INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

WHEN AN INDIVIDUAL SUSTAINS a submaximal isometric contraction until exhaustion, there is a progressive increase in motor unit activity that can include a change in the number of active motor units and a modulation of discharge rate (4, 11, 14, 15, 40). Despite maximal effort during a low-force contraction, the net motor unit activity at exhaustion, as indicated by the interference electromyogram (EMG), often does not achieve the maximal values recorded before the onset of the fatiguing contraction (2, 13). The existence of a deficit in activation at exhaustion suggests that endurance time could be prolonged if it is possible to alter the size of the deficit.

Such adjustments in motor unit activity during fatiguing contractions can be modified by chronic interventions. Our laboratory found, for example, that several weeks of limb immobilization alters the pattern of muscle activation and influences the endurance time during a submaximal contraction performed with the elbow flexor muscles (35, 36). When the fatiguing contraction was performed before the arm was immobilized in a plaster cast for 4 wk, all subjects exhibited the progressive increase in EMG during an isometric contraction that was sustained at 15% of the maximal voluntary contraction (MVC) force. Immediately after the cast was removed, 7 of the 12 subjects (6 women and 1 man) used intermittent muscle activation and a lower rate of increase in the EMG, and this was associated with a 220% increase in endurance time. The other five subjects, who were all men, used the same EMG pattern as they did before the intervention, and they experienced no change in endurance time.

These observations raised two questions: Does muscle activation change with repeat performances of the task, and is the effect related to the sex of the individual? The purpose of the study was to compare the patterns of muscle activation and endurance time during a submaximal fatiguing contraction of the elbow flexor muscles performed on three occasions by men and women. Because young women appear to use the biceps brachii muscle more often than young men during activities of daily living (21, 36), we expected to find that any adaptations in muscle activation would be greater for the women. As with the immobilization study, we found that some subjects did exhibit an adaptation in muscle activation. However, these responders included both men and women. A preliminary account of these results has been published in abstract form (18), and related data have been published elsewhere (17).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fourteen healthy adults [7 men and 7 women; age 27 ± 4 yr (SD), range 22-35 yr] volunteered to participate in the study. All the subjects were right handed with no known neurological disorders and were naive to the experimental protocol and procedures. The physical activity levels of the subjects were estimated with a questionnaire that evaluates a 7-day history of activity (33). Before participation in the study, all subjects gave informed consent, and the Human Subjects Committee at the University of Colorado approved the protocol.

Mechanical Recording

Electrical Recordings

Cardiovascular Measurements

Experimental Protocol

MVC force. Subjects performed three MVC trials with the elbow flexor muscles and three MVC trials with the elbow extensor muscles. The MVC task consisted of a gradual increase in force from zero to maximum over 3 s, with the maximal force held for 2-3 s. Subjects were able to observe the exerted force on a monitor and were verbally encouraged to achieve maximal force. Subjects rested for 60-90 s between trials. If the peak forces from two of the three trials were not within 5% of each other, additional trials were performed until this was accomplished. The greatest force achieved by the subject was taken as the MVC force and used as the reference to determine the target force for the fatiguing contraction. Subsequently, the MVC force for the elbow extensor muscles was determined by using the same procedures. Within 20 s of completion of the fatiguing contraction, an MVC was performed with the elbow flexor muscles.

Fatiguing contraction. The fatiguing contraction was performed at a target force of 20% MVC as determined from the MVC performed on that day. The absolute change in target force relative to session 1 was 0.7 ± 5.1 N and did not differ between the subjects who increased endurance time (responders) and those who did not (nonresponders) (Table 1). The subject was required to match the target force displayed on the monitor and was verbally encouraged to sustain the force for as long as possible. The contraction was terminated when the target force was <16% MVC force for ~3 s, despite strong verbal encouragement; this time was recorded as the endurance time. Subjects were not informed of the endurance times until completion of the third session.

Data Analysis

The MVC force was quantified as the average value over a 0.5-s interval that was centered about the peak force. Similarly, the maximal EMG for each muscle was determined as the average value over a 0.5-s interval that was centered about the peak rectified EMG.

The amplitude of the force fluctuations was quantified as the coefficient of variation (CV = SD/mean × 100) of the first, middle, and last 60 s of the fatiguing contraction. To compare the force fluctuations at the same absolute time across the three sessions, the CV was calculated for five 30-s epochs at the same absolute time relative to session 1 for each subject. The five 30-s epochs were the first 30 s; 15 s on both sides of the time at 25, 50, and 75% of endurance time; and the last 30 s of the endurance time for session 1.

The average EMG during the fatiguing contraction was quantified in two ways: 1) for statistical analyses, as averages of the rectified EMG (AEMG) for the first, middle, and last 60 s of the fatiguing contraction; and 2) for graphic presentation, as AEMG for every 1% of the endurance time. The EMG for each muscle was normalized to the maximum EMG (percentage of MVC) obtained during its respective MVC. The normalized EMGs for the short and long heads of biceps brachii did not differ during the fatiguing contraction on the basis of a statistical comparison; hence the two signals were averaged to yield a single value for biceps brachii.

Because low-force fatiguing contractions of long duration are characterized by bursts of EMG activity, we quantified the number and duration of these bursts throughout the fatiguing contraction for each muscle (Fig. 1). The rectified EMG signal was 1) smoothed with a low-pass filter at 2 Hz for surface EMG signals and at 3.8 Hz for the intramuscular EMG (brachialis), 2) differentiated to identify rapid changes in the EMG signal, and 3) divided by the AEMG so that muscles with different EMG amplitudes could be compared. A burst was identified when the smoothed, differentiated EMG signal increased by >0.4 s¹ for the surface EMG and 0.36 s¹ for the intramuscular EMG. These values represented ~3 SDs above the mean of the smoothed, differentiated EMG signal that had been calculated from 25 samples when the EMG signal displayed minimal bursting during the contraction. The end of a burst was identified as the time when the smoothed EMG signal decreased to the same amplitude as at the start of the burst. When the EMG signal did not decline to the same EMG amplitude at the start of the burst, however, the end of the burst was then identified as the time that the smoothed, differentiated EMG signal was most negative before the start of the next burst. This criterion represented the time at which the signal decreased most rapidly before the beginning of the next burst. Based on pilot observations, the start of two bursts was constrained to be >2 s apart and the minimum burst duration was 0.5 s. Burst rate and duration were compared at relative and absolute times during the fatiguing contractions.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Interference electromyogram (EMG) during a fatiguing contraction often exhibited bursts of activity. Record shows EMG activity in the short head of biceps brachii and the subsequent data analysis to identify the bursts of activity. First 2 traces indicate the interference and rectified EMGs, respectively. The rectified EMG was low-pass filtered at 2 Hz (3rd trace) and then differentiated and divided by the averages of the rectified EMG (AEMG; 4th trace) to identify rapid changes in the signal. A burst was identified when the smoothed and differentiated EMG signal exceeded a threshold set at 0.4 s1 (dashed horizontal line on 4th trace). In this example, the analysis identified 6 bursts of EMG activity. The beginning of each burst is identified by an arrow and the duration of the burst by the solid lines in the third trace. Trace at bottom indicates the vertical force exerted by the wrist.

Heart rate and mean arterial pressure during the fatiguing contraction were analyzed by comparing 10- to 15-s averages at 10% intervals throughout the endurance time. For each interval, the blood pressure signal was analyzed for the mean peaks [systolic blood pressure (SBP)], mean troughs [diastolic blood pressure (DBP)], and the number of pulses per second (multiplied by 60 to determine heart rate). Mean arterial pressure (MAP) was calculated for each epoch with the following equation: MAP = DBP + 1/3(SBP  DBP).

Statistical Analysis

A four-factor (muscle × response × session × time) ANOVA was used to compare the AEMG of the elbow flexor muscles and burst rate and burst rate at the same absolute time of session 1. Because bursts were absent during some time periods for most subjects, averages of the burst duration are reported and the results of independent t-tests reported where these analyses were possible. Post hoc analyses (Tukey-Kramer) were used to test for differences when appropriate. A significance level of P < 0.05 was used to identify statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The purpose of the study was to compare changes in muscle activation and endurance time across repeat performances of a fatiguing contraction. Because the changes in endurance time comprised a bimodal distribution, we categorized the subjects into those who increased endurance time across sessions (responders) and those who did not (nonresponders) (Fig. 2A). Accordingly, we first describe the changes in endurance time for these two groups of subjects before examining the associated changes in muscle activation.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   A: percent change in endurance time between sessions 1 and 3 for individual subjects assigned as either responders (n = 9) or nonresponders (n = 5). B: mean (± SE) endurance time for the fatiguing contractions across 3 sessions for the responders (n = 5 men, 4 women) and nonresponders (n = 2 men, 3 women). Endurance times for the responders and nonresponders were similar for session 1, but the responders increased endurance times in sessions 2 and 3 (P < 0.05).

The responders (5 men and 4 women) increased endurance time from session 1 to session 3 by >15% [60 ± 28% (SD), range 25-115%], whereas the nonresponders (2 men and 3 women) did not experience a change in endurance time (3 ± 11%, 12-15%). The endurance times for session 1 were similar for the responders and nonresponders (1,265 ± 753 vs. 1,153 ± 607 s; P > 0.05). An interaction between session and response (P < 0.05) indicated that the responders significantly increased endurance time from session 1 to session 3 (Fig. 2B). The estimated physical activity levels were not different for the responders (49.7 ± 11.0 kcal · day¹ · kg body mass¹) and the nonresponders (44.5 ± 8.6 kcal · day¹ · kg body mass¹).

The MVC force was similar for the responders (289 ± 117 N) and nonresponders (290 ± 140 N) before the fatiguing contraction and across sessions (P > 0.05). The MVC force after the fatiguing contraction was reduced to similar levels for both the responders (35.9 ± 18.8%) and the nonresponders (31.3 ± 16.6%; P > 0.05), and this decrease was consistent across sessions (Fig. 3, A and B).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Maximal voluntary contraction (MVC) forces before and immediately after the fatiguing contraction for the 3 sessions. Values are means ± SE. Responders (A) and nonresponders (B) had similar MVC forces before and after the fatiguing contraction (P > 0.05) for all 3 sessions.

RPE

Adjustments in AEMG

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Representative data for a fatiguing contraction performed by a male responder for session 1 (A) and session 3 (B). The interference EMG of the elbow flexors (4 traces at top of each panel) increased progressively throughout the contraction with significant bursting in the middle and later stages of the contraction. There was minimal EMG activity in the triceps brachii. Average force (traces at bottom) remained constant, but the fluctuations increased toward the end of each contraction. Duration of the fatiguing contraction for session 3 was 108% longer than that of session 1. Force fluctuations and bursts of EMG were less at the same absolute time in session 3 compared with session 1.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   AEMG (normalized to the maximal MVC value) of the elbow flexor and extensor muscles throughout the fatiguing contraction for responders (A-C) and nonresponders (D-F) across the 3 sessions. Each data point represents the mean AEMG for 1% of the endurance time. SE bars are shown for the first, middle, and last data points for each muscle. AEMG for biceps brachii represents the mean AEMG for the short and long heads.

Because endurance time increased across sessions for the responders but reached comparable terminal AEMG values, the rate of increase in AEMG declined across sessions (P < 0.05). For the responders, the average rate of increase in AEMG for the elbow flexor muscles was 0.87 ± 0.54%/min for session 1, 0.78 ± 0.64%/min for session 2, and 0.68 ± 0.62%/min for session 3. Furthermore, the average rate of increase in AEMG for nonresponders was greater than that for responders across sessions (P < 0.05). For the nonresponders, the average rate of increase in AEMG for the elbow flexor muscles was 2.00 ± 2.04%/min for session 1, 1.76 ± 1.57%/min for session 2, and 1.31 ± 1.81%/min for session 3. These rates for the nonresponders were influenced by the marked increase in terminal AEMG exhibited by some muscles (Fig. 5, D-F). For example, the average rate of increase in AEMG for the elbow flexor muscles of the nonresponders over the first 600 s was 1.23 ± 0.95%/min for session 1, 1.01 ± 0.95%/min for session 2, and 0.52 ± 0.61%/min for session 3.

The average rates of increase in AEMG varied among the individual muscles. For both groups of subjects, there was a close association between the AEMG values for biceps brachii and brachioradialis, except for the nonresponders near the end of the task in sessions 1 and 2 (Fig. 5, D and E). Similarly, the brachialis AEMG consistently began at a greater value (18 ± 6%) compared with the biceps brachii and brachioradialis (11 ± 3%), ended at a similar value (31 ± 10%), and increased at a lower average rate (P < 0.05). Nonetheless, there were differences among the muscles in the increase in AEMG of the individual muscles for the two groups of subjects. For the responders (Fig. 5, A-C), the average rate of increase in AEMG across sessions was similar for brachialis (0.60 ± 0.70%/min), whereas the average rate for biceps brachii and brachioradialis declined from session 1 (1.09 ± 0.50 and 0.96 ± 0.51%/min, respectively) to session 3 (0.66 ± 0.55 and 0.71 ± 0.62%/min, respectively). Although the average rate of increase in AEMG also declined for the nonresponders across sessions (Fig. 5, D-F), the rates did not achieve values similar to the responders, except for the brachialis muscle in session 3 (0.52 ± 0.69%/min). For example, the average rate of increase in AEMG of the nonresponders in session 3 was 2.19 ± 2.97%/min for biceps brachii and 1.22 ± 0.95%/min for brachioradialis. Qualitatively, the AEMG for the nonresponders appeared to change in all three sessions from the progressive linear increase at ~600 s when the AEMG reached ~20% MVC.

Compared with the elbow flexor muscles, the change in AEMG for the triceps brachii during the fatiguing contractions was much less (P < 0.05) and was similar for the responders and nonresponders (P > 0.05). The average values of the triceps brachii AEMG across all three sessions was 1.67 ± 0.93% for the first 60 s and increased to 3.78 ± 2.10% for the last 60 s (P < 0.05). This increase in triceps brachii AEMG (Fig. 5) was similar across sessions for both the responders (0.10 ± 0.10%/min) and nonresponders (0.17 ± 0.14%/min; P > 0.05). There was, however, a significant interaction between session and time, which indicated that the rate of increase in triceps brachii AEMG declined across sessions for both the responders and nonresponders. The average rate of increase in triceps brachii AEMG was 0.17 ± 0.13%/min for session 1 compared with 0.07 ± 0.05%/min for session 3.

Bursting of EMG Activity

The burst rate for all elbow flexor muscles and the triceps brachii increased progressively during the fatiguing contractions in all sessions for both the responders and nonresponders (P < 0.05; Fig. 6A). Burst rate was similar across all muscles (P > 0.05) and similar for responders and nonresponders (P > 0.05). There was, however, an interaction between time and session because the burst rate of the elbow flexor muscles in the last third of the endurance time of sessions 2 and 3 was less than that for session 1 (P < 0.05; Fig. 6A). In contrast, burst rate for triceps brachii was similar for session 1 (0.52 ± 0.69%/min), session 2 (0.44 ± 0.50%/min), and session 3 (0.43 ± 0.55%/min; P > 0.05). Both the responders and nonresponders exhibited these effects across sessions (P > 0.05).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Burst rate of the rectified EMG of the elbow flexor muscles for the fatiguing contractions. Values are means ± SE. There was no difference in the burst rate between the elbow flexor muscles. A: average burst rates for the first third, middle third, and last third of the endurance time for sessions 1-3. There was an increase in the burst rate during the fatiguing contraction (P < 0.05) for all 3 sessions. However, an interaction between session and time indicated that the burst rate in the last third of the endurance time was less in sessions 2 and 3 compared with session 1. B: average burst rates for sessions 1-3 at the same absolute endurance time of session 1. An interaction between session and time indicated that the burst rates in the middle third and last third of the endurance time in sessions 2 and 3 (on the basis of session 1) were less than that in session 1. C: average burst rates for the responders and nonresponders for sessions 1-3 over the same duration of endurance time of session 1. An interaction between session and response indicated a greater reduction in the burst rate across sessions for the responders compared with the nonresponders.

The burst rate was compared at the same absolute time across sessions on the basis of the endurance time of session 1 for each subject. There was a significant reduction in burst rate across sessions such that the burst rate at the same absolute time was less in sessions 2 and 3 compared with session 1 (Fig. 6B; P < 0.05). There was an interaction between session and time (P < 0.05) because burst rate was less for sessions 2 and 3 during the middle and last thirds of the endurance time (on the basis of the endurance time of session 1) (Fig. 6B; P < 0.05). Burst rate was similar for the elbow flexor muscles during sessions 2 and 3 (P > 0.05) when compared at the same absolute time of session 1. In contrast, the responders showed a greater reduction in average burst rate across sessions compared with the nonresponders at the same absolute time of session 1 (Fig. 6C; P < 0.05). There was a significant negative relation (r = 0.42; P < 0.05; n = 42) between endurance time and burst rate, which meant that longer endurance times were associated with a reduction in the rate of bursts in the AEMG.

These differences in burst rate were attributable to the two groups of subjects producing similar numbers of EMG bursts. Across all three sessions, the average number of bursts was 19 ± 29 for the responders and 10 ± 26 for the nonresponders (P > 0.05). Furthermore, the burst number did not change across sessions for the responders (21 ± 39, 17 ± 23, and 20 ± 24; P > 0.05) and the nonresponders (13 ± 15, 7 ± 8, 10 ± 11; P > 0.05). Because the responders increased endurance time, they were therefore able to delay the bursts of EMG activity.

The mean burst duration during the fatiguing contractions and across the three sessions was 5.8 ± 7.2 s. There was an increase in burst duration across the sessions for the elbow flexor muscles: session 1 (4.4 ± 2.2 s) was less than session 2 (7.8 ± 10.0 s; P = 0.02), and there was a trend toward significance in session 3 (6.6 ± 7.8 s; P = 0.08). The burst duration across sessions was similar for the responders (6.5 ± 9.1 s) and nonresponders (5.7 ± 4.7 s; P > 0.05), and the two groups showed a similar burst duration across the three sessions. Furthermore, the burst duration of the elbow flexor muscles was similar over the three sessions (P > 0.05): short head of biceps brachii (6.6 ± 8.1 s), long head of biceps brachii (5.9 ± 7.4 s), brachioradialis (7.0 ± 7.2 s), and brachialis (6.9 ± 11.5 s). The burst duration for triceps brachii was similar across sessions (3.4 ± 1.5, 4.1 ± 1.6, and 4.1 ± 2.6 s; P > 0.05) and similar between responders (4.1 ± 2.3 s) and nonresponders (3.7 ± 1.5 s; P > 0.05). The burst duration for triceps brachii (3.8 ± 1.9 s) tended to be briefer (P = 0.08) than that for the elbow flexor muscles (6.2 ± 7.8 s).

Force Fluctuations

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Coefficient of variation (CV) for force during the fatiguing contractions performed by the responders (A) and nonresponders (B) in the three sessions. Values are means ± SE for the first 60 s, middle 60 s, and last 60 s of the fatiguing contraction. The responders (A) and nonresponders (B) increased CV during the fatiguing contractions (P < 0.05). The responders had similar increases in CV for the 3 sessions. In contrast, the nonresponders had a reduction in CV at the end of sessions 2 and 3 compared with session 1 (P < 0.05).

The CV for force was compared between the responders and nonresponders at the same absolute time across sessions on the basis of the endurance time of session 1 for each subject. At the start of the fatiguing contraction, the CV for force in sessions 1-3 was similar for the responders (2.7 ± 1.0%) and nonresponders (2.7 ± 0.5%; P > 0.05). At the endurance time of session 1, however, the CV for sessions 2 and 3 was less than for session 1 (P < 0.05) for both the responders and nonresponders. The CV for the responders at the endurance time of session 1 was 9.3 ± 4.7% for session 1, 6.8 ± 3.0% for session 2, and 4.7 ± 1.9% for session 3. The CV for the nonresponders at the endurance time of session 1 was 17.2 ± 10.2% for session 1, 8.6 ± 3.6% for session 2, and 9.9 ± 3.0% for session 3. The attenuation of the increase in the CV for force across sessions was greater for the responders compared with the nonresponders (P < 0.05).

MAP and Heart Rate

Heart rate increased during the fatiguing contraction (P < 0.05), with a greater increase in session 1 compared with sessions 2 and 3, for both responders and nonresponders (P < 0.05). Heart rate was less for responders compared with nonresponders throughout the fatiguing contraction (P < 0.05), and this was consistent across all three sessions. At rest, heart rate was similar for responders and nonresponders (61 ± 9 vs. 67 ± 8 beats/min; P = 0.22). At the start of the contraction, the heart rate for responders was less than that of nonresponders (70 ± 9 vs. 82 ± 7 beats/min; P < 0.05) and also at exhaustion (95 ± 16 vs. 110 ± 12 beats/min; P < 0.05). The change in heart rate from the beginning to the end of the fatiguing contraction was similar for responders and nonresponders (35 ± 22 vs. 37 ± 10%; P > 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The purpose of this study was to compare the endurance time and muscle activation patterns of men and women during a fatiguing contraction performed on three occasions with the elbow flexor muscles. Endurance time increased across the three sessions for some men and women (responders) but not others (nonresponders). The responders exhibited a reduced rate of increase in muscle activation across sessions but similar levels of activation at exhaustion. Although the nonresponders displayed a significant reduction in the rate of increase in muscle activation across sessions, the rate of increase in AEMG was greater than that for the responders. The interference EMG comprised bursts of activity, which increased in number during the fatiguing contraction for both the responders and nonresponders. However, the increase in burst rate during the contraction declined across sessions, and there was a significant negative relation between burst rate and endurance time. In contrast to these differences in muscle activation, the pressor response was similar for the responders and nonresponders across sessions. Consequently, alterations in the level and pattern of muscle activation and not cardiovascular adjustments were associated with changes in endurance time for fatiguing contractions performed with the elbow flexor muscles. Furthermore, alterations in muscle activation and endurance time were not specific to either the sex or the strength (MVC force) of the individual.

Comparable Fatigue for the Responders and Nonresponders

Previous analyses of these data (17) have indicated that women had a longer endurance time than men and that this difference was associated with both a lower absolute target force during the sustained contraction and a reduced pressor response. The sex difference in endurance time, target force, and pressor response was consistent across sessions. In the present analysis, however, both men and women demonstrated the adaptations in endurance time across sessions while maintaining a constant target force across the three sessions. Thus, despite the acute adjustments over the three sessions for some men and women (responders) when performing the contraction at the same target force, the sex difference in endurance time persisted. These results indicate that the relatively acute adaptation of endurance time across sessions was independent of the sex of the subject and also the target force (Table 1). Consequently, the mechanisms underlying the longer endurance time of women compared with men must be different from those responsible for the increases in endurance time exhibited by the responders.

Adaptations of Endurance Time Were Not Associated With the Pressor Response

In this present study, both the responders and nonresponders had a reduced increase in the MAP and heart rate in session 3 compared with sessions 1 and 2. This reduction in the rate of increase in the pressor response suggests there was an adaptation in the central command or the metaboreflex in session 3 for both groups of subjects. Because the increase in MAP and heart rate was similar for the responders and nonresponders, however, the increase in the endurance time for the responders across sessions was independent of the adaptations in the pressor response. Consequently, the differential changes in endurance time for the responders and nonresponders could not be explained by circulatory adjustments or the metabolic state of the muscle. Furthermore, the mechanism for the sex difference in endurance time (17) that was related to the pressor response and the target force exerted was consistent across the three sessions and must be different from that for the increase in the endurance time of the responders.

Muscle Activation Differed for Responders and Nonresponders

We assessed triceps brachii activity during the sustained contractions because an increase in antagonist muscle activity can exacerbate the appearance of fatigue in an agonist muscle (29). However, the increase in the level of coactivation was minor and similar for both the responders and nonresponders across sessions. Consequently, coactivation of the antagonist muscle did not account for the different adaptations in endurance time for the responders and nonresponders.

The increase in EMG activity of an agonist muscle during a fatiguing contraction at a submaximal intensity is attributable to an increase in motor unit activity (7, 9, 26), perhaps including a contribution from a reduction in signal cancellation (6, 41). Because the modulation of discharge rate during such contractions is rather modest, the increase in AEMG is largely due to the recruitment of motor units (4, 5, 11, 14, 37, 38). The gradual increase in AEMG that we observed, therefore, likely involved the progressive recruitment of more fatigable motor units. Accordingly, the prolongation of endurance time across sessions should have been associated with a decline in the rate of increase in AEMG and a reduction in the rate of EMG bursts.

The results that were most consistent with this rationale were the more pronounced decline in burst rate across sessions for the responders and the negative relation between burst rate and endurance time. This effect is evident in the records of the individual shown in Fig. 4, which indicates that the number of bursts in session 1 (39, 18, 0, and 41 for brachialis, brachioradialis, biceps long, and biceps short, respectively) declined to almost zero (0, 0, 0, and 5, respectively) over the same absolute time in session 3. Because a burst of EMG activity corresponded to the transient recruitment of additional motor units, the decline in burst rate indicated a delay in the recruitment of more fatigable motor units and hence enabled the prolongation of endurance time.

Changes in the rate of increase in AEMG across sessions were also consistent with the rationale that the increase in endurance time was attributable to the activation of fewer fatigable motor units. Although the average rate of increase in AEMG decreased across sessions for both groups of subjects, the rate remained greater for the nonresponders. Furthermore, the AEMG for the nonresponders exhibited a qualitative change at ~20% MVC when it went from a linear increase to a much more variable signal, including marked increases in the terminal values for some muscles. The responders did not use this pattern of activity. The significance of this observation may be related to reports (16, 24, 30) that find target forces of 15-20% MVC force to distinguish between short and long endurance times for sustained contractions. On the basis of the cumulative sum of the tetanic forces of motor units, a force of 20% MVC would require the activation of about two-thirds of the motoneuron pool (10, 22), which can correspond to the region in which the fatigability of motor units changes (23). Once the descending drive is sufficient to recruit more fatigable motor units during a fatiguing contraction, they appear to be activated transiently in bursts of activity with burst rate increasing up to exhaustion.

This interpretation that the improvement in endurance time across sessions experienced by some subjects was due to adjustments made in motor unit activity raises a significant but unanswered question. What adjustment enabled the target force to be sustained with a lesser rate of increase in motor unit activity? A likely explanation, which awaits investigation, is that the input received by the motoneuron pool can change with practice, including changes in either the descending drive or the feedback from peripheral receptors. Our results indicate that the perceived effort of the responders was less during the contractions performed during sessions 2 and 3, although similar at exhaustion. These results further suggest that modulation of the excitatory descending drive partly contributed to an enhanced endurance time of the responders. In contrast, there was no difference in the pressor response between the two groups of subjects, which suggests that the feedback from group III and IV afferents did not contribute to the adaptation across sessions.

Fatiguing isometric contractions performed at low to moderate forces are characterized by a progressive increase in the amplitude of the force fluctuations (27, 36). This phenomenon appears to be adaptable because we observed that the increase in the fluctuations of the force exerted at the wrist during the fatiguing contractions declined in sessions 2 and 3 for the responders and nonresponders when compared with session 1. The adaptations in the force fluctuations paralleled the changes in AEMG across sessions for the two groups. Furthermore, there was a significant association between EMG burst rate and force fluctuations. Consequently, the mechanisms responsible for the adaptations in the force fluctuations and EMG activity may have been similar and underscore the modulation of the force fluctuations by central sources.

In summary, practice of a sustained isometric contraction at a submaximal intensity with the elbow flexor muscles enabled some men and women, but not others, to prolong the endurance time to exhaustion. This improvement in performance was independent of the sex of the individual, target force, and the pressor response. However, the changes in endurance time were associated with a reduced rate of increase in motor unit activity, as indicated by a decrease in EMG burst rate, a decline in the rate of increase in the AEMG, and an alteration in the distribution of AEMG activity among the synergist muscles. These results indicate that the maximum duration of a low-force contraction can be limited by the neural activation of muscle.