To assess the longitudinal dispersion of the stimulus induced by the magnetic coil, collision experiments were performed in seven normal ulnar nerves. A supramaximal electrical stimulus S1 was delivered at the wrist, and followed by a supramaximal stimulus S2 in the upper arm, which was either electrical (electrical collision studies), or magnetic (magnetic collision studies). The interstimulus interval was varied by 0.2 msec increments from the time of complete cancellation of the S2 evoked motor response onwards, to include the entire span of recovery of that compound motor action potential. Collision curves were obtained for both magnetic and electrical stimuli by plotting the amplitude of the motor response elicited by S2 as a function of the interstimulus interval. In all seven normal ulnar nerves, comparison of the collision curves showed that the S2 evoked motor response is restored significantly more slowly when magnetic stimulation is used. This finding is best explained by longitudinal dispersion of the stimulus induced by the magnetic coil relative to conventional electrical stimulation, the large fibers being stimulated further away from the coil than the small ones. This interpretation is confirmed by the findings obtained with the same method in two cases of ulnar neuropathy, and by comparison of different intensities of magnetic stimulation. Key words: Magnetic stimulation peripheral nerve physiology collision studies MUSCLE & NERVE 13~1076-1082 1990
SPATIAL DISPERSlON OF MAGNETIC STIMULATION IN PERIPHERAL DIDIERCROS, MD, TIMOTHY J. DAY, FRACP, and BHAGWAN T. SHAHANI, MD, DPhll
Magnetic stimulation of the central and peripheral nervous system is now technically possible.' The principle underlying this method is the induction of an electrical field within a volume conductor in response to a rapidly varying magnetic field. This electric field, in turn, induces a current, which changes the degree of polarization of excitable membranes as a function of the sign and number of electrical charges transported across these membranes. An action potential is elicited if a depolarization reaching threshold is obtained. One of the advantages of magnetic stimulation over electrical stimulation is its painlessness. AnFrom the Clinical Neurophysiology Laboratories and Neurology Service, Massachusetts General Hospital, and Department of Neurology, Harvard Medical School, Boston, Massachusetts. Acknowledgments: We are grateful to Mr. Kazutomo Yunokuchi, Visiting Scientist at the Francis Bittner National Magnet Laboratory, MIT, who kindly performed the current field mapping shown in Figure 5. Address reprint requests to Dr. Didier Cros, Clinical Neurophysiology Laboratories, Bigelow 11, Massachusetts General Hospital, Boston, MA 02114. Accepted for publication November 19, 1989. CCC 0148-639W90101101076-07 $04.00 0 1990 John Wiley & Sons, Inc.
1076
Spatial Dispersion of Magnetic Stimulation
other is the fact that the corresponding electrical field is not attenuated by high-resistance biological tissues, such as skin and bone, which makes it suitable for stimulation of deep neural structures. Magnetic stimulation of the cerebral cortex is thus readily performed with this method, and does not, unlike electrical stimulation, cause any significant discomfort. Studies of clinical relevance have already been performed in Europe.',' l Approval of cortical magnetic stimulation for clinical use by the Food and Drug Administration is still pending. The manufacturers and distributors of magnetic stimulators in the United States have therefore focused the attention of potential users on magnetic stimulation of the peripheral nervous system. We found that magnetic stimulation can be used to reliably stimulate motor fibers proximal to the brachial plexus when the coil is centered on the spinous processes of the cervical vertebrae, l o representing a valuable alternative to spinal root needle stimulation,* spinal root percutaneous electrical stimulation,' and to painful electrical stimulation at Erb's point. Magnetic stimulation can also be applied to peripheral nerves in the limbs, at sites where conven-
MUSCLE & NERVE
November 1990
120
100
w a
60
-
n 40
-
3
c -I E
a 20
-
0
- 1
1
0
I
1
2
3
IS1 (msec) tional electrical stimulation has been used satisfactorily for decades. Again, the advantage of magnetic stimulation over electrical stimulation is its painlessness, due to the induction of relatively small electrical currents in the skin, which has a high resistance. This is in contrast to the passage of large currents delivered percutaneously necessary to excite deep structures with conventional electrical stimulation. This relative advantage prompted further comparison of magnetic and electrical stimulation in peripheral nerves. A stimulus adequate for peripheral nerve studies should be (1) adjustable in intensity so that it recruits the entire population of motor fibers under investigation (supramaximal stimulus); (2) focal enough to avoid stimulus spread to neighboring nerves laterally; and (3) focal enough longitudinally, so that all nerve fibers within the stimulated nerve are activated at the same point, allowing for accurate distance measurement and conduction velocity determination. The first two points have already been a d d r e ~ s e d .We ~ . ~have investigated the latter question using a collision technique. MATERIALS AND METHODS
Subjects were laboratory personnel and the authors. A total of seven normal ulnar nerves, and two ulnar nerves with mild chronic compression neuropathy at the elbow were studied. T h e compound muscle action potentials (CMAPs) were recorded from the abductor digiti minimi muscle with surface disc electrodes with the usual tendon belly placement. N o volume conducted response
Spatial Dispersion of Magnetic Stimulation
FIGURE 1. Illustration of the method used to analyze the collision curves. The data points and the best fit curve (r > 0.99) are shown. The coordinates of three points are drawn. From left to right on the x axis, the x values corresponding to these three points are x , ~ , xs0, and x,, and represent the x values for S2 evoked amplitudes of 10, 50, and 90% of maximum, respectively. These x values are extrapolated from the best fit curves. Delta IS1 is obtained from x,-x,,. The origin on the x axis corresponds to the first measurable S2-evoked CMAP.
FIGURE 2. Collision study in which S1 was a supramaximal electrical stimulus delivered to the ulnar nerve at the wrist, and S2 was a magnetic stimulus delivered to the ulnar nerve in the upper arm. The Cadwell magnetic stimulator was used in this case. The compound motor action potential is recorded from the abductor digiti minimi. The motor response elicited by S1 is superimposed on the artifact corresponding to S2 stimulation. S2 was delivered at a constant interval relative to the onset of the sweep. S1 -S2 was increased by gradually shortening the interval between onset of the sweep and S1. A, C, and E: the interstimulus interval (ISI) was increased by 0.1 msec steps beginning at 3 msec (A: top trace) at which time no motor response was evoked by S2. The first response elicited by 52 is noted in the 2 bottom traces in A (IS1 3.4 and 3.5 msec, respectively). In 8, the IS1 increases from 3.6to 4.1 msec from top to bottom, and the response elicited by S2 gradually increases in amplitude. The IS1 increases from 4.2 to 4.7 in C, the compound motor response evoked by S2 has been restored to its maximum value. B, D, and F are superimpositions of the responses shown in A, C, and E, respectively. Calibrations: 10 mV and 5 msec.
MUSCLE & NERVE
November 1990
1077
120 1
.....
100 -
-8
A D O 0 0
0
-
80
0
0
.
a
5
40
.
D 0
D
b
2o
1
0
IS1 (msec)
120 100
11
!i _ _._401
B
~
~
1.
,
,
.
,
20
0 - 1
0
1
2
3
IS1 (msec)
was recordable from these electrodes on supramaximal stimulation of the median nerve. The recording equipment was a Mystro EMG machine (TECA Corp., Pleasantville, New York) equipped with a constant current stimulator. A Novametrix or Cadwell magnetic stimulator was used, the outer diameter of the coils being 14 cm and 9.5 cm, respectively. Collision studies were performed as follows: a supramaximal electrical stimulus (S 1) was delivered at the wrist, and was followed by a second stimulus (S2) delivered in the upper arm. T h e second stimulus was supramaximal in all instances. S2 was an electrical stimulus (S2e) in electrical collision studies (ECS), or a magnetic stimulus (S2m)
1078
Spatial Dispersion of Magnetic Stimulation
.
,
FIGURE 3. (A, B) Comparison of magnetic (open squares) and electrical (solid diamonds) collision curves in two normal ulnar nerves (A and 8). The solid squares indicate superimposition of an open square and a solid diamond. The origin on the x axis corresponds to the first measurable response evoked by S2. Note the steeper slope of the mid-portion of the electrical curves in both instances, and the plateaus in the magnetic collision curves.
in magnetic collision studies (MCS). The magnetic coil was applied tangentially over the medial aspect of the upper arm, and centered over the neurovascular bundle, The distal part of the outer rim of the coil was placed within a few centimeters of the site of application of the cathode. T h e maximal output of the magnetic stimulator was needed in most cases for supramaximal stimulation. T h e S1-S2 interstimulus interval (ISI) was gradually increased by 0.1 msec increments, from the time when the IS1 was too short for any response to develop in response to S2, to that of full recovery of the S2-elicited CMAP (the 100% value used to normalize the results) (Fig. 2). The normalized amplitude of the CMAP
MUSCLE & NERVE
November 1990
evoked by S2 was plotted as a function of the ISI. The resulting curve is sigmoid-shaped3 (Fig. l ) , and includes three consecutive segments corresponding to the cumulative effects of recruitment of fast, mid-range, and slow conducting alpha motor axons, respectively. Quantitative comparison of MCS and ECS in the same nerve was performed by determining (1) the slope of the tangent to the mid-portion of the curve, obtained from the derivative of the fourth order polynomial equation providing the best fit for the data points. T h e x value was extrapolated from the curve for a CMAP amplitude of 50% and (2) the increment in IS1 (delta ISI) necessary to increase the amplitude of the S2-evoked CMAP from 10% to 90%. The best fit curves were obtained with commercially available software (Cricketgraph, Cricket Software, Inc.) In all cases, the r value for the best fit curve was greater than 0.99. T h e statistical comparisons were conducted with the MannWhitney U test.
l2O1 100
:60{ 2
IS1 (msec)
FIGURE 4. Comparison of two magnetic collision curves performed at 100% (open squares) and 60% (solid diamonds) of maximum coil output. Solid squares indicate superimposition of an open square and a solid diamond. The amplitudes were normalized relative to the amplitude elicited by the 100% magnetic stimulus. The origin on the x axis is the first measurable response elicited by S2 with 100% output. Note the markedly different IS1 (0.6 msec) needed to elicit a recordable SP-evoked CMAP in both situations.
RESULTS
Magnetic and electrical collision curves were markedly different in all seven normal ulnar nerves (Fig. 3A and 3B). T h e rise time of the magnetic collision curve was always longer than that of the paired electrical collision curve. Occasionally, plateaus were noted in the magnetic collision curves (Fig. 3B), but were never seen in the electrical collision curves. T h e slope of the tangent to the mid-portion of the collision curve was steeper
Table 1. Comparison of magnetic and electrical collision studies. Nerve no. 1 2 3 4 5 6 7 8 9
Tangent theta E*
Tangent theta M
Delta lSEt
Delta IS1 M
73.7 77.8 92.5 63.8 86.1 308 69.5 86 103
48.1 33.1 60.4 44 42.4 35.1 58.2 73.8 86.6
1.21 1.13 1 1.57 1 1.75 1.32 1.04 0.84
2.15 3.97 1.49 2.05 2.05 2.50 1.63 1.19 1.05
"Tangent Theta is the slope of the best fit curve for the x value correspondrng to 50% of maximum amp/ifude with electrical (€j and magnetic (Mj stimulation tDelta IS/ refers to the increase in IS/ (in msec) needed to increase the CMAP evoked by S2 from 70 to 90% of the maximum value Nerves 7 to 7 were normal ulnar nerves Nerves 8 and 9 had mild cornpressron neuropathies at the elbow
Spatial Dispersion of Magnetic Stimulation
with electrical than with magnetic stimulation (P < 0.001), and the increment in IS1 needed to increase the amplitude of the S2 evoked CMAP from 10% to 90% was shorter with electrical than with magnetic stimulation (P < 0.002) (Table 1). Comparison of collision curves obtained with different intensities of magnetic stimulation with the coil kept in the same position showed a marked shift in IS1 needed to obtain an SP-elicited response. This indicates a marked difference in location of the virtual cathode exciting the largest motor axons in these two situations, The 0.6 msec difference in IS1 shown in Figure 3 corresponds to a longitudinal shift in position of the virtual cathode of 3.6 cm, as the intensity of the magnetic stimulus is increased from 60% to loo%, the maximum motor conduction velocity in this nerve being 60 m/sec (Fig. 4). Comparison of electrical and magnetic collision curves in two ulnar nerves with mild compression neuropathy at the elbow did not show the dramatic differences seen in normal ulnar nerves (Fig. 5 and Table 1). DISCUSSION
-
The objective of this study was to assess the longitudinal focalitv of mametic stimulation relative to that of conventional electrical stimulation in peripheral nerves. Percutaneous electrical stimulaI
MUSCLE & NERVE
November 1990
1079
A 120
3l 80 o0l Y
.
40
1
2o
o !
0
0
- , o 0
- 1
I
1
#
1
2
3
IS1 (msec)
120
B
1
-E.{. E3 60-
’
k
2 40-
5
20
0
-
0
8 0-
- -
a
1
I
tion results from the passage of an electrical current between cathode and anode within a volume conductor, and causes accumulation of depolarizing charges able to excite nerve fibers at a short distance from the cathode. It is useful to use the concept of “virtual cathode” as the exact site of stimulation of the nerve fibers. T h e virtual cathode lies in a position intermediate to that of the actual cathode and anode, and its location relative to that of the actual cathode varies slightly with the strength of the electrical stimulus. The assumption that the nerve fibers are stimulated as they pass under the cathode is the basis for distance measurements leading to conduction velocity determination. This approach has proven satis-
1080
Spatial Dispersion of Magnetic Stimulation
FIGURE 5. A and B. Comparison of magnetic and
factory for decades and remains unchallenged. Comparison of collision curves obtained with electrical and magnetic stimulation consistently revealed the same differences in seven normal ulnar nerves. These included a steeper rise time for the electrical than for the magnetic collision curve, corresponding to a longer IS1 span to restorethe amplitude of the CMAP to supramaximal level when a magnetic stimulus was used. Our interpretation of these findings is that there is spatial dispersion of the stimulus induced by the magnetic pulse along the nerve, resulting in activation of the fastest conducting motor fibers by a virtual cathode different from the virtual cathode activating slow conducting motor fibers. The larger the
MUSCLE & NERVE
November 1990
FIGURE 6. Current field profile induced by the Novametrix coil at 50% of maximum power in a rectangular tank filled with saline. The coil was horizontal 0.6 cm above the surface of the saline solution. The projection of the center of the coil was the dot centering the field. The measurements were performed 0.9 cm below the surface at points which were 1 cm apart. Each arrow represents an actual measurement and shows the direction and intensity of the current at this particular point. The current was measured along three lines above and two lines below the center of the coil. The bar also refers to the arrow length and represents 4.5 x 10e-2 Coulombs/square meter/phase.
motor fiber, the more distal the virtual cathode relative to the coil. This is probably due to the large size of the electrical field induced by the magnetic pulse compared to that of the field generated by electrical stimulation. In a large field, the gradual variation in current density may cause spatial dispersion of the stimulus because of the differing thresholds of individual nerve fibers as a function of their diameter. Thus, the largest fibers would be activated at the periphery of the field where the current density is not sufficient to excite smaller axons. Smaller fibers would be activated more centrally within the field, as the current density increases. Actual mapping of the current field elicited by a magnetic pulse in a tank reproducing the general shape of the volume conductor of the upper arm yielded data confirming this interpretation (Fig. 6). The data obtained by comparison of two different intensities of magnetic stimulation support this interpretation, since an increase in intensity from 60% to 100% results not only in an increase in amplitude of the evoked CMAP, but also in a shift of 3 cm of the virtual cathode activating the fastest-conducting motor fibers towards the periphery. In two ulnar nerves with mild, chronic compression at the elbow, the differences noted between magnetic and electrical collision curves were much less apparent than in normal ulnar nerves. It has been established that compression alters conduction in large diameter fibers, while smaller fibers are largely spared.539Depending on the se-
Spatial Dispersion of Magnetic Stimulation
verity of compression, segmental demyelination, focal remyelination or Wallerian degeneration will develop in large diameter motor fiber^.^ Any of these three lesions affecting the fastest conducting fibers results, at the whole nerve level, in lowering of the maximum conduction velocity with sparing of the minimum conduction velocity, since small motor fibers are largely intact. This corresponds to narrowing of the maximum-minimum range of velocities in the motor fibers to the detriment of the fastest conducting ones. l 2 The similarity of electrical and magnetic collision curves in ulnar neuropathy thus supports our interpretation of the differences between these curves in normal nerves. Since there are no functional fast-conducting fibers left in compressed ulnar nerves, the longitudinal dispersion of magnetic stimulation is no longer apparent. At this time, magnetic stimulation of peripheral nerves has several limitations: (i) spread to neighboring nerves4"; (ii) difficulties in obtaining supramaximal stimulation4; (iii) uncertainty in the location of the "virtual" cathode relative to the magnetic coil"'; and (iv) longitudinal dispersion of excitation of fibers of different diameter. These problems limit the routine use of magnetic stimulation in clinical studies of the peripheral nervous system. It is conceivable that further technical developments leading to coils generating smaller magnetic fields of similar magnitude will improve the focality of magnetic stimulation and make it more suitable for clinical studies.
MUSCLE 8, NERVE
November 1990
1081
REFERENCES
1. Barker AT, Jalinous R, Freeston IL: Non-invasive stimulation of human motor cortex. Lancet 1985;i:1106- 1107. 2. Berger AR, Busis NA, Logigian EL, Wierzbicka M, Shahani BT: Cervical root stimulation in the diagnosis of radiculopathy. Neurology 1987;37:329-332. 3. Domingue J, Shahani BT, Young RR: Conduction velocity in different diameter ulnar sensory and motor nerve fibers. Muscle Nerve 1980;3:437. 4. Evans BA, Litchy WJ, Daube JR: The utility of magnetic stimulation for routine peripheral nerve conduction studies. Muscle Nerve 1988;11:1074-1078. 5. Gilliatt RW: Chronic nerve compression and entrapment, in Sumner AJ (eds): The Physiology of Peripheral Nerve Diseuse, Philadelphia, Saunders, 1980, pp 316-339. 6. Hess CW, Mills KR, Murray NMF, Schriefer TN: Magnetic brain stimulation: Central motor conduction studies in Multiple Sclerosis. Ann Neurol 1987 ;22:744- 752. 7. Maccabee PJ, Amassian VE, Cracco RQ, Cadwell JA: An analysis of peripheral motor nerve stimulation in humans
1082
Spatial Dispersion of Magnetic Stimulation
using the magnetic coil. Electroenceph Clin Neurophysiol (in press). 8. Mills KR, Murray NMF: Electrical stimulation over the human vertebral column. Electroenceph Clin Neurophysiol 1986;63:582-589. 9. Ochoa J, Fowler TJ, Gilliatt RW: Anatomical changes in peripheral nerves compressed by a pneumatic tourniquet. J Anat 1972;113:433-455. 10. Santamaria J, King PL, Cros D, Chiappa KH: Cervical magnetic stimulation: roots or spinal nerves? Neurology 1988;38:199. 1 1 . Schriefer TN, Hess CW, Mills KR, Murray NMF: Central motor conduction studies in motor neuron disease using magnetic brain stimulation. Electroenceph Clin Neurophysiol (in press). 12. Shahani BT, Potts F, Juguilon A, Young RR: Maximalminimal motor nerve conduction and F response studies in normal subjects and patients with ulnar compression neuropathies. EEG Clin Neurophysiol 1980;50:172P.
MUSCLE & NERVE
November 1990
SPATIAL DISPERSlON OF MAGNETIC STIMULATION IN PERIPHERAL DIDIERCROS, MD, TIMOTHY J. DAY, FRACP, and BHAGWAN T. SHAHANI, MD, DPhll
Magnetic stimulation of the central and peripheral nervous system is now technically possible.' The principle underlying this method is the induction of an electrical field within a volume conductor in response to a rapidly varying magnetic field. This electric field, in turn, induces a current, which changes the degree of polarization of excitable membranes as a function of the sign and number of electrical charges transported across these membranes. An action potential is elicited if a depolarization reaching threshold is obtained. One of the advantages of magnetic stimulation over electrical stimulation is its painlessness. AnFrom the Clinical Neurophysiology Laboratories and Neurology Service, Massachusetts General Hospital, and Department of Neurology, Harvard Medical School, Boston, Massachusetts. Acknowledgments: We are grateful to Mr. Kazutomo Yunokuchi, Visiting Scientist at the Francis Bittner National Magnet Laboratory, MIT, who kindly performed the current field mapping shown in Figure 5. Address reprint requests to Dr. Didier Cros, Clinical Neurophysiology Laboratories, Bigelow 11, Massachusetts General Hospital, Boston, MA 02114. Accepted for publication November 19, 1989. CCC 0148-639W90101101076-07 $04.00 0 1990 John Wiley & Sons, Inc.
1076
Spatial Dispersion of Magnetic Stimulation
other is the fact that the corresponding electrical field is not attenuated by high-resistance biological tissues, such as skin and bone, which makes it suitable for stimulation of deep neural structures. Magnetic stimulation of the cerebral cortex is thus readily performed with this method, and does not, unlike electrical stimulation, cause any significant discomfort. Studies of clinical relevance have already been performed in Europe.',' l Approval of cortical magnetic stimulation for clinical use by the Food and Drug Administration is still pending. The manufacturers and distributors of magnetic stimulators in the United States have therefore focused the attention of potential users on magnetic stimulation of the peripheral nervous system. We found that magnetic stimulation can be used to reliably stimulate motor fibers proximal to the brachial plexus when the coil is centered on the spinous processes of the cervical vertebrae, l o representing a valuable alternative to spinal root needle stimulation,* spinal root percutaneous electrical stimulation,' and to painful electrical stimulation at Erb's point. Magnetic stimulation can also be applied to peripheral nerves in the limbs, at sites where conven-
MUSCLE & NERVE
November 1990
120
100
w a
60
-
n 40
-
3
c -I E
a 20
-
0
- 1
1
0
I
1
2
3
IS1 (msec) tional electrical stimulation has been used satisfactorily for decades. Again, the advantage of magnetic stimulation over electrical stimulation is its painlessness, due to the induction of relatively small electrical currents in the skin, which has a high resistance. This is in contrast to the passage of large currents delivered percutaneously necessary to excite deep structures with conventional electrical stimulation. This relative advantage prompted further comparison of magnetic and electrical stimulation in peripheral nerves. A stimulus adequate for peripheral nerve studies should be (1) adjustable in intensity so that it recruits the entire population of motor fibers under investigation (supramaximal stimulus); (2) focal enough to avoid stimulus spread to neighboring nerves laterally; and (3) focal enough longitudinally, so that all nerve fibers within the stimulated nerve are activated at the same point, allowing for accurate distance measurement and conduction velocity determination. The first two points have already been a d d r e ~ s e d .We ~ . ~have investigated the latter question using a collision technique. MATERIALS AND METHODS
Subjects were laboratory personnel and the authors. A total of seven normal ulnar nerves, and two ulnar nerves with mild chronic compression neuropathy at the elbow were studied. T h e compound muscle action potentials (CMAPs) were recorded from the abductor digiti minimi muscle with surface disc electrodes with the usual tendon belly placement. N o volume conducted response
Spatial Dispersion of Magnetic Stimulation
FIGURE 1. Illustration of the method used to analyze the collision curves. The data points and the best fit curve (r > 0.99) are shown. The coordinates of three points are drawn. From left to right on the x axis, the x values corresponding to these three points are x , ~ , xs0, and x,, and represent the x values for S2 evoked amplitudes of 10, 50, and 90% of maximum, respectively. These x values are extrapolated from the best fit curves. Delta IS1 is obtained from x,-x,,. The origin on the x axis corresponds to the first measurable S2-evoked CMAP.
FIGURE 2. Collision study in which S1 was a supramaximal electrical stimulus delivered to the ulnar nerve at the wrist, and S2 was a magnetic stimulus delivered to the ulnar nerve in the upper arm. The Cadwell magnetic stimulator was used in this case. The compound motor action potential is recorded from the abductor digiti minimi. The motor response elicited by S1 is superimposed on the artifact corresponding to S2 stimulation. S2 was delivered at a constant interval relative to the onset of the sweep. S1 -S2 was increased by gradually shortening the interval between onset of the sweep and S1. A, C, and E: the interstimulus interval (ISI) was increased by 0.1 msec steps beginning at 3 msec (A: top trace) at which time no motor response was evoked by S2. The first response elicited by 52 is noted in the 2 bottom traces in A (IS1 3.4 and 3.5 msec, respectively). In 8, the IS1 increases from 3.6to 4.1 msec from top to bottom, and the response elicited by S2 gradually increases in amplitude. The IS1 increases from 4.2 to 4.7 in C, the compound motor response evoked by S2 has been restored to its maximum value. B, D, and F are superimpositions of the responses shown in A, C, and E, respectively. Calibrations: 10 mV and 5 msec.
MUSCLE & NERVE
November 1990
1077
120 1
.....
100 -
-8
A D O 0 0
0
-
80
0
0
.
a
5
40
.
D 0
D
b
2o
1
0
IS1 (msec)
120 100
11
!i _ _._401
B
~
~
1.
,
,
.
,
20
0 - 1
0
1
2
3
IS1 (msec)
was recordable from these electrodes on supramaximal stimulation of the median nerve. The recording equipment was a Mystro EMG machine (TECA Corp., Pleasantville, New York) equipped with a constant current stimulator. A Novametrix or Cadwell magnetic stimulator was used, the outer diameter of the coils being 14 cm and 9.5 cm, respectively. Collision studies were performed as follows: a supramaximal electrical stimulus (S 1) was delivered at the wrist, and was followed by a second stimulus (S2) delivered in the upper arm. T h e second stimulus was supramaximal in all instances. S2 was an electrical stimulus (S2e) in electrical collision studies (ECS), or a magnetic stimulus (S2m)
1078
Spatial Dispersion of Magnetic Stimulation
.
,
FIGURE 3. (A, B) Comparison of magnetic (open squares) and electrical (solid diamonds) collision curves in two normal ulnar nerves (A and 8). The solid squares indicate superimposition of an open square and a solid diamond. The origin on the x axis corresponds to the first measurable response evoked by S2. Note the steeper slope of the mid-portion of the electrical curves in both instances, and the plateaus in the magnetic collision curves.
in magnetic collision studies (MCS). The magnetic coil was applied tangentially over the medial aspect of the upper arm, and centered over the neurovascular bundle, The distal part of the outer rim of the coil was placed within a few centimeters of the site of application of the cathode. T h e maximal output of the magnetic stimulator was needed in most cases for supramaximal stimulation. T h e S1-S2 interstimulus interval (ISI) was gradually increased by 0.1 msec increments, from the time when the IS1 was too short for any response to develop in response to S2, to that of full recovery of the S2-elicited CMAP (the 100% value used to normalize the results) (Fig. 2). The normalized amplitude of the CMAP
MUSCLE & NERVE
November 1990
evoked by S2 was plotted as a function of the ISI. The resulting curve is sigmoid-shaped3 (Fig. l ) , and includes three consecutive segments corresponding to the cumulative effects of recruitment of fast, mid-range, and slow conducting alpha motor axons, respectively. Quantitative comparison of MCS and ECS in the same nerve was performed by determining (1) the slope of the tangent to the mid-portion of the curve, obtained from the derivative of the fourth order polynomial equation providing the best fit for the data points. T h e x value was extrapolated from the curve for a CMAP amplitude of 50% and (2) the increment in IS1 (delta ISI) necessary to increase the amplitude of the S2-evoked CMAP from 10% to 90%. The best fit curves were obtained with commercially available software (Cricketgraph, Cricket Software, Inc.) In all cases, the r value for the best fit curve was greater than 0.99. T h e statistical comparisons were conducted with the MannWhitney U test.
l2O1 100
:60{ 2
IS1 (msec)
FIGURE 4. Comparison of two magnetic collision curves performed at 100% (open squares) and 60% (solid diamonds) of maximum coil output. Solid squares indicate superimposition of an open square and a solid diamond. The amplitudes were normalized relative to the amplitude elicited by the 100% magnetic stimulus. The origin on the x axis is the first measurable response elicited by S2 with 100% output. Note the markedly different IS1 (0.6 msec) needed to elicit a recordable SP-evoked CMAP in both situations.
RESULTS
Magnetic and electrical collision curves were markedly different in all seven normal ulnar nerves (Fig. 3A and 3B). T h e rise time of the magnetic collision curve was always longer than that of the paired electrical collision curve. Occasionally, plateaus were noted in the magnetic collision curves (Fig. 3B), but were never seen in the electrical collision curves. T h e slope of the tangent to the mid-portion of the collision curve was steeper
Table 1. Comparison of magnetic and electrical collision studies. Nerve no. 1 2 3 4 5 6 7 8 9
Tangent theta E*
Tangent theta M
Delta lSEt
Delta IS1 M
73.7 77.8 92.5 63.8 86.1 308 69.5 86 103
48.1 33.1 60.4 44 42.4 35.1 58.2 73.8 86.6
1.21 1.13 1 1.57 1 1.75 1.32 1.04 0.84
2.15 3.97 1.49 2.05 2.05 2.50 1.63 1.19 1.05
"Tangent Theta is the slope of the best fit curve for the x value correspondrng to 50% of maximum amp/ifude with electrical (€j and magnetic (Mj stimulation tDelta IS/ refers to the increase in IS/ (in msec) needed to increase the CMAP evoked by S2 from 70 to 90% of the maximum value Nerves 7 to 7 were normal ulnar nerves Nerves 8 and 9 had mild cornpressron neuropathies at the elbow
Spatial Dispersion of Magnetic Stimulation
with electrical than with magnetic stimulation (P < 0.001), and the increment in IS1 needed to increase the amplitude of the S2 evoked CMAP from 10% to 90% was shorter with electrical than with magnetic stimulation (P < 0.002) (Table 1). Comparison of collision curves obtained with different intensities of magnetic stimulation with the coil kept in the same position showed a marked shift in IS1 needed to obtain an SP-elicited response. This indicates a marked difference in location of the virtual cathode exciting the largest motor axons in these two situations, The 0.6 msec difference in IS1 shown in Figure 3 corresponds to a longitudinal shift in position of the virtual cathode of 3.6 cm, as the intensity of the magnetic stimulus is increased from 60% to loo%, the maximum motor conduction velocity in this nerve being 60 m/sec (Fig. 4). Comparison of electrical and magnetic collision curves in two ulnar nerves with mild compression neuropathy at the elbow did not show the dramatic differences seen in normal ulnar nerves (Fig. 5 and Table 1). DISCUSSION
-
The objective of this study was to assess the longitudinal focalitv of mametic stimulation relative to that of conventional electrical stimulation in peripheral nerves. Percutaneous electrical stimulaI
MUSCLE & NERVE
November 1990
1079
A 120
3l 80 o0l Y
.
40
1
2o
o !
0
0
- , o 0
- 1
I
1
#
1
2
3
IS1 (msec)
120
B
1
-E.{. E3 60-
’
k
2 40-
5
20
0
-
0
8 0-
- -
a
1
I
tion results from the passage of an electrical current between cathode and anode within a volume conductor, and causes accumulation of depolarizing charges able to excite nerve fibers at a short distance from the cathode. It is useful to use the concept of “virtual cathode” as the exact site of stimulation of the nerve fibers. T h e virtual cathode lies in a position intermediate to that of the actual cathode and anode, and its location relative to that of the actual cathode varies slightly with the strength of the electrical stimulus. The assumption that the nerve fibers are stimulated as they pass under the cathode is the basis for distance measurements leading to conduction velocity determination. This approach has proven satis-
1080
Spatial Dispersion of Magnetic Stimulation
FIGURE 5. A and B. Comparison of magnetic and
factory for decades and remains unchallenged. Comparison of collision curves obtained with electrical and magnetic stimulation consistently revealed the same differences in seven normal ulnar nerves. These included a steeper rise time for the electrical than for the magnetic collision curve, corresponding to a longer IS1 span to restorethe amplitude of the CMAP to supramaximal level when a magnetic stimulus was used. Our interpretation of these findings is that there is spatial dispersion of the stimulus induced by the magnetic pulse along the nerve, resulting in activation of the fastest conducting motor fibers by a virtual cathode different from the virtual cathode activating slow conducting motor fibers. The larger the
MUSCLE & NERVE
November 1990
FIGURE 6. Current field profile induced by the Novametrix coil at 50% of maximum power in a rectangular tank filled with saline. The coil was horizontal 0.6 cm above the surface of the saline solution. The projection of the center of the coil was the dot centering the field. The measurements were performed 0.9 cm below the surface at points which were 1 cm apart. Each arrow represents an actual measurement and shows the direction and intensity of the current at this particular point. The current was measured along three lines above and two lines below the center of the coil. The bar also refers to the arrow length and represents 4.5 x 10e-2 Coulombs/square meter/phase.
motor fiber, the more distal the virtual cathode relative to the coil. This is probably due to the large size of the electrical field induced by the magnetic pulse compared to that of the field generated by electrical stimulation. In a large field, the gradual variation in current density may cause spatial dispersion of the stimulus because of the differing thresholds of individual nerve fibers as a function of their diameter. Thus, the largest fibers would be activated at the periphery of the field where the current density is not sufficient to excite smaller axons. Smaller fibers would be activated more centrally within the field, as the current density increases. Actual mapping of the current field elicited by a magnetic pulse in a tank reproducing the general shape of the volume conductor of the upper arm yielded data confirming this interpretation (Fig. 6). The data obtained by comparison of two different intensities of magnetic stimulation support this interpretation, since an increase in intensity from 60% to 100% results not only in an increase in amplitude of the evoked CMAP, but also in a shift of 3 cm of the virtual cathode activating the fastest-conducting motor fibers towards the periphery. In two ulnar nerves with mild, chronic compression at the elbow, the differences noted between magnetic and electrical collision curves were much less apparent than in normal ulnar nerves. It has been established that compression alters conduction in large diameter fibers, while smaller fibers are largely spared.539Depending on the se-
Spatial Dispersion of Magnetic Stimulation
verity of compression, segmental demyelination, focal remyelination or Wallerian degeneration will develop in large diameter motor fiber^.^ Any of these three lesions affecting the fastest conducting fibers results, at the whole nerve level, in lowering of the maximum conduction velocity with sparing of the minimum conduction velocity, since small motor fibers are largely intact. This corresponds to narrowing of the maximum-minimum range of velocities in the motor fibers to the detriment of the fastest conducting ones. l 2 The similarity of electrical and magnetic collision curves in ulnar neuropathy thus supports our interpretation of the differences between these curves in normal nerves. Since there are no functional fast-conducting fibers left in compressed ulnar nerves, the longitudinal dispersion of magnetic stimulation is no longer apparent. At this time, magnetic stimulation of peripheral nerves has several limitations: (i) spread to neighboring nerves4"; (ii) difficulties in obtaining supramaximal stimulation4; (iii) uncertainty in the location of the "virtual" cathode relative to the magnetic coil"'; and (iv) longitudinal dispersion of excitation of fibers of different diameter. These problems limit the routine use of magnetic stimulation in clinical studies of the peripheral nervous system. It is conceivable that further technical developments leading to coils generating smaller magnetic fields of similar magnitude will improve the focality of magnetic stimulation and make it more suitable for clinical studies.
MUSCLE 8, NERVE
November 1990
1081
REFERENCES
1. Barker AT, Jalinous R, Freeston IL: Non-invasive stimulation of human motor cortex. Lancet 1985;i:1106- 1107. 2. Berger AR, Busis NA, Logigian EL, Wierzbicka M, Shahani BT: Cervical root stimulation in the diagnosis of radiculopathy. Neurology 1987;37:329-332. 3. Domingue J, Shahani BT, Young RR: Conduction velocity in different diameter ulnar sensory and motor nerve fibers. Muscle Nerve 1980;3:437. 4. Evans BA, Litchy WJ, Daube JR: The utility of magnetic stimulation for routine peripheral nerve conduction studies. Muscle Nerve 1988;11:1074-1078. 5. Gilliatt RW: Chronic nerve compression and entrapment, in Sumner AJ (eds): The Physiology of Peripheral Nerve Diseuse, Philadelphia, Saunders, 1980, pp 316-339. 6. Hess CW, Mills KR, Murray NMF, Schriefer TN: Magnetic brain stimulation: Central motor conduction studies in Multiple Sclerosis. Ann Neurol 1987 ;22:744- 752. 7. Maccabee PJ, Amassian VE, Cracco RQ, Cadwell JA: An analysis of peripheral motor nerve stimulation in humans
1082
Spatial Dispersion of Magnetic Stimulation
using the magnetic coil. Electroenceph Clin Neurophysiol (in press). 8. Mills KR, Murray NMF: Electrical stimulation over the human vertebral column. Electroenceph Clin Neurophysiol 1986;63:582-589. 9. Ochoa J, Fowler TJ, Gilliatt RW: Anatomical changes in peripheral nerves compressed by a pneumatic tourniquet. J Anat 1972;113:433-455. 10. Santamaria J, King PL, Cros D, Chiappa KH: Cervical magnetic stimulation: roots or spinal nerves? Neurology 1988;38:199. 1 1 . Schriefer TN, Hess CW, Mills KR, Murray NMF: Central motor conduction studies in motor neuron disease using magnetic brain stimulation. Electroenceph Clin Neurophysiol (in press). 12. Shahani BT, Potts F, Juguilon A, Young RR: Maximalminimal motor nerve conduction and F response studies in normal subjects and patients with ulnar compression neuropathies. EEG Clin Neurophysiol 1980;50:172P.
MUSCLE & NERVE
November 1990
