Figure 15.2. This figure shows the enhanced brain response in the chronic back pain patients with solicitous spouses to the back but not the finger stimulation. The brain activity was computed as the global field power related to the EEG gathered with 90 surface electrodes.
Direct verbal reinforcement of pain has been identified as an additional important modulator of the pain response. When patients and healthy controls were reinforced for increasing or decreasing their verbal pain responses both groups learned this task equally well, however, the patients showed a delay in the extinction of the response. When the somatosen-sory evoked potentials to the pain stimuli were examined, the late event-related responses (>200 ms) were unaltered and showed mainly habituation. However, the early response (N150) was affected by the conditioning procedure and remained high in the chronic pain group that had been reinforced for higher pain ratings thus indicating a direct effect of verbal reinforcement on the early cortical processing of nociceptive information (Flor et al., 2002b).
—♦—Punishing --■-■ Healthy controls —A -Solicitous
This lack of extinction in the cortical domain suggests that learning processes related to verbal and behavioral conditioning may exert long-lasting influences on the cortical response to pain-related stimuli and form implicit pain memories.
In addition to operant, classical conditioning has been identified as an important modulator of pain-related responses. This effect pertains not only to the ascending nociceptive system but also to descending pain-modulatory systems (Dubner and Ren, 1999). The fact that stress positively influences the pain response and activates the descending pain-inhibitory system has commonly been described by the term stress-induced analgesia or hypoalgesia. Animal studies have shown that stress analgesia can be conditioned and that some forms of both conditioned and unconditioned stress analgesia are mediated by the endogenous opioid system (Maier, 1989). It was recently shown that stress analgesia can also be classically conditioned in humans and that this conditioned analgesia is mediated by the release of endogenous opioids (Flor and Grusser, 1999; Flor et al., 2002a). To what extent deficient descending pain inhibition is involved in chronic pain has not yet been established nor do we know enough about the role of learning and memory process in the inhibition of pain. In an elegant study, Wunsch and colleagues (Wunsch et al., 2003) used pain as an conditioned and affective pictures as unconditioned stimuli (US) and showed that a classical conditioning procedure modified the pain ratings to a more or less intense perception depending on the nature of the US (aversive versus appetitive).
In summary, chronic pain states lead to the development of somatosensory pain memories that manifest themselves in alterations in the somatotopic map in somatosensory cortex as well as other brain areas related to the processing of pain. These plastic changes may contribute to hyperalgesic and allo-dynic states in the absence of or the presence of only minor peripheral nociceptive stimulation. These pain memories can be influenced by psychologic processes such as operant and classical conditioning that may establish additional and potentially more widespread implicit memories and may enhance existing memories. In addition to local representational changes, chronic states of pain are associated with increased cortical excitation that may significantly contribute to cortical reorganization. Pain-inhibitory systems are also influenced by learning and memory processes and may be altered in chronic pain.
As noted above, not only enduring nociceptive input but also the loss of input, for example, subsequent to amputation or nerve injury, can alter the cortical map. Several studies examined cortical reorganization after amputation in humans. These studies were instigated by the report of Ramachandran et al. (1992) that phantom sensation could be elicited in upper extremity amputees when they were stimulated in the face. There was a point-to-point correspondence between stimulation sites in the face and the localization of sensation in the phantom. Moreover, the sensations in the phantom matched the modality of the stimulation, for example warmth was perceived as a warm phantom sensation, painful touch was perceived as pain. The authors assumed that this phenomenon might be the perceptual correlate of the type of reorganization previously described in animal experiments. The invasion of the cortical hand or arm area by the mouth representation might lead to activity in the cortical amputation zone which would be projected into the no longer present limb. Subsequently, Elbert et al. (1994) and Yang et al. (1994) used a combination of magnetoencephalo-graphic recordings and structural magnetic resonance imaging - neuromagnetic source imaging - to test this hypothesis. They observed a significant shift of the mouth representation into the zone that formerly represented the now amputated hand or arm, however, this shift occurred in patients with and without phantom sensation referred from the mouth. Flor et al. (1995) showed that phantom limb pain rather than referred sensation was the perceptual correlate of these cortical reorganizational changes. Patients with phantom limb pain displayed a significant shift of the mouth into the hand representation whereas this was not the case in patients without phantom limb pain. The intensity of phantom limb pain was significantly positively correlated with the amount of displacement of the mouth representation. It was later shown that referred sensations as those described by Ramachandran et al. (1992) can also be elicited from areas far removed from the amputated limb (e.g., from the foot in arm amputees). Therefore, it was concluded that alterations in the organization of SI, where arm and foot are represented far apart, are most likely not the neuronal substrate of referred phantom sensations (Borsook et al., 1998; Grusser et al., 2001, 2004).
Similar results were obtained when the motor cortex was investigated. For example, an fMRI study where upper extremity amputees had to perform pucking lip movements showed that the representation of the lip in primary motor cortex had also shifted into the area that formerly occupied the amputated hand (Lotze et al., 2001; see Fig. 15.3). The magnitude of this shift was also highly significantly correlated with the amount of phantom limb pain experienced by the patients thus suggesting parallel processes in the somatosensory and motor system. A high concordance of changes in the somatosensory and the motor system was reported by Karl et al. (2001) who used transcranial magnetic stimulation to map the motor cortex and neuroelectric source imaging (that combines the determination of cortical sources by evoked potential recordings with structural magnetic resonance imaging) to map the somatosensory cortex. This close interconnection of changes in the somatosensory and motor system suggests that rehabilitative efforts directed at one modality may also affect the other.
The close association between cortical alterations and phantom limb pain was further underscored in a study by Birbaumer et al. (1997). In upper limb amputees, anesthesia of the brachial plexus lead to the elimination of phantom limb pain in about 50% of the amputees whereas phantom limb pain remained unchanged in the other half. Neuroelectric source imaging revealed that cortical reorganization was also reversed in those amputees that showed a reduction of phantom limb pain. Patients who continued to have phantom limb pain during the elimination of sensory input from the residual limb had an even more reorganized mouth representation. These data suggest that in some patients peripheral factors might be important in the maintenance of phantom limb pain whereas in others pain and reor-ganizational processes might have become independent of peripheral input. The importance of pain experiences prior to amputation was confirmed by a study of Nikolajsen et al. (2000a) who reported a close association between mechanical sensitivity
prior to amputation and early phantom limb pain. However, the authors only tested thresholds and not sensitization. Further research is needed to better clarify these relationships.
As Devor and Seltzer (1998), and others have pointed out it is to date not clear on which level of the neuraxis the cortical changes that have been observed in imaging studies originate. In addition to intracortical changes alterations might be present in the dorsal root ganglion, the dorsal horn, the brain stem or the thalamus. Several imaging studies (e.g., Willoch et al., 2000) have also shown that not only the primary and secondary somatosen-sory cortex and the posterior parietal cortex but also regions such as the insula and the anterior cin-gulate cortex are involved in the processing of phantom phenomena.
Similar alterations in the cortical processing of sensory information have recently also been reported in patients with complex regional pain syndromes (CRPS). In a magnetoencephalographic investigation, Juottonen et al. (2002) showed that in patients with CRPS the localization of the fingers of the affected hand had shifted closer together and that the magnitude of the magnetic response on the affected side was positively correlated with the amount of pain experienced by the patients. They also observed altered activity in the motor cortex suggesting a change of inhibition. Similarly, Pleger et al. (2004) found that CRPS patients revealed a marked hemispheric difference in the representation of the hand with the affected hand also being smaller in extent than the intact hand. The amount of hemispheric asymmetry was positively correlated with the amount of pain they experienced. Thus, both in phantom limb pain and in CRPS a smaller representation of the affected and completely or partially deafferented body part was associated with more pain whereas in chronic musculoskeletal pain, a larger representation was found to be associated with more pain. We do not yet fully understand the mechanisms underlying these types of relationship, however, the amount of use of the limb as well as the type of input from the limb or adjacent territory might play an important modulatory role.
Based on these findings and the results obtained on somatosensory memories in chronic back pain it can be assumed that prior pain memories might also be important in the development of phantom limb pain although they are surely not the only factors. Thus, when pain has occurred prior to the amputation, alterations in somatosensory cortex and other brain areas might have occurred that would later - when activated by neighboring input subsequent to the amputation - lead to the sensation of phantom limb pain (Flor, 2002). Initial evidence from a longitudinal study (Huse et al., in press) suggests that chronic pain before the amputation is a much more important predictor of later phantom limb pain than acute pain at the time of the amputation thus supporting this assumption. In addition, peripheral changes related to the amputation may contribute to enhanced cortical reorganization and phantom limb pain. To summarize, somatosensory pain memories represented by alterations in the topographic map of SI cortex may underlie the development of phantom limb pain and may contribute to other neuropathic pain syndromes such as CRPS. Longstanding states of chronic pain prior to the amputation may be instrumental in the formation of these pain memories by inducing representational and excitability changes. Deafferentation does not alter the original assignment of cortical representation zones to peripheral input zones but leads to double coding. Peripheral factors such as loss of C-fiber activity, spontaneous activity from neuroma or psychophysiologic activation may also influence the cortical representational changes (Calford and Tweedale, 1991; Spitzer et al., 1995). Learning processes are instrumental in the development and maintenance of these cerebral changes.
Whereas tension headache seems to have some similarity to musculoskeletal pain disorders, migraine headache and cluster headache are distinct categories of pain with different etiologies. In migraine headache, a consistent feature has been an inability to habituate to various types of sensory stimulation in the pain-free interval. For example, in migraine headache a lack of habituation to auditory or visual stimuli was observed for the pain-free interval (Gerber and Schoenen, 1998; Evers et al., 1999; see Ambrosini et al., 2003, for a review). This was assessed either by visual evoked potentials or by changes in the contingent negative variation, a slow cortical potential that is typically present in an S1-S2 task where S1 signals the occurrence of a second signal, S2 that is the sign for a response to be made by the subject. Usually, multiple presentations of the S1-S2 stimuli lead to habituation of the CNVV which was not found in migraine headache. In addition, a brainstem migraine generator has been proposed because alterations in brain stem activity have repeatedly been found during migraine attacks (Weiller et al., 1995; for a review see Sanchez del Rio and Linera, 2004). In cluster headache, several functional and structural imaging studies have identified a cluster headache-specific alteration in the hypothalamus (May et al., 1998, 1999; for a review see Goadsby et al., 2002). It is, however, not clear to what extent these changes are a cause or a consequence or merely an epiphenomenon of the pain that is experienced.
15.3 Influencing chronic pain and brain plasticity
The discussion in the preceding sections suggests that the alteration of somatosensory pain memories might be an influential method to reduce both chronic musculoskeletal and neuropathic pain. This could be achieved by altering the peripheral input that enters the brain region that coded a pain memory, for example by using electromyogram (EMG) or temperature biofeedback or other types of standard behavioral interventions (for reviews see Andrasik, 2003, 2004; Flor et al., 1992; Sherman, 1997) or by employing a sensory stimulation protocol that provides relevant correlated sensory input to the respective brain region. It would also be possible to directly alter the brain response to pain by providing feedback of event-related potential components or electroencephalogram (EEG) rhythms (Kropp et al.,
2002). Most of these methods have not yet been tested in a systematic manner and their effects on cortical reorganization are so far unknown. Alternatively, pharmacologic interventions could be used that prevent or reverse the establishment of central memory traces.
In phantom limb pain, it was assumed that the pain is maintained by cortical alterations fed by peripheral random input. In this case the provision of correlated input into the amputation zone might be an effective method to influence phantom limb pain. fMRI was used to investigate the effects of prosthesis use on phantom limb pain and cortical reorganization (Lotze et al., 1999). Patients who systematically used a myoelectric prosthesis that provides sensory and visual as well as motor feedback to the brain showed much less phantom limb pain and cortical reorganization than patients who used either a cosmetic or no prosthesis. The relationship between phantom limb pain and use of a myoelec-tric prosthesis was entirely mediated by cortical reorganization. When it was partialled out from the correlation, phantom limb pain and prosthesis use were no longer associated. This suggests that sensory input to the brain region that formerly represented the now absent limb may be beneficial in reducing phantom limb pain. These studies were performed in chronic phantom limb pain patients. An early fitting and training with a myoelectric prosthesis would probably be of great value not only in the rehabilitation of amputees but also in preventing or reversing phantom limb pain.
These assumptions were further confirmed in an intervention study where the patients received feedback on sensory discrimination of the residual limb (Flor et al., 2001). Eight electrodes were attached to the residual limb and provided high-intensity non-painful electric stimulation of varying intensity and location that led to the experience of intense phantoms. The patients were trained to discriminate the location or the frequency of the stimulation (alternating trials) of the stimulation and received feedback on the correct responses. The training was conducted for 90 min per day and was spread over a time of 2 weeks (10 days of training). Compared to a medically
0 Mouth pre I Thumb pre Mouth post Thumb post
0 Mouth pre I Thumb pre Mouth post Thumb post
Figure 15.4. Changes in cortical reorganization (distance of the hand and lip representation in millimeter) during the discrimination training (electrode montage is depicted on the left side).
treated control group that received an equal amount of attention the trained patients showed significantly better discrimination ability on the stump. They also experienced a more than 60% reduction of phantom limb pain and a significant reversal of cortical reorganization with a shift of the mouth representation back to its original location. The alterations in discrimination ability, pain and cortical reorganization were highly significantly correlated (see Fig. 15.4).
In a related study (Huse et al., 2001b) asynchronous tactile stimulation of the mouth and hand region was used over a time period of several weeks. This training was based on the idea that synchronous stimulation leads to fusion and asynchronous stimulation leads to a separation of cortical representation zones. In this case it was postulated that input from the mouth representation that would now activate the region that formerly represented the now amputated hand and arm would be eliminated and with it the phantom phenomena that would be projected to the amputated limb. This intervention also showed a reduction in phantom limb pain and cortical reorganization.
In addition to behavior, pharmacologic interventions may also be useful in the treatment of both chronic musculoskeletal and neuropathic pain. The prevention of pain memories might be possible by using pharmacologic agents that are known to also prevent or reverse cortical reorganization. Among these substances, gamma amino butyric acid (GABA) agonists, N-methyl-C-aspartate (NMDA) and alpha-amino-3-hydroxy-5methyl-4-isoxazole propionic acid (AMPA) receptor antagonists and anticholinergic substances seem to be the most promising. A recent double-blind placebo-controlled study that used the NMDA receptor antagonist memantine in the perioperative phase in acute amputations reported a decrease of the incidence of phantom limb pain from 72% to 20% 1 year after the amputation (Wiech et al., 2001; see Fig. 15.5). The pharmacologic intervention was most effective in those patients where treatment had begun before or immediately after the amputation. This study could explain why the results of different controlled prospective studies about the effect of preemptive analgesia initiated at least 24 h before the amputation on the incidence of phantom limb pain are inconsistent. For example, a well-controlled study by Nikolajsen et al. (1997) showed no effect of preemptive analgesia on phantom limb pain. If a preexisting pain memory is important in the development of phantom limb pain, the use of preemptive analgesia, that eliminates afferent barrage in the perioperative phase but will not alter previously formed neuronal changes may be ineffective.
Treatment of chronic phantom limb pain with pharmacologic agents has also yielded inconsistent
Incidence of phantom limb pain
Was this article helpful?
Free Your Mind And Achieve Peace. Discover How To Live In Peace And Harmony In A World Full Of Uncertainty And Dramatically Improve Your Quality Of Life Today. Finally You Can Fully Equip Yourself With These “Must Have” Tools For Achieving Peace And Calmness And Live A Life Of Comfort That You Deserve.