Reversal of muscle fatigue during 16 h of heavy intermittent cycle exercise

J Appl Physiol 97: 2166 2175, 2004; doi:10.1152/japplphysiol.00565.2004. Reversal of muscle fatigue during 16 h of heavy intermittent cycle exercise H. J. Green, T. A. Duhamel, S. Ferth, G. P. Holloway, M. M. Thomas, A. R. Tupling, S. M. Rich, and J. E. Yau Department of Kinesiology, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1 Submitted 1 June 2004; accepted in final form 11 July 2004 Green, H. J., T. A. Duhamel, S. Ferth, G. P. Holloway, M. M. Thomas, A. R. Tupling, S. M. Rich, and J. E. Yau. Reversal of muscle fatigue during 16 h of heavy, intermittent cycle exercise. J Appl Physiol 97: 2166 2175, 2004; doi:10.1152/japplphysiol. 00565.2004. This study examined the effects of extended sessions of heavy intermittent exercise on quadriceps muscle fatigue and weakness. Twelve untrained volunteers (10 men and 2 women), with a peak oxygen consumption of 44.3 2.3 ml kg 1 min 1, exercised at 91% peak oxygen consumption for 6 min once per hour for 16 h. Muscle isometric properties assessed before and after selected repetitions (R1, R2, R4, R7, R12, and R15) were used to quantitate fatigue (before vs. after repetitions) and weakness (before vs. before repetitions). Muscle fatigue at R1 was indicated by reductions (P 0.05) in peak twitch force (135 13 vs. 106 11 N) and by a reduction (P 0.05) in the force-frequency response, which ranged between 53% at 10 Hz (113 12 vs. 52.6 7.4 N) and 17% at 50 Hz (324 27 vs. 270 30 N). No recovery of force, regardless of stimulation frequency, was observed during the 54 min between R1 and R2. At R2 and for all subsequent repetitions, no reduction in force, regardless of stimulation frequency, was generally found after the exercise. The only exception was for R2, where, at 20 Hz, force was reduced (P 0.05) by 18%. At R15, force before repetitions for high frequencies (i.e., 100 Hz) returned to R1 (333 29 vs. 324 27 N), whereas force at low frequency (i.e., 10 Hz) was only partially (P 0.05) recovered (113 12 vs. 70 6.6 N). It is concluded that multiple sessions of heavy exercise can reverse the fatigue noted early and reduce or eliminate weakness depending on the frequency of stimulation. repetitive dynamic activity; isometric force; weakness IT IS WELL KNOWN THAT voluntary repetitive muscular activity ultimately results in a progressive inability to generate a required or desired amount of force necessary to perform a given task. This condition is commonly referred to as neuromuscular fatigue (7). It is also known that the failure in force-generating ability can persist for several hours and even days depending on the type of exercise (concentric vs. eccentric) and the specifics of the exercise protocol (intensity, duration, number of repetitions) (9). To differentiate between neuromuscular fatigue, which occurs during the exercise itself, and the persistent failure in force-generating potential observed after exercise, the term neuromuscular weakness has been used (7, 12). As might be expected, fatigue and weakness can display a variety of manifestations depending on the specifics of the task (6). The different mechanical manifestations suggest the differential involvement of selective sites in the motor command to the muscle and/or in the muscle itself (6). A major challenge in exercise physiology is to identify the specific sites responsible for the altered mechanical behavior observed during fatigue and weakness and the mechanisms responsible for the dysfunction observed in the particular sites involved. One of the most controversial issues is the relative roles of central command vs. the muscle cell itself in the disturbances observed with repetitive voluntary activity (19). There is evidence, particularly in heavy exercise, that inhibition in neural drive can be implicated in fatigue. It has also been proposed that the reduction in central command occurs as a consequence of the accumulation of selective metabolic by-products in the muscle that triggers central inhibition via reflexes generated from the muscle (20). It is also possible that, under some conditions, a failure in one or more excitation-contraction (E-C) process in muscle could be responsible for exercise-induced fatigue and weakness. High-intensity exercise has been shown to induce disturbances in the properties of the muscle compound action potential (M wave), which suggests that membrane excitability and the ability of the sarcolemma and T tubule to conduct action potentials may be compromised (25). There is also abundant evidence implicating a failure in sarcoplasmic reticulum (SR) Ca 2 handling as a site of fatigue (1). Repetitive exercise in known to depress Ca 2 uptake and Ca 2 release in humans (21). The alterations in these properties appear to disturb the cytosolic free Ca 2 ([Ca 2 f]) integral, resulting in a reduced activation of the myofibrillar complex (1). The alteration in specific E-C coupling processes could be responsible for different muscle mechanical responses under different activating conditions. Decreases in force during highfrequency stimulation, as an example, are most frequently attributed to the loss of membrane excitability (25). In contrast, the reduction in force at low frequency has been attributed to lower [Ca 2 f] levels (1). Interestingly, the persistence of this type of weakness after exercise suggests that SR impairment is not easily reversed (1). It has also been observed that, if the [Ca 2 f] is reduced sufficiently by exercise, depressions in high-frequency force can also be observed (1). Unclear from the studies performed to date are the characteristics of fatigue and weakness induced by repeated bouts of heavy exercise. A single bout of heavy exercise would be expected to induce fatigue, as indicated by a depression in both the maximal voluntary contraction (MVC) force and a shift in the force-frequency response to the left (3). If sufficient time is provided after the exercise to allow for full recovery of metabolites, weakness should persist (14) at least at low frequencies of stimulation. As the number of repetitions is increased, both fatigue and weakness should be exacerbated, resulting in a progressive fail- Address for reprint requests and other correspondence: H. J. Green, Dept. of Kinesiology, Univ. of Waterloo, Waterloo, Ontario, Canada N2L 3G1 (E-mail: green@healthy.uwaterloo.ca). The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 2166 8750-7587/04 $5.00 Copyright 2004 the American Physiological Society http://www.jap.org

ure in the mechanical response in response to both low and high frequencies of electrical stimulation (ES). The purpose of this study was to investigate the effects of repetitive intermittent exercise performed over an extended period of time on the mechanical manifestations of fatigue and weakness. We have hypothesized that, after exercise, fatigue would increase with the number of repetitions, resulting in progressive depressions in MVC force and in the force elicited at different frequencies of stimulation. We have also postulated that, during the recovery period, the low-frequency fatigue observed early in the repetitive exercise protocol would become more emphasized and ultimately result in weakness both during maximal voluntary effort and during different frequencies of stimulation. In this study, we have defined fatigue as a depression in mechanical responses observed after exercise and weakness as the persistent disturbance in the mechanical response observed at the end of the recovery period between exercise repetitions. METHODS Participants. Twelve university students (10 men and 2 women) volunteered for the study. As a condition of entry, all subjects were healthy (as determined by questionnaire) and not engaged in vigorous exercise on a regular basis. The mean age, height, and mass were 21.8 0.86 yr, 173.3 2.4 cm, and 70.2 3.6 kg, respectively. Peak aerobic power (V O2 peak), as determined by a progressive cycle exercise to fatigue, was 3.09 0.17 l/min. This study was approved by the Office of Research Ethics at the University of Waterloo, and all participants were fully informed of all experimental procedures and associated risks before written consent was obtained. Experimental design. Changes in fatigue and weakness in the quadriceps muscle were investigated in response to 16 repetitions of heavy cycling. Each repetition was performed for 6 min at 91% V O2 peak, as determined by the gas-exchange values determined during the final minute of exercise at the initial repetition. During the 54-min recovery period between repetitions, participants remained in the test area and were not permitted to exercise except as provided by the testing protocols. The basic testing protocol included the assessment of muscle mechanical behavior both before (B) and after exercise (A) at repetitions 1 (R1), 2 (R2), 4 (R4), 7 (R7), 12 (R12), and 15 (R15). Fatigue and weakness were determined by the disturbances in isometric force recorded both during MVC and ES at different frequencies. For this study, fatigue is defined as the loss of force occurring in response to each repetition (B vs. A), whereas weakness is defined as the persistent impairment in force observed between exercise repetitions (B vs. B). On the experimental day, each participant reported to the laboratory at 7 AM for preliminary preparation. The preliminary preparation involved insertion of catheters in the back of the hand for extraction of venous blood samples and preparation of sites on the vastus lateralis for harvesting muscle tissue (to be reported elsewhere). Before reporting to the laboratory, volunteers were required to ingest an Ensure (250 kcal) meal replacement consisting of 9.4 g of protein, 6.7 g of fat, and 38 g of carbohydrate (CHO) (Ross Products Division, Saint-Laurent, Canada). This was done to control for the preexercise nutritional intake between participants. Approximately 2 h elapsed between ingestion of the Ensure and the first exercise bout. During the 16-h experimental period, water and dietary supplements were not allowed during the first 2 h. Thereafter, water ingestion was ad libitum. After 2 h, the volunteers were allowed to consume selected vegetables, fruits, and Gatorade bars on a regulated basis. The macronutrient content of the foods ingested and the time of ingestion were recorded for each individual. Dietary composition as well as energy intake was calculated using nutritional analyses software REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY 2167 (ESHA-Diet Analysis Plus, version 5.0, Salem, OR). The analyses addressed different segments of the intermittent exercise repetitions as well as the combined 16-h characteristics. The environmental conditions remained constant throughout the experimental day. The temperature and relative humidity were 20 22 C and 39 48%, respectively. At least 2 wk before the intermittent exercise protocol, subjects performed progressive cycle exercise to fatigue for measurement of V O2 peak and related properties. The specifics of the protocol and the gas-collection system were as previously reported (23, 34). The oxygen consumption responses were plotted against power output and the relationship used to establish individual power outputs for the intermittent exercise sessions. The power output selected was designed to elicit 90% V O2 peak. The absolute power outputs selected remained constant for each of the work repetitions. For determinations of V O2 peak and repetitive exercise sessions, an electrically braked cycle ergometer (Siemens Elma 380 B) was used. The cycle was calibrated on a daily basis (V O2 peak tests) or at regular intervals (intermittent exercise). Seat height was individually adjusted and remained standardized for each repetition for each individual. All protocols were performed at a pedaling rate of 60 rpm. The intensity and duration of each of the intermittent work bouts were based on previous research from our laboratory (22, 36). We have found that, in untrained subjects, the protocol produces a pronounced strain on the metabolic systems. It was particularly important that subjects were able to perform complete multiple sessions of this protocol. Moreover, extensive adaptations also occurred in muscle metabolism and SR function when measured in the days after the intermittent exercise. Assessment of muscle function. Assessment of mechanical function was performed 10 15 min before the cycle exercise and 4 5 min after the exercise. The 4- to 5-min period, which was standardized for each volunteer, was needed to position and prepare for the measurements. The mechanical properties were based on isometric knee extension. For these measurements, the participant sat upright in a specially designed, straight-backed chair with hips and legs firmly secured, the knee at 90 to the thigh, and the arms folded across the chest. A 5-cm plastic cuff was positioned around the lower leg just proximal to the ankle malleoli and attached to a linear variable differential transducer. The linear variable transducer was amplified by a Daytronic carrier amplifier at 1 khz, converted to a digital signal, and fed into an IBM computer for analyses. Two aluminum chloride electrodes (8 13 cm) coated with warm electrode gel were used to deliver the electrical impulse to the quadriceps muscle. The ground electrode was placed centrally on the interior aspect of the thigh just above the patella while the active electrode was toward the hip on the proximal portion of the belly of the vastus lateralis. ES, applied to the right quadriceps muscle using a Grass model S48 stimulator with an isolation unit, was used to assess muscle mechanical behavior. The properties assessed were based on different frequencies of stimulation. To characterize the twitch properties, a single supramaximal ( 150 V) impulse with a 50- s duration was employed. The twitch properties assessed included the peak twitch force (P t), contraction time (CT), one-half relaxation time, the maximal rate of force development ( df/dt max), and the maximal rate of force decline ( df/dt max). Tetanic properties were based on stimulating the muscle at different frequencies (10, 20, 30, 50, and 100 Hz) using a pulse duration of 50 s and train durations of 1 s. The stimulation voltage employed for each participant was established during the days before the experimental day and was set to elicit 50% MVC at 100 Hz. The same absolute voltage was employed for each individual both B and A for each of the repetitions used for measurement. For the tetanic stimulations, peak force was measured as well as df/dt max and df/dt max for each frequency. A standardized sequence of measurements were performed both before and after selected repetitions of the exercise. The protocol consisted of two consecutive twitches (separated by 5 s) and stimulation at increasing frequencies, namely 10, 20, 30, 50, and 100

2168 REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY Hz. This was followed by two repetitions of MVC. A 30-s period separated each of the ES and MVC measurements. For the MVC, the subject was instructed to produce a maximal effort and to sustain the effort for 4 5 s. Verbal encouragement was provided. The time to administer the total protocol was 5 min. Additional details describing the measurement procedures and analyses of raw data (Labview 5.1 software routine) are as published earlier from our laboratory (18). We have previously determined the reliability of the forces produced by transcutaneous stimulation at different frequencies (26). Under the conditions of this experiment, reliabilities as calculated by the correlation coefficient (r), ranged between 0.81 and 0.96 for 10 and 100 Hz, respectively. We have also assessed membrane excitability at the time of the muscle function measurements using the properties of the muscle compound action potential (M wave). The properties assessed included the peak-to-peak amplitude (mv), duration (ms), and area ( V/s). The M wave was produced using a supramaximal twitch delivered to the vastus lateralis muscle as detailed previously by our group (18). No verification occurred to determine whether the M wave was maximized for each individual. However, in pilot work, we have determined that the voltage used for the twitch was well in excess of that needed to induce a maximal level for each of the M-wave properties of interest. Electromyogram recordings were made with 10-mm-diameter Ag- AgCl (Meditrace 60) surface electrodes. Electrodes were placed over the belly of the vastus medialis (interelectrode distance was 2 cm) with one ground electrode positioned on the lateral epicondycle of the fibia. The skin was shaved, abraded, and cleaned with alcohol before electrode placement. Electrode positioning was kept constant for each individual for each of the M-wave determinations. The electromyogram signal (bandwidth of 20 500 Hz) was passed through an alternating-current amplifier (National Instruments, AT-M10 16H multifunction board). The gain was calibrated to optimize signal amplitude for analog-to-digital conversion and collected at 2,048 Hz. Custom-modified National Institute of Allergy and Infectious Diseases software (National Instruments) was used to acquire electromyogram and analyze raw data (Labview 5.1 software routine). For assessment of the properties of the M wave, amplitude was defined as the sum of the absolute values for maximum and minimum points of the biphasic (one positive and one negative deflection); duration was defined as the time from baseline to baseline from the beginning to the end of the biphasic M wave, where the beginning is established as a positive deflection two standard deviations above baseline harmonic mean and the end as a return to baseline; area was calculated as the integral of the absolute value of the entire M wave. Data analyses. Two-way ANOVA procedures for repeated measurements were used to examine differences between exercise (B and A) and time (number of repetitions). Where significance was found, the Newman-Keuls technique was applied to determine which means were significantly different. Pearson s product-moment correlation coefficients were also calculated (using standardized techniques) between the change in MVC and the change in force elicited at each frequency of stimulation for each repetition and for the pooled sample. The probability for statistical significance was accepted at P 0.05. Data are represented as means SE. RESULTS Dietary analyses during intermittent exercise. The average total nutrient intake over the 16 h of intermittent exercise was 2,339 195 kcal (Table 1). The total energy derived from CHO (78.4 1.3%) far exceeded both protein (8.14 0.33%) and fat (13.6 0.54%). Also provided are the nutrient analyses for different segments of the intermittent exercise. As indicated, the percentage of CHO remained high over the different Table 1. Dietary analyses during the experimental period R1 R4 R5 R8 R9 R11 R12 R16 R1 R16 Protein, % 8.45 0.31 7.64 0.54 8.36 0.36 8.09 0.49 8.14 0.33 CHO, % 77.1 1.0 79.5 2.2 78.6 1.2 78.6 1.7 78.4 1.3 Fat, % 14.6 0.77 12.5 1.5 12.9 0.99 13.3 1.3 13.6 0.54 Total, kcal 644 83 465 53 596 72 634 70 2339 195 Values are means SE (n 11) of the percentage of total calories consumed from protein, fat, and carbohydrates (CHO). Total, total kilocalories consumed; R1 R4, repetitions 1 4; R5 R8, repetitions 5 8; R9 R11, repetitions 9 11; R12 R16, repetitions 12 16; R1 16, repetitions 1 16. segments. Caloric intake was also distributed throughout the day. MVC. The repetitive bouts of heavy exercise resulted in both fatigue and weakness (Fig. 1). Fatigue was observed after the first two repetitions, where the force produced was depressed by 14 and 18%, respectively. Weakness, defined as the persistent loss of force and assessed before performance of the heavy exercise bouts, was evident by the fourth repetition where reductions of 14 and 10% were observed compared with B at R1 and R2, respectively. The force exerted during MVC at B did not recover throughout the remaining repetitions. Twitch properties. The initial session of heavy exercise resulted in an 21% reduction in P t (Table 2). A reduction in P t was also observed after the second repetition of the exercise. Thereafter, no reductions in P t were observed with exercise regardless of the number of repetitions performed. The failure of the heavy exercise protocol to induce fatigue after the second exercise session coincided with a reduction in the force elicited before exercise. At B, force failed to recover during the first four repetitions. By R7, force was higher compared with the second repetition but still depressed compared with the value measured before R1. By R12, weakness was not detected since no differences were found with the force measured before the first bout of exercise. The two measures used to assess the kinetics of force development during P t, namely CT and df/dt max, showed a different response to the repetitive cycle exercise (Table 2). In contrast to CT, which was unaffected by exercise, df/dt max was altered both before and after exercise. Reductions in df/dt max were observed after the first two repetitions. Thereafter, the differences between before and after exercise for each of the repetitions were not significant. As with P t, reductions in df/dt max were initially observed before exercise at R2. As the number of repetitions increased, df/dt max increased such that, by R12, no differences were observed with the value recorded before R1. Both properties used to assess relaxation characteristics of the twitch were altered by the exercise protocol (Table 2). In the case of one-half relaxation time, only a main effect occurred when R1 was greater than R2. This effect persisted throughout the remaining sessions. Changes in df/dt max were also observed with exercise that depended on the number of repetitions. At R1 and R2, decreases were observed in df/ dt max. At each of the remaining repetitions, no differences were observed between B and A. By R12 and extending throughout R15, the fatigue observed at R1 and R2 was reversed such that no differences were noted with the values obtained before the first exercise bout.

Fig. 1. Effects of repetitive cycle exercise on maximal voluntary contractile force. Values are means SE; n 12. Repetitions, number of bouts of heavy cycle exercise; B, before exercise; A, after exercise. *Significantly different (P 0.05) from repetition 1. Significantly different (P 0.05) from repetition 2. # Significantly different (P 0.05) from B. Force-frequency responses. The heavy intermittent exercise protocol induced significant changes in isometric force that were dependent both on the frequency of stimulation and the number of repetitions of the exercise that were performed (Fig. 2). Low-frequency fatigue was induced by the first session of exercise as evidenced by the 53% reduction in force that occurred at a stimulation frequency of 10 Hz. The depression in force persisted throughout recovery and was evident at B for all of the remaining exercise sessions. However, some recovery in weakness occurred as the number of exercise bouts increased. Beginning at R7 and extending through R12 and R15, the force at B was higher than that observed at B for R2 and R4. The same general fatigue and weakness pattern was observed at 20-Hz stimulation. At the higher frequencies of stimulation, namely 50 and 100 Hz, some differences in the REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY Table 2. Effects of heavy repetitive exercise on twitch characteristics 2169 response pattern to the repetitions of heavy exercise were observed compared with the 10- and 20-Hz frequencies. Although a pronounced fatigue occurred in response to the first exercise bout, which resulted in a persistent weakness through bout R7, at bouts R12 and R15 the force exerted at B was not different from that observed at B before R1. For R12 and R15, force was not depressed with exercise despite the recovery that was found compared with the start of the repetitive exercise protocol. A typical response pattern illustrating the reduction in force at low and high frequencies with the 6 min of cycling exercise is presented in Fig. 3. Both df/dt max and df/dt max were also measured for each of the frequencies of stimulation. For df/dt max (Table 3), reductions were observed at R1 at all frequencies of stimulation. At each of the following repetitions, no change was observed in df/dt max with the exercise except at R2, where a reduction was found at 20 Hz. The failure of exercise to elicit reductions in df/dt max beyond R1 was associated with a failure of this property to recover between exercise sessions. Although some recovery at 10 Hz occurred with increases in the number of repetitions, full recovery was not observed by R15. The same general pattern was observed at each frequency of stimulation. Decreases in df/dt max were also observed with exercise (Table 4) that persisted during recovery. At R1, a pronounced depression occurred when assessed at 10 Hz. The depression in df/dt max persisted at B throughout the remaining repetitions, although some recovery was noted. No reductions occurred with exercise after R1. The same general response was found for all frequencies of stimulation. In general, changes were observed in the properties of the M wave with exercise and the number of repetitions (Fig. 4). For amplitude, a main effect was found with A generally being higher than B. For area, main effects of both exercise and repetitions were found. For exercise, B was less than A, whereas for repetitions R1 was greater than R2. At R3, lower values were observed than at R1 and R2, followed by a recovery to R1 at R12 and R15. For duration, a main effect of repetitions was observed where R1 was greater than R2. Repetitions, n 1 2 4 7 12 15 P t,n B 135 13 99.0 12* 104 13* 117 12* 131 12 123 8.4 A 106 11 86.3 9.7* 101 14 127 15* 116 11* 121 7.4* CT, ms B 96.2 2.9 96.2 2.5 94.9 3.4 97.5 2.8 88.7 7.9 98.3 2.7 A 102 4.7 102 2.9 98.9 3.7 99.0 3.3 95.6 2.3 95.8 1.5 RT1 2, ms B 75.8 11.6 61.6 6.1 60.6 5.5 58.8 5.6 57.6 5.1 57.9 4.4 A 72.5 6.6 67.4 6.2 66.6 5.3 65.4 7.1 62.2 5.2 60.1 4.4 dp/dt max, N/s B 2,546 299 1,814 244* 1,948 291* 2,146 261* 2,373 301 2,269 232 A 1,808 196 1,473 194* 1,794 268 2,348 321* 2,127 249* 2,202 186* dp/dt max, N/s B 1,683 279 1,314 286* 1,425 298 1,576 293 1,778 294 1,622 216 A 1,152 191 1,020 174 1,156 221 1,526 280* 1,310 197* 1,475 148* Values are means SE (n 12). n, No. of exercise sessions; B, preexercise; A, postexercise; P t, peak twitch tension; CT, contraction time; RT1 2, one-half relaxation time; dp/dt max, maximal rate of tension development; dp/dt max, maximal rate of tension relaxation. *Significantly different (P 0.05) from R1. Significantly different (P 0.05) from R2. Significantly different (P 0.05) from R4. Significantly different (P 0.05) from B. A main effect (P 0.05) of repetitions was found for 1 2 RT. For repetitions, R1 R2 R4 R12 R15.

2170 REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY Correlations between MVC and ES. The correlation was based on the changes observed in MVC force and the changes in force that occurred with different frequencies of stimulation as a result of the exercise (Table 5). In general, the change induced in MVC with exercise during R1 correlated with each frequency of stimulation. The correlates ranged from 0.95 (MVC vs. 20 Hz) to 0.52 (MVC vs. 100 Hz). With a few exceptions, no relationship existed between the changes in MVC force with any of the stimulation frequencies for each of the remaining repetitions. The overall correlation, determined by using the pooled data from each repetition, showed a significant relationship between MVC regardless of the frequency of stimulation. DISCUSSION Fig. 2. Effects of repetitive cycle exercise on isometric force at different frequencies of stimulation. Values are means SE; n 12. A: 10-Hz stimulation. B: 20-Hz stimulation. C: 30-Hz stimulation. D: 50-Hz stimulation. E: 100-Hz stimulation. *Significantly different (P 0.05) from repetition 1. Significantly different (P 0.05) from repetition 2. Significantly different (P 0.05) from repetition 4. # Significantly different (P 0.05) from B. In this study, we have been able to demonstrate that 16 bouts of cycling exercise performed for 6 min each hour at 91% V O2 peak resulted in adaptations that reduce or eliminate the fatigue and weakness induced as an early response. The potent effect of the exercise protocol on muscle fatigue and weakness in the unconditioned state can be appreciated by examining the mechanical response to ES before and after the first repetition of the exercise. After exercise, force was reduced across a range of frequencies, with the greatest relative reduction observed at 10 Hz (53%) and the lowest reduction observed at 100 Hz (17%). At R1, neuromuscular fatigue was also found as indicated by the 14% reduction in MVC. Weakness, defined as the persistent decrease in force during the 54-min period of recovery between exercise repetitions, was also pronounced after R1. Compared with R1, no recovery of force occurred at R2, regardless of the frequency of stimulation. In contrast, some recovery of MVC occurred since no differences were found between R1 and R2 before exercise. During the initial repetitions, the persistent weakness, resulting in low preexercise force levels, were in general associated with a reversal of fatigue since further depressions in force were not observed after exercise. As the number of repetitions of the heavy exercise increased, recovery of force was observed, with the magnitude depending on the frequency of stimulation. At the higher frequencies of 50 and 100 Hz, no differences in force were observed between R1 and R15 before exercise. At R15, fatigue induced by exercise was not observed at any frequency. The repetition-dependent effect on the mechanical behavior of the quadriceps is dramatic testimony to the rapidity with which adaptations can occur to protect the functional integrity of a muscle, which was initially perturbed by a heavy bout of exercise. To appreciate the nature of the adaptations that occur in the mechanical responses over the 16-h period of intermittent exercise, it is necessary to gain insight into the mechanisms responsible for fatigue with the first repetition. Although reductions in MVC occurred with R1 and R2, suggesting a possible inability of central neural processes to maximally activate the quadriceps muscle, our results clearly indicate that much of the functional loss is peripheral in nature. Because the fatigue, which appears to activate the muscle cell by stimulating nerve branches (5, 24), was characterized by using surface electrodes, failure at the neuromuscular junction must be added to the potential list of processes that may be involved in mechanical failure. Among the excitation sites and processes

REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY 2171 Fig. 3. A typical response pattern for a participant highlighting the changes in force occurring at low (10 Hz; left) and high (100 Hz; right) stimulation frequency with 6 min of cycling at 90% peak oxygen consumption. Data was obtained from repetition 1. that may be rate limiting are the sarcolemma and T tubules (which could result in an inability to conduct repetitive action potentials), the coupling between the T tubules and the Ca 2 release channel (CRC) of the SR (resulting in a disturbance in signal transmission to the CRC), and direct alterations to the SR itself (which could modify the recruitment of the contractile apparatus via alterations in the [Ca 2 ] f transient). Changes in [Ca 2 f] transient could occur via disruptions in Ca 2 release from the CRC and/or by disruptions in Ca 2 uptake. Direct changes in contractile function could occur via modifications in the regulatory proteins troponin and tropomyosin, or via changes in the contractile proteins themselves, namely actin and myosin (28). Collectively, these changes, regardless of the site involved, could compromise weak to strong actomyosin binding with consequent loss of force-generating capabilities (16). All of the sites and processes cited have been differentially implicated previously in the loss of muscle to generate a desired and expected force depending on the specifics of the task and the type of preparation employed (16). In an attempt to gain insight into how the changes in MVC force induced by the repetitive exercise correlated with the changes observed at the different frequencies of stimulation, correlation coefficients were calculated. In general, correlations were found only with the first repetition of the exercise. These results indicated that the loss of force assessed with an MVC is positively related to the force loss observed with each frequency of stimulation. This might be expected since MVC includes both central and peripheral components. After each of the remaining repetitions, correlations between MVC and ES, regardless of frequency, were generally not significant. This would suggest no relationship between the voluntary and involuntary induced effects of exercise on force loss. The general lack of a significant relationship after the first repetition was expected given the relatively small fatigue that occurred with each of the subsequent repetitions. It must also be emphasized that the correlations for each repetition are based on small sample sizes. To gain further insight into the potential rate-limiting sites for the fatigue observed in this study, it is necessary to distinguish between the task used to induce the fatigue and the task used to measure the fatigue. The task used to induce the fatigue involved a 6-min bout of heavy dynamic exercise. The measurement of fatigue was accomplished by measuring the isometric characteristics of the quadriceps muscle, which is known to be heavily involved in cycling (30). Moreover, the assessment of fatigue occurred 4 5 min after the end of the cycling exercise. This could have important implications to both the site and mechanism involved in fatigue. The type and duration of the task selected to induce fatigue are known to result in large disturbances in metabolic by-product accumulation (4). Although ATP remains relatively well protected, large decreases in phosphocreatine occur, accompanied by near stoichiometric increases in P i and creatine (4, 9). Significant recruitment of glycogenolysis and glycolysis also occurs, re- Table 3. Effects of heavy repetitive exercise on dp/dt max at different frequencies of stimulation Repetitions, n 1 2 4 7 12 15 10 Hz, N/s B 1,211 159 500 58 a 609 102 a 676 56 a,b 859 121 a,b,c 831 92 a,b,c A 470 47 e 428 52 518 80 787 106 a,b,c 712 97 a,b,c 756 76 a,b,c 20 Hz, N/s B 2,148 266 1,166 238 a 1,252 271 a 1,410 208 a 1,415 202 a 1,337 142 a A 977 160 e 838 138 e 1,026 229 1,431 255 a,b,c 1,215 203 b 1,245 130 a,b 30 Hz, N/s B 2,898 301 1,964 377 a 1,986 337 a 1,942 239 a 2,230 327 a 2,067 214 a A 1,699 270 e 1,767 352 1,749 314 2,342 430 a,b,c 1,951 308 1,943 176 50 Hz, N/s B 3,640 358 2,479 296 a 2,912 528 a,b 2,904 299 a,b 3,276 489 a,b 3,018 278 a,b A 2,516 296 e 2,413 331 2,760 461 3,229 495 a,b,c 3,013 487 a,b 3,000 245 a,b 100 Hz, N/s B 4,071 357 2,974 312 a 3,581 626 b 3,602 369 a 4,295 657 b,c,d 3,808 336 a A 3,221 388 e 3,076 415 a 3,208 456 3,802 415 a,b,c 3,927 563 a,b,c 3,843 301 a,b,c Values are means SE (n 12). 10, 20, 30, 50, and 100 Hz, frequency of electrical stimulation. a Significantly different (P 0.05) from R1. b Significantly different (P 0.05) from R2. c Significantly different (P 0.05) from R4. d Significantly different (P 0.05) from R7. e Significantly different (P 0.05) from B.

2172 REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY Table 4. Effects of heavy repetitive exercise on dp/dt max at different frequencies of stimulation Repetitions, n 1 2 4 7 12 15 10 Hz, N/s B 881 101 335 36 a 379 59 a 458 48 a 512 46 a,b 501 46 a,b A 383 54 e 311 50 375 68 550 64 a,b,c 446 60 b 479 58 b 20 Hz, N/s B 2,735 237 1,333 154 a 1,286 166 a 1,557 195 a 1,585 161 a 1,609 178 a A 1,137 162 e 1,022 142 a,c 1,049 150 1,468 192 a,b,c 1,403 175 b,c 1,494 159 a,b,c 30 Hz, N/s B 3,270 291 2,292 183 a 2,241 221 a 2,510 217 a 3,042 335 a,b,d 3,033 329 b,c,d A 2,107 190 e 2,030 208 2,085 204 2,635 246 a,b,c 2,746 331 a,b,c 2,866 280 a,b,c 50 Hz, N/s B 3,361 286 2,475 169 a 2,712 227 a 3,133 253 b,c 3,789 377 a,b,c,d 3,813 400 a,b,c,d A 2,543 190 e 2,545 177 2,804 241 3,518 320 a,b,c,e 3,598 383 a,b,c 3,865 346 a,b,c 100 Hz, N/s B 3,099 252 2,264 168 a 2,656 254 a 3,068 225 b,c 3,891 442 a,b,c,d 3,805 381 a,b,c,d A 2,412 207 e 2,380 162 2,595 216 3,365 251 a,b,c 3,615 360 a,b,c 3,829 318 a,b,c,d Values are means SE (n 12). 10, 20, 30, 50, and 100, frequency of electrical stimulation. a Significantly different (P 0.05) from R1. b Significantly different (P 0.05) from R2. c Significantly different (P 0.05) from R4. d Significantly different (P 0.05) from R7. e Significantly different (P 0.05) from B. sulting in substantially increased flux rates and production of lactic acid (4, 33). These by-products have been associated with muscle fatigue, operating potentially at one or more of all of the E-C sites reviewed (16). Inhibition of the catalytic activity of the cellular ATPases appears to be particularly vulnerable (27). As a consequence, disturbances in membrane transport for Na and K could affect membrane excitability and the generation of action potentials (10). Alternatively, an inhibition of the Ca 2 -ATPase activity could reduce Ca 2 sequestration into the SR, compromising Ca 2 storage and Ca 2 release (39). The accumulation of selected by-products could by themselves compromise Ca 2 release. As an example, P i, which is known to accumulate in the lumen of the SR, may combine with Ca 2 and reduce release rates (35). In addition, the accumulation of a range of by-products such as ADP, glucose-6-phosphate, and lactate, all of which are known to increase with the type of exercise protocol employed in this study, can affect the open state of the Ca 2 release channel and reduce Ca 2 release rates (35). Finally, metabolic by-product accumulation is known to inhibit actomyosin ATPase activity and to depress weak to strong binding transitions and isometric force (11). As a consequence of these changes, Ca 2 sensitivity is reduced, resulting in a reduction in force at a given [Ca 2 f] (28). Because our fatigue assessment was conducted 4 5 min after cycling, the phosphorylation potential of the muscle would be expected to be normalized, given the approximate half-life time for phosphocreatine regeneration, which has been estimated at 20 22 s (32). Lactic acid, on the other hand, which has a half-life time of 15 min (33), would be expected to be substantially elevated at the time of the fatigue assessment. The normalization of the phosphorylation potential during recovery in exercised muscle has also been shown to correlate with a similar fast component of mechanical recovery (3, 38). Despite the fast recovery that occurs early in exercise, a slower, more delayed phase of recovery also persists while the metabolic profile of the muscle is fully normalized (3). Given the role of muscle glycogen reserves in E-C coupling failure (8), we have also provided for intake of nutrients during the 16 h of exercise and recovery. Over the course of the experimental period, a total of 2,339 195 kcal were consumed, with CHO representing the dominant substrate. The fatigue that we have observed after exercise at R1 does not appear to be explained by disturbances in membrane excitability since it occurred in the absence of reductions in the properties of the M wave. No disturbances were found in the area, amplitude, or duration of the M wave when assessed using the supramaximal twitch. Alterations in the M wave, and specifically reductions in the amplitude, have been noted previously during heavy exercise in untrained subjects (2). These measurements were performed both while the exercise was in progress and for 15 20 min after the exercise (2). It is possible, given our protocol and the time delay between the end of exercise and measurement, that membrane excitability could have recovered. However, as emphasized, decreases in membrane excitability do not appear implicated in the fatigue that was observed at the time of the mechanical assessment. Alterations in the [Ca 2 f] transient, mediated by inhibition of SR Ca 2 release, appear to be at least partly responsible for the loss of the mechanical response. Reductions in Ca 2 release could occur as a result of reduced SR Ca 2 storage secondary to reductions in Ca 2 uptake (39) or to direct alterations in the CRC (15, 35). Previous studies using singlefiber measurements have demonstrated the association between reduced isometric force and depressed SR Ca 2 release during both heavy exercise and recovery (1). The persistent loss of force-generating capacity, which has been defined as weakness (13), was most conspicuous before R2. At this point, no change in force occurred compared with postexercise measurements despite a period of recovery that exceeded 30 min between the points of measurement. At the time of measurement, the phosphorylation potential and metabolic by-product concentrations would be normalized (9). The delayed recovery in mechanical function is consistent with a sustained depression in the Ca 2 transient mediated primarily by reductions in Ca 2 release (1). Particularly intriguing about our results was the observation that, when the 6-min session of exercise was performed with the muscle substantially weakened, no fatigue was generally observed. This would suggest that a critical functional reserve

REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY 2173 in E-C signal transduction remain protected, allowing the heavy exercise session to be performed. What these particular sites and processes are and the mechanisms defending further erosion are uncertain. As emphasized, Ca 2 release remains an inviting site to explain our observations. Before the last repetition of exercise, muscle weakness at the higher stimulation frequencies had been completely eliminated. Remaining, however, was a persistent weakness that was most pronounced at stimulation frequencies of 10 and 20 Hz. The loss of force at low frequencies in the absence of a loss at high frequencies has been attributed to a depression in Ca 2 release (1). The specificity observed is based on the magnitude of the reduction in Ca 2 release and the sigmoid nature of the force-[ca 2 f] curve. At low frequencies, which result in low force levels, a small reduction in [Ca 2 f] can have dramatic reductions in force given the steep nature of the curve (1). At higher stimulation frequencies and higher force levels, due to the asymptomatic properties of the curve, small depressions in [Ca 2 f] can be accommodated without significant effects on force (1). Our results are entirely consistent with this interpretation. A surprising and unexpected result was the observation that, by the last repetition, heavy exercise failed to induce fatigue despite either partial or full recovery of the force levels before exercise. In effect, these observations indicate that muscle function is protected during the heavy exercise in the absence or near absence of weakness. This finding is a dramatic illustration of the rapid adaptations that can be invoked in muscle to defend muscle contractility once compromised by the strain provided by the early exercise repetitions. Based on our analyses, although speculative, adaptations at the level of the CRC channel and Ca 2 release may be involved. This could occur either as a result of changes in the internal environment or as a result of modifications to the CRC itself. Our results provide an exciting opportunity to develop specific hypotheses and to probe for the underlying mechanisms that are involved. Conceivably, the fatigue observed at low frequencies of stimulation could be biased by other factors not resident in the high-frequency protocol. Unlike the stimulations at 50 and 100 Hz, which produce a fused tetanus, stimulations at 10 and 20 Hz produce an unfused tetanus. With an unfused tenanus, changes in the rate of tension development or tension decline could alter the peak force that was observed (17, 37). We have detected that accompanying the reduction in force at varying frequencies of stimulation after R1 were reductions in both df/dt max and df/dt max. The slower df/dt max would be expected to reduce force, whereas the slower df/dt max would be expected to promote increased force (37). Interestingly, Fig. 4. Effects of heavy intermittent exercise on the properties of the muscle compound action potential. Values are means SE; n 12. Details describing the measurement of amplitude (A), duration (B), and area (C) of the M wave appear in METHODS. A: for amplitude, a main effect (P 0.05) of exercise was observed. For exercise, B was greater than A. B: for duration, a main effect (P 0.05) of repetition was observed. For repetitions, repetition 1 was greater than repetition 2. C: for area, main effects (P 0.05) of both exercise and repetitions were observed. For exercise, B was less than A. For repetitions, repetitions 1, 7, 12, and 15 were greater than repetitions 2 and 4. Table 5. Effects of repeated bouts of heavy exercise on the correlations between the changes in MVC force and force at different frequencies of stimulation Repetitions, n 1 2 4 7 12 15 Overall MVC vs. 10 Hz 0.56* 0.21 0.11 0.00 0.08 0.39 0.54* MVC vs. 20 Hz 0.95* 0.17 0.12 0.79* 0.16 0.41 0.65* MVC vs. 30 Hz 0.74* 0.30 0.45 0.13 0.22 0.32 0.55* MVC vs. 50 Hz 0.86* 0.51* 0.08 0.26 0.34 0.30 0.56* MVC vs. 100 Hz 0.52* 0.56* 0.23 0.47 0.25 0.12 0.33* Values are correlations (r) derived from Pearson s product-moment calculations (n 12). MVC, maximal voluntary contractions; 10 Hz, 20 Hz, 30 Hz, 50 Hz, and 100 Hz, stimulation frequency in impulses/s; Overall, overall relationship between the changes for all repetitions of the exercise. *Significantly different from 0 (P 0.05).

2174 REPETITIVE EXERCISE AND MUSCLE CONTRACTILITY these effects were not observed for CT and one-half relaxation time, which are generally considered less sensitive measurements of the kinetics of the force transients. When these properties were examined at the low frequencies of stimulation, similar effects were observed for the twitch, namely a pronounced reduction in both df/dt max and df/dt max after R1. The net effect of the slowing of both of these properties would tend to cancel out, resulting in little effect in peak force. Interestingly, as weakness developed and force was compromised before exercise, a reduced df/dt max and df/dt max persisted. Although a number of processes combine to regulate the kinetics of force development and decay (16), the rates of Ca 2 sequestration and Ca 2 release by the SR remain important factors. Conceivably, changes in these properties are involved in the fatigue-induced responses in the behavior of force transients observed. By necessity, given the 16 h needed to complete the intermittent work protocol, the assessment of fatigue and weakness was made throughout the experimental day. Recent research investigating circadian rhythms in neuromuscular performance in human adductor muscle indicates increases in force and speed of contraction in the evening compared with the morning (29). If these results apply to larger muscles, such as the quadriceps, it is possible that the differences in weakness could be biased because of the circadian effects. The effects of diurnal variation on fatiguability, however, remain unknown. In this study, we have also included two women as part of the participants who volunteered for the study. It is possible that, in response to the intermittent cycling exercise, men would exhibit a greater decrement in neuromuscular performance than women, given recent research that has shown that the ankle dorsiflexors display a more rapid fatiguability in men (31). However, such a possibility remains only speculative given the differences in task requirements and the muscle group studied. The mechanical changes that we have observed with the intermittent exercise must also be put in context. The persistent weakness observed during MVC reflects the maximal forcegenerating capacity of the neuromuscular system. In contrast, due to safety concerns, the ES protocol generated a maximum of only 50% MVC. As a consequence, a large population of motor units are not activated during ES. This could lead to misleading conclusions regarding the magnitude and frequency-dependent nature of fatigue and weakness (12). However, it must be emphasized that, at least for the muscle fibers recruited by ES, the heavy intermittent exercise protocol produced novel findings regarding the nature and time course of adaptation. GRANTS The authors appreciate the financial support provided by the Natural Sciences and Engineering Research Council for this research. REFERENCES 1. Allen GA, Balnave CD, Chin ER, and Westerblad H. Failure of calcium release in muscle fatigue. In: Biochemistry of Exercise X, edited by Hargreaves M and Thompson M. Champaign, IL: Human Kinetics, 1998, p. 135 146. 2. Arnaud S, Zattara-Hartmann C, Tomei C, and Jammes Y. Correlation between muscle metabolism and changes in M-wave and surface electromyogram: dynamic constant load leg exercise in untrained subjects. Muscle Nerve 20: 1197 1199, 1997. 3. Baker AJ, Kostov KG, Miller RG, and Weiner MW. Slow force recovery after long-duration exercise: metabolic and activation factors in muscle fatigue. J Appl Physiol 74: 2294 2300, 1993. 4. Bergström J, Guarnieri G, and Hultman E. Carbohydrate metabolism and electrolyte changes in human muscle tissue during heavy work. J Appl Physiol 30: 122 125, 1971. 5. Bergström M and Hultman E. Contraction characteristics of the human quadriceps muscle during percutaneous electrical stimulation. Pflügers Arch 417: 136 141, 1990. 6. Bigland-Ritchie B, Garland SJ, Garland SJ, and Walsh ML. Taskdependent factors in fatigue of human voluntary contractions. Adv Exp Med Biol 384: 361 380, 1995. 7. Bigland-Ritchie B and Woods JJ. Changes in muscle contractile properties and neural control during human muscle fatigue. Muscle Nerve 7: 691 699, 1984. 8. Chin ER and Allen DG. Effects of reduced muscle glycogen concentration on force, Ca 2 -release and contractile protein function in intact mouse skeletal muscle. J Physiol 498: 17 29, 1997. 9. Clarkson P and Sayers SP. Etiology of exercise-induced muscle damage. Can J Appl Physiol 24: 234 238, 1999. 10. Clausen T, Nielsen OB, Harrison AP, Flatman JA, and Overgaard K. The Na -K pump and muscle excitability. Acta Physiol Scand 162: 183 190, 1998. 11. Cooke R and Pate E. The inhibition of muscle contraction by the by-products of ATP hydrolysis. In: Biochemistry of Exercise VII, edited by Taylor B. Champaign, IL: Human Kinetics, 1990, p. 59 72. 12. Davies CTM and White MJ. Muscle weakness following dynamic exercise in humans. J Appl Physiol 53: 236 241, 1982. 13. Edwards RHT. Physiological analysis of skeletal muscle weakness and fatigue. Clin Sci Mol Med 54: 463 470, 1978. 14. Edwards RHT, Hill DK, Jones DA, and Merton PA. Fatigue of long duration in human skeletal muscle after exercise. J Physiol 272: 769 778, 1977. 15. Favero TG. Sarcoplasmic reticulum Ca 2 release and muscle fatigue. J Appl Physiol 87: 471 483, 1999. 16. Fitts RH. Cellular mechanisms of muscle fatigue. Physiol Rev 74: 49 94, 1994. 17. Fowles JR and Green HJ. Co-existence of potentiation and low-frequency fatigue during voluntary exercise in human skeletal muscle. Can J Physiol Pharmacol 81: 1092 1100, 2003. 18. Fowles JR, Green HJ, Tupling R, O Brien S, and Roy BD. Human neuromuscular fatigue is associated with altered Na -K -ATPase activity following isometric exercise. J Appl Physiol 92: 1585 1593, 2002. 19. Gandevia SC. Neural control in human muscle fatigue: changes in muscle afferents, motoneurones and motocortical drive. Acta Physiol Scand 162: 275 283, 1998. 20. Garland SJ and Gossen ER. The muscular wisdom hypothesis in human muscle fatigue. Exerc Sport Sci Rev 30: 45 49, 2002. 21. Green HJ. Cation pumps in skeletal muscle: potential role in muscle fatigue. Acta Physiol Scand 162: 201 213, 1998. 22. Green HJ, Tupling R, Roy B, O Toole D, Burnett M, and Grant S. Adaptations in skeletal muscle exercise metabolism to a sustained session of heavy intermittent exercise. Am J Physiol Endocrinol Metab 278: E118 E126, 2000. 23. Hughson RL, Kowalchuk JM, Prime WM, and Green HJ. Open-circuit gas exchange analysis in the non-steady state. Can J Appl Sport Sci 5: 15 18, 1980. 24. Hultman E, Sjöholm H, Jäderholm EK, and Krynicki J. Evaluation of methods for electrical stimulation of human skeletal muscle in situ. Pflügers Arch 398: 139 141, 1983. 25. Jones DA. High- and low-frequency fatigue revisited. Acta Physiol Scand 156: 265 270, 1996. 26. Jones S, Green H, and Ashton N. Reproducibility of muscle force output during submaximal and transcutaneous stimulation (Abstract). Med Sci Sports Exerc 15: 146, 1983. 27. Korge P and Campbell KB. The importance of the ATPase microenvironment in muscle fatigue: a hypothesis. Int J Sports Med 16: 172 179, 1995. 28. MacIntosh BR. Role of calcium sensitivity modulation in skeletal muscle performance. News Physiol Sci 18: 222 225, 2003. 29. Martin A, Carpentier A, Guissard N, Van Hoecke J, and Duchateau J. Effect of time of day on force variation in a human muscle. Muscle Nerve 22: 1380 1387, 1999. 30. Patla AE. Some neuromuscular strategies characterizing adaptation process during prolonged activity in humans. Can J Sport Sci 12: 33S 44S, 1982.