We compared magnetic stimulation using different coil designs (2 rounded coils and a butterfly-prototypecoil) with electrical stimulation of the median and ulnar nerves in 5 normal subjects. Using magnetic stimulation we were able to record technically satisfactory maximal sensory and motor responses only with the butterfly coil. Submaximal electrical stimuli preferentially activated sensory rather than motor axons, but submaximal magnetic stimuli did not. The onset latency, amplitude, area and duration of responses elicited electrically or magnetically with the butterfly coil during routine sensory and motor nerve conduction studies were similar, and motor and sensory conduction velocities were comparable when studied over long segments of nerve. However, the motor conduction velocities with magnetic and electrical stimulation differed by as much as 18 m/sec in the across-elbow segment of ulnar nerve. Thus, recent developments in magnetic stimulator design have improved the focality of the stimulus, but the present butterfly coil design cannot replace electrical stimulation for the detection of focal changes in nerve conduction velocity at common entrapment sites, such as in the across-elbow segment of the ulnar nerve. Key words: magnetic stimulation peripheral nerve nerve conduction study MUSCLE & NERVE 13~957-963 1990
A COMPARISON OF MAGNETIC AND ELECTRICAL STIMULATION OF PERIPHERAL NERVES RICHARD K. OLNEY, MD, YUEN T. SO, MD, PhD, DOUGLAS S. GOODIN, MD, and MICHAEL J. AMINOFF, MD, FRCP
Stimulation of the peripheral nervous system by a magnetic field was first demonstrated in a frog nerve-muscle preparation in 1959,5 and soon thereafter in mixed human nerve.2 These stimulators generated sinusoidal magnetic fields capable of producing observable muscle contractions. Magnetic stimulation of the peripheral nervous system with successful recording of compound muscle action potentials was first reported in 1982.' This was accomplished by developing a stimulator that generated a single pulsed magnetic field. Such stimulators are now being marketed commercially for routine use on the peripheral
From the Department of Neurology, School of Medicine, University of California, San Francisco, San Francisco, California. Acknowledgment: We are grateful to Cadwell Laboratories for providing us with the prototype butterfly coil for magnetic stimulation. Presented in part at the American Association of Electromyography and Electrodiagnosis, Washington, DC, September 15, 1989. Address reprint requests to Dr. Richard K. Olney, Box 01 14, Rm M-794, Department of Neurology, University of California, San Francisco, San Francisco, CA 94143. Accepted for publication November 13, 1989 CCC 0148-639W90/0100957-07 $04.00 0 1990 John Wiley & Sons, Inc.
Magnetic Stimulation
nervous system, but their clinical utility has not been established. Previous reports have evaluated round and pointed coils. Whereas some authors have found that one type of round coil was unable to selectively stimulate peripheral nerves supramaximally,3 others have reported success in focally stimulating the peripheral nervous system with a different type of round coil.'*' We report our experience with two types of these coils and with a prototype butterfly-shaped coil. METHODS
We studied 5 subjects with a mean age of 37 years (range: 34 to 4 1 years), all of whom gave informed voluntary consent to participate in these studies. They denied any symptoms of neuromuscular disease and had no history of any chronic medical illness.
Subjects.
ElectrophyslologlcalTechniques. Compound muscle action potentials (CMAPs) were recorded with active surface electrodes over abductor pollicis brevis (APB) and abductor digiti minimi (ADM) muscles, and the reference surface electrodes over the tendons of these muscles, after stimulation of
MUsZLE & NERVE
October 1990
957
the median and ulnar nerves, respectively. Sensory nerve action potentials of the median and ulnar nerves were recorded with ring electrodes around the second and fifth digits, respectively, with the active electrode separated from the reference by 4 cm. These 4 compound action potentials were recorded concomitantly, using a standard electromyograph system with 4 amplifiers (Dantec 1500, Dantec Electronics Inc., Allendale, NJ). In all subjects, supramaximal electrical stimuli were delivered percutaneously to the median nerve at the wrist (8 cm proximal to the active recording electrode over APB) and at the elbow (antecubital fossa), and to the ulnar nerve at the wrist (8 cm proximal to the active recording electrode over ADM), below the elbow (5 cm distal to the medial epicondyle) and above the elbow (10 cm proximal to the below- elbow stimulation site) with a constant current stimulator (Dantec 15 E 07). Magnetic stimuli were delivered to the same 5 sites with 3 different stimulator heads. T h e intensities of these magnetic stimuli either were sufficient to elicit a maximal response or were 100% of stimulator output. The Novometrix Model 200 Magstim (marketed by Nicolet Biomedical Instruments, Madison, WI) was used on 3 subjects with the round stimulator coil head; this coil with an outside diameter of 14 cm has a maximum field strength of 1.5 Tesla. The Cadwell MES-10 (Cadwell Laboratories, Kennewick, WA) was used on 3 subjects with a rounded stimulator coil; this coil has an outside diameter of 9.5 cm for the round portion and a 1.5 cm pointed extension at its tip. This coil, that will be referred to as a round coil in the remainder of this article, has a maximum field strength of 2.0 Tesla. T h e third stimulator used was a Cadwell prototype “butterfly” head driven by the same Cadwell MES-10. This stimulator head has 2 round coils that each have an outside diameter of 5 cm; these 2 round coils are joined at one slightly flattened edge with the planes of the coils angled 18 degrees and with the power cable attaching at this line of intersection, giving an appearance similar to a butterfly. This latter stimulator with a maximum field strength of 2.0 Tesla was used on all 5 subjects. Selectivity of median nerve stimulation was determined by the absence of an ulnar sensory nerve action potential after averaging 4 to 8 responses and the absence of a hypothenar compound muscle action potential. Selectivity of ulnar nerve stimulation was determined primarily by the absence of a median sensory nerve action potential, but the size of the volume conducted thenar response
958
Magnetic Stimulation
with magnetic stimulation was also assessed and compared with the thenar response obtained following electrical stimulation of the ulnar nerve to help determine the selectivity of the former. For quantitative comparisons, the compound action potential waveforms were transferred from the Dantec 1500 through a parallel digital interface into an Apple IIe computer. The onset latency of the potentials, and the amplitude, area, and duration of their negative phase were measured by the computer, after the onset and termination of the negative phase were marked with operator-controlled cursors. The effect of stimulus intensity on response size was assessed for the median nerve at the wrist in all 5 subjects, using the electrical stimulator and the butterfly magnetic head. First, maximal responses were obtained using stimuli that were clearly supramaximal. The stimulation intensity was then reduced until the intensity was found that just elicited maximal motor and sensory responses, so that any further reduction in stimulation intensity produced submaximal responses. This intensity was labeled loo%, because it was maximal but not supramaximal. Finally, motor and sensory responses were recorded with stimulation at intensities of loo%, 75%, 50%, 25%, and 0%. The amplitude and area of each of these responses was determined as described above, and analyzed as the percent of the maximal amplitude and area, so that motor and sensory as well as magnetic and electrical data could be compared in similar units. T h e effect of surface position for stimulation on response size was assessed for the median nerve at the wrist in 2 subjects. First, electrical stimulation was used to localize the transverse point at the wrist where a submaximal stimulus produced the largest motor response. This was labeled the “0” point. Second, 0.5 cm intervals were marked across the wrist, with “positive” designating distance toward the lateral (radial) border and “negative” designating more medial (ulnar) displacements. Third, the maximal stimulus intensity was determined at the “ 0 point for electrical stimulation, as previously described, and the amplitude of the CMAP over APB elicited by this stimulus was considered 100% maximum. Fourth, the amplitude of the CMAP over APB was measured and expressed as a percentage of the 100% maximum response, after stimulation at each 0.5 cm interval. Finally, the last 2 steps were repeated using magnetic stimulation with the Cadwell round and butterfly coils.
MUSCLE & NERVE
October 1990
The results of routine sensory and motor nerve conduction studies were compared quantitatively on all 5 subjects, using electrical stimulation and the Cadwell butterfly coil. The midpoint of the line connecting the leading edge of the 2 round coils was positioned at the same point as the electrical cathode. The differences in the motor latencies obtained with magnetic and electrical stimulation at the same site were converted into estimated distances between the depolarization loci, based on the electrically determined motor nerve conduction velocity for the forearm segment. Comparable calculations were also performed on the differences in sensory latencies using the electrically determined sensory conduction velocity. Descriptive statistics and the Mann-Whitney nonparametric test were used for statistical analysis.
Statistical Analysis.
RESULTS
We compared different angles of orientation for each of the round coils, using the terminology proposed by others.6 We were able to obtain maximal amplitudes and areas of compound muscle action potentials in comparison with those obtained with electrical stimulation using tangential-edge magnetic stimulation. By contrast, using the orthogonal-longitudinal orientation described by others,6 we were unable to achieve maximal responses. These responses sometimes approached 80% maximal following stimulation at the wrist with the Cadwell round coil, but were less than 10% maximal following stimulation in the region of the elbow with either round coil. The tilted-longitudinal orientation sometimes produced amplitudes and areas similar to tangential-edge, but the intensity of stimulation varied markedly with minor changes in tilt-angle and surface position, so that reproducible maximal responses could only be obtained with the tangential-edge orientation.
Orientation of Round Coils.
We were unable to obtain selective maximal activation of either the median o r ulnar nerves in 3 subjects with any orientation of the Novometrix coil at the wrist o r elbow. We were, however, able to obtain maximal responses with selective stimulation of at the wrist with the the median and ulnar Cadwell round coil, and % ’’ to 95% responses with selective stimulation of the median nerve at the elbow. We were unable to obtain se-
Comparisonof Three Magnetic Stimulator Coils.
Magnetic Stimulation
lective, near-maximal stimulation of the ulnar nerve below or above the elbow. The Cadwell butterfly coil, by contrast, permitted selective stimulation of the nerves at intensities sufficient to elicit maximal responses at all 5 sites. Both Cadwell coils, but not the Novometrix coil, had sufficiently low magnetic field interference to allow the recording of digital nerve action potentials following stimulation at the wrist. The Effect of Surface Position for Stimulation on Response Size. The effective field of stimulation for
the 2 Cadwell coils was compared with electrical stimulation by plotting the amplitude of compound muscle action potential (expressed as percent maximum) obtained when the point of stimulation was displaced transversely from the point of maximal response (Fig. 1). Using this method, the effective field of stimulation for the butterfly magnet was larger than for the electrical stimulus, but was smaller than for the round coil. Thus, there is greater spatial selectivity for the magnetic field produced by the butterfly coil in comparison with the round coil, although neither had the narrow effective field achieved using electrical stimulation. Stimulus Intensity Study. T h e effects of changing the stimulus intensity on the amplitude of the response are shown in Figure 2. The effects on area were the same as those on amplitude. Compared to maximal responses, the amplitude and area of the sensory responses are greater than for the motor responses with electrical, but not magnetic, stimulation at 25% and 50% intensity (P < 0.05).
-
100
‘
80
.-
0
Electrical
A Round
X
s
60
a
U
.--
* 3
40
!
20
0 -4
-2
0
2
4
Displacement (em) FIGURE 1. The effect of surface position for stimulation on response size. The point at the wrist where submaximal electrical stimulation produces the largest response is “0”;displacement is positive if lateral (radial) and negative if medial (ulnar).
MUSCLE & NERVE
October 1990
959
100
-
h
E
g
.-X
80-
60-
E 40._ V
-
c)
-€
20A
" T
0
/ _
25
Magnetic: A Magnetic: Electrical: A Electrical: 0
. 50
.
. 75
.
Motor Sensory Motor Sensory
. lo0
Intensity (% maximum) FIGURE 2. The effects of changing the stimulus intensity on the amplitude of the response is illustrated for both magnetic (open points) and electrical (filled points) stimulation on sensory (triangular points) and motor (square points) responses.
3 mm behind the point activated by the electrical stimulus. Since the leading edge of the butterfly magnet is 10 mm in front of the junction point of its 2 round coils, the magnetic stimulus usually activates the peripheral nerve between these 2 points of the coil. T h e sensory and motor conduction velocities were also in reasonably close agreement, except for the short across-elbow segment of the ulnar nerve (Fig. 3). The sensory conduction velocities derived from the two types of stimulation did not differ by more than 5 m/sec. The motor conduction velocities differed by less than 3 m/sec for the above-elbow to wrist segment of the ulnar nerve, by less than 7 m/sec for the belowelbow to wrist segment of the ulnar nerve and the median nerve, but by up to 18 m/sec for the across-elbow segment of the ulnar nerve. DISCUSSION
The differences in amplitude and area of sensory and motor responses are not significant with either magnetic or electrical stimuli at 75% of maximal stimulus intensity (P > 0.05). Motor responses to magnetic and electrical stimulation were comparable (both in amplitude and area) at submaxima1 stimulus intensities. By contrast, submaximal electrical stimulation produced a larger sensory response than the comparable magnetic stimulus at 25% and 50% intensity (P < 0.05), but not at 75% (P > 0.05). Thus, submaximal electrical stimuli preferentially activate sensory axons, whereas submaximal magnetic stimuli d o not. Selective supramaximal stimulation was successfully performed at all 5 sites with both the butterfly magnet and the electrical stimulator, and no significant differences were found in amplitude, area or duration of motor and sensory responses obtained with the two types of stimuli (Table 1). Similarly, no statistically significant differences were detected in response latency or conduction velocity during sensory and motor studies. There was close agreement between the latencies obtained with the two types of stimulation for both sensory and motor responses (as illustrated in Fig. 3 for the motor latencies). T h e latencies obtained with electrical stimulation were 0.04 0.1 msec shorter (mean ? 1 SD) than those obtained with the butterfly coil, and the differences varied from -0.2 to 0.2 msec. By converting these latency differences into distance, the leading edge of the magnet can be estimated to typically activate peripheral nerve Routine Nerve Conduction Studies.
*
960
Magnetic Stimulation
In this article we have compared the clinical utility of three magnetic coils (two different round coils and one butterfly-shaped coil) with standard electrical stimulation in normal subjects. The primary aim was to assess their suitability for the performance of routine median and ulnar nerve conduction studies. We consider the minimal requirement for a suitable stimulator to be the ability to provide selective, supramaximal stimulation of these nerves at different points along their course with a minimum of patient discomfort. Our experience with the Novametrix round coil and stimulator was similar to that of other^,^ in that it did not fulfill this minimum requirement for either nerve at the wrist or in the region of the elbow with any orientation of the coil. The Cadwell round coil did deliver supramaximal stimulation in the tangential-edge orientation, but it was selective only at some sites. In contrast, the Cadwell butterfly-prototype coil consistently provided selective supramaximal stimulation at all sites. The better performance of the butterfly-shaped, as opposed to the'round-shaped, coil was due to improved focality rather than strength of the magnetic stimulus. As illustrated (see Fig. l), the butterfly design provided a stimulus that was more focal than the round coil, but less focal than the electrical stimulator. Previous authors have also compared magnetic and electrical stimulation of the motor axons of peripheral nerves and found results similar to ours. 173r6 However, unlike these other authors, we also evaluated the performance of magnetic stimulators during routine sensory nerve conduction studies. Using the Novametrix stimulator, we were
MUSCLE & NERVE
October 1990
Table 1. The results of routine nerve conduction studies are expressed as the mean
Test
Place of stimulation
Nerve
Amplitude (mV or uV) Median Ulnar Area (mV . msec or uV . msec) Median Ulnar
Duration (msec) Median Ulnar
Onset Latency (msec) Median Ulnar
Motor
Sensory Magnetic
Electrical
Magnetic
Wrist Elbow Wrist Below-elbow Above-elbow
9.4f 1.7 9.0f 1.8 10.5f 2.2 9.7f 2.9 9.4f 2.8
9.5-c 2.2 8.9f 2.3 10.4 f 2.5 10.4f 2.0 9.9f 1.8
36f 18
35 f 19
31
30 f 15
Wrist Elbow Wrist Below-elbow Above-elbow
26.2f 3.9 25.7f 6.4 31.6 f 5.4 29.7f 8.5 28.6f 8.1
25.7f 5.7 23.8f 6.6 31.3 f 5.5 32.1 f 5.5 31.4f 5.2
22 f 9
Wrist Elbow Wrist Below-elbow Above-elbow
4.7f 0.5 4.8 f 0.2 4.9f 0.5 5.3 f 0.7 5.4f 0.6
4.7 f 0.4 4.8f 0.4 4.9 f 0.5 5.2f 0.5 5.4f 0.5
1.2f 0.1
1.2f 0.1
1.2 t 0.2
1.2 f 0.3
-
-
Wrist Elbow Wrist Below-elbow Above-elbow
3.3f 0.4 7.2f 0.7 2.9f 0.1 6.0f 0.5 7.7f 0.7
3.3f 0.3 7.1 f 0.7 2.9 f 0.1 6.0f 0.4 7.5 f 0.7
2.3f 0.3
2.3f 0.3
Distal Forearm Distal Below-elbow to wrist Across-el bow Above-elbow to wrist
unable to record sensory responses with stimulation at the wrist due to the large stimulus artifact produced by the magnetic field. With either of the Cadwell coils, by contrast, we were able to obtain selective, maximal sensory responses with stimulation of the median and ulnar nerves at the wrist. Comparing the ability of magnetic and electrical stimulators to activate sensory and motor axons, we found that only electrical stimulation preferentially activated sensory axons at some submaximal stimulus intensities. The reason for this difference between magnetic and electrical stimuli in the fiber population activated by submaximal stimuli is uncertain. This phenomenon does not affect routine sensory studies, because these are performed at supramaximal stimulus intensities. However, because selective stimulation of sensory axons at submaximal intensity is necessary in the performance of H-reflex studies, electrical stimulation may be the method of choice in these circumstances.
Magnetic Stimulation
1 SD.
Electrical
Conduction Velocity (mkec) Median Ulnar
f
f
12
11
21 c 9
19 2 8
-
-
-
-
-
-
2.5f 0.5
2.5f 0.4
-
-
64 2 5 -
65 f 2 68 & 16 65 f 5
f
-
65 f 5
67 f 5 60 f 9 64 f 6
21
-
-
-
-
-
65 f 2
-
-
60 f 8
-
-
64 f 3
-
59 f 6 -
We also compared the onset latencies of the sensory and motor response and the calculated conduction velocities derived using the butterfly coil with those from electrical stimulation. As illustrated in Figure 3 for the motor responses, the onset latencies were quite comparable, whereas the calculated conduction velocities had a wider range of differences, especially for the short acrosselbow segment of the ulnar nerve. Statistically significant differences were not found in the conduction velocities for any nerve segment. We also compared the differences in our conduction velocities derived from magnetic and electrical stimulation with the differences in conduction velocities derived from electrical stimulation on two separate occasions. Under the latter circumstances, others have reported differences in motor conduction velocities of 5% to 15%, or 5 to 11 r n / ~ e c , ~ and we have observed intertrial differences in antidromic sensory conduction velocities of up to 14 m/sec (unpublished data). Based on this range of
MUSCLE & NERVE
October 1990
961
Median Nerve
B
A
Latency (msec): Electrical Stimulation
CV (mlsec): Electrical Stimulation
Ulnar Nerve
“
I
0 Across Elbow A Above-Elbow to Wrist I /
ow. 0
, 2
. , . , . , 4
6
0
.
1
1
0
Latency (msec): Electrical Stimulation
V isi 25
. . . .
,
. . . . ,
50
75
.
.
.
..I
100
CV (mlsec):Electrical Stimulation
FIGURE 3. The results of motor latencies (A and C) and nerve conduction velocities (B and D) for median (A and B) and ulnar (C and D) nerves are compared for the butterfly magnet and electrical stimuli.
differences, magnetic stimulation with the butterfly coil provides findings comparable to electrical stimulation for sensory and motor conduction velocities for most segments, where we obtained differences of less than 7 m/sec. However, even the current butterfly stimulator cannot be used as a substitute for electrical stimulation when performing routine motor conduction studies on short segments, such as the across-elbow segment of the ulnar nerve, where we observed differences of up to 18 m/sec. In summary, recent developments in magnetic stimulator design have improved the focality of the stimulus, so that selective supramaximal stimuli may be delivered to commonly studied peripheral nerves in the upper limbs. Our study demon-
strates that the results of magnetic stimulation are sufficiently comparable to those obtained with electrical stimulation that the use of magnetic stimulation may be a reasonable alternative to electrical stimulation in certain situations. However, we have not demonstrated any particular advantage of magnetic stimulation. Futhermore, the calculated motor nerve conduction velocities that are based on magnetic stimulation have a broader range of variability than velocities derived from electrical stimulation. This presently limits the clinical utility of routine magnetic stimulation in detecting focal changes in nerve conduction velocity at common entrapment sites, such as the across-elbow segment of the ulnar nerve.
REFERENCES
1. Amassian VE, Maccabee PJ, Cracco RQ: Focal stimulation of human peripheral nerve with the magnetic coil: A comparison with electrical stimulation. Ex$ Neurol 1989; 103~282-289.
962
Magnetic Stimulation
2. Bickford RG, Guidi M, Fortesque P, Swenson M: Magnetic stimulation of human peripheral nerve and brain: response enhancement by combined magnetoelectrical technique. Neurosurgery 1987;20:110- 116.
MUSCLE & NERVE
October 1990
3. Evans BA, Litchy WJ, Daube JR: The utility of magnetic stimulation for routine peripheral nerve conduction studies. Muscle N m e 1988;11:1074-1078. 4. Halar EM, Venkatesh B: Nerve conduction velocity measurements: improved accuracy using superimposed response waves. Arch Phys Med Rehabil 1976;57:451-457. 5. Kolin A, Brill NQ, Broberg PJ: Stimulation of irritable tissues by means of an alternating magnetic field. Proc SOLExp Biol Med 1959;102:251-253.
Magnetic Stimulation
6. Maccabee PJ, Amassian VE, Cracco RQ, Cadwell JA: An analysis of peripheral motor nerve stimulation in humans using the magnetic coil. Electroencephalogr Clin Neurophysiol 1988;70:524-533. 7. Polson MJR, Barker AT, Freeston IL: Stimulation of nerve trunks with time-varying magnetic fields. Med Biol Eng Cm@t 1982;20:243-244.
MUSCLE & NERVE
October 1990
963
A COMPARISON OF MAGNETIC AND ELECTRICAL STIMULATION OF PERIPHERAL NERVES RICHARD K. OLNEY, MD, YUEN T. SO, MD, PhD, DOUGLAS S. GOODIN, MD, and MICHAEL J. AMINOFF, MD, FRCP
Stimulation of the peripheral nervous system by a magnetic field was first demonstrated in a frog nerve-muscle preparation in 1959,5 and soon thereafter in mixed human nerve.2 These stimulators generated sinusoidal magnetic fields capable of producing observable muscle contractions. Magnetic stimulation of the peripheral nervous system with successful recording of compound muscle action potentials was first reported in 1982.' This was accomplished by developing a stimulator that generated a single pulsed magnetic field. Such stimulators are now being marketed commercially for routine use on the peripheral
From the Department of Neurology, School of Medicine, University of California, San Francisco, San Francisco, California. Acknowledgment: We are grateful to Cadwell Laboratories for providing us with the prototype butterfly coil for magnetic stimulation. Presented in part at the American Association of Electromyography and Electrodiagnosis, Washington, DC, September 15, 1989. Address reprint requests to Dr. Richard K. Olney, Box 01 14, Rm M-794, Department of Neurology, University of California, San Francisco, San Francisco, CA 94143. Accepted for publication November 13, 1989 CCC 0148-639W90/0100957-07 $04.00 0 1990 John Wiley & Sons, Inc.
Magnetic Stimulation
nervous system, but their clinical utility has not been established. Previous reports have evaluated round and pointed coils. Whereas some authors have found that one type of round coil was unable to selectively stimulate peripheral nerves supramaximally,3 others have reported success in focally stimulating the peripheral nervous system with a different type of round coil.'*' We report our experience with two types of these coils and with a prototype butterfly-shaped coil. METHODS
We studied 5 subjects with a mean age of 37 years (range: 34 to 4 1 years), all of whom gave informed voluntary consent to participate in these studies. They denied any symptoms of neuromuscular disease and had no history of any chronic medical illness.
Subjects.
ElectrophyslologlcalTechniques. Compound muscle action potentials (CMAPs) were recorded with active surface electrodes over abductor pollicis brevis (APB) and abductor digiti minimi (ADM) muscles, and the reference surface electrodes over the tendons of these muscles, after stimulation of
MUsZLE & NERVE
October 1990
957
the median and ulnar nerves, respectively. Sensory nerve action potentials of the median and ulnar nerves were recorded with ring electrodes around the second and fifth digits, respectively, with the active electrode separated from the reference by 4 cm. These 4 compound action potentials were recorded concomitantly, using a standard electromyograph system with 4 amplifiers (Dantec 1500, Dantec Electronics Inc., Allendale, NJ). In all subjects, supramaximal electrical stimuli were delivered percutaneously to the median nerve at the wrist (8 cm proximal to the active recording electrode over APB) and at the elbow (antecubital fossa), and to the ulnar nerve at the wrist (8 cm proximal to the active recording electrode over ADM), below the elbow (5 cm distal to the medial epicondyle) and above the elbow (10 cm proximal to the below- elbow stimulation site) with a constant current stimulator (Dantec 15 E 07). Magnetic stimuli were delivered to the same 5 sites with 3 different stimulator heads. T h e intensities of these magnetic stimuli either were sufficient to elicit a maximal response or were 100% of stimulator output. The Novometrix Model 200 Magstim (marketed by Nicolet Biomedical Instruments, Madison, WI) was used on 3 subjects with the round stimulator coil head; this coil with an outside diameter of 14 cm has a maximum field strength of 1.5 Tesla. The Cadwell MES-10 (Cadwell Laboratories, Kennewick, WA) was used on 3 subjects with a rounded stimulator coil; this coil has an outside diameter of 9.5 cm for the round portion and a 1.5 cm pointed extension at its tip. This coil, that will be referred to as a round coil in the remainder of this article, has a maximum field strength of 2.0 Tesla. T h e third stimulator used was a Cadwell prototype “butterfly” head driven by the same Cadwell MES-10. This stimulator head has 2 round coils that each have an outside diameter of 5 cm; these 2 round coils are joined at one slightly flattened edge with the planes of the coils angled 18 degrees and with the power cable attaching at this line of intersection, giving an appearance similar to a butterfly. This latter stimulator with a maximum field strength of 2.0 Tesla was used on all 5 subjects. Selectivity of median nerve stimulation was determined by the absence of an ulnar sensory nerve action potential after averaging 4 to 8 responses and the absence of a hypothenar compound muscle action potential. Selectivity of ulnar nerve stimulation was determined primarily by the absence of a median sensory nerve action potential, but the size of the volume conducted thenar response
958
Magnetic Stimulation
with magnetic stimulation was also assessed and compared with the thenar response obtained following electrical stimulation of the ulnar nerve to help determine the selectivity of the former. For quantitative comparisons, the compound action potential waveforms were transferred from the Dantec 1500 through a parallel digital interface into an Apple IIe computer. The onset latency of the potentials, and the amplitude, area, and duration of their negative phase were measured by the computer, after the onset and termination of the negative phase were marked with operator-controlled cursors. The effect of stimulus intensity on response size was assessed for the median nerve at the wrist in all 5 subjects, using the electrical stimulator and the butterfly magnetic head. First, maximal responses were obtained using stimuli that were clearly supramaximal. The stimulation intensity was then reduced until the intensity was found that just elicited maximal motor and sensory responses, so that any further reduction in stimulation intensity produced submaximal responses. This intensity was labeled loo%, because it was maximal but not supramaximal. Finally, motor and sensory responses were recorded with stimulation at intensities of loo%, 75%, 50%, 25%, and 0%. The amplitude and area of each of these responses was determined as described above, and analyzed as the percent of the maximal amplitude and area, so that motor and sensory as well as magnetic and electrical data could be compared in similar units. T h e effect of surface position for stimulation on response size was assessed for the median nerve at the wrist in 2 subjects. First, electrical stimulation was used to localize the transverse point at the wrist where a submaximal stimulus produced the largest motor response. This was labeled the “0” point. Second, 0.5 cm intervals were marked across the wrist, with “positive” designating distance toward the lateral (radial) border and “negative” designating more medial (ulnar) displacements. Third, the maximal stimulus intensity was determined at the “ 0 point for electrical stimulation, as previously described, and the amplitude of the CMAP over APB elicited by this stimulus was considered 100% maximum. Fourth, the amplitude of the CMAP over APB was measured and expressed as a percentage of the 100% maximum response, after stimulation at each 0.5 cm interval. Finally, the last 2 steps were repeated using magnetic stimulation with the Cadwell round and butterfly coils.
MUSCLE & NERVE
October 1990
The results of routine sensory and motor nerve conduction studies were compared quantitatively on all 5 subjects, using electrical stimulation and the Cadwell butterfly coil. The midpoint of the line connecting the leading edge of the 2 round coils was positioned at the same point as the electrical cathode. The differences in the motor latencies obtained with magnetic and electrical stimulation at the same site were converted into estimated distances between the depolarization loci, based on the electrically determined motor nerve conduction velocity for the forearm segment. Comparable calculations were also performed on the differences in sensory latencies using the electrically determined sensory conduction velocity. Descriptive statistics and the Mann-Whitney nonparametric test were used for statistical analysis.
Statistical Analysis.
RESULTS
We compared different angles of orientation for each of the round coils, using the terminology proposed by others.6 We were able to obtain maximal amplitudes and areas of compound muscle action potentials in comparison with those obtained with electrical stimulation using tangential-edge magnetic stimulation. By contrast, using the orthogonal-longitudinal orientation described by others,6 we were unable to achieve maximal responses. These responses sometimes approached 80% maximal following stimulation at the wrist with the Cadwell round coil, but were less than 10% maximal following stimulation in the region of the elbow with either round coil. The tilted-longitudinal orientation sometimes produced amplitudes and areas similar to tangential-edge, but the intensity of stimulation varied markedly with minor changes in tilt-angle and surface position, so that reproducible maximal responses could only be obtained with the tangential-edge orientation.
Orientation of Round Coils.
We were unable to obtain selective maximal activation of either the median o r ulnar nerves in 3 subjects with any orientation of the Novometrix coil at the wrist o r elbow. We were, however, able to obtain maximal responses with selective stimulation of at the wrist with the the median and ulnar Cadwell round coil, and % ’’ to 95% responses with selective stimulation of the median nerve at the elbow. We were unable to obtain se-
Comparisonof Three Magnetic Stimulator Coils.
Magnetic Stimulation
lective, near-maximal stimulation of the ulnar nerve below or above the elbow. The Cadwell butterfly coil, by contrast, permitted selective stimulation of the nerves at intensities sufficient to elicit maximal responses at all 5 sites. Both Cadwell coils, but not the Novometrix coil, had sufficiently low magnetic field interference to allow the recording of digital nerve action potentials following stimulation at the wrist. The Effect of Surface Position for Stimulation on Response Size. The effective field of stimulation for
the 2 Cadwell coils was compared with electrical stimulation by plotting the amplitude of compound muscle action potential (expressed as percent maximum) obtained when the point of stimulation was displaced transversely from the point of maximal response (Fig. 1). Using this method, the effective field of stimulation for the butterfly magnet was larger than for the electrical stimulus, but was smaller than for the round coil. Thus, there is greater spatial selectivity for the magnetic field produced by the butterfly coil in comparison with the round coil, although neither had the narrow effective field achieved using electrical stimulation. Stimulus Intensity Study. T h e effects of changing the stimulus intensity on the amplitude of the response are shown in Figure 2. The effects on area were the same as those on amplitude. Compared to maximal responses, the amplitude and area of the sensory responses are greater than for the motor responses with electrical, but not magnetic, stimulation at 25% and 50% intensity (P < 0.05).
-
100
‘
80
.-
0
Electrical
A Round
X
s
60
a
U
.--
* 3
40
!
20
0 -4
-2
0
2
4
Displacement (em) FIGURE 1. The effect of surface position for stimulation on response size. The point at the wrist where submaximal electrical stimulation produces the largest response is “0”;displacement is positive if lateral (radial) and negative if medial (ulnar).
MUSCLE & NERVE
October 1990
959
100
-
h
E
g
.-X
80-
60-
E 40._ V
-
c)
-€
20A
" T
0
/ _
25
Magnetic: A Magnetic: Electrical: A Electrical: 0
. 50
.
. 75
.
Motor Sensory Motor Sensory
. lo0
Intensity (% maximum) FIGURE 2. The effects of changing the stimulus intensity on the amplitude of the response is illustrated for both magnetic (open points) and electrical (filled points) stimulation on sensory (triangular points) and motor (square points) responses.
3 mm behind the point activated by the electrical stimulus. Since the leading edge of the butterfly magnet is 10 mm in front of the junction point of its 2 round coils, the magnetic stimulus usually activates the peripheral nerve between these 2 points of the coil. T h e sensory and motor conduction velocities were also in reasonably close agreement, except for the short across-elbow segment of the ulnar nerve (Fig. 3). The sensory conduction velocities derived from the two types of stimulation did not differ by more than 5 m/sec. The motor conduction velocities differed by less than 3 m/sec for the above-elbow to wrist segment of the ulnar nerve, by less than 7 m/sec for the belowelbow to wrist segment of the ulnar nerve and the median nerve, but by up to 18 m/sec for the across-elbow segment of the ulnar nerve. DISCUSSION
The differences in amplitude and area of sensory and motor responses are not significant with either magnetic or electrical stimuli at 75% of maximal stimulus intensity (P > 0.05). Motor responses to magnetic and electrical stimulation were comparable (both in amplitude and area) at submaxima1 stimulus intensities. By contrast, submaximal electrical stimulation produced a larger sensory response than the comparable magnetic stimulus at 25% and 50% intensity (P < 0.05), but not at 75% (P > 0.05). Thus, submaximal electrical stimuli preferentially activate sensory axons, whereas submaximal magnetic stimuli d o not. Selective supramaximal stimulation was successfully performed at all 5 sites with both the butterfly magnet and the electrical stimulator, and no significant differences were found in amplitude, area or duration of motor and sensory responses obtained with the two types of stimuli (Table 1). Similarly, no statistically significant differences were detected in response latency or conduction velocity during sensory and motor studies. There was close agreement between the latencies obtained with the two types of stimulation for both sensory and motor responses (as illustrated in Fig. 3 for the motor latencies). T h e latencies obtained with electrical stimulation were 0.04 0.1 msec shorter (mean ? 1 SD) than those obtained with the butterfly coil, and the differences varied from -0.2 to 0.2 msec. By converting these latency differences into distance, the leading edge of the magnet can be estimated to typically activate peripheral nerve Routine Nerve Conduction Studies.
*
960
Magnetic Stimulation
In this article we have compared the clinical utility of three magnetic coils (two different round coils and one butterfly-shaped coil) with standard electrical stimulation in normal subjects. The primary aim was to assess their suitability for the performance of routine median and ulnar nerve conduction studies. We consider the minimal requirement for a suitable stimulator to be the ability to provide selective, supramaximal stimulation of these nerves at different points along their course with a minimum of patient discomfort. Our experience with the Novametrix round coil and stimulator was similar to that of other^,^ in that it did not fulfill this minimum requirement for either nerve at the wrist or in the region of the elbow with any orientation of the coil. The Cadwell round coil did deliver supramaximal stimulation in the tangential-edge orientation, but it was selective only at some sites. In contrast, the Cadwell butterfly-prototype coil consistently provided selective supramaximal stimulation at all sites. The better performance of the butterfly-shaped, as opposed to the'round-shaped, coil was due to improved focality rather than strength of the magnetic stimulus. As illustrated (see Fig. l), the butterfly design provided a stimulus that was more focal than the round coil, but less focal than the electrical stimulator. Previous authors have also compared magnetic and electrical stimulation of the motor axons of peripheral nerves and found results similar to ours. 173r6 However, unlike these other authors, we also evaluated the performance of magnetic stimulators during routine sensory nerve conduction studies. Using the Novametrix stimulator, we were
MUSCLE & NERVE
October 1990
Table 1. The results of routine nerve conduction studies are expressed as the mean
Test
Place of stimulation
Nerve
Amplitude (mV or uV) Median Ulnar Area (mV . msec or uV . msec) Median Ulnar
Duration (msec) Median Ulnar
Onset Latency (msec) Median Ulnar
Motor
Sensory Magnetic
Electrical
Magnetic
Wrist Elbow Wrist Below-elbow Above-elbow
9.4f 1.7 9.0f 1.8 10.5f 2.2 9.7f 2.9 9.4f 2.8
9.5-c 2.2 8.9f 2.3 10.4 f 2.5 10.4f 2.0 9.9f 1.8
36f 18
35 f 19
31
30 f 15
Wrist Elbow Wrist Below-elbow Above-elbow
26.2f 3.9 25.7f 6.4 31.6 f 5.4 29.7f 8.5 28.6f 8.1
25.7f 5.7 23.8f 6.6 31.3 f 5.5 32.1 f 5.5 31.4f 5.2
22 f 9
Wrist Elbow Wrist Below-elbow Above-elbow
4.7f 0.5 4.8 f 0.2 4.9f 0.5 5.3 f 0.7 5.4f 0.6
4.7 f 0.4 4.8f 0.4 4.9 f 0.5 5.2f 0.5 5.4f 0.5
1.2f 0.1
1.2f 0.1
1.2 t 0.2
1.2 f 0.3
-
-
Wrist Elbow Wrist Below-elbow Above-elbow
3.3f 0.4 7.2f 0.7 2.9f 0.1 6.0f 0.5 7.7f 0.7
3.3f 0.3 7.1 f 0.7 2.9 f 0.1 6.0f 0.4 7.5 f 0.7
2.3f 0.3
2.3f 0.3
Distal Forearm Distal Below-elbow to wrist Across-el bow Above-elbow to wrist
unable to record sensory responses with stimulation at the wrist due to the large stimulus artifact produced by the magnetic field. With either of the Cadwell coils, by contrast, we were able to obtain selective, maximal sensory responses with stimulation of the median and ulnar nerves at the wrist. Comparing the ability of magnetic and electrical stimulators to activate sensory and motor axons, we found that only electrical stimulation preferentially activated sensory axons at some submaximal stimulus intensities. The reason for this difference between magnetic and electrical stimuli in the fiber population activated by submaximal stimuli is uncertain. This phenomenon does not affect routine sensory studies, because these are performed at supramaximal stimulus intensities. However, because selective stimulation of sensory axons at submaximal intensity is necessary in the performance of H-reflex studies, electrical stimulation may be the method of choice in these circumstances.
Magnetic Stimulation
1 SD.
Electrical
Conduction Velocity (mkec) Median Ulnar
f
f
12
11
21 c 9
19 2 8
-
-
-
-
-
-
2.5f 0.5
2.5f 0.4
-
-
64 2 5 -
65 f 2 68 & 16 65 f 5
f
-
65 f 5
67 f 5 60 f 9 64 f 6
21
-
-
-
-
-
65 f 2
-
-
60 f 8
-
-
64 f 3
-
59 f 6 -
We also compared the onset latencies of the sensory and motor response and the calculated conduction velocities derived using the butterfly coil with those from electrical stimulation. As illustrated in Figure 3 for the motor responses, the onset latencies were quite comparable, whereas the calculated conduction velocities had a wider range of differences, especially for the short acrosselbow segment of the ulnar nerve. Statistically significant differences were not found in the conduction velocities for any nerve segment. We also compared the differences in our conduction velocities derived from magnetic and electrical stimulation with the differences in conduction velocities derived from electrical stimulation on two separate occasions. Under the latter circumstances, others have reported differences in motor conduction velocities of 5% to 15%, or 5 to 11 r n / ~ e c , ~ and we have observed intertrial differences in antidromic sensory conduction velocities of up to 14 m/sec (unpublished data). Based on this range of
MUSCLE & NERVE
October 1990
961
Median Nerve
B
A
Latency (msec): Electrical Stimulation
CV (mlsec): Electrical Stimulation
Ulnar Nerve
“
I
0 Across Elbow A Above-Elbow to Wrist I /
ow. 0
, 2
. , . , . , 4
6
0
.
1
1
0
Latency (msec): Electrical Stimulation
V isi 25
. . . .
,
. . . . ,
50
75
.
.
.
..I
100
CV (mlsec):Electrical Stimulation
FIGURE 3. The results of motor latencies (A and C) and nerve conduction velocities (B and D) for median (A and B) and ulnar (C and D) nerves are compared for the butterfly magnet and electrical stimuli.
differences, magnetic stimulation with the butterfly coil provides findings comparable to electrical stimulation for sensory and motor conduction velocities for most segments, where we obtained differences of less than 7 m/sec. However, even the current butterfly stimulator cannot be used as a substitute for electrical stimulation when performing routine motor conduction studies on short segments, such as the across-elbow segment of the ulnar nerve, where we observed differences of up to 18 m/sec. In summary, recent developments in magnetic stimulator design have improved the focality of the stimulus, so that selective supramaximal stimuli may be delivered to commonly studied peripheral nerves in the upper limbs. Our study demon-
strates that the results of magnetic stimulation are sufficiently comparable to those obtained with electrical stimulation that the use of magnetic stimulation may be a reasonable alternative to electrical stimulation in certain situations. However, we have not demonstrated any particular advantage of magnetic stimulation. Futhermore, the calculated motor nerve conduction velocities that are based on magnetic stimulation have a broader range of variability than velocities derived from electrical stimulation. This presently limits the clinical utility of routine magnetic stimulation in detecting focal changes in nerve conduction velocity at common entrapment sites, such as the across-elbow segment of the ulnar nerve.
REFERENCES
1. Amassian VE, Maccabee PJ, Cracco RQ: Focal stimulation of human peripheral nerve with the magnetic coil: A comparison with electrical stimulation. Ex$ Neurol 1989; 103~282-289.
962
Magnetic Stimulation
2. Bickford RG, Guidi M, Fortesque P, Swenson M: Magnetic stimulation of human peripheral nerve and brain: response enhancement by combined magnetoelectrical technique. Neurosurgery 1987;20:110- 116.
MUSCLE & NERVE
October 1990
3. Evans BA, Litchy WJ, Daube JR: The utility of magnetic stimulation for routine peripheral nerve conduction studies. Muscle N m e 1988;11:1074-1078. 4. Halar EM, Venkatesh B: Nerve conduction velocity measurements: improved accuracy using superimposed response waves. Arch Phys Med Rehabil 1976;57:451-457. 5. Kolin A, Brill NQ, Broberg PJ: Stimulation of irritable tissues by means of an alternating magnetic field. Proc SOLExp Biol Med 1959;102:251-253.
Magnetic Stimulation
6. Maccabee PJ, Amassian VE, Cracco RQ, Cadwell JA: An analysis of peripheral motor nerve stimulation in humans using the magnetic coil. Electroencephalogr Clin Neurophysiol 1988;70:524-533. 7. Polson MJR, Barker AT, Freeston IL: Stimulation of nerve trunks with time-varying magnetic fields. Med Biol Eng Cm@t 1982;20:243-244.
MUSCLE & NERVE
October 1990
963
