HUBAL, MONICA J.; RUBINSTEIN, SCOTT R.; CLARKSON, PRISCILLA M.
Strenuous exercise involving lengthening muscle actions causes temporary muscle damage, as indicated by direct markers such as ultrastructural disruptions (11,12) and indirect markers such as declines in muscle function and the appearance of myofibrillar proteins in the blood (1,6,21,28). Marked muscle dysfunction, especially the profound loss in voluntary force-producing capability after lengthening exercise, is considered one of the best indirect indicators of muscle damage in humans (28), but it is also highly variable among subjects, even in populations exposed to similar exercise protocols. For example, in a recent study of 157 subjects who completed 50 maximal lengthening actions of the elbow flexors, responses ranged from no strength loss to > 90% strength loss immediately after exercise (7).
Although the underlying mechanisms driving strength loss after lengthening actions have been investigated (29), they remain poorly understood, and little is known regarding the large intersubject variability in strength loss. The capacity to generate force in skeletal muscle is a product of both neural drive from the brain to the muscle and peripheral contractile function. Variability in strength losses could potentially originate with changes either in the neural or peripheral systems. Decreased neural drive to muscle can occur through a reduction in the number of activated motor units or a decreased motoneuron firing rate (13). Impaired peripheral muscle function could result from many different mechanisms, including damaged contractile proteins, damaged sarcolemma, disrupted excitation-contraction coupling, and altered intracellular environment, such as altered pH or ATP levels (2,10,30).
Exercise consisting of lengthening actions causes immediate and prolonged deficits in peripheral muscle function during the days after exercise, as noted in many studies (15,17,19). This is most often observed as decreased torque in response to motor nerve or surface stimulation. For example, Prasartwuth et al. (19) have documented declines in twitch torque and alterations in twitch contractile properties, occurring immediately after exercise. Studies have also demonstrated alterations in central function after exercise (19) and during recovery over the following days (8,9). However, only one study to date has examined whether alterations in neural drive may explain the high intersubject variability in strength loss.
Sayers et al. (27) found no differences in EMG/torque at 36 h after exercise between groups with large or small strength losses, implying that differences in strength loss were not generated by differing levels of central function between groups. However, the Sayers et al. (27) study did not directly test either central activation or peripheral function, nor did it test the relationship of neural drive to strength losses before 36 h postexercise. Therefore, it is currently unknown whether the documented decreases in central activation (19) contribute to early strength losses in humans after lengthening exercise.
The purpose of this study was to directly examine central and peripheral muscle fatigue after lengthening actions of the elbow flexors in objectively defined groups stratified by postexercise strength loss. We hypothesized that the higher-strength loss group would demonstrate similar central function but impaired peripheral function compared with lower responders, providing evidence that variability in postexercise strength loss is generated within the periphery and not via impaired neural signaling to muscle.
Forty-six healthy young (age 18-29 yr) adult men and women participated in the study. All subjects signed an informed consent document approved by the University of Massachusetts institutional review board. Subjects had not resistance-trained with their upper extremities in the past 6 months and had no prior history of musculoskeletal injuries to the nondominant shoulder, elbow, or wrist.
During visit 1, subjects performed a baseline maximal voluntary contraction (MVC) test. MVC was also collected on visits 2 and 3, with 48 h between sessions. Muscle function properties (using both voluntary and stimulated contractions) and electromyographic data were also collected on baseline visits 2 and 3. After 10 min of rest after baseline testing on visit 3, subjects performed 50 maximal lengthening actions of the elbow flexors. The specific postexercise sequence for muscle function testing was identical to that used during baseline testing. MVC was also collected during recovery on postexercise days 1-4 (visits 4-7), with 24 h between sessions. The overall study timeline is depicted in Figure 1A.
FIGURE 1-A, Study ti...Image Tools
Elbow-flexion strength testing.
Strength testing consisted of measures of elbow-flexion maximal voluntary isometric contractile torque, both on an isokinetic dynamometer (MVC) and on a custom-designed contractile property apparatus (maximal isometric torque (MIT)). The MIT device was warranted because the physical setup of the Biodex for elbow-flexion testing prohibited the collection of clean EMG signal from the biceps (because this area is covered tightly with a Velcro strap). Furthermore, the MIT device allowed for better stabilization and isolation of the elbow flexors, and the MIT device was more sensitive to lower forces such as those that result from twitch stimuli. It should be noted that MVC and MIT measured isometric strength at slightly different body positions; the MIT device secured each subject in 90° elbow flexion and 90° shoulder flexion, and the MVC was collected at 90° elbow flexion and 80° shoulder flexion.
MVC was assessed on an isokinetic dynamometer (Biodex System 3; Biodex, Shirley, NY) and preceded all other measures. Each subject performed five 3-s contractions with 1 min of rest between contractions. For the postexercise MVC, only a single 3-s contraction was collected, immediately after the final lengthening action. All maximal isometric strength efforts were accompanied by strong verbal and visual feedback to ensure maximal efforts. Data were sampled at 100 Hz, and a 100-point moving average was used to smooth the data for analysis. MVC was defined as the average of the maximal torque for the highest three trials (of five) or the single highest point (postexercise MVC).
MIT was collected on a custom-built device in which the subject was seated in an upright position. The chest was securely stabilized by a wide Velcro strap, and the nondominant arm was secured in the apparatus. The hand and wrist were set in an orthosis (#1011, Orthomerica Products Inc., Newport Beach, CA) and placed so that the midpoint of the wrist fit snugly against a SM 250 force transducer (Interface, Scotsdale, AZ). Both seat height and transducer height were adjusted for each subject so that the force transducer was at the level of the styloid processes of the radius and ulna. A padded metal plate against the back of the orthosis stabilized the arm during contractions.
The signal from the force transducer was amplified (Grass P15 Amplifier; Grass Instrument Company, Braintree, MA), converted to digital signals using an A/D board (Model # CIO-DAS08/Jr-AO 12-bit; Computerboards, Inc., Mansfield, MA), and interfaced to a personal computer using Win30 hardware and Dasylab software (DASYTEC USA, Amherst, NH). Baseline signal (i.e., tension at rest caused by pressure from the back plate) was subtracted from all measures of torque before subsequent analysis.
Three 3-s trials were collected with 1 min of rest between trials. Data were sampled at 1000 Hz. MIT was defined as the peak force of the force/time output after smoothing data to a modified Gaussian curve, which allowed us to reliably measure what torque could be sustained for a 1-s time period. Torque (N·m) was calculated for each subject by multiplying the peak force applied at the force transducer by the distance (m) from the elbow to the center of the force transducer, which could be adjusted vertically to account for different arm lengths.
Elbow-flexion contractile property testing.
Stimulated torque was measured on two baseline days and after exercise on the custom-designed apparatus. Stimuli were created via a constant-current high-voltage stimulator (DS7AH; Digitimer Ltd., Hertfordshire, England). Stimulation to the musculocutaneous nerve at Erb's point was applied via a Teca stimulator wand (Model # 9523-1; Teca, Pleasantville, NY) with an interelectrode spacing of 2.5 cm. Stimuli were triggered using custom-written Dasylab programs.
To prepare the subject for electromyography, the skin over the biceps brachii muscle of the nondominant arm was abraded and cleaned with alcohol pads (#326895; Becton Dickinson Co., Franklin Lakes, NJ) and then allowed to dry. Two reusable silver-silver chloride electrodes (Safelead F-E9; Astro-Med, West Warwick, RI) coated with a conductivity gel (Teca, Pleasantville, NY) were attached to the skin over the midline and parallel to the fibers of the biceps brachii muscle. Electrodes were attached to the skin using electrode washers (E410 washers; In Vivo Metric, Ukiah, CA). An interelectrode (center to center) distance of 20 mm was used, with the distal electrode placed approximately midlength of the muscle. A ground electrode was also taped to the skin overlying the anterior deltoid of the same arm used for EMG measurements. To improve the accuracy of the EMG electrode placements, electrode placement for all three electrodes were marked each day with indelible ink. Reliability of repeated electrode preparations is reported in the Results section.
The EMG data were sampled at 1000 Hz, wide-band filtered (10 Hz to 10 kHz), and amplified (×500) using a preamplifier (Grass P15 Amplifier; Grass Instrument Company, Braintree, MA). Amplified EMG signals were then converted to digital signals using an A/D converter board (Model # CIO-DAS08/Jr-AO 12-bit; Computerboards, Inc., Mansfield, MA) and Dasylab software (DASYTEC USA, Amherst, NH). The voluntary surface EMG signal (sEMG) was analyzed using MATLAB software (Mathworks, Inc., Natick, MA). The evoked compound muscle action potential (CMAP) was measured as the peak-to-peak amplitude of the EMG signal elicited from a single supramaximal stimulus (0.1-ms duration). The voluntary sEMG waveform was then full-wave rectified and integrated (iEMG) for a 0.5-s interval beginning with the onset of contraction and normalized to the peak torque generated. Twitch torque was calculated from the force channel during CMAP stimulation (torque = force × distance from elbow to force transducer).
Neural activation assessment.
The assessment of neural activation and peripheral contractile properties followed the same protocol progression during each testing session (Fig. 1B). First, stimulator output was optimized by increasing voltage until the greatest CMAP peak-to-peak amplitude was achieved using stimuli 0.1 ms in duration with 30 s between contractions. Typically, two to four contractions were used to optimize output and ensure maximal twitches. After a 1-min rest period, a 50-Hz, 500-ms stimulus (TET) was then applied to resting muscle with a stimulator output 20% greater than optimal CMAP voltage to ensure supramaximal response. One minute after the single stimulated train of resting muscle, we used the interpolated train method to assess central activation. Subjects were instructed to perform a maximal isometric contraction, and a 50-Hz, 500-ms train was triggered by the investigator when the force-time output showed a plateau of maximal strength output. The central activation ratio (CAR) of the elbow flexors was computed as the ratio of MIT to peak total force (MIT/MIT + supramaximal train torque) during the interpolated train (16). Finally, subjects performed three 3-s maximal isometric torque (MIT) trials with 1 min of rest between attempts. The final MIT trials were done to ensure that there were no differences between groups from immediately after exercise to 5 min after exercise. Resultant data were then imported into curve-analysis software (Tablecurve 2D V5.0, SYSTAT, Richmond, CA) or MATLAB software (Mathworks, Inc., Natick, MA) for analysis.
Measures of maximal eccentric torque (MET) during exercise were collected on an isokinetic dynamometer (Biodex System 3, Shirley, NY). The exercise protocol involved 50 maximal lengthening actions of the elbow flexors of the nondominant arm grouped in five sets of 10 actions each, with 12 s between repetitions and 2 min of rest between sets. The action speed was 60°·s−1 through the subject's range of motion. This exercise protocol was adapted from our previous work (22,24-27) as one that produces moderate to high levels of muscle damage. A single MVC contraction at a 90° elbow angle was done immediately after the final lengthening action.
Data analysis and cluster analysis.
Reliability of all measures was assessed using intraclass correlation coefficients for baseline measures (three baseline measures for MVC and two baseline measures each for MIT, MET, CMAP, EMG, torque after tetanic stimuli (TET), and torque after twitch stimuli (TWT)). Two-factor ANOVA (time × gender) were also used for each of these factors between baseline days to ensure that there were no significant differences between baseline days. We included gender as a factor in these ANOVA because of the presence of more women in the higher-responder (HR) group and because average strength in men is greater than that of women.
Subjects were grouped according to their strength loss response to lengthening actions. Initial groupings were based on MVC strength loss at 0 h after exercise so that we could assess the contributions of neural and peripheral changes to the extent of muscle dysfunction after exercise. Because strength losses at this time point (0 h) could also potentially include metabolic muscle fatigue unrelated to muscle damage, we then repeated the cluster analysis using the 24-h MVC measure. Cluster analysis was done using Cluster V2.2 (Eisen software, Berkeley, CA), and Treeview V1.6 (Eisen software, Berkeley, CA) was then used to visualize the resultant clusters. This protocol allowed us to test for contributions of the central nervous system and the periphery on both immediate (0 h) and lasting (24 h) muscle dysfunction.
To test our hypothesis, MVC, MIT, MET, CMAP, EMG, TET, and TWT were assessed using two-way ANOVA (group × time) with repeated factors over time. CAR was assessed via the nonparametric Wilcoxon signed rank test. Differences were deemed significant at P < 0.05. Significant differences reported in the ANOVA tests were further tested using Tukey's honestly significant difference (HSD) post hoc tests.
Subject characteristics and baseline measures.
Forty-six subjects (22 males, 24 females) completed the study. Subjects' physical characteristics are presented in Table 1. Intraclass correlations for the baseline days were: MVC (R = 0.98), MIT (R = 0.95), TET (R = 0.96), TWT (R = 0.85), CMAP (R = 0.94), and EMG (R = 0.86). ANOVA for the baseline days detected no significant effects of time. Thus, the reliability of the measures was judged to be good.
Cluster analysis grouping.
To objectively identify higher responders to exercise, we grouped the entire cohort using mathematical cluster analysis. Primary cluster analysis was done on the postexercise MVC (percent loss from baseline) measure taken immediately after the final lengthening action. Three clusters were apparent (Fig. 2), comprising a group of nonresponders (NR; N = 3; postexercise strength greater than preexercise strength), a group of lower responders (LR; N = 22; average loss = 23%), and a group of higher responders (HR; N = 21; average loss = 49%) (Fig. 3). Furthermore, we ran the same analysis with the 24-h MVC time point (LR average = 17%; HR average = 33%) with identical results, meaning that those identified as higher responders on the basis of strength loss immediately after exercise were still identified as higher responders on the basis of their persistent strength loss at 24 h. Results were also consistent using MIT loss (at 5 min after exercise) as the clustering variable.
Although there was a sex-distribution difference between groups (6 men/15 women in HR; 16 men/6 women in LR), there were no between-group differences in age (P = 0.93), height (P = 0.11), weight (P = 0.30), or arm length (P = 0.11). In addition, there were no demographic differences between the NR group and the HR or LR groups for these measures (Table 1).
Because the aim of the study was to examine the contributions of central and peripheral changes to strength loss, nonresponders (no strength loss) were subsequently eliminated from further data analysis because of absence of strength loss.
Central and Peripheral Muscle Function between HR and LR.
Absolute values for all strength measures are presented in Table 2. The repeated-measures ANOVA of MVC at six time points (pre- to immediately postexercise and for 4 d after exercise) showed significant effects (P < 0.05) for time, group, and interaction. Post hoc analysis (Tukey HSD test) determined that HR and LR were significantly different with regard to MVC at the time points from immediately after exercise through 48 h postexercise only. Relative MVC loss is depicted in Figure 3 and is the measure on which clustering was performed. As such, it was expected that these values would be different between groups immediately after exercise. Postexercise MVC losses for the cluster groups have been described in a previous section. The LR group returned to baseline values at 72 h after exercise, whereas the HR group was significantly impaired for 96 h after exercise.
The ANOVA of MIT from pre- to postexercise showed a significant main effect for time (P < 0.05) and interaction term (P < 0.05), but no group main effect (P = 0.24). The HR group lost an average of 32.4 ± 2.5% MIT, whereas the LR lost an average of 19.4 ± 2.8% MIT at 5 min after exercise (P < 0.05).
The ANOVA of MET from pre- to postexercise showed a significant main effect for time (P < 0.05) and group (P < 0.05), but no interaction effect (P = 0.28). However, relative MET loss was significantly different between groups (P < 0.05). The HR group lost an average of 43.2± 3.8% MET, whereas the LR group lost an average of 30.3 ± 3.9% MET at the end of the exercise protocol.
Stimulated muscle function.
Absolute values for all evoked muscle function measures are presented in Table 2. The ANOVA of TWT from pre- to postexercise showed a significant main effect for time (P < 0.05), but no group (P = 0.08) or interaction effects (P = 0.33). However, relative TWT loss was significantly different between groups (P < 0.05). The HR group lost an average of 54.7 ± 3.6% TWT, whereas the LR lost an average of 42.4 ± 4.6% TWT after exercise.
The ANOVA of TET from pre- to postexercise showed a significant main effect for time (P < 0.01) and interaction (P < 0.05), but no group effect (P = 0.08). Relative TET loss was significantly different between groups (P < 0.05; Fig. 4).
FIGURE 4-Torque deri...Image Tools
In summary, all measures of voluntary and evoked muscle function after lengthening exercise were decreased in the entire cohort, and relative torque losses were exacerbated in the HR group compared with the LR group.
Neural and central activation measures.
Absolute values for all neural measures are presented in Table 3. The ANOVA of EMG from pre- to postexercise showed no significant differences for time (P = 0.09), group (P = 0.19), or interaction (P = 0.74). Relative EMG losses were not different between groups (P = 0.08; 3.9 ± 7.4% loss in LR and 20.3 ± 5.0% loss in HR).
The ANOVA of CMAP from pre- to postexercise showed no significant differences for group (P = 0.159) or interaction (P = 0.88). There was a significant time effect (P < 0.05). Relative CMAP losses were not significantly different between groups (P = 0.75; 16.0 ± 5.4% loss in LR and 19.0 ± 7.8% loss in HR). Overall, groups demonstrated reductions of approximately 17% in CMAP after exercise.
The Wilcoxon test analysis of CAR from pre- to postexercise showed no significant main effect for group (P = 0.19), Fig. 5). Both groups decreased slightly over time for CAR.
FIGURE 5-Central act...Image Tools
In summary, of the measures of neural and central activation after lengthening exercise, only CMAP was significantly decreased in the entire cohort (with trends for reduction in EMG and CAR), and no measures were different between HR and LR.
Although it is well known that strenuous lengthening actions can result in pronounced losses in muscle strength (1,6,21), a large variability exists in the severity of the strength loss, even when subjects are exposed to a standardized exercise stimulus (3,5,7,14,18,23). For example, in this study, the strength loss of the entire sample ranged from no strength loss (NR group) to 62%. The purpose of this study was to examine central and peripheral neuromuscular parameters that might contribute to the variability in voluntary strength loss after lengthening actions.
We compared neuromuscular function in two groups of subjects separated by higher versus lower strength loss after maximal lengthening actions of the elbow flexors. The main finding of this study was that central function (i.e., neural drive to muscle) was similar between groups, despite large differences in voluntary function. Conversely, all measures of peripheral function (i.e., twitch and tetanus torques generated via direct nerve stimulation) were impaired more in those with greater voluntary strength loss, indicating that the mechanism driving early strength losses after lengthening actions is within the muscle and not within the central pathways that activate the muscle.
We first stratified groups on the basis of the loss in maximal isometric voluntary contraction torque (MVC) immediately after exercise, a highly reliable measure (typical intraclass R > 0.95) that is also considered one of the best indirect indicators of muscle damage in humans (28). The clustering procedures identified three groups of subjects: higher responders (average torque loss = 49%), lower responders (average torque loss = 23%), and a group of nonresponders (postexercise MVC ≥ baseline). Subsequent cluster analysis using the 5-min postexercise torque loss (MIT) and the 24-h postexercise MVC loss revealed that these groupings are stable until 2 d after exercise. Groups were not different from each other according to any demographic measure (i.e. age, height, weight) or any baseline measure, although there seemed to be a difference in sex distribution; the higher responders included 7 men and 14 women, whereas the lower responders included 14 men and 8 women.
Several studies have described the disturbance of neuromuscular variables after lengthening exercise (8,9,19,20), suggesting that central failure might contribute to strength loss after exercise and during recovery. Prasartwuth et al. (19) describe decreases in voluntary activation in the elbow flexors immediately after lengthening exercise, especially at short muscle lengths (20). However, studies of central disturbances during recovery have demonstrated equivocal results. Prasartwuth et al. (19,20) describe changes persisting up to 24 h after exercise, but they found no differences in voluntary activation from 2 to 8 d after exercise. Similarly, Endoh et al. (9) did not find changes in voluntary activation during MVC at 2 and 4 d after exercise. However, Deschenes et al. (8) have reported decreased neuromuscular efficiency (iEMG/torque) for 10 d after exercise-induced damage in the quadriceps, indicating that the muscle required increased neural drive to maintain any given torque output. Therefore, studies directly examining voluntary activation have found decreases during the early recovery phase, when strength loss is maximal (6) (i.e., immediately after exercise to 24 h after exercise), but not from 2 d onwards. To date, only Sayers et al. (27) have addressed the contribution of central function after lengthening exercise in groups of higher and lower strength loss, and this study has reported no differences between groups for neuromuscular efficiency (EMG/MVC torque) at 36 h after exercise. Therefore, this study is the first to address possible contributions of voluntary activation dysfunction to early losses in strength after lengthening exercise.
The current study has assessed voluntary activation by superimposing a tetanic stimulation on a MVC, which has been found to be a valid measure of muscle activation (4). We found that subjects were able to produce high levels of activation in the biceps during MVC before exercise, and although there was a trend for a reduction in the CAR in the entire cohort (independent of group) after exercise, the CAR still exceeded 95% activation. We also have demonstrated that CMAP and EMG activity were not different between groups before or after exercise. Thus, the current study provides direct evidence that central function does not contribute to the variability in early strength losses after exercise.
In summary, the data from this study demonstrate that although some central changes occur after lengthening actions, these changes are not associated with early strength losses after exercise. Conversely, peripheral function is significantly affected by lengthening actions and is exaggerated in those with the greatest strength losses, indicating that factors driving variability in early strength losses are located within the periphery.
This work was supported by an American College of Sports Medicine Foundation Research Grant and a National Strength and Conditioning Association Graduate Student Research Grant. The authors would like to thank Dr. Jane Kent-Braun and Dr. David Russ for their valuable advice throughout the study; Laura Gould, Steve Trahan, and Stephanie Dallaire for their help in collecting the data; and Ian Lanza and Katherine Eck for their help in data processing.
1. Allen, D. G. Eccentric muscle damage: mechanisms of early reduction of force. Acta Physiol. Scand.
2. Allen, D. G. Skeletal muscle function: role of ionic changes in fatigue, damage and disease. Clin. Exp. Pharmacol. Physiol.
3. Beaton, L. J., M. A. Tarnopolsky, and S. M. Phillips. Variability in estimating eccentric contraction-induced muscle damage and inflammation in humans. Can. J. Appl. Physiol.
4. Behm, D., K. Power, and E. Drinkwater. Comparison of interpolation and central activation ratios as measures of muscle inactivation. Muscle Nerve
5. Clarkson, P. M., and C. Ebbeling. Investigation of serum creatine kinase variability after muscle-damaging exercise. Clin. Sci. (Lond).
6. Clarkson, P. M., and M. J. Hubal. Exercise-induced muscle damage in humans. Am. J. Phys. Med. Rehabil.
7. Clarkson, P. M., E. P. Hoffman, E. Zambraski, H. Gordish-Dressman, A. Kearns, M. Hubal, B. Harmon, and J. M. Devaney. ACTN3 and MLCK genotype associations with exertional muscle damage. J. Appl. Physiol.
8. Deschenes, M. R., R. E. Brewer, J. A. Bush, R. W. McCoy, J. S. Volek, and W. J. Kraemer. Neuromuscular disturbance outlasts other symptoms of exercise-induced muscle damage. J. Neurol. Sci.
9. Endoh, T., T. Nakajima, M. Sakamoto, and T. Komiyama. Effects of muscle damage induced by eccentric exercise on muscle fatigue. Med. Sci. Sports Exerc.
10. Fitts, R. H. Cellular mechanisms of muscle fatigue. Physiol. Rev.
11. Friden, J., M. Sjostrom, and B. Ekblom. Myofibrillar damage following intense eccentric exercise in man. Int. J. Sports Med.
12. Friden, J. Changes in human skeletal muscle induced by long-term eccentric exercise. Cell Tissue Res.
13. Gandevia, S. C. Spinal and supraspinal factors in human muscle fatigue. Physiol. Rev.
14. Gulbin, J. P., and P. T. Gaffney. Identical twins are discordant for markers of eccentric exercise-induced muscle damage. Int. J. Sports Med.
15. Jones, D. A., D. J. Newham, and C. Torgan. Mechanical influences on long-lasting human muscle fatigue and delayed-onset pain. J. Physiol.
16. Kent-Braun, J. A. Central and peripheral contributions to muscle fatigue in humans during sustained maximal effort. Eur. J. Appl. Physiol. Occup. Physiol.
17. Newham, D. J., K. R. Mills, B. M. Quigley, and R. H. Edwards. Pain and fatigue after concentric and eccentric muscle contractions. Clin. Sci. (Lond).
18. Nosaka, K., and P. M. Clarkson. Variability in serum creatine kinase response after eccentric exercise of the elbow flexors. Int. J. Sports Med.
19. Prasartwuth, O., J. L. Taylor, and S. C. Gandevia. Maximal force, voluntary activation and muscle soreness after eccentric damage to human elbow flexor muscles. J. Physiol.
20. Prasartwuth, O., T. J. Allen, J. E. Butler, S. C. Gandevia, and J. L. Taylor. Length-dependent changes in voluntary activation, maximum voluntary torque and twitch responses after eccentric damage. J. Physiol.
571(Pt 1):243-252, 2006 Feb 15.
21. Proske, U., and T. J. Allen. Damage to skeletal muscle from eccentric exercise. Exerc. Sport Sci. Rev.
22. Saxton, J. M., P. M. Clarkson, R. James, M. Miles, M. Westerfer, S. Clark, and A. E. Donnelly. Neuromuscular dysfunction following eccentric exercise. Med. Sci. Sports Exerc.
23. Sayers, S. P., P. M. Clarkson, P. A. Rouzier, and G. Kamen. Adverse events associated with eccentric exercise protocols: six case studies. Med. Sci. Sports Exerc.
24. Sayers, S. P., P. M. Clarkson, and J. Lee. Activity and immobilization after eccentric exercise: I. Recovery of muscle function. Med. Sci. Sports Exerc.
25. Sayers, S. P., and P. M. Clarkson. Force recovery after eccentric exercise in males and females. Eur. J. Appl. Physiol.
26. Sayers, S. P., C. A. Knight, P. M. Clarkson, E. H. Van Wegen, and G. Kamen. Effect of ketoprofen on muscle function and sEMG activity after eccentric exercise. Med. Sci. Sports Exerc.
27. Sayers, S. P., C. A. Knight, and P. M. Clarkson. Neuromuscular variables affecting the magnitude of force loss after eccentric exercise. J. Sports Sci.
28. Warren, G. L., D. A. Lowe, and R. B. Armstrong. Measurement tools used in the study of eccentric contraction-induced injury. Sports Med.
29. Warren, G. L., C. P. Ingalls, D. A. Lowe, and R. B. Armstrong. What mechanisms contribute to the strength loss that occurs during and in the recovery from skeletal muscle injury? J. Orthop. Sports Phys. Ther.
30. Westerblad, H., and D. G. Allen. Cellular mechanisms of skeletal muscle fatigue. Adv. Exp. Med. Biol.
538:563-570, 2003. discussion 571.