PhD Thesis
The role of transcranial magnetic
stimulation in the investigation of the human motor system
Zsuzsanna Arányi, MD 2002
Semmelweis University, Budapest Faculty of Medicine
Department of Neurology
Consultant: Anita Kamondi, MD, PhD
Semmelweis University, School of PhD Studies in Neuroscience Committee for PhD examination:
Professor Péter Rajna, MD, DSc Professor Ferenc Mechler, MD, DSc Sámuel Komoly, MD, DSc
Referees:
Péter Diószeghy, MD, PhD Ferenc Nagy, MD, PhD
A transzkraniális mágneses ingerlés szerepe a humán motoros rendszer vizsgálatában
Dr. Arányi Zsuzsanna 2002
Semmelweis Egyetem, Budapest Általános Orvostudományi Kar Neurológiai Klinika
Témavezető: Dr. Kamondi Anita, PhD
Semmelweis Egyetem, Idegtudományok Doktori Iskolája Szigorlati bizottság:
Prof. Rajna Péter, MTA Doktora Prof. Mechler Ferenc, MTA Doktora Dr. Komoly Sámuel, MTA Doktora Hivatalos bírálók:
Dr. Diószeghy Péter, PhD Dr. Nagy Ferenc, PhD
INTRODUCTION AND BACKGROUND... 4
AIMS... 23
TRANSCRANIAL MAGNETIC STIMULATION IN RESEARCH: VOLITIONAL CONTROL OF PROXIMAL AND DISTAL ARM MUSCLES... 26
Facilitation of motor evoked potentials in proximal and distal arm muscles... 26
Ipsilateral / transcallosal responses in proximal and distal arm muscles... 38
Transcallosal inhibition during effort induced mirror movements... 54
Conclusions: motor control of proximal and distal arm muscles... 66
TRANSCRANIAL MAGNETIC STIMULATION IN CLINICAL DIAGNOSIS... 67
Assessment of the corticospinal tract... 67
Assessment of the facial nerve ... 88
CONCLUSION AND FUTURE DIRECTIONS... 104
SUMMARY... 105
ÖSSZEFOGLALÁS... 106
ABBREVIATIONS... 107
ACKNOWLEDGMENTS... 108
PUBLICATIONS... 109
REFERENCES... 112
INTRODUCTION AND BACKGROUND
Non-invasive transcranial electric stimulation of the human brain was first reliably achieved in 1980 by Merton and Morton.1 They were able to elicit contralateral electromyographic responses with a very short latency, thereafter named motor evoked potentials (MEP), compatible with the activation of the paucisynaptic, fast-propagating corticospinal tract. Electric stimulation involves however the use of current with very high intensity, which is painful and poorly tolerated by subjects. As a major advent in the field, Barker et al.2 introduced transcranial magnetic stimulation (TMS) in 1985.
This revolutionary new technique allows painless, non-invasive excitation of neural structures located deep in the body or covered by bone, such as the cortex, spinal nerve roots and the intracranial portion of the facial nerve. It has received wide attention and has been applied since in many different areas, including the research of nervous system physiology in humans and the diagnosis, monitoring and therapy of nervous system dysfunction.
After an overview of the technical and neurophysiological aspects of TMS, and of its various applications, we present our experimental work with transcranial magnetic stimulation concerning the volitional motor control of arm muscles and an analysis of our experience with TMS in the clinical setting. We thereby aim to illustrate the versatility and value of TMS both in the realm of research and clinical work.
Biophysical principles of magnetic stimulation
The technique of transcranial magnetic stimulation is based on the phenomenon of electromagnetic induction, first described by Michael Faraday in 1831.3 Accordingly, if a very brief, but strong electric current is passed through a coil of wire it generates a changing (time-varying) magnetic field (Fig. 1), which in turn induces a current in an adjacent wire circuit or volume conductor (Faraday's law).
Upon this principle, Barker et al. in Sheffield, UK constructed in 1985 a magnetic stimulator for the purpose of non-invasive stimulation of the human brain.2 The stimulator essentially consists of a bank of capacitors that is charged and discharged during the stimulation process, and of a stimulating coil, attached to the stimulator via about a meter long cable (Fig. 2). The storage capacitor system can be charged to a set level, determined by the examiner using controls on the front panel, up to a maximum of 2,800 volts. The levels are expressed in percentage of maximum and can be increased in increments as small as
1%. If the charged stimulator receives a trigger input signal, the energy stored in the
capacitor is discharged into the stimulating coil. The stored energy is transferred to the coil in approximately 0.1 ms and then returned to the instrument to reduce coil
heating. Most commercially available magnetic stimulators utilise an electronic discharge switch, which conducts current only in one direction and prevents current reversal, thereby producing a monophasic discharge current pulse (hence a monophasic
Figure 2. Main unit of a magnetic stimulator Magstim 200 of the Magstim Company Ltd., Spring Gardens, Whitland, UK.
(Reproduced from R. Jalinous, Guide to Magnetic Stimulation, The Magstim Company Limited, Spring Gardens, UK, 1995.)
Figure 1. Generation of magnetic field.
(Reproduced from R. Jalinous, Guide to Magnetic Stimulation, The Magstim Company Limited, Spring Gardens, UK, 1995.)
magnetic pulse). Biphasic or polyphasic pulses are less accurate and produce more click noise and heat.
The brief and strong discharge current of up to 5,000-8,000 amps flowing through the stimulating coil generates a magnetic pulse with a fast rise time (0.1 ms) and slower decay (up to 1 ms) (Fig. 3), and a peak magnetic field power of 1.5-4 Tesla. Magnetic field crosses high resistance
tissues unattenuated, as the scalp or the skull, without activating pain receptors, rendering TMS a well tolerable method. The rapidly changing magnetic field (the rising part of the magnetic pulse) induces a brief electric current in the neural tissue, which acts as a volume conductor. The intensity
of the induced current is proportional to the power of the magnetic field; it is in the range of 1-20 mA/cm2, similar to that used in conventional electric stimulation. Field strength falls off rapidly with distance (as the inverse square of distance), therefore stimulation strength is at its highest close to the coil surface. If the induced electric current is strong enough, it depolarises and discharges neural membrane.
The stimulating coil, housed in moulded plastic covers, consists of one or more tightly wound and well insulated copper windings, together with safety switches and
temperature sensors. There are a number of different stimulating coils available, designed for different purposes. The shape and size determines the output (peak magnetic field) of the coil and its stimulating characteristics (strength, depth of
penetration, size of stimulated tissue).4 The peak magnetic field power is related to coil size: smaller coils have a stronger field power (stimulation strength), but strength falls off more rapidly, therefore depth of penetration is smaller.
Figure 3. Monophasic magnetic field pulse. (Reproduced from R. Jalinous, Guide to Magnetic Stimulation, The Magstim Company Limited, Spring Gardens, UK, 1995.)
The coil most frequently employed for stimulation with the Magstim 200 stimulator is the large 90 mm round coil, with a peak magnetic field power of 2.0 Tesla (Fig. 5). The current induced by the round coil follows more or less the trajectory of the coil
windings, i.e. a current loop parallel to the coil is generated in the brain (Fig. 4). The magnitude of the induced current is equal at any point along the loop (around the coil winding), but it is zero in the central axis of the coil; by contrast the power of magnetic field is maximum at the centre of the coil. The
direction of the current flow in the tissue is opposite to that in the coil windings. It follows from the above that if a large round coil is placed on
the vertex in a tangential orientation for the purpose of motor cortex stimulation, it will activate both hemispheres underlying the coil windings, the motor cortices on either side. However, the hemispheres are preferentially activated by a current flowing in the posterior-anterior direction, therefore depending on the direction of current flow within the coil one hemisphere will be stimulated stronger than the other. In case of the most widely used Magstim 200 stimulator, that produces a monophasic magnetic pulse, if side A of the coil is facing upward, the current is flowing counter-clockwise within the coil and clock-wise in the brain, which leads to a preferential, but not exclusive
stimulation of the left hemisphere; and vice versa with side B. Polyphasic magnetic pulses are not direction sensitive, because either the first or the opposite polarity second phase may be effective.
The 90 mm round coil is also used to excite the spinal nerve roots and the intracranial portion of the facial nerve. For root stimulation it is placed tangentially on the
appropriate part of the spine, with side A facing upward for the stimulation of the roots on the right side, and vice versa. It has been shown that excitation of the roots occurs at
Figure 4. Current loops induced by tangential (1), saggital (2) and coronal (3) orientation of the round coil. (Reproduced from K.H. Chiappa, Evoked Potentials in Clinical Medicine, Raven Press, New York 1990.)
the level of the intervertebral foramina, bypassing the intraspinal segments.5,6,7 To date stimulation of the spinal cord itself has not yet proved possible. Excitation of the intracranial portion of the facial nerve is achieved by placing the coil behind the ear, in the parieto-occipital region.8
Large round coils are unsuited for focal, more circumscribed stimulation of the brain.
The so-called double coil (figure-of-eight, butterfly coil), containing two adjacent windings with opposing current directions, has been devised for this purpose (Fig. 5).
The standard 70 mm double-coil has a peak magnetic field power of 2.2 Tesla. Such a coil induces two current loops that are superimposed at the junction of the two loops, where a maximum in the magnitude of the induced electric field is formed. This results in a small preferential site of stimulation under the intercept (centre) of the coil.
Nonetheless, it is important to see that weaker stimulation can still take place under either side of the windings. The natural curvature of the head, however, helps keep the outer edges away from other areas of the cortex, further improving accuracy of
stimulation.
The double coil allows the isolated stimulation of one hemisphere; furthermore, by moving the coil along the central gyrus it is even possible to activate muscle groups separately within a limb. For example, if the centre of the double coil is located about 4 cm lateral to the vertex, the proximal arm muscles, whereas at 6 cm the small hand muscles will be activated. It has been found that using a double-coil the lowest
threshold responses are elicited when the centre of the coil is rotated around 45 degrees from the midline (the handle of the coil forms a 45˚ angle with the midline), thereby inducing a current perpendicular to the central sulcus. If the handle is pointing posterior, the direction of the induced current is posterior-anterior (preferential for the excitation of the cortex), and vice versa.9
Figure 5. 3D representation of the magnetic field produced by a 90 mm round coil (left) and a 70 mm double coil (right). (Reproduced from R. Jalinous, Guide to Magnetic Stimulation, The Magstim Company Limited, Spring Gardens, UK, 1995.)
The double cone coil is a less commonly used coil, where two large cup shaped windings are positioned side by side in an angle with a flat central section. The angled sides closely fit the patient's head, allowing a more efficient magnetic field coupling to the head. In particular this coil is suited for stimulation within the central fissure and the brain stem.
In spite of a relatively focal stimulation achieved by the double-coil, as opposed to the round coil, none of the coils provide a truly focal stimulation and indeed the precise site of stimulation (depolarisation) is always an approximation. For that reason, magnetic stimulation is less suited for the investigation of peripheral nerves, where the calculation of conduction velocity requires a positional accuracy to a few millimetres. Nonetheless, small round coils (40-50 mm) that have a stronger peak magnetic field power (4.1-3.6 Tesla), concentrated on a smaller area (i.e. stimulation is more focal than with large round coils), have been used for peripheral nerve stimulation, for example at Erb's point.
Physiology of magnetic stimulation
Magnetic stimulation of the cortex activates the rapidly conducting corticospinal tract, resulting in a contralateral muscle twitch. It has been shown that a single stimulus (magnetic pulse) triggers repetitive discharges of the pyramidal neurons, leading to multiple, up to 8 descending volleys in the corticospinal tract, separated by 1-1.5 ms intervals.10 The first wave is directly triggered in the pyramidal axon ('D' for direct wave); the following ones are transsynaptically elicited in the same neuron via one or more cortical interneurons ('I' for indirect wave). Temporal and spatial summation of these descending corticospinal waves of excitation takes place in the spinal alpha- motoneuron pool, leading to a progressive depolarisation of alpha-motoneurons until their threshold is reached and their action potential discharges. Thus, the degree of summation (number of descending waves) needed for discharge (i.e. muscle twitch) depends on the excitability status of the alpha-motoneurons.
A major difference in the mechanism of magnetic and electric brain stimulation
concerns the site of stimulation. Magnetic stimulation, as explained above, activates the pyramidal neurons not only directly at its axon, but also transsynaptically via
interneurons. Electric stimulation on the other hand excites only the pyramidal axon deeper in the white matter, thereby producing only D-waves and resulting in the shortening of the central motor conduction time by about 2 ms, as compared with magnetic stimulation.11 It also follows that the overall excitability level of the cortex and its spontaneous fluctuation influences responses to magnetic stimuli to a greater degree than to electric stimuli. This spontaneous fluctuation of excitability is partly reflected in the observed trial-to-trial variability in the latency and amplitude of MEPs elicited by a magnetic cortical stimulus of the same intensity.12 The difference in the effects of electric and magnetic cortical stimulation has been exploited in numerous instances to investigate cortical, as opposed to subcortical or spinal contribution in various phenomena.
The muscle compound action potential elicited by cortical stimulation is recorded by surface electrodes on the target muscle as the motor evoked potential. The parameters of MEPs measured include the onset (cortico-muscular) latency, the amplitude (baseline to
peak or peak to peak) and stimulation threshold. Since the pioneering studies of Merton and Morton, it has been observed that MEPs elicited during the voluntary contraction of the target muscle have a lower
stimulus threshold, are shorter in latency and larger in amplitude, in comparison to MEPs obtained during full muscular relaxation.13,14 This is termed facilitation, it reflects an increased excitability of the motor pathway during
voluntary contraction (Fig. 6). The most conspicuous aspect and best measure of facilitation is the increase in amplitude, which can be up to several fold in magnitude.
Latency shortening can reach 3-4
ms. Maximal facilitation occurs with the contraction of the target muscle, however there are numerous other facilitatory manoeuvres. As the amount of facilitation is an
indication of the excitability and its changes in the motor pathway, it is a tool to explore the physiology of motor control. The phenomenon of facilitation has been, accordingly, investigated in different muscles under different conditions; a detailed discussion will follow later.
The cortico-muscular latency of the MEP comprises the conduction time of both the corticospinal tract and the alpha-motoneuron. The conduction time in the corticospinal tract (central motor conduction time- CMCT) can be calculated by subtracting the conduction time of the alpha-motoneuron, which is obtained by magnetic or high- voltage electric stimulation of spinal roots, or using F-waves. It is generally 4-8 ms for the upper limb recorded from hand muscles, and 13-17 ms for the lower limbs recorded from the tibialis anterior muscle. CMCT is a reliable, if not the best parameter providing information on the integrity of the corticospinal tract; it is thus mainly used in the clinical setting, to assess corticospinal tract function. It is important to stress that the
Figure 6. Note the substantial increase in amplitude and reduction in latency of the MEP during contraction of the target muscle as opposed to muscle at rest.
interpretation of CMCT differs from that of the conduction time (conduction velocity) of peripheral nerves. Slowed peripheral conduction usually reflects demyelination in the peripheral nerve, whereas prolonged CMCT can be both a sign of demyelination and axon loss in the corticospinal tract. As explained above, the MEP is ultimately the result of the discharge of the alpha motoneuron pool, which requires the temporal and spatial summation of descending corticospinal volleys. Central demyelination and axon loss can both result in impaired summation. Demyelination leads to slowed conduction and thus summation takes longer; axon loss leads to a smaller number of descending volleys and therefore more consecutive volleys are needed for summation. The same principle applies to the interpretation of MEP amplitudes. Slowed central conduction, or
conduction failure caused by demyelinative conduction block or axon loss can all lead to insufficient summation, thus lowered amplitudes or absent responses. Nonetheless, an MEP of normal amplitude but prolonged CMCT speaks more for central demyelination without substantial conduction failure, whereas an MEP of very low amplitude with near normal CMCT speaks more for conduction failure due to central axon loss.
The interpretation of the MEP amplitude is further hampered by two factors. The amplitude of the MEP is usually smaller than the amplitude of the response evoked by peripheral nerve stimulation. Even with facilitation, the amplitude ratio between the response to peripheral and cortical stimulation can be as little as 18% in healthy individuals.15 Furthermore, the amplitude of the MEP varies considerably from one stimulus to another.12,16 It has been shown that the smaller amplitude of the MEP is the result of phase cancellation of the action potentials caused by desynchronisation of conduction occurring within the corticospinal tract or at spinal level.17 Variability of this desynchronisation also plays a role in the variability of MEP amplitudes.17 All these circumstances impede detection of central conduction failure in disease states. To circumvent this problem, recently the so-called triple stimulation technique (TST) has been developed.17 This technique involves the application of three stimuli with delays, leading to two collisions. The first is the magnetic cortical stimulus, followed by an electric stimulus at the wrist (recording is done from the abductor digiti minimi muscle).
The action potentials descending from the cortex collide with and cancel the antidromic potentials evoked at the wrist somewhere along the arm. After another delay a third
(electric) stimulus is applied at Erb's point, and it is this response that is finally studied;
its amplitude is compared to the amplitude obtained without magnetic cortical
stimulation. This way TST transforms the desynchronised activity of the corticospinal tract into the synchronised activity of the peripheral nerve, avoiding the effect of phase cancellation. If there is 'surviving' ascending activity after the first collision on the arm (i.e. activity from the wrist escaping collision with the cortical response), it collides with the response descending from Erb's point, thus reducing its size. In other words, if not all the axons are excited by the magnetic cortical stimulation the response from Erb's point will be smaller than without cortical stimulation. The degree of amplitude reduction of the Erb potential is an indication of the degree of central conduction failure; it quantifies the proportion of motor units activated by the cortical stimulus. It has been shown by TST that in healthy individuals magnetic cortical stimulation is able to excite virtually all motoneurons supplying the target muscle.
The threshold of cortical stimulation, together with the amplitude, is an indication of the excitability of the entire motor pathway, i.e. both of the cortex and spinal motoneuron pool. It is described as a percentage of the maximum stimulator output. Thresholds for muscles at rest and during contraction naturally differ, as facilitation reduces the threshold. A standard procedure for threshold measurement is recommended: threshold is defined as the stimulator output at which a response of at least 100 µV is elicited in 50% of 10 to 20 consecutive stimuli.18
Magnetic cortical stimulation produces not only excitatory, but also inhibitory phenomena in contralateral muscles. If stimulation is done while the subject is
voluntarily contracting the target muscle, MEP is followed by a transient inhibition of ongoing muscle activity, lasting for about 100-200 ms, referred to as the silent period (SP).19,20 It is generally proposed that the first 50 ms of the silent period reflects both cortical and spinal inhibitory mechanisms, whereas the latter 100 ms or more
exclusively reflects cortical mechanisms. Spinal mechanisms implicated in the first 50 ms include Renshaw inhibition and activation of inhibitory Ia interneurons by
descending input from the motor cortex. Both F-waves and H reflexes are inhibited during this period.21 The latter two-thirds of the SP is presumed to be cortical in origin.
This is derived from studies showing that a second magnetic stimulus given during this part of the SP elicits an inhibited MEP or none at all.20 Moreover, the SP elicited by electric stimulation is shorter than that elicited by a magnetic pulse, consistent with a lesser degree of contribution from cortical inhibitory mechanisms.20 The duration of SP is strongly influenced by stimulation intensity: SP duration increases linearly with increasing stimulation intensity even when preceding MEP amplitudes saturate.20 There is strong evidence that the MEP and the SP are physiologically dissociated, reflecting separate mechanisms.19 The silent period has a lower threshold and may occur without a preceding MEP. Thus, excitatory and inhibitory messages either run through separate descending routes from the cortex or travel along the same tract but are discharged at different thresholds.
The silent period has been applied to investigate the cortical inhibitory processes both in healthy individuals and in disease states; a shortening of the SP signifies damage to cortical inhibitory mechanisms. For example, it is shortened in lesions of the primary sensorimotor cortex22, but prolonged with lesions elsewhere, including the cerebellum, premotor cortex, parietal and temporal lobes, internal capsule, and thalamus.23,24 Also, it is shortened during motor tasks in writer's cramp.25
MEPs and silent periods are recorded in muscles contralateral to the hemisphere being stimulated. If recordings are done in ipsilateral muscles, early excitatory potentials (MEPs), reflecting ipsilateral fast conducting corticospinal connections, seldom occur, however inhibitory (silent) periods are consistently observed, especially in distal hand muscles.26,27 Meyer et al.27 showed that this response is absent in patients with
congenital agenesis or other abnormalities of the corpus callosum. It is now widely accepted that the origin of this inhibition is transcallosal, hence the name transcallosal inhibition. A more detailed discussion of transcallosal inhibition follows later.
Paired-pulse techniques
So far only single pulse magnetic stimulation has been addressed. It is possible to deliver magnetic stimuli pairs with two independent magnetic stimulators, connected to each other via a 'BiStim Module'. The BiStim Module triggers stimulation in both
stimulators with an interstimulus interval (ISI) set by the examiner. This is typically between 1 and 50 ms. The output of each stimulator can be individually adjusted. The stimuli can be discharged through the individual coils of the two stimulators, or through one coil attached to the BiStim Module (typically double coils are used for these
experiments). In the former case the stimuli are delivered at different sites of the brain, e.g. opposing hemispheres; in the latter stimuli are delivered to the same site.
Generally speaking, magnetic stimuli pairs are used to examine the effect of the first stimulus (conditioning stimulus) on the second one (test stimulus). The effect constitutes the change in amplitude of the MEP evoked by the test stimulus preceded by a
conditioning stimulus, in comparison to the amplitude of the MEP evoked by the test stimulus alone, without the conditioning stimulus. If the amplitude becomes smaller, the conditioning stimulus exerts an inhibitory effect, if larger the effect is facilitatory. There are a number of different ways to apply paired-pulse techniques. They differ in the site of stimulation, whether one or two coils are used, in the intensity of the conditioning and test stimuli with respect to motor threshold, and in interstimulus intervals.
Kujirai et al.28 performed the first systematic investigation of the paired-pulse technique. They applied a pair of stimuli through the same coil with varying interstimulus intervals, setting the conditioning stimulus to subthreshold intensity, followed by a test stimulus at suprathreshold intensity. They found that with ISIs of 1-6 ms a subthreshold conditioning stimulus inhibits the suprathreshold test stimulus. At ISIs of 7-20 ms, the test response is facilitated. These effects were termed as
intracortical inhibition and intracortical facilitation respectively. The arguments supporting the notion that the excitatory and inhibitory activity are each produced in separate neuronal circuits and that both reflect activity within cortical rather than subcortical circuits are the following: The low intensity of the conditioning stimulus does not evoke any descending activity in corticospinal axons, as shown by the lack of effect on the H-reflex, which is consistent with a cortical site of action. The inhibitory effect does not depend on the direction of the induced current, whereas the facilitatory effect is optimal with a posterior-anterior induced current as compared with a lateral- medial induced current, speaking for a separate circuitry for both effects.29
In addition to the investigation of facilitation and silent period, this technique provides further means to investigate the excitability changes of intracortical circuitries in conscious man under different circumstances and conditions. It has been applied in a number of neurological disorders and was found for example that the inhibitory phase may be diminished in various epilepsies30, Parkinson's disease31 and focal dystonia32. Intracortical facilitation appears to be increased in stiff-man's syndrome.33 A further and potentially very useful application of this paradigm relates to studies of the effects of various drugs on intracortical inhibition and facilitation.34 Modulators of
neurotransmission such as GABAergic drugs (benzodiazepines, baclofen) and antiglutamergic drugs (memantine, riluzole) increase intracortical inhibition.
Zolmitriptan (a serotonin receptor agonist) reduces intracortical inhibition. Conversely, drugs whose primary mechanism of action is targeted at inhibiting voltage-gated sodium channels (carbamazepine, phenytoin) affect threshold but specifically do not affect intracortical inhibition or facilitation.
Ferbert et al.35 used a modification of the paired-pulse technique to examine the effect of one hemisphere on the other. They applied a pair of suprathreshold stimuli on the hand area of opposing motor cortices and examined the effect of the conditioning stimulus ipsilateral to the target muscle on the test stimulus contralateral to the target muscle. They found that with ISIs of over 6 ms the conditioning stimulus inhibited the test response. They reasoned that this inhibition is probably produced via a transcallosal pathway; hence it is another way to demonstrate transcallosal inhibition.
In the paradigm investigating local intracortical inhibition and facilitation, a
subthreshold stimulus is followed by a suprathreshold stimulus through the same coil. In the paradigm of Felbert et al., two suprathreshold stimuli are applied to opposing
hemispheres. A third possibility is to apply pairs of identical near-threshold stimuli through the same coil at very short ISIs. At specific ISIs facilitation occurs, the size of the responses to each pair of stimuli was greater than the algebraic sum of responses to each stimulus alone. This was found at ISIs reflecting I wave periodicity, at 1.0 to 1.5, 2.5 to 3.0 and 4.5 or greater ms. Hence the phenomenon is called temporal I wave
facilitation.36,37 Using the paired pulse technique, it is also possible to demonstrate convergence of cortico-cortical inputs from premotor and parietal areas onto the motor cortex, referred to as spatial facilitation.38 In this paradigm, pairs of identical near- threshold stimuli were delivered through two coils located caudal and rostral to the optimal scalp position for the target muscle. If the stimulus was given alone at either of these positions, no response or very small response was produced. Given together, with ISIs of 0.5 to 1 or 2 ms, a clear response was obtained. The functional implication of this experimental observation is that commanding a voluntary movement from anterior frontal regions interacts with parietal information systems as to the state of body when movement is commanded.
The paired-pulse technique has also proved useful in intraoperative monitoring of corticospinal tract function during spinal or cerebral surgery.39,40 The purpose of intraoperative monitoring is to preserve function and prevent injury to the nervous system at a time when clinical examination is not possible. For a long time
somatosensory evoked potentials (SEPs) have been the standard tool for intraoperative monitoring of the spinal cord. Damage to the corticospinal tract can however occur without changes in SEP recordings, which points to the importance of monitoring corticospinal function as well. A major obstacle, however, in the utilisation of magnetic stimulation during surgery is that anaesthetic agents have significant suppressive effects on MEPs; MEPs can be completely eliminated during general anaesthesia, due to the reduced excitability of the cortex.41,42 Experiments show that D waves remain intact, but I waves are reduced or abolished, thus summation in the alpha motoneuron pool is insufficient.43 It was found that paired pulses or pulse trains of 3-6 pulses with short ISIs counteract the depressive effect of general anaesthesia.44 It mimics the natural multi-volley descending I wave activity and temporal summation thus occurs. A new generation of magnetic stimulators was devised for this purpose, called the Magstim QuadroPulse. This instrument is capable of producing trains of up to four pulses with very close (1 ms) ISIs. It contains four capacitor banks that can be discharged in sequence, or simultaneously to produce very high power levels. Using this method stable and large responses are obtained during anaesthesia.
Repetitive transcranial magnetic stimulation
Rapid rate or repetitive magnetic stimulation (rTMS) is defined as stimulation with a rate exceeding 1 Hz, but not normally more than 50 Hz. Unlike the paired-pulse technique, rapid rate stimulation does not allow individual adjustment of pulse intensity.
Rapid rate stimulators have lower output power, but are capable of delivering high frequency stimulation, with about 70% of the energy recovered after each pulse.
Conventional single pulse stimulators have a higher output but take at least a second to recharge, which does not allow stimulation with a frequency higher than 1 Hz. Rapid rate stimulators are subject to significant heating, which is prevented by a cooling system.
In contrast to single pulse stimulation in most instances, rTMS can transiently suppress (deactivate) certain higher cortical functions. It thus opened a new field in cognitive research and in the investigation of lateralized properties of dominant and non-dominant hemispheres. Studies have demonstrated for example that rTMS of the left temporal lobe produces counting errors or speech arrest45 and transient neglect46 or recall deficit47 may follow right parietal lobe stimulation. The identification of the dominant
hemisphere with rTMS would be of clinical value as well in certain situations.
Repetitive TMS in cognitive research is still a fairly new area and its role is yet to be determined.
Another, highly publicized area of application of rTMS is the treatment of patients suffering from major depression. In contrast to single, paired-pulse stimulation or rTMS in cognitive research, rTMS in this case is applied with the intention to produce a lasting (beneficial) effect on central nervous system function. Pascual-Leone et al.
demonstrated the role of the left prefrontal cortex in mood regulation using rTMS.48 Healthy subjects reported a significant increase in happiness following rTMS of the left prefrontal cortex, but not the right. There have been a number of studies since
examining the effect of rTMS on depressed patients and the results are controversial, improve in mood has not been observed in all studies. In depressed patients rTMS is performed on the left prefrontal cortex repeatedly, usually in daily sessions.49,50 To date, rTMS as currently performed does not reach the efficacy of electroconvulsive therapy
(ECT) and it cannot be considered as a replacement for ECT. The place of rTMS in the treatment of depression needs further investigation.
There have been also controversial results concerning the therapeutic effects of rTMS in patients with Parkinson's disease. There are observations that rTMS produces at least a temporary clinical improvement of symptoms51,52, however appropriate control
experiments have been a problem. In a recent review on TMS and Parkinson's disease, the therapeutic role of rTMS is still considered as an open question, needing further evaluation.53 TMS, however, has been successfully applied in the demonstration of multiple functional alterations of the corticospinal pathway in Parkinson's disease. For example, there is a shortening of the cortical silent period and a reduction in
intracortical inhibition in patients with Parkinson's disease.54 Levo-dopa restores these alterations.
Cortical mapping with transcranial magnetic stimulation
Using magnetic stimulation it is possible to map the cortico-motor
representation areas of muscles. Mapping studies are performed using a double-coil with the help of a grid of points 1 cm apart, placed on the scalp around the optimal position for producing MEPs in the target muscle. At each point several stimuli are delivered and the evoked responses are averaged. The number of positions from which an MEP can be elicited defines the cortico-motor representation area of the muscle under investigation (Fig. 7). This technique has proved especially useful in the
demonstration of both short and long-term plasticity, reorganisation of the motor cortex following damage to the brain, deafferentation or acquisition of new motor skills.
Changes occurring within the time span of minutes can be detected with TMS.
Figure 7. Construction of a motor output map. Stimulation is performed using a figure-of-eight coil at the points of the grid (left). The amplitudes of the corresponding MEPs (middle) serve to interpolate
isoamplitude lines to obtain a cortical motor map (right). The map's centre of gravity, the 'hot spot' is where the MEP has the largest amplitude or the lowest threshold. (Reproduced from Rösler et al., Transcranial magnetic brain stimulation: a tool to investigate central motor pathways. News Physiol Sci 2001, 16:297-302.)
An demonstration of long-term (structural) motor reorganisation was published by Mano et al.55 They examined patients with traumatic cervical root avulsion causing complete paralysis of the arm. Surgical anastomosis of an intercostal nerve into the musculocutaneous nerve was performed. Several months after the operation motor units were recorded in the biceps muscle, however they were only activated during
respiration and in response to cortical magnetic stimulation, voluntary activation was not possible. Gradually, 1-2 years after the operation, during which time the patients received intensive feedback therapy, voluntary elbow movements were observed. At this time the cortical representation area of the biceps shifted from the normal intercostal nerve region to a more lateral position, representing the arm area of the motor cortex.
An example of rapid cortical modulation is presented in the study of Pascual-Leone et al. with subjects learning simple one handed, five-finger piano exercises.56 The cortical motor areas of the contralateral long finger extensor and flexor muscles were
determined before and following a daily exercise of 2 hours over a course of 5 days.
The cortical maps enlarged substantially (ten-fold or even more) and motor threshold decreased. Interestingly, mental practice alone resulted in a map enlargement of similar, but slightly smaller magnitude. In another study Pascual-Leone et al. examined blind subjects proficient in Braille reading and found that the cortical map of the reading hand
was significantly larger than that of the non-reading hand or in control subjects.57 This difference was more robust following a day of intense reading. An even more rapid cortical modulation takes place following transient deafferentation of the arm below the elbow through nerve block. It was accompanied by a gradual and reversible
enlargement of the cortical representation area of the muscles immediately proximal the anaesthetised forearm.58 Conversely, the cortical map of the tibialis anterior muscle diminished in size as the result of immobilisation of the ankle due to bone fracture, without peripheral nerve lesion.59 The reduction could be quickly reversed by return of voluntary muscle contraction. It is suggested that rapid cortical modulation is achieved by unmasking of latent intracortical connections and long-term potentiation.60,61
Transcranial magnetic stimulation in clinical diagnosis
In the previous sections a number of examples were given that highlight the role of transcranial magnetic stimulation in the investigation of the physiological and
pathophysiological mechanisms of the nervous system. Its role in monitoring nervous system function during surgery and steps in the direction of therapeutic applications of TMS have been mentioned as well. TMS has also been successfully applied to aid the clinician in the diagnosis of neurological disorders. For reasons mentioned previously, TMS is less apt for the examination of the peripheral nervous system and it is applied mainly for the assessment of the corticospinal tract, both in disorders of the brain and the spinal cord. As with other evoked potential techniques, the value of TMS concerns mainly the demonstration of subclinical dysfunction of the corticospinal tract.
Peripheral applications of TMS include assessment of nerve segments inaccessible to conventional electric stimulation, such as spinal nerve roots or the intracranial portion of the facial nerve. All of these will be addressed in more detail later on.
Safety issues
Based on the wealth of experience accumulated since the introduction of commercial magnetic stimulators in the mid-1980s, single pulse transcranial magnetic stimulation, defined as non-rhythmic stimulation delivered at rates not exceeding 1 Hz, is considered as a safe method with virtually no side effects.62 Epileptic seizures have occurred extremely rarely, and always in the context of pre-existing epilepsy or stroke.63
Single pulse TMS does not alter cognition or memory. By contrast, repetitive
transcranial magnetic stimulation, defined as stimulation with rates exceeding 1 Hz, has resulted in seizures even in healthy individuals.64 High frequency stimulation, long trains of stimuli and short inter-train intervals carry a higher risk for provoking a seizure. Safety guidelines, with respect to the frequency and duration of trains of impulses, have been worked out.65 Repetitive TMS is also known to cause transient neurological deficits, exploited in research as mentioned above, but to date permanent adverse consequences have not been observed, even in patients after receiving as many as 50,000 cumulative rTMS stimuli for the treatment of depression. Nonetheless, caution is advisable with the use of rTMS and researchers should effectuate long-term follow-up of subjects or patients receiving rTMS.
Subjects with metallic objects or implants in the region exposed to magnetic stimulation (scalp, brain, neck, vertebral spine, sacrum) are excluded from magnetic stimulation studies, as they can be moved or disrupted by the strong magnetic field. These objects include aneurysm clips, cochlear implants, cardiac pacemakers, bullets or their
remnants, bone plates, foreign bodies in the eye, screws, spinal rods or wires in the spine. The magnitude of physical force the magnetic field exerts on objects depends on their size, conductivity and ferromagnetic characteristics. The effect on larger objects is more significant, up to 20 cm within the vicinity of the coil. Patients with skull defects, instability of the cervical spine are also excluded from magnetic stimulation. Epilepsy, pregnancy and young age are considered as relative contraindications, however in most laboratories these patients do not routinely undergo examination.
AIMS
Transcranial magnetic stimulation in research: investigation of volitional control of proximal and distal arm muscles
Differences in the volitional control of proximal and distal arm muscles in humans have long been a subject of interest. Distal muscles are involved in more precise, finely tuned and independent movements as opposed to proximal muscles;
proximal muscles on the other hand have a greater role in bilaterally synchronous postural responses. Moreover, the neurological syndromes that result from discharging or destructive lesions of the brain have unequal effects on these muscle groups. The distal muscles are usually the first to be involved by epileptiform activity66, are more severely affected by stroke than proximal muscles67, and they are more likely to be involved in mirror movements as well68.
These clinical observations are commonly believed to be due to differences in the corticospinal projections to these muscles. Much of our knowledge is based on experimental and anatomical knowledge derived from studies on primates. Cortical stimulation studies on monkeys have shown that monosynaptic projection to spinal motoneurons is more powerful for distal muscles, and provided evidence of bilaterally distributed pathways relaying in the brainstem for proximal muscles.69 The area of motor cortex associated with individual fingers was found to be greater than that for other parts of the arm. The existence of ipsilateral corticospinal projections and their difference in distribution to proximal and distal muscles in both primates and humans have also been often addressed, also with respect to their role in functional recovery after central lesions. It was shown in monkeys that 8% of the neurons in the primary motor cortex became active in relation to ipsilateral movement.70 Anatomical data have also demonstrated pathways from the motor cortex to the ipsilateral spinal cord in animals69 and humans71. In humans it is estimated that about 10% of the corticospinal tract remains uncrossed.72 Another point of difference between proximal and distal muscles is that, as shown in animal experiments, interhemispheric, callosal connections between the hand areas are fewer than for proximal muscles.73,74
One of the best ways to study human corticospinal projections, and volitional motor control in general, is the stimulation of the motor cortex. Before the development of non-invasive methods of stimulating the human brain, most of the information about the outflow of the human motor cortex was derived from direct intraoperative cortical stimulation studies.75 These studies appeared to support the distal predominance of corticospinal projection, as suggested by recordings in monkeys. Information was however limited, since only a limited number of sites could be explored in any one patient. Thus, the introduction of transcranial magnetic stimulation opened a new era in the study of human motor cortex in healthy human subjects.
With transcranial magnetic stimulation the corticospinal tract can easily be activated and short-latency, probably monosynaptic responses can be recorded in practically every single muscle studied. It is however important to take into account that TMS acts in a non-specific manner upon the motor cortex. It generates activation of many
motoneuron pools simultaneously, including agonists and antagonists. Furthermore, as described earlier, a single magnetic stimulus produces multiple descending volleys in the corticospinal tract. However, despite the complex nature of the descending input elicited by TMS, it has been shown that TMS produces orderly recruitment and rate coding of spinal motoneurons similar to that found for voluntary activation76, referred to as Henneman's size principle77. Accordingly, excitatory input into the spinal
motoneuron pool recruits motoneurons in a size-related fashion: small, low threshold motoneurons are recruited first, followed by large, high-threshold motoneurons. It is the similarity between the recruitment pattern of TMS and voluntary input, which renders TMS particularly suitable for the investigation of volitional motor control. While most of the attention was centred on distal hand muscles up to now, a systematic
investigation of the control of proximal arm muscles is lacking.
Using the method of transcranial magnetic stimulation, we have set out to examine and to compare several aspects of the volitional control of proximal and distal arm muscles:
1. Investigation of contralateral corticospinal projections and excitability changes of the motor pathway during different tasks in proximal and distal arm muscles.
2. Investigation of ipsilateral projections and transcallosal influences in both proximal and distal hand muscles.
3. Investigation of the modulation of transcallosal influences.
In the first part of the thesis results of these experiments are presented and discussed. A methodological consideration restricted our investigation to arm muscles. The cortical motor representation area of the arm lies exposed on the convexity of the hemisphere, whereas that of the leg is hidden in the interhemispherical fissure. Thus, in case of leg muscles selective stimulation of different muscle groups is impossible.
Transcranial magnetic stimulation in clinical diagnosis
As with all new methods, the diagnostic significance of transcranial magnetic stimulation is still under evaluation and is a matter of discussion. TMS has been applied mainly in neurological disorders such as multiple sclerosis, amyotrophic lateral sclerosis and stroke; in other disorders its role is less clear. We have undertaken to analyse the results of patients who were examined by TMS in our department over the course of 4 years:
1. Analysis of MEPs of patients referred to us for the assessment of the corticospinal tract. Particular attention was paid to patients with cervical
spondylosis, to test the sensitivity of MEP in detecting compressive myelopathy secondary to cervical spondylosis, even before clinical pyramidal signs appear.
2. Analysis of the results of patients referred to us for the assessment of the facial nerve, using TMS.
The second part of the thesis concerns the presentation of this analysis. We have endeavoured thereby to provide further clues as to the place of TMS in the clinical diagnostic work-up.
TRANSCRANIAL MAGNETIC STIMULATION IN RESEARCH:
VOLITIONAL CONTROL OF PROXIMAL AND DISTAL ARM MUSCLES
Facilitation of motor evoked potentials in proximal and distal arm muscles
The amount and pattern of facilitation with any manoeuvre or task reflects the excitability and its changes in the motor pathway under that particular condition. It is a question whether the site of facilitation or change in excitability resides in the cortex or in the spinal cord. In other words, voluntary activation of the muscle may enable a larger descending volley to be produced by the magnetic stimulus reflecting an increased excitability of the cortex. Alternatively, the descending volley may remain unchanged but excite a larger population of spinal motoneurons because their thresholds have been lowered by the contraction. As a third possibility, a change in excitability may take place on both levels. It is however most likely that the site of facilitation depends on the type of facilitatory manoeuvre.
Hess et al. examined the mechanism and possible site of facilitation during contraction of the target muscle in case of small hand muscles.16 They found that the force of voluntary contraction (level of background electromyographic- EMG activity) related linearly to the degree of facilitation when using weak stimuli. Spinal reflex mechanisms are enhanced as well and in a similar linear fashion by pre-existing motoneuronal activity.78 Thus, they concluded that a similar spinal mechanism determines the efficacy of the descending excitatory volley set up by brain stimulation. Facilitation could also be produced by the contraction of homologous hand muscles contralateral to the target muscle, which could be explained by the irradiation of excitability in the spinal
motoneuron pool to the motoneurons of the target muscle.
In addition to this spinal mechanism, there seems to be another mechanism of
facilitation, which comes into action at very low background forces. Hess et al. found that after a steep increase, facilitation becomes saturated already at very low
background force levels (less than 10% of maximal voluntary force) if the target muscle itself or another hand muscle on the same hand contracts. However, this marked
facilitation could not be produced to the same extent by electric stimulation of the brain, where the linear relationship between background force and MEP amplitude is seen over a larger range.79 They concluded that when a subject focuses his attention on the motor performance of a particular hand, then there is a preferential rise in excitability of the pathways to that hand, which does not depend on the degree of pre-existing
voluntary force exerted; this mechanism would most probably reside in the cortex. The fact that it is not seen with electric stimulation of the brain would favour this idea, since electric stimulus with its presumed direct action on pyramidal axons, in part bypasses the intracortical neuronal elements, which would allow for such a regulation of the descending volleys. Furthermore, they also showed that muscle-afferent input, which could theoretically influence both spinal motoneurons and pyramidal cells, was not needed for facilitation since the illusory movement in the phantom limb of an amputee could also facilitate responses on the contralateral side.80
Kischka et al. compared the pattern of facilitation in a small hand muscle and a proximal muscle (biceps brachii) during steady innervation.81 They found that in the biceps saturation of MEP amplitudes occurs later and shows a more gradual increment with background contraction force, as opposed to hand muscles. They explained this difference by a smaller contribution of cortical facilitatory mechanisms, in line with a less pronounced corticospinal control of proximal muscles. Similarly, Mathis et al.
found that facilitation in the deltoid muscle seems to be mainly of spinal origin, since both voluntary contraction of cortical origin and involuntary contraction (Kohnstamm phenomenon) of presumably subcortical origin produced the same amount of
facilitation.82
As mentioned before, the most effective way to facilitate an MEP is the voluntary contraction of the target muscle, but there are numerous other methods. Facilitation of
MEPs can be produced by observing movements performed by others83 and by non- specific manoeuvres such as protruding the tongue or counting aloud84. Furthermore, facilitation occurred during the period just prior to the movement period (preparatory phase)85 and also by just thinking about the movement86. In all these conditions an increase in cortical excitability appears to play the major role.
In summary, it is seen that in addition to the contraction of the target muscle there are a number of different facilitatory manoeuvres, and that the mechanism and site of
facilitation depends on various factors, such as the type of facilitatory manoeuvre, the target muscle, the amount of background force and even the timing of the stimulus.
There is however a further factor influencing facilitation, namely the type of motor task the target muscle is involved in. Task-related differential facilitation has been described when comparing various types of simple tasks, complex tasks and simplex versus complex tasks. Simple tasks are such where only the target muscle is contracting, whereas in complex tasks several muscles are simultaneously active in addition to the target muscle. For example, Flament et al. examined facilitation in the first dorsal interosseus muscle, when it performed an isolated finger abduction (simple task) and when it was involved in a power grip (complex task).87 They found that facilitation was greater during a power grip, and they attributed this difference to a cortical mechanism.
Another example is the modulation of facilitation in hand and forearm muscle across the different phases of a 'reach-grasp-lift task'.88 A cortical change in excitability was
proposed as well for the varying facilitation across this complex task. However, the interpretation of facilitation in the target muscle in complex tasks is impeded by the effect of the contraction of the other muscles and the afferent influences from joint and skin receptors.
It is reasonable to assume that variation of the excitability at either the cortical or the spinal level across different tasks accounts for the task-dependent differences in MEP amplitudes. In other words, the cortical stimulation will lead to the discharge of a preselected set of pyramidal cells and/or spinal motoneurons depending on the task, possibly reflecting an imprinted motor pattern. Our aim was to examine the differences of these 'patterns' between proximal and distal arm muscles by examining the
differential facilitatory effect of simple tasks, i.e. dynamic versus steady state contraction. A dynamic contraction refers to a contraction with slowly and smoothly increasing force; in a steady state the contraction force is held constant.
Methods
23 healthy subjects participated in this study. None of them had a history of previous neurological disorders, implanted metal in the eye or in the brain or a cardiac pacemaker. Informed consent was given by all subjects and the study was approved by the local ethics committee. Transcranial magnetic stimulation was performed by
Magstim 200 stimulator (The Magstim Company Ltd., Spring Gardens, Whitland, UK), using a 70 mm double-coil (figure-of-eight coil). The coil was positioned tangentially on the skull, contralateral to the target muscle with the coil current in the intercept flowing in an anterior-posterior direction, resulting in a posterior-anterior current flow in the brain tissue. The optimal coil placement for eliciting contralateral responses was determined in each subject by searching the coil position that yielded the lowest stimulation threshold during minimal constant innervation of the target muscle.
Stimulus threshold was defined as the intensity of stimulator output with which MEPs were elicited in 50% of the trials.18 In most subjects the optimal position of the coil was 4 cm lateral to the vertex for the deltoid muscle and 6 cm lateral for the abductor digiti minimi (ADM) muscle. The experiments were carried out with a comparatively low stimulus intensity of 5% above stimulus threshold. The low stimulus intensity was chosen to avoid saturation effects. Recordings were made from the deltoid and ADM muscles using surface electrodes in the conventional belly-tendon arrangement. Along with the electromyographic (EMG) measurements, under non-isometric circumstances the abduction angle of the arm or finger was recorded and calibrated in degrees by use of a precision rotatory potentiometer (Burster, Gernsbach, Germany). Under isometric circumstances force level was measured by a force transducer. The mechanical signals were also fed into an oscilloscope in front of the subject, to allow a visual feedback of the movement, position or the force level of the finger or arm. The EMG and
mechanical signals were DC-amplified using a 1902 programmable signal conditioner (CED®, Cambridge, UK.) and sampled at 4 kHz by a stand alone AD converter (MacLab, ADInstruments Pty Ltd., Castle Hill, NSW, Australia) connected to a
personal computer (Macintosh, Apple Computer Inc., Cupertino, CA, USA). The
deltoid and ADM muscles were examined in separate experiments. The data were stored on hard disc for later off-line analysis. Additional filtering was usually avoided, but occasionally 50-100 Hz high-pass filtering was needed to eliminate low-frequency artefacts in the EMG traces.
For the deltoid muscle in the non-isometric experiment (Fig. 8), the subjects sat on a chair with their arm hanging freely along their
side. In the dynamic task, the subjects slowly abducted the arm. The abduction angle was measured by the precision potentiometer connected to a lever that moved together with the arm. The magnetic cortical stimulus was triggered automatically using a custom made level discriminator, when the arm reached a preset abduction angle. Recordings were made at three different abduction angles between 5º and 45º. In the steady task, the subjects held the arm steadily abducted for at least 3 s before
the stimuli were given manually by the examiner, at roughly the same abduction angles as for the dynamic task. Abduction angles were slightly adjusted to yield similar background EMG levels as with the dynamic task. Five recordings were collected for each abduction angle in each task.
In the isometric experiment, the arm was fixed along the side of the subject at the wrist and attached to a force transducer. First maximal voluntary force was determined. The dynamic and steady tasks were performed essentially the same way as in the non- isometric experiment, only the arm was not allowed to move and stimulation was carried out at different force levels not abduction angles. In the dynamic task, the subjects slowly increased the force level, until a preset force level was reached, and stimulation was triggered by a custom made level discriminator. Three different abduction forces between 2.5 and 20% of maximal voluntary force were examined. In
Figure 8. Schematic representation of the experiment on the deltoid muscle.
the steady task, the subjects held the force steadily for at least 3 s before the stimuli were given by the examiner, at roughly the same force levels as for the dynamic task.
Five recordings were collected for each force level in each task.
In three subjects magnetic stimulation of the brain stem according to Ugawa et al.89 was also performed using a double cone coil (consisting of two large loops set at an angle of 110º), under non-isometric conditions and at one angle (5º), recording form the deltoid.
The centre of the coil was positioned at the inion.
For the ADM muscle (Fig. 9), measurements were only performed during isometric conditions, since preliminary experiments showed that the background EMG activity under non-isometric circumstances is not comparable in the dynamic and steady tasks.
The left forearm and hand was fastened on a platform, with the fifth finger attached to a force transducer. In the dynamic task, subjects were asked to slowly increase the abduction force of the little finger until stimulation was triggered automatically by a custom made level discriminator at a preset force level. Three force levels were
examined. In the steady task, the subjects maintained the force at the same levels and the examiner performed the stimulation manually. Ten trials were collected for each force level in each task.
The EMG and the mechanical signals (i.e. abduction angle and force level) were both recorded, separately for each stimulus. Rectification and averaging of the recordings corresponding to the same set of conditions was performed off-line. Recording epochs of 500 ms duration were taken, of which 100 ms preceded the stimulus. The background EMG activity was defined as the mean amplitude of the rectified and averaged EMG
Figure 9. Schematic representation of the experiment on the ADM.
during the 100 ms preceding the stimulus. The mean amplitude of the MEP was
similarly calculated from the rectified and averaged traces. Latencies were not analysed.
Comparison of tasks, statistical analysis of MEP amplitudes was performed using the non-parametric methods the Wilcoxon's signed-rank test (paired) or the Mann-Whitney rank-sum test (unpaired). The level of significance was set at P<0.05.
Results
For the deltoid, in each subject in each trial a clear short latency MEP, followed by a silent period was recorded (Fig. 10), with an average cortico-muscular onset latency of 8.2 ms (SD 1.5). MEP amplitudes increased linearly with background EMG under both isometric and non-isometric conditions, and in both dynamic and steady tasks (Fig. 11). Note that background EMG levels tended to be greater during isometric conditions. At identical background EMG levels, the MEP amplitudes were
significantly greater during the dynamic than during the steady tasks (Fig. 11). There was no essential difference in this respect between non-isometric and isometric conditions. At the highest level of background EMG, which was only reached under isometric conditions, the task-dependent difference between the dynamic and steady tasks did not reach statistical difference.
Magnetic brain stem stimulation was performed in three subjects, recording from the deltoid. Brain stem MEPs could be recorded in each subject, with latencies slightly shorter than those elicited by cortical stimulation. Due to the small number of subjects, here the MEPs were not averaged and statistical comparison between tasks was
performed separately for each subject using individual trials. The amplitudes of these brain stem MEPs were significantly greater during the dynamic task than during the steady task in all 3 subjects (Fig. 12).
Figure 10. Recordings from the deltoid muscle (non-isometric experiment) from one subject. Stimulation at 5° abduction angle. The upper trace shows the rectified and averaged EMG (n = 5 sweeps). Stimulus is at 0 ms time, followed by a motor evoked potential at 8 ms, the mean amplitude of which was taken for further analysis. The lower trace shows the simultaneously recorded position of the arm in degrees of abduction angle. Note the difference in the MEP amplitude of the responses in the two tasks, in spite of identical background EMG level and equal stimulus intensity.
Figure 11. Results from the deltoid muscle under non-isometric (a) and isometric (b) conditions in 11 subjects. Mean MEP amplitudes (±SEM) are plotted against background EMG (±SEM). Note the greater MEP amplitudes during the dynamic tasks at identical background EMG levels, which is most
pronounced at low background EMG levels. Wilcoxon P values are given.
Figure 12. Brain stem stimulation in 3 subjects: relationship between MEP amplitudes and background EMG separately for each subject. Amplitudes were significantly greater if stimulation was performed during the dynamic arm abduction (subject 1, P< 0.001; subject 2, P= 0.015; subject 3, P< 0.001; Mann–
Whitney test). The average background EMG levels were not statistically different in subjects 2 and 3, but slightly different in subject 1 (P= 0.03; Mann–Whitney test).
For the ADM, in each subject a clear short latency MEP, followed by a silent period was recorded with an average cortico-muscular onset latency of 18.5 ms (SD 1.5). MEP amplitudes increased linearly with the background EMG in both dynamic and steady tasks. In contrast to the deltoid muscle, MEP amplitudes in the ADM showed no significant difference between the dynamic and steady tasks (Fig. 13).
Figure 13. Results from ADM in 8 subjects. The mean amplitudes (±SEM) of the MEP are plotted against the mean background EMG (±SEM) separately for the steady and dynamic task under isometric
conditions. No task-dependent differences of the MEP amplitudes were observed (P> 0.05; Wilcoxon's test).
Discussion
The results presented above demonstrate that in the deltoid muscle the degree of facilitation (MEP amplitudes) depend not only on the level of contraction (background EMG activity), but also on the type of contraction. In other words, not only the quantity of the ongoing muscle activity but also the quality determined the amount of facilitation.
Specifically, facilitation was greater during a dynamic contraction, when the stimulus was applied during an increasing force, than during a steady contraction, when the force was not changing, despite comparable background EMG levels. The difference was seen with both cortical and brain stem stimulation of the corticospinal tract. In contrast to the deltoid, this task-dependent difference in facilitation was not observed in the ADM.
The number and firing frequency of suprathreshold motoneurons activated by the voluntary contraction is reflected in the amount of EMG activity just prior to the stimulus (background EMG activity). The size of the MEP superponated on the background EMG activity can be regarded as a rough estimate of the still excitable (recruitable) portion of the pools of pyramidal cells and/or spinal motoneurons that has not yet reached firing threshold during voluntary activation. This 'recruitable portion' has been termed as the subliminal fringe.90 In other words, the degree of facilitation depends on the size of the subliminal fringe at the time of the stimulus, which in turn is an indication of the change in excitability.
The larger MEP size in the dynamic task of the deltoid suggests that a greater number of motoneurons was near firing threshold and could be brought to discharge by the
stimulus. Such a pool of motoneurons, the subliminal fringe, appears to be greater during a dynamic contraction. The question arises whether this change occurs at the level of the cortex or the spinal cord. The same pattern of facilitation was detected with stimulation at the level of the brain stem, which strongly implies that the site of this 'extra' facilitation (i.e. the increased size of the subliminal fringe) resides in the spinal cord. Abbruzzese et al. examined likewise task-dependent differences in the
brachioradial muscle during shortening as compared to static and lengthening
