Electroencephalography and clinical Neurophysiology, 83 (1992) 192-200 © 1992 Elsevier Scientific Publishers Ireland, Ltd. 0013-4649/92/$05.l)0
192
EEG 91663
Cortical s o m a t o s e n s o r y m a g n e t i c r e s p o n s e s in m u l t i p l e sclerosis * Jari Karhu, Riitta Hari, Jyrki P. MS.kel~i a, Juha Huttunen a and Jukka Knuutila Low Temperature Laboratory, Helsinki Unicersity of Technology, 02150 Espoo (Finland), and ~ Department of Neurology, Helsinki Unicersity Central Hospital, 00290 Helsinki (Finland) (Accepted for publication: 10 April 1992)
Summary Somatosensory evoked magnetic fields (SEFs) to contralateral median and ulnar nerve stimulation were analyzed in 10 patients with multiple sclerosis and in 8 healthy controls. SEFs were recorded with a 24-channel SQUID gradiometer over both hemispheres. Seven patientsshowed abnormally large-amplitude SEF deflections at 60-80 msec; 5 of them had multiple lesions around lateral ventricles in magnetic resonance imaging. In 2 patients with plaques at the level of 3rd and 4th ventricles and medulla, the 30 msec responses were enlarged. The equivalent sources of 20 msec and 30-80 msec responses were in the primary hand sensorimotor cortex both in patients and in control subjects. The results suggest that early and middle-latency SEFs reflect parallel processing of somatosensory input. Recording of middle-latency evoked responses, electric or m~ignetic, may give additional information about the somatosensory function in multiple sclerosis. Key words: Multiple sclerosis; Evoked responses; Magnetoencephalography; Median nerve; Ulnar nerve; Primary somatosensory cortex
The diagnosis of multiple sclerosis (MS) is based primarily on fluctuating clinical findings suggesting disseminated lesions of the white matter at several levels of the central nervous system. The diagnosis is supported by cerebrospinal fluid analysis and increasingly by magnetic resonance imaging (MRI) (Young et al. 1981). Visual, auditory and somatosensory evoked potentials (SEPs) are used routinely to assess the integrity of sensory pathways (Namerow 1968; Halliday et al. 1973; Chiappa 1980), and central motor pathways can be evaluated by magnetic stimulation (Mills and Murray 1985). Combined use of these tests often enables the early diagnosis of MS (Poser et al. 1983). The clinical usefulness of median nerve SEPs in diagnostics of MS has been examined exhaustively, with most studies focusing on deflections within the first 25 msec after stimulus onset (for example, Mastaglia et al. 1977; Chiappa 1980). Both latency increase and amplitude reduction have been observed already in the P14 deflection (Garcia Larrea and Mau-
Correspondence to: J. Karhu, Low Temperature Laboratory, Helsinki University of Technology, 02150 Espoo (Finland). Fax: 358-0-4512969. * This study was financially supported by the Academy of Finland and by the Sigrid Jus61ius Foundation. A preliminary report of this study has been presented in abstract form (Karhu et al. 1991).
gui~re 1988). A delayed or absent parietal 20 msec deflection (N20) has been found in 40-90% of patients. N20 probably represents cortical synaptic events triggered by the primary input volley, whereas the later deflections may originate through both cortico-cortical connections and separate thalamo-cortical projections (Desmedt and Cheron 1980; Yamada 1988). The considerable inter- and intrasubject variability and the uncertainty of the exact origin have reduced the clinical usefulness of later SEP deflections, even though some MS patients show abnormalities limited to them (Yamada et al. 1982). Magnetoencephalography (MEG) allows selective studies of each hemisphere; activity on the opposite side does not interfere with the interpretation of the data (for reviews, see Williamson and Kaufman 1981; Hari and Lounasmaa 1989). The equivalent sources of somatosensory evoked magnetic fields (SEFs), peaking 20-150 msec after median and ulnar nerve stimulation, are located in the hand sensorimotor cortex (Hari et al. 1984; Wood et al. 1985; Huttunen et al. 1987; Tiihonen et al. 1989; Suk et al. 1991); some magnetic deflections around 100 msec emerge from the second somatosensory cortex SII (Hari et al. 1984, 1990). We report here 24-channel SEF recordings of 10 patients with clinically definite MS. Their responses to median and ulnar nerve stimulation differed significantly from those of the control group at both early (20 msec) and middle latencies (30-80 msec). The SEF results were also compared with MRI findings.
SEFs IN MULTIPLE SCLEROSIS
193
nerves at the wrist through transcutaneous electrodes. The stimuli elicited visible muscle twitches in the small hand muscles. Each hand was stimulated separately on the side contralateral to recording.
Methods
Subjects Ten patients (2 females, 8 males; mean age 34 years, range 22-56 years; mean length 179 cm, range 163-187 cm) with recently diagnosed multiple sclerosis and 8 healthy control subjects (2 females, 6 males; mean age 32 years, range 22-49 years; mean length 179 cm, range 155-193 cm) were studied. All patients had symptoms and signs of lesion from at least two levels of the central nervous system (see Table I) and an increased cerebro-spinal fluid/serum IgG ratio (in patient 3, the IgG ratio was not available). The 20 msec deflections of the median nerve SEPs were abnormal in 5 out of the 8 patients studied in the hospital. The 0.02 T MRI showed bilateral periventricular lesions in 5 patients (subgroup MRI + in Tables II and III); in patient 6 a plaque was observed in the midbrain. Three subjects had normal 0.02 T MRI. In one of them, 1.5 T MRI illustrated a plaque at the 4th ventricle level and two medullary plaques. In one patient, MRI was not available; X-ray computed tomography was normal. The subjects gave their informed consent for the measurements and the experimental protocol was accepted by the Ethical Committee of the Helsinki University Central Hospital.
Recordings The recordings were made in a magnetically shielded room. During the measurements the subject was lying on a wooden bed with eyes open and the head stabilized with a vacuum cast. Control experiments were made in two healthy subjects to evaluate the effect of attention and vigilance, which affect the middle-latency SEPs (Desmedt and Tomberg 1989). The same stimuli were delivered during 4 behavioral conditions: (1) subject was reading, (2) subject was instructed to count the number of weaker median nerve stimuli in the set, although such were not delivered, (3) subject was counting the ulnar nerve stimuli, (4) subject was counting the number of a-vowels in Finnish songs that were played to him through an earphone. Before recording, about 100 electric pulses were delivered to both nerves to familiarize the subject with the stimulation. SEFs to contralateral stimulation were measured separately over the left and right hemispheres. The 24-channel SQUID gradiometer (Ahonen et al. 1991) has planar flux transformers: the tangential derivatives, 0Br/0x and 0 B r / 0 y of B r, the magnetic field perpendicular to the skull, are recorded with two orthogonal figure-of-8 loops simultaneously at 12 locations, 3 cm apart (see Fig. 1). The planar gradiometer records the maximum signal just above a current dipole, and its
Stimuli Constant-current pulses of 0.3 msec duration (Grass $88 stimulator, Grass SIU 4678 isolation unit and Grass CCU 1A constant current unit) were delivered once every 505 msec alternately to median and ulnar TABLE I
Main symptoms, MRI lesions, IgG indexes, and median nerve SEF findings (N20m and P60m) in all patients. In symptoms R and L refer to the right and left sides of the body, in SEFs to the right and left hemisphere. Normal range for IgG index is 0.34-0.60, for SEF latencies and amplitudes control group mean + 2.5 S.D.). Pt 1 2 3 4 5 6 7 8 9 10
Main symptoms
MRI lesions
IgG index
N20m
P60m
L leg weakness Lhermitte + R leg paresthesia L leg weakness L mild hemiparesis Bilateral leg stiffness R leg paresis Urge incontinence Double vision L hand clumsiness R leg paresis L hand paresis Optic neuritis L mild hemiparesis L hand clumsiness R lowered vision L hand clumsiness R lowered vision R tongue paresthesia R hand clumsiness
bilateral periventricular bilateral periventricular bilateral periventricular not available (CT normal) bilateral periventricular at 3rd ventricle level
1.1
L delayed R delayed L delayed and suppressed R delayed and suppressed L delayed R normal L delayed and suppressed R delayed and suppressed L normal R normal L normal R suppressed L delayed and suppressed R suppressed L normal R normal L normal R suppressed L normal R normal
L enlarged R enlarged L enlarged R enlarged L normal R enlarged L enlarged R enlarged L enlarged R normal L normal R absent L normal R absent L absent R normal L absent R normal L normal R normal
1.3 not available 1.4 3.7 1.2
normal (low quality)
0.8
bilateral periventricular normal
0.7
one at 4th ventr, level two medullary
0.6
1.7
194 J. K A R H U El" AL
PATIENT
33 ms I
,~-~.-.
8v~~.~
9
~
, o ~
39 ms
63 ms
E 100 fT/cm
I74 6t
~
/
m 20 ms
1 2}-,,,
891q
\
CONTROL ~ 1 --,.f'~.-'~
Y L 2 ~
20 ms 45 ms 59 ms 89 ms
~
7 ~
]8~.~~ 30 ms
9 ~
1
[ 100 fT/cm 11 ~,~/"'"~,,P'~. 1 I"-1
20 ms
2
~
0
~
f
SEFs IN M U L T I P L E SCLEROSIS
195
PATIENT
CONTROL
Fig. 2. Isocontour maps calculated from the 24-channel recordings of Fig. 1. Shadowed areas indicate magnetic flux out of the head, white areas flux into the head. The isocontour lines are separated by 30 fT, except in the 63 msec m a p of the patient, marked by " s t e p x 2 " (60 fT). The locations and orientations of equivalent dipoles are indicated by arrows on the maps. The approximate area covered by the maps is shown in the insert.
coverage (O = 12.5 cm) allowed the source areas of SEFs to be determined with one positioning of the instrument. Measuring the SEFs to ulnar and median nerve stimulation from both hemispheres took about 1 h, which we consider acceptable for clinical studies. The exact locations and orientations of the sensors with respect to the head were determined by analyzing the magnetic signals resulting from currents in 3 small coils fixed on known locations on the scalp (Knuutila et al. 1987).
Signal analysis The recording passband was 0.05-500 Hz (3 dB points, high-pass roll-off 35 d B / d e c a d e and low-pass over 80 d B / d e c a d e ) . The signals were digitized at 2 kHz and about 100 responses per nerve were averaged on-line at each location. The first two responses of the stimulus block were rejected from the analysis. Each
measurement was repeated once to see the replicability of the responses. The analysis period was 140 msec, and response amplitudes were measured with respect to a 20 msec prestimulus baseline. The channel pair with the maximum total signal (~/(OBr/0x)2+ (0Br/Oy) 2) in each hemisphere was chosen for latency and amplitude measurements. To obtain isocontour maps from the gradient measurements, minimum-norm current estimates (MNE) corresponding to the measured data were first computed. Thereafter, B r generated by the MNE was determined (Ilmoniemi 1991). This procedure accomplishes a 3-dimensional interpolation suitable for the magnetic field. The deviations of the sensitivity axes of the gradiometer from the radial direction are also taken into account. Equivalent current dipoles (ECDs) were determined
Fig. 1. Magnetic responses of patient 2 and control subject 3 to stimulation of the left median nerve at the wrist. The responses were recorded over the right hemisphere with the 24-channel gradiometer. The approximate locations of the channels are shown in the insert. The upper traces of each response pair show the field gradient in the vertical (y) and the lower in the horizontal (x) direction. The positions of the instrument in the patient and in the control subject were within 1 cm from each other with respect to the external landmarks of the skull. Two records (N = 100) are superimposed. Responses were digitally low-pass filtered at 140 Hz.
196
J. K A R H U ET AL.
TABLE II Mean ± S.E.M. latencies and amplitudes for main deflections of control group, patient group and subgroup of patients with periventicular lesions in the MRI (MRI + ). N indicates the number of analyzed hemispheres. Statistical significances of differences between control and patient groups are marked by asterisks: * < 0.05, ** < 0.01 and *** < 0.001 (2-tailed t test for group differences). Latency (msec)
Amplitude ( f T / c m )
Controls
Patients
Patients/MRI +
Controls
20.7_+0.3 N = 16 33.8 ± 0.9 N=I6 56.6 ± 1.4 N = 13
23.2±0.9 * N = 20 37.4+ 1.2 * N=I7 57.3 ± 1.7 N = 16
23.6± 1.1 N = 10 37.5+ 1.8 ** N=8 57.4 ± 2.0 N = 10
73.0_+6.9
21.5±0.4 N = 16 34.6± 1.0 N = 16 57.6± 1.6 N = 14
24.4± 1.2 * N = 19 36.4_+ 1.4 N = 17 62.1 ± 1.5 N = 16
25.2± 1.7 * N = 10 37.9± 2.0 N = 10 61.7± 2.1 N = 10
Patients
Patients/MR1 +
M e d i a n nerce
N20m P30m P60m
55.5± 8.8
62.7± 7.6
114.0± 16.0 **
73.1 _+6.9
140.9 ± 19.9 **
51.7± 8.1 80.6± 16.1 163.5 _+28.7 ** *
Ulnar nert~e
N20m P30m P60m
with a least-squares search from the original data, applying the measured locations and orientations of the sensors. In the analysis, a spherical conductor model was used with a radius of 10 cm, corresponding to the local radius of curvature of the head in the measurement area. Only ECDs which accounted for over 80% of the measured field variance were included in further analysis.
59.3±7.2
41.5_+ 5.4
48.5+ 8.1
44.7_+5.2
72.4_+9.8 *
61.5_+ 10.8
62.5 ±7.6
111.0-+ 13.9 **
116.9± 11.7 ***
Results
Fig. 1 shows median nerve SEFs of a patient and a control subject over the right hemisphere. SEFs of the control subject contain 5 main deflections: N20m, P30m, N45m, P60m and P85m peaking at 20, 30, 45, 59 and 89 msec, respectively. In the patient, N20m is suppressed and delayed (latency 33 msec), and P60m at
TABLE III Mean _+S.E.M. strengths (Q) and depths of ECDs at the time of main deflections of control group, patient group, and patient group with periventricular lesions in the MRI (MRI +). N indicates the number of analyzed hemispheres. Dipoles with goodness-of-fit values over 80% are included in the analysis. Statistical significances of differences between control and patient groups are marked by asterisks: * < 0.05, ** < 0.01 and ***< 0.001 (2-tailed t test). Q (nAm)
Depth (mm)
Controls
Patients
Patients/MRI +
Controls
Patients
Patients/MRI +
18.4 ± 1.5 N=16 22.5 ± 3.2 N=13 23.6 ± 3.2 N=13
19.0 _+2.9 N=12 33.4 ± 3.5 * N=17 41.5 ± 4.6 ** N=14
21.0 ± 3.5 N=8 29.5 ± 5.1 N=8 47.1 ± 3.6 *** N=8
30.1 ± 1.6
28. l _+ 1.7
31.3 ± 2.1
27.3 ± 1.8
30.1 _+ 1.3
31.7 ± 2.6
28.5 _+2.4
31.0 ± 2.2
30.7 ± 2.7
13.1 + 1.4 N = 15 15.5 ± 1.6 N=16 21.8 ± 1.9 N=14
13.5+2.1 N = 10 20.3 _+2.4 N=17 34.5 _+5.3 * N=16
15.9_+3.0 N= 6 17.7 ± 3.3 N=9 36.6 ± 7.6 * N=8
25.1 ± 1.7
29.4_+2.1
31.0+_3.1
27.3 _+ 1.6
26.4 _+2.2
25.9 _+3.2
28.4 ± 1.6
29.2 ± 2.7
29.1 ± 4.5
M e d i a n nerce
N20m P30m P60m
Ulnar nert~e
N20m P30m P60m
----}
Fig. 3. Responses to contralateral median and ulnar nerve stimulations over both hemispheres in patients 1-10 and controls 1-8. Two records (N = 100) are superimposed. The approximate measurement location is shown in the insert.
197
SEFs IN MULTIPLE SCLEROSIS PATIENTS median nerve
CONTROLS ulnar nerve
median nerve
ulnar nerve
t
3 °
~,c 5
6
6/
+
9
0 20 ms
,,,,,4
0 20 ms
I00 fT/cm
20 ms
0 20 ms CONTROLS
PATIENTS median nerve
median nerve
ulnar nerve
ulnar nerve
& L
L ,
,
6
7
8~ 0 20 ms
100 fT/cm
o 20 ms
0 20 ms
0 20 ms
198
63 msec is markedly enlarged; N45m and P85m are not seen. Fig. 2 shows magnetic field patterns to median nerve stimulation during the main deflections of Fig. 1. In the control subject, the field patterns are dipolar and ECDs explain 88-97% of them. The single current dipole is thus considered to model adequately the underlying brain currents. In the patient, no dipolar pattern is found at latencies 18-35 msec, and the field at 63 msec (P60m) is conspicuously stronger than that of the control. ECDs account for 89% of the measured field at 39 msec (P30m) and for 96% at 63 msec. The locations and orientations of ECDs suggest the underlying sources to be in the hand area of the primary sensorimotor cortex both in the patient and in the control subject. Fig. 3 shows examples of median and ulnar nerve SEFs for all patients and control subjects over both hemispheres. In patients 1, 2, 3, 4 and 7 N20m was delayed for both median and ulnar nerve stimulations (Table I). The delay of N20m was significant in the patient group compared with the controls ( P < 0.05, 2-tailed t test for group differences, cf., Table lI). If the field pattern at latencies 18-35 msec was not explained satisfactorily with an ECD, N20m was considered to be suppressed. This happened in patients 2, 4, 6, 7 and 9 for median nerve and in patients 5, 8 and 10 for ulnar nerve stimulations. Therefore, the current dipole model was unsatisfactory during N20m for 18 out of 40 nerves, and at least one abnormal N20m was found in each patient. In control subjects, N45m was detected to stimulation of both nerves in 15/16 of examined hemispheres. Its mean + S.E.M. latency was 46.9 + 0.9 msec to median and 44.4 _+ 1.0 msec to ulnar nerve stimulation. In patients, N45m was seen bilaterally only in patient 10. P60m was enhanced, i.e., exceeded the mean amplitude of controls by more than 2.5 S.D. (see Table II), for both nerves in patients 1-5 and for ulnar nerve in patient 6. Altogether, P60m was enhanced in 11 out of 20 hemispheres studied. In 3 of the 4 subjects with large bilateral P60m, MR1 showed multiple bilateral plaques around lateral ventricles (patients 1, 2 and 3); in patient 4, MRI was not available. In patients 5 and 8, who also had bilateral periventricular lesions, P60m was enlarged unilaterally. The mean amplitude of P60m was nearly 2-fold in the patient group compared with the controls ( P < 0.01) and over 2-fold in the MRI + subgroup ( P < 0.001). Analysis of equivalent current dipoles gave corresponding results (see Table liD. Both patients with plaques below or at the 3rd ventricle level (patients 6 and 10) had bilaterally enlarged P30m. P85m was detected in 10 out of 20 of examined hemispheres of patients and in 14 out of 16 of controls, with significantly larger amplitude in patients than in controls (mean amplitudes differed at P < 0.01 for
J. K A R H U ET AL.
ULNAR NERVE
MEDIANNERVE • patient 0 controlsubject
Yl
Y
2
N20m
-2 N20m
2 cm
I
r
o 2 cm
X
~o ° -
O
Q
•
Y
N > 20 nAm
Y 2
-2
P60m I
2 cm
o--~---I~ x e•Ol
P60m t ° ' ~ o ~
2 cm ~x
Fig. 4. Source locations of N20m and P60m to median and ulnar nerve stimulations in control subjects (open circles) and patients (filled circles) in the left hemisphere. Mean source locations (larger circles) and m e a n equivalent current dipole orientations and strengths are also shown. The coordinate system is demonstrated in the center: the y-axis connects vertex and ear canal and the origin is 50 m m from vertex towards the ear canal.
median nerve, and at P < 0.05 for ulnar nerve). There was no further amplitude enhancement in the MRI + group. The P85m latencies did not differ significantly between patients and controls (median nerve: controls 85.1 + 2.2 msec, patients 85.5 + 2.0 msec; ulnar nerve: controls 83.1 + 2.1 msec, patients 86.2 _+_1.3 msec). The middle-latency SEFs around 60 msec were not altered by attention in control experiments described in Methods. Fig. 4 shows the locations of ECDs for N20m and P60m to median and ulnar nerve stimulation in all subjects. The mean source locations did not differ significantly between patients and controls. The ECDs were on average 30 mm beneath the scalp (see Table III). All sources, also those of P30m, N45m and P85m, were in an area of 4 cm × 4 cm with respect to the external landmarks of the skull.
Discussion
Focal demyelination can decrease conduction velocity, abolish responses to high-frequency impulse trains, or cause a total conduction block (McDonald 1963). These abnormalities are limited to the demyelinated
SEFs IN MULTIPLE SCLEROSIS
zone whereas the other regions of the nerve fiber conduct in a relatively normal manner. However, loss of synchrony between impulses in different fibers can cause temporal dispersion of the compound response of the nerve (for a review, see Waxman 1988), which affects evoked cortical responses. Thus, in addition to prolonged latencies, changes in SEP and SEF wave forms are to be expected in MS. The latencies of N20m deflections agreed in general with earlier clinical median nerve SEP recordings in our MS patients, even though successive evoked potential recordings can vary considerably despite the stable clinical deficit (Aminoff et al. 1984). Most SEP abnormalities are obtained from limbs showing sensory impairment, but there are reports of delayed or abolished SEPs from limbs with no sensory deficit in the area of nerve tested (Namerow 1968). In the present study, no obvious relationship was found between the clinical signs in upper extremities and the lateralization of SEF findings. The enhancement of P60m was unexpected: it was evident in all patients with bilateral periventricular MRI lesions (see Fig. 3). This might reflect relative enhancement of P60m due to disappearance of N45m, which is superimposed on the P60m slope. Another possibility is that increased temporal dispersion of the early input volleys to the cortex decreases later inhibition in sensorimotor cortex through, as yet, unknown mechanisms. SEPs may be recorded over the scalp when sensory nerve action potentials are absent, presumably because responses arising f r o m a few normally conducting axons are synchronized and amplified in the central somatosensory pathways (Parry and Aminoff 1987). Central amplification seems to be impaired in patients with MS (Eisen et al. 1982). An increased amplification is thus an unlikely explanation of the enlarged middlelatency SEFs, especially because the early cortical components are often simultaneously attenuated. Delay of N20m did not necessarily result in delay of P60m, and patients 3 and 5 with periventricular MRI findings had enhanced P60m, even though their N20m was normal. Therefore, in addition to sequential cortico-cortical connections (Allison et al. 1989; Garraghty et al. 1990), also other afferent inputs besides those responsible of N20m contribute to later deflections, possibly via separate thalamo-cortical projections. Evidently early and middle-latency SEP and SEF deflections reflect at least partially parallel processing of somatosensory information and could both be evaluated in MS patients. Correspondingly, separate components of visual evoked potentials can be independently affected in multiple sclerosis (Ghilardi et al. 1991). MS lesions visible in MRI are often located in the periventricular areas, at or close to the somatosensory thalamo-cortical radiation (Turano et al. 1991). There-
199
fore, alteration of middle-latency somatosensory responses should be relatively frequent. That such changes have not been reported in SEPs is probably not purely due to differences between the L E G and MEG. Clinical SEPs are often measured with a highpass filter at 10-20 Hz, which attenuates middle-latency deflections. Furthermore, the analysis periods are often shorter than 50 msec. In fact, enlarged 60 msec responses have been observed in wide-passband SEP and SEF recordings of MS patients, as well as in one patient with a tumor in the periventricular area (A. Asgan, R. Bain and Y. Yang, Glasgow, personal communication). According to our results, middle-latency SEFs are altered in MS, especially in patients with lesions around the lateral ventricles. These responses seem to originate in the primary sensorimotor cortex, and evaluation of both them and their electric counterparts may give additional information about somatosensory function in MS. Prof. O.V. Lounasmaa and Dr. J. Partanen made valuable comments on the manuscript.
References Ahonen, A., H~im~il~iinen, M., Kajola, M., Knuutila, J., Lounasmaa, O., Simola, J., Tesche, C. and Vilkman, V. Multichannel SQUID systems for brain research. IEEE Trans. Magn., 1991, 27: 27862792. Allison, T., McCarthy, G., Wood, C., Darcey, T., Spencer, D. and Williamson, P. Human cortical potentials evoked by stimulation of the median nerve. I. Cytoarchitectonic areas generating shortlatency activity. J. Neurophysiol., 1989, 62: 694-710. Aminoff, M.J., Davis, S.L. and Panitch, H.S. Serial evoked potential studies in patients with definite multiple sclerosis. Clinical relevance. Arch. Neurol., 1984, 41: 1197-1202. Chiappa, K.H. Pattern shift visual, brainstem auditory and short latency somatosensory evoked potentials in multiple sclerosis. Neurology, 1980, 30: 110-123. Desmedt, J. and Cheron, G. Somatosensory evoked potential to finger stimulation in healthy octogenarians and in young adults: wave forms, scalp topography and transit times of parietal and frontal connections. Electroenceph. clin. Neurophysiol., 1980, 50: 404-425. Desmedt, J. and Tomberg, C. Mapping early somatosensory evoked potentials in selective attention: critical evaluation of control conditions used for titrating by difference the cognitive P30, P40, P100 and N140. Electroenceph. clin. Neurophysiol., 1989, 74: 321-346. Eisen, A., Purves, S. and Hoirch, R.T. Central nervous system amplification: its potential in the diagnosis of early multiple sclerosis. Neurology, 1982, 32: 359-364. Garcla Larrea, L. and Maugui~re, F. Latency and amplitude abnormalities of the scalp far-field P14 to median nerve stimulation in multiple sclerosis. A SEP study of 122 patients recorded with a non-cephalic reference montage. Electroenceph. clin. Neurophysiol., 1988, 71: 180-186. Garraghty, P., Florence, S. and Kaas, J. Ablations of areas 3a and 3b of monkey somatosensory cortex abolish cutaneous responsivity in area 1. Brain Res., 1990, 528: 165-169.
200 Ghilardi, M.F., Sartucci, F., Brannan, J.R., Onofrj, M.C., BodisWollner, I., Mylin, L. and Stroch, R. N70 and P100 can be independently affected in multiple sclerosis. Electroenceph. clin. Neurophysiol., 1991, 80: 1-7. Halliday, A.M., McDonald, W.I. and Mushin, J. The visual evoked response in the diagnosis of multiple sclerosis. Br. Med. J., 1973, 4: 661-664. Hari, R. and Lounasmaa, O.V. Recording and interpretation of cerebral magnetic fields. Science, 1989, 244: 432-436. Hari, R., Reinikainen, K., Kaukoranta, E., H~im~il~iinen, M., I1moniemi, R., Penttinen, A., Salminen, J. and Teszner, D. Somatosensory evoked cerebral magnetic fields from SI and SII in man. Electroenceph. clin, Neurophysiol., 1984, 57: 254-263. Hari, R., H~im~il~iinen,M., H~im~il~iinen,H., Kekoni, J., Sams, M. and Tiihonen, J. Separate finger representations at the human second somatosensory cortex. Neuroscience, 1990, 37: 245-249. Huttunen, J., Hari, R. and Leinonen, L. Cerebral magnetic responses to stimulation of ulnar and median nerves. Electroenceph. clin. Neurophysiol., 1987, 66: 391-400. llmoniemi, R.J. Estimates of neuronal current distributions. Acta Oto-Laryngol. (Stockh.), 1991, Suppl. 491: 80-87. Karhu, J., Hari, R., M~ikel~i,J.P., Huttunen, J. and Knuutila, J. MEG findings in multiple sclerosis. In: 8th International Conference on Biomagnetism, Book of Abstracts, 1991: 251-252. Knuutila, J., Ahlfors, S., Ahonen, A., H~illstr6m, J., Kajola, M., Lounasmaa, O.V., Tesche, C. and Vilkman, V. A large-area low-noise seven-channel DC SQUID magnetometer for brain research. Rev. Sci. Instrum., 1987, 58: 2145-2156. Mastaglia, F.L., Black, J.L., Cala, L.A. and Collins, D.W.K. Evoked potentials, saccadic velocities and computerized tomography in diagnosis of multiple sclerosis. Br. Med. J., 1977, 1: 1315-1317. McDonald, W.I. The effects of experimental demyelination on conduction in peripheral nerve: a histological and electrophysiological study. II. Electrophysiological observations. Brain, 1963, 86: 501-524. Mills, K.R. and Murray, N.M.F. Corticospinal tract conduction time in multiple sclerosis. Ann. Neurol., 1985, 18: 601-605. Namerow, N.S. Somatosensory evoked responses in multiple sclerosis with varying sensory loss. Neurology, 1968, 18: 1197-1204.
J. KARHU ET AL. Parry, G.J. and Aminoff, M.J. Somatosensory evoked potentials in chronic acquired demyelinating peripheral neuropathy. Neurology, 1987, 37: 313-316. Poser, C.M., Paty, D.W., Schernberg, L., McDonald, W.J., Davis, F.A., Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberherg, D.H. and Tourtellotte, W.W. New diagnostic criteria for multiple sclerosis - guidelines for research protocols. Ann. Neurol., 1983, 13: 227-231. Suk, J., Ribary, U., Cappell, J., Yamamoto, T. and Llinfis, R. Anatomical localization revealed by MEG recordings of the human somatosensory system. Electroenceph. clin. Neurophysiol., 1991, 78: 185-196. Tiihonen, J., Hari, R. and H~im~il~iinen, M. Early deflections of cerebral magnetic responses to median nerve stimulation. Electroenceph, clin. Neurophysiol., 1989, 74: 290-296. Turano, G., Jones, S.J., Miller, D.H., Du Boulay, G.H., Kakigi, R. and McDonald, W.I. Correlation of SEP abnormalities with brain and cervical cord MRI in multiple sclerosis. Brain, 1991~ 114: 663-681. Waxman, S.G. Clinical course and electrophysiology of multiple sclerosis. In: S.G. Waxman (Ed.), Advances in Neurology, Vol. 47: Functional Recovery in Neurological Disease. Raven Press, New York, 1988: 157-184. Williamson, S.J. and Kaufman, L. Biomagnetism. J. Magn. Magn. Mat., 1981, 22: 129-202. Wood, C.C., Cohen, D., Cuffin, B.N., Yarita, M. and Allison, T. Electric sources in the human somatosensory cortex: identification by combined magnetic and potential field recordings. Science, 1985, 227: 1051-1053. Yamada, T. The anatomic and physiologic bases of median nerve somatosensory evoked potentials. Neurol. Clin., 1988, 6: 705-733. Yamada, T., Shivapour, E., Wilkinson, J.T. and Kimura, J. Shortand long-latency somatosensory evoked potentials in multiple sclerosis. Arch. Neurol., 1982, 39: 88-94. Young, I.R., Hall, A.S., Pallis, C.A., Legg, N.J., Byder, G.M. and Steiner, R.E. Nuclear magnetic resonance imaging of the brain in multiple sclerosis. Lancet, 1981, ii: 1063-1066.
192
EEG 91663
Cortical s o m a t o s e n s o r y m a g n e t i c r e s p o n s e s in m u l t i p l e sclerosis * Jari Karhu, Riitta Hari, Jyrki P. MS.kel~i a, Juha Huttunen a and Jukka Knuutila Low Temperature Laboratory, Helsinki Unicersity of Technology, 02150 Espoo (Finland), and ~ Department of Neurology, Helsinki Unicersity Central Hospital, 00290 Helsinki (Finland) (Accepted for publication: 10 April 1992)
Summary Somatosensory evoked magnetic fields (SEFs) to contralateral median and ulnar nerve stimulation were analyzed in 10 patients with multiple sclerosis and in 8 healthy controls. SEFs were recorded with a 24-channel SQUID gradiometer over both hemispheres. Seven patientsshowed abnormally large-amplitude SEF deflections at 60-80 msec; 5 of them had multiple lesions around lateral ventricles in magnetic resonance imaging. In 2 patients with plaques at the level of 3rd and 4th ventricles and medulla, the 30 msec responses were enlarged. The equivalent sources of 20 msec and 30-80 msec responses were in the primary hand sensorimotor cortex both in patients and in control subjects. The results suggest that early and middle-latency SEFs reflect parallel processing of somatosensory input. Recording of middle-latency evoked responses, electric or m~ignetic, may give additional information about the somatosensory function in multiple sclerosis. Key words: Multiple sclerosis; Evoked responses; Magnetoencephalography; Median nerve; Ulnar nerve; Primary somatosensory cortex
The diagnosis of multiple sclerosis (MS) is based primarily on fluctuating clinical findings suggesting disseminated lesions of the white matter at several levels of the central nervous system. The diagnosis is supported by cerebrospinal fluid analysis and increasingly by magnetic resonance imaging (MRI) (Young et al. 1981). Visual, auditory and somatosensory evoked potentials (SEPs) are used routinely to assess the integrity of sensory pathways (Namerow 1968; Halliday et al. 1973; Chiappa 1980), and central motor pathways can be evaluated by magnetic stimulation (Mills and Murray 1985). Combined use of these tests often enables the early diagnosis of MS (Poser et al. 1983). The clinical usefulness of median nerve SEPs in diagnostics of MS has been examined exhaustively, with most studies focusing on deflections within the first 25 msec after stimulus onset (for example, Mastaglia et al. 1977; Chiappa 1980). Both latency increase and amplitude reduction have been observed already in the P14 deflection (Garcia Larrea and Mau-
Correspondence to: J. Karhu, Low Temperature Laboratory, Helsinki University of Technology, 02150 Espoo (Finland). Fax: 358-0-4512969. * This study was financially supported by the Academy of Finland and by the Sigrid Jus61ius Foundation. A preliminary report of this study has been presented in abstract form (Karhu et al. 1991).
gui~re 1988). A delayed or absent parietal 20 msec deflection (N20) has been found in 40-90% of patients. N20 probably represents cortical synaptic events triggered by the primary input volley, whereas the later deflections may originate through both cortico-cortical connections and separate thalamo-cortical projections (Desmedt and Cheron 1980; Yamada 1988). The considerable inter- and intrasubject variability and the uncertainty of the exact origin have reduced the clinical usefulness of later SEP deflections, even though some MS patients show abnormalities limited to them (Yamada et al. 1982). Magnetoencephalography (MEG) allows selective studies of each hemisphere; activity on the opposite side does not interfere with the interpretation of the data (for reviews, see Williamson and Kaufman 1981; Hari and Lounasmaa 1989). The equivalent sources of somatosensory evoked magnetic fields (SEFs), peaking 20-150 msec after median and ulnar nerve stimulation, are located in the hand sensorimotor cortex (Hari et al. 1984; Wood et al. 1985; Huttunen et al. 1987; Tiihonen et al. 1989; Suk et al. 1991); some magnetic deflections around 100 msec emerge from the second somatosensory cortex SII (Hari et al. 1984, 1990). We report here 24-channel SEF recordings of 10 patients with clinically definite MS. Their responses to median and ulnar nerve stimulation differed significantly from those of the control group at both early (20 msec) and middle latencies (30-80 msec). The SEF results were also compared with MRI findings.
SEFs IN MULTIPLE SCLEROSIS
193
nerves at the wrist through transcutaneous electrodes. The stimuli elicited visible muscle twitches in the small hand muscles. Each hand was stimulated separately on the side contralateral to recording.
Methods
Subjects Ten patients (2 females, 8 males; mean age 34 years, range 22-56 years; mean length 179 cm, range 163-187 cm) with recently diagnosed multiple sclerosis and 8 healthy control subjects (2 females, 6 males; mean age 32 years, range 22-49 years; mean length 179 cm, range 155-193 cm) were studied. All patients had symptoms and signs of lesion from at least two levels of the central nervous system (see Table I) and an increased cerebro-spinal fluid/serum IgG ratio (in patient 3, the IgG ratio was not available). The 20 msec deflections of the median nerve SEPs were abnormal in 5 out of the 8 patients studied in the hospital. The 0.02 T MRI showed bilateral periventricular lesions in 5 patients (subgroup MRI + in Tables II and III); in patient 6 a plaque was observed in the midbrain. Three subjects had normal 0.02 T MRI. In one of them, 1.5 T MRI illustrated a plaque at the 4th ventricle level and two medullary plaques. In one patient, MRI was not available; X-ray computed tomography was normal. The subjects gave their informed consent for the measurements and the experimental protocol was accepted by the Ethical Committee of the Helsinki University Central Hospital.
Recordings The recordings were made in a magnetically shielded room. During the measurements the subject was lying on a wooden bed with eyes open and the head stabilized with a vacuum cast. Control experiments were made in two healthy subjects to evaluate the effect of attention and vigilance, which affect the middle-latency SEPs (Desmedt and Tomberg 1989). The same stimuli were delivered during 4 behavioral conditions: (1) subject was reading, (2) subject was instructed to count the number of weaker median nerve stimuli in the set, although such were not delivered, (3) subject was counting the ulnar nerve stimuli, (4) subject was counting the number of a-vowels in Finnish songs that were played to him through an earphone. Before recording, about 100 electric pulses were delivered to both nerves to familiarize the subject with the stimulation. SEFs to contralateral stimulation were measured separately over the left and right hemispheres. The 24-channel SQUID gradiometer (Ahonen et al. 1991) has planar flux transformers: the tangential derivatives, 0Br/0x and 0 B r / 0 y of B r, the magnetic field perpendicular to the skull, are recorded with two orthogonal figure-of-8 loops simultaneously at 12 locations, 3 cm apart (see Fig. 1). The planar gradiometer records the maximum signal just above a current dipole, and its
Stimuli Constant-current pulses of 0.3 msec duration (Grass $88 stimulator, Grass SIU 4678 isolation unit and Grass CCU 1A constant current unit) were delivered once every 505 msec alternately to median and ulnar TABLE I
Main symptoms, MRI lesions, IgG indexes, and median nerve SEF findings (N20m and P60m) in all patients. In symptoms R and L refer to the right and left sides of the body, in SEFs to the right and left hemisphere. Normal range for IgG index is 0.34-0.60, for SEF latencies and amplitudes control group mean + 2.5 S.D.). Pt 1 2 3 4 5 6 7 8 9 10
Main symptoms
MRI lesions
IgG index
N20m
P60m
L leg weakness Lhermitte + R leg paresthesia L leg weakness L mild hemiparesis Bilateral leg stiffness R leg paresis Urge incontinence Double vision L hand clumsiness R leg paresis L hand paresis Optic neuritis L mild hemiparesis L hand clumsiness R lowered vision L hand clumsiness R lowered vision R tongue paresthesia R hand clumsiness
bilateral periventricular bilateral periventricular bilateral periventricular not available (CT normal) bilateral periventricular at 3rd ventricle level
1.1
L delayed R delayed L delayed and suppressed R delayed and suppressed L delayed R normal L delayed and suppressed R delayed and suppressed L normal R normal L normal R suppressed L delayed and suppressed R suppressed L normal R normal L normal R suppressed L normal R normal
L enlarged R enlarged L enlarged R enlarged L normal R enlarged L enlarged R enlarged L enlarged R normal L normal R absent L normal R absent L absent R normal L absent R normal L normal R normal
1.3 not available 1.4 3.7 1.2
normal (low quality)
0.8
bilateral periventricular normal
0.7
one at 4th ventr, level two medullary
0.6
1.7
194 J. K A R H U El" AL
PATIENT
33 ms I
,~-~.-.
8v~~.~
9
~
, o ~
39 ms
63 ms
E 100 fT/cm
I74 6t
~
/
m 20 ms
1 2}-,,,
891q
\
CONTROL ~ 1 --,.f'~.-'~
Y L 2 ~
20 ms 45 ms 59 ms 89 ms
~
7 ~
]8~.~~ 30 ms
9 ~
1
[ 100 fT/cm 11 ~,~/"'"~,,P'~. 1 I"-1
20 ms
2
~
0
~
f
SEFs IN M U L T I P L E SCLEROSIS
195
PATIENT
CONTROL
Fig. 2. Isocontour maps calculated from the 24-channel recordings of Fig. 1. Shadowed areas indicate magnetic flux out of the head, white areas flux into the head. The isocontour lines are separated by 30 fT, except in the 63 msec m a p of the patient, marked by " s t e p x 2 " (60 fT). The locations and orientations of equivalent dipoles are indicated by arrows on the maps. The approximate area covered by the maps is shown in the insert.
coverage (O = 12.5 cm) allowed the source areas of SEFs to be determined with one positioning of the instrument. Measuring the SEFs to ulnar and median nerve stimulation from both hemispheres took about 1 h, which we consider acceptable for clinical studies. The exact locations and orientations of the sensors with respect to the head were determined by analyzing the magnetic signals resulting from currents in 3 small coils fixed on known locations on the scalp (Knuutila et al. 1987).
Signal analysis The recording passband was 0.05-500 Hz (3 dB points, high-pass roll-off 35 d B / d e c a d e and low-pass over 80 d B / d e c a d e ) . The signals were digitized at 2 kHz and about 100 responses per nerve were averaged on-line at each location. The first two responses of the stimulus block were rejected from the analysis. Each
measurement was repeated once to see the replicability of the responses. The analysis period was 140 msec, and response amplitudes were measured with respect to a 20 msec prestimulus baseline. The channel pair with the maximum total signal (~/(OBr/0x)2+ (0Br/Oy) 2) in each hemisphere was chosen for latency and amplitude measurements. To obtain isocontour maps from the gradient measurements, minimum-norm current estimates (MNE) corresponding to the measured data were first computed. Thereafter, B r generated by the MNE was determined (Ilmoniemi 1991). This procedure accomplishes a 3-dimensional interpolation suitable for the magnetic field. The deviations of the sensitivity axes of the gradiometer from the radial direction are also taken into account. Equivalent current dipoles (ECDs) were determined
Fig. 1. Magnetic responses of patient 2 and control subject 3 to stimulation of the left median nerve at the wrist. The responses were recorded over the right hemisphere with the 24-channel gradiometer. The approximate locations of the channels are shown in the insert. The upper traces of each response pair show the field gradient in the vertical (y) and the lower in the horizontal (x) direction. The positions of the instrument in the patient and in the control subject were within 1 cm from each other with respect to the external landmarks of the skull. Two records (N = 100) are superimposed. Responses were digitally low-pass filtered at 140 Hz.
196
J. K A R H U ET AL.
TABLE II Mean ± S.E.M. latencies and amplitudes for main deflections of control group, patient group and subgroup of patients with periventicular lesions in the MRI (MRI + ). N indicates the number of analyzed hemispheres. Statistical significances of differences between control and patient groups are marked by asterisks: * < 0.05, ** < 0.01 and *** < 0.001 (2-tailed t test for group differences). Latency (msec)
Amplitude ( f T / c m )
Controls
Patients
Patients/MRI +
Controls
20.7_+0.3 N = 16 33.8 ± 0.9 N=I6 56.6 ± 1.4 N = 13
23.2±0.9 * N = 20 37.4+ 1.2 * N=I7 57.3 ± 1.7 N = 16
23.6± 1.1 N = 10 37.5+ 1.8 ** N=8 57.4 ± 2.0 N = 10
73.0_+6.9
21.5±0.4 N = 16 34.6± 1.0 N = 16 57.6± 1.6 N = 14
24.4± 1.2 * N = 19 36.4_+ 1.4 N = 17 62.1 ± 1.5 N = 16
25.2± 1.7 * N = 10 37.9± 2.0 N = 10 61.7± 2.1 N = 10
Patients
Patients/MR1 +
M e d i a n nerce
N20m P30m P60m
55.5± 8.8
62.7± 7.6
114.0± 16.0 **
73.1 _+6.9
140.9 ± 19.9 **
51.7± 8.1 80.6± 16.1 163.5 _+28.7 ** *
Ulnar nert~e
N20m P30m P60m
with a least-squares search from the original data, applying the measured locations and orientations of the sensors. In the analysis, a spherical conductor model was used with a radius of 10 cm, corresponding to the local radius of curvature of the head in the measurement area. Only ECDs which accounted for over 80% of the measured field variance were included in further analysis.
59.3±7.2
41.5_+ 5.4
48.5+ 8.1
44.7_+5.2
72.4_+9.8 *
61.5_+ 10.8
62.5 ±7.6
111.0-+ 13.9 **
116.9± 11.7 ***
Results
Fig. 1 shows median nerve SEFs of a patient and a control subject over the right hemisphere. SEFs of the control subject contain 5 main deflections: N20m, P30m, N45m, P60m and P85m peaking at 20, 30, 45, 59 and 89 msec, respectively. In the patient, N20m is suppressed and delayed (latency 33 msec), and P60m at
TABLE III Mean _+S.E.M. strengths (Q) and depths of ECDs at the time of main deflections of control group, patient group, and patient group with periventricular lesions in the MRI (MRI +). N indicates the number of analyzed hemispheres. Dipoles with goodness-of-fit values over 80% are included in the analysis. Statistical significances of differences between control and patient groups are marked by asterisks: * < 0.05, ** < 0.01 and ***< 0.001 (2-tailed t test). Q (nAm)
Depth (mm)
Controls
Patients
Patients/MRI +
Controls
Patients
Patients/MRI +
18.4 ± 1.5 N=16 22.5 ± 3.2 N=13 23.6 ± 3.2 N=13
19.0 _+2.9 N=12 33.4 ± 3.5 * N=17 41.5 ± 4.6 ** N=14
21.0 ± 3.5 N=8 29.5 ± 5.1 N=8 47.1 ± 3.6 *** N=8
30.1 ± 1.6
28. l _+ 1.7
31.3 ± 2.1
27.3 ± 1.8
30.1 _+ 1.3
31.7 ± 2.6
28.5 _+2.4
31.0 ± 2.2
30.7 ± 2.7
13.1 + 1.4 N = 15 15.5 ± 1.6 N=16 21.8 ± 1.9 N=14
13.5+2.1 N = 10 20.3 _+2.4 N=17 34.5 _+5.3 * N=16
15.9_+3.0 N= 6 17.7 ± 3.3 N=9 36.6 ± 7.6 * N=8
25.1 ± 1.7
29.4_+2.1
31.0+_3.1
27.3 _+ 1.6
26.4 _+2.2
25.9 _+3.2
28.4 ± 1.6
29.2 ± 2.7
29.1 ± 4.5
M e d i a n nerce
N20m P30m P60m
Ulnar nert~e
N20m P30m P60m
----}
Fig. 3. Responses to contralateral median and ulnar nerve stimulations over both hemispheres in patients 1-10 and controls 1-8. Two records (N = 100) are superimposed. The approximate measurement location is shown in the insert.
197
SEFs IN MULTIPLE SCLEROSIS PATIENTS median nerve
CONTROLS ulnar nerve
median nerve
ulnar nerve
t
3 °
~,c 5
6
6/
+
9
0 20 ms
,,,,,4
0 20 ms
I00 fT/cm
20 ms
0 20 ms CONTROLS
PATIENTS median nerve
median nerve
ulnar nerve
ulnar nerve
& L
L ,
,
6
7
8~ 0 20 ms
100 fT/cm
o 20 ms
0 20 ms
0 20 ms
198
63 msec is markedly enlarged; N45m and P85m are not seen. Fig. 2 shows magnetic field patterns to median nerve stimulation during the main deflections of Fig. 1. In the control subject, the field patterns are dipolar and ECDs explain 88-97% of them. The single current dipole is thus considered to model adequately the underlying brain currents. In the patient, no dipolar pattern is found at latencies 18-35 msec, and the field at 63 msec (P60m) is conspicuously stronger than that of the control. ECDs account for 89% of the measured field at 39 msec (P30m) and for 96% at 63 msec. The locations and orientations of ECDs suggest the underlying sources to be in the hand area of the primary sensorimotor cortex both in the patient and in the control subject. Fig. 3 shows examples of median and ulnar nerve SEFs for all patients and control subjects over both hemispheres. In patients 1, 2, 3, 4 and 7 N20m was delayed for both median and ulnar nerve stimulations (Table I). The delay of N20m was significant in the patient group compared with the controls ( P < 0.05, 2-tailed t test for group differences, cf., Table lI). If the field pattern at latencies 18-35 msec was not explained satisfactorily with an ECD, N20m was considered to be suppressed. This happened in patients 2, 4, 6, 7 and 9 for median nerve and in patients 5, 8 and 10 for ulnar nerve stimulations. Therefore, the current dipole model was unsatisfactory during N20m for 18 out of 40 nerves, and at least one abnormal N20m was found in each patient. In control subjects, N45m was detected to stimulation of both nerves in 15/16 of examined hemispheres. Its mean + S.E.M. latency was 46.9 + 0.9 msec to median and 44.4 _+ 1.0 msec to ulnar nerve stimulation. In patients, N45m was seen bilaterally only in patient 10. P60m was enhanced, i.e., exceeded the mean amplitude of controls by more than 2.5 S.D. (see Table II), for both nerves in patients 1-5 and for ulnar nerve in patient 6. Altogether, P60m was enhanced in 11 out of 20 hemispheres studied. In 3 of the 4 subjects with large bilateral P60m, MR1 showed multiple bilateral plaques around lateral ventricles (patients 1, 2 and 3); in patient 4, MRI was not available. In patients 5 and 8, who also had bilateral periventricular lesions, P60m was enlarged unilaterally. The mean amplitude of P60m was nearly 2-fold in the patient group compared with the controls ( P < 0.01) and over 2-fold in the MRI + subgroup ( P < 0.001). Analysis of equivalent current dipoles gave corresponding results (see Table liD. Both patients with plaques below or at the 3rd ventricle level (patients 6 and 10) had bilaterally enlarged P30m. P85m was detected in 10 out of 20 of examined hemispheres of patients and in 14 out of 16 of controls, with significantly larger amplitude in patients than in controls (mean amplitudes differed at P < 0.01 for
J. K A R H U ET AL.
ULNAR NERVE
MEDIANNERVE • patient 0 controlsubject
Yl
Y
2
N20m
-2 N20m
2 cm
I
r
o 2 cm
X
~o ° -
O
Q
•
Y
N > 20 nAm
Y 2
-2
P60m I
2 cm
o--~---I~ x e•Ol
P60m t ° ' ~ o ~
2 cm ~x
Fig. 4. Source locations of N20m and P60m to median and ulnar nerve stimulations in control subjects (open circles) and patients (filled circles) in the left hemisphere. Mean source locations (larger circles) and m e a n equivalent current dipole orientations and strengths are also shown. The coordinate system is demonstrated in the center: the y-axis connects vertex and ear canal and the origin is 50 m m from vertex towards the ear canal.
median nerve, and at P < 0.05 for ulnar nerve). There was no further amplitude enhancement in the MRI + group. The P85m latencies did not differ significantly between patients and controls (median nerve: controls 85.1 + 2.2 msec, patients 85.5 + 2.0 msec; ulnar nerve: controls 83.1 + 2.1 msec, patients 86.2 _+_1.3 msec). The middle-latency SEFs around 60 msec were not altered by attention in control experiments described in Methods. Fig. 4 shows the locations of ECDs for N20m and P60m to median and ulnar nerve stimulation in all subjects. The mean source locations did not differ significantly between patients and controls. The ECDs were on average 30 mm beneath the scalp (see Table III). All sources, also those of P30m, N45m and P85m, were in an area of 4 cm × 4 cm with respect to the external landmarks of the skull.
Discussion
Focal demyelination can decrease conduction velocity, abolish responses to high-frequency impulse trains, or cause a total conduction block (McDonald 1963). These abnormalities are limited to the demyelinated
SEFs IN MULTIPLE SCLEROSIS
zone whereas the other regions of the nerve fiber conduct in a relatively normal manner. However, loss of synchrony between impulses in different fibers can cause temporal dispersion of the compound response of the nerve (for a review, see Waxman 1988), which affects evoked cortical responses. Thus, in addition to prolonged latencies, changes in SEP and SEF wave forms are to be expected in MS. The latencies of N20m deflections agreed in general with earlier clinical median nerve SEP recordings in our MS patients, even though successive evoked potential recordings can vary considerably despite the stable clinical deficit (Aminoff et al. 1984). Most SEP abnormalities are obtained from limbs showing sensory impairment, but there are reports of delayed or abolished SEPs from limbs with no sensory deficit in the area of nerve tested (Namerow 1968). In the present study, no obvious relationship was found between the clinical signs in upper extremities and the lateralization of SEF findings. The enhancement of P60m was unexpected: it was evident in all patients with bilateral periventricular MRI lesions (see Fig. 3). This might reflect relative enhancement of P60m due to disappearance of N45m, which is superimposed on the P60m slope. Another possibility is that increased temporal dispersion of the early input volleys to the cortex decreases later inhibition in sensorimotor cortex through, as yet, unknown mechanisms. SEPs may be recorded over the scalp when sensory nerve action potentials are absent, presumably because responses arising f r o m a few normally conducting axons are synchronized and amplified in the central somatosensory pathways (Parry and Aminoff 1987). Central amplification seems to be impaired in patients with MS (Eisen et al. 1982). An increased amplification is thus an unlikely explanation of the enlarged middlelatency SEFs, especially because the early cortical components are often simultaneously attenuated. Delay of N20m did not necessarily result in delay of P60m, and patients 3 and 5 with periventricular MRI findings had enhanced P60m, even though their N20m was normal. Therefore, in addition to sequential cortico-cortical connections (Allison et al. 1989; Garraghty et al. 1990), also other afferent inputs besides those responsible of N20m contribute to later deflections, possibly via separate thalamo-cortical projections. Evidently early and middle-latency SEP and SEF deflections reflect at least partially parallel processing of somatosensory information and could both be evaluated in MS patients. Correspondingly, separate components of visual evoked potentials can be independently affected in multiple sclerosis (Ghilardi et al. 1991). MS lesions visible in MRI are often located in the periventricular areas, at or close to the somatosensory thalamo-cortical radiation (Turano et al. 1991). There-
199
fore, alteration of middle-latency somatosensory responses should be relatively frequent. That such changes have not been reported in SEPs is probably not purely due to differences between the L E G and MEG. Clinical SEPs are often measured with a highpass filter at 10-20 Hz, which attenuates middle-latency deflections. Furthermore, the analysis periods are often shorter than 50 msec. In fact, enlarged 60 msec responses have been observed in wide-passband SEP and SEF recordings of MS patients, as well as in one patient with a tumor in the periventricular area (A. Asgan, R. Bain and Y. Yang, Glasgow, personal communication). According to our results, middle-latency SEFs are altered in MS, especially in patients with lesions around the lateral ventricles. These responses seem to originate in the primary sensorimotor cortex, and evaluation of both them and their electric counterparts may give additional information about somatosensory function in MS. Prof. O.V. Lounasmaa and Dr. J. Partanen made valuable comments on the manuscript.
References Ahonen, A., H~im~il~iinen, M., Kajola, M., Knuutila, J., Lounasmaa, O., Simola, J., Tesche, C. and Vilkman, V. Multichannel SQUID systems for brain research. IEEE Trans. Magn., 1991, 27: 27862792. Allison, T., McCarthy, G., Wood, C., Darcey, T., Spencer, D. and Williamson, P. Human cortical potentials evoked by stimulation of the median nerve. I. Cytoarchitectonic areas generating shortlatency activity. J. Neurophysiol., 1989, 62: 694-710. Aminoff, M.J., Davis, S.L. and Panitch, H.S. Serial evoked potential studies in patients with definite multiple sclerosis. Clinical relevance. Arch. Neurol., 1984, 41: 1197-1202. Chiappa, K.H. Pattern shift visual, brainstem auditory and short latency somatosensory evoked potentials in multiple sclerosis. Neurology, 1980, 30: 110-123. Desmedt, J. and Cheron, G. Somatosensory evoked potential to finger stimulation in healthy octogenarians and in young adults: wave forms, scalp topography and transit times of parietal and frontal connections. Electroenceph. clin. Neurophysiol., 1980, 50: 404-425. Desmedt, J. and Tomberg, C. Mapping early somatosensory evoked potentials in selective attention: critical evaluation of control conditions used for titrating by difference the cognitive P30, P40, P100 and N140. Electroenceph. clin. Neurophysiol., 1989, 74: 321-346. Eisen, A., Purves, S. and Hoirch, R.T. Central nervous system amplification: its potential in the diagnosis of early multiple sclerosis. Neurology, 1982, 32: 359-364. Garcla Larrea, L. and Maugui~re, F. Latency and amplitude abnormalities of the scalp far-field P14 to median nerve stimulation in multiple sclerosis. A SEP study of 122 patients recorded with a non-cephalic reference montage. Electroenceph. clin. Neurophysiol., 1988, 71: 180-186. Garraghty, P., Florence, S. and Kaas, J. Ablations of areas 3a and 3b of monkey somatosensory cortex abolish cutaneous responsivity in area 1. Brain Res., 1990, 528: 165-169.
200 Ghilardi, M.F., Sartucci, F., Brannan, J.R., Onofrj, M.C., BodisWollner, I., Mylin, L. and Stroch, R. N70 and P100 can be independently affected in multiple sclerosis. Electroenceph. clin. Neurophysiol., 1991, 80: 1-7. Halliday, A.M., McDonald, W.I. and Mushin, J. The visual evoked response in the diagnosis of multiple sclerosis. Br. Med. J., 1973, 4: 661-664. Hari, R. and Lounasmaa, O.V. Recording and interpretation of cerebral magnetic fields. Science, 1989, 244: 432-436. Hari, R., Reinikainen, K., Kaukoranta, E., H~im~il~iinen, M., I1moniemi, R., Penttinen, A., Salminen, J. and Teszner, D. Somatosensory evoked cerebral magnetic fields from SI and SII in man. Electroenceph. clin, Neurophysiol., 1984, 57: 254-263. Hari, R., H~im~il~iinen,M., H~im~il~iinen,H., Kekoni, J., Sams, M. and Tiihonen, J. Separate finger representations at the human second somatosensory cortex. Neuroscience, 1990, 37: 245-249. Huttunen, J., Hari, R. and Leinonen, L. Cerebral magnetic responses to stimulation of ulnar and median nerves. Electroenceph. clin. Neurophysiol., 1987, 66: 391-400. llmoniemi, R.J. Estimates of neuronal current distributions. Acta Oto-Laryngol. (Stockh.), 1991, Suppl. 491: 80-87. Karhu, J., Hari, R., M~ikel~i,J.P., Huttunen, J. and Knuutila, J. MEG findings in multiple sclerosis. In: 8th International Conference on Biomagnetism, Book of Abstracts, 1991: 251-252. Knuutila, J., Ahlfors, S., Ahonen, A., H~illstr6m, J., Kajola, M., Lounasmaa, O.V., Tesche, C. and Vilkman, V. A large-area low-noise seven-channel DC SQUID magnetometer for brain research. Rev. Sci. Instrum., 1987, 58: 2145-2156. Mastaglia, F.L., Black, J.L., Cala, L.A. and Collins, D.W.K. Evoked potentials, saccadic velocities and computerized tomography in diagnosis of multiple sclerosis. Br. Med. J., 1977, 1: 1315-1317. McDonald, W.I. The effects of experimental demyelination on conduction in peripheral nerve: a histological and electrophysiological study. II. Electrophysiological observations. Brain, 1963, 86: 501-524. Mills, K.R. and Murray, N.M.F. Corticospinal tract conduction time in multiple sclerosis. Ann. Neurol., 1985, 18: 601-605. Namerow, N.S. Somatosensory evoked responses in multiple sclerosis with varying sensory loss. Neurology, 1968, 18: 1197-1204.
J. KARHU ET AL. Parry, G.J. and Aminoff, M.J. Somatosensory evoked potentials in chronic acquired demyelinating peripheral neuropathy. Neurology, 1987, 37: 313-316. Poser, C.M., Paty, D.W., Schernberg, L., McDonald, W.J., Davis, F.A., Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberherg, D.H. and Tourtellotte, W.W. New diagnostic criteria for multiple sclerosis - guidelines for research protocols. Ann. Neurol., 1983, 13: 227-231. Suk, J., Ribary, U., Cappell, J., Yamamoto, T. and Llinfis, R. Anatomical localization revealed by MEG recordings of the human somatosensory system. Electroenceph. clin. Neurophysiol., 1991, 78: 185-196. Tiihonen, J., Hari, R. and H~im~il~iinen, M. Early deflections of cerebral magnetic responses to median nerve stimulation. Electroenceph, clin. Neurophysiol., 1989, 74: 290-296. Turano, G., Jones, S.J., Miller, D.H., Du Boulay, G.H., Kakigi, R. and McDonald, W.I. Correlation of SEP abnormalities with brain and cervical cord MRI in multiple sclerosis. Brain, 1991~ 114: 663-681. Waxman, S.G. Clinical course and electrophysiology of multiple sclerosis. In: S.G. Waxman (Ed.), Advances in Neurology, Vol. 47: Functional Recovery in Neurological Disease. Raven Press, New York, 1988: 157-184. Williamson, S.J. and Kaufman, L. Biomagnetism. J. Magn. Magn. Mat., 1981, 22: 129-202. Wood, C.C., Cohen, D., Cuffin, B.N., Yarita, M. and Allison, T. Electric sources in the human somatosensory cortex: identification by combined magnetic and potential field recordings. Science, 1985, 227: 1051-1053. Yamada, T. The anatomic and physiologic bases of median nerve somatosensory evoked potentials. Neurol. Clin., 1988, 6: 705-733. Yamada, T., Shivapour, E., Wilkinson, J.T. and Kimura, J. Shortand long-latency somatosensory evoked potentials in multiple sclerosis. Arch. Neurol., 1982, 39: 88-94. Young, I.R., Hall, A.S., Pallis, C.A., Legg, N.J., Byder, G.M. and Steiner, R.E. Nuclear magnetic resonance imaging of the brain in multiple sclerosis. Lancet, 1981, ii: 1063-1066.
