Brain (1992), 115, 1947-1961
STIMULATION OF CORTICOSPINAL PATHWAYS AT THE LEVEL OF THE PYRAMIDAL DECUSSATION IN NEUROLOGICAL DISORDERS t
by YOSHIKAZU UGAWA, KIEKO GENBA, TORU MANNEN ICHIRO KANAZAWA
and
{From the Department of Neurology, Institute of Brain Research, School of Medicine, University of Tokyo) SUMMARY A newly developed technique of brainstem stimulation was applied in 14 normal subjects and 23 patients with various neurological disorders. The electromyographic (EMG) responses of limb muscles following cortical, brainstem and cervical stimulation were recorded. The cortical-brainstem conduction time and brainstem-cervical conduction time were then calculated from the difference in latency between the two sites of stimulation. From the regression lines of the relationship between the normal conduction times in the first dorsal interosseous muscle and the length of the descending tracts, the site of activation by brainstem stimulation was estimated to lie near the cervical-medullary junction. The most distal lesion causing prolongation of cortical-brainstem conduction time was a small cerebral infarction in the lower pons. Herniation of the third cervical spinal disc was the most rostral lesion resulting in delayed brainstem-cervical conduction time and normal cortical-brainstem conduction time. These observations suggest that activation occurs at the level of the cervical-medullary junction where the pyramidal decussation lies. The conduction velocities of the activated tracts estimated from the regression lines for normal individuals were 57—92 m/s. In patients with supratentorial lesions, the threshold for brainstem stimulation was abnormally high. The abnormal findings in this test were correlated significantly with the clinical pyramidal signs. This suggests that the EMG responses elicited by brainstem stimulation are mediated mainly by the corticospinal tract. We conclude that the brainstem stimulation technique would be clinically useful for localization of lesions in the corticospinal tract; the primary lesion can be localized whether above or below the pyramidal decussation. INTRODUCTION
The technique of percutaneous electrical stimulation has been widely used to evaluate central motor conduction in patients with various neurological diseases. Central motor conduction time, which is usually used in these studies, is calculated by subtracting the latency of the muscle response to spinal stimulation from the latency of the response produced by cortical stimulation (Marsden et al., 1982; Ugawa et al., 1989a). Abnormalities in the central motor pathways are expressed as prolongation of central motor conduction time or a lack of response evoked by cortical stimulation (Ugawa et al., 1988a,b). Such abnormal findings are seen in patients with lesions within the central motor pathways anywhere between the motor cortex and the spinal cord. Therefore, the usual method using central motor conduction time is not sufficient in Correspondence to: Dr Y. Ugawa, Department of Neurology, Institute of Brain Research, School of Medicine, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113, Japan. © Oxford University Press 1992
1948
Y. UGAWA AND OTHERS
clinical practice for localizing lesions in the central motor pathways. It has been reported recently that the descending motor tracts can be activated at the level of the pyramidal decussation in humans using high voltage electrical stimulation across the base of the skull (Ugawa et al., 1991). However, it is still unclear which pathways are activated and where the activation occurs in this new method. We have used normal volunteers in studies aimed at clarifying the site of activation and for establishing normal values. We have applied this new technique to localizing lesions within the central motor pathways in patients with various kinds of neurological disorders. In this paper, we describe the estimation of the site of activation using this method and its potential usefulness in clinical neurophysiology. SUBJECTS The subjects were 23 patients with various neurological disorders: nine with cerebrovascular disease, four with cervical spondylotic myelopathy, three with amyotrophic lateral sclerosis, one with clinically definite multiple sclerosis and one with leukodystrophy. Four patients with spinocerebellar degeneration and one with Fisher's syndrome, showing no pyramidal signs, were also included. In order to obtain normal values, 14 healthy volunteers (10 men and four women, aged 2 6 - 4 2 yrs and 154-188 cm in height) were studied. All of the subjects gave informed consent and the procedures were approved by the Ethical Committee of Tokyo University. No side-effects were noted in any individual. METHODS Surface electromyographic (EMG) activities were recorded simultaneously from biceps brachii, extensor carpi radialis, flexor carpi ulnaris, first dorsal interosseous and tibialis anterior muscles in normal volunteers, using surface cup electrodes. The first dorsal interosseous muscle was studied in all patients, and also some of the other muscles listed above were selected for investigation according to the clinical features of the patients. Electromyographic signals were amplified with filters set at 80 Hz and 3 kHz, recorded and analysed by a Signal Processor 7T18 (NEC San-Ei) or a Neuropack 4 (Nihon Kohden). We performed electrical stimulation of three sites within the central motor pathways (motor cortex, brainstem and cervical spinal cord or roots) using a high voltage electrical stimulator described previously (Ugawa et al., 1989a). This stimulator is capable of delivering stimuli of up to 1000 V with a decay time constant of 50 jis. Stimulus intensities are expressed as the voltage charged by the stimulator. For cortical activation of arm muscles, the cathode was placed at the vertex and the anode 7 cm lateral. These electrode polarities were reversed in order to activate leg muscles. For cervical stimulation, the anode was fixed over the fifth cervical spinous process and the cathode over the first thoracic spinous process. The descending motor tracts were activated at the level of the brainstem (brainstem stimulation) using a technique described recently (Ugawa et al., 1991). The stimulating electrodes were fixed with collodion on both sides of the head at the angle formed by the posterior edge of the mastoid process and the anterior part of the occipital bone (described as site B in Ugawa et al., 1991). The anode was on the right side and the cathode on the left. For every stimulation, the leg muscles were preactivated by slight voluntary contraction. For cortical and brainstem stimulation, the arm muscles were studied in an active state. However, they were relaxed during cervical stimulation. Stimulus intensity was adjusted to produce a response of maximum amplitude and shortest latency. When a maximum response could not be elicited, we used the maximum output of our stimulator. In all experiments, stimulation on each site was repeated at least twice to assess the reproducibility of the findings. The following three conduction times were calculated for each muscle. The cortical-brainstem conduction time was obtained by subtracting the latency of the response produced by brainstem stimulation (brainstem latency) from that of the response evoked by cortical stimulation (cortical latency). The brainstem-cervical conduction time was calculated by subtracting the latency of the response elicited by cervical stimulation (cervical latency) from the brainstem latency. The conventional central motor conduction time of arm muscles was also obtained by subtracting the cervical latency from the cortical latency. The results obtained from
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1949
the normal volunteers were used as normative data for these conduction times. The above conduction times were considered to be prolonged if they exceeded the upper limit of the normal range (mean +2.5 SD). We studied correlations between the physiological findings in the patients and their clinical signs. RESULTS
Normal subjects Consistent motor responses could be obtained from all muscles in all normal subjects. The thresholds for activation of any muscle were 200 — 350 V. The average latencies and conduction times for every muscle are presented in Table 1. These values all correspond well with those in the previous report (Ugawa et al., 1991). The site of activation in the brainstem stimulation was estimated from our data for latencies in normal subjects. The majority of the brainstem-evoked EMG responses were thought to be mediated by the corticospinal tract (Ugawa et al., 1991). In the first dorsal interosseous muscle especially, almost all components of the potentials are mediated by the corticospinal tract because this tract projects strongly to the intrinsic hand muscles (Clough et al., 1968; Alstermark and Sasaki, 1985; Iwatsubo et al., 1990). Therefore, we used the data obtained from first dorsal interosseous muscles for estimating the level in the corticospinal tract where activation occurred. The actual time in which the nerve impulses are conducted from the motor cortex to the activation site of the brainstem stimulation is the measured cortical-brainstem conduction time. On the other hand, brainstem-cervical conduction time includes synaptic delay at the alpha motor neuron and conduction time along the motor roots in the spinal canal, as well as the conduction time from the brainstem to the spinal cord. The conduction time in the motor roots is reported to be 0.6—1.4 ms (Mills and Murray, 1986; Ugawa et al., 1989a,b). Our mean value was about 1 ms (Ugawa et al., \9%9a,b). If we assume that the synaptic delay is 1 ms, then the actual conduction time from brainstem to spinal cord is calculated as 'measured brainstem-cervical conduction time — 1 ms (conduction time along motor roots) - 1 ms (synaptic delay)', i.e. brainstem-cervical conduction time — 2 ms. The mean distance from the cervical-medullary junction to the first thoracic spinal cord (innervation level for first dorsal interosseous) is 10.9 cm (Donaldson and Davis, 1903). The mean length of the corticospinal tract from the cortex to the cervical-medullary junction is estimated to be 14.2 cm from anatomical data (Shimada, 1939). This approximation of the distance from the motor cortex to the first thoracic spinal level
TABLE I. NORMAL LATENCIES AND CONDUCTION TIMES OF RESPONSES TO BRAINSTEM STIMULATION IN VARIOUS MUSCLES
Biceps Forearm flexor Forearm extensor First dorsal interosseous Tibialis anterior
CTX (ms)
BST (ms)
CV (ms)
CTX-BST CT (ms)
10.2±1.0 13.6± 1.2 13.7±!.O 19.2± 1.9 25.5 ±1.5
8.5±0.4 ll.7±l.l 1I.9±1.O 17.3±1.5 23.9± 1.7
5.3±0.3 8.4±1.0 8.2±1.0 13.6±1.6 21.3±1.5
1.8±0.3 1.9±0.4 1.7±0.3 1.8±0.3 1.7±0.3
BST-CV CT (ms) 3.I±O.2 3.3±0.9 3.5±0.5 3.7±0.5 2.3±0.5
Values are mean ±SD. The values are taken from 16 subjects. CTX = cortical latency; BST = brainstem latency; CV = Cervical latency; CTX-BST CT = cortical-brainstem conduction time; BST-CV CT = brainstem-cervical conduction time.
1950
Y. UGAWA AND OTHERS
(25.1 cm = 10.9 cm+14.2 cm) is consistent with another estimation of the distance from the parietal cortex to the first thoracic level of 24 cm (Desmedt and Cheron, 1983). Assuming that conduction velocity in the corticospinal tract is the same at all levels from the cortex to the spinal cord, we plotted the distance from the cervical-medullary junction as a function of the conduction times in 14 first dorsal interosseous muscles. (Fig. 1.). The site where activation occurred during brainstem stimulation was estimated to be between 2.2 cm rostral and 1.5 cm caudal from the cervical-medullary junction. This was compatible with the previous conclusion that the corticospinal tract was activated at the level of the pyramidal decussation (Ugawa et al., 1991). The conduction velocity in the stimulated tract was also estimated to be 5 7 - 9 2 m/s from the slopes of the regression lines. This calculation of the conduction velocity gave only an approximate value, since it was based on average lengths in post-mortem specimens. Despite this, these values fitted well with other estimates of conduction in the descending motor tract in humans (Marsden et al., 1982; Ugawa et al., 1989a). Patients We first present five typical cases. Patient 1 (Case 2 in Table 2). A 62-yr-old woman developed flaccid tetraparesis. T2-weighted magnetic resonance images (MRIs) of the brain demonstrated a high
Cortex 150-, 100,. 50-
^^^*^*^
Estimated * ofactivati
0-50100150-3
VSRH^V •
i
-2
-1
Conduction time from cortex (ms)
1
TI
2
Conduction time to spinal level (ms)
FIG. 1. The relationship between conduction time in the first dorsal interosseous muscle of 14 different subjects and the length of the descending tract. The average length from the motor cortex to the cervical-medullary junction was 142 mm (Shimada, 1939) and that from the cervical-medullary junction to the first thoracic spinal cord (Tl) innervating the first dorsal interosseous muscle is 109 mm (Donaldson and Davis, 1903). The y-axis represents the length from the cervical-medullary junction. The x-axis represents the conduction time. The conduction time from the motor cortex to the cervical-medullary junction is plotted on the minus side. That from the cervical-medullary junction to the innervating spinal level, calculated as 'brainstem-cervical conduction time —2', is plotted on the plus side. The points where the regression lines cross the y-axis are the sites where activation occurs in brainstem stimulation. These points are from — 15 mm to +22 mm, which means from 22 mm rostral to 15 mm caudal to the cervical-medullary junction. The slopes of the lines are - 9 2 to - 5 7 m/s.
B R A I N S T E M S T I M U L A T I O N IN N E U R O L O G I C A L D I S O R D E R S
1951
TABLE 2. THE PHYSIOLOGICAL RESULTS IN THE FIRST DORSAL INTEROSSEOUS MUSCLE IN 23 PATIENTS Case
Diagnosis (level of lesion)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Cerebrovascular disease (subcortex) Cerebrovascular disease (pons) Cerebrovascular disease (subcortex) + leprosy Cervical spondylotic myelopathy (C4, 5) Cervical spondylotic myelopathy (C6, 7) Amyotrophic lateral sclerosis Amyotrophic lateral sclerosis Amyotrophic lateral sclerosis Multiple sclerosis Leukodystrophy Cerebrovascular disease (cortex) Cerebrovascular disease (pons) Cervical spondylotic myelopathy (C5, 6) Cervical spondylotic myelopathy (C3) Shy-Drager syndrome Shy-Drager syndrome Thalamic infarction Subcortical infarction Cerebellar haemorrhage Cerebellar haemorrhage Late onset cerebellar cortical atrophy Late onset cerebellar cortical atrophy Fisher's syndrome
Normal values (mean ±SD) Mean +2.5 SD
CTX-BST CT (ms) 4.01 3.31 3.41 2.1 1.8 12.71 4.11 4.61 NR NR NR NR NR NR 1.2 1.6 1.5 1.8 1.7 1.5 1.5 1.9 1.5
BST-CV CT (ms)
1.8±0.3 2.6
3.4±0.4 4.4
3.9 4.2 4.51 5.21 4.5t 4.61 4.3 4.1
NR NR NR NR NR NR 4.2 3.7 3.1 3.2 3.7 2.5 3.6 3.5 3.0
CMCT (ms) 7.91 7.51 7.91 7.31 6.3 17.31 8.41 8.71 16.31 9.41 13.21 8.41 8.71 5.4 5.3 4.6 5.0 5.4 4.0 5.1 5.4 4.5 5.3±0.6 6.8
CTX-BST CT = cortical-brainstem conduction time; BST-CV CT = brainstem-cervical conduction time; CMCT = central motor conduction time; NR = no response in brainstem stimulation; * = CMCT could not be obtained due to lack of cortical response; I = prolonged.
intensity area in the lower pontine region without any other abnormalities (Fig. 2). The threshold for brainstem stimulation (450 V) was increased. The cortical-brainstem conduction time of the first dorsal-interosseous was 3.3 ms, which was longer than the mean +2.5 SD of the normative data (Fig. 3). The lesion in the lower pons resulted in prolongation of the cortical-brainstem conduction time, suggesting that activation occurred at a point below the lower pons in brainstem stimulation. Patient 2 (Case 5 in Table 2). A 50-yr-old woman with cervical spondylosis was admitted to our hospital. Myelography disclosed compression on the spinal cord by osteophytes at the levels of the sixth and seventh cervical vertebrae. Figure 4 shows the responses of the first dorsal interosseous muscles. The brainstem-cervical conduction time (4.5 ms) was significantly delayed, even though central motor conduction time (6.3 ms) was within the normal range. Using brainstem stimulation, we were able to detect prolonged conduction which had not been revealed by the conventional technique using only central motor conduction time. Patient 3 (Case 14 in Table 2). In this patient with spastic paraparesis, compression of the spinal cord was demonstrated by MRI at the level of the third cervical vatebra (Fig. 5). The responses of biceps brachii, first dorsal interosseous and tibialis anterior
1W2
Y. UGAWA AND O T H E R S
FIG. 2. Axial T2-weighted MRI of the brainstem in Case 2. The image was obtained with a Siemens Magnetom scanner with a superconducting magnet of 1.5 Tesla field strength, spin echo T2-weighted (TR = 3000 ms, TE = 90 ms). There is an increased signal intensity in the lower pontine base. This indicates that the corticospinal tract is affected at the level of the lower pons.
muscles are presented in Fig. 6. In the arm muscles (biceps brachii and first dorsal interosseous), the central motor conduction times were prolonged. This indicated that there was a lesion within the central motor pathways at a level rostral to the spinal levels innervating the biceps brachii (C5, 6). However, brainstem stimulation produced no responses in these arm muscles. Stimulation at 700 V evoked no measurable responses, and only the cervical nerve roots were activated by stimulation at 750 V or more. The cervical nerve roots were activated by spreading currents from brainstem stimulation before the activation of some descending tracts which must have a lower threshold in normal subjects. In the arm muscles, therefore, our new technique demonstrated an
B R A I N S T E M S T I M U L A T I O N IN N E U R O L O G I C A L
DISORDERS
1953
Conical
Brainstem
1 mV
Cervical
10 ms FIG. 3. Responses of the first dorsal interosseous muscle following electrical stimulation in Case 2. The corticalbrainstem conduction time (3.3 ms) is abnormally prolonged, while the brainstem-cervical conduction time is slightly but not abnormally delayed.
abnormality in the central motor pathways because of the high threshold for brainstem stimulation, although it did not contribute to localization of the lesion. However, studies on leg muscles provided more accurate localization of the lesion in this patient. The brainstem-cervical conduction time of tibialis anterior was prolonged markedly (7.2 ms), while the cortical-brainstem conduction time was normal (1.6 ms). The lesion was thought to be located below the site of activation by brainstem stimulation. This case suggested that brainstem stimulation occurred at a point rostral to the level of the C3 vertebra. Patient 4 (Case 6 in Table 2). Figure 7 shows the responses in the first dorsal interosseous muscle in a patient with amyotrophic lateral sclerosis, later confirmed by autopsy. Cervical stimulation elicited a very small action potential with a normal latency. Much smaller responses were provoked by cortical or brainstem stimulation. In particular, cortical stimulation activated a very small number of motor units. The threshold for brainstem stimulation (600 V) was higher than in normal subjects. In recordings (Fig. 7) using an abnormally high intensity of stimulation (800 V), a small early component was elicited in addition to a brainstem response. This component was thought to be a potential evoked by activation of the cervical nerve roots by the spreading currents. The onset latency of the brainstem response was measured as 17 ms. The central motor conduction time was markedly prolonged (17.3 ms), as was cortical-brainstem conduction time (12.7 ms). Brainstem-cervical conduction time (4.6 ms) was also prolonged, but to a much lesser extent than cortical-brainstem conduction time. Patient 5 (Case 3 in Table 2). A 76-yr-old man with leprosy showed distal limb weakness and sensory disturbance. Neither deep tendon reflexes nor extensor plantar
1954
Y. UGAWA AND O T H E R S
Cortical
Brainstem
Cervical 0.5 mV/div
5 ms/div FIG. 4. Electromyographic responses from the first dorsal interosseous muscle in a patient with cervical spondylosis (Case 5). The brainstem-cervical conduction time abnormally delayed (4.5 ms), even though the central motor conduction time (6.3 ms) is not prolonged.
FIG. 5. Sagittal T,-weighted MRI image (TR = 500 ms. TE = 17 ms) of the cervical spinal cord of Case 14. demonstrating compression of the spinal cord by a C3, 4 herniated disc.
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1955
Biceps Cortical Cervical
Extensor carpi radial is Cortical
Cervical 1 mV
10 ms/div Tibialis anterior
Cortical
f*=^
• ••
•^*+~'
1
32.5
1
30.9
f
23.7 X/V
Brainstem
W
Cervical 1 mV
10 ms/div FIG. 6. Responses from the biceps brachii, extensor carpi radialis and tibialis anterior muscles following electrical stimulation. The central motor conduction times of two upper-limb muscles are abnormally prolonged. This suggests that there is a lesion within the central motor pathways at a level rostral to the C5, 6 spinal cord. Brainstem stimulation provokes no responses in the upper limb muscles. In contrast, the responses are evoked in the tibialis anterior. They show an abnormally prolonged brainstem-cervical conduction time and a normal cortical-brainstem conduction time.
responses were elicited due to marked muscular atrophy. The diagnosis was confirmed as leprous neuropathy by the presence of bacilli of Mycobacterium leprae in the biopsied sural nerve. Figure 8 shows the muscle action potentials in the first dorsal interosseous in this patient. The spinal latency (22 ms) was markedly prolonged compared with normal values
1956
Y. UGAWA AND OTHERS
Cortical —
Brainstem
Cervical
1 mV
20 ms FIG. 7. Responses of the first dorsal interosseous muscle following electrical stimulation in Case 6. Cortical and brainstem stimulation elicits very small EMG potentials. The cortical-brainstem conduction time (12.7 ms) is markedly prolonged, and the brainstem-cervical conduction time is also delayed.
Cortical
Brainstem
Cervical 1 mV
20 ms
FIG. 8. Two superimposed single EMG responses from the first dorsal interosseous muscle following motor cortical, brainstem and cervical electrical stimulation in Case 3. The response evoked by cervical stimulation is markedly delayed and small due to severe neuropathy. In addition, cortical-brainstem conduction time is considerably prolonged (3.4 ms) and brainstem-cervical conduction time is also prolonged (4.5 ms).
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1957
(mean ±SD: 13.6± 1.6 ms), compatible with severe neuropathy. Brainstem stimulation elicited a small response using an abnormally high intensity of stimulation (600 V). The cortical latency was variable {see Fig. 8). However, the shortest latency obtained by cortical stimulation was 29.9 ms. Therefore, even the shortest cortical-brainstem conduction time (3.4 ms) was significantly prolonged. The brainstem-cortical conduction time was also prolonged slightly (4.5 ms). Prolonged brainstem-cervical conduction time does not necessarily indicate the presence of lesions in the central motor pathways, since it includes the conduction time through the motor roots in the spinal canal (Aizawa et al., 1988). However, the prolongation of the cortical-brainstem conduction time suggested that there were lesions in the central motor pathways. Computerized tomography scans of the brain showed multiple subcortical low density areas, even though there were no clinical pyramidal signs. In this case, physiological techniques revealed some dysfunction of the central motor pathways, which was clinically masked by the severe muscular atrophy. Supratentorial lesions caused prolongation of cortical-brainstem conduction time and raised the threshold for brainstem stimulation. The results of conduction time studies in the first dorsal interosseous of 23 patients are summarized in Table 2. The cortical-brainstem conduction times were prolonged in three cases of cerebrovascular disease (Cases 1-3). Definite lesions were demonstrated by MRI in all of them. The lesions were found in the subcortical white matter in Cases 1 and 3. In Case 2, a small infarction was revealed in the lower pons by MRI. In these three cases of cerebrovascular disease with prolonged cortical-brainstem conduction time, the most caudal lesion was the lower pontine infarction. The threshold for brainstem stimulation was abnormally high even in patients with supratentorial cerebrovascular disease (Cases 1, 2). In all of the patients, the brainstem-cervical conduction time was also prolonged, although the cortical-brainstem conduction time was prolonged more prominently than the brainstem-cervical conduction time. The three cases of amyotrophic lateral sclerosis (Cases 6 — 8) showed prolongation of the cortical-brainstem conduction time. In all of them, the brainstem-cervical conduction time was also delayed but to a lesser extent than the cortical-brainstem conduction time. In two of the patients with cervical spondylotic myelopathy (Cases 4, 5), the brainstemcervical conduction times exceeded the upper limit of the normal range, whereas the cortical-brainstem conduction times were normal. In Case 5, the brainstem-cervical conduction time was prolonged, although the central motor conduction time was within the normal range. In the other two patients with cervical spondylotic myelopathy (Cases 13, 14) the cortical-brainstem conduction times were normal, but the brainstem-cervical conduction times of the lower limb muscles were abnormally prolonged, while brainstem stimulation provoked no responses in the upper limb muscles. The most rostral lesion was located at the level of the third cervical vertebra (Case 14) in the four patients who had abnormally prolonged brainstem-cervical conduction times and normal corticalbrainstem conduction times. The responses evoked by brainstem stimulation were not obtainable in the first dorsal interosseous muscles of six patients (Cases 9-14). The brainstem responses were masked by those produced by activation of the cervical nerve roots, or they were not elicited. The thresholds for activating the descending tracts were higher than those for activating the cervical roots due to the lesions in the descending tracts. The threshold for brainstem
1958
Y. UGAWA AND OTHERS
response was again abnormally increased in a patient with a supratentorial cerebrovascular disease (Case 11). Furthermore, as mentioned above, in two (Cases 13, 14) of these six patients, the responses evoked by brainstem stimulation were obtained in the lower limb muscles even though they were not recordable in the first dorsal interosseus muscle. In the other four patients (Cases 9 — 12), no responses were elicited even in the lower limb muscles. Some abnormal findings were revealed by the present method in all patients who had some clinical pyramidal signs (Cases 1 -14). In contrast, no abnormalities were found in patients with only autonomic failure (Cases 15, 16) and in those with ataxia (Cases 19—23). The conduction times were all normal even in patients with cerebrovascular disease who had neither paresis nor pyramidal signs (Cases 17, 18). This significant correlation between the abnormal findings revealed by this technique, such as prolonged conduction time or no obtainable response, and clinical pyramidal signs is consistent with the previous results (Ugawa et al., 1988a,b). DISCUSSION
The new technique described here provides an opportunity to estimate more accurately the localization of lesions within the central motor pathways. Only five normal volunteers were studied using this method previously (Ugawa et al., 1991). The normative data from the larger number of normal subjects provided here should be helpful when applying this technique in clinical practice. From a comparison of the cortical-brainstem conduction time with the results of the intraoperative recordings of the descending volleys, it was concluded that the site of activation upon brainstem stimulation was the pyramidal decussation (Ugawa et al., 1991). We have also estimated the site of activation to lie at the level of the cervicalmedullary junction, by using the regression lines between the latencies of the first dorsal interosseous muscles in normal individuals and anatomical data. In patients, the most rostral lesion causing prolongation of only the brainstem-cervical conduction time was positioned at the level of the third cervical vertebra, and the most caudal one causing prolonged cortical-brainstem conduction time was a cerebral infarction in the lower pons. The site of activation, therefore, was thought to lie below the medulla and rostral to the high cervical spinal cord. These results are all consistent. We conclude that brainstem stimulation occurs around the level of the cervical-medullary junction where the pyramidal decussation lies. The results of collision experiments using cortical and brainstem stimulation and the threshold differences among muscles indicated that the majority of the brainstem-evoked EMG responses were mediated by the corticospinal tract (Ugawa et al., 1991). Our estimated conduction velocity of the tract activated by this method was also comparable to that of the corticospinal tract. In patients, the abnormal findings revealed using this technique were significantly correlated with the clinical pyramidal signs. The thresholds for brainstem stimulation were abnormally high in the patients with supratentorial lesions. This suggests the likelihood that the tract activated by brainstem stimulation originates from some supratentorial area and terminates at a more caudal point than the cervicalmedullary junction. Therefore, the corticospinal tract is the most probable pathway other than the reticulospinal, vestibulospinal or long propriospinal tract. From all of the above-
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1959
mentioned results, it is concluded that the EMG potentials evoked by brainstem stimulation are mediated mainly by the corticospinal tract. The cortical-brainstem conduction times were delayed in the patients whose lesions were positioned at levels rostral to the cervical-medullary junction (Cases 1 —3) or in those whose motor cortical neurons (Cases 6 — 8) were primarily degenerated. In some of them, the brainstem-cervical conduction times were also delayed. Several mechanisms are thought to contribute to the prolongation of these conduction times (Thompson et al., 1987). These are (i) slowing of conduction in the largest diameter fibres of the corticospinal tract (e.g. demyelinating disease); (ii) reduction in size (and number) of excitatory postsynaptic potentials generated by cortical or brainstem stimulation (e.g. amyotrophic lateral sclerosis); (iii) reduction in the number of descending volleys in cortical stimulation due to cortical interneuronal damage (e.g. cerebrovascular disease). Activation of descending tracts other than the corticospinal tract might also contribute to prolongation of the conduction times. The observation that the cortical-brainstem conduction time was much more delayed than the brainstem-cervical conduction time can be explained by some of the above mechanisms (e.g. damage of the cortical interneurons), or by the fact that a brainstem stimulus is capable of recruiting a greater proportion of corticospinal fibres than a cortical stimulus. This observation may be in part due to the fact that the brainstem-cervical conduction time includes the conduction time through the motor roots which are not affected by lesions in the central motor pathways. In patients, a combination of different mechanisms could explain the delayed conduction times. Irrespective of the mechanism, the prolongation of the cortical-brainstem conduction time was thought to mean that the corticospinal tract was affected at a level rostral to the cervical-medullary junction. In these patients, the delayed brainstem-cervical conduction time did not necessarily indicate that a primary lesion was present below the cervical-medullary junction. On the other hand, in the four patients with cervical spondylotic myelopathy (Cases 4, 5, 13, 14), the brainstem-cervical conduction times were abnormally prolonged even though the cortical-brainstem conduction times were normal. Therefore, the finding that the brainstem-cervical conduction time is delayed and the cortical-brainstem conduction time is within the normal range might suggest that the primary lesion was positioned at a level below the cervical-medullary junction. From all of the above results, we conclude that the brainstem stimulation technique can provide more accurate localization of lesions within the central motor pathways than the conventional method. It can reveal whether the primary lesion is situated above or below the cervical-medullary junction. There are a few problems in applying this technique to patients with neurological disorders, mainly because of the need to use a very high intensity of stimulation. First, the responses directly from the cervical roots mask the brainstem responses in the upperlimb muscles. This phenomenon probably occurred in two (Cases 13, 14) of our 14 patients (Cases 1 — 14) who had some clinical pyramidal signs. In these two patients (c/Case 14), the brainstem responses could not be recorded in the upper limb muscles, whereas they were evoked in the lower limb muscles. We were then able to calculate the cortical-brainstem conduction time and the brainstem-cervical conduction time from the results obtained in the lower limb muscles. Therefore, study of the lower limb muscles is necessary in such patients. We sometimes had difficulty in judging the onset of the brainstem responses, since they were associated with small responses from the cervical
1960
Y. UGAWA AND OTHERS
roots, as demonstrated in Fig. 7. Accordingly, it is advisable to measure the onset latency from many responses in the same muscle evoked by stimulations at various intensities. The second problem is the variability of the site of excitation in brainstem stimulation. The site of activation was proved to be constant in normal individuals (Ugawa et al., 1991). However, it would not necessarily be constant during stimulation at an abnormally high voltage, and might be shifted more distally. In cortical stimulation at a very high voltage, excitation of the pyramidal tract axons would occur at some point within the internal capsule or even within the brainstem (Burke et al., 1990). A similar phenomenon may occur in brainstem stimulation. This shift might occur especially when the pyramidal tracts are pathologically hypo-excitable at the level of the pyramidal decussation; the more distal normo-excitable site would be activated. Because of these shifts, it is necessary to exercise care in interpreting the values of the conduction times. The third problem is that the brainstem responses in the patients might be mediated by tracts other than the corticospinal tract. Stimulation at very high voltage might activate such tracts whose thresholds are higher than that of the corticospinal tract. The responses mediated by such tracts would be seen if the corticospinal tract were affected. Thus, it must be borne in mind that conduction times do not always reflect the conduction in the corticospinal tract in pathological states. If we interpret the results on the basis of this knowledge, the few problems mentioned above would not affect the use of the technique. We conclude that the present technique would be very useful for neurophysiological evaluation of the descending motor pathways in patients with neurological disorders. REFERENCES AIZAWA H, UGAWA Y, GENBA K, SHIMPO T, MANNEN T (1988) Percutaneous electrical stimulation (PES)
and SEP in peripheral neuropathies. Clinical Neurology, Tokyo, 28, 447—452 (in Japanese). ALSTERMARK B, SASAKI S (1985) Integration in descending motor pathways controlling the forelimb in the cat. 13. Corticospinal effects in shoulder, elbow, wrist, and digit motoneurones. Experimental
Brain Research, 59, 353-364. BURKE D, HICKS RG, STEPHEN JPH (1990) Corticospinal volleys evoked by anodal and cathodal stimulation to the human motor cortex. Journal of Physiology, London, 425, 283—299. CLOUGH JMF, KERNELL D, PHILLIPS CG (1968) The distribution of monosynaptic excitation from the pyramidal tract and from primary spindle afferents to motoneurones of the baboon's hand and forearm. Journal of Physiology, London, 198, 145 — 166. DESMEDT JE, CHERON G (1983) Spinal and far-field components of human somatosensory evoked potentials to posterior tibial nerve stimulation analysed with oesophageal derivations and non-cephalic reference recording. Electroencephalography and Clinical Neurophysiology, 56, 635—651. DONALDSON HH, DAVIS DJ (1903) A description of charts showing the areas of the cross sections of the human spinal cord at the level of each spinal nerve. Journal of Comparative Neurology, 13, 19—40. IWATSUBO T, KUZUHARA S, KANEMITSU A, SHIMADA H, TOYOKURA Y (1990) Corticofugal projections
to the motor nuclei of the brainstem and spinal cord in humans. Neurology, Cleveland, 40, 309-312. MARSDEN CD, MERTON PA, MORTON HB (1982) Percutaneous stimulation of spinal cord and brain; pyramidal tract conduction velocities in man. Journal of Physiology, London, 328, 6P. MILLS KR, MURRAY NMF (1986) Electrical stimulation over the human vertebral column: which neural elements are excited? Electroencephalography and Clinical Neurophysiology, 63, 582—589. SHIMADA Y (1939) Brain and spinal cord in Japanese. In: Jinruigaku-senshi-koza, Volume 3. Tokyo: Yuhikaku, pp. 1 - 2 5 0 . THOMPSON PD, DAY BL, ROTHWELL JC, DICK JPR, COWAN JMA, ASSELMAN P et al. (1987) The inter-
pretation of electromyographic responses to electrical stimulation of the motor cortex in diseases of the upper motor neurone. Journal of the Neurological Sciences, 80, 91 — 110.
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1961
UGAWA Y, SHIMPO T, MANNEN T (1988a) Central motor conduction in cerebrovascular disease and motor neuron disease. Acta Neurologica Scandinavica, 78, 297—306. UGAWA Y, KOHARA N, SHIMPO T, MANNEN T (19886) Central motor and sensory conduction in
adrenoleukomyeloneuropathy, cerebrotendinous xanthomatosis, HTLV-1-associated myelopathy and tabes dorsal is. Journal of Neurology, Neurosurgery, and Psychiatry, 51, 1069-1074. UGAWA Y, GENBA K, SHIMPO T, MANNEN T (1989a) Physiologic analysis of central motor pathwayssimultaneous recording from multiple relaxed muscles. European Neurology, 29, 135-140. UGAWA Y, ROTHWELL JC, DAY BL, THOMPSON PD, MARSDEN CD (1989*) Magnetic stimulation over
the spinal enlargements. Journal of Neurology, Neurosurgery, and Psychiatry, 52, 1025-1032. UGAWA Y, ROTHWELL JC, DAY BL, THOMPSON PD, MARSDEN CD (1991) Percutaneous electrical stimula-
tion of corticospinal pathways at the level of the pyramidal decussation in humans. Annals of Neurology, 29, 418-427. (Received November 7, 1991. Revised February 2, 1992. Accepted May 19, 1992)
STIMULATION OF CORTICOSPINAL PATHWAYS AT THE LEVEL OF THE PYRAMIDAL DECUSSATION IN NEUROLOGICAL DISORDERS t
by YOSHIKAZU UGAWA, KIEKO GENBA, TORU MANNEN ICHIRO KANAZAWA
and
{From the Department of Neurology, Institute of Brain Research, School of Medicine, University of Tokyo) SUMMARY A newly developed technique of brainstem stimulation was applied in 14 normal subjects and 23 patients with various neurological disorders. The electromyographic (EMG) responses of limb muscles following cortical, brainstem and cervical stimulation were recorded. The cortical-brainstem conduction time and brainstem-cervical conduction time were then calculated from the difference in latency between the two sites of stimulation. From the regression lines of the relationship between the normal conduction times in the first dorsal interosseous muscle and the length of the descending tracts, the site of activation by brainstem stimulation was estimated to lie near the cervical-medullary junction. The most distal lesion causing prolongation of cortical-brainstem conduction time was a small cerebral infarction in the lower pons. Herniation of the third cervical spinal disc was the most rostral lesion resulting in delayed brainstem-cervical conduction time and normal cortical-brainstem conduction time. These observations suggest that activation occurs at the level of the cervical-medullary junction where the pyramidal decussation lies. The conduction velocities of the activated tracts estimated from the regression lines for normal individuals were 57—92 m/s. In patients with supratentorial lesions, the threshold for brainstem stimulation was abnormally high. The abnormal findings in this test were correlated significantly with the clinical pyramidal signs. This suggests that the EMG responses elicited by brainstem stimulation are mediated mainly by the corticospinal tract. We conclude that the brainstem stimulation technique would be clinically useful for localization of lesions in the corticospinal tract; the primary lesion can be localized whether above or below the pyramidal decussation. INTRODUCTION
The technique of percutaneous electrical stimulation has been widely used to evaluate central motor conduction in patients with various neurological diseases. Central motor conduction time, which is usually used in these studies, is calculated by subtracting the latency of the muscle response to spinal stimulation from the latency of the response produced by cortical stimulation (Marsden et al., 1982; Ugawa et al., 1989a). Abnormalities in the central motor pathways are expressed as prolongation of central motor conduction time or a lack of response evoked by cortical stimulation (Ugawa et al., 1988a,b). Such abnormal findings are seen in patients with lesions within the central motor pathways anywhere between the motor cortex and the spinal cord. Therefore, the usual method using central motor conduction time is not sufficient in Correspondence to: Dr Y. Ugawa, Department of Neurology, Institute of Brain Research, School of Medicine, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113, Japan. © Oxford University Press 1992
1948
Y. UGAWA AND OTHERS
clinical practice for localizing lesions in the central motor pathways. It has been reported recently that the descending motor tracts can be activated at the level of the pyramidal decussation in humans using high voltage electrical stimulation across the base of the skull (Ugawa et al., 1991). However, it is still unclear which pathways are activated and where the activation occurs in this new method. We have used normal volunteers in studies aimed at clarifying the site of activation and for establishing normal values. We have applied this new technique to localizing lesions within the central motor pathways in patients with various kinds of neurological disorders. In this paper, we describe the estimation of the site of activation using this method and its potential usefulness in clinical neurophysiology. SUBJECTS The subjects were 23 patients with various neurological disorders: nine with cerebrovascular disease, four with cervical spondylotic myelopathy, three with amyotrophic lateral sclerosis, one with clinically definite multiple sclerosis and one with leukodystrophy. Four patients with spinocerebellar degeneration and one with Fisher's syndrome, showing no pyramidal signs, were also included. In order to obtain normal values, 14 healthy volunteers (10 men and four women, aged 2 6 - 4 2 yrs and 154-188 cm in height) were studied. All of the subjects gave informed consent and the procedures were approved by the Ethical Committee of Tokyo University. No side-effects were noted in any individual. METHODS Surface electromyographic (EMG) activities were recorded simultaneously from biceps brachii, extensor carpi radialis, flexor carpi ulnaris, first dorsal interosseous and tibialis anterior muscles in normal volunteers, using surface cup electrodes. The first dorsal interosseous muscle was studied in all patients, and also some of the other muscles listed above were selected for investigation according to the clinical features of the patients. Electromyographic signals were amplified with filters set at 80 Hz and 3 kHz, recorded and analysed by a Signal Processor 7T18 (NEC San-Ei) or a Neuropack 4 (Nihon Kohden). We performed electrical stimulation of three sites within the central motor pathways (motor cortex, brainstem and cervical spinal cord or roots) using a high voltage electrical stimulator described previously (Ugawa et al., 1989a). This stimulator is capable of delivering stimuli of up to 1000 V with a decay time constant of 50 jis. Stimulus intensities are expressed as the voltage charged by the stimulator. For cortical activation of arm muscles, the cathode was placed at the vertex and the anode 7 cm lateral. These electrode polarities were reversed in order to activate leg muscles. For cervical stimulation, the anode was fixed over the fifth cervical spinous process and the cathode over the first thoracic spinous process. The descending motor tracts were activated at the level of the brainstem (brainstem stimulation) using a technique described recently (Ugawa et al., 1991). The stimulating electrodes were fixed with collodion on both sides of the head at the angle formed by the posterior edge of the mastoid process and the anterior part of the occipital bone (described as site B in Ugawa et al., 1991). The anode was on the right side and the cathode on the left. For every stimulation, the leg muscles were preactivated by slight voluntary contraction. For cortical and brainstem stimulation, the arm muscles were studied in an active state. However, they were relaxed during cervical stimulation. Stimulus intensity was adjusted to produce a response of maximum amplitude and shortest latency. When a maximum response could not be elicited, we used the maximum output of our stimulator. In all experiments, stimulation on each site was repeated at least twice to assess the reproducibility of the findings. The following three conduction times were calculated for each muscle. The cortical-brainstem conduction time was obtained by subtracting the latency of the response produced by brainstem stimulation (brainstem latency) from that of the response evoked by cortical stimulation (cortical latency). The brainstem-cervical conduction time was calculated by subtracting the latency of the response elicited by cervical stimulation (cervical latency) from the brainstem latency. The conventional central motor conduction time of arm muscles was also obtained by subtracting the cervical latency from the cortical latency. The results obtained from
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1949
the normal volunteers were used as normative data for these conduction times. The above conduction times were considered to be prolonged if they exceeded the upper limit of the normal range (mean +2.5 SD). We studied correlations between the physiological findings in the patients and their clinical signs. RESULTS
Normal subjects Consistent motor responses could be obtained from all muscles in all normal subjects. The thresholds for activation of any muscle were 200 — 350 V. The average latencies and conduction times for every muscle are presented in Table 1. These values all correspond well with those in the previous report (Ugawa et al., 1991). The site of activation in the brainstem stimulation was estimated from our data for latencies in normal subjects. The majority of the brainstem-evoked EMG responses were thought to be mediated by the corticospinal tract (Ugawa et al., 1991). In the first dorsal interosseous muscle especially, almost all components of the potentials are mediated by the corticospinal tract because this tract projects strongly to the intrinsic hand muscles (Clough et al., 1968; Alstermark and Sasaki, 1985; Iwatsubo et al., 1990). Therefore, we used the data obtained from first dorsal interosseous muscles for estimating the level in the corticospinal tract where activation occurred. The actual time in which the nerve impulses are conducted from the motor cortex to the activation site of the brainstem stimulation is the measured cortical-brainstem conduction time. On the other hand, brainstem-cervical conduction time includes synaptic delay at the alpha motor neuron and conduction time along the motor roots in the spinal canal, as well as the conduction time from the brainstem to the spinal cord. The conduction time in the motor roots is reported to be 0.6—1.4 ms (Mills and Murray, 1986; Ugawa et al., 1989a,b). Our mean value was about 1 ms (Ugawa et al., \9%9a,b). If we assume that the synaptic delay is 1 ms, then the actual conduction time from brainstem to spinal cord is calculated as 'measured brainstem-cervical conduction time — 1 ms (conduction time along motor roots) - 1 ms (synaptic delay)', i.e. brainstem-cervical conduction time — 2 ms. The mean distance from the cervical-medullary junction to the first thoracic spinal cord (innervation level for first dorsal interosseous) is 10.9 cm (Donaldson and Davis, 1903). The mean length of the corticospinal tract from the cortex to the cervical-medullary junction is estimated to be 14.2 cm from anatomical data (Shimada, 1939). This approximation of the distance from the motor cortex to the first thoracic spinal level
TABLE I. NORMAL LATENCIES AND CONDUCTION TIMES OF RESPONSES TO BRAINSTEM STIMULATION IN VARIOUS MUSCLES
Biceps Forearm flexor Forearm extensor First dorsal interosseous Tibialis anterior
CTX (ms)
BST (ms)
CV (ms)
CTX-BST CT (ms)
10.2±1.0 13.6± 1.2 13.7±!.O 19.2± 1.9 25.5 ±1.5
8.5±0.4 ll.7±l.l 1I.9±1.O 17.3±1.5 23.9± 1.7
5.3±0.3 8.4±1.0 8.2±1.0 13.6±1.6 21.3±1.5
1.8±0.3 1.9±0.4 1.7±0.3 1.8±0.3 1.7±0.3
BST-CV CT (ms) 3.I±O.2 3.3±0.9 3.5±0.5 3.7±0.5 2.3±0.5
Values are mean ±SD. The values are taken from 16 subjects. CTX = cortical latency; BST = brainstem latency; CV = Cervical latency; CTX-BST CT = cortical-brainstem conduction time; BST-CV CT = brainstem-cervical conduction time.
1950
Y. UGAWA AND OTHERS
(25.1 cm = 10.9 cm+14.2 cm) is consistent with another estimation of the distance from the parietal cortex to the first thoracic level of 24 cm (Desmedt and Cheron, 1983). Assuming that conduction velocity in the corticospinal tract is the same at all levels from the cortex to the spinal cord, we plotted the distance from the cervical-medullary junction as a function of the conduction times in 14 first dorsal interosseous muscles. (Fig. 1.). The site where activation occurred during brainstem stimulation was estimated to be between 2.2 cm rostral and 1.5 cm caudal from the cervical-medullary junction. This was compatible with the previous conclusion that the corticospinal tract was activated at the level of the pyramidal decussation (Ugawa et al., 1991). The conduction velocity in the stimulated tract was also estimated to be 5 7 - 9 2 m/s from the slopes of the regression lines. This calculation of the conduction velocity gave only an approximate value, since it was based on average lengths in post-mortem specimens. Despite this, these values fitted well with other estimates of conduction in the descending motor tract in humans (Marsden et al., 1982; Ugawa et al., 1989a). Patients We first present five typical cases. Patient 1 (Case 2 in Table 2). A 62-yr-old woman developed flaccid tetraparesis. T2-weighted magnetic resonance images (MRIs) of the brain demonstrated a high
Cortex 150-, 100,. 50-
^^^*^*^
Estimated * ofactivati
0-50100150-3
VSRH^V •
i
-2
-1
Conduction time from cortex (ms)
1
TI
2
Conduction time to spinal level (ms)
FIG. 1. The relationship between conduction time in the first dorsal interosseous muscle of 14 different subjects and the length of the descending tract. The average length from the motor cortex to the cervical-medullary junction was 142 mm (Shimada, 1939) and that from the cervical-medullary junction to the first thoracic spinal cord (Tl) innervating the first dorsal interosseous muscle is 109 mm (Donaldson and Davis, 1903). The y-axis represents the length from the cervical-medullary junction. The x-axis represents the conduction time. The conduction time from the motor cortex to the cervical-medullary junction is plotted on the minus side. That from the cervical-medullary junction to the innervating spinal level, calculated as 'brainstem-cervical conduction time —2', is plotted on the plus side. The points where the regression lines cross the y-axis are the sites where activation occurs in brainstem stimulation. These points are from — 15 mm to +22 mm, which means from 22 mm rostral to 15 mm caudal to the cervical-medullary junction. The slopes of the lines are - 9 2 to - 5 7 m/s.
B R A I N S T E M S T I M U L A T I O N IN N E U R O L O G I C A L D I S O R D E R S
1951
TABLE 2. THE PHYSIOLOGICAL RESULTS IN THE FIRST DORSAL INTEROSSEOUS MUSCLE IN 23 PATIENTS Case
Diagnosis (level of lesion)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Cerebrovascular disease (subcortex) Cerebrovascular disease (pons) Cerebrovascular disease (subcortex) + leprosy Cervical spondylotic myelopathy (C4, 5) Cervical spondylotic myelopathy (C6, 7) Amyotrophic lateral sclerosis Amyotrophic lateral sclerosis Amyotrophic lateral sclerosis Multiple sclerosis Leukodystrophy Cerebrovascular disease (cortex) Cerebrovascular disease (pons) Cervical spondylotic myelopathy (C5, 6) Cervical spondylotic myelopathy (C3) Shy-Drager syndrome Shy-Drager syndrome Thalamic infarction Subcortical infarction Cerebellar haemorrhage Cerebellar haemorrhage Late onset cerebellar cortical atrophy Late onset cerebellar cortical atrophy Fisher's syndrome
Normal values (mean ±SD) Mean +2.5 SD
CTX-BST CT (ms) 4.01 3.31 3.41 2.1 1.8 12.71 4.11 4.61 NR NR NR NR NR NR 1.2 1.6 1.5 1.8 1.7 1.5 1.5 1.9 1.5
BST-CV CT (ms)
1.8±0.3 2.6
3.4±0.4 4.4
3.9 4.2 4.51 5.21 4.5t 4.61 4.3 4.1
NR NR NR NR NR NR 4.2 3.7 3.1 3.2 3.7 2.5 3.6 3.5 3.0
CMCT (ms) 7.91 7.51 7.91 7.31 6.3 17.31 8.41 8.71 16.31 9.41 13.21 8.41 8.71 5.4 5.3 4.6 5.0 5.4 4.0 5.1 5.4 4.5 5.3±0.6 6.8
CTX-BST CT = cortical-brainstem conduction time; BST-CV CT = brainstem-cervical conduction time; CMCT = central motor conduction time; NR = no response in brainstem stimulation; * = CMCT could not be obtained due to lack of cortical response; I = prolonged.
intensity area in the lower pontine region without any other abnormalities (Fig. 2). The threshold for brainstem stimulation (450 V) was increased. The cortical-brainstem conduction time of the first dorsal-interosseous was 3.3 ms, which was longer than the mean +2.5 SD of the normative data (Fig. 3). The lesion in the lower pons resulted in prolongation of the cortical-brainstem conduction time, suggesting that activation occurred at a point below the lower pons in brainstem stimulation. Patient 2 (Case 5 in Table 2). A 50-yr-old woman with cervical spondylosis was admitted to our hospital. Myelography disclosed compression on the spinal cord by osteophytes at the levels of the sixth and seventh cervical vertebrae. Figure 4 shows the responses of the first dorsal interosseous muscles. The brainstem-cervical conduction time (4.5 ms) was significantly delayed, even though central motor conduction time (6.3 ms) was within the normal range. Using brainstem stimulation, we were able to detect prolonged conduction which had not been revealed by the conventional technique using only central motor conduction time. Patient 3 (Case 14 in Table 2). In this patient with spastic paraparesis, compression of the spinal cord was demonstrated by MRI at the level of the third cervical vatebra (Fig. 5). The responses of biceps brachii, first dorsal interosseous and tibialis anterior
1W2
Y. UGAWA AND O T H E R S
FIG. 2. Axial T2-weighted MRI of the brainstem in Case 2. The image was obtained with a Siemens Magnetom scanner with a superconducting magnet of 1.5 Tesla field strength, spin echo T2-weighted (TR = 3000 ms, TE = 90 ms). There is an increased signal intensity in the lower pontine base. This indicates that the corticospinal tract is affected at the level of the lower pons.
muscles are presented in Fig. 6. In the arm muscles (biceps brachii and first dorsal interosseous), the central motor conduction times were prolonged. This indicated that there was a lesion within the central motor pathways at a level rostral to the spinal levels innervating the biceps brachii (C5, 6). However, brainstem stimulation produced no responses in these arm muscles. Stimulation at 700 V evoked no measurable responses, and only the cervical nerve roots were activated by stimulation at 750 V or more. The cervical nerve roots were activated by spreading currents from brainstem stimulation before the activation of some descending tracts which must have a lower threshold in normal subjects. In the arm muscles, therefore, our new technique demonstrated an
B R A I N S T E M S T I M U L A T I O N IN N E U R O L O G I C A L
DISORDERS
1953
Conical
Brainstem
1 mV
Cervical
10 ms FIG. 3. Responses of the first dorsal interosseous muscle following electrical stimulation in Case 2. The corticalbrainstem conduction time (3.3 ms) is abnormally prolonged, while the brainstem-cervical conduction time is slightly but not abnormally delayed.
abnormality in the central motor pathways because of the high threshold for brainstem stimulation, although it did not contribute to localization of the lesion. However, studies on leg muscles provided more accurate localization of the lesion in this patient. The brainstem-cervical conduction time of tibialis anterior was prolonged markedly (7.2 ms), while the cortical-brainstem conduction time was normal (1.6 ms). The lesion was thought to be located below the site of activation by brainstem stimulation. This case suggested that brainstem stimulation occurred at a point rostral to the level of the C3 vertebra. Patient 4 (Case 6 in Table 2). Figure 7 shows the responses in the first dorsal interosseous muscle in a patient with amyotrophic lateral sclerosis, later confirmed by autopsy. Cervical stimulation elicited a very small action potential with a normal latency. Much smaller responses were provoked by cortical or brainstem stimulation. In particular, cortical stimulation activated a very small number of motor units. The threshold for brainstem stimulation (600 V) was higher than in normal subjects. In recordings (Fig. 7) using an abnormally high intensity of stimulation (800 V), a small early component was elicited in addition to a brainstem response. This component was thought to be a potential evoked by activation of the cervical nerve roots by the spreading currents. The onset latency of the brainstem response was measured as 17 ms. The central motor conduction time was markedly prolonged (17.3 ms), as was cortical-brainstem conduction time (12.7 ms). Brainstem-cervical conduction time (4.6 ms) was also prolonged, but to a much lesser extent than cortical-brainstem conduction time. Patient 5 (Case 3 in Table 2). A 76-yr-old man with leprosy showed distal limb weakness and sensory disturbance. Neither deep tendon reflexes nor extensor plantar
1954
Y. UGAWA AND O T H E R S
Cortical
Brainstem
Cervical 0.5 mV/div
5 ms/div FIG. 4. Electromyographic responses from the first dorsal interosseous muscle in a patient with cervical spondylosis (Case 5). The brainstem-cervical conduction time abnormally delayed (4.5 ms), even though the central motor conduction time (6.3 ms) is not prolonged.
FIG. 5. Sagittal T,-weighted MRI image (TR = 500 ms. TE = 17 ms) of the cervical spinal cord of Case 14. demonstrating compression of the spinal cord by a C3, 4 herniated disc.
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1955
Biceps Cortical Cervical
Extensor carpi radial is Cortical
Cervical 1 mV
10 ms/div Tibialis anterior
Cortical
f*=^
• ••
•^*+~'
1
32.5
1
30.9
f
23.7 X/V
Brainstem
W
Cervical 1 mV
10 ms/div FIG. 6. Responses from the biceps brachii, extensor carpi radialis and tibialis anterior muscles following electrical stimulation. The central motor conduction times of two upper-limb muscles are abnormally prolonged. This suggests that there is a lesion within the central motor pathways at a level rostral to the C5, 6 spinal cord. Brainstem stimulation provokes no responses in the upper limb muscles. In contrast, the responses are evoked in the tibialis anterior. They show an abnormally prolonged brainstem-cervical conduction time and a normal cortical-brainstem conduction time.
responses were elicited due to marked muscular atrophy. The diagnosis was confirmed as leprous neuropathy by the presence of bacilli of Mycobacterium leprae in the biopsied sural nerve. Figure 8 shows the muscle action potentials in the first dorsal interosseous in this patient. The spinal latency (22 ms) was markedly prolonged compared with normal values
1956
Y. UGAWA AND OTHERS
Cortical —
Brainstem
Cervical
1 mV
20 ms FIG. 7. Responses of the first dorsal interosseous muscle following electrical stimulation in Case 6. Cortical and brainstem stimulation elicits very small EMG potentials. The cortical-brainstem conduction time (12.7 ms) is markedly prolonged, and the brainstem-cervical conduction time is also delayed.
Cortical
Brainstem
Cervical 1 mV
20 ms
FIG. 8. Two superimposed single EMG responses from the first dorsal interosseous muscle following motor cortical, brainstem and cervical electrical stimulation in Case 3. The response evoked by cervical stimulation is markedly delayed and small due to severe neuropathy. In addition, cortical-brainstem conduction time is considerably prolonged (3.4 ms) and brainstem-cervical conduction time is also prolonged (4.5 ms).
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1957
(mean ±SD: 13.6± 1.6 ms), compatible with severe neuropathy. Brainstem stimulation elicited a small response using an abnormally high intensity of stimulation (600 V). The cortical latency was variable {see Fig. 8). However, the shortest latency obtained by cortical stimulation was 29.9 ms. Therefore, even the shortest cortical-brainstem conduction time (3.4 ms) was significantly prolonged. The brainstem-cortical conduction time was also prolonged slightly (4.5 ms). Prolonged brainstem-cervical conduction time does not necessarily indicate the presence of lesions in the central motor pathways, since it includes the conduction time through the motor roots in the spinal canal (Aizawa et al., 1988). However, the prolongation of the cortical-brainstem conduction time suggested that there were lesions in the central motor pathways. Computerized tomography scans of the brain showed multiple subcortical low density areas, even though there were no clinical pyramidal signs. In this case, physiological techniques revealed some dysfunction of the central motor pathways, which was clinically masked by the severe muscular atrophy. Supratentorial lesions caused prolongation of cortical-brainstem conduction time and raised the threshold for brainstem stimulation. The results of conduction time studies in the first dorsal interosseous of 23 patients are summarized in Table 2. The cortical-brainstem conduction times were prolonged in three cases of cerebrovascular disease (Cases 1-3). Definite lesions were demonstrated by MRI in all of them. The lesions were found in the subcortical white matter in Cases 1 and 3. In Case 2, a small infarction was revealed in the lower pons by MRI. In these three cases of cerebrovascular disease with prolonged cortical-brainstem conduction time, the most caudal lesion was the lower pontine infarction. The threshold for brainstem stimulation was abnormally high even in patients with supratentorial cerebrovascular disease (Cases 1, 2). In all of the patients, the brainstem-cervical conduction time was also prolonged, although the cortical-brainstem conduction time was prolonged more prominently than the brainstem-cervical conduction time. The three cases of amyotrophic lateral sclerosis (Cases 6 — 8) showed prolongation of the cortical-brainstem conduction time. In all of them, the brainstem-cervical conduction time was also delayed but to a lesser extent than the cortical-brainstem conduction time. In two of the patients with cervical spondylotic myelopathy (Cases 4, 5), the brainstemcervical conduction times exceeded the upper limit of the normal range, whereas the cortical-brainstem conduction times were normal. In Case 5, the brainstem-cervical conduction time was prolonged, although the central motor conduction time was within the normal range. In the other two patients with cervical spondylotic myelopathy (Cases 13, 14) the cortical-brainstem conduction times were normal, but the brainstem-cervical conduction times of the lower limb muscles were abnormally prolonged, while brainstem stimulation provoked no responses in the upper limb muscles. The most rostral lesion was located at the level of the third cervical vertebra (Case 14) in the four patients who had abnormally prolonged brainstem-cervical conduction times and normal corticalbrainstem conduction times. The responses evoked by brainstem stimulation were not obtainable in the first dorsal interosseous muscles of six patients (Cases 9-14). The brainstem responses were masked by those produced by activation of the cervical nerve roots, or they were not elicited. The thresholds for activating the descending tracts were higher than those for activating the cervical roots due to the lesions in the descending tracts. The threshold for brainstem
1958
Y. UGAWA AND OTHERS
response was again abnormally increased in a patient with a supratentorial cerebrovascular disease (Case 11). Furthermore, as mentioned above, in two (Cases 13, 14) of these six patients, the responses evoked by brainstem stimulation were obtained in the lower limb muscles even though they were not recordable in the first dorsal interosseus muscle. In the other four patients (Cases 9 — 12), no responses were elicited even in the lower limb muscles. Some abnormal findings were revealed by the present method in all patients who had some clinical pyramidal signs (Cases 1 -14). In contrast, no abnormalities were found in patients with only autonomic failure (Cases 15, 16) and in those with ataxia (Cases 19—23). The conduction times were all normal even in patients with cerebrovascular disease who had neither paresis nor pyramidal signs (Cases 17, 18). This significant correlation between the abnormal findings revealed by this technique, such as prolonged conduction time or no obtainable response, and clinical pyramidal signs is consistent with the previous results (Ugawa et al., 1988a,b). DISCUSSION
The new technique described here provides an opportunity to estimate more accurately the localization of lesions within the central motor pathways. Only five normal volunteers were studied using this method previously (Ugawa et al., 1991). The normative data from the larger number of normal subjects provided here should be helpful when applying this technique in clinical practice. From a comparison of the cortical-brainstem conduction time with the results of the intraoperative recordings of the descending volleys, it was concluded that the site of activation upon brainstem stimulation was the pyramidal decussation (Ugawa et al., 1991). We have also estimated the site of activation to lie at the level of the cervicalmedullary junction, by using the regression lines between the latencies of the first dorsal interosseous muscles in normal individuals and anatomical data. In patients, the most rostral lesion causing prolongation of only the brainstem-cervical conduction time was positioned at the level of the third cervical vertebra, and the most caudal one causing prolonged cortical-brainstem conduction time was a cerebral infarction in the lower pons. The site of activation, therefore, was thought to lie below the medulla and rostral to the high cervical spinal cord. These results are all consistent. We conclude that brainstem stimulation occurs around the level of the cervical-medullary junction where the pyramidal decussation lies. The results of collision experiments using cortical and brainstem stimulation and the threshold differences among muscles indicated that the majority of the brainstem-evoked EMG responses were mediated by the corticospinal tract (Ugawa et al., 1991). Our estimated conduction velocity of the tract activated by this method was also comparable to that of the corticospinal tract. In patients, the abnormal findings revealed using this technique were significantly correlated with the clinical pyramidal signs. The thresholds for brainstem stimulation were abnormally high in the patients with supratentorial lesions. This suggests the likelihood that the tract activated by brainstem stimulation originates from some supratentorial area and terminates at a more caudal point than the cervicalmedullary junction. Therefore, the corticospinal tract is the most probable pathway other than the reticulospinal, vestibulospinal or long propriospinal tract. From all of the above-
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1959
mentioned results, it is concluded that the EMG potentials evoked by brainstem stimulation are mediated mainly by the corticospinal tract. The cortical-brainstem conduction times were delayed in the patients whose lesions were positioned at levels rostral to the cervical-medullary junction (Cases 1 —3) or in those whose motor cortical neurons (Cases 6 — 8) were primarily degenerated. In some of them, the brainstem-cervical conduction times were also delayed. Several mechanisms are thought to contribute to the prolongation of these conduction times (Thompson et al., 1987). These are (i) slowing of conduction in the largest diameter fibres of the corticospinal tract (e.g. demyelinating disease); (ii) reduction in size (and number) of excitatory postsynaptic potentials generated by cortical or brainstem stimulation (e.g. amyotrophic lateral sclerosis); (iii) reduction in the number of descending volleys in cortical stimulation due to cortical interneuronal damage (e.g. cerebrovascular disease). Activation of descending tracts other than the corticospinal tract might also contribute to prolongation of the conduction times. The observation that the cortical-brainstem conduction time was much more delayed than the brainstem-cervical conduction time can be explained by some of the above mechanisms (e.g. damage of the cortical interneurons), or by the fact that a brainstem stimulus is capable of recruiting a greater proportion of corticospinal fibres than a cortical stimulus. This observation may be in part due to the fact that the brainstem-cervical conduction time includes the conduction time through the motor roots which are not affected by lesions in the central motor pathways. In patients, a combination of different mechanisms could explain the delayed conduction times. Irrespective of the mechanism, the prolongation of the cortical-brainstem conduction time was thought to mean that the corticospinal tract was affected at a level rostral to the cervical-medullary junction. In these patients, the delayed brainstem-cervical conduction time did not necessarily indicate that a primary lesion was present below the cervical-medullary junction. On the other hand, in the four patients with cervical spondylotic myelopathy (Cases 4, 5, 13, 14), the brainstem-cervical conduction times were abnormally prolonged even though the cortical-brainstem conduction times were normal. Therefore, the finding that the brainstem-cervical conduction time is delayed and the cortical-brainstem conduction time is within the normal range might suggest that the primary lesion was positioned at a level below the cervical-medullary junction. From all of the above results, we conclude that the brainstem stimulation technique can provide more accurate localization of lesions within the central motor pathways than the conventional method. It can reveal whether the primary lesion is situated above or below the cervical-medullary junction. There are a few problems in applying this technique to patients with neurological disorders, mainly because of the need to use a very high intensity of stimulation. First, the responses directly from the cervical roots mask the brainstem responses in the upperlimb muscles. This phenomenon probably occurred in two (Cases 13, 14) of our 14 patients (Cases 1 — 14) who had some clinical pyramidal signs. In these two patients (c/Case 14), the brainstem responses could not be recorded in the upper limb muscles, whereas they were evoked in the lower limb muscles. We were then able to calculate the cortical-brainstem conduction time and the brainstem-cervical conduction time from the results obtained in the lower limb muscles. Therefore, study of the lower limb muscles is necessary in such patients. We sometimes had difficulty in judging the onset of the brainstem responses, since they were associated with small responses from the cervical
1960
Y. UGAWA AND OTHERS
roots, as demonstrated in Fig. 7. Accordingly, it is advisable to measure the onset latency from many responses in the same muscle evoked by stimulations at various intensities. The second problem is the variability of the site of excitation in brainstem stimulation. The site of activation was proved to be constant in normal individuals (Ugawa et al., 1991). However, it would not necessarily be constant during stimulation at an abnormally high voltage, and might be shifted more distally. In cortical stimulation at a very high voltage, excitation of the pyramidal tract axons would occur at some point within the internal capsule or even within the brainstem (Burke et al., 1990). A similar phenomenon may occur in brainstem stimulation. This shift might occur especially when the pyramidal tracts are pathologically hypo-excitable at the level of the pyramidal decussation; the more distal normo-excitable site would be activated. Because of these shifts, it is necessary to exercise care in interpreting the values of the conduction times. The third problem is that the brainstem responses in the patients might be mediated by tracts other than the corticospinal tract. Stimulation at very high voltage might activate such tracts whose thresholds are higher than that of the corticospinal tract. The responses mediated by such tracts would be seen if the corticospinal tract were affected. Thus, it must be borne in mind that conduction times do not always reflect the conduction in the corticospinal tract in pathological states. If we interpret the results on the basis of this knowledge, the few problems mentioned above would not affect the use of the technique. We conclude that the present technique would be very useful for neurophysiological evaluation of the descending motor pathways in patients with neurological disorders. REFERENCES AIZAWA H, UGAWA Y, GENBA K, SHIMPO T, MANNEN T (1988) Percutaneous electrical stimulation (PES)
and SEP in peripheral neuropathies. Clinical Neurology, Tokyo, 28, 447—452 (in Japanese). ALSTERMARK B, SASAKI S (1985) Integration in descending motor pathways controlling the forelimb in the cat. 13. Corticospinal effects in shoulder, elbow, wrist, and digit motoneurones. Experimental
Brain Research, 59, 353-364. BURKE D, HICKS RG, STEPHEN JPH (1990) Corticospinal volleys evoked by anodal and cathodal stimulation to the human motor cortex. Journal of Physiology, London, 425, 283—299. CLOUGH JMF, KERNELL D, PHILLIPS CG (1968) The distribution of monosynaptic excitation from the pyramidal tract and from primary spindle afferents to motoneurones of the baboon's hand and forearm. Journal of Physiology, London, 198, 145 — 166. DESMEDT JE, CHERON G (1983) Spinal and far-field components of human somatosensory evoked potentials to posterior tibial nerve stimulation analysed with oesophageal derivations and non-cephalic reference recording. Electroencephalography and Clinical Neurophysiology, 56, 635—651. DONALDSON HH, DAVIS DJ (1903) A description of charts showing the areas of the cross sections of the human spinal cord at the level of each spinal nerve. Journal of Comparative Neurology, 13, 19—40. IWATSUBO T, KUZUHARA S, KANEMITSU A, SHIMADA H, TOYOKURA Y (1990) Corticofugal projections
to the motor nuclei of the brainstem and spinal cord in humans. Neurology, Cleveland, 40, 309-312. MARSDEN CD, MERTON PA, MORTON HB (1982) Percutaneous stimulation of spinal cord and brain; pyramidal tract conduction velocities in man. Journal of Physiology, London, 328, 6P. MILLS KR, MURRAY NMF (1986) Electrical stimulation over the human vertebral column: which neural elements are excited? Electroencephalography and Clinical Neurophysiology, 63, 582—589. SHIMADA Y (1939) Brain and spinal cord in Japanese. In: Jinruigaku-senshi-koza, Volume 3. Tokyo: Yuhikaku, pp. 1 - 2 5 0 . THOMPSON PD, DAY BL, ROTHWELL JC, DICK JPR, COWAN JMA, ASSELMAN P et al. (1987) The inter-
pretation of electromyographic responses to electrical stimulation of the motor cortex in diseases of the upper motor neurone. Journal of the Neurological Sciences, 80, 91 — 110.
BRAINSTEM STIMULATION IN NEUROLOGICAL DISORDERS
1961
UGAWA Y, SHIMPO T, MANNEN T (1988a) Central motor conduction in cerebrovascular disease and motor neuron disease. Acta Neurologica Scandinavica, 78, 297—306. UGAWA Y, KOHARA N, SHIMPO T, MANNEN T (19886) Central motor and sensory conduction in
adrenoleukomyeloneuropathy, cerebrotendinous xanthomatosis, HTLV-1-associated myelopathy and tabes dorsal is. Journal of Neurology, Neurosurgery, and Psychiatry, 51, 1069-1074. UGAWA Y, GENBA K, SHIMPO T, MANNEN T (1989a) Physiologic analysis of central motor pathwayssimultaneous recording from multiple relaxed muscles. European Neurology, 29, 135-140. UGAWA Y, ROTHWELL JC, DAY BL, THOMPSON PD, MARSDEN CD (1989*) Magnetic stimulation over
the spinal enlargements. Journal of Neurology, Neurosurgery, and Psychiatry, 52, 1025-1032. UGAWA Y, ROTHWELL JC, DAY BL, THOMPSON PD, MARSDEN CD (1991) Percutaneous electrical stimula-
tion of corticospinal pathways at the level of the pyramidal decussation in humans. Annals of Neurology, 29, 418-427. (Received November 7, 1991. Revised February 2, 1992. Accepted May 19, 1992)
