-existent in the congenitally aplastic patient (Brugger et al., 2000). This indicates that sensorimotor memory participates in the experience of phantom limb phenomena. In patients with congenitally absent limbs, there is no residual M1 or S1 memory; yet illusory limb movement and sensation are present. Thus, it is logical to conclude that many factors, both innate and experiential, contribute to cortical functional organization.

Traumatic injury and temporal passage as factors in reorganization

Both the amount of limb stimulation prior to amputation and the elapsed time post limb-loss play a role in the reorganization of cortical structures. Qi, Stepniewska and Kaas (2000) examined cortical reorganization (of M1) in adult macaque monkeys years after forelimb or hindlimb amputation due to traumatic injury. Two of the monkeys studied were injured as adults, one as a juvenile and one as an infant. Muscle movements elicited through intracortical microstimulation (ICMS) of distinct M1 regions were observed in amputee and normal macaque monkeys. In each of the monkeys, stimulation of regions devoted to the missing limb resulted in movement of more proximal remaining body parts (Qi, Stepniewska, & Kaas, 2000). Though all subjects exhibited similar patterns of M1 reorganization, ICMS results indicated that early injury combined with a more limited amputation (i.e., finger removal as opposed to entire hand or forearm loss) promote greater cortical reorganization with fewer ineffective M1 sites (Qi, Stepniewska, & Kaas, 2000). The observed reorganization of deefferented M1 regions across numerous life stages indicates that cortical function, though less malleable in later life, always retains some degree of plasticity.

M1 reorganization in traumatic injury amputees

The M1 reorganization observed in Qi, Stepniewska, and Kaas' study (2000) has been supported by numerous TMS and ICMS experiments on human and non-human primate (i.e., monkey and galagos) amputees. Primary motor cortices of squirrel monkeys and galagos have four distinct functional sections hindlimb (i.e., leg and tail), trunk, forelimb (i.e., forelimbs and digits) and orofacial (i.e., face and mouth). The hindlimb region of M1 is most medial, extending laterally into the trunk, forelimb and, most laterally, orofacial areas. In an experiment similar to the macaque monkey study (Qi, Stepniewska, & Kaas, 2000), three squirrel monkeys and two galagos forelimb and hindlimb amputees were observed during M1 ICMS (Wu and Kaas, 1999).

Results showed that microstimulation of deefferented M1 regions evoked movement in the contralateral stump and more proximal remaining limb. Though no unresponsive zones of deefferented cortex were observed, much higher currents to M1 were required to evoke movement than in the contralateral, intact limb, M1. These data imply that the reorganized cortex threshold is elevated in comparison to efferented, normal M1. Reassigned cortex is less sensitive to stimulation, thus contributing to the fickle and transient nature of phantom limb sensation.

Also, in the forelimb amputee, ICMS of forelimb M1 evoked movement in the stump, shoulder, trunk and orofacial muscles (Wu and Kaas, 1999). These data seemingly contradict Brugger et al.'s (2000) case study that did not find facial representation invading deefferented cortex. The results of these two studies, however, are not incompatible; rather, it may be understood that injury during different developmental stages, of varying severity and across species results in distinct reorganization patterns.

Additionally, stimulation of deefferented forelimb M1 cortex resulted in elicited movement of the shoulder and stump in about 80% of tested regions. In only the remaining 20% of test sites were trunk and orofacial movement observed (Wu and Kaas, 1999). Similar results are seen in the subjects with hindlimb loss, except the surrounding tail and lower trunk areas extend into the hindlimb cortex. The small percentage of trunk and orofacial stimulation evoked by deefferented forelimb cortex hints at the reduced prevalence of such reorganization.

Tantamount to these animal studies were human focal TMS experiments examining electromyographic (EMG) responses in muscles proximal to the forearm amputation site (Röricht, Meyer, Niehaus, & Brandt, 1999). In order to determine the existence and extent of M1 plasticity in humans, researchers observed fifteen participants exhibiting different levels of amputation (forearm, upper arm and shoulder). All subjects had undergone amputation 20 to 65 years prior to the study and each exhibited cortical reorganization of deefferented M1.

Reorganization patterns, however, varied according to the level of amputation. In the forearm amputee, the forearm region of contra-amputational M1 stimulated the biceps near the stump. Those with upper arm and shoulder level amputations exhibited increased stimulation of the nearby deltoids and trapezoids in response to contralateral activation of M1 forearm areas (Röricht et al., 1999). Upper arm and shoulder amputees had lost the biceps on the amputation side, explaining why M1 reorganization does not include the biceps. On the other hand, lower arm amputees had remaining deltoid and trapezoid muscles, yet the biceps gained the majority of deefferented responsiveness. Thus, the difference in organizational pattern of deefferented M1 forearm cortex must be directed and influenced, to some extent, by the level of amputation.

As both human and primate studies showed similar cortical reorganization patterns following amputation (Qi, Stepniewska, & Kaas, 2000; Wu and Kaas, 1999; Röricht et al., 1999), it may be concluded that limb amputation results in primary motor cortex reorganization. Regardless of species, this includes evoked stump movement during amputated limb cortex activation. Many factors, including species, age at time of injury, elapsed time since amputation and extent of limb loss (Qi, Stepniewska, & Kaas, 2000; Wu and Kaas, 1999; Röricht et al., 1999), determine the pattern and extent of altered evoked M1 response.

S1 reorganization in traumatic injury amputees

Although primary motor cortex reorganization explains the ability to move phantom limbs, this does not explain the initial phantom limb sensation. Understanding the need to examine amputation effects on deafferented somatosensory cortex (S1), Wu and Kaas (1999) observed S1 activation from cutaneous stimulation in galagos. Different areas of the stump, shoulder, chest, neck and lower face of forelimb amputee galagos were stimulated using a fine probe to isolate distinct receptive areas. By stimulating such minute body regions, a detailed and precise map of receptive fields could be recorded. Microelectrode recordings of S1 activation indicated the receptive field of that cortical region. A receptive field was defined as the skin area where near-threshold stimuli evoked S1 response. In areas where gentle cutaneous stimulation did not evoke S1 activity, joint manipulation was employed. Thus less sensitive skin regions were stimulated through more intense movements and could be mapped, with a higher threshold, to the S1 cortex.

Just as M1 reorganization of deefferented cortex left almost no unresponsive zones (Wu & Kaas, 1999), S1 reassignment occurred in nearly all deafferented regions. Most S1 areas devoted to forelimb sensation in the hemisphere contralateral to amputation responded to light, cutaneous stimulation of the stump and adjoining shoulder. Also similar to M1 data, some areas were receptive to chest, neck and lower face stimulation, but were highly infrequent. The hindlimb amputee exhibited a related pattern, with lower body parts (i.e., stump, tail and lower trunk) claiming control of deafferented S1 regions. Reorganization of S1 did not extend beyond the borders of deafferented regions into normal cortex (Wu and Kaas, 1999). These data indicated that, in galagos, both M1 and S1 reassignment following limb amputation produce nearly equivalent activities.

Additionally, neuroelectric source imaging of somatosensory cortex in human amputees further confirmed the galagos experimental data. In a comparable human arm amputee study (Karl, Birbaumer, Lutzenberger, Cohen, & Flor, 2001), cutaneous stimulation was applied to the first and fifth digits of the intact hand and both corners of the lower lip (amputation and non-amputation sides). Due to the proximity of lip and hand representations on the sensorimotor cortices, the lower lip was selected for this study.

While applying these stimuli, electrodes in a cap on the subject's head recorded somatosensory evoked potentials (SEPs). Comparison of activated regions in both hemispheres provided interesting results. Through stimulation of the intact hand, an S1 hand representation area was determined for each subject. This area was used to delineate hand representation area of deafferented cortex and data indicated that some hand area was responsive to mouth stimulation. Though lip stimulation invaded few regions of deafferented cortex, there was a statistically significant increase of such areas in patients reporting phantom limb pain (PLP) (Karl et al., 2001). While this experiment indicates similar S1 reorganization in non-human primates (i.e., galagos) and humans, the human ability to communicate experience affords experimenters functional data that may explain how cortical reorganization is consciously manifest.

Implications of similar M1 and S1 reorganization patterns

It is clear from the above analyses of M1 and S1 reorganization following limb amputation that certain patterns emerge in both cortices. As such, it is necessary to determine what is causing the similar reassignment of deefferented and deafferented cortex. One theory is that connections between the somatosensory and motor cortex synchronize organizational activities. The somatosensory cortex has definite projections into motor cortex, specifically into layers II and III of M1. These layers are closely associated with motor cortex layer V output neurons (Chen, Corwell, Yaseen, Hallett, & Cohen, 1998; Karl et al., 2001). Studies have shown that stimulation of somatosensory cortex can generate long-term potentiation (LTP) in motor cortex (Chen et al., 1998). When certain stimuli cause greater synaptic strength along a specific pathway, thus generating greater postsynaptic responses following subsequent stimulation, it is referred to as LTP. Projections from somatosensory cortex to motor cortex have also been implicated in some motor skill learning (Chen et al., 1998). Based on the observed concordance between M1 and S1 reassignment and neural connections indicated through scientific study, it may be concluded that deefferented M1 reorganization results from deafferented S1 region reassignment.

A comparative TMS study of M1 and neuroelectric source imaging of S1 in arm amputees (Karl et al., 2001) provided additional evidence for this theory. Results indicated a medial displacement of face area S1, specifically regarding lip representation. This was positively correlated with medial displacement of face motor representations (in M1) of the zygomaticus muscle located near the corners of the mouth. Such pronounced concordance between M1 and S1 plasticity strongly supports the theory of connectedness between the two areas.

Also confirming the connection between S1 and M1 cortex were fMRI studies of unilateral arm amputees hand representations in both hemispheres (Lotze et al., 2001). Amputees performed movements with the intact hand and imagined equivalent movements of the phantom hand. As expected, during imagined hand movement, fMRI analysis showed M1 and S1 activation in the hand area contralateral to the amputation. Even more striking were the observations of primary motor and somatosensory cortices in amputees when compared with normal controls. In the control group hand movement task, greatest M1 and S1 activation was observed contralateral to the dominant hand. Surprisingly, amputee subjects M1 and S1 activation patterns of intact hand cortex matched those found in the dominant hemisphere of the control group. Dominance in amputees was unaffected by hand dominance prior to injury and did not depend on which hand (dominant or non-dominant) was amputated (Lotze et al., 2001). In every case studied, reorganization of motor and somatosensory cortex devoted to the intact hand allowed it to function cortically as would a dominant hand in normals. Similarly, monkey studies show that, following sensory training, subjects are more cortically and physically sensitive to finger stimulation in the trained receptive fields (Kaas, 2001). Functionally, this is exhibited in patients who learn to write and perform other daily activities with the non-dominant, remaining hand following amputation. These data provided evidence that, following limb loss, even normally afferented and efferented amputee S1 and M1 are subject to change. This implies that plasticity is, to a certain degree, dependent on limb use. The equivalence between observed changes in both types of cortex further confirmed connection theories.

What is phantom limb pain?

Phantom limb pain (PLP) is reported in 50% to 80% of all amputation patients (Lotze et al., 2001). It is often described as cramping pain similar to hyperflexion of intact digits that dig into the surface of the palm or sole of the foot. Patients also describe burning, stabbing and crushing sensations in the phantom limbs (Bharwani, Rajagopal, & Ray, 2003). PLP can range from an annoying sensation to unbearable pain and is experienced in both short and extended durations. It is unclear exactly what causes PLP, but studies have concluded that it is not directly related to age, sex, region amputated or elapsed time since amputation (Bharwani et al., 2003).

Numerous fMRI and TMS studies of human amputees have found certain neural correlates reliably connected to the experience of phantom limb pain. General results of three neuroimaging studies all show a high correlation between PLP reports and increased M1 and S1 reorganization (Karl et al., 2001; Lotze et al., 2001; Flor, Elbert, Knecht, Wienbruch, Pantev, Birbaumer et al., 1995). fMRI comparison of upper limb amputees, with and without phantom limb pain, showed a basic difference between the two groups M1/S1 reorganization. Only participants reporting PLP exhibited a shift of the neighboring lip representation area into silent M1 and S1 hand regions (Lotze et al., 2001; Karl et al., 2001). Additionally, the magnitude of medial displacement of the lip region was highly correlated to the strength of PLP (Lotze et al., 2001). In one study (Flor et al., 1995), the calculated correlation between the amount of cortical reorganization and magnitude of PLP was r = + 0.93. Statistically, correlation is rated on a continuum from no correlation = 0) to correlations of greater strength culminating at r = 1. Thus, the near linear positive correlation found by Flor et al. (1995) indicates that the two variables have a strong direct relationship. Though correlation does not imply causation, it is difficult to ignore the strong relationship implied by these data. It appears that the simultaneous activation of hand and mouth representation areas of amputee patients co-occurs with, and even perhaps results in, PLP.

Viewed from another perspective, Karl et al. (2001) raised the point that those who reported less PLP exhibited less cortical reorganization. In this TMS and neuromagnetic source imaging experiment, S1 and M1 activity was examined in 5 PLP arm amputees and 5 non-PLP subjects. Only the participants reporting PLP exhibited medial displacement of the face area in S1 and of the zygomaticus and depressor labii inferioris muscle regions of M1. These studies all support the distressing hypothesis that greater plasticity in amputee cortex is, in the case of PLP, maladaptive.

Proposed PLP treatment methods based on the neurophysiological basis hypothesis

Since the exact cause of PLP remains unclear, it is difficult to develop a treatment that effectively treats the adverse effects of phantom limb phenomena without additional, unwanted side effects. Procedures and medications accepted for PLP treatment do not help all patients and often have other drawbacks, for example, "dead" phantom limb sensation (Roux, Ibarrola, Lazorthes, & Berry, 2001). Some of the more common PLP medications are carbamazepine, ketamine and tricyclic anti-depressants (Roux et al., 2001a), which act as painkillers. Surgical procedures such as S1 area removal, thalamotomies, deep brain stimulation, and chronic spine stimulation attempt to pinpoint the neural correlates of PLP and counter their actions. Chronic motor cortex stimulation, a procedure that has been used for years as a pain reduction method in post-stroke and trigeminal neuralgia patients, was recently investigated for use in PLP control (Roux et al., 2001a; Roux et al., 2001b).

Researchers (Roux et al., 2001a; Roux et al., 2001b) believe that chronic motor cortex stimulation somehow enhances M1/S1 cortex inhibition and this controls PLP. Different hypotheses have been proposed regarding the mechanism of action that causes such inhibition. Some believe that chronic M1 stimulation inhibits the responsiveness of hyperactive deafferented neurons, while others have proposed that the influence of pain inhibiting S1 neurons is increased following constant stimulation. Furthermore, some studies indicated that chronic motor cortex stimulation activates corticospinal neurons that then reduce PLP phenomena. Additionally, a fourth school of thought suggests that this procedure increases the activity of the brainstem and thalamus, thus inhibiting nociceptive reflexes (Roux et al., 2001a). The numerous theories and divergent mechanisms of action proposed for this treatment highlight the complexity of the phantom limb pain phenomenon and its treatment.

Research correlating PLP with M1/S1 activity (Karl et al., 2001; Lotze et al., 2001; Flor et al., 2001) has contributed to neuroscientists and doctors positive attitudes regarding the use of chronic motor cortex stimulation in amputees. This invasive treatment modality is proposed only for patients with intractable pain unresponsive to drug treatment. Four such patients were administered chronic motor cortex stimulation in two experimental studies. One patient described the phantom pain as a snake constantly biting his phantom hand, accompanied by unbearable clenching spasms (Roux et al., 2001a). The severe pain was debilitating for all patients included in the two studies. Postoperatively, the subjects were tracked to determine the procedure's success or failure (Roux et al., 2001a; Roux et al., 2001b).

Preoperatively, patients could voluntarily move the phantom limb. fMRI data of stimulated M1 during virtual movement directed subsequent electrode placement. Each participant had an electrode plate inserted below the skull and secured above the dura. The plate was positioned according to the stimulation patterns unique to each amputee and had specific stimulation parameters set based on individual pain control needs (Roux et al., 2001a; Roux et al., 2001b). In one patient, satisfactory pain control was recorded merely three days after surgery. Though the pain was reduced, the phantom limb sensation did not disappear. Rather, the subject felt as though the missing arm were "dead," hanging limp and stiff at his side (Roux et al., 2001a). He was troubled by his lost ability to evoke controlled imaginary movement of the phantom limb.

Successful pain control continued for the first three months following the procedure, however, PLP returned and stimulation parameters were adjusted to regain stable control. Reports ten months later indicated continued successful pain control (Roux et al., 2001a). The necessary adjustment of initial stimulation parameters may indicate procedural initiation of cortical plasticity. Though the parameters may be altered to reduce PLP when needed, the experimental appearance of reorganization indicates that this treatment may only yield temporary PLP relief.

Conclusion

Until recently, there was general scientific consensus that the mature brain is relatively rigid in its functional organization. With the advent of modern neuroimaging techniques, studies have provided sufficient evidence to support a theory of adult brain plasticity. Significant functional reorganization of S1 and M1 cortex in amputee subjects has been identified in numerous experiments.

Data indicate that many factors influence the degree and pattern of cortical reassignment. Though the early theory of perceptuomotor memory was displaced by the appearance of phantom sensation in congenitally aplastic patients, subsequent evidence points to combined genetic and experiential influences in neural adaptation. Additionally, age at amputation, time elapsed since injury, and the extent of amputation all contribute to the final path of M1 and S1 reorganization.

Many neuroimaging studies have shown that primary somatosensory and motor cortex reassignment are pivotal in the development of phantom limb phenomena. When a limb is amputated, M1 and S1 are affected in functionally opposite manners. While amputated extremities no longer send sensory signals to S1 regions, M1 areas cannot send efferents because the receptive motor neurons have been lost. Despite these basic differences, the topographical mapping of representation areas and neural connections between M1 and S1 lead to similar reorganization patterns.

On both M1 and S1 gyri, the arm/hand region is abutted on either side by face and trunk representational areas. One theory explaining the invasion of stump, trunk and orofacial regions into amputated-hand cortex (Brugger et al., 2000; Qi, Stepniewska, & Kaas, 200; Wu & Kaas, 1999) posits that when synaptic efferents and afferents are removed, synapses with neighboring neurons are strengthened. This theory also accounts for lower trunk area expansion in lower limb amputees.

Orofacial and trunk receptive areas were observed in only 20% of forelimb deefferented M1. The other 80% of test regions resulted in stump and shoulder movement (Wu & Kaas, 1999). Both invading representational areas are adjacent to the silent M1 region. The small ratio of trunk/orofacial to stump/shoulder appropriation of forelimb M1, however, indicates possible adaptive and use-dependent factors influencing reorganization. Patients who have lost a hand require more specialized stump movement for the completion of daily activities and, ultimately, for survival. It is unclear whether the brain senses limb loss and adapts accordingly or if increased stump use strengthens synaptic connections that were previously inhibited. This same question may be posed regarding the hand dominance effect observed in amputees. Though a unilateral arm amputee's intact hand exhibits an activation pattern equivalent to the dominant hand in normals, whether use determines dominance or vice versa has yet to be determined.

Experimenters have shown use-dependent plasticity in normal subjects, extrapolating from these data the hypothesis that amputee cortical reorganization results from a similar mechanism. A common observation in motor and perceptual skill acquisition is that, as training improves task performance, the quantity of neurons devoted to the task increases (Kaas, 2001, as reported by Li, Walters, McCandlish, & Johnson, 1996). These findings indicate that repetitive practice stimulates plasticity in normal brains. If, in normal brains, task repetition reassigns neurons to a specific function, then it is reasonable that amputees' increased dependence on stump and intact hand movements directs M1/S1 reorganization.

A major breakthrough in the study of phantom limb pain's neural origins was data correlating M1/S1 reorganization and PLP magnitude (Lotze et al., 2001). Only patients reporting PLP exhibited medial displacement of the lip representation area into silent M1/S1 hand cortex (Lotze et al., 2001; Karl et al., 2001). Based on these data, experimenters proposed the use of chronic motor cortex stimulation in PLP treatment. fMRI determined receptive imagined hand movement areas and these images directed electrode placement on the dura. This surgical procedure is highly invasive and dangerous, thus it is prescribed only in the most severe PLP cases. Even so, following this procedure, PLP control is not guaranteed. In fact, early data indicate that neural reassignment may be induced by chronic M1 stimulation. If cortical plasticity renders original electrical pulses insufficient for PLP regulation, then the stimulation parameters can be reset.

Since treatment of PLP with chronic stimulation has only recently been studied, it remains unclear whether stimulation parameter readjustment will prove adequate over long periods of time. Short-term results from the experimental treatment are encouraging. However, in order to determine their long-term validity, postoperative data must be reported for many more years. Also, to delineate and gauge the universality of such experimental findings, scientists must repeat the study in many more patients. Only then will scientists be able to venture more detailed theories as to the neural mechanisms underlying phantom limb sensation and PLP.

While phantom limb sensation studies have shown that some neural plasticity is maladaptive, as in the case of PLP, some changes are adaptive. For example, the hand dominance effect observed in unilateral arm amputees allows patients a functional advantage if the remaining hand was originally non-dominant. Since each neuron has many synaptic connections, it is inevitable that a strengthened, or new, connection will have repercussions extending beyond the single synapse. The difficulty appears when neuroscientists attempt to disentangle adaptive from maladaptive cortical reassignment. As technology evolves and more studies of cortical plasticity are performed, neuroscientists will continue to unravel the pathways governing functional cortical reassignment. Once the true neural correlates of plasticity are understood, clinicians will be able to develop treatments enhancing the beneficial aspects of cortical reorganization, while minimizing and preventing unwanted ones.


References
Bharwani, I., Rajagopal, A., Ray, J. (2003). Use of Calcitonin to treat phantom limb pain. Hospital Physician, 46-50.
Brugger, P., Kollias, S., Müri, R., Crelier, G., Hepp-Reymond, M.-C., & Regard, M. (2000). Beyond re-membering Phantom sensations of congenitally absent limbs. PNAS, 97:11, 6167-6172.
Chen, R., Corwell, B., Yaseen, Z., Hallett, M., Cohen, L., (1998). Mechanisms of cortical reorganization in lower-limb amputees. The Journal of Neuroscience, 18:9, 3443-3450.
Flor, H., Elbert, T., Knecht, S., Wienbruch, C., Pantev, C., Birbaumer, N., Larbig, W., & Taub, E. (1995). Phantom-limb pain as a perceptual correlate of cortical reorganization following arm amputation. Nature, 375, 482-484.
Gazzaniga, M.S., Ivry, R.B., and Mangun, G.R. (2002). Cognitive neuroscience. New York W.W. Norton & Company, Inc.
Kaas, J. (2001). The mutability of sensory representations after injury in adult mammals. In C. A. Jones & J.C. McEachern (Eds.), Toward a theory of neural plasticity (pp.323-334). Philadelphia Psychology Press.
Karl, A., Birbaumer, N., Lutzenberger, W., Cohen, L., and Flor, H. (2001). Reorganization of motor and somatosensory cortex in upper extremity amputees with phantom limb pain. The Journal of Neuroscience, 21:10, 3609-3618.
Lotze, M., Flor, H., Grodd, W., Larbig, W., & Birbaumer, N. (2001). Phantom movements and pain An fMRI study in upper limb amputees. Brain, 124:11, 2268-2277.
Qi, H.-X., Stepniewska, I., & Kaas, J. (2000). Reorganization of primary motor cortex in adult macaque monkeys with long-standing amputations. Journal of Neurophysiology, 84:4, 2133-2147.
Röricht, S., Meyer, B., Niehaus, L., & Brandt, S. (1999). Long-term reorganization of motor cortex outputs after arm amputation. Neurology, 53:106.
Roux, F.-E., Ibarolla, D., Lazorthes, Y., & Berry, I. (2001a). Chronic motor cortex stimulation for phantom limb pain A functional magnetic resonance imaging study Technical case report. Neurosurgery, 48:3, 681-688.
Roux, F.-E., Ibarolla, D., Lazorthes, Y., & Berry, I. (2001b). Virtual movements activate primary sensorimotor areas in amputees Report of three cases. Neurosurgery, 49:3, 736-742.
W., C., Wu, H., & Kaas, J. (1999). Reorganization in primary motor cortex of primates with long-standing therapeutic amputations. The Journal of Neuroscience, 19:17, 7679-7697.

MG