Fatigability and variable-frequency train stimulation of human skeletal muscles. (Research Report).

Surface electrical stimulation is often applied to human skeletal muscles that differ in fiber-type composition. For example, the tibialis anterior (TA) muscle is sometimes stimulated in an effort to control foot drop. (1) However, we have observed that the quadriceps femoris (QF) muscle is probably the muscle most often activated with electrical stimulation during rehabilitation, thereby serving as the test bed for protocol development. Whether protocols developed using a given muscle can be applied to another muscle has seldom been questioned, despite the fact that the characteristics of individual fibers contribute to a muscle's ability to develop and maintain force. (2) The observation that, on average, the TA muscle has 50% relatively more slow fibers than the QF muscle relative to each muscle's overall composition suggests that this "slow" human skeletal muscle would be more fatigue resistant than its "fast" counterpart, the QF muscle, due to the low energy cost of contraction of slow fibers. (3-5)

The relationship between the supply of and the demand for energy for contraction has a marked influence on one type of fatigue. (6) The ratio of succinate dehydrogenase (SDH) to actomyosin adenosine triphosphatase activity (qATPase) has been used as an estimate of this relationship. (3,6) Gregory et al (3) recently reported that average fiber SDH:qATPase ratio was similar for the TA and vastus lateralis (VL) muscles, suggesting that the TA muscle may not be more fatigue resistant than the QF muscle. The average fiber in the QF muscle has greater capacity for aerobic-oxidative energy supply to balance out the high energy cost of contraction of fast fibers. (3,4) Thus, the TA muscle, with 75% slow fibers and about 60% relative slow fiber cross-sectional area (CSA), may have the same capacity as the QF muscle (40% slow fibers and 35% relative slow fiber CSA) to maintain force during stimulation. (3,7-10)

The ability of skeletal muscle to maintain tension also can be influenced by factors independent of its inherent characteristics, such as fiber-type composition or metabolic enzyme profile. Nowhere do we believe is this more evident than during electrical stimulation. The asynchronous and orderly (slow to fast) recruitment of motor units that occurs during voluntary activation is absent during electrical stimulation. (11) This lack of asynchrony and orderly recruitment contributes to the increased fatigability observed with electrical stimulation when compared with volitional effort. (11,12) Variable-frequency trains (VFTs) have been used to counter the increased fatigue associated with conventional electrical stimulation, which uses constant-frequency trains (CFTs). (13-15) The VFT takes advantage of the "catchlike" property of skeletal muscle, which is the tension enhancement realized when a brief interpulse interval (IPI) is added to the beginning of a relatively low frequency train. (16) Variable-frequency trains enhance the rate of rise in torque as well as peak torque, and thereby the torque-time integral, especially in fatigued muscle. (14,17)

Slow and fast fibers also differ in the frequency of stimulation required to evoke a tetanic contraction. (16-18) The VFT typically consists of 2 frequencies that are above and below the frequency required for tetanic contraction. Thus, the difference in tetanic frequencies of different muscles may translate into distinct ranges of frequencies for VFTs to effectively enhance the torquetime integral. The frequency required to obtain a tetanic contraction may be reduced due to the slowing of contraction that occurs with fatigue. (19) Thus, increasing the frequency by using VFTs may not produce tension enhancement because the CFT, which was initially subtetanic, may now be sufficient to cause tetanic contraction in the fatigued muscle. The simplest way to exhibit this concept is to assume that a "slow," fatigued muscle has a tetanic frequency of 15 Hz. If a train of pulses were applied to this muscle at 15 Hz, increasing the rate of the first 2 pulses to 50 Hz would not improve rise time or peak torque.

Few studies on animals have investigated the efficacy of VFT stimulation in slow muscle. Rabbit TA muscle (99% fast fibers) stimulated 24 hours per day for greater than 4 weeks at a low frequency to convert to slow muscle loses the advantage offered by the VFT. (20) However, Burke et al (17) showed that slow motor units of the cat demonstrate the "catchlike" property. To our knowledge, there have been no direct comparisons between human skeletal muscles that have large differences in fiber type concerning the effectiveness of VFTs to enhance the torque-time integrals as compared with CFTs. We could find only one published study of the "catchlike" property of the TA muscle, representative of a "slow" human skeletal muscle, in this case examining energy cost of contraction. (21)

In our study, we tested 2 hypotheses. The first hypothesis was that fatigue during CFT stimulation would not differ between the TA and QF muscles. The second hypothesis was that force augmentation by a VFT would not differ between these muscles when they are fatigued.

Method

Subjects

Ten recreationally active men with no history of lower-extremity pathology participated in the study (mean age = 25 years, SD = 4, range = 19-31; mean height = 179 cm, SD = 5, range = 170-188; mean body mass = 80 kg, SD = 15, range = 63-111). The left QF muscle and the right TA muscle of each subject were stimulated. All subjects gave written informed consent prior to testing.

QF Muscle Experimental Setup

The QF muscle was stimulated essentially as described previously. (11,22-25) Subjects were seated in a custom-built chair with the left hip and knee secured at approximately 90 degrees of flexion. The leg was secured to a rigid lever arm with an inelastic strap in an effort to ensure that the knee extensors could perform only isometric contractions. A Rice Lake 2000A load cell * was attached via a steel rod to the rigid lever arm on one end and an immovable support on the other. The line of pull on the load cell was perpendicular to the rigid lever arm. Thus, the torque produced about the knee was calculated as the product of the load cell force and the length of the lever arm between axis of rotation and the load cell attachment. Two 8- x 10-cm surface electrodes ([dagger]) were placed on the distal vastus medialis muscle and the proximal VL muscle. This electrode placement has been shown to allow recruitment of a large portion of the QF muscle fibers. (11,25)

TA Muscle Experimental Setup

Subjects lay supine with their leg resting in a custom-built apparatus that secured the hip at 20 degrees of flexion, the knee at 35 degrees of flexion, and the ankle at 15 degrees of plantar flexion. The foot was secured to a rigid lever arm with an inelastic strap in an effort to ensure that the ankle dorsiflexors could perform only isometric contractions. The load cell was attached by a steel rod to a rigid lever arm on one end and an immovable support on the other. The line of pull on the load cell was perpendicular to the rigid lever arm. Torque was Calculated as the product of the load cell force and the length of the lever arm between the axis of rotation and the load cell attachment. Two 5- x 5-cm surface electrodes ([dagger]) were placed on the proximal and distal portions of the TA muscle.

Electrical Stimulation and Force Recordings

A Digitimer DS7AH electrical stimulator ([double dagger]) was triggered by a personal computer with a Keithley KPCI 3108 A/D board ([section]) and a customized program written with TestPoint software (version 4.0). ([parallel]) The stimulator delivered six 200-microsecond square-wave pulses with either a CFT or a VFT. The CFT consisted of 6 pulses separated by 70-millisecond IPIs. The VFT had a 5-millisecond IPI between the first and second pulses followed by 4 additional pulses separated by 70-millisecond IPIs. The VFT, with only one brief IPI, was chosen because it has been reported to augment force in a fatigued muscle. (14) Torque from the load cell was sampled at 10 kHz by computer using the KPCI 3108 A/D board.

Experimental Procedure

Identical procedures were followed for the QF and TA muscles. Subjects were secured for testing of either the QF or TA muscle as described. Maximal voluntary isometric contractions (MVICs) were performed, and peak torque was recorded. The current necessary to elicit approximately 25% of MVIC was then determined, and the subjects rested for 3 to 5 minutes. The muscle was then potentiated with 6-pulse CFTs that were delivered once every 5 seconds until torque no longer increased. When the muscle was fully potentiated, a CFT and a VFT were delivered. The muscle was then fatigued using 180 six-pulse CFTs delivered at a 50% duty cycle. Immediately following the fatigue protocol, a CFT and a VFT were delivered in random order.

Torque-time integral, peak torque, and the time from 20% to 80% of peak torque (T20-T80) were determined from the torque recordings. A 3-way analysis of variance (type of stimulation train x muscle x time) for repeated measures over muscle and time was run on each variable using SPSS software (version 10.0). (#)

Results

Muscle Fatigue

Stimulation prior to fatigue resulted in greater absolute peak torque (in newton-meters) for the QF muscle than for the TA muscle, but torque relative to MVIC was not different between the 2 muscles (QF muscle mean torque = 27% of MVIC, SD = 3% of MVIC, range = 23%-33% of MVIC; TA muscle mean torque = 28% of MVIC, SD = 10% of MVIC, range = 20%-50% of MVIC) (Fig. 1). Relative torque decline during the 180 CFTs was not shown to be different between the 2 muscles (QF muscle mean torque decline = 57%, SD = 7%, range = 42%-65%; TA muscle mean torque decline = 55%, SD = 9%, range = 42%-66% (Fig. 2).

[FIGURES 1-2 OMITTED]

Torque-Time Integral

There was a 3-way interaction among type of stimulation train, muscle, and time for the torque-time integral (P = .05). A 2-way interaction between type of stimulation train and time (P[less than or equal to].001) was partly responsible for the 3-way interaction. The torque-time integral for the VFT and CFT prior to fatigue differed by 5% or less for either muscle (Table). After the fatigue protocol, however, the VFT enhanced the torque-time integral Of the QF and TA muscles by 23% and 29%, respectively (Table). There was also a 2-way interaction between muscle and time (P[less than or equal to].001) when the torque-time integral was expressed in absolute values. This interaction reflected the larger initial value and greater absolute decrease for the QF muscle compared with the TA muscle (Table). When data were expressed in a relative sense (ie, percentage of MVIC), force loss was not different between the 2 muscles. There was no 2-way interaction between type of stimulation train and muscle, indicating that augmentation of the torque-time integral by VFTs was not shown to be different between the 2 muscles.

Peak Torque

Greater peak torque could contribute to the augmented torque-time integral of fatigued muscle evoked by a VFT. There was a type of stimulation train x time interaction (P = .028) for peak torque, but there was no 3-way interaction (type of stimulation train x muscle x time) (Table). The VFT and CFT evoked comparable peak torques prior to fatigue (within 5%) (Table). However, VFTs evoked 25% and 29%, respectively, greater peak torque than CFTs in the fatigued QF and TA muscles (Table). There was no type of stimulation train x muscle interaction, indicating that the effect of the VFT was not different between muscles. As with the torque-time integral, there was a muscle x time interaction (P[less than or equal to].001) for peak torque. Again, this interaction reflected the greater absolute (initial and decrease) peak torque for the QF muscle than for the TA muscle, whereas the relative decrease was not different between muscles (Table, Figs. 1 and 2).

Time From 20% to 80% of Peak Torque

A more rapid rise time also could account for the greater torque-time integral that was evoked in fatigued muscle by the VFT. As for peak torque, there was a type of stimulation train x time interaction (P[less than or equal to].001) for T20-T80, but there was no 3-way interaction for type of stimulation train x muscle x time. In essence, the more rapid T20-T80 for the VFT was not prolonged as much by fatigue as it was for the CFT (Table). There were no muscle x time and type of stimulation train x muscle interactions for T20-T80, suggesting that the slowing of contraction and the effect of the VFT were not different between muscles (Table). There was a muscle effect (P[less than or equal to].001), indicating that the QF muscle was faster than the TA muscle (Table).

Discussion

We had 2 major findings that supported our hypotheses. First, relative force loss over 180 CFTs did not differ between the TA and QF muscles. Second, both muscles demonstrated augmentation of the torque-time integral by VFTs, due to greater peak torques and more rapid rates of contraction.

Based on fiber type alone, the TA muscle, in theory, should fatigue less than the QF muscle because its greater relative amount of slow fibers would be expected to have less energy demand during contraction. (3) Studies of single fibers from human muscle have shown that the ATP cost of generating force during isometric contraction is lower in slow fibers than in fast fibers. (5) The TA muscle, we believe, would have been expected to have greater capacity for aerobic-oxidative energy supply because slow human fibers have generally been reported to have higher mitochondrial content. (3,26) However, this may not necessarily be the case, as recently reported by Gregory et al. (3) Average fiber SDH activity was not different in the TA and VL muscles, with the latter muscle being representative of the QF muscle. The relationship between aerobic-oxidative energy supply (SDH) and energy demand of contraction, as reflected by actomyosin myofibrillar ATPase activity (qATPase), may be more indicative of fatigue resistance than mitochondrial content. (27) Because of these findings and because the average fiber SDH:qATPase ratio did not differ between the TA and QF muscles, Gregory et al hypothesized that fatigue would be comparable between muscles. Our results strongly support this hypothesis because the TA and QF muscles each showed about 55% force loss over 180 CFTs.

We believe that aerobic-oxidative energy supply would make a meaningful contribution of ATP during the 2 minutes of intermittent contractions imposed on the TA and QF muscles in our study. We believe the protocol we used would allow sufficient time and blood flow for cellular respiration to increase energy supply and thereby influence performance. Dudley et al (28) previously showed that the sensitivity of respiratory control is so robust that mitochondrial content influences metabolic responses during the first several seconds of intermittent contractions. However, we believe this would not be the case during continuous stimulation, which is sometimes used to examine muscular endurance, because such contractions compromise blood flow and thereby aerobic-oxidative energy supply. (29)

The muscles we studied were stimulated with a current sufficient to evoke about 25% of MVIC with 6-pulse trains of about 14 Hz. Potentiation increased torque a few percent, and torque was not different for VFTs and CFTs after potentiation. Nonetheless, it is highly unlikely that more than 50% of the mass of either muscle was being stimulated. (11,25) Thus, it might be predicted that mainly fast fibers were being activated in both muscles, based on their fiber-type composition and because it is often suggested that surface electrical stimulation reverses recruitment order, with the result being comparable rise times between the TA and QF muscles. (30-32) However, the results of studies that have utilized techniques such as muscle biopsy to investigate muscle fiber glycogen depletion after electrical stimulation suggest a random pattern of activation. (12,23)

Kim et al (12) have shown that surface electrical stimulation is rather random, with both fast and slow fibers being activated. The mechanical response depends on a muscle's architecture and relative proportion of fast and slow fibers. Our results, we believe, support the contention of a random pattern because rise time was about 40% longer in the potentiated, unfatigued "slow" TA muscle than in the "fast" QF muscle (Table). This is almost identical to the 50% more slow fibers in the TA muscle compared with the QF muscle. If random activation were the case, we also would predict that fatigue would not differ between the 2 muscles just as we found.

We attempted to determine whether augmentation of the torque-time integral by VFTs would differ between the TA and QF muscles. Fast muscle, with its need for a relatively higher frequency to produce tetany, theoretically is well designed to take advantage of the short IPI of VFTs performance is augmented. In contrast, the inherent contractile properties of slow muscle may (20) or may not (17) limit the influence of VFTs. For example, rise time and peak torque of a slow, fatigued muscle with an assumed tetanic frequency of 15 Hz would not increase if such a muscle were stimulated with a 50-Hz train. After potentiation, we found both comparable peak torque and torque-time integral between CFT and VFT activation in the TA muscle or the QF muscle even though the VFT was 65 milliseconds shorter. This occurred because the VFT evoked a more rapid rise time than the CFT. Rise time was markedly prolonged with fatigue for the CFT for both muscles, as expected. (34) There was little change for the VFT; thus, its more rapid rise time and greater peak torque resulted in a greater torque-time integral even though the VFT's duration was 20% shorter than that for the CFT. In addition, this effect was not different between the TA and QF muscles. The TA and QF muscles showed 29% and 23% more augmentation, respectively, in the torque-time integral with VFTs. Corresponding increases in peak torque were 29% and 25%. The 14.3-Hz stimulation from the CFT is probably closer to the fusion frequency of the TA muscle than that of the QF muscle, based on our rise time data, yet the VFT still caused an increase in the T20-T80 for the TA muscle. This finding suggests to us that the slow fibers in the TA muscle were probably not being activated at a frequency sufficient to evoke tetanus, as we believe was the case for the fast fibers that were stimulated. As a result, the brief IPI at the beginning of the train augmented rise time and peak torque.

The extent of fatigue and force augmentation by VFTs that we found is comparable to that reported previously for the QF muscle by Binder-MacLeod and colleagues. (14,15) For example, they reported that a VFT with one brief IPI can enhance force-time integrals in the fatigued QF muscle from 19% to 36%. (14,15) Russ and Binder-MacLeod (15) reported that the peak force of the QF muscle declined by about 44% and that there was subsequent force augmentation with VFTs for up to 37 minutes. We showed declines in peak torque of about 55%, and the VFT enhanced the torque-time integral of the fatigued QF muscle by about 23% when compared with a CFT; these percentages are well within previously reported values.

Conclusions

We believe that our results have clinical implications. The TA muscle represents two thirds of the dorsiflexors, which control foot drop and aid in toe clearance during gait. (35) Approximately 20% of people who have had a stroke have some degree of foot drop, (1) and people with other diagnoses often have foot drop, including those with incomplete spinal cord injury, multiple sclerosis, and acquired brain injury. Thus, there is a large group of patients undergoing rehabilitation who potentially could benefit from application of functional electrical stimulation to the dorsiflexors during ambulation. We showed torque augmentation with identical VFT stimulation variables in 2 skeletal muscles with different fiber types. Thus, protocols developed on the QF muscle also may be effective on the TA muscle, despite the large difference in fiber-type composition.

The results of our study indicate that fatigue during intermittent surface electrical stimulation was not different between the QF and TA muscles. Our results suggest that VFTs can augment force of fatigued human skeletal muscle irrespective of fiber type. Because the TA and QF muscles represent the "extremes" of fiber-type composition for most human skeletal muscles (see Gregory et al (3)) except under the most trying conditions (for a review, see Castro et al (23)), application of VFTs as done in our study may prove beneficial for assisting human skeletal muscles to perform functional movements with less fatigue.

Table.

Mechanical Responses of the Quadriceps Femoris (QF) and Tibialis
Anterior (TA) Muscles of 10 Subjects With No History of Lower-Extremity
Pathology to Variable-Frequency Train (VFT) and Constant-Frequency
Train (CFT) Surface Electrical Stimulation Immediately Before and After
180 CFTs (a)

                                             Before CFTs

Variable                  Muscle    Train    [bar]X     SD    Range

Torque-time integral      QF        CFT       20.1       4     15-27
  (b,c,d) (Nm*s)                    VFT       19.1       4     15-26
                          TA        CFT        4.9       2      2-9
                                    VFT        4.8       2      2-8
Peak torque (Nm) (c,d)    QF        CFT       54.3      11     41-73
                                    VFT       56.3      12     41-76
                          TA        CFT       14.2       2     12-16
                                    VFT       15.0       2     12-17
T20-T80 (ms) (d,e)        QF        CFT      132.0      16    109-167
                                    VFT       85.0      20     53-111
                          TA        CFT      179.0      19    154-200
                                    VFT      121.0      16     90-148

                                             After CFTs

Variable                  Muscle    Train    [bar]X     SD    Range

Torque-time integral      QF        CFT       11.0       4      8-19
  (b,c,d) (Nm*s)                    VFT       13.5       4     11-21
                          TA        CFT        2.4       1      1-5
                                    VFT        3.1       1      1-5
Peak torque (Nm) (c,d)    QF        CFT       28.3       9     20-46
                                    VFT       35.5      10     26-54
                          TA        CFT        6.6       2      4-12
                                    VFT        8.5       3      6-13
T20-T80 (ms) (d,e)        QF        CFT      187.0      18    156-210
                                    VFT       96.0      26     56-129
                          TA        CFT      212.0      22    182-249
                                    VFT      141.0      39    101-183

(a) CFT=six 200-microsecond square waves separated by 70 milliseconds,
VFT=first interpulse interval only 5 milliseconds, T20-T80=time from
20% to 80% of peak torque.

(b) Three-way interaction, P=.05.

(c) Two-way interaction, less pretest-posttest change for VFT than for
CFT, P<.029.

(d) Two-way interaction, QF muscle showed greater decline than TA
muscle due to greater absolute pretest value,
P[less than or equal to] .001 (relative decrease in QF muscle=relative
decrease in TA muscle).

(e) QF muscle faster than TA muscle, P[less than or equal to] .001.

* Rice Lake Weighing Systems, 230 W Coleman St, Rice Lake, WI 54868.

([dagger]) Uni-Patch, PO Box 1271, 1313 W Grant Blvd, Wabasha, MN 55981.

([double dagger]) Digitimer Ltd, 37 Hydeway, Welwyn Garden City, Hertfordshire, AL7 3BE, England.

([section]) Keithley Instruments, 28775 Aurora Rd, Cleveland, OH 44139.

([parallel]) Capital Equipment Carp, 900 Middlesex Turnpike, Billerica, MA 01821.

(#) SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

References

(1) Taylor PN, Burridge JH, Dunkerley AL, et al. Clinical use of the Odstock dropped foot stimulator: its effect on the speed and effort of walking. Arch Phys Med Rehabil. 1999;80:1577-1583.

(2) Fitts RH. Cellular mechanisms of muscle fatigue. Physiol Rev. 1994; 74:49-94.

(3) Gregory CM, Vandenborne K, Dudley GA. Metabolic enzymes and phenotypic expression among human locomotor muscles. Muscle Nerve. 2001;24:387-393.

(4) Crow MT, Kushmerick MJ. Chemical energetics of slow- and fast-twitch muscles of the mouse. J Gen Physiol. 1982;79:147-166.

(5) Han YS, Proctor DN, Geiger PC, Sieck GC. Reserve capacity for ATP consumption during isometric contraction in human skeletal muscle fibers. J Appl Physiol. 2001;90:657-664.

(6) van der Laarse WJ, Diegenbach PC, Maslam S. Quantitative histochemistry of three mouse hind-limb muscles: the relationship between calcium-stimulated myofibrillar ATPase and succinate dehydrogenase activities. Histochem J. 1984;16:529-541.

(7) Johnson MA, Polgar J, Weightman D, Appleton D. Data on the distribution of fibre types in thirty-six human muscles: an autopsy study. J Neurol Sci. 1973;18:111-129.

(8) Hather BM, Tesch PA, Buchanan P, Dudley GA. Influence of eccentric actions on skeletal muscle adaptations to resistance training. Acta Physiol Scand, 1991;143:177-185.

(9) Kent-Braun JA, Ng AV, Castro M, et al. Strength, skeletal muscle composition, and enzyme activity in multiple sclerosis. J Appl Physiol. 1997;83:1998-2004.

(10) Staron RS, Hagerman FC, Hikida RS, et al. Fiber type composition of the vastus lateralis muscle of young men and women. J Histochem Cytochem. 2000;48:623-629.

(11) Adams GR, Harris RT, Woodard D, Dudley GA. Mapping of electrical muscle stimulation using MRI. J Appl Physiol. 1993;74: 532-537.

(12) Kim CK, Bangsbo J, Strange S, et al. Metabolic response and muscle glycogen depletion pattern during prolonged electrically induced dynamic exercise in man. Scand J Rehabil Med. 1995;27:51-58.

(13) Binder-Macleod SA, Barker CB III. Use of a catchlike property of human skeletal muscle to reduce fatigue. Muscle Neroe. 1991;14: 850-857.

(14) Binder-Macleod SA, Lee SC, Baadte SA. Reduction of the fatigue-induced force decline in human skeletal muscle by optimized stimulation trains. Arch Phys Med Rehabil. 1997;78:1129-1137.

(15) Russ DW, Binder-Macleod SA. Variable-frequency trains offset low-frequency fatigue in human skeletal muscle. Muscle Nerve. 1999;22: 874-882.

(16) Burke RE, Rudomin P, Zajac RE. The effect of activation history on tension production by individual muscle units. Brain Res. 1976;109: 515-529.

(17) Burke RE, Rudomin P, Zajac FE. Catch property in single mammalian motor units. Science. 1970;168:122-124.

(18) Bellemare F, Woods JJ, Johansson R, Bigland-Ritchie B. Motor-unit discharge rates in maximal voluntary contractions of three human muscles. J Neurophysiol. 1983;50:1380-1392.

(19) Bigland-Ritchie B, Johansson R, Lippold OC, et al. Changes in motoneurone firing rates during sustained maximal voluntary contractions. J Physiol. 1983;340:335-346.

(20) Kwende MM, Jarvis JC, Salmons S. The input-output relations of skeletal muscle. Proc R Soc Lond B Biol Sci. 1995;261(1361):193-201.

(21) Ratkevicius A, Quistorff B. Metabolic costs of force generation for constant-frequency and catchlike-inducing electrical stimulation in human tibialis anterior muscle. Muscle Nerve. 2002;25:419-426.

(22) Castro MJ, Apple DF Jr, Rogers S, Dudley GA. Influence of complete spinal cord injury on skeletal muscle mechanics within the first 6 months of injury. Eur J Appl Physiol. 2000;81:128-131.

(23) Castro MJ, Apple DF Jr, Staron RS, et al. Influence of complete spinal cord injury on skeletal muscle within 6 months of injury. J Appl Physiol. 1999;86:350-358.

(24) Harris RT, Dudley GA. Factors limiting force during slow, shortening actions of the quadriceps femoris muscle group in vivo. Acta Physiol Scand. 1994;152:63-71.

(25) Hillegass EA, Dudley GA. Surface electrical stimulation of skeletal muscle after spinal cord injury. Spinal Cord. 1999;37:251-257.

(26) Kraus WE, Torgan CE, Taylor DA. Skeletal muscle adaptation to chronic low-frequency motor nerve stimulation. Exerc Sport Sci Rev. 1994;22:313-360.

(27) van der Laarse WJ, Lannergren J, Diegenbach PC. Resistance to fatigue of single muscle fibres from Xenopus related to succinate dehydrogenase and myofibrillar ATPase activities. Exp Physiol. 1991;76: 589-596.

(28) Dudley GA, Tullson PC, Terjung RL. Influence of mitochondrial content on the sensitivity of respiratory control. J Biol Chem. 1987;262: 9109-9114.

(29) Gaviria M, Ohanna F. Variability of the fatigue response of paralyzed skeletal muscle in relation to the time after spinal cord injury: mechanical and electrophysiological characteristics. Eur J Appl Physiol Occup Physiol. 1999;80:145-153.

(30) Enoka RM. Muscle strength and its development: new perspectives. Sports Med. 1988;6:146-168.

(31) Sinacore DR, Delitto A, King DS, Rose SJ. Type II fiber activation with electrical stimulation: a preliminary report. Phys Ther. 1990;70: 416-422.

(32) Trimble MH, Enoka RM. Mechanisms underlying the training effects associated with neuromuscular electrical stimulation. Phys Ther. 1991;71:273-280; discussion 280-272.

(33) Knaflitz M, Merletti R, De Luca CJ. Inference of motor unit recruitment order in voluntary and electrically elicited contractions. J Appl Physiol. 1990;68:1657-1667.

(34) Bigland-Ritchie B, Johansson R, Lippold OC, Woods JJ. Contractile speed and EMG changes during fatigue of sustained maximal voluntary contractions. J Neurophysiol. 1983;50:313-324.

(35) Fukunaga T, Roy RR, Shellock FG, et al. Specific tension of human plantar flexors and dorsiflexors. J Appl Physiol. 1996;80:158-165.

CS Bickel, PT, PhD, is Post-doctoral Research Fellow, Department of Exercise Science, The University of Georgia, 300 River Rd, Athens, GA 30602 (USA) ([e-mail],edu). Address all correspondence to Dr Bickel.

JM Slade, MA, is a doctoral student in the Department of Exercise Science, The University of Georgia.

GL Warren, PhD, is Associate Professor, Department of Physical Therapy, Georgia State University, Atlanta, Ga.

GA Dudley, PhD, is Distinguished Research Professor and Professor of Exercise Science, The University of Georgia, and Research Scientist, Shepherd Center, Atlanta, Ga.

All authors provided concept/idea/research design and writing. Dr Bickel, Ms Slade, and Dr Warren provided data collection. Dr Bickel and Dr Dudley provided data analysis, project management, and fund procurement. Dr Dudley provided facilities/equipment and institutional liaisons.

All methods were approved by the institutional review boards at The University of Georgia and Shepherd Center.

This research was funded by the Shepherd Center (JMS, CSB, and GAD), the Foundation for Physical Therapy (CSB), and the National Institutes of Health (grants HD37439-S1 and HD39676 to GAD).

This article was submitted June 7, 2002, and was accepted November 22, 2002.