- Email: [email protected]
An analysis of peripheral motor nerve stimulation in humans using the magnetic coil
524
Electroencephalograph)' and chnicai Neurophysiology, 1988, 7 0 : 5 2 4 - 5 3 3 Elsevier Scientific Publishers Ireland, Ltd.
EEG 03629
An analysis of peripheral motor nerve stimulation in humans using the magnetic coil P.J. Maccabee, V.E. Amassian *, R.Q. Cracco and J.A. Cadwell * * Departments of Neurology and * Physiology, State University of New York Health Science Center, Brookli, n, N Y (U.S.A.), and * * Cadwell Laboratories, Kennewick, WA (U.S.A.) (Accepted for publication: 25 March 1988)
Summa~
We compared conventional electrical and magnetic coil (MC) stimulation of distal median nerve in 10 normal subjects and 1 patient. Orthogonal (90 ° to volar forearm)-longitudinal (the plane of the MC aligned with the long axis of nerve or wire), tilted (to 45 ° ) longitudinal, and tangential edge orientations elicited maximal or near maximal c o m p o u n d motor axon potentials (CMAPs) without simultaneous co-activation of ulnar nerve. Transverse and symmetrical tangential orientations were inefficient. A simulation study of an ideal volume conductor confirmed these findings by predicting that the m a x i m u m current density was near the outer edge of the MC and not at the center where the magnetic flux intensity is maximal. An orthogonal-longitudinal MC induces a current in the adjacent volume conductor (for example elbow or wrist), which flows in the same circular direction as in the MC. This differs from a tangentially orientated MC which classically elicits current flow in the volume conductor opposite in circular direction to that in the MC. Amplitude and latency of the C M A P were both altered, but not identically, by changing the intensity of MC and cathodal stimuli. Rotating an orthogonal-longitudinal MC through 180 o thus reversing the direction of current flow, elicited single fiber muscle action potentials whose peak latencies differed at most by 100 ~sec. Thus, the (virtual) cathode and anode are significantly closer (i.e., 5 - 6 mm) with MC than with electrical stimulation where they are at least 20 m m apart. A disadvantage of MC stimulation is the imprecision in defining exactly where the distally propagating nerve impulse originates. In different subjects, using m a x i m u m output and orthogonal or tilted (to 4 5 ° ) longitudinal orientations, the calculated site of excitation in the median nerve varied 2-15 m m distal to the midpoint of the contacting edge of the MC. This limits the usefulness of the MC in its current configuration for determining distal motor latencies. Future advances in MC design may overcome these difficulties.
Key words: Peripheral motor nerve stimulation; Magnetic coil; (Humans)
Following the pioneering construction of a magnetic stimulator by Bickford and Fremming (1965), improvements in design by Barker et al. (1985) opened up numerous possibilities for clinical neurophysiological investigation (Barker et al. 1986; Hess et al. 1986). Using fiat (tangential) MC orientations, a changing magnetic field was passed through high resistance human skull; the induced currents excited relatively large regions of motor Correspondence to: Paul J. Maccabee, M.D., Department of Neurology, State University of New York Health Science Center at Brooklyn, 450 Clarkson Avenue, Box 35, Brooklyn, NY 11203 (U.S.A.).
cortex, producing non-focal movements in the contralateral limb. Amassian et al. (1987) subsequently showed that restricted regions of motor cortex can be stimulated if the MC is positioned orthogonally (more lateral and nearly vertical to motor cortex), and energized at threshold intensity. This orientation can elicit a movement predominantly localized to a contralateral single digit, presumably because an electrical field is induced approximating the long axis of corticospinal neurons and their trajectory in white matter. Early in our investigations of peripheral motor nerve stimulation (Maccabee et al. 1987a, b), a tangentially positioned MC geometrically bisecting the
0013-4649/88/$03.50 © 1988 Elsevier Scientific Publishers Ireland, Ltd.
P E R I P H E R A L M O T O R N E R V E S T I M U L A T I O N BY MC
median nerve over volar forearm generated a rela. tively low amplitude compound motor action potential (CMAP). Moreover, structures including muscle and tendon were also activated, often painfully. However, with an orthogonally positioned MC, the induced electric field was oriented parallel to the peripheral nerve, and maximal CMAP amplitudes were elicited without any accompanying local discomfort. This report establishes some relationships between MC orientation and peripheral, distal motor nerve stimulation, comparing conventional electrical cathodal and MC stimuli. Issues are examined such as shape of MC, evoked CMAP onset latency as a function of stimulus intensity, site of nerve impulse generation in relation to MC location and orientation, effect of reversing the MC and thus reversing the polarity of the induced electric field, and co-activation of other forearm nerves. These studies suggest certain principles in the utilization of the MC which may be useful as it evolves into a commonplace neurophysiological tool.
525 CURRENT DENSITY ¢m.
0.15 •
~x
4
0.25
""
,,
0.35 ~/ ' // 0 50/ ,' 0 66-~---------~, /j 1.00
". \\
=i ,~i ~ f
.*--
x~~ 1 ¼ l , =i
•
I
•
t/ ii
i i; ~ I
t ~
t
/,
x
~
/
ol
,
3.5cm
-~~ I= ~1
I
EPOXY MOLDING CONCENTRIC COIL
Fig. 1. Single isocurrent density lines plotted in a hemi-infinite, volume conductor for a 7 cm diameter MC (see Results). (Modified from Fig. A-I in Cadwell Instruction Manual, 1987.)
pointed MC with maximal outer diameters of 8.6 by 10.6 cm. The maximum output measured at the center of each coil was approximately 2.2 Tesla. The current density distribution for the round MC
Methods Clinical studies were performed on the distal right median nerves of 10 healthy young adults (age 25-43 years, mean age 34; 6 males, 4 females), and on 1 patient with lumbosacral radiculopathy (female, age 64 years). The data in 5 of these subjects were selected from a larger group of 20 normals that is to be reported separately (Maccabee et al. in prep.). All subjects gave informed consent, and the investigation was authorized by our hospital's Institutional Review Board. Electrical stimuli were rectangular pulses, 0.1-0.2 msec in duration, delivered through Ag-AgC1 disks, 1 cm in diameter and 3 cm apart (cathode distal), which were embedded in a plastic mount. Typically, the cathode was located 6.5 cm proximal to the active recording electrode over the thenar eminence. The reference electrode was attached to the distal anterior thumb. In all studies, electrical stimulus intensity was raised to supramaximal and CMAPs were conventionally amplified. Magnetic fields were generated by either a round MC with an outer diameter of 9.2 cm or a
/ /"
0.20
ROUND COIL
POINTED COIL
L. . . . . . . . : ~9.3 cm LONGITUDINAL (o" TOWIREAXIS) ORTHOGONAL (90° TOHORIZONTAL)
WiRE
LONGITUDINAL (O" TO WIREAXIS) (45"TO HORZ I ONTAL) TANGENTIAL EDGE . (O"TO WIREAXIS) ( (7 TOHORIZONTAL)
WIRE
~
AXIS OF COIL
WIRE-
~.'~
(45" TO WIRE AXIS)
WIRE
TRANSVE AR XS IS E COIL (90" TO WIREAXIS) SYMMETRICAL TANGENTIAL
WiRE ~
WIRE
Fig. 2. Voltages induced in a wire by various orientations of both MCs at 100% of maximum output. Positivity is up.
526
P.J. MACCABEE ET AL.
ROUND COIL
COILOUTPUT60% WRIST X- SECTION
5 mV
POINTED COIL
2ms
5mV
2 ms
FLAT EOOE
°
°.o.O.,ZONTAL
90°to HORIZONTAL ~ O* to NERVE AXIS
900 to HORIZONTAL ~ b / 45"to NERVE AXIS . ~ L
TRANSVERSE
90",0 ,0R,ZONTAL A /
Fig. 3. Effect of various orientations of the two MCs on amplitude of CMAPs elicited by median nerve stimulation• In this and all subsequent EMG recordings, positivity is down. was calculated (see Appendix) and is illustrated in Fig. 1. The intensity control on the MC stimulator was linearly related to output. The duty cycle for 100% output was 1.5 sec. All MC stimuli were delivered at rates slower than once every 4 sec. The voltages induced by various orientations of the MC were evaluated in a linear segment of wire approximately twice as long as the outer coil diameter (Fig. 2). Similarly, the influence of MC orientation was studied on distal median nerve (Fig. 3). The uninsulated wire (no. 16 American wire gauge) was secured to a square cardboard box (dimensions in Fig. 2) whose ends were led to an amplifier and storage oscilloscope, positivity being displayed upwards. (In all E M G recordings negativity was up.) An MC positioned perpendicular to the plane of the box or to the volar forearm was designated as orthogonal. An MC aligned with the long axis and placed directly over wire or distal median nerve was designated as longitudinal. The MC orientations tested included: orthogonal-longitudinal, 45 o lateral tilt longitudinal
(half-way between orthogonal and tangential edge), tangential edge, orthogonal-45 o longitudinal, orthogonal-transverse, and symmetrical tangential (Fig. 2). Unless stated otherwise, the MC was orientated so that the current in it flowed counterclockwise when it was positioned orthogonal-longitudinal to the distal right median nerve and viewed from its lateral aspect (left upper trace, Fig. 6). When the MC was at the wrist, the shock artifact was unacceptably large with the reverse orientation. To compare reliably C M A P onset latencies obtained with electrical cathodal and MC stimuli, the middle of the contacting edge of each coil was marked by a line, which could then be used as a reference. In 2 normal subjects, cathodal stimulation was compared with round and pointed MC stimulation at various orientations and stimulus-intensity relations were also evaluated. In the 1 patient, a single fiber muscle action potential was recorded from abductor pollicis brevis (electrode: SF25, Teca; bandpass: 500-3000 Hz). Evoked single fiber muscle action
P E R I P H E R A L M O T O R N E R V E S T I M U L A T I O N BY MC
potentials (SFMAPs) could then be compared when an orthogonal-longitudinally orientated round MC was reversed, thus reversing the induced current. When using this recording electrode, the contacting edge of the MC had to be located at the elbow rather than the wrist to reduce the stimulus artifact. When the MC was physically reversed, the reference line at the contacting edge was carefully replaced at the same distance from the needle recording electrode. The presumed locality of MC stimulation was evaluated as follows: (a) In 3 subjects, an orthogonal-longitudinal MC was moved transversely in 0.5 cm increments, between radial and ulnar sides of the volar forearm, eliciting a CMAP at each location. (b) In 5 subjects, the median nerve was stimulated at the elbow and wrist with the MC in 3 positions (orthogonaMongitudinal, 45 ° lateral-longitudinal, tangential edge) while simultaneously recording from thenar and hypothenar muscles. The site of nerve impulse generation at the wrist corresponding to round MC location and geometry was also assessed. Initial experiments comparing electrical cathodal and MC stimuli revealed that the onset latencies of maximal amplitude CMAPs were frequently identical when the reference line (middle of the contacting edge) of an orthogonal-longitudinal coil was located 1.5-2.0 cm proximal to the cathode. Therefore, CMAPs were elicited with the reference line of the round MC arbitrarily displaced 1.5 cm proximal to the corresponding cathodal stimulation site (located 6.5 cm proximal to the thenar recording electrode). Latencies and amplitudes were measured with a cursor; integrated area was measured between 2 cursor locations.
Results
(1) Current density distribution induced in volume by the M C Typically, the greatest intensity of magnetic field strength in the vicinity of a round MC is at its center. This notably contrasts with the estimated isocurrent density distribution in an infinite volume conductor where current density is negligi-
527
ble in a central cylindrical shaft (Fig. 1). The induced current peaks closer to the outer edge of the MC than the inner edge, attenuating by 80% nearly 4 cm away from an MC that is 7 cm in diameter. Similar intensity plots are not available with an orthogonal orientation of the MC. However, qualitatively, the current intensity is significantly reduced. (2) The effect of M C orientation on voltage induced in a wire The relationship between MC orientation over a linear segment of wire to the induced voltage was evaluated for 6 different MC positions (Fig. 2). The magnitude of the induced voltage was maximal when the MC was held orthogonaMongitudinal. When the MC was tilted 45 ° laterally, but remained longitudinal and contacted the wire, or when it was positioned tangentially so that its flat edge was directly over the wire, the induced voltage was attenuated by approximately 75%. Rotating an orthogonally orientated MC by 45 ° and then by 90 ° to the long axis of the wire progressively diminished the induced voltage. Finally, when the MC was symmetrically positioned tangentially over the wire, the induced voltage was negligible. For all of these orientations, there was no significant difference between the round or pointed MC. The induced voltage flux consisted of a short initial phase extending from 0 to 70 /~sec, followed by a slightly smaller amplitude, reversed phase extending from 70 to 200 t~sec. Because of its longer duration, a greater charge transfer occurred during the reversal phase. Thereafter the flux was damped and progressively lower in amplitude, terminating at approximately 750 ~sec, i.e., beyond the sweep duration in Fig. 2. The possibility that these results were due to the wire and the oscilloscope forming a loop which interacted with the MC was rejected by obtaining analogous results when using an approximately equilateral triangle of wire, each side being a little over 70 cm. (3) Effect of M C orientation over distal median nerve Given the changes in voltage induced in a wire by various MC orientations, changes in excitation
528
P.J. MACCABEE ET AL.
of motor axons of the median nerve might also be expected with changes in orientation of the MC. In Fig. 3, the largest amplitude CMAPs were evoked when the edge of the MC was directly over nerve (tangential edge) and when the MC was tilted laterally to 45 ° (remaining longitudinal to nerve). When the MC was orthogonal and longitudinal to nerve, the CMAP was reduced by 25%. However, which of these 3 orientations yielded the largest amplitude CMAP varied among different individuals (Table I). The tangential edge orientation tended to be the most effective and the longitudinal, the least. The evoked CMAP amplitude was much diminished when the orthogonally orientated MC was rotated 45 ° or 90 o (i.e., transverse) to the long axis of the median nerve. When the MC was placed flat and symmetrical over the median nerve (as in the bottom trace, Fig. 2), the CMAP amplitude was suboptimal. Such stimulation was often accompanied by moderate to severe pain, presumably because additional, local structures were stimulated.
TABLE I
(4) Locus of nerve excitation with longitudinal orientations of the M C In order to locate where the nerve is excited by the MC, we compared CMAP latencies to MC and conventional electrical cathodal stimulation. Because the latency of the CMAP is affected by the intensity of stimulation, response amplitudestimulus intensity relationships were compared with both types of stimulation. Theoretically, changes in CMAP onset latency with changes in stimulus intensity reflect differences in delay for discharge of the membrane capacity and increase in Na conductance to firing level (Hodgkin and Huxley 1952). Additionally, electrical stimuli of increasing intensity are well known to excite underlying nerve at progressively greater distances from the cathode (Wiederholt 1970). In the intensity series (Fig. 4), the MC was held orthogonal and longitudinal to distal median nerve and the output was set to approximate the CMAP amplitudes obtained with cathodal stimuli. With cathodal stimuli, increasing the intensity above threshold yielded a significant shortening in latency. However, the shortening in latency was considerably less when the CMAP amplitude ex-
ceeded 50% of maximum. Similarly, the latency shortened with increased MC stimulation; however, for comparable CMAP amplitudes, the two types of stimulation did not result in identical shortening. Therefore, the latencies of maximal and near-maximal CMAPs to cathodal and MC stimulation were compared. In normal subjects the contacting edge of the MC was arbitrarily displaced 1.5 cm proximal to the cathode; orthogonal-longitudinal or 45 ° lateral-longitudinal, orientations were used to elicit the CMAPs. To avoid any possible complications due to coactivation of the ulnar nerve, 5 normal subjects were selected with ratios of magnetically to electrically ( M / E ) elicited CMAP amplitudes ranging from 0.88 to 1.0 (Table II). Although the Pearson linear correlation coefficient between M / E CMAP amplitude and integrated area ratios was rho = 0.86, this was not statistically significant at P < 0.05. The onset latencies of the CMAPs with MC stimuli were 0.2, 0.1, 0.1, 0, and 0 msec, respectively, greater than the corresponding cathodal onset latencies. Motor conduction velocity was obtained with cathodal stimulation of median nerve at elbow and wrist. The product of motor
Magnetical/electrical ( M / E ) ratios of CMAP amplitudes are shown in 5 normal subjects. The right median nerve was stimulated electrically at the wrist with the cathode 6.5 cm proximal to the active, thenar recording electrode. The reference line on the contacting edge of the MC (100% output) was located 1.5 cm proximal to the cathode. The MC was oriented in 3 positions as indicated. Simultaneous recordings from hypothenar muscles showed no evidence of ulnar nerve coactivation (Fig. 8). Subject
M / E amplitude ratios Orthogonal
45 ° lateral
Tangential
Wrist
1 2 3 4 5
1.13 1.00 0.87 0.48 0.00
1.00 1.00 0.87 1.09 0.64
1.08 1.06 1.00 1.03 0.68
Elbow
1 2 3 4 5
1.00 0.72 1.00 0.64 0.00
1.05 0.96 0.97 0.95 0.66
1.09 0.96 1.00 0.99 0.97
P E R I P H E R A L M O T O R N E R V E S T I M U L A T I O N BY MC BIPOLAR CATHODAL
ROUND COIL
5 mV
2 ms
.
.
. . . .
.
.
.
529
5 mV
I
.
. ~ ~
500pV
POINTED COIL
I ms
i
2 ms > - -
"
5mV
2ms
500~V
I ms
"
500~V
I ms
Fig. 4. Effect of altering cathodal and MC stimulus intensities on C M A P onset latencies and amplitudes. In bottom traces, amplification and sweep speed increased 10 x and 2 × , respectively.
conduction velocity and the difference between MC and cathodal CMAP onset latencies gave a range of distances proximal to the cathode for the impulse origin in the nerve. Fig. 5 shows the relationship between MC location and calculated range of impulse generation sites; the approximate locus of median nerve depolarization occurred 2-15 mm distal to the center of the contacting edge of the MC.
Having found that the origin of the distally propagating impulse can be approximately located in relation to the longitudinally oriented MC, is it possible to define the interpolar distance between virtual cathode and anode? Fig. 6, left traces, shows the theoretical reversal of the induced electric field when the longitudinally oriented MC is rotated 180 ° over the median nerve at the elbow. However, the CMAPs usually failed to show a
T A B L E II The fight median nerve was stimulated at the wrist as described in Methods. C M A P amplitude, integrated area, onset latency, and duration were measured with a cursor. M a x i m u m motor conduction velocity was obtained with cathodal stimuli applied to wrist and elbow. The effective site of impulse propagation was calculated using the cathode as reference (see text and Fig. 5). Although M / E C M A P amplitude and integrated area ratios are highly correlated (rho = 0.86), they are not statistically significant. Subject
6 7 8 9 10
M/E
M/E
CMAP
amplitude ratio
area ratio
Duration (msec)
Onset latency (msec)
Distance from
MC
Cathodal
MC
Cathodal
1.00 1.00 0.96 0.94 0.88
1.10 1.04 1.05 0.93 0.92
6.4 5.1 5.3 5.8 5.0
6.2 5.2 5.2 5.6 4.8
3.9 3.3 3.2 3.4 3.3
3.9 3.1 3.1 3.3 3.3
A
CV (m/sec)
Displacement (ram)
midpoint of contacting edge (mrn)
0.0 0.2 0.1 0.1 0.0
60.4 65.8 59.7 64.7 58.5
0.0 13.1 6.0 6.5 0.0
15 1.9 9.0 8.5 15
530
P.J. MACCABEE ET AL. COL~OO'/,~TPUT 5mY
POINTED
ROUN~ z~
5mv
2ms
p
ad = 'V'a t
Z 3
o .....,/ FLEX~ CARPl RADIALI$ PALMARI$ LONGUS
I
PROXIMAL
I
I
!
~
,
AN E
DISTAL
i
2
:
92mm
Fig. 7. Effect of transversely moving orthogonal-longitudinal orientated MCs across the forearm from radial to ulnar aspects. At right, numbered traces show responses to stimulating at the sites indicated on the diagram at the left.
,,
I5~
"',
RIGHT MEDIANNERVE,WRIST LONGITUDINALCOIL: ORTHOGONALTO 45o
Fig. 5. The effective site of nerve impulse generation in relation to MC and electrical anode and cathode placements over the median nerve. significant difference in l a t e n c y (i.e., m o r e t h a n 100/~sec) when the M C o r i e n t a t i o n y i e l d e d clockwise versus c o u n t e r c l o c k w i s e d i r e c t i o n of c u r r e n t
PROPOSED DIRECTION OF CURRENT FLOW
VIRTUAL ANODE PROXIMAL
6
!i::i:i CATHODE
[ ] RANGE
VIRTUAL CATHODE PROXIMAL
4
STIM.: RT. MEDIAN NERVE AT ELBOW RECORD:SF EMG FROM RT. APe IOO V 2ml.
VIRTUAL ANODE DISTAL
VIRTUAL CATHODE DISTAL
Fig. 6. Effect of direction of current flow in the MC on the latency of single fiber muscle action potentials. At left, diagram depicts the proposed direction of induced current flow in the volume conductor with an orthogonal-longitudinally oriented MC near the elbow. Reversing the direction of current in the MC reverses the circular direction of the induced current (see text). At fight, single fiber muscle action potentials (SFMAPs), in the abductor pollicis brevis, elicited by an MC before and after it is physically rotated through 180 ° to reverse the direction of current flow.
flow. T h e p o s s i b l e p r o b l e m of exciting different m o t o r a x o n s w i t h i n the p o p u l a t i o n with the two M C o r i e n t a t i o n s was a v o i d e d in the p a t i e n t b y r e c o r d i n g a single fiber muscle a c t i o n p o t e n t i a l (Fig. 6, right traces). A c o m p a r i s o n of S F M A P s revealed a slight d i f f e r e n c e in onset latency, app r o x i m a t e l y 100 y s e c earlier with clockwise current, which g e n e r a t e s a distal cathode. Thus, the virtual c a t h o d e a n d a n o d e are s i t u a t e d no more than 0.1 msec times the c o n d u c t i o n velocity apart, i.e., 5 - 6 m m . (5) Nerve excitation lateral to the M C K e e p i n g either the r o u n d or the p o i n t e d M C o r i e n t a t e d o r t h o g o n a l a n d l o n g i t u d i n a l to the long axis of the m e d i a n nerve, the M C was m o v e d f r o m r a d i a l to the u l n a r a s p e c t across the forearm. In 3 subjects, s u b s t a n t i a l C M A P s were elicited only w h e n the c o n t a c t i n g edge (12 m m thick) was directly over the nerve or when the M C was ins i n u a t e d b e t w e e n the t e n d o n s of the flexor carpi r a d i a l i s a n d p a l m a r i s longus (Fig. 7), that is, the closer the M C was to the m e d i a n nerve, the greater the C M A P a m p l i t u d e . M o v i n g the M C in a r a d i a l o r u l n a r d i r e c t i o n a w a y f r o m the m e d i a n nerve led to a p r e c i p i t o u s d r o p in C M A P a m p l i t u d e . (6) Evaluation o f C M A P amplitude, area, and ulnar nerve coactivation In 6 subjects, the m e d i a n nerve was s t i m u l a t e d with 3 M C o r i e n t a t i o n s ( o r t h o g o n a l - l o n g i t u d i n a l , 45 ° lateral t i l t - l o n g i t u d i n a l , t a n g e n t i a l edge) at the e l b o w a n d wrist, s i m u l t a n e o u s l y r e c o r d i n g from t h e n a r a n d h y p o t h e n a r muscles (Fig. 8). I n gen-
PERIPHERAL MOTOR NERVE STIMULATION BY MC eral, the largest M / E C M A P amplitude ratios were obtained with 45 ° lateral tilt-longitudinal and tangential edge orientations. Unexpectedly, the C M A P amplitudes evoked by M C stimuli at the wrist were sometimes slightly greater than those evoked by supramaximal cathodal stimuli (e.g., 6% in Fig. 9, top left and middle records, and Table I). In subjects selected for M / E amplitude ratios ~< 1 (Table II), the corresponding area ratios were more often >/1, because the duration of the C M A P negative phase was usually increased with M C stimulation. W h e n with supramaximal stimulation, the M / F ratio is greater than 1, ulnar coactivation by the M C must be excluded as an explanation. In Fig. 9, top left and middle records, the h y p o t h e n a r C M A P shows no evidence of local activation, the small initially positive deflection reflecting distant activity. True coactivation of h y p o t h e n a r muscles is readily elicited by moving the tangential edge from the optimal site over the median nerve closer to the ulnar border (right record). However, the possibility remains that the h y p o t h e n a r C M A P is not an adequate sample of muscles activated by ulnar m o t o r axons. F o r example, if ulnar m o t o r M.C. LON61TUDINAL, ORTHOGONAL
BIPOLAR - C A T H O D A L STI MULATION
5my //
',\
=
- :
~'/
- t
\\ \
8LE M.C. LONGITUDINAL, 4 5 °
f
t
531 ~ ELECTRICAL MEDIAN NERVE
T..A.
- TANGENTIAL EDGE - OPTIMAL LOCATION 13ram from MEDIAN NERVE ULNAR BORDER M.C. (lO0°/=) M.C (fO0%)
--
i} ov TANGENTIAL EDGE OVER ULNAR NERVE - - ~ M.C. (90-100%)
M.C. (22-40%)
T II 2 rnV
POLLICIS
--
2 mV
Fig. 9. Thenar CMAPs without ulnar coactivation, elicited by electrical and MC stimulation of distal median nerve. Top superimposed records: left, responses to several including supramaximal intensities; middle, maximum response to MC at optimal location; right, ulnar coactivation when MC moved closer to ulnar border. Throughout top records, MC tilted towards radial border. Bottom superimposed records: left, maximal responses to MC stimulation at 90 and 100% output, MC tilted towards ulnar border; right, threshold responses to MC stimulation at 22-40% output, progressing in 2% steps. Gains increased 10 ×, compared to left records.
axons innervating a d d u c t o r pollicis had the lowest threshold, coactivation of the ulnar nerve could occur without a detectable h y p o t h e n a r C M A P . This possibility was evaluated by simultaneously recording from a d d u c t o r pollicis and h y p o t h e n a r muscles with the M C over ulnar nerve and tilted away from the median nerve. Even at high recording sensitivity (Fig. 9, b o t t o m right), no evidence was f o u n d of a lower threshold for a d d u c t o r pollicis.
M.C. TANGENTIAL EDGE
Discussion
Fig. 8. Simultaneous recordings from thenar and hypothenar muscles with electrical and MC stimulation of median nerve at elbow and at wrist to show lack of ulnar coactivation.
R u s h t o n (1927) showed that electrical stimulation of peripheral nerve is most efficient when the imposed direction of current flow is aligned with the long axis of the nerve being stimulated. Similarly, we f o u n d the induced voltages and C M A P s were largest when the M C - i n d u c e d electrical field was orientated parallel to the long axis of a metallic c o n d u c t o r (Fig. 2) and the median nerve, respectively (Fig. 3). Some control of the orientation of
532 the induced electric field is obtainable with the MC because, when it is flat against the volume conductor, the induced current flows in concentric rings parallel to the plane of the MC and in an opposite direction to that in the MC. However, when the MC is orthogonal to the volume conductor, not only is the distribution of current flow more restricted, but it flows in the same circular direction in the volume conductor as in the MC. Thus, the current in the MC and the volume conductor flow in opposite directions at the edge of the volume conductor (Fig. 6). Given an appropriate orientation, usually the MC can maximally excite the median nerve without activation of ulnar nerve (Fig. 8 and Table I). However, supramaximal stimulation cannot be guaranteed in all individuals with the currently available MCs and stimulator outputs. The marked attenuation in CMAP when an orthogonal-longitudinally oriented MC is moved laterally away from the underlying median nerve (Fig. 7) demonstrates the focal properties of the orthogonallongitudinal orientation. Frequently, the largest CMAPs occurred when the longitudinally orientated MC was tilted 45 ° possibly because: (1) the flat outer edge of the MC was insinuated between tendons and thus was closer to the underlying median nerve, and (2) a larger fraction of the magnetic flux is appropriately orientated to enter the volume conductor. A potential disadvantage of stimulation by the MC, even when orientated orthogonally, is the imprecision in defining exactly where the distally propagating nerve impulse arises. In different individuals, when using supramaximal stimuli, the site varied 2-15 mm from the midpoint of the contacting edge. Because of the longer conduction distance, this error would not be serious when stimulating nerve roots, but would be in measuring distal motor latencies. A surprising result was the proximity, less than 5 - 6 mm, of the virtual cathode to the anode induced by an orthogonal-longitudinally oriented MC. By contrast, a much greater interpolar distance, at least 20 mm, is typically used with electrical stimulation. As a result, the direction of current flow in the orthogonally orientated MC is much less critical for distal peripheral nerve stimulation than with conventional electrical stimula-
P.J. MACCABEEET AL. tion. Another unexpected finding was the larger CMAP amplitude (Fig. 9 and Table I) and area (Table II), that may be obtained with MC versus clearly supramaximal electrical stimulation. Possible explanations include: (1) The MC elicits double discharges at short intervals in some motor units. However single fiber muscle action potential recordings did not reveal double discharges. (2) The MC excites both median and ulnar nerves. However, ulnar coactivation was excluded in recordings from hypothenar muscles. No evidence was found that ulnar motor axons supplying other muscles (e.g., adductor pollicis) had a lower threshold than those supplying hypothenar muscle. Therefore the hypothenar muscle appeared to be a valid test for ulnar coactivation. (3) The MC excites motor axons in a temporal pattern which results in less phase cancellation between extracellular action potentials generated by different motor units. For example, where a motor end-plate region of a muscle fiber of motor unit A overlies a portion of a muscle fiber of motor unit B remote from its motor end-plate, phase cancellation of the extracellular electrical signals would tend to occur if both motor axons were synchronously excited. Summation would be promoted if the timing of excitation of the two units was changed, possibly because of the temporal characteristics of the MC-induced electric field. Although the two MCs tested had different shapes, we did not find any marked differences in their properties when stimulating peripheral motor nerves. The orthogonal-longitudinal orientation used in our study was advantageous for focal stimulation despite coupling only a small fraction of the magnetic flux to the volume conductor. Nevertheless, further improvement in MC shape is desirable to increase the focality of stimulation. Furthermore, a larger output is required to ensure supramaximal nerve stimulation in all subjects, an essential requirement for clinical use.
Appendix Calculation of current density (1) The MC was assumed to rest over a hemiinfinite volume conductor. A sense coil (3 loops of
PERIPHERAL MOTOR NERVE STIMULATION BY MC
wire, 1 cm in diameter) was held parallel to the flat plane of the MC at various locations consisting of a grid of spatial points surrounding the MC. At each point the sense coil measured the rate of change of flux (dq,/dt), which is an indirect measure of the peak flux density (/~). (2) Since current flows in concentric circles or annular disks parallel to the MC, and the voltage around any single disc is proportional to total flux (H, where H = B * 7rr2), then the voltage (V) is given by: V = K ~ /~iri2 i~1
where r i = radius of each disk; K = constant; n = number of disks. (3) The resistance (w) is defined as the path length in tissue (2¢r * radius) * resistivity. Assuming unity cross-sectional area for the conductor, the resistance is therefore proportional to the radius (r) of each disk, or w i = r i. (4) Current density at a given radius is then computed from Ohm's law
K i-
e w
i
Bir i=l ri
(5) The current density was computed for all points in the grid and those points with the same current density were plotted as an isocurrent density line (Fig. 1). The authors gratefully acknowledge the technical assistance of Drs. Bradley P. Grayum, Shrinath Kamat, Ken Kaplove,
533 Krishna Nalluri, Vasanthi Nalluri and Lyzette Velazquez. Sam Polaniecki, of the Dept. of Scientific and Medical Instrumentation, helped in the measurement of voltages induced in wire. We thank also Ms Madge Edmead for secretarial assistance.
References Amassian, V.E., Cadwell, J., Cracco, R.Q. and Maccabee, P.J. Focal cerebral and peripheral nerve stimulation in man with the magnetic coil. J. Physiol. (Lond.), 1987, 390: 24P. Barker, A.T., Jalinous, R. and Freeston, I.L. Non-invasive stimulation of human motor cortex. Lancet, 1985, i: 1106-1107. Barker, A.T., Freeston, I.L, Jalinous, R. and Jarrat, J.A. Clinical evaluation of conduction time measurements in central motor pathways using magnetic stimulation of the human brain. Lancet, 1986, i: 1325-1326. Bickford, R.G. and Fremming, B.D. Neuronal stimulation by pulsed magnetic fields in animals and man. In: Digest of the 6th Int. Conf. on Medical Electronics and Biological Engineering, Tokyo, 1965: 112. Hess, C.W., Mills, K.R. and Murray, N.M.F. Measurement of central motor conduction in multiple sclerosis by magnetic brain stimulation. Lancet, 1986, i: 355-358. Hodgkin, A.L and Huxley, Aft. A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. (Lond.), 1952, 117: 500-544. Maccabee, P.J., Amassian, V.E. and Cracco, R.Q. A comparison of electrical and magnetic coil stimulation of peripheral nerves in man. Neurosci. Abstr., 1987a, 13: 1265. Maccabee, P.J., Amassian, V.E. and Cracco, R.Q. Focal stimulation of peripheral nerves using the magnetic coil. Muscle Nerve, 1987b, 10: 642-643. Rushton, W.A.H. Effect upon the threshold for nervous excitation of the length of nerve exposed and the angle between current and nerve. J. Physiol. (Lond.), 1927, 63: 357-377. Wiederholt, W.C. Threshold and conduction velocity in isolated mixed nerves. Neurology, 1970, 20: 347-352.
Electroencephalograph)' and chnicai Neurophysiology, 1988, 7 0 : 5 2 4 - 5 3 3 Elsevier Scientific Publishers Ireland, Ltd.
EEG 03629
An analysis of peripheral motor nerve stimulation in humans using the magnetic coil P.J. Maccabee, V.E. Amassian *, R.Q. Cracco and J.A. Cadwell * * Departments of Neurology and * Physiology, State University of New York Health Science Center, Brookli, n, N Y (U.S.A.), and * * Cadwell Laboratories, Kennewick, WA (U.S.A.) (Accepted for publication: 25 March 1988)
Summa~
We compared conventional electrical and magnetic coil (MC) stimulation of distal median nerve in 10 normal subjects and 1 patient. Orthogonal (90 ° to volar forearm)-longitudinal (the plane of the MC aligned with the long axis of nerve or wire), tilted (to 45 ° ) longitudinal, and tangential edge orientations elicited maximal or near maximal c o m p o u n d motor axon potentials (CMAPs) without simultaneous co-activation of ulnar nerve. Transverse and symmetrical tangential orientations were inefficient. A simulation study of an ideal volume conductor confirmed these findings by predicting that the m a x i m u m current density was near the outer edge of the MC and not at the center where the magnetic flux intensity is maximal. An orthogonal-longitudinal MC induces a current in the adjacent volume conductor (for example elbow or wrist), which flows in the same circular direction as in the MC. This differs from a tangentially orientated MC which classically elicits current flow in the volume conductor opposite in circular direction to that in the MC. Amplitude and latency of the C M A P were both altered, but not identically, by changing the intensity of MC and cathodal stimuli. Rotating an orthogonal-longitudinal MC through 180 o thus reversing the direction of current flow, elicited single fiber muscle action potentials whose peak latencies differed at most by 100 ~sec. Thus, the (virtual) cathode and anode are significantly closer (i.e., 5 - 6 mm) with MC than with electrical stimulation where they are at least 20 m m apart. A disadvantage of MC stimulation is the imprecision in defining exactly where the distally propagating nerve impulse originates. In different subjects, using m a x i m u m output and orthogonal or tilted (to 4 5 ° ) longitudinal orientations, the calculated site of excitation in the median nerve varied 2-15 m m distal to the midpoint of the contacting edge of the MC. This limits the usefulness of the MC in its current configuration for determining distal motor latencies. Future advances in MC design may overcome these difficulties.
Key words: Peripheral motor nerve stimulation; Magnetic coil; (Humans)
Following the pioneering construction of a magnetic stimulator by Bickford and Fremming (1965), improvements in design by Barker et al. (1985) opened up numerous possibilities for clinical neurophysiological investigation (Barker et al. 1986; Hess et al. 1986). Using fiat (tangential) MC orientations, a changing magnetic field was passed through high resistance human skull; the induced currents excited relatively large regions of motor Correspondence to: Paul J. Maccabee, M.D., Department of Neurology, State University of New York Health Science Center at Brooklyn, 450 Clarkson Avenue, Box 35, Brooklyn, NY 11203 (U.S.A.).
cortex, producing non-focal movements in the contralateral limb. Amassian et al. (1987) subsequently showed that restricted regions of motor cortex can be stimulated if the MC is positioned orthogonally (more lateral and nearly vertical to motor cortex), and energized at threshold intensity. This orientation can elicit a movement predominantly localized to a contralateral single digit, presumably because an electrical field is induced approximating the long axis of corticospinal neurons and their trajectory in white matter. Early in our investigations of peripheral motor nerve stimulation (Maccabee et al. 1987a, b), a tangentially positioned MC geometrically bisecting the
0013-4649/88/$03.50 © 1988 Elsevier Scientific Publishers Ireland, Ltd.
P E R I P H E R A L M O T O R N E R V E S T I M U L A T I O N BY MC
median nerve over volar forearm generated a rela. tively low amplitude compound motor action potential (CMAP). Moreover, structures including muscle and tendon were also activated, often painfully. However, with an orthogonally positioned MC, the induced electric field was oriented parallel to the peripheral nerve, and maximal CMAP amplitudes were elicited without any accompanying local discomfort. This report establishes some relationships between MC orientation and peripheral, distal motor nerve stimulation, comparing conventional electrical cathodal and MC stimuli. Issues are examined such as shape of MC, evoked CMAP onset latency as a function of stimulus intensity, site of nerve impulse generation in relation to MC location and orientation, effect of reversing the MC and thus reversing the polarity of the induced electric field, and co-activation of other forearm nerves. These studies suggest certain principles in the utilization of the MC which may be useful as it evolves into a commonplace neurophysiological tool.
525 CURRENT DENSITY ¢m.
0.15 •
~x
4
0.25
""
,,
0.35 ~/ ' // 0 50/ ,' 0 66-~---------~, /j 1.00
". \\
=i ,~i ~ f
.*--
x~~ 1 ¼ l , =i
•
I
•
t/ ii
i i; ~ I
t ~
t
/,
x
~
/
ol
,
3.5cm
-~~ I= ~1
I
EPOXY MOLDING CONCENTRIC COIL
Fig. 1. Single isocurrent density lines plotted in a hemi-infinite, volume conductor for a 7 cm diameter MC (see Results). (Modified from Fig. A-I in Cadwell Instruction Manual, 1987.)
pointed MC with maximal outer diameters of 8.6 by 10.6 cm. The maximum output measured at the center of each coil was approximately 2.2 Tesla. The current density distribution for the round MC
Methods Clinical studies were performed on the distal right median nerves of 10 healthy young adults (age 25-43 years, mean age 34; 6 males, 4 females), and on 1 patient with lumbosacral radiculopathy (female, age 64 years). The data in 5 of these subjects were selected from a larger group of 20 normals that is to be reported separately (Maccabee et al. in prep.). All subjects gave informed consent, and the investigation was authorized by our hospital's Institutional Review Board. Electrical stimuli were rectangular pulses, 0.1-0.2 msec in duration, delivered through Ag-AgC1 disks, 1 cm in diameter and 3 cm apart (cathode distal), which were embedded in a plastic mount. Typically, the cathode was located 6.5 cm proximal to the active recording electrode over the thenar eminence. The reference electrode was attached to the distal anterior thumb. In all studies, electrical stimulus intensity was raised to supramaximal and CMAPs were conventionally amplified. Magnetic fields were generated by either a round MC with an outer diameter of 9.2 cm or a
/ /"
0.20
ROUND COIL
POINTED COIL
L. . . . . . . . : ~9.3 cm LONGITUDINAL (o" TOWIREAXIS) ORTHOGONAL (90° TOHORIZONTAL)
WiRE
LONGITUDINAL (O" TO WIREAXIS) (45"TO HORZ I ONTAL) TANGENTIAL EDGE . (O"TO WIREAXIS) ( (7 TOHORIZONTAL)
WIRE
~
AXIS OF COIL
WIRE-
~.'~
(45" TO WIRE AXIS)
WIRE
TRANSVE AR XS IS E COIL (90" TO WIREAXIS) SYMMETRICAL TANGENTIAL
WiRE ~
WIRE
Fig. 2. Voltages induced in a wire by various orientations of both MCs at 100% of maximum output. Positivity is up.
526
P.J. MACCABEE ET AL.
ROUND COIL
COILOUTPUT60% WRIST X- SECTION
5 mV
POINTED COIL
2ms
5mV
2 ms
FLAT EOOE
°
°.o.O.,ZONTAL
90°to HORIZONTAL ~ O* to NERVE AXIS
900 to HORIZONTAL ~ b / 45"to NERVE AXIS . ~ L
TRANSVERSE
90",0 ,0R,ZONTAL A /
Fig. 3. Effect of various orientations of the two MCs on amplitude of CMAPs elicited by median nerve stimulation• In this and all subsequent EMG recordings, positivity is down. was calculated (see Appendix) and is illustrated in Fig. 1. The intensity control on the MC stimulator was linearly related to output. The duty cycle for 100% output was 1.5 sec. All MC stimuli were delivered at rates slower than once every 4 sec. The voltages induced by various orientations of the MC were evaluated in a linear segment of wire approximately twice as long as the outer coil diameter (Fig. 2). Similarly, the influence of MC orientation was studied on distal median nerve (Fig. 3). The uninsulated wire (no. 16 American wire gauge) was secured to a square cardboard box (dimensions in Fig. 2) whose ends were led to an amplifier and storage oscilloscope, positivity being displayed upwards. (In all E M G recordings negativity was up.) An MC positioned perpendicular to the plane of the box or to the volar forearm was designated as orthogonal. An MC aligned with the long axis and placed directly over wire or distal median nerve was designated as longitudinal. The MC orientations tested included: orthogonal-longitudinal, 45 o lateral tilt longitudinal
(half-way between orthogonal and tangential edge), tangential edge, orthogonal-45 o longitudinal, orthogonal-transverse, and symmetrical tangential (Fig. 2). Unless stated otherwise, the MC was orientated so that the current in it flowed counterclockwise when it was positioned orthogonal-longitudinal to the distal right median nerve and viewed from its lateral aspect (left upper trace, Fig. 6). When the MC was at the wrist, the shock artifact was unacceptably large with the reverse orientation. To compare reliably C M A P onset latencies obtained with electrical cathodal and MC stimuli, the middle of the contacting edge of each coil was marked by a line, which could then be used as a reference. In 2 normal subjects, cathodal stimulation was compared with round and pointed MC stimulation at various orientations and stimulus-intensity relations were also evaluated. In the 1 patient, a single fiber muscle action potential was recorded from abductor pollicis brevis (electrode: SF25, Teca; bandpass: 500-3000 Hz). Evoked single fiber muscle action
P E R I P H E R A L M O T O R N E R V E S T I M U L A T I O N BY MC
potentials (SFMAPs) could then be compared when an orthogonal-longitudinally orientated round MC was reversed, thus reversing the induced current. When using this recording electrode, the contacting edge of the MC had to be located at the elbow rather than the wrist to reduce the stimulus artifact. When the MC was physically reversed, the reference line at the contacting edge was carefully replaced at the same distance from the needle recording electrode. The presumed locality of MC stimulation was evaluated as follows: (a) In 3 subjects, an orthogonal-longitudinal MC was moved transversely in 0.5 cm increments, between radial and ulnar sides of the volar forearm, eliciting a CMAP at each location. (b) In 5 subjects, the median nerve was stimulated at the elbow and wrist with the MC in 3 positions (orthogonaMongitudinal, 45 ° lateral-longitudinal, tangential edge) while simultaneously recording from thenar and hypothenar muscles. The site of nerve impulse generation at the wrist corresponding to round MC location and geometry was also assessed. Initial experiments comparing electrical cathodal and MC stimuli revealed that the onset latencies of maximal amplitude CMAPs were frequently identical when the reference line (middle of the contacting edge) of an orthogonal-longitudinal coil was located 1.5-2.0 cm proximal to the cathode. Therefore, CMAPs were elicited with the reference line of the round MC arbitrarily displaced 1.5 cm proximal to the corresponding cathodal stimulation site (located 6.5 cm proximal to the thenar recording electrode). Latencies and amplitudes were measured with a cursor; integrated area was measured between 2 cursor locations.
Results
(1) Current density distribution induced in volume by the M C Typically, the greatest intensity of magnetic field strength in the vicinity of a round MC is at its center. This notably contrasts with the estimated isocurrent density distribution in an infinite volume conductor where current density is negligi-
527
ble in a central cylindrical shaft (Fig. 1). The induced current peaks closer to the outer edge of the MC than the inner edge, attenuating by 80% nearly 4 cm away from an MC that is 7 cm in diameter. Similar intensity plots are not available with an orthogonal orientation of the MC. However, qualitatively, the current intensity is significantly reduced. (2) The effect of M C orientation on voltage induced in a wire The relationship between MC orientation over a linear segment of wire to the induced voltage was evaluated for 6 different MC positions (Fig. 2). The magnitude of the induced voltage was maximal when the MC was held orthogonaMongitudinal. When the MC was tilted 45 ° laterally, but remained longitudinal and contacted the wire, or when it was positioned tangentially so that its flat edge was directly over the wire, the induced voltage was attenuated by approximately 75%. Rotating an orthogonally orientated MC by 45 ° and then by 90 ° to the long axis of the wire progressively diminished the induced voltage. Finally, when the MC was symmetrically positioned tangentially over the wire, the induced voltage was negligible. For all of these orientations, there was no significant difference between the round or pointed MC. The induced voltage flux consisted of a short initial phase extending from 0 to 70 /~sec, followed by a slightly smaller amplitude, reversed phase extending from 70 to 200 t~sec. Because of its longer duration, a greater charge transfer occurred during the reversal phase. Thereafter the flux was damped and progressively lower in amplitude, terminating at approximately 750 ~sec, i.e., beyond the sweep duration in Fig. 2. The possibility that these results were due to the wire and the oscilloscope forming a loop which interacted with the MC was rejected by obtaining analogous results when using an approximately equilateral triangle of wire, each side being a little over 70 cm. (3) Effect of M C orientation over distal median nerve Given the changes in voltage induced in a wire by various MC orientations, changes in excitation
528
P.J. MACCABEE ET AL.
of motor axons of the median nerve might also be expected with changes in orientation of the MC. In Fig. 3, the largest amplitude CMAPs were evoked when the edge of the MC was directly over nerve (tangential edge) and when the MC was tilted laterally to 45 ° (remaining longitudinal to nerve). When the MC was orthogonal and longitudinal to nerve, the CMAP was reduced by 25%. However, which of these 3 orientations yielded the largest amplitude CMAP varied among different individuals (Table I). The tangential edge orientation tended to be the most effective and the longitudinal, the least. The evoked CMAP amplitude was much diminished when the orthogonally orientated MC was rotated 45 ° or 90 o (i.e., transverse) to the long axis of the median nerve. When the MC was placed flat and symmetrical over the median nerve (as in the bottom trace, Fig. 2), the CMAP amplitude was suboptimal. Such stimulation was often accompanied by moderate to severe pain, presumably because additional, local structures were stimulated.
TABLE I
(4) Locus of nerve excitation with longitudinal orientations of the M C In order to locate where the nerve is excited by the MC, we compared CMAP latencies to MC and conventional electrical cathodal stimulation. Because the latency of the CMAP is affected by the intensity of stimulation, response amplitudestimulus intensity relationships were compared with both types of stimulation. Theoretically, changes in CMAP onset latency with changes in stimulus intensity reflect differences in delay for discharge of the membrane capacity and increase in Na conductance to firing level (Hodgkin and Huxley 1952). Additionally, electrical stimuli of increasing intensity are well known to excite underlying nerve at progressively greater distances from the cathode (Wiederholt 1970). In the intensity series (Fig. 4), the MC was held orthogonal and longitudinal to distal median nerve and the output was set to approximate the CMAP amplitudes obtained with cathodal stimuli. With cathodal stimuli, increasing the intensity above threshold yielded a significant shortening in latency. However, the shortening in latency was considerably less when the CMAP amplitude ex-
ceeded 50% of maximum. Similarly, the latency shortened with increased MC stimulation; however, for comparable CMAP amplitudes, the two types of stimulation did not result in identical shortening. Therefore, the latencies of maximal and near-maximal CMAPs to cathodal and MC stimulation were compared. In normal subjects the contacting edge of the MC was arbitrarily displaced 1.5 cm proximal to the cathode; orthogonal-longitudinal or 45 ° lateral-longitudinal, orientations were used to elicit the CMAPs. To avoid any possible complications due to coactivation of the ulnar nerve, 5 normal subjects were selected with ratios of magnetically to electrically ( M / E ) elicited CMAP amplitudes ranging from 0.88 to 1.0 (Table II). Although the Pearson linear correlation coefficient between M / E CMAP amplitude and integrated area ratios was rho = 0.86, this was not statistically significant at P < 0.05. The onset latencies of the CMAPs with MC stimuli were 0.2, 0.1, 0.1, 0, and 0 msec, respectively, greater than the corresponding cathodal onset latencies. Motor conduction velocity was obtained with cathodal stimulation of median nerve at elbow and wrist. The product of motor
Magnetical/electrical ( M / E ) ratios of CMAP amplitudes are shown in 5 normal subjects. The right median nerve was stimulated electrically at the wrist with the cathode 6.5 cm proximal to the active, thenar recording electrode. The reference line on the contacting edge of the MC (100% output) was located 1.5 cm proximal to the cathode. The MC was oriented in 3 positions as indicated. Simultaneous recordings from hypothenar muscles showed no evidence of ulnar nerve coactivation (Fig. 8). Subject
M / E amplitude ratios Orthogonal
45 ° lateral
Tangential
Wrist
1 2 3 4 5
1.13 1.00 0.87 0.48 0.00
1.00 1.00 0.87 1.09 0.64
1.08 1.06 1.00 1.03 0.68
Elbow
1 2 3 4 5
1.00 0.72 1.00 0.64 0.00
1.05 0.96 0.97 0.95 0.66
1.09 0.96 1.00 0.99 0.97
P E R I P H E R A L M O T O R N E R V E S T I M U L A T I O N BY MC BIPOLAR CATHODAL
ROUND COIL
5 mV
2 ms
.
.
. . . .
.
.
.
529
5 mV
I
.
. ~ ~
500pV
POINTED COIL
I ms
i
2 ms > - -
"
5mV
2ms
500~V
I ms
"
500~V
I ms
Fig. 4. Effect of altering cathodal and MC stimulus intensities on C M A P onset latencies and amplitudes. In bottom traces, amplification and sweep speed increased 10 x and 2 × , respectively.
conduction velocity and the difference between MC and cathodal CMAP onset latencies gave a range of distances proximal to the cathode for the impulse origin in the nerve. Fig. 5 shows the relationship between MC location and calculated range of impulse generation sites; the approximate locus of median nerve depolarization occurred 2-15 mm distal to the center of the contacting edge of the MC.
Having found that the origin of the distally propagating impulse can be approximately located in relation to the longitudinally oriented MC, is it possible to define the interpolar distance between virtual cathode and anode? Fig. 6, left traces, shows the theoretical reversal of the induced electric field when the longitudinally oriented MC is rotated 180 ° over the median nerve at the elbow. However, the CMAPs usually failed to show a
T A B L E II The fight median nerve was stimulated at the wrist as described in Methods. C M A P amplitude, integrated area, onset latency, and duration were measured with a cursor. M a x i m u m motor conduction velocity was obtained with cathodal stimuli applied to wrist and elbow. The effective site of impulse propagation was calculated using the cathode as reference (see text and Fig. 5). Although M / E C M A P amplitude and integrated area ratios are highly correlated (rho = 0.86), they are not statistically significant. Subject
6 7 8 9 10
M/E
M/E
CMAP
amplitude ratio
area ratio
Duration (msec)
Onset latency (msec)
Distance from
MC
Cathodal
MC
Cathodal
1.00 1.00 0.96 0.94 0.88
1.10 1.04 1.05 0.93 0.92
6.4 5.1 5.3 5.8 5.0
6.2 5.2 5.2 5.6 4.8
3.9 3.3 3.2 3.4 3.3
3.9 3.1 3.1 3.3 3.3
A
CV (m/sec)
Displacement (ram)
midpoint of contacting edge (mrn)
0.0 0.2 0.1 0.1 0.0
60.4 65.8 59.7 64.7 58.5
0.0 13.1 6.0 6.5 0.0
15 1.9 9.0 8.5 15
530
P.J. MACCABEE ET AL. COL~OO'/,~TPUT 5mY
POINTED
ROUN~ z~
5mv
2ms
p
ad = 'V'a t
Z 3
o .....,/ FLEX~ CARPl RADIALI$ PALMARI$ LONGUS
I
PROXIMAL
I
I
!
~
,
AN E
DISTAL
i
2
:
92mm
Fig. 7. Effect of transversely moving orthogonal-longitudinal orientated MCs across the forearm from radial to ulnar aspects. At right, numbered traces show responses to stimulating at the sites indicated on the diagram at the left.
,,
I5~
"',
RIGHT MEDIANNERVE,WRIST LONGITUDINALCOIL: ORTHOGONALTO 45o
Fig. 5. The effective site of nerve impulse generation in relation to MC and electrical anode and cathode placements over the median nerve. significant difference in l a t e n c y (i.e., m o r e t h a n 100/~sec) when the M C o r i e n t a t i o n y i e l d e d clockwise versus c o u n t e r c l o c k w i s e d i r e c t i o n of c u r r e n t
PROPOSED DIRECTION OF CURRENT FLOW
VIRTUAL ANODE PROXIMAL
6
!i::i:i CATHODE
[ ] RANGE
VIRTUAL CATHODE PROXIMAL
4
STIM.: RT. MEDIAN NERVE AT ELBOW RECORD:SF EMG FROM RT. APe IOO V 2ml.
VIRTUAL ANODE DISTAL
VIRTUAL CATHODE DISTAL
Fig. 6. Effect of direction of current flow in the MC on the latency of single fiber muscle action potentials. At left, diagram depicts the proposed direction of induced current flow in the volume conductor with an orthogonal-longitudinally oriented MC near the elbow. Reversing the direction of current in the MC reverses the circular direction of the induced current (see text). At fight, single fiber muscle action potentials (SFMAPs), in the abductor pollicis brevis, elicited by an MC before and after it is physically rotated through 180 ° to reverse the direction of current flow.
flow. T h e p o s s i b l e p r o b l e m of exciting different m o t o r a x o n s w i t h i n the p o p u l a t i o n with the two M C o r i e n t a t i o n s was a v o i d e d in the p a t i e n t b y r e c o r d i n g a single fiber muscle a c t i o n p o t e n t i a l (Fig. 6, right traces). A c o m p a r i s o n of S F M A P s revealed a slight d i f f e r e n c e in onset latency, app r o x i m a t e l y 100 y s e c earlier with clockwise current, which g e n e r a t e s a distal cathode. Thus, the virtual c a t h o d e a n d a n o d e are s i t u a t e d no more than 0.1 msec times the c o n d u c t i o n velocity apart, i.e., 5 - 6 m m . (5) Nerve excitation lateral to the M C K e e p i n g either the r o u n d or the p o i n t e d M C o r i e n t a t e d o r t h o g o n a l a n d l o n g i t u d i n a l to the long axis of the m e d i a n nerve, the M C was m o v e d f r o m r a d i a l to the u l n a r a s p e c t across the forearm. In 3 subjects, s u b s t a n t i a l C M A P s were elicited only w h e n the c o n t a c t i n g edge (12 m m thick) was directly over the nerve or when the M C was ins i n u a t e d b e t w e e n the t e n d o n s of the flexor carpi r a d i a l i s a n d p a l m a r i s longus (Fig. 7), that is, the closer the M C was to the m e d i a n nerve, the greater the C M A P a m p l i t u d e . M o v i n g the M C in a r a d i a l o r u l n a r d i r e c t i o n a w a y f r o m the m e d i a n nerve led to a p r e c i p i t o u s d r o p in C M A P a m p l i t u d e . (6) Evaluation o f C M A P amplitude, area, and ulnar nerve coactivation In 6 subjects, the m e d i a n nerve was s t i m u l a t e d with 3 M C o r i e n t a t i o n s ( o r t h o g o n a l - l o n g i t u d i n a l , 45 ° lateral t i l t - l o n g i t u d i n a l , t a n g e n t i a l edge) at the e l b o w a n d wrist, s i m u l t a n e o u s l y r e c o r d i n g from t h e n a r a n d h y p o t h e n a r muscles (Fig. 8). I n gen-
PERIPHERAL MOTOR NERVE STIMULATION BY MC eral, the largest M / E C M A P amplitude ratios were obtained with 45 ° lateral tilt-longitudinal and tangential edge orientations. Unexpectedly, the C M A P amplitudes evoked by M C stimuli at the wrist were sometimes slightly greater than those evoked by supramaximal cathodal stimuli (e.g., 6% in Fig. 9, top left and middle records, and Table I). In subjects selected for M / E amplitude ratios ~< 1 (Table II), the corresponding area ratios were more often >/1, because the duration of the C M A P negative phase was usually increased with M C stimulation. W h e n with supramaximal stimulation, the M / F ratio is greater than 1, ulnar coactivation by the M C must be excluded as an explanation. In Fig. 9, top left and middle records, the h y p o t h e n a r C M A P shows no evidence of local activation, the small initially positive deflection reflecting distant activity. True coactivation of h y p o t h e n a r muscles is readily elicited by moving the tangential edge from the optimal site over the median nerve closer to the ulnar border (right record). However, the possibility remains that the h y p o t h e n a r C M A P is not an adequate sample of muscles activated by ulnar m o t o r axons. F o r example, if ulnar m o t o r M.C. LON61TUDINAL, ORTHOGONAL
BIPOLAR - C A T H O D A L STI MULATION
5my //
',\
=
- :
~'/
- t
\\ \
8LE M.C. LONGITUDINAL, 4 5 °
f
t
531 ~ ELECTRICAL MEDIAN NERVE
T..A.
- TANGENTIAL EDGE - OPTIMAL LOCATION 13ram from MEDIAN NERVE ULNAR BORDER M.C. (lO0°/=) M.C (fO0%)
--
i} ov TANGENTIAL EDGE OVER ULNAR NERVE - - ~ M.C. (90-100%)
M.C. (22-40%)
T II 2 rnV
POLLICIS
--
2 mV
Fig. 9. Thenar CMAPs without ulnar coactivation, elicited by electrical and MC stimulation of distal median nerve. Top superimposed records: left, responses to several including supramaximal intensities; middle, maximum response to MC at optimal location; right, ulnar coactivation when MC moved closer to ulnar border. Throughout top records, MC tilted towards radial border. Bottom superimposed records: left, maximal responses to MC stimulation at 90 and 100% output, MC tilted towards ulnar border; right, threshold responses to MC stimulation at 22-40% output, progressing in 2% steps. Gains increased 10 ×, compared to left records.
axons innervating a d d u c t o r pollicis had the lowest threshold, coactivation of the ulnar nerve could occur without a detectable h y p o t h e n a r C M A P . This possibility was evaluated by simultaneously recording from a d d u c t o r pollicis and h y p o t h e n a r muscles with the M C over ulnar nerve and tilted away from the median nerve. Even at high recording sensitivity (Fig. 9, b o t t o m right), no evidence was f o u n d of a lower threshold for a d d u c t o r pollicis.
M.C. TANGENTIAL EDGE
Discussion
Fig. 8. Simultaneous recordings from thenar and hypothenar muscles with electrical and MC stimulation of median nerve at elbow and at wrist to show lack of ulnar coactivation.
R u s h t o n (1927) showed that electrical stimulation of peripheral nerve is most efficient when the imposed direction of current flow is aligned with the long axis of the nerve being stimulated. Similarly, we f o u n d the induced voltages and C M A P s were largest when the M C - i n d u c e d electrical field was orientated parallel to the long axis of a metallic c o n d u c t o r (Fig. 2) and the median nerve, respectively (Fig. 3). Some control of the orientation of
532 the induced electric field is obtainable with the MC because, when it is flat against the volume conductor, the induced current flows in concentric rings parallel to the plane of the MC and in an opposite direction to that in the MC. However, when the MC is orthogonal to the volume conductor, not only is the distribution of current flow more restricted, but it flows in the same circular direction in the volume conductor as in the MC. Thus, the current in the MC and the volume conductor flow in opposite directions at the edge of the volume conductor (Fig. 6). Given an appropriate orientation, usually the MC can maximally excite the median nerve without activation of ulnar nerve (Fig. 8 and Table I). However, supramaximal stimulation cannot be guaranteed in all individuals with the currently available MCs and stimulator outputs. The marked attenuation in CMAP when an orthogonal-longitudinally oriented MC is moved laterally away from the underlying median nerve (Fig. 7) demonstrates the focal properties of the orthogonallongitudinal orientation. Frequently, the largest CMAPs occurred when the longitudinally orientated MC was tilted 45 ° possibly because: (1) the flat outer edge of the MC was insinuated between tendons and thus was closer to the underlying median nerve, and (2) a larger fraction of the magnetic flux is appropriately orientated to enter the volume conductor. A potential disadvantage of stimulation by the MC, even when orientated orthogonally, is the imprecision in defining exactly where the distally propagating nerve impulse arises. In different individuals, when using supramaximal stimuli, the site varied 2-15 mm from the midpoint of the contacting edge. Because of the longer conduction distance, this error would not be serious when stimulating nerve roots, but would be in measuring distal motor latencies. A surprising result was the proximity, less than 5 - 6 mm, of the virtual cathode to the anode induced by an orthogonal-longitudinally oriented MC. By contrast, a much greater interpolar distance, at least 20 mm, is typically used with electrical stimulation. As a result, the direction of current flow in the orthogonally orientated MC is much less critical for distal peripheral nerve stimulation than with conventional electrical stimula-
P.J. MACCABEEET AL. tion. Another unexpected finding was the larger CMAP amplitude (Fig. 9 and Table I) and area (Table II), that may be obtained with MC versus clearly supramaximal electrical stimulation. Possible explanations include: (1) The MC elicits double discharges at short intervals in some motor units. However single fiber muscle action potential recordings did not reveal double discharges. (2) The MC excites both median and ulnar nerves. However, ulnar coactivation was excluded in recordings from hypothenar muscles. No evidence was found that ulnar motor axons supplying other muscles (e.g., adductor pollicis) had a lower threshold than those supplying hypothenar muscle. Therefore the hypothenar muscle appeared to be a valid test for ulnar coactivation. (3) The MC excites motor axons in a temporal pattern which results in less phase cancellation between extracellular action potentials generated by different motor units. For example, where a motor end-plate region of a muscle fiber of motor unit A overlies a portion of a muscle fiber of motor unit B remote from its motor end-plate, phase cancellation of the extracellular electrical signals would tend to occur if both motor axons were synchronously excited. Summation would be promoted if the timing of excitation of the two units was changed, possibly because of the temporal characteristics of the MC-induced electric field. Although the two MCs tested had different shapes, we did not find any marked differences in their properties when stimulating peripheral motor nerves. The orthogonal-longitudinal orientation used in our study was advantageous for focal stimulation despite coupling only a small fraction of the magnetic flux to the volume conductor. Nevertheless, further improvement in MC shape is desirable to increase the focality of stimulation. Furthermore, a larger output is required to ensure supramaximal nerve stimulation in all subjects, an essential requirement for clinical use.
Appendix Calculation of current density (1) The MC was assumed to rest over a hemiinfinite volume conductor. A sense coil (3 loops of
PERIPHERAL MOTOR NERVE STIMULATION BY MC
wire, 1 cm in diameter) was held parallel to the flat plane of the MC at various locations consisting of a grid of spatial points surrounding the MC. At each point the sense coil measured the rate of change of flux (dq,/dt), which is an indirect measure of the peak flux density (/~). (2) Since current flows in concentric circles or annular disks parallel to the MC, and the voltage around any single disc is proportional to total flux (H, where H = B * 7rr2), then the voltage (V) is given by: V = K ~ /~iri2 i~1
where r i = radius of each disk; K = constant; n = number of disks. (3) The resistance (w) is defined as the path length in tissue (2¢r * radius) * resistivity. Assuming unity cross-sectional area for the conductor, the resistance is therefore proportional to the radius (r) of each disk, or w i = r i. (4) Current density at a given radius is then computed from Ohm's law
K i-
e w
i
Bir i=l ri
(5) The current density was computed for all points in the grid and those points with the same current density were plotted as an isocurrent density line (Fig. 1). The authors gratefully acknowledge the technical assistance of Drs. Bradley P. Grayum, Shrinath Kamat, Ken Kaplove,
533 Krishna Nalluri, Vasanthi Nalluri and Lyzette Velazquez. Sam Polaniecki, of the Dept. of Scientific and Medical Instrumentation, helped in the measurement of voltages induced in wire. We thank also Ms Madge Edmead for secretarial assistance.
References Amassian, V.E., Cadwell, J., Cracco, R.Q. and Maccabee, P.J. Focal cerebral and peripheral nerve stimulation in man with the magnetic coil. J. Physiol. (Lond.), 1987, 390: 24P. Barker, A.T., Jalinous, R. and Freeston, I.L. Non-invasive stimulation of human motor cortex. Lancet, 1985, i: 1106-1107. Barker, A.T., Freeston, I.L, Jalinous, R. and Jarrat, J.A. Clinical evaluation of conduction time measurements in central motor pathways using magnetic stimulation of the human brain. Lancet, 1986, i: 1325-1326. Bickford, R.G. and Fremming, B.D. Neuronal stimulation by pulsed magnetic fields in animals and man. In: Digest of the 6th Int. Conf. on Medical Electronics and Biological Engineering, Tokyo, 1965: 112. Hess, C.W., Mills, K.R. and Murray, N.M.F. Measurement of central motor conduction in multiple sclerosis by magnetic brain stimulation. Lancet, 1986, i: 355-358. Hodgkin, A.L and Huxley, Aft. A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. (Lond.), 1952, 117: 500-544. Maccabee, P.J., Amassian, V.E. and Cracco, R.Q. A comparison of electrical and magnetic coil stimulation of peripheral nerves in man. Neurosci. Abstr., 1987a, 13: 1265. Maccabee, P.J., Amassian, V.E. and Cracco, R.Q. Focal stimulation of peripheral nerves using the magnetic coil. Muscle Nerve, 1987b, 10: 642-643. Rushton, W.A.H. Effect upon the threshold for nervous excitation of the length of nerve exposed and the angle between current and nerve. J. Physiol. (Lond.), 1927, 63: 357-377. Wiederholt, W.C. Threshold and conduction velocity in isolated mixed nerves. Neurology, 1970, 20: 347-352.