Conduction along the sural nerve was studied in 64 normal subjects using near-nerve electrodes. Conduction velocities over the same nerve segments were calculated: (1) from the latency recorded from a site of stimulation to a site of recording (1R-method); and (2) from the difference in latency between 2 recording sites, the site of stimulation being situated elsewhere along the nerve (2R-method). Consistently faster velocities were seen with the 2R-method and could best be explained by a fixed delay of about 0.15 ms at the stimulus site (latency of activation, utilization time). This delay was markedly prolonged when a ramp rather than a rectangular stimulus was applied, though fast fibers were excited with both types of stimuli. The delay is thought to be dependent on the relationship between the density of current at the stimulus site and the threshold of responding fibers. Q 1992 John Wiley & Sons, Inc. Key words: compound sensory action potential conduction velocity stirnulus delay rectangular/ramp stimulation MUSCLE &? NERVE 1 5 8 1 3 4 2 1 1992
THE INFLUENCE OF THE STIMULUS ON NORMAL SURAL NERVE CONDUCTION VELOCITY: A STUDY OF THE LATENCY OF ACTlVATlON CHRISTIAN KRARUP, MD, STEVEN H. HOROWITZ, MD, and KRlSTlAN DAHL, BSc
T h e sensory conduction velocity (CV) is an important parameter in the electrophysiological evaluation of peripheral nerve (dys)function. Its value over a given nerve segment is determined by dividing the conduction distance by the latency between the site of stimulation and the site of recording (one recording point method, 1Rmethod) or by the differential latency between two sites of recording (two recording point method, 2R-method). A few studies have compared the CVs calculated by these 2 methods over the same
From the laboratories of Clinical Neurophysiology Department of Neurology, Long Island Jewish medical Center. the Long Island Campus lor the Albert Einstein College of Medicine, New Hyde Park (Dr. Horowitz); Department of Medicine (Neurology), Brigham and Women’s Hospital and the Department of Neurology, Harvard Medical School, Boston, Massachusetts (Dr Krarup); and the Department of Clinical Neurophysiology, University Hospital, Rigshospitalet, Denmark (Drs Krarup and Dahl). Presented in part at the annual meeting of the AAEM, Vancouver, 1991 Acknowledgment. We are grateful to Dr. W. Trojaborg for comments on the manuscript and to Prof A Rosenfalck for allowing us to include some subjects studied in collaboration with Aalborg University and forming the basis of a Masters thesis Address reprint requests to C Krarup, MD, Department of Clinical Neurophysiology, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen 0, Denmark. Accepted lor publication December 4, 1991 CCC 0148-639X1921070813-09$04.00 0 1992 John Wiley & Sons, lnc
Stimulus Latency of Activation
nerve segment2r14and found little or no difference. However, in a study of conduction along multiple segments of the sural nerve in normal subjects,” it was noted that CV values were dependent not only on the relative position of the segment in a proximo-distal direction but also on the method of calculation: the CV calculated by the 2K-method was consistently faster than the CV calculated by the 1K-method. If substantiated, this finding has implications for the routine evaluation of the CV in peripheral nerve assessment. I t was the aim of this study to examine the extent of this observation over different nerve segments irrespective of position. Our results suggest that the influence on CV could be explained by a delay of about 0.15 ms at the stimulus site (latency of activation, utilization time).
MATERIALS AND METHODS
The sural nerve was studied in 64 normal subjects (29 females, 35 males) with an average age of 40.0 years (range 11 to 80 years). They were selected from our larger material of 273 ~ u b j e c t s ’using ~ the criteria that the CV along various portions of the sural nerve from malleolus lateralis (ML) to lower calf (LC) and midcalf (MC) could be determined by both the 1R- and the 2R-methods. The
MUSCLE & NERVE
July 1992
813
temperature was kept above 35°C by an infrared, automatically controlled heating element. Electrode Placement. The sural nerve was stimulated and the evoked action potentials recorded through near-nerve needle electrodes as previously d e ~ c r i b e d . 'In ~ brief, the orthodromic SAP was evoked at dorsum pedis (DP) and recorded at ML, LC, and MC and evoked at ML and recorded at LC and MC (Fig. 1A). In 42 of the subjects, the nerve was further stimulated at LC and the SAP recorded at MC. An antidromic SAP evoked at LC was recorded at ML in 23 subjects.
The stimulus was a rectangular pulse, 0.2 ms in duration, applied from a constant current stimulator. At DP the stimulating electrodes were placed longitudinally at a distance of 10 to 15 mm, the cathode being closer to the recording sites. At ML and LC, the electrodes were placed transversely, the cathode (C) (Fig. 1B) close to the nerve and the anode (A) at a medial transverse distance of' 30 to 40 mm. In order to record the stimulus itself, the voltage drop through the tissue was measured via two needle electrodes placed between the cathode and the anode at the ML (electrodes 1-2, Fig. 1B). The stimulus strength was determined at threshold (0.1 to 0.5
Stimulation.
mA), to ensure proper electrode placement, and then increased to 10% above maximum for supramaximal stimulation (2 to 5 mA). In 5 subjects, the amplitude, area, duration, and CV of the SAP evoked by stimuli of different shapes were examined. The SAP was evoked at ML and recorded at LC and MC; the current stimulus was delivered from a normal constant current stimulator and, in addition, from a batterydriven stimulus isolator (BAK BSI-1) which was slaved to a signal generator. 'The injected stimulus was recorded and, in addition, the shape of the voltage drop through the tissue was recorded via two electrodes placed between the stimulus electrodes at ML (electrodes 1-2, Fig. 1B). The supramaximal current necessary to ensure maximal responses by a rectangular pulse with a duration of 0.2 ms was first determined. Then, ramp stimuli, which terminated vertically with predetermined durations, were injected and the amplitude of the ramp stimulus was selected to keep the area of injected current constant (see Fig. 4). At constant area, the amplitude of a ramp stimulus with a duration of 0.4 ms was the same as that of a rectangular pulse (defined as 1.0), and the angle was therefore defined as 45". The angle of the ramp of a pulse with, for example, a duration of' 0.8 ms would then be:
SURAL NERVE Midcalf, MC
L o w Calf, LC
Malleolus Lateralis, M L
D o r s u m Pedis, DP __*
FIGURE 1. Schematic representation of recording and stimulation sites along the sural nerve. (A) responses were evoked (S) at the dorsum of the foot (DP), the lateral malleolus (ML), the low calf (LC), and recorded (R) in various combinations of orthodromic and antidromic measurements at ML, LC, and the midcalf (MC). (B) The recording and the stimulating electrodes were placed transversely, the active recording electrode (AR) or the cathode (C) being close to the nerve, and the reference electrode (RR) or the anode (A) at a medial transverse distance of about 30 mm. The same electrode pair could be used for both recording and stimulation, as indicated by the switches to the right. The stimulus voltage in the tissue was recorded at the ankle via the electrodes (1 and 2) placed between the stimulus leads.
814
Stimulus Latency of Activation
MUSCLE & NERVE
July 1992
tan-'{0.5*1 amplitude/2*1 duration}= 14.0' and that of a ramp with a duration of 0.2 ms: tan-'{2/0.5}
=
76.0'.
Recording. T h e SAP was recorded through the electrodes placed close to the nerve at ML, LC, and MC (AR, Fig. 1B). The reference electrode (RR) was placed transversely at a distance of 30 to 40 mm from the recording electrode to obtain a unipolar recording. T h e SAP was amplitifed (Dantec 15C02,200 to 4000 Hz, 3 dB down), averaged (8 to 64 responses) in a Dantec averager with a sampling interval of' 10 ps, or a Nicolet digital oscilloscope 4094C with a sampling interval of 5 ps, and recorded at a sweep speed of 5 1 msicm. A sweep speed of 1 ms/cm allowed resolution at 0.03 ms. The peak-to-peak amplitude of the main component of the SAP was measured. In experiments where the sural nerve was excited using ramp stimuli, the duration and area of the negative phase of the SAP were measured using the digital oscilloscope. The latencies to the first positive peaks of the triphasic SAPS were measured and converted to the corresponding conduction velocities. With the 1R-method, the CV was calculated by dividing the conduction distance of a given nerve segment by the latency of' the evoked SAP between a site of stimulation and a site of recording at each segment end (Fig. 2). With the 2R-
4
2
S
4
6 "8
FIGURE 2. (Left) Orthodromic sural sensory action potential (SAP) evoked at ML (S, lateral malleolus) and recorded at MC (R, midcalf). The conduction velocity (CV) was calculated by the 1R-method (see text). (Right) SAP evoked at PD (S, dorsum pedis) and recorded at ML (RI) and MC (R2).The CV along DP-ML was calculated by the 1 R- and along ML-MC by the 2Rmethod.
segment, the CV was calculated by stimulating at a site distal to that segment, with both sets of electrodes at either end of the segment now serving as recording electrodes, and dividing the conduction distance by the differential latency of the SAP between these sites (Fig. 2). RESULTS
Conduction Velocities Calculated by the 1R- and 2RMethods. The CV from ML to MC calculated by
the 2K-method was 7.4 k 0.7% greater ( P < 0.0001, paired t-test; Fig. 2, Table 1) than by the 1R-method. Similarly, the CV along the segments ML-LC and LC-MC with stimulation at DP, ML, and LC was significantly ( P < 0.0001) greater when calculated by the 2R- than the 1R-method. The reduction in CV calculated by the 1R-method
Table 1. Conduction velocities along normal sural nerve segments as calculated by the 1R- and 2R-methods. Segment of sural nerve* ML-LC, 71
Site of stimulation* DP (64)11
?
0.5 mm
CV (rn/s)
Ca1c.t
53.6 t 0.7
2R
Diff.*
LC-MC, 60 CV (m/s) 57.4
?
0.9
-t
Calc.
47.9
f
0.5
1R
ML-MC, 131
Diff.
2R
5.8 ? 0.5 (Diff.) ML (64)
0.5 mrn
?
CV (rn/s)
Calc.
55.1 ? 0.6
2R
0.9 ? 0.7 (NS) 56.5 2 0.8
2R
0.6 mm Diff
3.8 f 0.5 (Diff.) 51.3 2 0.5
1R
7.3 ? 1.0s (Diff.)
LC (42)
49.6
f
0.8
1R
*Srtes of stimulation and recording: DP, dorsum pedis, ML, malleolus lateralis, LC. low calf; MC,midcalf. ?Method of calcuiation- conduction velocity (CV) between a site of stimulation and of recording (IR-method), and between two sites of recording (2R-method) #Difference in CV. GCornoarison between 42 subjects in whom DP, ML, and LC a// were sites of stimulation iiNumber of subjects Drff indicates a significant difference by pared t-test ( P < 0 0001) The difference between two values for the segment LC-MC calculated by the 2R-method was not significant (NS)
Stimulus Latency of Activation
MUSCLE & NERVE
July 1992
815
was about twice as pronounced along LC-MC compared with ML-MC (P < 0.0005, Table 1). In contrast, the CV along from LC to MC, calculated by the 2R-method, was similar (P < 0.2, Table 1) with stimulation at DP or ML. When calculated both by the 1R- and the 2Rmethods, the CV was lower along distally compared with more proximally placed nerve segments. Calculated by the PR-method, the SAP evoked at DP was conducted 3.8 -+ 1.0 m/s faster (P < 0.001) along LC-MC than along ML-LC. Using the 1R-method, the CV was 1.5 2 0.5 m/s ( P < 0.006) greater along LC-MC than d o n g ML-LC. The CV along ML-LC was 1.7 k 0.5 m/s greater ( P < 0.002) than along the further distally placed segment DP-ML. Latencies Measured by the 1 R - and the 2R-Method.
T o further analyze the basis for these differences in CV, the latency differences between the two methods of calculation were compared. Using analysis of variance (ANOVA) with repeated measures, latency differences along a given segment (ML-LC, LC-ML, LC-MC, and ML-MC) were evaluated by the 1R- and PR-methods, and also whether such differences were quantitatively dependent on the length and position of the segment. For the ML-LC and LC-MC segments, the latency was dependent on the type of measurement (P < 0.0001). As expected from the CV findings, the latency was shorter when measured by the 2R- than by the 1R-method. This difference between the 1K- and the PR-methods was independent (P > 0.2) of the site and length of the nerve segment, raising the possibility that the difference was a fixed quantity. This was further
-0.05
0.05
0.15
0.25
STIMULUS DELAY (ms)
0.35
supported by the finding that the combined latency of the orthodromic response from LC to MC and the antidromic response from LC to ML (i.e., two 1R-measurements) was significantly longer ( P < 0.0001) than that measured by the 1R-method from ML to MC. Combining the findings from all 64 subjects, the difference in latency between the 1R- and 2R-methods averaged 0.15 ? 0.01 ms (range -0.05 to +0.35 ms, Fig. 3A). That the additional latency associated with the stimulus was a constant value was also supported by the finding that the CV was relatively more reduced when calculated by the 1R-method the shorter the distance between the site of stimulation and the site of recording (Fig. 3B). Latency of the SAP Evoked by Submaximal and Max. imal Stimulation. In 1'7 subjects, the latency from
ML to MC was measured at a stimulus intensity which evoked a response which, on average, was 19% of that evoked by the supramaximal stimulus. The latency of the submaximal response was, on average, 3 1% longer ( P = 0.01) than that of the supramaximal response.
*
The Influence of the Configuration of the Stimulus.
The voltage drop of a rectangular pulse through the tissue had a rise time to 90% of maximum of 5 1 0 to 12 ks, which was slightly longer than that of the injected stimulus, 5 5 to 10 ps. The amplitude of the recorded stimulus indicated an impedance of about 100 to 120 ilt through the tissue, corresponding to the distance of 1 to 2 cm between the electrodes (1-2, Fig. 1B). When the ramp stimuli were compared with the recorded pulses, the shapes were closely similar, indicating
50
70
90
110
130
150
DISTANCE (mm)
FIGURE 3. (A) Frequency distribution of the stimulus delay in 64 subjects. ( 6 ) Relationship between the percentage difference in conduction velocity between the 2R- and the 1R-method, % difference = (CV(2R) - CV(IR))iOO/CV(lR), and the length of the nerve segment. The power function was calculated using the least square method after logarithmic transformation.The error bars indicate the SEM.
816
Stimulus Latency of Activation
MUSCLE 8, NERVE
July 1992
that there was little distortion of the time course of the voltage in the tissue (Fig. 4). The average CV determined by the 1Rmethod from the site of stimulation at ML to LC and to MC decreased linearly with decreasing angle of the stimulus (Figs. 5 and 6F), though the CV determined by the 2R-method along the segment LC-MC decreased only minimally (Fig. 6F). In the example (Fig. 5), the CV by the 1R-method decreased markedly, whereas the CV by the 2Rmethod was slightly variable. The amplitude of the SAP decreased logarithmically with stimulation by ramps with small angles, and this reduction occurred in parallel with marked prolongation of the duration of the SAP (Figs. 5, 6B, and E). T h e area of the response decreased much
4 A 1
Injected
stimulus
0.4ms
-
0.6ms
less-though in parallel with the peak-to-peak amplitude (Fig. 6C and D). DISCUSSION
The conduction velocity along a given segment of nerve was dependent on the method of calculation being lower between a site of stimulation and a site of recording than between two sites of recording. This difference indicates that the CV obtained in clinical assessment must be viewed according to the method of calculation. On the other hand, the discrepancy decreased with the distance over which conduction was assessed; at a distance of 15 to 20 cm, the difference between the two methods would be about 596, i.e., close to
4
3
2 14'
330
0.8ms
0.5ms
0.5ms
5
,
6
7
a
FIGURE 4. The injected current stimulus (upper trace in each panel) and the voltage drop through the tissue recorded via electrodes 1 and 2 (lower trace in each panel) (see Fig. lB) when the nerve was stimulated at lateral malleolus. The numbers indicate the sequence of study and the degrees indicate the slope of the angle.
Stimulus Latency of Activation
MUSCLE & NERVE
July 1992
817
4 01. 4
2
45.9 I
3
s 6
55.8
I\
7
YV
PV
FIGURE 5. In each panel, the SAP evoked at malleolus lateralis was recorded at low calf (upper trace), at midcalf (middle trace), and, in addition, the voltage drop through the tissue at the stimulus site (lower trace) was recorded. The conduction velocity above the upper trace was calculated by the 1R-method and between the upper and the middle trace by the 2R-method. The number above each panel indicates the sequence of study (see Fig. 4).
the limit of accuracy in clinical studies. Several factors should be considered to explain the difference in CV at different methods of calculation. Temporal dispersion: As the SAP latency is measured to its first positive peak, which is a summation of potentials from rapidly conducting large myelinated fibers in the nerve, temporal dispersion would cause the positive phase to be selectively derived from the fastest fibers to a greater degree at longer conduction distances. That the distance over which conduction is measured has an influence on the compound nerve action potential was noted in earlier s t ~ d i e s . ~ ~ " T . "h~e' ~ CV along a given portion of the nerve calculated from the 2R-method might be greater than from the 1R-method because of greater distance from the cathode. Thus, the 7% greater CV along ML-MC when the SAP was evoked at DP than at ML (Ta-
818
Stimulus Latency of Activation
ble 1) might be due to a shift of 0.5 to 1 km ofthe fiber diameter, which determines the first positive peak, considerin a conversion factor of 4.3 m/s per Consequently, with increasing distance from the cathode, the CV along a given portion of the nerve should increase toward a limit determined by the diameter of the largest fibers in the nerve. However, the CV calculated from the 2R-method along the segment LC-MC was similar when the SAP was evoked at DP and at ML (Table l), even though the distance between the two sites of stimulation was 88 rt_ 1 mm. Thus, the difference in CV according to the method of calculation over the relatively short distances examined here, could not be explained on the basis of temporal dispersion. At a markedly longer conduction distance, the effect of temporal dispersion became more noticeable, as shown by the mark-
MUSCLE 8, NERVE
July 1992
: : : c
1
1001
200
90 80 1
\
70 -
60 1 5 1
2
3
4
5
6
7
8
STIMULUS SEQUENCE
i
n
a
fn
0 ~I ~ 1 ~ l ~ , ~ , ~ l , 0 10 20 30 40 50 60 70 80 90 100 ANGLE (degree)
-
a
fn
1.
7 5 [ . , , 1 , 1 ~ , , , , 1 , 1 , [ , 1 50 ~ 0 10 20 30 40 50 60 70 80 90 100 80 ANGLE (degree)
135
I I
I
85
a
I
90
,.
%
,
95
j
I
,
.
100
.,. 105
SAP AREA ("A)
105
130
s
,
n
80 :
Y
i
>
5 s w
125
120
>
115
100 95 90
110 105
85
100 95 10 20 30 40 50 60 70 80 90 ANGLE (degree)
0
10 20 30 40 50 60 70 80 90 100 ANGLE (degree)
FIGURE 6. The influence of the stimulus configuration on the SAP. (A) The percentage difference from rectangular supramaximal stimulus: (a),angle; (+), area of injected stimulus; (0),area of recorded stimulus; (U),duration of stimulus and the sequence of the stimulus series (see also Figs. 4 and 5). The mean amplitude (B),area (C), duration (E),and conduction velocity (F) of the SAP evoked and )midcalf (0-0as )functions of the stimulus angle. (D) The relationship at malleolus lateralis and recorded at low calf (H between the amplitude and the area. The error bars indicate the SEM. The curves were calculated using the least square method in indicates the conduction velocity between LC and MC calculated by the (B), (C), and (E) after logarithmic transformation. (F) (DU) 2R-method.
edly polyphasic shape of the SAP when recorded proximally in the lower extremity or spinal root (cf. Fig. 3 in reference 15). Displacement of the site of stimulation: A "virtual
Stimulus Latency of Activation
cathode effect" occurs because a strong stimulus may excite the nerve at a site away from the position of the cathode. Usually, this effect tends to shorten the latency and to overestimate of the CV,
MUSCLE & NERVE
July 1992
819
l
,
l
and the slower conduction with the 1R-method would only occur if the virtual cathode is displaced away from the recording electrode. ACcording to this assumption, collision might occur with stimulation at LC and simultaneous recording of the antidromic response at ML and the orthodromic response at MC. Even though a delay occurred in both directions of recording, occlusion did not occur indicating that the “virtual cathode” was not displaced away from the site of recording. Using near-nerve needle electrodes for stimulation has the advantage that smaller stimulus curr e n t ~ * ’ ~are ” necessary to activate the nerve and, hence, the likelihood of pronounced stimulus spread is reduced. The average reduction in latency when the submaximal was compared with the supramaximal stimulus was 0.07 2 0.03 ms, which corresponded to 3% of the overall latency and similar to the 0.05 to 0.15 ms reduction in latency found by Gilliatt et al.14 Assuming that fast fibers were activated at submaximal stimulation, the “virtual cathode” appeared to be displaced slightly proximally toward the recordin site as would be expected from other However, as indicated below, the delay of excitation may be longer at low levels of stimulation causing a slightly longer latency. Delay of excitation at the site of stimulation: An excitation or conduction delay at the cathode ( “ N u t z ~ e i t ” ’or ~ utilization time) is possible if nerve excitation does not occur at the rising front of a rectangular stimulus, or if a period of acceleration occurs before a plateau conduction at maximal speed is achieved. Thus, in experimental nerve conduction studies, the delay was considered to be about 0.1 ms.” Variable delays of the nerve action potential evoked by electrical stimulation are described in early studies on conduction in peripheral l 2 and are thought to be due to delays in the excitation process. In theoretical models of compound nerve action potentials, a delay due to the stimulus was introduced by Cummins et a1.,8 whereas this factor was ignored in other When noted, the delay was considered to be longer in small, slowly conducting fibers with a higher thresholdg than in large fibers“ and more pronounced at threshold than at maximal ~tirnulation.~ Even though the delay may be reduced by increasing the depolarizing voltage, the exact relationship between these two factors is difficult to determine due to the spread of the stimulus at high stimulus i n t e n ~ i t y . ~ ~ ’ ~ ~ ~ ~ From the differences in latencies at different
820
Stimulus Latency of Activation
nerve segments in this study, the delay was calculated at about 0.15 ms which is similar to the utilization time cited in the literature (cf. Kimura“). The delay ranged from -0.05 to 0.35 ms in our 64 normal subjects. This variability may be due to at least three causes: (1) inaccuracies in measurement, conduction was measured over short distances which increased the relevance of even small inaccuracies in the measurement of distance or latency; (2) effects of stimulus levels, even though the stimulus was adjusted to be supramaximal, the density of depolarizing current at the membrane may have been variable causing differences in the delay; or (3) activation of fibers of dfferent diameter, the stimulus at more distal sites may, in some instances, have activated smaller fibers than those activated at the more proximal site, thus reducing or even reversing the difference in velocity calculated by the 2K- and 1R-rnethod~.’~ Recording of the actual voltage in the tissue induced by the applied stimulus indicated that the delay in excitation at electrical stimulation with rectangular stimuli could not be explained on the basis of‘ a delay in the rise of the depolarizing voltage. The rise time of the stimulus was measured at about 10 ps, which is an upper limit. In order to measure a true, and probably faster rise time, an upper limiting frequency of >100,000 Hz of the recording set-up would have been necessary. Hence, it seems most likely that the delay is due to an interplay between density of the depolarizing current and the resulting membrane reaction. The delay in excitation was markedly dependent on the rise of the stimulus at different angles. This delay was increased at even very small reductions from 90” to 76”, and the relationship between the delay and the angle was linearly reciprocal, suggesting that there was a continuous relationship between the stimulus and the delay rather than an abrupt difference between the ramp and the rectangular stimulus. With stimulation at ML the CV, when calculated by the 2Rmethod, between the two recording sites LC and MC changed little if at all, indicating that the reduction in CV from ML to LC was due primarily to a prolongation of the delay of activation of large fibers rather than to activation of thinner fibers. T h e amplitude of the SAP decreased markedly with stimulation by ramps with small angles (Fig. 6B), and this reduction in amplitude was associated with marked prolongation of the duration of the SAP (Fig. 6E), reducing the area of the response much less than the peak-to-peak amplitude (Fig. 6C and D). Taken together with the ev-
MUSCLE & NERVE
July 1992
idence of preserved activation of large fibers, this finding suggests that fibers with smaller diameters were more delayed by the slope of the stimulus than fast-conducting large fibers. The difference in effect corresponds to earlier findings that the utilization time is longer in small than large myelinated fibers.“ In cat experiments where the stimulus was applied through circumferential leads to the nerve placed within a silicone nerve cuff,” the delay as calculated from differences between the 1R- and 0.006 ms (n = 36, P < 2R-methods was 0.06 0.001) (Krarup, unpublished results). In this setup, the current density at the stimulation site was increased markedly as indicated by the low supramaximal current (0.3 to 0.6 mA) due to the insulating properties of the silicone cuff. Moreover, the CV of the fastest fibers is almost twice that in humans. These differences suggest that the delay in excitation is a variable quantity dependent on both the fibers in the nerve and the stimulus applied.
*
~
~~
REFERENCES 1. Behse F: Morphometric studies on the human sural nerve. Acta Neurol Scand 1990;82:(suppl. 132)5-38. 2. Behse F, Buchthal F: Normal sensory conduction in the nerves of the leg in man. J Neurol Neurosurg Psychiatry 1971;34:404-414. 3. Behse F, Buchthal F: Sensory action potentials and biopsy of the sural nerve in neuropathy. Bruin 1978;101:473493. 4. Blair EA, Erlanger J: A comparison of the characteristics of axons through their individual electrical responses. Am J Physiol 1933;106:524-564. 5. Blair EA, Erlanger J: O n the process of excitation by brief shocks in axons. A m ] Physiol 1935;114:309-316. 6. Buchthal F, Rosenfalck A: Evoked action potentials and conduction velocity in human sensory nerves. Bruin Res 1966;3:1- 122. 7. Buchthal F, Rosenfalck A, Behse F: Sensory potentids of normal and diseased nerves, in Dyck PJ, Thomas PK, Lambert
Stimulus Latency of Activation
EH, Bunge R (eds): Pripheral Neuropathy (2 ed). Philadelphia, Saunders, 1984, pp 981-1015. 8. Cumrnins KL, Perkel DH, Dorfrnan LJ: Nerve fiber conductiori-velocity distributions. I. Estimation based on the single fiber and compound action potentials. Electroencephalogr Clin Neurophysiol 19 79 ;46 :634 - 646. 9. Dawson GD: T h e relative excitability and conduction velocity of sensory and motor nerve fibres in man. J Physiol (Londj 1956;131:436-451. 10. Erlanger J , Blair EA: The configuration of axon and “sitnple” nerve action potentials. A m J Physiol 1933;106:565570. 11. Gasser HS, Erlanger J: T h e role played by the sizes of the constituent fibers of a nerve trunk in determining the form of its action potential wave. Am J Physiol 1927;80:522-547. 12. Gasser HS, Grundfest H: Axon diameters in relation to the spike dimensions and the conduction velocity in marnmalian A fibers. Am J Physiol 1939;127:393-414. 13. Gilderneister M: Die allgemeinen Gesetze des elektrischen Reizes. I. Die Nutzzeit und ihre Gesetze. Z Biol 1913;62:358-396. 14. Gilliatt RW, Melville ID, Velate AS, Willison RG: A study of normal nerve action potentials using an averaging technique (barrier grid storage tube). ] Neurol Neurosurg Psychiatry 1965;28:191- 200. 15. Horowitz SH, Krarup C: Conduction studies of the normal s u r d nerve. Muscle Nerue 1992;15;374-383. 16. Kimura J: Electrodiagnosis in Diseases of Nerue and Muscle: Princzples and Practice ( 2 ed). Philadelphia, FA Davis, 1989, pp 629- 630. 17. Krarup C, Loeb GE: Conduction studies in peripheral cat nerve using implanted electrodes: I. Methods and findings in controls. Muscle Nerve 1988;11:922-932. 18. McDonald WI: T h e effects of experimental demyelination on conduction in peripheral nerve: A histological and electrophysiological study. 11. Electrophysiological observations. Brain 1963;86:501-524. 19. Podivinski F: The effect of stimulus intensity on latency of nerve action potential in healthy subjects and in patients with peripheral nerve lesions. Neurology 1965;15: 10591062. 20. Schoonhoven R, Schellens RLLA, Stegeman DF, GabreelsFesten AAWM: Sensory potentials and sural nerve biopsy: a model evaluation. Muscle Nerve 1987;10:246-262. 21. Stegeman DF, d e Weerd JPC: Modelling compound action potentials of peripheral nerves in situ. 1. Model description: evidence for a non-linear relation between fibre diameter and velocity. Ebctroencephalogr Clin Neurophysiol 1982;54:436-448. 22. Wiederholt WC: Threshold and conduction velocity in isolated mixed mammalian nerves. Neurology 1970;20:347352.
MUSCLE & NERVE
July 1992
821
THE INFLUENCE OF THE STIMULUS ON NORMAL SURAL NERVE CONDUCTION VELOCITY: A STUDY OF THE LATENCY OF ACTlVATlON CHRISTIAN KRARUP, MD, STEVEN H. HOROWITZ, MD, and KRlSTlAN DAHL, BSc
T h e sensory conduction velocity (CV) is an important parameter in the electrophysiological evaluation of peripheral nerve (dys)function. Its value over a given nerve segment is determined by dividing the conduction distance by the latency between the site of stimulation and the site of recording (one recording point method, 1Rmethod) or by the differential latency between two sites of recording (two recording point method, 2R-method). A few studies have compared the CVs calculated by these 2 methods over the same
From the laboratories of Clinical Neurophysiology Department of Neurology, Long Island Jewish medical Center. the Long Island Campus lor the Albert Einstein College of Medicine, New Hyde Park (Dr. Horowitz); Department of Medicine (Neurology), Brigham and Women’s Hospital and the Department of Neurology, Harvard Medical School, Boston, Massachusetts (Dr Krarup); and the Department of Clinical Neurophysiology, University Hospital, Rigshospitalet, Denmark (Drs Krarup and Dahl). Presented in part at the annual meeting of the AAEM, Vancouver, 1991 Acknowledgment. We are grateful to Dr. W. Trojaborg for comments on the manuscript and to Prof A Rosenfalck for allowing us to include some subjects studied in collaboration with Aalborg University and forming the basis of a Masters thesis Address reprint requests to C Krarup, MD, Department of Clinical Neurophysiology, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen 0, Denmark. Accepted lor publication December 4, 1991 CCC 0148-639X1921070813-09$04.00 0 1992 John Wiley & Sons, lnc
Stimulus Latency of Activation
nerve segment2r14and found little or no difference. However, in a study of conduction along multiple segments of the sural nerve in normal subjects,” it was noted that CV values were dependent not only on the relative position of the segment in a proximo-distal direction but also on the method of calculation: the CV calculated by the 2K-method was consistently faster than the CV calculated by the 1K-method. If substantiated, this finding has implications for the routine evaluation of the CV in peripheral nerve assessment. I t was the aim of this study to examine the extent of this observation over different nerve segments irrespective of position. Our results suggest that the influence on CV could be explained by a delay of about 0.15 ms at the stimulus site (latency of activation, utilization time).
MATERIALS AND METHODS
The sural nerve was studied in 64 normal subjects (29 females, 35 males) with an average age of 40.0 years (range 11 to 80 years). They were selected from our larger material of 273 ~ u b j e c t s ’using ~ the criteria that the CV along various portions of the sural nerve from malleolus lateralis (ML) to lower calf (LC) and midcalf (MC) could be determined by both the 1R- and the 2R-methods. The
MUSCLE & NERVE
July 1992
813
temperature was kept above 35°C by an infrared, automatically controlled heating element. Electrode Placement. The sural nerve was stimulated and the evoked action potentials recorded through near-nerve needle electrodes as previously d e ~ c r i b e d . 'In ~ brief, the orthodromic SAP was evoked at dorsum pedis (DP) and recorded at ML, LC, and MC and evoked at ML and recorded at LC and MC (Fig. 1A). In 42 of the subjects, the nerve was further stimulated at LC and the SAP recorded at MC. An antidromic SAP evoked at LC was recorded at ML in 23 subjects.
The stimulus was a rectangular pulse, 0.2 ms in duration, applied from a constant current stimulator. At DP the stimulating electrodes were placed longitudinally at a distance of 10 to 15 mm, the cathode being closer to the recording sites. At ML and LC, the electrodes were placed transversely, the cathode (C) (Fig. 1B) close to the nerve and the anode (A) at a medial transverse distance of' 30 to 40 mm. In order to record the stimulus itself, the voltage drop through the tissue was measured via two needle electrodes placed between the cathode and the anode at the ML (electrodes 1-2, Fig. 1B). The stimulus strength was determined at threshold (0.1 to 0.5
Stimulation.
mA), to ensure proper electrode placement, and then increased to 10% above maximum for supramaximal stimulation (2 to 5 mA). In 5 subjects, the amplitude, area, duration, and CV of the SAP evoked by stimuli of different shapes were examined. The SAP was evoked at ML and recorded at LC and MC; the current stimulus was delivered from a normal constant current stimulator and, in addition, from a batterydriven stimulus isolator (BAK BSI-1) which was slaved to a signal generator. 'The injected stimulus was recorded and, in addition, the shape of the voltage drop through the tissue was recorded via two electrodes placed between the stimulus electrodes at ML (electrodes 1-2, Fig. 1B). The supramaximal current necessary to ensure maximal responses by a rectangular pulse with a duration of 0.2 ms was first determined. Then, ramp stimuli, which terminated vertically with predetermined durations, were injected and the amplitude of the ramp stimulus was selected to keep the area of injected current constant (see Fig. 4). At constant area, the amplitude of a ramp stimulus with a duration of 0.4 ms was the same as that of a rectangular pulse (defined as 1.0), and the angle was therefore defined as 45". The angle of the ramp of a pulse with, for example, a duration of' 0.8 ms would then be:
SURAL NERVE Midcalf, MC
L o w Calf, LC
Malleolus Lateralis, M L
D o r s u m Pedis, DP __*
FIGURE 1. Schematic representation of recording and stimulation sites along the sural nerve. (A) responses were evoked (S) at the dorsum of the foot (DP), the lateral malleolus (ML), the low calf (LC), and recorded (R) in various combinations of orthodromic and antidromic measurements at ML, LC, and the midcalf (MC). (B) The recording and the stimulating electrodes were placed transversely, the active recording electrode (AR) or the cathode (C) being close to the nerve, and the reference electrode (RR) or the anode (A) at a medial transverse distance of about 30 mm. The same electrode pair could be used for both recording and stimulation, as indicated by the switches to the right. The stimulus voltage in the tissue was recorded at the ankle via the electrodes (1 and 2) placed between the stimulus leads.
814
Stimulus Latency of Activation
MUSCLE & NERVE
July 1992
tan-'{0.5*1 amplitude/2*1 duration}= 14.0' and that of a ramp with a duration of 0.2 ms: tan-'{2/0.5}
=
76.0'.
Recording. T h e SAP was recorded through the electrodes placed close to the nerve at ML, LC, and MC (AR, Fig. 1B). The reference electrode (RR) was placed transversely at a distance of 30 to 40 mm from the recording electrode to obtain a unipolar recording. T h e SAP was amplitifed (Dantec 15C02,200 to 4000 Hz, 3 dB down), averaged (8 to 64 responses) in a Dantec averager with a sampling interval of' 10 ps, or a Nicolet digital oscilloscope 4094C with a sampling interval of 5 ps, and recorded at a sweep speed of 5 1 msicm. A sweep speed of 1 ms/cm allowed resolution at 0.03 ms. The peak-to-peak amplitude of the main component of the SAP was measured. In experiments where the sural nerve was excited using ramp stimuli, the duration and area of the negative phase of the SAP were measured using the digital oscilloscope. The latencies to the first positive peaks of the triphasic SAPS were measured and converted to the corresponding conduction velocities. With the 1R-method, the CV was calculated by dividing the conduction distance of a given nerve segment by the latency of' the evoked SAP between a site of stimulation and a site of recording at each segment end (Fig. 2). With the 2R-
4
2
S
4
6 "8
FIGURE 2. (Left) Orthodromic sural sensory action potential (SAP) evoked at ML (S, lateral malleolus) and recorded at MC (R, midcalf). The conduction velocity (CV) was calculated by the 1R-method (see text). (Right) SAP evoked at PD (S, dorsum pedis) and recorded at ML (RI) and MC (R2).The CV along DP-ML was calculated by the 1 R- and along ML-MC by the 2Rmethod.
segment, the CV was calculated by stimulating at a site distal to that segment, with both sets of electrodes at either end of the segment now serving as recording electrodes, and dividing the conduction distance by the differential latency of the SAP between these sites (Fig. 2). RESULTS
Conduction Velocities Calculated by the 1R- and 2RMethods. The CV from ML to MC calculated by
the 2K-method was 7.4 k 0.7% greater ( P < 0.0001, paired t-test; Fig. 2, Table 1) than by the 1R-method. Similarly, the CV along the segments ML-LC and LC-MC with stimulation at DP, ML, and LC was significantly ( P < 0.0001) greater when calculated by the 2R- than the 1R-method. The reduction in CV calculated by the 1R-method
Table 1. Conduction velocities along normal sural nerve segments as calculated by the 1R- and 2R-methods. Segment of sural nerve* ML-LC, 71
Site of stimulation* DP (64)11
?
0.5 mm
CV (rn/s)
Ca1c.t
53.6 t 0.7
2R
Diff.*
LC-MC, 60 CV (m/s) 57.4
?
0.9
-t
Calc.
47.9
f
0.5
1R
ML-MC, 131
Diff.
2R
5.8 ? 0.5 (Diff.) ML (64)
0.5 mrn
?
CV (rn/s)
Calc.
55.1 ? 0.6
2R
0.9 ? 0.7 (NS) 56.5 2 0.8
2R
0.6 mm Diff
3.8 f 0.5 (Diff.) 51.3 2 0.5
1R
7.3 ? 1.0s (Diff.)
LC (42)
49.6
f
0.8
1R
*Srtes of stimulation and recording: DP, dorsum pedis, ML, malleolus lateralis, LC. low calf; MC,midcalf. ?Method of calcuiation- conduction velocity (CV) between a site of stimulation and of recording (IR-method), and between two sites of recording (2R-method) #Difference in CV. GCornoarison between 42 subjects in whom DP, ML, and LC a// were sites of stimulation iiNumber of subjects Drff indicates a significant difference by pared t-test ( P < 0 0001) The difference between two values for the segment LC-MC calculated by the 2R-method was not significant (NS)
Stimulus Latency of Activation
MUSCLE & NERVE
July 1992
815
was about twice as pronounced along LC-MC compared with ML-MC (P < 0.0005, Table 1). In contrast, the CV along from LC to MC, calculated by the 2R-method, was similar (P < 0.2, Table 1) with stimulation at DP or ML. When calculated both by the 1R- and the 2Rmethods, the CV was lower along distally compared with more proximally placed nerve segments. Calculated by the PR-method, the SAP evoked at DP was conducted 3.8 -+ 1.0 m/s faster (P < 0.001) along LC-MC than along ML-LC. Using the 1R-method, the CV was 1.5 2 0.5 m/s ( P < 0.006) greater along LC-MC than d o n g ML-LC. The CV along ML-LC was 1.7 k 0.5 m/s greater ( P < 0.002) than along the further distally placed segment DP-ML. Latencies Measured by the 1 R - and the 2R-Method.
T o further analyze the basis for these differences in CV, the latency differences between the two methods of calculation were compared. Using analysis of variance (ANOVA) with repeated measures, latency differences along a given segment (ML-LC, LC-ML, LC-MC, and ML-MC) were evaluated by the 1R- and PR-methods, and also whether such differences were quantitatively dependent on the length and position of the segment. For the ML-LC and LC-MC segments, the latency was dependent on the type of measurement (P < 0.0001). As expected from the CV findings, the latency was shorter when measured by the 2R- than by the 1R-method. This difference between the 1K- and the PR-methods was independent (P > 0.2) of the site and length of the nerve segment, raising the possibility that the difference was a fixed quantity. This was further
-0.05
0.05
0.15
0.25
STIMULUS DELAY (ms)
0.35
supported by the finding that the combined latency of the orthodromic response from LC to MC and the antidromic response from LC to ML (i.e., two 1R-measurements) was significantly longer ( P < 0.0001) than that measured by the 1R-method from ML to MC. Combining the findings from all 64 subjects, the difference in latency between the 1R- and 2R-methods averaged 0.15 ? 0.01 ms (range -0.05 to +0.35 ms, Fig. 3A). That the additional latency associated with the stimulus was a constant value was also supported by the finding that the CV was relatively more reduced when calculated by the 1R-method the shorter the distance between the site of stimulation and the site of recording (Fig. 3B). Latency of the SAP Evoked by Submaximal and Max. imal Stimulation. In 1'7 subjects, the latency from
ML to MC was measured at a stimulus intensity which evoked a response which, on average, was 19% of that evoked by the supramaximal stimulus. The latency of the submaximal response was, on average, 3 1% longer ( P = 0.01) than that of the supramaximal response.
*
The Influence of the Configuration of the Stimulus.
The voltage drop of a rectangular pulse through the tissue had a rise time to 90% of maximum of 5 1 0 to 12 ks, which was slightly longer than that of the injected stimulus, 5 5 to 10 ps. The amplitude of the recorded stimulus indicated an impedance of about 100 to 120 ilt through the tissue, corresponding to the distance of 1 to 2 cm between the electrodes (1-2, Fig. 1B). When the ramp stimuli were compared with the recorded pulses, the shapes were closely similar, indicating
50
70
90
110
130
150
DISTANCE (mm)
FIGURE 3. (A) Frequency distribution of the stimulus delay in 64 subjects. ( 6 ) Relationship between the percentage difference in conduction velocity between the 2R- and the 1R-method, % difference = (CV(2R) - CV(IR))iOO/CV(lR), and the length of the nerve segment. The power function was calculated using the least square method after logarithmic transformation.The error bars indicate the SEM.
816
Stimulus Latency of Activation
MUSCLE 8, NERVE
July 1992
that there was little distortion of the time course of the voltage in the tissue (Fig. 4). The average CV determined by the 1Rmethod from the site of stimulation at ML to LC and to MC decreased linearly with decreasing angle of the stimulus (Figs. 5 and 6F), though the CV determined by the 2R-method along the segment LC-MC decreased only minimally (Fig. 6F). In the example (Fig. 5), the CV by the 1R-method decreased markedly, whereas the CV by the 2Rmethod was slightly variable. The amplitude of the SAP decreased logarithmically with stimulation by ramps with small angles, and this reduction occurred in parallel with marked prolongation of the duration of the SAP (Figs. 5, 6B, and E). T h e area of the response decreased much
4 A 1
Injected
stimulus
0.4ms
-
0.6ms
less-though in parallel with the peak-to-peak amplitude (Fig. 6C and D). DISCUSSION
The conduction velocity along a given segment of nerve was dependent on the method of calculation being lower between a site of stimulation and a site of recording than between two sites of recording. This difference indicates that the CV obtained in clinical assessment must be viewed according to the method of calculation. On the other hand, the discrepancy decreased with the distance over which conduction was assessed; at a distance of 15 to 20 cm, the difference between the two methods would be about 596, i.e., close to
4
3
2 14'
330
0.8ms
0.5ms
0.5ms
5
,
6
7
a
FIGURE 4. The injected current stimulus (upper trace in each panel) and the voltage drop through the tissue recorded via electrodes 1 and 2 (lower trace in each panel) (see Fig. lB) when the nerve was stimulated at lateral malleolus. The numbers indicate the sequence of study and the degrees indicate the slope of the angle.
Stimulus Latency of Activation
MUSCLE & NERVE
July 1992
817
4 01. 4
2
45.9 I
3
s 6
55.8
I\
7
YV
PV
FIGURE 5. In each panel, the SAP evoked at malleolus lateralis was recorded at low calf (upper trace), at midcalf (middle trace), and, in addition, the voltage drop through the tissue at the stimulus site (lower trace) was recorded. The conduction velocity above the upper trace was calculated by the 1R-method and between the upper and the middle trace by the 2R-method. The number above each panel indicates the sequence of study (see Fig. 4).
the limit of accuracy in clinical studies. Several factors should be considered to explain the difference in CV at different methods of calculation. Temporal dispersion: As the SAP latency is measured to its first positive peak, which is a summation of potentials from rapidly conducting large myelinated fibers in the nerve, temporal dispersion would cause the positive phase to be selectively derived from the fastest fibers to a greater degree at longer conduction distances. That the distance over which conduction is measured has an influence on the compound nerve action potential was noted in earlier s t ~ d i e s . ~ ~ " T . "h~e' ~ CV along a given portion of the nerve calculated from the 2R-method might be greater than from the 1R-method because of greater distance from the cathode. Thus, the 7% greater CV along ML-MC when the SAP was evoked at DP than at ML (Ta-
818
Stimulus Latency of Activation
ble 1) might be due to a shift of 0.5 to 1 km ofthe fiber diameter, which determines the first positive peak, considerin a conversion factor of 4.3 m/s per Consequently, with increasing distance from the cathode, the CV along a given portion of the nerve should increase toward a limit determined by the diameter of the largest fibers in the nerve. However, the CV calculated from the 2R-method along the segment LC-MC was similar when the SAP was evoked at DP and at ML (Table l), even though the distance between the two sites of stimulation was 88 rt_ 1 mm. Thus, the difference in CV according to the method of calculation over the relatively short distances examined here, could not be explained on the basis of temporal dispersion. At a markedly longer conduction distance, the effect of temporal dispersion became more noticeable, as shown by the mark-
MUSCLE 8, NERVE
July 1992
: : : c
1
1001
200
90 80 1
\
70 -
60 1 5 1
2
3
4
5
6
7
8
STIMULUS SEQUENCE
i
n
a
fn
0 ~I ~ 1 ~ l ~ , ~ , ~ l , 0 10 20 30 40 50 60 70 80 90 100 ANGLE (degree)
-
a
fn
1.
7 5 [ . , , 1 , 1 ~ , , , , 1 , 1 , [ , 1 50 ~ 0 10 20 30 40 50 60 70 80 90 100 80 ANGLE (degree)
135
I I
I
85
a
I
90
,.
%
,
95
j
I
,
.
100
.,. 105
SAP AREA ("A)
105
130
s
,
n
80 :
Y
i
>
5 s w
125
120
>
115
100 95 90
110 105
85
100 95 10 20 30 40 50 60 70 80 90 ANGLE (degree)
0
10 20 30 40 50 60 70 80 90 100 ANGLE (degree)
FIGURE 6. The influence of the stimulus configuration on the SAP. (A) The percentage difference from rectangular supramaximal stimulus: (a),angle; (+), area of injected stimulus; (0),area of recorded stimulus; (U),duration of stimulus and the sequence of the stimulus series (see also Figs. 4 and 5). The mean amplitude (B),area (C), duration (E),and conduction velocity (F) of the SAP evoked and )midcalf (0-0as )functions of the stimulus angle. (D) The relationship at malleolus lateralis and recorded at low calf (H between the amplitude and the area. The error bars indicate the SEM. The curves were calculated using the least square method in indicates the conduction velocity between LC and MC calculated by the (B), (C), and (E) after logarithmic transformation. (F) (DU) 2R-method.
edly polyphasic shape of the SAP when recorded proximally in the lower extremity or spinal root (cf. Fig. 3 in reference 15). Displacement of the site of stimulation: A "virtual
Stimulus Latency of Activation
cathode effect" occurs because a strong stimulus may excite the nerve at a site away from the position of the cathode. Usually, this effect tends to shorten the latency and to overestimate of the CV,
MUSCLE & NERVE
July 1992
819
l
,
l
and the slower conduction with the 1R-method would only occur if the virtual cathode is displaced away from the recording electrode. ACcording to this assumption, collision might occur with stimulation at LC and simultaneous recording of the antidromic response at ML and the orthodromic response at MC. Even though a delay occurred in both directions of recording, occlusion did not occur indicating that the “virtual cathode” was not displaced away from the site of recording. Using near-nerve needle electrodes for stimulation has the advantage that smaller stimulus curr e n t ~ * ’ ~are ” necessary to activate the nerve and, hence, the likelihood of pronounced stimulus spread is reduced. The average reduction in latency when the submaximal was compared with the supramaximal stimulus was 0.07 2 0.03 ms, which corresponded to 3% of the overall latency and similar to the 0.05 to 0.15 ms reduction in latency found by Gilliatt et al.14 Assuming that fast fibers were activated at submaximal stimulation, the “virtual cathode” appeared to be displaced slightly proximally toward the recordin site as would be expected from other However, as indicated below, the delay of excitation may be longer at low levels of stimulation causing a slightly longer latency. Delay of excitation at the site of stimulation: An excitation or conduction delay at the cathode ( “ N u t z ~ e i t ” ’or ~ utilization time) is possible if nerve excitation does not occur at the rising front of a rectangular stimulus, or if a period of acceleration occurs before a plateau conduction at maximal speed is achieved. Thus, in experimental nerve conduction studies, the delay was considered to be about 0.1 ms.” Variable delays of the nerve action potential evoked by electrical stimulation are described in early studies on conduction in peripheral l 2 and are thought to be due to delays in the excitation process. In theoretical models of compound nerve action potentials, a delay due to the stimulus was introduced by Cummins et a1.,8 whereas this factor was ignored in other When noted, the delay was considered to be longer in small, slowly conducting fibers with a higher thresholdg than in large fibers“ and more pronounced at threshold than at maximal ~tirnulation.~ Even though the delay may be reduced by increasing the depolarizing voltage, the exact relationship between these two factors is difficult to determine due to the spread of the stimulus at high stimulus i n t e n ~ i t y . ~ ~ ’ ~ ~ ~ ~ From the differences in latencies at different
820
Stimulus Latency of Activation
nerve segments in this study, the delay was calculated at about 0.15 ms which is similar to the utilization time cited in the literature (cf. Kimura“). The delay ranged from -0.05 to 0.35 ms in our 64 normal subjects. This variability may be due to at least three causes: (1) inaccuracies in measurement, conduction was measured over short distances which increased the relevance of even small inaccuracies in the measurement of distance or latency; (2) effects of stimulus levels, even though the stimulus was adjusted to be supramaximal, the density of depolarizing current at the membrane may have been variable causing differences in the delay; or (3) activation of fibers of dfferent diameter, the stimulus at more distal sites may, in some instances, have activated smaller fibers than those activated at the more proximal site, thus reducing or even reversing the difference in velocity calculated by the 2K- and 1R-rnethod~.’~ Recording of the actual voltage in the tissue induced by the applied stimulus indicated that the delay in excitation at electrical stimulation with rectangular stimuli could not be explained on the basis of‘ a delay in the rise of the depolarizing voltage. The rise time of the stimulus was measured at about 10 ps, which is an upper limit. In order to measure a true, and probably faster rise time, an upper limiting frequency of >100,000 Hz of the recording set-up would have been necessary. Hence, it seems most likely that the delay is due to an interplay between density of the depolarizing current and the resulting membrane reaction. The delay in excitation was markedly dependent on the rise of the stimulus at different angles. This delay was increased at even very small reductions from 90” to 76”, and the relationship between the delay and the angle was linearly reciprocal, suggesting that there was a continuous relationship between the stimulus and the delay rather than an abrupt difference between the ramp and the rectangular stimulus. With stimulation at ML the CV, when calculated by the 2Rmethod, between the two recording sites LC and MC changed little if at all, indicating that the reduction in CV from ML to LC was due primarily to a prolongation of the delay of activation of large fibers rather than to activation of thinner fibers. T h e amplitude of the SAP decreased markedly with stimulation by ramps with small angles (Fig. 6B), and this reduction in amplitude was associated with marked prolongation of the duration of the SAP (Fig. 6E), reducing the area of the response much less than the peak-to-peak amplitude (Fig. 6C and D). Taken together with the ev-
MUSCLE & NERVE
July 1992
idence of preserved activation of large fibers, this finding suggests that fibers with smaller diameters were more delayed by the slope of the stimulus than fast-conducting large fibers. The difference in effect corresponds to earlier findings that the utilization time is longer in small than large myelinated fibers.“ In cat experiments where the stimulus was applied through circumferential leads to the nerve placed within a silicone nerve cuff,” the delay as calculated from differences between the 1R- and 0.006 ms (n = 36, P < 2R-methods was 0.06 0.001) (Krarup, unpublished results). In this setup, the current density at the stimulation site was increased markedly as indicated by the low supramaximal current (0.3 to 0.6 mA) due to the insulating properties of the silicone cuff. Moreover, the CV of the fastest fibers is almost twice that in humans. These differences suggest that the delay in excitation is a variable quantity dependent on both the fibers in the nerve and the stimulus applied.
*
~
~~
REFERENCES 1. Behse F: Morphometric studies on the human sural nerve. Acta Neurol Scand 1990;82:(suppl. 132)5-38. 2. Behse F, Buchthal F: Normal sensory conduction in the nerves of the leg in man. J Neurol Neurosurg Psychiatry 1971;34:404-414. 3. Behse F, Buchthal F: Sensory action potentials and biopsy of the sural nerve in neuropathy. Bruin 1978;101:473493. 4. Blair EA, Erlanger J: A comparison of the characteristics of axons through their individual electrical responses. Am J Physiol 1933;106:524-564. 5. Blair EA, Erlanger J: O n the process of excitation by brief shocks in axons. A m ] Physiol 1935;114:309-316. 6. Buchthal F, Rosenfalck A: Evoked action potentials and conduction velocity in human sensory nerves. Bruin Res 1966;3:1- 122. 7. Buchthal F, Rosenfalck A, Behse F: Sensory potentids of normal and diseased nerves, in Dyck PJ, Thomas PK, Lambert
Stimulus Latency of Activation
EH, Bunge R (eds): Pripheral Neuropathy (2 ed). Philadelphia, Saunders, 1984, pp 981-1015. 8. Cumrnins KL, Perkel DH, Dorfrnan LJ: Nerve fiber conductiori-velocity distributions. I. Estimation based on the single fiber and compound action potentials. Electroencephalogr Clin Neurophysiol 19 79 ;46 :634 - 646. 9. Dawson GD: T h e relative excitability and conduction velocity of sensory and motor nerve fibres in man. J Physiol (Londj 1956;131:436-451. 10. Erlanger J , Blair EA: The configuration of axon and “sitnple” nerve action potentials. A m J Physiol 1933;106:565570. 11. Gasser HS, Erlanger J: T h e role played by the sizes of the constituent fibers of a nerve trunk in determining the form of its action potential wave. Am J Physiol 1927;80:522-547. 12. Gasser HS, Grundfest H: Axon diameters in relation to the spike dimensions and the conduction velocity in marnmalian A fibers. Am J Physiol 1939;127:393-414. 13. Gilderneister M: Die allgemeinen Gesetze des elektrischen Reizes. I. Die Nutzzeit und ihre Gesetze. Z Biol 1913;62:358-396. 14. Gilliatt RW, Melville ID, Velate AS, Willison RG: A study of normal nerve action potentials using an averaging technique (barrier grid storage tube). ] Neurol Neurosurg Psychiatry 1965;28:191- 200. 15. Horowitz SH, Krarup C: Conduction studies of the normal s u r d nerve. Muscle Nerue 1992;15;374-383. 16. Kimura J: Electrodiagnosis in Diseases of Nerue and Muscle: Princzples and Practice ( 2 ed). Philadelphia, FA Davis, 1989, pp 629- 630. 17. Krarup C, Loeb GE: Conduction studies in peripheral cat nerve using implanted electrodes: I. Methods and findings in controls. Muscle Nerve 1988;11:922-932. 18. McDonald WI: T h e effects of experimental demyelination on conduction in peripheral nerve: A histological and electrophysiological study. 11. Electrophysiological observations. Brain 1963;86:501-524. 19. Podivinski F: The effect of stimulus intensity on latency of nerve action potential in healthy subjects and in patients with peripheral nerve lesions. Neurology 1965;15: 10591062. 20. Schoonhoven R, Schellens RLLA, Stegeman DF, GabreelsFesten AAWM: Sensory potentials and sural nerve biopsy: a model evaluation. Muscle Nerve 1987;10:246-262. 21. Stegeman DF, d e Weerd JPC: Modelling compound action potentials of peripheral nerves in situ. 1. Model description: evidence for a non-linear relation between fibre diameter and velocity. Ebctroencephalogr Clin Neurophysiol 1982;54:436-448. 22. Wiederholt WC: Threshold and conduction velocity in isolated mixed mammalian nerves. Neurology 1970;20:347352.
MUSCLE & NERVE
July 1992
821
