- Email: [email protected]
Cortical reflex myoclonus studied with cortical electrodes
Clinical Neurophysiology 110 (1999) 1521±1530
Cortical re¯ex myoclonus studied with cortical electrodes P. Ashby*, R. Chen, R. Wennberg, A.M. Lozano, A.E. Lang Playfair Neuroscience Unit, Toronto Hospital, Western Division, University of Toronto, 399 Bathurst Street, EC8-005 Toronto, Ontario M5T 2S8, Canada Accepted 12 April 1999
Abstract Objective: To study the mechanism of cortical re¯ex myoclonus. Methods: A patient with stimulus sensitive myoclonus of the left foot had an array of subdural electrodes placed over the right sensorimotor cortex. Results: Stimulation through one of the electrodes (contact 13) facilitated leg muscles with the shortest latency and was presumed to lie over the motor cortex. Tibial nerve stimulation evoked a potential with the shortest latency 1 cm further posteriorly (contacts 11±12). These contacts were presumed to lie over the sensory cortex. The potential at 11±12 was followed by a much larger potential that reversed polarity at contact 13. Back averaging from spontaneous myoclonic jerks showed a cortical premovement potential which reversed polarity at contact 13. The threshold for the motor evoked potential in leg muscles evoked by transcranial magnetic stimulation was lower on the affected side. Electrical stimulation through contact 13 produced cortical potentials that could be recorded at adjacent contacts. The combination of a positive potential followed by a negative potential recurred at ~35±40 ms intervals, each positive potential generating a myoclonic jerk. Additional waves resembling I waves accompanied only the ®rst positive potential. Surgical removal of the cortex under electrode 13 abolished the myoclonus. Conclusions: The myoclonic jerks arose in the motor cortex. We postulate that there is increased excitability or synchronization of the cortical neurons at that site. The spontaneous, peripherally induced and recurrent cortical potentials and myoclonic jerks can occur without participation of the circuitry of the presumed I waves. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Cortical myoclonus; Cortical stimulation; Cortical inhibition; Motor cortex; Transcranial magnetic stimulation
1. Introduction Cortical re¯ex myoclonus (Hallett et al., 1979; Obeso et al., 1985) has the following features. The myoclonic jerks are produced by voluntary movements or sensory stimulation. They involve a few contiguous, usually distal, muscles. The EMG bursts are brief (10±30 ms) and are preceded by a large cortical potential. The condition may be related to epilepsy partialis continua in which continuous myoclonic jerks occur (Obeso et al., 1985). The abnormality giving rise to cortical myoclonus may be in the sensory cortex (Cowan et al., 1986; Rothwell et al., 1986), the motor cortex (Mima et al., 1998) or both (Uesaka et al., 1996). The reason for the increase in cortical excitability is unknown. There may be local or widespread pathological changes in the cerebral cortex (Cowan et al., 1986) or the main pathological ®ndings may be elsewhere, for example in the cerebellum (Bhatia et al., 1995). We inves* Corresponding author. Tel.: 1 1-416-603-5017; fax: 1 1-416-6035768. E-mail address: [email protected] (P. Ashby)
tigated a patient with cortical re¯ex myoclonus by recording and stimulating through electrodes placed directly on the cortex. 1.1. Case history 1.1.1. History When studied in 1998, the patient was a 66-year-old woman with myoclonus of the left leg. She had a past history of celiac disease and in 1994 was found to have a T cell lymphoma, treated with Prednisone and Chlorambucil with apparent resolution. In 1993 she started to have twitching of the left leg which ®rst occurred when she moved. This progressed so that touching the leg or attempting to move the limb voluntarily produced a succession of myoclonic jerks. Eventually she was unable to stand or walk. She had no sensory symptoms. About 2 months previously she noticed, in addition, that she was unable to dorsi¯ex her left foot. She was not aware of any numbness of the foot. 1.1.2. Examination Power, tone and coordination in the upper limbs and right
1388-2457/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(99)00101-7
CLINPH 98180
1522
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
tibial nerve produced normal sensory evoked potentials with an onset of 32 ms, peaks at 40 and 62 ms and amplitudes of 3±10 mV. Stimulation of the left posterior tibial nerve, however, produced very large evoked potentials with an onset of 30 ms and with two to 3 additional components each followed by EMG bursts in the left leg muscles. The evoked potentials were larger when slow stimulus rates were used. For example, at 0.2 Hz, evoked potentials had an amplitude of 50 uV while stimulation at 1 Hz resulted in evoked potentials with amplitudes 10±25 uV. The evoked potentials from the upper limbs were equal on the two sides.
Fig. 1. Position of the electrode contacts on the medial surface of the right hemisphere. The contacts were 1 cm apart and 0.5 cm in diameter. Contact 13 was considered to lie over the leg area of the motor cortex and contact 12 over the leg area of the sensory cortex.
leg were normal. In the left leg bulk was normal. There were jerky involuntary movements of the left leg, which included plantar ¯exion and eversion of the foot, and ¯exion of the hip and knee. These jerks, occurring 2 or 3 at a time, were precipitated by attempted movements or by touching the leg. There was weakness of the left tibialis anterior and extensor hallucis longus, but not of peroneus longus. Coordination was impaired on the heel-shin test with several side-to-side jerks resembling cerebellar ataxia (but dif®cult to distinguish from action myoclonus). The upper limb re¯exes were normal, but the adductor, hamstrings, knee jerk and ankle jerk were all brisker on the left, often with double jerks. Both plantar re¯exes were ¯exor. Sensation was entirely normal. Vibration threshold was 0.5 microns on the ®ngers and toes, joint position normal in the left big toe, temperature and pin prick normal in the left leg. There was no cortical sensory loss in the left foot and no hemi-extinction. She was unable to walk or even transfer because of the intense jerking of the left leg. 1.1.3. Investigations 1.1.3.1. MRI MRI, including T1 weighted, T2 weighted, proton density and FLAIR sequences in the sagittal, axial and coronal planes plus high resolution axial and coronal inversion recovery sequences carried out prior to surgery, showed scattered small areas of altered signal in the whitematter (`microvascular') lesions but no local pathology in the interhemispheric ®ssure (including the presumed location of the motor cortex where contact 13 was later placed). 1.1.3.2. Evoked potentials Stimuli were applied to the posterior tibial nerve and recordings were made from C3CZ and C4-CZ. Stimulation of the (normal) right posterior
1.1.3.3. Magnetic stimulation over the cortex Transcranial magnetic stimulation with a circular coil (Caldwell MES 10) resulted in normal motor evoked potentials (MEPs) in the small muscles of both hands (resting threshold 44%, latency 19.6 ms). A MEP was evoked in the left quadriceps (threshold 64%) with a latency of 28 ms followed by 3 or more large MEPs with latencies 55, 90 and 110 ms. On the right, a quadriceps MEP was only rarely obtained at this intensity and required voluntary activation of the muscle. 1.1.3.4. EMG Nerve conduction studies were normal except for a partial left peroneal nerve lesion at the head of the ®bula. There was denervation of tibialis anterior but not peroneus longus. This was considered to account entirely for weakness of dorsi¯exion of the left foot. 1.1.3.5. EEG Routine scalp EEG demonstrated a mild to moderate intermittent non-epileptiform disturbance of cerebral activity recorded independently over both frontocentrotemporal regions, suggestive of dysfunction affecting subcortical white matter. No epileptiform activity was recorded. 1.1.4. Outcome The cortex under contact 13 (see below) was resected and the myoclonus was abolished. Unfortunately a local cerebral vein thrombosis occurred postoperatively, producing leftleg weakness. The weakness gradually resolved. Six months after the operation she was walking with a cane. The myoclonus had not returned. Histology of the resected cortex showed patchy recent infarction probably related to the surgical procedure. One small blood vessel contained amyloid. 2. Materials and methods 2.1. Surgical procedure A right craniotomy was performed and two rows of 8 electrodes were placed over the medial aspect of the hemisphere in the interhemispheric ®ssure (Fig. 1). Contacts 1±8 were below contact 9±16, the higher numbers were anterior.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
1523
Fig. 2. (A) Electrocorticographic (EcoG) recordings show bursts of multiple spikes and intervening slow waves maximal at contacts 13 . 14 s 6 (average referential montage; time bar interval 2 s). (B) Combined bipolar and monopolar referential (to lateral convexity inferior parietal electrode) montages at faster sweep speed show positive phase reversal of multiple spike bursts at contact 13. Time bar interval 2 s.
Five rows of contacts were also placed over the lateral aspect of the right hemisphere. Flexible connecting cables were lead out through the skull for stimulation and recording.
2.2. Recordings Cortical activity was ampli®ed 500 000 times with band pass 2.5 Hz±2.5 kHz, monitored on an oscilloscope and
1524
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
3.1. Electrocorticography (EcoG)
Fig. 3. Mapping the leg area of the motor cortex. Modulation of voluntary EMG in the left peroneus longus muscle by single stimuli (100 ms, 8 mA) given through the contacts indicated on the left (1, anode; -, cathode). The short latency facilitation of voluntary EMG is greatest (and has the shortest latency, 32.5 ms) with cathodal stimulation at contact 13. Each trace is the average of 50 sweeps of recti®ed EMG. Note the long inhibition following the facilitation in traces 4 and 5.
digitized at 2 kHz or 20 kHz. EMG activity, recorded with surface electrodes, was ampli®ed 5000 to 10 000 times with band pass 2.5 Hz±2.5 kHz and digitized at 2 kHz. The traces were averaged or differentiated and superimposed using data processing software (Cambridge Electronic Design Signal and Spike 2). 2.3. Stimulation Stimuli were delivered to the cortical leads with an isolated, constant current stimulator (World Precision Instrument, #A360). A pulse duration of 100 ms or 1 ms was used. 2.4. Cortical mapping First, as a screening procedure, single stimuli were delivered through each electrode in turn and the thresholds for visible contractions of various muscle groups noted. Then, recording electrodes were placed over certain leg muscles (peroneus longus and abductor hallucis). While the subject maintained a 30% maximum voluntary contraction of a chosen muscle, 50 sweeps of recti®ed EMG were averaged in relation to cortical stimulation. The latency and magnitude of the short latency facilitation resulting from these stimuli were recorded. 3. Results The position of the electrode contacts on the medial aspect of the right hemisphere traced from a skull X-ray is shown in Fig. 1.
Bursts of multiple spikes, associated with myoclonic jerks of the left leg, were recorded localized to contacts 13 and 14, with occasional minimal extension to involve contact 6. The spikes were electropositive at the cortical surface (with positive phase reversal at 13), and tended to occur in bursts of 4±6 spikes, each spike separated by 35±40 ms (,25±30 Hz). The bursts were typically followed by a surface negative slow wave (Fig. 2). Occasionally, runs of lower amplitude fast sinusoidal activity at ,35±40 Hz were discernable in the same location, either overriding the bursts of multiple spikes and slow waves or in isolation. When movement of the left leg elicited an especially long-lasting episode of re¯ex myoclonus (e.g. upon helping the patient move from bed to chair) the intervening slow waves became less pronounced and the bursts of multiple spikes organized into what amounted to electrographic seizure activity. 3.2. Cortical mapping Fig. 3 shows the results of stimulating between adjacent contacts (100 ms, 8 mA) during a 30% maximum voluntary contraction of the peroneus longus. The facilitation of voluntary EMG had the greatest amplitude and the shortest latency (32.5 ms) with stimulation at contacts 12 1 13 2 (Fig. 3). With monopolar stimulation, contact 13 gave the largest facilitation. Cathodal stimulation was slightly more effective than anodal stimulation. The leg area of the motor cortex was presumed to lie under contact 13. The averaged cortical potentials recorded in response to left tibial nerve stimulation (n 100) are shown in Fig. 4. The earliest potential occurred 31.9 ms after the stimulus at contacts 11±12. These contacts were presumed to lie over the sensory cortex. This initial potential was followed by a much larger potential (onset 35 ms) which reversed phase at contact 13. The differentiated traces showed no evidence for higher frequency components. 3.3. Cortical potentials preceding myoclonic jerks A typical series of `spontaneous' myoclonic jerks is shown in Fig. 5. In this single sweep cortical potentials, showing phase reversal with positivity at contact 13, were followed by EMG bursts in peroneus longus ~35 ms later. In this recording the mean interval between successive cortical potentials was 39.1 ms (range 16.9 to 63.8 ms, n 15). By triggering the sweep from EMG bursts in peroneus longus the preceding cortical potentials could be seen by backaveraging. The latency from the ®rst major de¯ection to the onset of EMG was ,35 ms, similar to the latency of the facilitation of EMG by stimulating through contact 13 (not shown). The differentiated, superimposed traces revealed no higher frequency potentials.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
1525
Fig. 4. Somatosensory evoked potentials in response to stimulation of the left tibial nerve at the ankle (given 10 ms after the start of the sweep) recorded from contacts (top down) 11±12, 12±13, 13±14, 14±15. Note that the earliest evoked response (1) occurs at contacts 11±12 (latency 31.9 ms). This is followed, 3.5 ms later (2), by a very large potential which shows phase reversal indicating positivity at contact 13. Vertical calibration IV represents 200 mV. Upward de¯ections are negative.
3.4. Conditioning the motor evoked potential from adjacent contacts The MEP evoked by stimulating through 12 1 13 - (4.6 mA, 1 ms) was conditioned by stimuli given at other contacts that were of equal intensity but subthreshold for a MEP. With conditioning stimuli at 14 1 15 - and 4 1 5 -, the MEP was depressed at 3 ms and facilitated at 10 ms intervals. Conditioning stimuli at 10 1 11 - did not produce the facilitation at 10 ms. (not shown). 3.5. Local cortical potentials produced by cortical stimulation Cathodal monopolar stimuli were given through contact 13 while averaging the cortical potentials at surrounding contacts 11±12, 12±5 and 5±14. Stimuli at contact 13 generated a positive potential at 14 (with up to 4 subcomponents) within the ®rst 10 ms. This was followed by a slow negative
wave and often a second and third positive potential (each lasting ~10 ms) each followed by a negative wave (Fig. 6). These positive peaks occurred at a mean of 43.9 (SE 0:3) ms and at a mean of 78.5 (SE 0:7) ms (n 44) after the stimulus. The mean interval between the second and third positive potentials was 34.6 (SE 0:3) ms. Each of the positive potentials was associated with a myoclonic jerk. The latency of the second positive potential was variable. When it occurred early, it was smaller. The amplitude and the frequency of occurrence were largest at ~45 ms. In some sweeps, there was no second positive potential. In these sweeps the initial positive potential (the one occurring in the ®rst 10 ms after the stimulus) was larger, broader and associated with a deeper subsequent negative potential. In these sweeps there was only a single myoclonic jerk that was related to the stimulus and the ®rst positive potential. In the sweeps with second and third positive potentials, the ®rst potential and the ®rst myoclonic jerk were smaller (Fig. 7). Paired stimulation at a condition-test interval of 10 ms
1526
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
Fig. 5. Typical sequence of spontaneous myoclonic jerks preceded by cortical potentials with phase reversal indicating positivity at contact 13 (single sweep). Upper trace: EMG from peroneus longus. Lower 3 traces: EEG recorded from (top down) cortical contacts 13±14, 12±13, 11±12. Vertical calibration 100 mV.
caused the initial positive potential to be larger and the second positive potential (expected ,40 ms later) was again absent. The subcomponents of the ®rst potential produced by
stimulation at contact 13 could be better seen by differentiating the traces (Fig. 8) or recording them with band pass 500 Hz±2.5 kHz (Fig. 9). The positive peaks had latencies of 2.3, 3.7, 5.3 and 6.5 ms (mean interval between peaks 1:4
Fig. 6. Cortical potentials recorded at contacts 5±14 in response to single stimuli (n 100) given through contact 13. Upward de¯ections are positive. There is a positive potential in the ®rst 10 ms with 4 subcomponents. This is followed by a long negative wave and, on many sweeps, second and third positive potentials each followed by long negative waves. The second positive wave is smaller if it occurs early.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
1527
latency. This contact was presumed to lie over the motor cortex. Stimulation of the tibial nerve evoked a cortical potential with the shortest latency at contacts 11±12. These contacts were presumed to lie over the sensory cortex. 4.2. Site of origin of the myoclonus
Fig. 7. Same data as Fig. 6. Average of traces without (top) and with (bottom) the second positive potential. When the second positive potential was absent, the ®rst positive and negative wave and the ®rst myoclonic jerk were larger. The top trace of each pair represents the recordings from contacts 5±14. The bottom trace of each pair is the averaged recti®ed EMG from abductor hallucis. Calibration: 5 V represents 100 uV.
ms). With paired stimulation at intervals of 2±20 ms, the amplitude of the ®rst of these subcomponents was unchanged but the later subcomponents were suppressed. They recovered successively over ~20 ms (Figs. 10 and 11). These 4 waves occurred immediately after the stimulus but did not accompany subsequent positive waves (Fig. 8), nor were they discernable on the superimposed, differentiated records of the cortical potentials accompanying spontaneous jerks or those induced by peripheral nerve stimulation. Stimulation at contact 12 produced an initial positive potential, followed by a negative potential and occasionally subsequent positive potentials, but these were all were delayed , 10 ms compared to those elicited at contact 13. No myoclonic jerks occurred. Stimulation at contact 14 generated an initial wave (positive at 13) but no subsequent waves or myoclonic jerks.
4. Discussion 4.1. Cortical mapping Stimulation through contact 13 facilitated the motoneurons of leg muscles with the lowest threshold and shortest
The myoclonic jerks were preceded by cortical potentials that reversed polarity at contact 13. These could be seen on individual sweeps and on back-averages triggered from EMG bursts. The latency from the cortical spike to the EMG burst was similar to that of the facilitation of voluntary EMG produced by stimulating through this contact, suggesting that both used the same fast-conducting efferent pathway, likely the corticospinal tract. Although the `giant' cortical potentials and myoclonic jerks could be produced by tibial nerve stimulation, the sensory cortex did not appear to be the site of the major abnormality. The earliest sensory evoked potential was at contacts 11±12 but the `giant' potentials occurred ~3 ms later at contact 13. Direct stimulation through 13 produced a typical cluster of myoclonic jerks while stimulation at the same intensity through adjacent contacts, including 11 and 12, did not. Furthermore, conditioning the motor potential evoked by stimulation at contact 13 with a preceding stimulus at 11±12 did not produce more facilitation than stimulation through other contacts such as 14±15. In this case, as in others described previously (Mima et al., 1998), the myoclonus appeared to originate in the motor cortex. 4.3. Excitability of the corticospinal neurons The surface electropositivity of the multiple spike bursts indicates that the depolarization of cortical neurons was occurring at the cell body. Potentials generated in the apical dendrites result in surface electronegativity. This is consistent with the observation that the spike potentials in a similar case reversed phase 0.5±1 mm below the cortical surface and disappeared a few millimeters deeper (Kugelberg and Widen, 1954). It is likely that the affected area of the right motor cortex was more excitable than normal. Stimulation at contact 13 but not at other sites produced a positive potential followed by a myoclonic jerk. The single stimulus evidently led to the synchronous discharge of many cortical neurons, implying that these neurons were more excitable or more readily synchronized (Obeso et al., 1985). The motor potentials evoked in leg muscles by transcranial magnetic stimulation had a lower threshold on that side. This may re¯ect an alteration in sodium channel dynamics; this threshold is increased by sodium channel blocking antiepileptic drugs (Ziemann et al., 1996; Chen et al., 1997). 4.4. Excitability of cortical interneuron systems Three interneuron systems in the cortex under contact 13 appeared to be functional, although we cannot say if they were up or down regulated. (1) The MEP produced by
1528
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
Fig. 8. Same data as Fig. 6, differentiated to accentuate the 4 potentials occurring within the ®rst 10 ms after the stimulus. These potentials do not accompany the subsequent positive potentials at ~50 ms and ~90 ms.
stimulation at 13 was followed by a long silent period (Fig. 3) which has been attributed to inhibition within the cortex. (2) Short latency inhibition and facilitation, similar to that
described with paired magnetic stimuli (Kujirai et al., 1993) could be demonstrated by conditioning the MEP produced by stimulation at 12 1 13 2 with stimulation at neighbouring sites. These inhibitory and facilitatory interneurons were
Fig. 9. Cortical potentials produced by stimulating through (top down) contacts 14, 13, 12. The recordings were made with a high-pass ®lter of 500 Hz from contacts (top down) 6±15, 5±14, 4±13. Note the series of waves produced by stimulation through contact 13. The positive peaks have latencies of 2.3, 3.7, 5.3 and 6.5 ms after the stimulus (the mean interpeak interval 1:4 ms).
Fig. 10. Cortical potentials, recorded at contacts 5±14 with band width 500 Hz-±2.5 kHz, resulting from paired stimulation (two equal pulses 100 ms, 9 mA) through contact 13. Note that, at the 5 ms interstimulus interval, the second and third potentials are diminished but the ®rst is not. At the 10 ms interstimulus interval, the second and third potential are slightly larger than those produced by the ®rst stimulus but the fourth potential is still reduced.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
Fig. 11. Amplitude (expressed as a percent of control) of each of the 4 cortical potentials at various condition-test intervals. Note that the ®rst potential is not depressed and that the others recover at successively longer intervals becoming transiently facilitated at a condition-test interval of 10 ms.
evidently present. In cortical myoclonus less inhibition of the MEP in the period from 1 to 5 ms following a subthreshold transcranial magnetic stimulus has been reported, but only in the type in which the abnormal activity spread across the cortex (Brown et al., 1996). (3) The positive wave produced by direct cortical stimulation was accompanied by a series of smaller potentials with latencies similar to the succession of I waves that follow a cortical stimulus. When paired stimuli were given, the later potentials were suppressed. The order of recovery of these waves was similar to that reported for I waves following paired transcranial magnetic stimulation (Di Lazzaro et al., 1998). We postulate that these potentials are the cortical counterpart of I waves and surmise that the I wave circuitry is intact in this area of cortex. In patients with cortical mycolonus it has been noted that the EMG bursts resulted from a succession of brief facilitations of motoneurons at ~3.5 ms intervals (~half the frequency of the presumed I waves seen here (Brown and Marsden, 1996). In the present case, however, there was no trace of the presumed I waves on the differentiated traces of cortical potentials associated with myoclonic jerks occurring spontaneously or in response to peripheral stimuli, so these interneurons did not seem to be involved in the generation of myoclonus. 4.5. Why do the myoclonic jerks recur? Whether occurring spontaneously or elicited by afferent or cortical stimulation, the cortical potentials and subsequent myoclonic jerks often recurred 2±3 times at ~35-40 ms intervals. This was not because the ®rst jerk resulted in a sensory volley that triggered another jerk - the interval was too short. The shortest turn-around time from cortex to the foot and back to cortex would be 31:9 1 3 1 35 ~70ms.
1529
Furthermore, the ®rst of a succession of cortical potentials often did not give rise to an EMG burst (e.g. Fig. 5). Direct stimulation of the cortex gave rise to a succession of positive potentials (at ,35±40 ms intervals) each followed by a myoclonic jerk. Each positive wave lasted about 10 ms and was followed by a longer negative wave which it evidently generated. The excitability of the cortex appeared to be reduced during the negative wave: the positive potentials occurred with increasing probability towards the end of the negative wave and were smaller if they occurred earlier (Fig. 6). In sweeps in which the second positive wave was absent, the initial positive wave and the subsequent negative wave were larger (Fig. 7). The same phenomena occurred with paired stimuli 10 ms apart. The initial positive potential and the presumed I waves were larger and subsequent positive potentials were abolished. It may be that the additional inhibition related to a large positive potential curtails further myoclonic jerks. It is not possible by recording ®eld potentials such as these to distinguish between the many possible underlying mechanisms. This recurrent activity could re¯ect the tendency for single neurons to oscillate or the behaviour of feed-back circuitry. Kanthasamy et al. (1996), for example, reported that rhythmic mycolonus could be induced in normal rats by injecting the GABA-A antagonist bicuculine into the reticular nucleus of the thalamus, and postulated that disinhibition of the reticular nucleus, and its calciumdependent burst ®ring, generated myoclonus by oscillations in a thalamo-cortical loop. In the present case, the absence of presumed I waves on the second and third positive potentials (Fig. 8) indicates that the recurrent activity does not result from the indirect activation of cortical neurons via the interneurons responsible for the I waves. On the other hand, the positive and negative potentials recorded at the cortex associated with the recurrent myoclonic jerks suggest that local events, such as an abnormally synchronous discharge and subsequent after hyperpolarization and reexcitation of cortical neurons, might be a suf®cient explanation. References Bhatia KP, Brown P, Gregory R, Lennox GG, Manji H, Thompson PD, Ellison DW. Progressive myoclonic ataxia associated with coeliac disease. The myoclonus is of cortical origin, but the pathology is in the cerebellum, Brain 1995;118:1087±1093. Brown P, Marsden CD. Rhythmic cortical and muscle discharge in cortical myoclonus. Brain 1996;119:1307±1316. Brown P, Ridding MC, Werhahn KJ, Rothwell JC, Marsden CD. Abnormalities of the balance between inhibition and excitation in the motor cortex of patients with cortical myoclonus. Brain 1996;119:309±317. Chen R, Samii A, Canos M, Wasserman EM, Hallett M. Effects of phenytoin on cortical excitability in humans. Neurology 1997;49:881±883. Cowan JMA, Rothwell JC, Wise RJS, Marsden CD. Electrophysiological and positron emission studies in a patient with cortical myoclonus, epilpsia partialis continua and motor epilepsy. J Neurol Neurosurg Psychiatry 1986;49:796±807. Di Lazzaro V, Restuccia D, Oliviero A, Pro®ce P, Ferrara L, Insola A, Mazzone P, Tonali P, Rothwell JC. Magnetic transcranial stimulation
1530
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
at intensities below active motor threshold activates intracortical inhibitory circuits. Exp Brain Res 1998;119:265±268. Hallett M, Chadwick D, Marsden CD. Cortical re¯ex myoclonus. Neurology 1979;29:1107±1125. Kanthasamy AG, Matsumoto RR, Truong DD. Animal model of myoclonus. Clin Neurosci 1996;3:236±245. Kugelberg E, Widen L. Epilepsia Partialis Continua. Electroenceph clin Neurophysiol 1954;6:503±506. Kujirai T, Caramia MD, Rothwell JC, Day BL, Thompson PD, Ferberts S, Wroe S, Asselman P, Marsden CD. Corticospinal inhibition in human motor cortex. J Physiol 1993;471:501±519. Mima T, Nagamine T, Nishitani N, Mikuni N, Ikeda A, Fukuyama H,
Takigawa T, Kimura J, Shibasaki H. Cortical myoclonus. Sensorimotor hyperexcitability. Neurology 1998;50:933±942. Obeso JA, Rothwell JC, Marsden CD. The spectrum of cortical myoclonus. Brain 1985;108:193±224. Rothwell JC, Obeso JA, Marsden CD. Electrophysiology of somatosensory re¯ex myoclonus. Adv Neurol 1986;43:385±398. Uesaka Y, Terao Y, Ugawa Y, Yumoto M, Hanajima R, Kanazawa I. Magnetoencephalographic analysis of cortical myoclonic jerks. Electroenceph clin Neurophysiol 1996;99:141±148. Ziemann U, Lonnecker S, Steinhoff BJ, Paulus W. Effects of antiepileptic drugs on motor cortex excitability in humans: a transcranial magnetic stimulation study. Ann Neurol 1996;40:367±378.
Cortical re¯ex myoclonus studied with cortical electrodes P. Ashby*, R. Chen, R. Wennberg, A.M. Lozano, A.E. Lang Playfair Neuroscience Unit, Toronto Hospital, Western Division, University of Toronto, 399 Bathurst Street, EC8-005 Toronto, Ontario M5T 2S8, Canada Accepted 12 April 1999
Abstract Objective: To study the mechanism of cortical re¯ex myoclonus. Methods: A patient with stimulus sensitive myoclonus of the left foot had an array of subdural electrodes placed over the right sensorimotor cortex. Results: Stimulation through one of the electrodes (contact 13) facilitated leg muscles with the shortest latency and was presumed to lie over the motor cortex. Tibial nerve stimulation evoked a potential with the shortest latency 1 cm further posteriorly (contacts 11±12). These contacts were presumed to lie over the sensory cortex. The potential at 11±12 was followed by a much larger potential that reversed polarity at contact 13. Back averaging from spontaneous myoclonic jerks showed a cortical premovement potential which reversed polarity at contact 13. The threshold for the motor evoked potential in leg muscles evoked by transcranial magnetic stimulation was lower on the affected side. Electrical stimulation through contact 13 produced cortical potentials that could be recorded at adjacent contacts. The combination of a positive potential followed by a negative potential recurred at ~35±40 ms intervals, each positive potential generating a myoclonic jerk. Additional waves resembling I waves accompanied only the ®rst positive potential. Surgical removal of the cortex under electrode 13 abolished the myoclonus. Conclusions: The myoclonic jerks arose in the motor cortex. We postulate that there is increased excitability or synchronization of the cortical neurons at that site. The spontaneous, peripherally induced and recurrent cortical potentials and myoclonic jerks can occur without participation of the circuitry of the presumed I waves. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Cortical myoclonus; Cortical stimulation; Cortical inhibition; Motor cortex; Transcranial magnetic stimulation
1. Introduction Cortical re¯ex myoclonus (Hallett et al., 1979; Obeso et al., 1985) has the following features. The myoclonic jerks are produced by voluntary movements or sensory stimulation. They involve a few contiguous, usually distal, muscles. The EMG bursts are brief (10±30 ms) and are preceded by a large cortical potential. The condition may be related to epilepsy partialis continua in which continuous myoclonic jerks occur (Obeso et al., 1985). The abnormality giving rise to cortical myoclonus may be in the sensory cortex (Cowan et al., 1986; Rothwell et al., 1986), the motor cortex (Mima et al., 1998) or both (Uesaka et al., 1996). The reason for the increase in cortical excitability is unknown. There may be local or widespread pathological changes in the cerebral cortex (Cowan et al., 1986) or the main pathological ®ndings may be elsewhere, for example in the cerebellum (Bhatia et al., 1995). We inves* Corresponding author. Tel.: 1 1-416-603-5017; fax: 1 1-416-6035768. E-mail address: [email protected] (P. Ashby)
tigated a patient with cortical re¯ex myoclonus by recording and stimulating through electrodes placed directly on the cortex. 1.1. Case history 1.1.1. History When studied in 1998, the patient was a 66-year-old woman with myoclonus of the left leg. She had a past history of celiac disease and in 1994 was found to have a T cell lymphoma, treated with Prednisone and Chlorambucil with apparent resolution. In 1993 she started to have twitching of the left leg which ®rst occurred when she moved. This progressed so that touching the leg or attempting to move the limb voluntarily produced a succession of myoclonic jerks. Eventually she was unable to stand or walk. She had no sensory symptoms. About 2 months previously she noticed, in addition, that she was unable to dorsi¯ex her left foot. She was not aware of any numbness of the foot. 1.1.2. Examination Power, tone and coordination in the upper limbs and right
1388-2457/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(99)00101-7
CLINPH 98180
1522
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
tibial nerve produced normal sensory evoked potentials with an onset of 32 ms, peaks at 40 and 62 ms and amplitudes of 3±10 mV. Stimulation of the left posterior tibial nerve, however, produced very large evoked potentials with an onset of 30 ms and with two to 3 additional components each followed by EMG bursts in the left leg muscles. The evoked potentials were larger when slow stimulus rates were used. For example, at 0.2 Hz, evoked potentials had an amplitude of 50 uV while stimulation at 1 Hz resulted in evoked potentials with amplitudes 10±25 uV. The evoked potentials from the upper limbs were equal on the two sides.
Fig. 1. Position of the electrode contacts on the medial surface of the right hemisphere. The contacts were 1 cm apart and 0.5 cm in diameter. Contact 13 was considered to lie over the leg area of the motor cortex and contact 12 over the leg area of the sensory cortex.
leg were normal. In the left leg bulk was normal. There were jerky involuntary movements of the left leg, which included plantar ¯exion and eversion of the foot, and ¯exion of the hip and knee. These jerks, occurring 2 or 3 at a time, were precipitated by attempted movements or by touching the leg. There was weakness of the left tibialis anterior and extensor hallucis longus, but not of peroneus longus. Coordination was impaired on the heel-shin test with several side-to-side jerks resembling cerebellar ataxia (but dif®cult to distinguish from action myoclonus). The upper limb re¯exes were normal, but the adductor, hamstrings, knee jerk and ankle jerk were all brisker on the left, often with double jerks. Both plantar re¯exes were ¯exor. Sensation was entirely normal. Vibration threshold was 0.5 microns on the ®ngers and toes, joint position normal in the left big toe, temperature and pin prick normal in the left leg. There was no cortical sensory loss in the left foot and no hemi-extinction. She was unable to walk or even transfer because of the intense jerking of the left leg. 1.1.3. Investigations 1.1.3.1. MRI MRI, including T1 weighted, T2 weighted, proton density and FLAIR sequences in the sagittal, axial and coronal planes plus high resolution axial and coronal inversion recovery sequences carried out prior to surgery, showed scattered small areas of altered signal in the whitematter (`microvascular') lesions but no local pathology in the interhemispheric ®ssure (including the presumed location of the motor cortex where contact 13 was later placed). 1.1.3.2. Evoked potentials Stimuli were applied to the posterior tibial nerve and recordings were made from C3CZ and C4-CZ. Stimulation of the (normal) right posterior
1.1.3.3. Magnetic stimulation over the cortex Transcranial magnetic stimulation with a circular coil (Caldwell MES 10) resulted in normal motor evoked potentials (MEPs) in the small muscles of both hands (resting threshold 44%, latency 19.6 ms). A MEP was evoked in the left quadriceps (threshold 64%) with a latency of 28 ms followed by 3 or more large MEPs with latencies 55, 90 and 110 ms. On the right, a quadriceps MEP was only rarely obtained at this intensity and required voluntary activation of the muscle. 1.1.3.4. EMG Nerve conduction studies were normal except for a partial left peroneal nerve lesion at the head of the ®bula. There was denervation of tibialis anterior but not peroneus longus. This was considered to account entirely for weakness of dorsi¯exion of the left foot. 1.1.3.5. EEG Routine scalp EEG demonstrated a mild to moderate intermittent non-epileptiform disturbance of cerebral activity recorded independently over both frontocentrotemporal regions, suggestive of dysfunction affecting subcortical white matter. No epileptiform activity was recorded. 1.1.4. Outcome The cortex under contact 13 (see below) was resected and the myoclonus was abolished. Unfortunately a local cerebral vein thrombosis occurred postoperatively, producing leftleg weakness. The weakness gradually resolved. Six months after the operation she was walking with a cane. The myoclonus had not returned. Histology of the resected cortex showed patchy recent infarction probably related to the surgical procedure. One small blood vessel contained amyloid. 2. Materials and methods 2.1. Surgical procedure A right craniotomy was performed and two rows of 8 electrodes were placed over the medial aspect of the hemisphere in the interhemispheric ®ssure (Fig. 1). Contacts 1±8 were below contact 9±16, the higher numbers were anterior.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
1523
Fig. 2. (A) Electrocorticographic (EcoG) recordings show bursts of multiple spikes and intervening slow waves maximal at contacts 13 . 14 s 6 (average referential montage; time bar interval 2 s). (B) Combined bipolar and monopolar referential (to lateral convexity inferior parietal electrode) montages at faster sweep speed show positive phase reversal of multiple spike bursts at contact 13. Time bar interval 2 s.
Five rows of contacts were also placed over the lateral aspect of the right hemisphere. Flexible connecting cables were lead out through the skull for stimulation and recording.
2.2. Recordings Cortical activity was ampli®ed 500 000 times with band pass 2.5 Hz±2.5 kHz, monitored on an oscilloscope and
1524
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
3.1. Electrocorticography (EcoG)
Fig. 3. Mapping the leg area of the motor cortex. Modulation of voluntary EMG in the left peroneus longus muscle by single stimuli (100 ms, 8 mA) given through the contacts indicated on the left (1, anode; -, cathode). The short latency facilitation of voluntary EMG is greatest (and has the shortest latency, 32.5 ms) with cathodal stimulation at contact 13. Each trace is the average of 50 sweeps of recti®ed EMG. Note the long inhibition following the facilitation in traces 4 and 5.
digitized at 2 kHz or 20 kHz. EMG activity, recorded with surface electrodes, was ampli®ed 5000 to 10 000 times with band pass 2.5 Hz±2.5 kHz and digitized at 2 kHz. The traces were averaged or differentiated and superimposed using data processing software (Cambridge Electronic Design Signal and Spike 2). 2.3. Stimulation Stimuli were delivered to the cortical leads with an isolated, constant current stimulator (World Precision Instrument, #A360). A pulse duration of 100 ms or 1 ms was used. 2.4. Cortical mapping First, as a screening procedure, single stimuli were delivered through each electrode in turn and the thresholds for visible contractions of various muscle groups noted. Then, recording electrodes were placed over certain leg muscles (peroneus longus and abductor hallucis). While the subject maintained a 30% maximum voluntary contraction of a chosen muscle, 50 sweeps of recti®ed EMG were averaged in relation to cortical stimulation. The latency and magnitude of the short latency facilitation resulting from these stimuli were recorded. 3. Results The position of the electrode contacts on the medial aspect of the right hemisphere traced from a skull X-ray is shown in Fig. 1.
Bursts of multiple spikes, associated with myoclonic jerks of the left leg, were recorded localized to contacts 13 and 14, with occasional minimal extension to involve contact 6. The spikes were electropositive at the cortical surface (with positive phase reversal at 13), and tended to occur in bursts of 4±6 spikes, each spike separated by 35±40 ms (,25±30 Hz). The bursts were typically followed by a surface negative slow wave (Fig. 2). Occasionally, runs of lower amplitude fast sinusoidal activity at ,35±40 Hz were discernable in the same location, either overriding the bursts of multiple spikes and slow waves or in isolation. When movement of the left leg elicited an especially long-lasting episode of re¯ex myoclonus (e.g. upon helping the patient move from bed to chair) the intervening slow waves became less pronounced and the bursts of multiple spikes organized into what amounted to electrographic seizure activity. 3.2. Cortical mapping Fig. 3 shows the results of stimulating between adjacent contacts (100 ms, 8 mA) during a 30% maximum voluntary contraction of the peroneus longus. The facilitation of voluntary EMG had the greatest amplitude and the shortest latency (32.5 ms) with stimulation at contacts 12 1 13 2 (Fig. 3). With monopolar stimulation, contact 13 gave the largest facilitation. Cathodal stimulation was slightly more effective than anodal stimulation. The leg area of the motor cortex was presumed to lie under contact 13. The averaged cortical potentials recorded in response to left tibial nerve stimulation (n 100) are shown in Fig. 4. The earliest potential occurred 31.9 ms after the stimulus at contacts 11±12. These contacts were presumed to lie over the sensory cortex. This initial potential was followed by a much larger potential (onset 35 ms) which reversed phase at contact 13. The differentiated traces showed no evidence for higher frequency components. 3.3. Cortical potentials preceding myoclonic jerks A typical series of `spontaneous' myoclonic jerks is shown in Fig. 5. In this single sweep cortical potentials, showing phase reversal with positivity at contact 13, were followed by EMG bursts in peroneus longus ~35 ms later. In this recording the mean interval between successive cortical potentials was 39.1 ms (range 16.9 to 63.8 ms, n 15). By triggering the sweep from EMG bursts in peroneus longus the preceding cortical potentials could be seen by backaveraging. The latency from the ®rst major de¯ection to the onset of EMG was ,35 ms, similar to the latency of the facilitation of EMG by stimulating through contact 13 (not shown). The differentiated, superimposed traces revealed no higher frequency potentials.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
1525
Fig. 4. Somatosensory evoked potentials in response to stimulation of the left tibial nerve at the ankle (given 10 ms after the start of the sweep) recorded from contacts (top down) 11±12, 12±13, 13±14, 14±15. Note that the earliest evoked response (1) occurs at contacts 11±12 (latency 31.9 ms). This is followed, 3.5 ms later (2), by a very large potential which shows phase reversal indicating positivity at contact 13. Vertical calibration IV represents 200 mV. Upward de¯ections are negative.
3.4. Conditioning the motor evoked potential from adjacent contacts The MEP evoked by stimulating through 12 1 13 - (4.6 mA, 1 ms) was conditioned by stimuli given at other contacts that were of equal intensity but subthreshold for a MEP. With conditioning stimuli at 14 1 15 - and 4 1 5 -, the MEP was depressed at 3 ms and facilitated at 10 ms intervals. Conditioning stimuli at 10 1 11 - did not produce the facilitation at 10 ms. (not shown). 3.5. Local cortical potentials produced by cortical stimulation Cathodal monopolar stimuli were given through contact 13 while averaging the cortical potentials at surrounding contacts 11±12, 12±5 and 5±14. Stimuli at contact 13 generated a positive potential at 14 (with up to 4 subcomponents) within the ®rst 10 ms. This was followed by a slow negative
wave and often a second and third positive potential (each lasting ~10 ms) each followed by a negative wave (Fig. 6). These positive peaks occurred at a mean of 43.9 (SE 0:3) ms and at a mean of 78.5 (SE 0:7) ms (n 44) after the stimulus. The mean interval between the second and third positive potentials was 34.6 (SE 0:3) ms. Each of the positive potentials was associated with a myoclonic jerk. The latency of the second positive potential was variable. When it occurred early, it was smaller. The amplitude and the frequency of occurrence were largest at ~45 ms. In some sweeps, there was no second positive potential. In these sweeps the initial positive potential (the one occurring in the ®rst 10 ms after the stimulus) was larger, broader and associated with a deeper subsequent negative potential. In these sweeps there was only a single myoclonic jerk that was related to the stimulus and the ®rst positive potential. In the sweeps with second and third positive potentials, the ®rst potential and the ®rst myoclonic jerk were smaller (Fig. 7). Paired stimulation at a condition-test interval of 10 ms
1526
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
Fig. 5. Typical sequence of spontaneous myoclonic jerks preceded by cortical potentials with phase reversal indicating positivity at contact 13 (single sweep). Upper trace: EMG from peroneus longus. Lower 3 traces: EEG recorded from (top down) cortical contacts 13±14, 12±13, 11±12. Vertical calibration 100 mV.
caused the initial positive potential to be larger and the second positive potential (expected ,40 ms later) was again absent. The subcomponents of the ®rst potential produced by
stimulation at contact 13 could be better seen by differentiating the traces (Fig. 8) or recording them with band pass 500 Hz±2.5 kHz (Fig. 9). The positive peaks had latencies of 2.3, 3.7, 5.3 and 6.5 ms (mean interval between peaks 1:4
Fig. 6. Cortical potentials recorded at contacts 5±14 in response to single stimuli (n 100) given through contact 13. Upward de¯ections are positive. There is a positive potential in the ®rst 10 ms with 4 subcomponents. This is followed by a long negative wave and, on many sweeps, second and third positive potentials each followed by long negative waves. The second positive wave is smaller if it occurs early.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
1527
latency. This contact was presumed to lie over the motor cortex. Stimulation of the tibial nerve evoked a cortical potential with the shortest latency at contacts 11±12. These contacts were presumed to lie over the sensory cortex. 4.2. Site of origin of the myoclonus
Fig. 7. Same data as Fig. 6. Average of traces without (top) and with (bottom) the second positive potential. When the second positive potential was absent, the ®rst positive and negative wave and the ®rst myoclonic jerk were larger. The top trace of each pair represents the recordings from contacts 5±14. The bottom trace of each pair is the averaged recti®ed EMG from abductor hallucis. Calibration: 5 V represents 100 uV.
ms). With paired stimulation at intervals of 2±20 ms, the amplitude of the ®rst of these subcomponents was unchanged but the later subcomponents were suppressed. They recovered successively over ~20 ms (Figs. 10 and 11). These 4 waves occurred immediately after the stimulus but did not accompany subsequent positive waves (Fig. 8), nor were they discernable on the superimposed, differentiated records of the cortical potentials accompanying spontaneous jerks or those induced by peripheral nerve stimulation. Stimulation at contact 12 produced an initial positive potential, followed by a negative potential and occasionally subsequent positive potentials, but these were all were delayed , 10 ms compared to those elicited at contact 13. No myoclonic jerks occurred. Stimulation at contact 14 generated an initial wave (positive at 13) but no subsequent waves or myoclonic jerks.
4. Discussion 4.1. Cortical mapping Stimulation through contact 13 facilitated the motoneurons of leg muscles with the lowest threshold and shortest
The myoclonic jerks were preceded by cortical potentials that reversed polarity at contact 13. These could be seen on individual sweeps and on back-averages triggered from EMG bursts. The latency from the cortical spike to the EMG burst was similar to that of the facilitation of voluntary EMG produced by stimulating through this contact, suggesting that both used the same fast-conducting efferent pathway, likely the corticospinal tract. Although the `giant' cortical potentials and myoclonic jerks could be produced by tibial nerve stimulation, the sensory cortex did not appear to be the site of the major abnormality. The earliest sensory evoked potential was at contacts 11±12 but the `giant' potentials occurred ~3 ms later at contact 13. Direct stimulation through 13 produced a typical cluster of myoclonic jerks while stimulation at the same intensity through adjacent contacts, including 11 and 12, did not. Furthermore, conditioning the motor potential evoked by stimulation at contact 13 with a preceding stimulus at 11±12 did not produce more facilitation than stimulation through other contacts such as 14±15. In this case, as in others described previously (Mima et al., 1998), the myoclonus appeared to originate in the motor cortex. 4.3. Excitability of the corticospinal neurons The surface electropositivity of the multiple spike bursts indicates that the depolarization of cortical neurons was occurring at the cell body. Potentials generated in the apical dendrites result in surface electronegativity. This is consistent with the observation that the spike potentials in a similar case reversed phase 0.5±1 mm below the cortical surface and disappeared a few millimeters deeper (Kugelberg and Widen, 1954). It is likely that the affected area of the right motor cortex was more excitable than normal. Stimulation at contact 13 but not at other sites produced a positive potential followed by a myoclonic jerk. The single stimulus evidently led to the synchronous discharge of many cortical neurons, implying that these neurons were more excitable or more readily synchronized (Obeso et al., 1985). The motor potentials evoked in leg muscles by transcranial magnetic stimulation had a lower threshold on that side. This may re¯ect an alteration in sodium channel dynamics; this threshold is increased by sodium channel blocking antiepileptic drugs (Ziemann et al., 1996; Chen et al., 1997). 4.4. Excitability of cortical interneuron systems Three interneuron systems in the cortex under contact 13 appeared to be functional, although we cannot say if they were up or down regulated. (1) The MEP produced by
1528
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
Fig. 8. Same data as Fig. 6, differentiated to accentuate the 4 potentials occurring within the ®rst 10 ms after the stimulus. These potentials do not accompany the subsequent positive potentials at ~50 ms and ~90 ms.
stimulation at 13 was followed by a long silent period (Fig. 3) which has been attributed to inhibition within the cortex. (2) Short latency inhibition and facilitation, similar to that
described with paired magnetic stimuli (Kujirai et al., 1993) could be demonstrated by conditioning the MEP produced by stimulation at 12 1 13 2 with stimulation at neighbouring sites. These inhibitory and facilitatory interneurons were
Fig. 9. Cortical potentials produced by stimulating through (top down) contacts 14, 13, 12. The recordings were made with a high-pass ®lter of 500 Hz from contacts (top down) 6±15, 5±14, 4±13. Note the series of waves produced by stimulation through contact 13. The positive peaks have latencies of 2.3, 3.7, 5.3 and 6.5 ms after the stimulus (the mean interpeak interval 1:4 ms).
Fig. 10. Cortical potentials, recorded at contacts 5±14 with band width 500 Hz-±2.5 kHz, resulting from paired stimulation (two equal pulses 100 ms, 9 mA) through contact 13. Note that, at the 5 ms interstimulus interval, the second and third potentials are diminished but the ®rst is not. At the 10 ms interstimulus interval, the second and third potential are slightly larger than those produced by the ®rst stimulus but the fourth potential is still reduced.
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
Fig. 11. Amplitude (expressed as a percent of control) of each of the 4 cortical potentials at various condition-test intervals. Note that the ®rst potential is not depressed and that the others recover at successively longer intervals becoming transiently facilitated at a condition-test interval of 10 ms.
evidently present. In cortical myoclonus less inhibition of the MEP in the period from 1 to 5 ms following a subthreshold transcranial magnetic stimulus has been reported, but only in the type in which the abnormal activity spread across the cortex (Brown et al., 1996). (3) The positive wave produced by direct cortical stimulation was accompanied by a series of smaller potentials with latencies similar to the succession of I waves that follow a cortical stimulus. When paired stimuli were given, the later potentials were suppressed. The order of recovery of these waves was similar to that reported for I waves following paired transcranial magnetic stimulation (Di Lazzaro et al., 1998). We postulate that these potentials are the cortical counterpart of I waves and surmise that the I wave circuitry is intact in this area of cortex. In patients with cortical mycolonus it has been noted that the EMG bursts resulted from a succession of brief facilitations of motoneurons at ~3.5 ms intervals (~half the frequency of the presumed I waves seen here (Brown and Marsden, 1996). In the present case, however, there was no trace of the presumed I waves on the differentiated traces of cortical potentials associated with myoclonic jerks occurring spontaneously or in response to peripheral stimuli, so these interneurons did not seem to be involved in the generation of myoclonus. 4.5. Why do the myoclonic jerks recur? Whether occurring spontaneously or elicited by afferent or cortical stimulation, the cortical potentials and subsequent myoclonic jerks often recurred 2±3 times at ~35-40 ms intervals. This was not because the ®rst jerk resulted in a sensory volley that triggered another jerk - the interval was too short. The shortest turn-around time from cortex to the foot and back to cortex would be 31:9 1 3 1 35 ~70ms.
1529
Furthermore, the ®rst of a succession of cortical potentials often did not give rise to an EMG burst (e.g. Fig. 5). Direct stimulation of the cortex gave rise to a succession of positive potentials (at ,35±40 ms intervals) each followed by a myoclonic jerk. Each positive wave lasted about 10 ms and was followed by a longer negative wave which it evidently generated. The excitability of the cortex appeared to be reduced during the negative wave: the positive potentials occurred with increasing probability towards the end of the negative wave and were smaller if they occurred earlier (Fig. 6). In sweeps in which the second positive wave was absent, the initial positive wave and the subsequent negative wave were larger (Fig. 7). The same phenomena occurred with paired stimuli 10 ms apart. The initial positive potential and the presumed I waves were larger and subsequent positive potentials were abolished. It may be that the additional inhibition related to a large positive potential curtails further myoclonic jerks. It is not possible by recording ®eld potentials such as these to distinguish between the many possible underlying mechanisms. This recurrent activity could re¯ect the tendency for single neurons to oscillate or the behaviour of feed-back circuitry. Kanthasamy et al. (1996), for example, reported that rhythmic mycolonus could be induced in normal rats by injecting the GABA-A antagonist bicuculine into the reticular nucleus of the thalamus, and postulated that disinhibition of the reticular nucleus, and its calciumdependent burst ®ring, generated myoclonus by oscillations in a thalamo-cortical loop. In the present case, the absence of presumed I waves on the second and third positive potentials (Fig. 8) indicates that the recurrent activity does not result from the indirect activation of cortical neurons via the interneurons responsible for the I waves. On the other hand, the positive and negative potentials recorded at the cortex associated with the recurrent myoclonic jerks suggest that local events, such as an abnormally synchronous discharge and subsequent after hyperpolarization and reexcitation of cortical neurons, might be a suf®cient explanation. References Bhatia KP, Brown P, Gregory R, Lennox GG, Manji H, Thompson PD, Ellison DW. Progressive myoclonic ataxia associated with coeliac disease. The myoclonus is of cortical origin, but the pathology is in the cerebellum, Brain 1995;118:1087±1093. Brown P, Marsden CD. Rhythmic cortical and muscle discharge in cortical myoclonus. Brain 1996;119:1307±1316. Brown P, Ridding MC, Werhahn KJ, Rothwell JC, Marsden CD. Abnormalities of the balance between inhibition and excitation in the motor cortex of patients with cortical myoclonus. Brain 1996;119:309±317. Chen R, Samii A, Canos M, Wasserman EM, Hallett M. Effects of phenytoin on cortical excitability in humans. Neurology 1997;49:881±883. Cowan JMA, Rothwell JC, Wise RJS, Marsden CD. Electrophysiological and positron emission studies in a patient with cortical myoclonus, epilpsia partialis continua and motor epilepsy. J Neurol Neurosurg Psychiatry 1986;49:796±807. Di Lazzaro V, Restuccia D, Oliviero A, Pro®ce P, Ferrara L, Insola A, Mazzone P, Tonali P, Rothwell JC. Magnetic transcranial stimulation
1530
P. Ashby et al. / Clinical Neurophysiology 110 (1999) 1521±1530
at intensities below active motor threshold activates intracortical inhibitory circuits. Exp Brain Res 1998;119:265±268. Hallett M, Chadwick D, Marsden CD. Cortical re¯ex myoclonus. Neurology 1979;29:1107±1125. Kanthasamy AG, Matsumoto RR, Truong DD. Animal model of myoclonus. Clin Neurosci 1996;3:236±245. Kugelberg E, Widen L. Epilepsia Partialis Continua. Electroenceph clin Neurophysiol 1954;6:503±506. Kujirai T, Caramia MD, Rothwell JC, Day BL, Thompson PD, Ferberts S, Wroe S, Asselman P, Marsden CD. Corticospinal inhibition in human motor cortex. J Physiol 1993;471:501±519. Mima T, Nagamine T, Nishitani N, Mikuni N, Ikeda A, Fukuyama H,
Takigawa T, Kimura J, Shibasaki H. Cortical myoclonus. Sensorimotor hyperexcitability. Neurology 1998;50:933±942. Obeso JA, Rothwell JC, Marsden CD. The spectrum of cortical myoclonus. Brain 1985;108:193±224. Rothwell JC, Obeso JA, Marsden CD. Electrophysiology of somatosensory re¯ex myoclonus. Adv Neurol 1986;43:385±398. Uesaka Y, Terao Y, Ugawa Y, Yumoto M, Hanajima R, Kanazawa I. Magnetoencephalographic analysis of cortical myoclonic jerks. Electroenceph clin Neurophysiol 1996;99:141±148. Ziemann U, Lonnecker S, Steinhoff BJ, Paulus W. Effects of antiepileptic drugs on motor cortex excitability in humans: a transcranial magnetic stimulation study. Ann Neurol 1996;40:367±378.