| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Neuroimaging in Rehabilitation |
AJ Butler, PT, PhD, is Assistant Professor, Department of Rehabilitation Medicine, Center for Rehabilitation Medicine, Emory University School of Medicine, 1441 Clifton Road NE, Atlanta, GA 30322 (USA)
SL Wolf, PT, PhD, FAPTA, is Professor, Department of Rehabilitation Medicine, Emory University School of Medicine
Address all correspondence to Dr Butler at: andrew.butler{at}emory.edu
Submitted September 15, 2006;
Accepted January 10, 2007
 | Abstract |
|---|
| Introduction |
|---|
This article is intended to serve several purposes that are presented in a logical sequence. First, we explore the fundamental nature and mechanisms of plasticity. This discussion is followed by a brief introduction to TMS techniques and physiological effects of magnetic stimulation in healthy adults. We then apply this foundation to TMS studies of plasticity in subjects who were healthy. Next, we describe how TMS induces plasticity within the human brain. This description is followed by a survey of stimulation techniques that can serve as potential therapeutic tools for promoting favorable plasticity, initially within a variety of neurological disorders and subsequently within the process of neurological rehabilitation. Finally, we look toward the future, speculating on how novel and far-reaching approaches with TMS could influence human brain plasticity. As a result, readers should have a firm understanding of neuroplasticity that may lead to a better understanding of the human nervous system and the relevance of neuroplasticity to clinical rehabilitation.
| Fundamentals of Plasticity |
|---|
If plasticity does characterize the capacity of the brain to change and is an intrinsic but persistent property of the nervous system, one relevant application of the underlying principles may involve the acquisition of new skills, especially in response to changes in the environment. This mechanism may be the basis for growth, development, and learning. For example, when a new skill is acquired, the function of the neural network is determined by the most dominant input that it receives; the input can be altered by certain behaviors. Additionally, and particularly germane to the practice of neurorehabilitation scientists across disciplines, there is the possibility that plastic changes underlie the mechanisms by which the recovery of function occurs after central nervous system (CNS) or peripheral nervous system injury11,12; this point is explored more completely later.
It is known that the functional organization of the cerebral cortex is plastic; that is, changes in organization occur throughout the life span in response to the numerous events that define experiences. The potential for reorganization has been demonstrated in both sensory and motor areas of the adult cortex as a consequence of trauma, pathological changes, manipulation of sensory experience, or learning. These changes can be evaluated only when referenced against an extensive collection of experimental data that have identified a repeatable representation pattern (eg, somatotopic, tonotopic, or retinotopic pattern) from which changes can be detected. Although assessing the scope of such changes is often at the edge of current technical capabilities, there are striking examples of significant and rapid changes, such as the increased size of the trained hand motor representation following 5 consecutive days of piano exercise (2 hours per day) compared with the size of the untrained hand motor representation.13
Alterations in cortical organization are known to emerge through changes in synaptic efficacy within the cortex and elsewhere in the nervous system. Furthermore, these changes have been linked closely to 2 phenomena, long-term potentiation (LTP) and long-term depression (LTD). Long-term potentiation, the long-lasting enhancement of synaptic transmission first reported by Bliss and colleagues more than 30 years ago,14,15 has been the focus of an enormous amount of investigation (Fig. 1). Figure 1 shows that there is a clear, long-lasting potentiation (up to 4 hours) of responses following trains of stimuli given at 15 per second to the hippocampal formation of awake, active rabbits. Long-term potentiation has long been regarded, along with its counterpart, LTD, the weakening of a neuronal synapse that lasts from hours to days, as a potential mechanism for memory formation and learning. Figure 2 shows a model of normal synaptic transmission (Fig. 2A) and the rise of Ca2+ levels in the dendritic spine, triggering the induction of LTP (Fig. 2B).
|
|
Uncovering of latent or existing connections.
Uncovering or unmasking preexisting connections in the primary motor cortex (M1)16,17 could be a mechanism for rapid (early) plasticity in response to manipulations of sensory inputs18,19 or motor outputs20,21 of cortical representational maps.
The somatosensory cortex of adult mammals has been shown to have the capacity to reorganize itself when inputs are removed through cutting of afferent nerves or amputation of a part of the body. The area of the cortex that normally would respond to stimulation from the missing input can become responsive to inputs from other parts of the body surface. Plastic changes were shown to occur in the primary somatosensory cortex of the flying fox following amputation of the single exposed digit on the forelimb.22 Immediately after amputation, neurons in the area of the cortex receiving input from the missing digit were not silent but responded to stimulation from adjoining regions of the digit, hand, arm, and wing. In the week following amputation, the enlarged receptive fields shrank until they covered only the skin around the amputation wound. The immediate response was interpreted as a removal of inhibition, and the subsequent shrinking of the fields might have been attributable to reestablishment of the inhibitory balance in the affected cortex and its inputs.
Activation of existing but silent synapses.
Alternatively, the activation of existing but silent synapses could serve as a mechanism for the induction of rapid plasticity. Silent synapses are connections between neurons displaying no 
-amino-5-hydroxy-3-methyl-4-isoxazole propionic acid (AMPA)–mediated glutamate response23,24; presynaptic transmitter release would not result in a rapid potential shift in the target neuron. The AMPA receptor is a non-N-methyl-D-aspartic–type ionotropic transmembrane receptor for glutamate that mediates fast synaptic transmission in the CNS. Its name is derived from its ability to be activated by the artificial glutamate analog AMPA. Receptors for AMPA are found in many parts of the brain and are the most commonly found receptors in the nervous system. The "awakening" of silent synapses by the insertion of postsynaptic AMPA receptors25–28 is a mechanism proposed to account for the rapid increases in synaptic efficacy that have been observed experimentally.
Silent synapses have been implicated in brain plasticity in both young and mature animals.29 There is convincing evidence for the occurrence of silent synapses in the developing nervous system,23,24 but as maturation progresses, silent synapses become rare27,30 and presumably are replaced by active ones. The unmasking of any silent synapses that are present could support functional reorganization. The silent synapse mechanism may be relevant to the immature human nervous system and hence rehabilitation potential in young patients with cerebral palsy but is a less likely candidate mechanism for the cortical changes seen in older adults during recovery from stroke.
Activity-dependent synaptic plasticity and LTP.
The most widely studied but controversial mechanism for supporting representational plasticity is LTP,14,31 especially as a critical link between behavioral change and synaptic function. For the hippocampal cortex, neocortex, and amygdala, there is now more than 30 years of evidence supporting a possible role of LTP in learning and memory. Population measures of neuronal cells have indicated that LTP and LTD operate during learning to modify synaptic efficacy.32 Certain forms of learning lead to an enhancement of synaptic responses in a variety of brain structures.33–35 Recently, LTP was implicated in the learning of new motor skills,36 and there is compelling evidence that LTP is the mechanism involved in natural learning. In the study by Rioult-Pedotti et al,36 rats were trained for 5 successive days to reach with their preferred forelimb into a box and retrieve small food pellets. Grasp attempts began during the first session, and the success rate improved during the first 3 training days and then became asymptotic (Fig. 3A). The results reported by Rioult-Pedotti et al36 indicated that increased synaptic efficacy with initial skill learning as well as skill performance was maintained. Learning specifically strengthened extracellular field potentials in the M1 forelimb region (Fig. 3B). There were no interhemispheric differences in the hind limb region or in paired control rats. The data suggested that synapses are modifiable; they are modified with learning and are strengthened through an LTP-like mechanism.
|
Enriched Environments
Exposure to enriched environments results in a larger number of synapses per neuron,46 increases in dendritic spine density,47 and changes in dendritic spine morphology.48,49 In a study by Biernaskie and Corbett,50 animals that had lesions and that were exposed to an enriched environment showed enhanced dendritic spine complexity and length compared with animals exposed to a "standard" environment. These results suggested that enrichment combined with task-specific rehabilitative therapy is capable of augmenting intrinsic neuronal plasticity within noninjured, functionally connected brain regions as well as promoting an enhanced functional outcome.
Thus far, the general concept of cortical plasticity as it pertains to the motor cortex and its role in motor skill learning and more general principles concerning synaptic plasticity have been introduced. The possible mechanisms for plasticity have been discussed as they relate to motor system function, skill learning, and rehabilitation. These constructs underlying plasticity are precursors to the description of a noninvasive technique, TMS, a tool that provides a valuable method for exploring and understanding cortical plasticity in humans.
Examples of Plastic Change
Some of the most convincing evidence that learning and practice influence cortical organization and that learning operates through LTP- and LTD-mediated mechanisms has been obtained in studies of the motor cortex. This work is significant to physical therapists because knowing that potentiation has been engaged implies that the impaired or damaged motor cortex can be restructured through appropriate physical rehabilitation methods or through other means (eg, pharmacological or magnetic stimulation) that alter the mechanisms accounting for LTP and LTD.
The functional topography (ie, the graphic delineation of spatial architecture of the cortex usually on 2-dimensional maps) of M1 can be modified by peripheral or central injury, electrical stimulation, pharmacological manipulations, or experience. Behaviorally or experimentally induced reorganization of M1 output maps is characterized by shifts in borders between different motor representations. For example, M1 representations undergo rapid reorganization within hours of the occurrence of peripheral nerve lesions.20,51,52 Changes in cortical output maps can be induced with prolonged changes in limb positions,53,54 supporting the conclusion that sensory feedback derived from joint or muscle afferents is important in shaping M1 representations.
The primary motor cortex is also a site in which reorganization occurs during the acquisition or practice of motor skills. In a study in which intracortical microstimulation techniques (and not TMS) were used, skilled finger use in monkeys expanded the digit representation in M1.55 Skill learning–induced changes in M1 also were detected at the single-cell level in primates.56 Monkeys learned to adapt their reaching movements to externally applied force fields. The firing rate and the tuning of individually recorded cells in M1 changed during the period of adaptation to new force fields. A group of these cells (memory cells) retained the newly acquired activation pattern even after the force field was shut off and the monkey's hand trajectory returned to the control condition. Other memory cells that normally were directionally untuned became directionally tuned with the acquisition of the new skill and remained tuned to the direction of the arm movement after the force field was shut off. These data provide evidence for single-cell plasticity in M1. In humans, M1 representations also appeared to be enlarged or rearranged during motor learning.57–60 Further, roles of M1 in early motor consolidation61 and in motor memory58 have been demonstrated in humans.
In rats, learning a skilled reaching task but not an unskilled reaching task led to significant increases in the mean areas of the wrist and digit representations at the expense of the size of the shoulder representation; these results demonstrated that training-induced map reorganization was characterized by an expansion of "trained" representations into "untrained" representations without an overall increase in map size62 as rats accrued skilled distal forelimb movements. Moreover, such changes may well be driven by the specificity or "challenge" contained within the task, such as the skill set required to reach. For example, Kleim et al63 demonstrated a significant increase in the volume of neurons within the caudal forebrain area of rats trained to retrieve food pellets from a rotating disk (skilled reaching), but this result was not obtained when rats used a total forelimb lever press to obtain the food reward (unskilled reaching).
These results indicated that representational plasticity is driven by skill acquisition, learning, or practice of a newly acquired action and not by simple repetitive motor activity64,65 and suggested that only specific patterns of activity are capable of producing functional M1 plasticity. The implications of these observations for the provision and progression of rehabilitative training procedures are profoundly clear.
| Introduction to TMS Techniques and Physiological Effects on Adults Who Are Healthy |
|---|
Electromagnetic induction allows current to be directed through a handheld copper stimulation coil, which produces a transient magnetic field (Fig. 4B). When held over the scalp, the rapidly changing magnetic field induces a small electric current (Fig. 4E) in underlying brain tissue; this current produces a depolarization of nerve cells that results in the stimulation or disruption of brain activity, depending on the frequency and intensity of stimulation as well as the location of the stimulating probe. When applied over M1 at low stimulus intensities, single-pulse TMS is thought to stimulate the corticospinal tract indirectly (transsynaptically) through horizontal fiber depolarization.71,72 The neurons activated depend on the size, shape, orientation, and intensity of the stimulus waveform that are produced by the magnetic stimulator.73 The resultant efferent volleys can be recorded as MEPs with surface or indwelling electrodes at peripheral target muscles.
|
Biphasic stimulus pulses are more efficient in stimulating the brain than monophasic pulses, even when the initial phase of the stimulus is the same size,74,75 because the charge transfer is maximal in the swing between the first and second phases of a biphasic pulse.73
The delivery of TMS often is described on the basis of the frequency of the cortical stimulation. Repetitive (or rapid-rate) TMS usually refers to the application of TMS at frequencies above 1 Hz and often is applied in treatment studies (see below). The application of TMS at frequencies of 1 Hz or below may be referred to as slow or low-frequency TMS and often is used in motor cortex mapping procedures.8
In the context of physical therapy, the need to understand the relationship between an intervention and its effect on movement capabilities would make TMS a most appealing tool for studying cortical reorganization. Different TMS parameters are used to investigate motor system excitability. The "hot spot" (the most active scalp position for the target muscle) for motor stimulation is defined as the location at which the minimal stimulus intensity needed to produce an evoked motor response (the motor threshold) is the lowest from among all of the locations surveyed but at which the highest-amplitude response at that stimulus intensity also is obtained.76 Specifically, the resting motor threshold for the hot spot is defined as the minimum TMS intensity required to elicit at least 5 MEPs (
50 µV) in 10 consecutive stimuli at rest.77
A principal measure is the area of motor output representation, often referred to as an MEP map. The MEP map refers to the area on the scalp surface from which MEPs in the target muscle can be obtained. For this, multiple scalp sites are stimulated by moving the stimulation coil along a grid. Other important measures include MEP latency, location of the amplitude-weighted center of gravity (COG) of the motor output map,4 MEP amplitudes (at rest and sometimes with facilitation), and MEP recruitment curves.65,78,79 The location of the COG of the MEP map corresponds to the scalp location at which the largest number of the most excitable corticospinal neurons can be stimulated. Therefore, changes in the COG should indicate true changes in the topographical organization of motor cortex representations.
Therapeutic studies in which TMS is used as an outcome measure have been undertaken, and an examination of their relative strengths and weaknesses seems appropriate because the data generated from such studies have profound implications for the interpretation of cortical reorganization following the application of neurorehabilitative procedures.
Transcranial magnetic stimulation has become appreciated as an important treatment modality for a variety of psychiatric diseases, including major depressive disorder, schizophrenia, and obsessive-compulsive disorder.8 Transcranial magnetic stimulation also has become an important evaluative tool80 and potential predictor of stroke recovery.81 However, TMS currently is not approved by the US Food and Drug Administration (FDA) for the treatment of these disorders in the United States. Single-pulse TMS has achieved FDA approval for the stimulation of peripheral nerves and muscles in the United States (but not for use on the brain). Therefore, TMS can be used as a tool for the evaluation of nerve root and plexus lesions. In November 2005, licensing for magnetic stimulation was granted by Health Canada for the assessment of neurological and muscular functions. Here we describe the use of TMS-derived mapping as an outcome measure.
As mentioned previously, the MEP map refers to the area on the scalp surface from which MEPs in the target muscle can be obtained. For this, multiple scalp sites are stimulated by moving the stimulation coil along a grid. Motor evoked potentials are recorded from electrodes placed strategically over a muscle of interest as maps related to specific movements are charted (Fig. 5). Transcranial magnetic stimulation mapping of motor cortical areas follows the basic principles of Penfield82 and is based on the idea of stimulating different regions of the brain and measuring the motor effects. Maps are generated by quantifying the motor effects and relating these to the scalp sites stimulated. Such maps indicate the region of the scalp in which stimulation can evoke a response in a muscle of interest and, therefore, are related only indirectly to the origins of the projection in the underlying motor cortex.83 Motor evoked potentials are elicited by providing a temporally varying current passed through a coil to induce an electric field in the underlying brain when the coil is placed over the appropriate cortical location, such as the motor cortex. When the MEPs are displayed as a function of a Cartesian coordinate system, a motor map can be created (Fig. 6).
|
|
This anatomical certainty is important because the electrode placement traditionally used during TMS studies may lead to cortical MEPs that may be derived from more than one muscle85; therefore, the relationship of the stimulation to the response should not necessarily be expressed as a muscle but as a movement. Moreover, if electrodes used to record MEPs are not placed specifically over the muscle in question, then the volume of the conducted response actually may represent accumulated responses from several muscles.86 This concern is justified because monitoring of the cortical representation of movement before and after an intervention with TMS as the assessment vehicle should yield data relevant to the intent of the treatment. Thus, for example, a treatment designed to relax finger, thumb, or wrist flexion motions while enhancing the counterpart extension activities should be reflected in MEPs that include the relevant movements and not the counterproductive movements.87,88
As mentioned above, TMS-derived maps are related only indirectly to the origins of the projection in the underlying motor cortex. Nevertheless, TMS-derived maps can provide at least a gross idea of the somatotopic pattern of the human motor cortex and reveal the best points for activating muscles in the shoulder, arm, and hand as well as in the face, arm, and leg.4 The process of finding the best points for activating various muscles can be standardized by marking a matrix of points on the scalp and then plotting the amplitudes of electromyographic responses obtained in various muscles at each point with a Cartesian coordinate system (Fig. 6).
Such maps provide 3 pieces of information: the optimal position at which to obtain the largest response (the so-called hot spot), the COG of the area, and the area of the scalp from which responses can be obtained. The optimal position, or hot spot, presumably corresponds to the location of the most excitable population of neurons that project to the target muscle. Specific details regarding the derivation of these measures can be found elsewhere.4,86
The area of the representation is more complex and depends on 2 factors: the true area of the cortex on which neurons that project to the target muscle are located and the stimulus intensity used to produce the map. If the stimulus intensity is too low, then the total extent of the map may be underestimated because less excitable elements will not be recruited. If the stimulus intensity is too high, then the area will be overestimated because the stimulus current will spread beyond the point of stimulation.89 These 2 opposing factors are difficult to reconcile with TMS, leaving the exact meaning of the map area that is recorded difficult to interpret.73
Nearly all mapping studies recognize these inherent limitations in technique; therefore, they focus not on the absolute size of a map but on changes in the map resulting, for example, from a stroke or intervention. With the exception of a few studies, most interventions change the size of the map without affecting its hot spot.90–94 In such cases, changes in the map size are best observed when participants are in a relaxed position and the muscles being tested are in a state of inaction. However, observations of motor excitability at rest present challenges to interpretation. In a recent study, Darling et al95 showed that the variability of motor potentials evoked by TMS depended on muscle activation. They showed that the relative variability of single MEPs at a constant stimulus intensity and prestimulus muscle electromyographic activation was lower during maintained 5% and 10% contraction levels than during 0% contraction levels.95 Therefore, maintaining a stable low-intensity contraction helps to stabilize cortical and spinal excitability. This observation suggests that comparisons of physiological changes during recovery in people with neuropathology may be influenced by the resting state of muscle contraction.
| TMS for Predicting Functional Recovery After Stroke |
|---|
Several investigators have examined the correlation between TMS-derived map characteristics after stroke and the extent of motor recovery in humans.102–104 Pennisi et al81 demonstrated that complete hand paralysis in association with the absence of early MEPs (within 48 hours of ictus) predicted poor neurological recovery at 1 year in 15 subjects after stroke (middle cerebral artery infarct). Conversely, the preservation of motor potentials evoked by TMS in the early period after stroke may portend good functional recovery.72,105 Other investigators72,81,106,107 have reported relationships between the rate and extent of recovery after stroke and changes in the presence of MEPs, length of time for conduction from cortex to muscle, MEP latency, excitability threshold, and MEP amplitude. The absence of a response to TMS, a long duration of MEP latency, and a lengthened conduction time (compared with those of people who are healthy) in the early period after an injury are predictive of reduced hand motor function recovery.
In monohemispheric infarctions, decreased affected hemisphere motor output area and increased excitability thresholds for paretic muscles have been observed repeatedly in TMS-derived maps obtained during the subacute and chronic phases for patients with stroke.70,108 These electrophysiological changes presumably are related to motor impairment and may be secondary to neuronal damage, disuse, unbalanced transcallosal inhibition from the less affected hemisphere, or other, unidentified mechanisms.109
| Responses to Repetitive Task Practice |
|---|
The area of APB muscle representation in the affected hemisphere increased significantly immediately after training but then decreased toward baseline after 1 day. Increased affected hemisphere motor output area was associated with improved dexterity on a clinical measure (the Nine-Hole Peg Test) in 7 of the subjects, although the amount of clinical improvement did not correlate with the extent of change in the area. The excitability threshold at the hot spot and the COG were unchanged after training, possibly signifying that the enlargement in the affected hemisphere was attributable to increased excitability at the edges of the map. The rapid change detected in the TMS-derived map after a brief training session suggested that functional, rather than structural, mechanisms were involved. Potential mechanisms discussed by the authors included the modulation of inhibitory gamma-aminobutyric acid transmission at the borders of the motor map and alterations in glutamate transmission.108 Gamma-aminobutyric acid is the chief inhibitory neurotransmitter in the vertebrate CNS.
| TMS-Derived Mapping in Constraint-Induced Movement Therapy (CIMT) |
|---|
After CIMT, TMS parameters showed no change in thresholds but significant increases in MEP amplitudes and APB muscle motor output area in the affected hemisphere, possibly indicating increased excitability of surrounding neuronal networks. The unaffected hemisphere motor output areas were smaller after the training period, presumably because of decreased use of the less affected upper extremity, normalization of the unaffected hemisphere APB muscle representation, or increased transcallosal inhibition of the unaffected hemisphere by the affected hemisphere.
Shifts in the COG were significant (in the medial-lateral axis) only for the affected hemisphere, suggesting the possible recruitment of adjacent areas along the motor cortex. All subjects showed significant improvement in their use of the affected extremity, but scores on the Motor Activity Log,117 a 6-point subjective impression of how well and how often movement is observed in the affected arm during basic activities of daily living, did not correlate with the degree of map changes. Liepert et al108 suggested that physical therapy induces use-dependent reorganization which supports recovery-associated plastic changes.
In a follow-up study,70 clinical (Motor Activity Log) and TMS measurements were obtained at multiple time points before and after CIMT in 13 patients with chronic stroke (>6 months). Again, the affected hemisphere showed a smaller area of APB muscle representation at baseline, with a near doubling of the area after CIMT. Motor Activity Log improvements were maintained at the later measurement points. However, a return toward baseline in the area of the affected hemisphere APB muscle representation was seen at the 4-week and 6-month TMS sessions, indicating a possible "normalization after therapy-induced hyper-excitability"70 through improved synaptic efficacy or the relegation of motor function to non–TMS-accessible regions.
Several mechanisms have been purported to explain the TMS-derived map changes observed following CIMT.70 The intervention may have produced long-lasting changes in the cortical inhibitory network or perhaps the enhancement of synaptic strength within preexisting synaptic connections. Given that the unaffected limb was immobilized during the intervention, it is possible that the initial changes were related to a use-dependent mechanism.
| Induction of Plasticity With Cortical Stimulation |
|---|
| rTMS |
|---|
3 Hz) rTMS has been shown to increase contralateral motor cortex excitability, whereas low-frequency (
1 Hz) rTMS decreases contralateral motor cortex activity (MEP). Mechanisms similar to LTP and LTD are thought to be involved in the generation of these effects.127–129 High-frequency rTMS increases overall corticospinal synaptic activity,130 as expressed through changes in blood flow and metabolism and as measured by positron emission tomography and functional magnetic resonance imaging, whereas low-frequency rTMS tends to reduce synaptic activity in targeted brain areas.126 Moreover, it seems that modulatory effects extend beyond a targeted area and involve various cortical and subcortical regions functionally related to the targeted area.126,131 | Modulation of the Motor Cortex |
|---|
|
Up-Regulation of the Affected Motor Cortex
Enhancement of the ability of peri-infarcted and non–primary motor cortex regions of the affected hemisphere to respond to motor training or other neurorehabilitative interventions may be important because recent observations showed that increases in a number of growth-related processes likely contribute to behavioral recovery (Fig. 7, item 4). These processes may take place at the rim of tissue surrounding a cortical infarct.110,135–137 Cortical stimulation can modify activity in the motor cortex in animals138 and modulates cortical plasticity in humans. For example, TMS synchronously applied to a human motor cortex engaged in a motor training task enhanced use-dependent plasticity in the contralateral hand.139 This outcome provides evidence for the role of TMS in enhancing use-dependent plasticity and has implications for treatment methods aimed at facilitating motor recovery after stroke. This notion was tested recently in a therapeutic trial. Repetitive TMS or sham stimulation was applied over the stroke-affected motor cortex daily for 10 days to 2 randomly assigned groups of 26 patients with acute ischemic stroke. Patients otherwise continued their normal treatments. Disability measures—such as the Scandinavian Stroke Scale, the National Institutes of Health Stroke Scale, and the Barthel Index—applied before rTMS, at the end of the last rTMS session, and 10 days later showed that rTMS stimulation improved patients’ scores more than sham stimulation.140
The implication of these findings is that noninvasive cortical stimulation could represent an adjuvant to motor training in efforts to recover lost function after cortical lesions such as stroke. Consistent with this view, recent studies showed that noninvasive transcranial direct current stimulation can enhance motor function in people who are able-bodied141 and patients with chronic stroke.142
Given that there are several options for increasing and decreasing the levels of excitability and synaptic activation of the motor cortex in order to promote and facilitate plastic changes and consequently to improve motor learning in people who are healthy and in people with stroke, effects similar to those obtained with direct electrical cortical stimulation (through surgically implanted epidural electrodes) without the risks inherent in surgery might be expected. Moreover, rTMS might effectively replace the need for surgical procedures in at least a subset of patients. Although transcranial direct current stimulation has been shown to be effective, it produces a current that is dispersed through the cortex, thereby posing a challenge to the identification of muscle-specific changes and the exact anatomical substrate influenced by the stimulus.
| TMS as a Potential Therapeutic Tool for Promoting Beneficial Plasticity |
|---|
It is clear that TMS can produce effects not only at the site of stimulation but also at distant connected sites. Thus, stimulation of M1 affects spinal motor neurons and muscle through at least 2 synaptic linkages. The same is true of central connections. For example, stimulation over the motor cortex in one hemisphere affects the excitability of contralateral motor areas through transcallosal connections,130,143 stimulation over the frontal eye fields affects metabolic activity in the parieto-occipital cortex,131 and stimulation over the premotor cortex affects the excitability of M1.144
The effect of TMS is proportional to the level of neuronal excitability at the time at which the stimulus is applied. Thus, motor potentials evoked in actively contracting muscles are larger than those evoked in muscles at rest. The same principle applies to central pathways. The excitability of the transcallosal connections between the motor cortexes changes depending on whether people contract one hand or both hands while performing a task. This mechanism suggests the possibility of increasing the specificity of targeting of particular connections by applying rTMS when a person performs a behavioral task. For example, if a person with upper-limb hemiparesis had poor individual finger movements but finger tapping was maintained, then the application of stimulation during finger tapping might prove beneficial to the individual finger movements because of the mechanism outlined. For a further discussion of the mechanism involved, refer to work done by Liepert et al,145 Shimizu et al,146 and De Gennaro et al.147
Perhaps the most problematic question regarding the therapeutic use of rTMS concerns the duration of its effect. In all studies of participants who were healthy, effects have lasted between 30 minutes and 1 hour. The limited period of time following the removal of rTMS during which to modulate the excitability of the motor cortex has led several groups148–151 to use repeated (daily) administration of rTMS to prolong benefits through the summation of responses.
Five consecutive sessions of rTMS increased the magnitude and duration of the motor effects in patients with stroke.152 Fifteen patients with chronic stroke were randomly assigned to receive active or sham rTMS of the unaffected hemisphere. Compared with sham rTMS, active rTMS resulted in a significant improvement in motor function performance in the affected hand that lasted for 2 weeks. There was a significant correlation between improvement in motor function performance and change in corticospinal excitability in the affected hemisphere. These results support and extend the findings of previous studies of rTMS in patients with stroke because 5 consecutive sessions of rTMS increased the magnitude and duration of the motor effects.
Heightened excitability typically is expressed through electrophysiological differences, but few studies have addressed behavioral enhancements of the contralateral limb. Thus, work on the motor system commonly has used MEP threshold, MEP amplitude, paired-pulse testing, or silent-period duration as a measure of the effects of rTMS. However, relatively few studies have tested whether any of these measures is behaviorally relevant. In subjects who were healthy, finger tapping speed decreased after low-frequency magnetic stimulation at 0.9 Hz for 15 minutes (810 pulses) over the motor cortex.153 In contrast, peak force and peak acceleration were not affected by application to the hand representation of the right M1 of rTMS at 1 Hz for 15 minutes at an intensity of 115% of the individual resting motor threshold.154 After the application of subthreshold rTMS at 1 Hz, patients with dystonia showed a significant reduction in mean writing pressure that was associated with clear but transient improvement.155
We have offered support for TMS as a potential therapeutic tool through the promotion of beneficial plasticity in the human brain. In some patients, rTMS can reinforce deficient neuronal pathways and may improve behavior temporarily. The effects of TMS can be produced not only at the site of stimulation but also at distant connected sites, a finding that could have potential implications for therapeutic use in patients with Parkinson disease. Next, we speculate on the future uses of TMS.
| Future Study and Influence of TMS on Human Brain Plasticity |
|---|
Transcranial magnetic stimulation can be used to map the cortex and assess its excitability and resultant changes following interventions in many patient populations. Changes in map size correlate with improved recovery in patients with stroke.108 Other investigators156 have used TMS to assess cortical plasticity and function in people with incomplete tetraplegia. Transcranial magnetic stimulation has been used to assess changes in inhibitory and excitatory activities in the motor cortex in patients with stroke and to evaluate whether these changes are related to the extent of a patient's recovery of function.157 The ability to accurately assess the physiological mechanisms of recovery with TMS will provide rehabilitation therapists with an opportunity to generate interventions tailored to the specific physiology of an individual patient.
The potential uses of rTMS as a therapeutic tool include producing effects on the cerebral cortex that outlast the stimulus. There is a need to define clearly the stimulation parameters (such as frequency, duration, and interpulse interval) for specific brain regions and specific patient populations before rTMS can be used safely in clinics.
The excitatory changes mentioned above last only minutes at the longest. Therefore, although these phenomena may represent precursors of LTD and LTP, they may result in less durable changes. Studies exploring the combination of TMS and dopaminergic agents in an effort to enhance synaptic plasticity and improve function in patients with chronic stroke are under way.
Especially exciting for therapists is the combination of TMS and physical therapy interventions. Several studies have demonstrated that rTMS is capable of improving symptoms temporarily in a variety of neurological disorders, including movement disorders, depression, epilepsy, stroke and, more recently, chronic pain conditions.151,155,158–160 However, the effects are unreliable, modest, and short-lived. Perhaps one aim of the therapeutic application of TMS should be to help the brain reach a state in which it learns better. Once an optimal state of learning is reached, interventions can proceed. This strategy may allow physical therapy interventions to be more efficient.
In summary, understanding of the basic properties of TMS and its application to therapeutics is still elementary and currently provides only suggestions. The possibility for implementation by physical therapists appears to warrant further exploration.
| Footnotes |
|---|
Both authors provided concept/idea/research design, writing, data collection and analysis, subjects, project management, and fund procurement.
This work was presented as part of a platform series on neuroimaging of stroke rehabilitation at the Combined Sections Meeting of the American Physical Therapy Association; February 23–27, 2005; New Orleans, La.
| References |
|---|
 TopAbstractIntroductionFundamentals of Plasticity |
|---|
Introduction
Fundamentals of Plasticity
