Using a new collision method, we measured motor nerve conduction velocities of the ulnar nerve in the forearm and the action potential amplitude of the abductor digiti minimi muscle on 60 adults, ages 20 to 82 years and apparently free from diseases of the peripheral nervous system. Both maximal and minimal motor nerve conduction velocities were linear functions of age; 64.42 - 0.05 age and 60.45 - 0.12 age, respectively. The percentage of the minimal to the maximal motor nerve conduction velocities was expressed as 94.45 - 0.13 age. The maximum amplitude of evoked muscle action potentials was also correlated with age. This novel method may be useful in detecting pathology of motor nerve fibers which results in a decrease in submaximal conduction velocities. Key words: collision method motor conduction velocity ulnar nerve abductor digiti minimi muscle MUSCLE & NERVE 14:647453 1991
NORMAL MAXIMAL AND MINIMAL MOTOR NERVE CONDUCTION VELOCITIES IN ADULTS DETERMINED BY A COLLISION METHOD KEISUKE ARASAKI, MD, PhD, MASAKAZU IIJIMA, MD, and TAKA0 NAKANISHI, MD, PhD
T h e conventional nerve conduction study measures only the maximal conduction velocity, which represents a minority of the fastest-conducting fibers in the human peripheral nerve. Therefore, it is unable to detect pathology of the majority of nerve fibers with submaximal conduction velocities. In order to compensate for this deficiency, two methods have been invented: one employs computer analysis of the compound nerve action p ~ t e n t i a l and , ~ the other is a collision method using paired stimuli.72s912 There are fundamentally two kinds of collision methods for measuring submaximal motor nerve
From the Neurophysiology Laboratory, Department of Neurology, University of Tsukuba, Tsukuba-City, Japan. Acknowledgment: The authors thank Drs. Ingrid L Kwee (University of California, Davis) and Forbes H. Norris, Jr (Pacific Medical Center, San Francisco) for their critical comments and assistance in improving English syntax. The authors are also grateful to Ms. Hiroko Akita for her ardent technical assistance. This study was supported by research grants from the Research Committee for Specific Diseases, the Ministry of Health and Welfare, and the Ministry of Education Japanese Government. Address reprint requests to Dr Arasaki. Department of Neurology, Institute of Clinical Medicine, University of Tsukuba 1-1-1 Tennodai, TsukubaCity, Ibaraki-Pref. 305, Japan Accepted for publication July 11, 1990. CCC 0148-639X/91/070647-07 $04.00 0 1991 John Wiley & Sons, Inc.
Motor Conduction Velocity
conduction velocities. In one kind, a pair of proximal and distal supramaximal stimuli with various inter-stimulus intervals (ISIs) is used.738Because of the refractory period of nerve fibers repeatedly stimulated by these supramaximal shocks? and distortion of the test muscle response,' conduction velocity calculated by Hopf s technique7 is subject to error. Although Ingram's method* takes into consideration the refractory period and the muscle response distortion, in practice, it is difficult to decide the minimal IS1 for obtaining the maximum motor nerve conduction velocity (maxMCV). In the another method, distal submaximal and proximal supramaximal stimuli are used. This method was first described in an animal study by Gilliatt et a16 and has recently been reevaluated and applied to human studies by Nakanishi et al.11312The maximal and minimal motor nerve conduction velocities (minMCV) thus measured are not influenced by the refractory period, and no tedious correction process is required.' In our previous study'* on maxMCV and minMCV of patients with amyotrophic lateral sclerosis using the latter collision method, we used agematched control subjects. The purpose of the present study was to obtain normative data on maxMCV and minMCV of the ulnar nerve in subjects from all adult age groups with no peripheral nervous system disease.
MUSCLE & NERVE
July 1991
647
MATERIALS AND METHODS
We examined 60 adults, aged 20 to 82 years (mean 47.1 years), who had no clinical evidence of peripheral nervous system disease, such as compressive neuropathy of the ulnar nerve. Subjects consisted of 32 healthy volunteers (mean age 43.8), who have given informed consent, and 28 patients with Parkinson’s disease (mean age 56.1). The ulnar nerve of each subject was tested by the collision method detailed elsewhere. l 1 For stimulation, we clamped the right hand and arm onto a rigid board to prevent movement. Cathode electrodes were then placed at 3 points over the ulnar nerve: at about 10 mm (Sl) and 70 mm (S2) above the distal wrist crease and at the elbow (S3). At each point of stimulation, an anode electrode was placed lateral to the cathode. The duration of the stimulating electric pulse was 0.1 ms, and the intensity was independently adjusted so that supramaximal shocks were given to 2 proximal points (S2 and S3) and submaximal ones to the distal point (Sl). This submaximal stimulation at S 1 preceded supramaximal stimulation at S2 or S3 by 0.5 ms in order to obtain a clear separation between the 2 stimuli. One series of the shocks consisted of a single stimulation at S1; another series was paired stimuli at S1 and S2, and another second paired shocks at S1 and S3. We repeated 10 to 20 series with different stimulus intensities at S 1. For recording motor responses, 2 chloridecoated silver disk electrodes on the skin overlay the belly of the abductor digiti minimi and the proximal phalanx of the fifth finger. These two recording electrodes were connected to a differential amplifier (WPI DAM-60) with a frequency response of 1.0 Hz and 3 kHz ( 3 dB down), and the output was then led through an A-D converter to a computer (Macintosh IIcx) with a horizontal resolution of 25 pdpoint with 256 points per channel. T o obtain better signal-to-noise ratio, 10 to 20 consecutive responses were averaged. All the muscle evoked action potentials were stored on floppy disks. Off-line computer analysis was performed by commercially available softwares (CricketGraph and StatView SE+Graphics). A submaximal motor response to a single stimulation at S1 (MSli’) was digitally subtracted from a mixed motor response following paired stimulation at S2 and S1 or S3 and S1 ([MS2 f MSli]’ o r [MS3 + MSli]’). The resulting [MS2 + MSli]’ - MSli’ or [MS3 MSli]’ - MSli’ should show muscle activity by
+
648
Motor Conduction Velocity
orthodromic motor impulses evoked at S2 or S3 and free from collision (MS2’ or MS3’). The onset latency of muscle activity was digitally measured: the wave form of each motor response was transformed to data sets composed of time and the corresponding amplitude. The first data set which had its amplitude above the range of baseline noise was defined as the onset of the motor response. Conduction velocities of motor nerve fibers giving rise to MS2’ and MS3’ were calculated by dividing the distance between S2 and S3 by the difference between the onset latencies of MS2’ and MS3’. Skin temperature between S2 and S3 was measured with a surface thermister, and conduction velocities were corrected for temperature according to the following equation: corrected conduction velocity = observed conduction velocity + Q (34.0 - observed skin temperature). We used 2.4 m/s per degree C as the Q-value for the fastestconducting motor fibers,’ whereas the Q-value for slow-conducting motor fibers was estimated to be 2.2 m/s per degree C in this study (see discussion). Results were printed out by a laser printer (LaserWriter I1 NTX). For a better statistical description of the relationship between such variables as the nerve conduction velocity and age, both simple and quadratic regressions were tested. The first null hypothesis was that age had no effect on nerve conduction velocities. We defined the Fa ratio as follows:
Fa
SSJDF SS,IDF
= ___
at degrees of freedom (OF) 1 (n-2, linear regression) or 2 (n-3, quadratic regression); SS,. was the variation explained by a particular regression equation, and SS, the variation unexplained by the equation. The calculated ratio and its probability were used to obtain the analysis of variance (ANOVA) table. T o test the second null hypothesis that a quadratic regression did not fit the observed data better than the linear one, we performed the F test using the Fr ratio defined as below:
where SS,(Q, was the sample variation explained by the quadratic regression, SS,(L) the variation explained by the linear regression, and SSe(Q, the variation not explained by the quadratic regres-
MUSCLE & NERVE
July 1991
sion. The degrees of freedom of the Fr ratio was 1 and n-3. Amplitudes of the maximal motor responses to the distal stimulation (S 1) were measured. Although the amplitude of a single response revealed by a collision was extremely variable, the peak-to-peak amplitude of averaged responses to the most proximal stimulation (S3) could be measured digitally. RESULTS
With small increments of stimulation intensities at S1 (5% to 10% of the dif-
MaxMCV and minMCV.
-2.0
4 0
B
-,
ference between the threshold and the maximal intensities), the early parts of MS2' and MS3' showed decreasing amplitudes and step-wise prolongation of their onset latencies. Thus, 2 major components of motor nerve conduction velocities were disclosed (Figs. 1A and 1B). The fastest conduction velocity of the fast component was equal to the conventional maxMCV (Fig. IA). It ranged from 55.4 to 66.8 m/s (mean, 61.9 m/s). For healthy volunteers and Parkinsonian patients, the means were 62.4 and 61.5 m/s, respectively, which was not statistically significant ( P > 0.05). The maxMCV could be expressed as a
10
2o
Timeinms
10
2'
Timeinms
s1=1M) v
2.0
E
-2.0 0
FIGURE 1. (A) Muscle responses to 52 (thin line) and 53 (thick line) shocks with 51 of 0 V. Closed arrows show the onset of the fastest component used to calculate the maxMCV. (B) Response to 52 (thin line) and 53 (thick line) stimulation after adding 51 of 100 V. Open arrows designate the onset of the slower component used to obtain the minMCV.
Motor Conduction Velocity
MUSCLE & NERVE
July 1991
649
6515 5.
0
A v)
0
\
6
v
5 E
5 0. 4 54 0-
3 5. 3 04
10
20
30
40
50
60
70
I
80
90
age FIGURE 2. Scattergram of the maxMCV (open circles) and minMCV (closed circles) (rn/s) against age in years. Solid lines designate linear regressions for both MCVs. Broken lines indicate 99% confidence bands for absolute motor conduction velocities.
linear function of age: 64.42 - 0.05 age (Fig. 2, open circles) with an Fa ratio 7.28 ( P < 0.005) in the ANOVA table. A t-value €or the slope coe€ficient was 2.70 ( P < 0.01). On the other hand, the best-estimated quadratic regression equation of 61.25 - 0.1 1 age - 0.002 age‘ had t-values of 1.13 and 1.67 ( P > 0.1). The Fr ratio for evaluating better adequacy of these 2 regression equations was 2.80 (P > 0.05). Therefore, this quadratic regression did not fit the observed data better than the linear one. We designated the fastest conduction velocity of the slow component (Fig. 1B) as the minMCV. In the 60 subjects studied, it varied between 45.9 and 63.7 mls (mean, 55.0 mls). When analyzed separately, the mean minMCVs in healthy subjects and Parkinsonian patients were 55.7 and 54.3 mls, respectively, which were not statistically different ( P > 0.05). The minMCV could also be described as a linear function of age: 60.45 - 0.12 age (Fig. 2, closed circles) with an Fu ratio of 17.68 ( P < 0.001) in ANOVA. The t-value of this slope coef-
650
Motor Conduction Velocity
ficient was 4.20 ( P < 0.001). The quadratic regression that best-fit the observed data was 57.70 0.03 age - 0.002 age’, the coefficient t-values of which were 0.19 and 0.99, respectively ( P > 0.1). The Fr ratio was 1.25 ( P > 0.1), which led to the conclusion that this quadratic equation was apparently not better than the linear one. The range of conduction velocity differences between maxMCV and minMCV in each individual, both corrected for temperature, was between 2.6 m/s and 12.0 m/s (mean, 6.9 m/s). Its frequency distribution is shown in Figure 3. The ratio of minMCV to maxMCV in each subject was also calculated and expressed in percentage. It ranged from 74% to 97% (mean, 89%). ‘This ratio had an inverse relationship to age (an Fu value with P < 0.005 in ANOVA) and was expressed as a linear regression, 94.45 - 0.13 age (Fig. 4). This suggested that age-related slowing of motor nerve conduction velocities was more
Comparison Between maxMCV and minMCV.
MUSCLE & NERVE
July 1991
*
2
3
5
4
6
7
8
25.6 mV (11.8 5.5 mV). As in previous rep o r t ~ , ~there ” ~ was a negative correlation between age and muscle action potential amplitudes. The linear equation, 22,23 - 0.14 age, had an Fa ratio of 32.85 (P < 0.001), whereas the Fr ratio in comparison with the quadratic regression, 25.69 0.32 age + 0.002 age2, was 2.18 ( P > 0.05). Therefore, the quadratic regression was not more appropriate than the linear one. Peak-to-peak amplitudes of the subtracted motor responses to S3 ([MS3 + MSli]’ - MSli’) were also measured. They ranged between 0.3 mV and 3.8 mV (mean, 1.8 mV). T h e percentage cancellation of the maximal motor response amplitude, at which minMCV was obtained, was also calculated. It varied from 85% to 98% (mean 89%).
1 0 1 1 1 2
9
maxMCV - minMCV (m/s) FIGURE 3. Frequency histogram showing differences between maxMCV and minMCV. “Count” designates the number of subjects having a given range of each difference.
prominent in motor fibers having slower conduction velocities. Amplitude of Muscle Action Potentials. Peak-topeak amplitudes of the muscle action potentials to the supramaximal stimuli at S1 varied from 2.6 to
1001
----___ A
,s
Y
OI
7 0. 6 5.
6 01 10
20
30
40
50
60
70
80
90
age
FIGURE 4. Scattergram of the ratio of the rninMCV to the maxMCV against age. A solid line shows the linear regression line. Broken lines are 99% confidence bands for the ratio.
Motor Conduction Velocity
DISCUSSION
We measured skin temperature in order to standardize conduction velocities of motor nerve fibers for temperature. Skin temperature may well be lower than true tissue temperature around the ulnar nerve. Therefore, it should be explicitly stated that motor conduction velocities reported in the present study are those standardized to skin temperature of 34.0”C between S2 and S3 and not to true tissue temperature. We adopted Q-values of 2.4 and 2.2 m/s per “C for the maxMCV and minMCV, respectively. Many previous studies have shown the linear relationship between skin temperature and ~ ~ x M C V ,al~” though data supporting the nonlinear relationship have been published re~ent1y.I~ It is generally agreed that maxMCV changes by 2.4 m/s per degree C.’ However, it is not known whether the conduction velocity of the slower-conducting motor nerve fibers also changes proportional to the rate of skin temperature changes. Fast and slow motor fibers may have the same temperature coefficient (QlO), defined as the ratio of the conduction velocity at a given temperature to the velocity at a 10°C-lower t e m p e r a t ~ r e The . ~ minMCV was about 90% of the maxMCV in the present study. This is why a Q-value for the minMCV was estimated as 90% of that for the maxMCV, ie, 2.2 m/s per degree C. In the present study, minMCV was obtained at about 89% cancellation of the maximal motor response, and the difference between maxMCV and minMCV in each individual ranged from 2.6 mls to 12.0 m/s with the mean 6.9 m/s (Fig. 3). Hopf7 reported that differences between the fastest and slowest motor conduction velocities in humans
MUSCLE & NERVE
July 1991
651
were between 4 and 7 m/s. However, in a recent human study' with a modified Hopf m e t h ~ d , ~ conduction velocities were analyzed in the 5% to 95% cancellation range of the maximal motor response area; these were reported to be more variable, between 3.6 and 12.8 m/s. Considering that we used a collision technique different from that used by these authors,72sthese reported velocity differences are remarkably similar to our values. Both maxMCV and minMCV were inversely related to age of the subjects in the present study (Fig. 2). Most of the previous studies on maxMCV showed a linear relationship with age, 1 3 , 1 9 although a recent extensive study demonstrated that quadratic regression may be more adequate. Our results are consistent with a linear relationship between age and conduction velocity. Furthermore, we found that the slope coefficient of minMCV was larger than that of maxMCV, which suggested that the rate of decrement in the motor conduction velocity in aging might be more pronounced in slower-conducting nerve fibers (Fig. 4). The pathophysiology leading to reduction in motor nerve conduction velocities with ageing is probably multifactorial. The steady change in the conduction velocity after the age of 20 may be due to a gradual alteration in the properties of the axonal membrane as a part of the systemic ageing process. After 60 years of age, however, shortening of the internodal length'^''^'' and focal demyelination of nerve fibers' may also contribute. Our findings suggested that slower-conducting motor nerve fibers are more vulnerable to the ageing process. This may seem incompatible with the morphological findings associated with human ageing in the sural nerve,'4918a purely sensory nerve in man, consisting of a decrease in large myelinated nerve fibers, which may well result in a decrease in the maximal conduction velocity. However, this apparent contradiction can be explained in two ways: first, we examined motor rather than sensory nerve fibers; second, the anatomic observation that shortening of the internodal length associated with ageing is more prominent in medium and smaller myelinated nerve fibers, l o which are generally slower conducting.*' The physiological properties of motor units
having minMCV have not been elucidated in the present study. However, we found that minMCV was obtained at about 89% cancellation of the maximal motor response, which may be of some significance in this context. It was previously reported, using a collision method, that the area of muscle action potentials decreased rapidly as their latencies prolonged up to about the 86% to 89% cancellation point, above which the rate of decrease slowed. The authors suggested that this phenomenon could be explained by ( 1 ) skewed distribution of nerve fiber diameters, (2) larger nerve fibers contributing to a greater extent to the area, or (3) muscle synchronization.16 The modern concept of functional subtypes of motor units allows us to state that the second suggestion is correct and, in addition, the change in slope at 86% to 89% cancellation is probably due to the fact that there are two kinds of motor units, large and small. Therefore, we suggest that minMCV may correspond to the conduction velocity of the motor axons belonging to relatively small motor units. The range of human axonal motor conduction velocities in conjunction with the physiological properties of single motor units was reported only recently by Dengler et al.4 The maximal conduction velocity in their study was 62 m/s so that the mean minMCV of 55.0 m/s, ie, 82% of the maximal conduction velocity observed in the present study, may well be equivalent to about 50 m/s in their study.4 Furthermore, Dengler et a14 showed that motor units with their axonal conduction velocity below 50 m/s had much smaller twitch force and lower recruitment threshold in a single subject. Therefore, minMCV may represent motor units with smaller twitch force and lower recruitment t h r e ~ h o l d . ~ In summary, minMCV may be equivalent to the nerve conduction velocity of the relatively small motor units, which have small twitch forces and low recruitment thresholds. However, in order to correlate minMCV precisely to a certain type of motor unit, the other physiological properties of single motor units, such as twitch contraction time, fatigue index, and tetanus tension, need to be studied with concomitant measurement of minMCV in the same subject.
REFERENCES 1. Arnold N , Harriman DGF: The incidence of abnormality in control human peripheral nerves studied by single axon dissection. J Neurol Neurosurg Psychiaty 1970; 33:55-61.
652
Motor Conduction Velocity
2. Borg J: Refractory period of single motor nerve fibres in man. J Neurol Neurosurg Psychaaty 1984; 47:344-348. 3. de Jesus, J r PV, Hausrnanowa-Petrusewicz I, Barchi RL:
MUSCLE & NERVE
July 1991
4. 5. 6. 7. 8.
9. 10. 11.
The effect of cold on nerve conduction of human slow and fast nerve fibers. Neurology 1973; 23:1182- 1189. Dengler R, Stein RB, Thomas CK: Axonal conduction velocity and forces of single human motor units. Mwcle Nerve 1988; 11:136-145. Dorfman LJ: The distribution of conduction velocities (DCV) in peripheral nerves: A review. Mwcle Nerve 1984; 7:2- 11. Gilliatt RW, Hopf HC, Rudge P, Baraitser: Axonal velocities of motor units in the hand and foot muscles of the baboon. J Neural Neurosurg Psychiatry 1976; 29:249- 258. Hopf HC: Untersuchungen ueber die Untershiede in der Leitgeschwindigkeit motorischer Nervenfasern beim Menschen. Dtsch 2 Nervenheilkd 1962; 183:579-588. Ingram DA, Davis GR, Swash M: Motor nerve conduction velocity distribution in man: results of a new computerbased collision technique. Electroenceph Clin Neurophysiol 1987; 66~235-243. Kimura J: Principles and pitfalls of nerve conduction studies. Ann Neural 1984; 16:415-429. Lascelles RG, Thomas PK: Changes due to age in internoda1 length in the sural nerve in man. J Neural Neurosurg Psychiatry 1966; 29:40-44. Nakanishi T, Tamaki M, Mizusawa H, Akatsuka T, Kinoshita T : An experimental study for analyzing nerve conduction velocity. Electroenceph Clin Neurophysiol 1986; 631484-487.
Motor Conduction Velocity
12. Nakanishi T, Tamaki M, Arasaki K: Maximal and minimal motor nerve conduction velocities in amyotrophic lateral sclerosis. Neurology 1989; 39:580-583. 13. Norris AH, Shock NW, Wagman IH: Age changes in the maximum conduction velocity of motor fibers of human ulnar nerves. J Appl Physiol 1953; 5:589-593. 14. Ochoa J, Mair WGP: The normal sural nerve in man. 11. Changes in the axons and Schwann cells due to ageing. Acta Neuropathol (Berl) 1969; 13:217-239. 15. Taylor PK: Non-linear effects of age on nerve conduction in adults. J Neurol Neurosurg Psychiatry 1984; 66:223-234. 16. Thomas PK, Sears TA, Gilliatt RW: The range of conduction velocity in normal motor nerve fibers to the small muscles of the hand and foot. J Neural Neurosurg Psychiatry 1959; 22:175-181. 17. Todnem K, Knudsen G, Riise T , Nyland H, Aarli JA: The non-linear relationship between nerve conduction velocity and skin temperature. J Neural Neurosurg Psychiatry 1989; 521497-50 1 . 18. Tohgi H, Tsukagoshi H, Toyokura Y: Quantitative changes with age in normal sural nerves. Acta Neuropathol (Berl) 1977; 38:213-220. 19. Wagman IH, Lesse H: Maximum conduction velocities of motor fibers of ulnar nerve in human subjects of various ages and sizes.J Neurophysiol 1952; 15:235-244. 20. Waxman SG: Determinants of conduction velocity in myelinated nerve fibers. Muscle Nerue 1980; 3:141- 150.
MUSCLE & NERVE
July 1991
653