Journal of Neurophysiology

Load Type Influences Motor Unit Recruitment in Biceps Brachii During a Sustained Contraction

Stéphane Baudry, Thorsten Rudroff, Lauren A. Pierpoint, Roger M. Enoka

Abstract

Twenty subjects participated in four experiments designed to compare time to task failure and motor-unit recruitment threshold during contractions sustained at 15% of maximum as the elbow flexor muscles either supported an inertial load (position task) or exerted an equivalent constant torque against a rigid restraint (force task). Subcutaneous branched bipolar electrodes were used to record single motor unit activity from the biceps brachii muscle during ramp contractions performed before and at 50 and 90% of the time to failure for the position task during both fatiguing contractions. The time to task failure was briefer for the position task than for the force task (P = 0.0002). Thirty and 29 motor units were isolated during the force and position tasks, respectively. The recruitment threshold declined by 48 and 30% (P = 0.0001) during the position task for motor units with an initial recruitment threshold below and above the target force, respectively, whereas no significant change in recruitment threshold was observed during the force task. Changes in recruitment threshold were associated with a decrease in the mean discharge rate (−16%), an increase in discharge rate variability (+40%), and a prolongation of the first two interspike intervals (+29 and +13%). These data indicate that there were faster changes in motor unit recruitment and rate coding during the position task than the force task despite a similar net muscle torque during both tasks. Moreover, the results suggest that the differential synaptic input observed during the position task influences most of the motor unit pool.

INTRODUCTION

The nervous system controls the force a skeletal muscle exerts by modulating the number of motor units that are activated (motor unit recruitment) and the rate at which motor neurons discharge action potentials (rate coding) (Kernell 2006). Motor unit recruitment is orderly and based on in the relative sizes of the motor neurons (Henneman 1957), and the force at which a motor unit is activated is known as its recruitment threshold. In humans, recruitment threshold is determined during a voluntary contraction in which force is increased gradually (Desmedt and Godaux 1978; Enoka et al. 1989). The recruitment threshold depends on the sum of the forces developed by the units activated before the test unit and on the extrinsic synaptic input received by the motor neurons (Desmedt and Godaux 1978; Farina et al. 2009). The synaptic input also determines the discharge characteristics of the motor neuron (Kernell 1965; Schwindt 1973).

Both motor unit recruitment and rate coding are modulated during sustained contractions (Carpentier et al. 2001; Christova and Kossev 2001; Enoka et al. 1989; Garland et al. 1994) and, as emphasized in a recent review (Enoka and Duchateau 2008), the changes depend on the characteristics of the task being performed. It has been difficult to determine the relative influence of changes in motor unit recruitment and rate coding on the decline in force that occurs during a fatiguing contraction because muscle fatigue can be caused by the impairment of multiple mechanisms (Enoka and Stuart 1992; Gandevia 2001). One experimental strategy that is used to evaluate the relative significance of such adjustments is to compare the rate of change in selected parameters during two tasks that have similar mechanical requirements but different performance characteristics (Maluf and Enoka 2005).

A number of studies have used this strategy to examine the adjustments that occur when individuals push against a rigid restraint (force task) compared with producing the same net muscle torque to maintain the position of the limb while supporting a mass (position task) (Hunter et al. 2002; Klass et al. 2008b; Maluf et al. 2005; Mottram et al. 2005; Rudroff et al. 2005, 2007). Although the decline in the maximal muscle torque at task failure is similar for the two tasks, the time to task failure is briefer and the rate of fatigue development is faster for the position task compared with the force task. EMG recordings of interference signals and single motor unit action potentials suggest that the motor unit pool is recruited more rapidly during the position task (Maluf et al. 2005; Mottram et al. 2005). The greater recruitment of motor units during the position task seems necessary to accommodate the faster decline in discharge rate during this task so that the target force can be maintained (Mottram et al. 2005). Such an interpretation predicts that the recruitment thresholds of motor units, even those not activated at the beginning of the fatiguing contraction, should decline more rapidly during the position task. We tested this hypothesis by comparing the change in motor unit recruitment threshold and discharge characteristics when the force and position tasks were sustained for a similar duration. Some of these data have been presented previously in abstract form (Baudry et al. 2008b).

METHODS

After informed consent was obtained, experiments were conducted on 24 subjects (9 women) between 18 and 37 yr of age (mean, 25.0 ± 5.7 yr). None of the participants reported any signs of neurological disorder or cardiovascular disease. Subjects were all right-handed and were asked to refrain from exercising the arm muscles for 24 h before testing. The Human Subjects Committee at the University of Colorado in Boulder approved the experimental procedures.

Experimental apparatus

Subjects were seated upright in an adjustable chair with the left forearm vertical and supinated and the upper arm horizontal (1.57 rad of flexion at the shoulder and elbow). The left hand and forearm were placed in a modified wrist-hand-thumb orthosis (Orthoamerica, Newport Beach, CA) to limit hand movement and to restrain movement about the pronation-supination axis of the forearm. The orthosis was connected to the force transducer during the force task and used to suspend an inertial load during the position task. Force was measured with a JR-3 Force-Moment Sensor (900-N range, 90.0 N/V, Woodland, CA) during the force task, and the signal was displayed on a monitor [1% maximal voluntary contraction (MVC)/cm] placed in front of the subject at eye level and stored on a personal computer. The force under the elbow joint was measured with an Entran transducer (ELW-D1–2001, 0.27 mV/N, Hampton, VA). Elbow joint angle was measured during the position task with an electrogoniometer (SG110 and K800, Biometrics, Cwmfelinfach, Gwent, UK) secured to the lateral aspect of the elbow joint. The output of the goniometer was recorded, displayed on a monitor (0.017 rad/cm), and stored on a personal computer. An inertial load equivalent to 15% MVC force was suspended from the subject's wrist via a line-and-pulley system for the position task with the line attached at the same location that contacted the force transducer during the force task (Rudroff et al. 2007). A load cell (SBO-200, Transducer Techniques, Temecula, CA) was attached in series with the inertial load to measure force fluctuations during the position task. The force and position signals were digitized at 200 samples/s.

EMG recordings

EMG signals were recorded with bipolar surface electrodes (Ag-AgCl, 8 mm diam; 20-mm distance between electrodes) that were placed over the short and long heads of biceps brachii, brachioradialis, triceps brachii (lateral head), and anterior head of the deltoid muscle. The EMG signals were amplified (×2,000), band-pass filtered (13–1,000 Hz; Coulbourn Instruments, Allentown, PA), and recorded on a personal computer. The EMG signals were digitized at 2,000 samples/s.

Muscle fiber action potentials from single motor units in biceps brachii were recorded with branched bipolar electrodes (stainless steel, 50 μm diam; California Fine Wire, Grover Beach, CA). The electrode contained three exposed areas where the insulation was removed: two on one wire (branched) and one on the other wire (monopolar) directly across from the midpoint of the branched pair (Enoka et al. 1988). Each exposed area had a width of 1 mm, and the space between each exposed area was 1 mm. The length of the branched (2 short-circuited surfaces) and the monopolar-exposed surfaces was 3 mm. The wires were secured together with all-purpose gel, leaving the recording sites exposed. A disposable 25-gauge hypodermic needle was used to insert the branched bipolar electrode under the skin (not penetrating the fascia) and over the belly of the biceps brachii muscle for a distance of 3–8 cm and was removed before recording motor-unit activity. The electrode was positioned perpendicular to the direction of the muscle fibers.

Two electrodes were inserted subcutaneously in each experiment: one over the short head of biceps brachii and another over the long head of biceps brachii. The short and long heads were distinguished by palpation. However, the discharge of action potentials by motor units was only recorded from one electrode in each experiment. The selected signal was based on the signal-to-noise ratio of the action potentials of at least one motor unit. The signal was digitized at 20,000 samples/s. Single motor unit recordings were amplified (1,000–2,000 times), band-pass filtered (20–8,000 Hz), displayed on an oscilloscope, and stored on a computer.

Methods

Each subject visited the laboratory on five occasions with ≥72 h between each visit. The first visit was a familiarization session in which subjects were introduced to the equipment and procedures. The next two sessions involved performing the force and position tasks to failure. The force task required the subject to maintain a force that was equal to 15% of the MVC force. The position task involved maintaining the elbow joint at a right angle while supporting an inertial load that was equivalent to 15% MVC force. Subjects were provided with visual feedback of the force (1% MVC/cm) exerted at the wrist during the force task and of the elbow angle (0.017 rad/cm) during the position task. These visual gain settings resulted in a similar amount of on-screen movement of the cursor caused by the typical fluctuations observed in the force and position tasks, respectively (Mottram et al. 2005). The criteria for terminating the force task were an inability to sustain the force within 5% of the target value for >3 s without correction, despite strong verbal encouragement, or lifting the elbow away from the elbow force transducer for >3 s. The criterion for ending the position task was an inability to maintain the elbow angle within 0.2 rad of the target for >3 s, despite the urging of the investigators. The required posture of the forearm was monitored visually by one of the investigators for the duration of the position task.

The third and fourth sessions were used to measure the recruitment and derecruitment thresholds and discharge characteristics of single motor units during the force or position task performed for 90% of the time to failure for the position task (Fig. 1). The recruitment threshold of a single motor unit was defined as the force at which the discharge rate of the isolated motor unit was minimal and repetitive (Spiegel et al. 1996). The recruitment threshold was measured during three to five ramp contractions that consisted of gradually increasing the force exerted by the elbow flexor muscles to a level that was sufficient to observe a repetitive discharge of an isolated motor unit. Subjects were given visual feedback of the force exerted by the wrist during the ramp contractions. The isolated motor unit was characterized by the shape and amplitude of its action potential, which was monitored on an oscilloscope during the ramp contractions. The target force for the ramp contractions was set at twice the estimated force recruitment threshold force, and subjects were asked to reach this level in ∼6 s and to decrease the force similarly. This approach made it possible to analyze the discharge of other motor units off-line.

FIG. 1.

Experimental protocol and motor unit recordings. A: representative data for averaged EMG (aEMG) and force for 1 subject when performing the position task to 90% of the time to failure. The maximal voluntary contractions (MVCs) have been removed to improve the clarity of the figure. The arrows on the force trace indicate the recruitment threshold for 1 motor unit from the short head of the biceps brachii with an initial recruitment threshold of 56% MVC, which decreased to 31% MVC at the end of the sustained contraction. B–D: expended time scale during the ramp contractions showing the force (top trace), subcutaneous EMG recording with the 1st 5 action potentials of the motor unit indicated by the arrows in A, and instantaneous discharge rate (bottom trace) before the sustained contraction (B), and at 50 (C) and 90% (D) of the time to failure. Note the decline in motor unit recruitment threshold and instantaneous discharge rate for the 2 1st interspike intervals, whereas the slope of force development remained similar between the different time points.

The MVC involved a gradual increase in torque from zero to maximum over 3 s and sustaining the maximal torque for ∼3 s with the elbow flexor muscles while the subjects were verbally encourage to achieve maximal force. At least two trials were performed, with subjects resting for 90–120 s between trials to minimize fatigue. When the peak torques were within 5% of each other, the greater value was taken as the maximum and used as a reference for the submaximal contractions. Otherwise, additional trials were performed until the 5% criterion was achieved. In addition, single MVCs were performed with the elbow extensors and the shoulder flexors to obtain reference EMG values for triceps brachii (lateral head) and anterior deltoid activity, respectively. One of the experimenters restrained limb movement during these two supplementary MVCs.

Data analysis

Surface EMG signals were rectified and averaged (aEMG) during a 0.5-s epoch from the peak EMG during the MVCs. The EMG activity of the elbow flexor muscles, anterior deltoid, and triceps brachii muscles during the sustained submaximal contractions were rectified and averaged over the first 30 s, 15 s on either side of 25, 50, and 75% of time to failure, and the last 30 s of the contraction. The EMG activity was normalized to the peak EMG obtained during the MVC. The CV for force (SD divided by the mean) was calculated for similar epochs. The relation between the aEMG activity of the elbow flexor muscles (% MVC) and time (s) was calculated for each subject. Because this relation was linear (r2 > 0.82), the slope of this linear relation was used as an index of the rate of change in aEMG during the sustained contractions.

Action potentials discharged by single motor units in biceps brachii during the ramp contractions were discriminated using a computerized, spike-sorting algorithm (Spike2, version 5.02, Cambridge Electronic Design), which identified the potentials belonging to a single motor unit based on waveform amplitude, duration, and shape. Thereafter, spike-by-spike discrimination was performed visually, and interspike intervals (ISIs) <40 or >200 ms were discarded. The first and last four successive intervals that met these criteria during the ascending and descending phases of the ramp contractions were used to calculate the recruitment threshold, mean discharge rate, CV for interspike interval, and durations of the first and last four interspike intervals (Klass et al. 2008a; Van Cutsem and Duchateau 2005; Van Cutsem et al. 1998). The interspike intervals were averaged for each interval across the three to five ramp contractions performed before, during, and after the sustained contraction. Newly recruited motor units that were identified during the ramp contractions performed at 50 and 90% of the time to failure were similarly characterized. These additional motor units, which were not identified before task initiation, were considered newly recruited if they discharged action potentials for at least two of the three to five ramp contractions. However, the newly recruited motor units were so few (3 for the force task and 6 for the position task) that they could not be characterized by statistical analysis.

Statistics

Paired Student's t-test were used to compare initial values of the MVC torque, target force, and time to task failure. Changes in MVC torque were compared with a two-way ANOVA with repeated measures (task × time). Changes in aEMG activity during MVCs and the fatiguing contractions were compared using a three-way ANOVA with repeated measures (task × muscle × time). Linear slope analysis for the rate of increase in aEMG was compared using a paired Student's t-test. The changes in recruitment threshold, mean discharge rateand, discharge rate variability were analyzed by a three-way ANOVA (task × initial recruitment threshold × time) with repeated measures on time. Changes in the duration of the first and last four interspike intervals were tested by a four-way ANOVA (task × initial recruitment threshold × time × interspike interval). The initial recruitment threshold factor was used to distinguish between motor units with an initial recruitment threshold below the target force and those with an initial recruitment threshold above the target force.

When a significant main effect was found with an ANOVA, a Tukey post hoc test was used to identify the significant differences among the selected means. The level of statistical significance was set at P < 0.05 for all comparisons. Unless otherwise specified, values are expressed by mean ± SD in the text and means ± SE in the figures.

RESULTS

The four experimental sessions provided data on the relative times to failure and accompanying adjustments when the force and position tasks were sustained to failure and changes in motor unit activity when the two tasks were performed for the same absolute duration.

Force and position tasks to failure

Twenty of the 24 subjects who participated in the study were included in the statistical analysis. Data from four subjects were excluded because of methodological issues in recording the surface EMG or motor unit activity. Despite a similar target force during the force (32.4 ± 10.9 N) and position (34.6 ± 12.8 N) tasks (paired t-test, P = 0.16) and comparable criteria for task termination, the time to failure was briefer for the position task (1,225 ± 666 s, range: 518–3,496 s) than for the force task (1,904 ± 944 s, range: 626–3,661 s; paired t-test, P = 0.0002). Immediately after task failure, however, MVC torque was reduced to a similar level for the force and position tasks (69.0 ± 11.8 and 67.3 ± 9.4% of initial, respectively; P = 0.60). Similarly, aEMG of the elbow flexor muscles during the MVC was reduced to a similar level after the force and position tasks (77.1 ± 23.6 and 75.9 ± 21.6% of initial, respectively; time main effect, P = 0.0001).

Because there was no difference in the changes in aEMG activity among the three elbow flexor muscles during each sustained contraction, the results were collapsed across the three muscles. At the beginning of the fatiguing contraction, aEMG activity for the three elbow flexor muscles was similar for the force (10.3 ± 4.6% MVC) and position tasks (11.7 ± 7.3% MVC; P = 0.26) and increased similarly (time main effect, P = 0.0001) to 21.7 ± 10.7 and 25.1 ± 9.5% MVC, respectively, at task failure (Fig. 2A). However, the rate of increase in aEMG was greater (paired t-test, P = 0.0001) during the position task (0.018 ± 0.017% MVC/s) than during the force task (0.006 ± 0.008% MVC/s). The aEMG activity in triceps brachii began at 9.3 ± 7.6 and 10.0 ± 7.8% MVC for the force and position task, respectively, and increased significantly throughout the task (time main effect, P = 0.004) to a final value of 14.9 ± 12.4% MVC with no difference between the two tasks (task main effect, P = 0.45). The aEMG for anterior deltoid did not differ between the two tasks at the beginning of the two tasks (task × time, P = 0.68; collapsed data: 15.1 ± 7.1% MVC) and did not increase during either task. The CV for force increased during both tasks to reach 385 ± 252 (force task) and 580 ± 332% (position task) of the initial values (Fig. 2B), with a greater increase for the position task than the force task (Tukey post hoc test, P = 0.0001).

FIG. 2.

Change in EMG amplitude and force fluctuations during the 2 fatiguing contractions. The average rectified aEMG for the elbow flexor muscles (collapsed across the 3 elbow flexor muscles and expressed as means ± SE; A) and the average CV for force (B) during the force (•) and position tasks (○) sustained to failure. The values have been averaged over 30-s intervals for each task at 5 time points that correspond to the relative times at the start and 25, 50, 75, and 100% of time to task failure. *P < 0.05 compared with initial values; †P < 0.05 compared with force task values at the same relative time point.

Motor unit recording during the force and the position tasks

Figure 3 shows the change in EMG activity for the elbow flexor muscles when the force and the position tasks were sustained to 90% of the time to failure for the position task. Consistent with the results in Fig. 2, the aEMG activity increased to a greater value at the end of the position task compared with the force task (task × time, P = 0.005). The aEMG value for the elbow flexor muscles at the 90% time was significantly greater for the position task compared with the force task (Tukey post hoc test, P = 0.03), even though the target torque and absolute time were similar for both tasks. These changes were accompanied by a significantly different decrease in MVC torque (task × time, P = 0.0002). At the end of the position task, the MVC torque was decreased to a greater extent (70.2 ± 8.7% of initial value) compared with force task (85.1 ± 8.4% of initial value; P = 0.0002), whereas the MVC force was decreased to the same extent for the force (87.8 ± 9.2% initial; P = 0.001) and the position tasks (83.5 ± 8.1% initial; P = 0.0001) at the 50% time.

FIG. 3.

Change in EMG amplitude during the 2 contractions sustained to 90% of the time to failure for the position task. The average rectified aEMG for the elbow flexor muscles (collapsed across the 3 elbow flexor muscles and expressed as means ± SE) during the force (•) and position tasks (○) sustained to 90% of the time to failure for position task. The values have been averaged over 30-s intervals for each task at four time points that correspond to the relative times at the start and 25, 75, and 90% of time to task failure. The values for the 50% time point have been averaged over 15-s intervals at the end of the 1st part of the contraction and at the beginning of the 2nd part. *P < 0.05 compared with initial values; †P < 0.05 compared with force task values at the same relative time point.

Single motor units were isolated during the force and position tasks sustained to 90% of the time to failure for the position task and changes in recruitment threshold and discharge characteristics during ramp contractions were determined. A total of 59 motor units were identified during the ramp contractions that preceded the sustained contraction and the ramp contractions at the 50 and 90% times of the sustained contraction. Thirty motor units were isolated during the force task (13 in the long head of biceps brachii) and 29 during the position task (13 in the long head of biceps brachii). Because no difference was observed for the motor unit characteristics in the long and short heads of the biceps brachii, these data were collapsed across location. Before the sustained contraction, the recruitment threshold was 25.7 ± 15.3% (range: 3–49% MVC) for the motor units in the force task and 23.9 ± 15.3% MVC (range: 6–56% MVC) for motor units in the position task. These values did not differ between task (unpaired t-test, P = 0.97). The motor units recorded in each task were assigned to two groups (Table 1): those with an initial recruitment threshold below the target force (14 and 13 motor units for the force and position tasks, respectively) and those with recruitment thresholds above the target force (16 motor units for both force and position tasks). The rate of torque increase during the ramp contractions was similar during the force and position tasks performed before the sustained contraction (6.4 ± 3.2 and 7.1 ± 4.9% MVC/s, respectively) and did not change across the experiment (task × time, P = 0.56) or during the descending phase of the ramp contractions. It was only possible, however, to determine the force and discharge rate at derecruitment of 15 (7 with a recruitment threshold below the target force) and 14 motor units (6 with a recruitment threshold below the target force) in the force and position tasks, respectively (Table 1).

View this table:
TABLE 1.

Motor unit characteristics before the sustained contraction

Recruitment and derecruitment thresholds

Most motor units (25 of 29) showed a decrease in recruitment threshold during the position task, whether or not the initial value for recruitment threshold was below the target force, whereas the change in recruitment threshold was less pronounced (19 of 30 motor units) during the force task (Fig. 4). Recruitment threshold decreased during the position task from 9.8 to 5.1% MVC force (Tukey post hoc test, P = 0.03) for motor units with an initial recruitment threshold below the target force (Fig. 5, bottom left) and from 32.0 to 21.9% MVC (Tukey post hoc test, P = 0.0003) for motor units with an initial recruitment threshold above the target force (Fig. 5, top left). In contrast, no significant changes were observed in recruitment threshold during the force task (Fig. 5, left). Among the motor units with recruitment thresholds above the target force (16 for both tasks), three and seven were activated after the beginning of the force and position tasks, respectively.

FIG. 4.

Change in recruitment threshold of individual motor units during the force and position tasks. Recruitment threshold of single motor units in biceps brachii with an initial recruitment threshold above the target force of 15% MVC force (A and B) and with an initial recruitment threshold below the target force (C and D) during the force (left) and position tasks (right).

FIG. 5.

Change in recruitment threshold during the force and position tasks. Changes in average recruitment (left) and derecruitment threshold (right) of single motor units with an initial recruitment threshold either above (top) or below (bottom) the target force during the force (•) and position tasks (○) before and at 50 and 90% of the time to failure for the position task (means ± SE). *P < 0.05 compared with initial values.

The derecruitment threshold also declined during the sustained contraction without a significant difference between tasks (time main effect, P = 0.0002; Fig. 5, right). The absence of significant task difference when the derecruitment threshold was expressed relative to the MVC force likely rises from the few number of motor units available for the measurement of the derecruitment threshold. When changes were expressed relative to the initial value, however, a task main effect (P = 0.0004) indicated a greater decline in the derecruitment threshold for the position task than for the force task (Fig. 5, right). At the end of the sustained contraction, the derecruitment threshold was reduced by 11.2 ± 29.2 and 27.4 ± 25.2% for the force and position tasks, respectively.

Motor unit discharge rate

The mean discharge rate for the first four interspike intervals during the ramp contractions decreased progressively during the position task but not during the force task (task × time, P = 0.02; Fig. 6, top). At the end of the position task, the mean discharge rate was significantly depressed to 84.5 ± 14.1% initial (Tukey post hoc test, P = 0.03) and differed from the value observed at the end of the force task (Tukey post hoc test, P = 0.02) for which there was no significant change (Table 1). The decline in the mean discharge rate for the position task was accompanied by an increase in the discharge variability as indicated by the CV for ISI (task × time, P = 0.02; Fig. 6, bottom). The CV for ISI increased by 40.1 ± 12.8% at the end of the position task (Tukey post hoc test, P = 0.005), whereas it did not change significantly for the force task (12.9 ± 10.7%, P = 0.81). The analysis of individual ISIs for the first four interspike intervals showed a prolongation of the first (36.4 ± 36.0%; Tukey post hoc test, P = 0.0002) and second (18.8 ± 32.0%; Tukey post hoc test, P = 0.0002) ISIs at the end of the position task (Table 2). The ISIs did not change during the force task. There were no significant changes at derecruitment in any of these parameters during either task (Fig. 6, right).

FIG. 6.

Change in mean discharge rate and CV for interspike interval at recruitment and derecruitment for the force and position tasks sustained to 90% of the time to failure for the position task. Top: the average mean discharge rate (means ± SE) for the 1st 4 interspike intervals at recruitment (left) and the last 4 interspike intervals at derecruitment (right) for the force (•) and position tasks (○) sustained to 90% of the time to failure for position task. Bottom: the average CV for interspike interval (means ± SE) for the 1st 4 interspike intervals at recruitment (left) and the last 4 interspike intervals at derecruitment (right) for the force (•) and position tasks (○) sustained to 90% of the time to failure for position task. *P < 0.05 compared with initial values; †P < 0.05 compared with force task values at the same relative time point.

View this table:
TABLE 2.

Duration (ms) of the 1st 4 ISIs at recruitment during the force and position tasks

There were significant associations between changes in the recruitment threshold and in discharge rate characteristics (Fig. 7). When considering all motor units during the force and position tasks, changes in recruitment threshold were significantly correlated with changes in mean discharge rate (P = 0.005), discharge variability (P = 0.008), and duration of the first interspike interval at recruitment (P = 0.0001). The latter relation was also significant when only the motor units isolated during the position task were included in the regression (y = 0.66x + 27.7; r = 0.54; P = 0.002).

FIG. 7.

Relations between the relative changes in recruitment threshold and in discharge characteristics of motor units at the end of the force and position tasks sustained to 90% of the time to failure for the position task. The change in recruitment threshold of motor units isolated during the force (•) and the position (○) tasks was associated with changes in the mean discharge rate (P = 0.02; A), discharge variability (P = 0.02; B), and duration of the 1st interspike interval (P = 0.008; C). Changes are expressed as a percentage of the initial values.

DISCUSSION

There were two key findings in this study: 1) a decline in recruitment threshold of motor units during the position task, but not the force task, whether or not the motor units were activated during the sustained contraction; and 2) the decline in recruitment threshold for the position task was accompanied by a prolongation of the two first interspike intervals and increased discharge variability at recruitment. These results extend previous work on the influence of load type on the rate at which neuromuscular adjustments occur during a sustained, submaximal contraction by showing that the changes in motor unit recruitment and rate coding can involve much of the motor unit pool.

Motor unit recruitment threshold

Consistent with previous work, the time to task failure was briefer for the position task than the force task when performed with the elbow flexor muscles at different intensities and in various limb postures (Hunter et al. 2002; Klass et al. 2008b; Maluf et al. 2005; Rudroff et al. 2005, 2007). In this study, the target torque was 15% MVC and the time to task failure was longer for both tasks compared with those reported by Hunter et al. (2002) at a similar contraction intensity. This difference in time to failure for the two studies is likely attributable to the influence of arm posture on the demands placed on shoulder muscles (Rudroff et al. 2007). When the subjects sustained the two tasks to failure in this study, the MVC torque and the associated aEMG were reduced to the same extent at the end of both tasks, indicating a similar decline in the force capacity at the end of the sustained contractions (Gandevia et al. 1996; Søgaard et al. 2006; Zijdewind et al. 1998). The similar decrease in force capacity at failure, however, was achieved more rapidly during the position task. Consistent with the assumption that the recruitment threshold of a motor unit mainly depends on the force exerted by the motor units previously activated (Farina et al. 2009), a faster decline in the force contributed by the active motor units requires a faster recruitment of the motor unit pool to sustain the target force. Accordingly, the recruitment and derecruitment thresholds of motor units decreased more rapidly during the position task with no significant changes during the force task (Fig. 4).

A decline in recruitment threshold was also observed for motor units that were not activated during the sustained contraction. Although the upper limit of motor unit recruitment in biceps brachii is ∼88% MVC, ≥80% of the motor unit pool is recruited below 60% MVC force (see Fig. 1B in Kukulka and Clamann 1981), which suggests that the decrease in recruitment threshold during the position task influenced most of the motor units in the pool as the decrease was observed for motor units with recruitment thresholds from 6 to 56% MVC force. The more rapid recruitment of motor units during the position task is consistent with the greater rate of increase in the CV for force during this task (Fig. 2B) (Ebenbichler et al. 2000; Löscher et al. 1996).

The faster decrease in recruitment threshold during the position task in this study differs from the similar decline observed in the two tasks by Mottram et al. (2005). Such a divergence is likely attributable to differences in the experimental protocol. Mottram and colleagues adjusted the target force and the duration of the contraction relative to the recruitment threshold of the motor unit, whereas an absolute target force and a relative duration was used in this study. As a consequence, the average duration of the contraction was 161 s for Mottram et al. and 1,103 s in this study. Furthermore, Mottram et al. stated that the target force was set 3.5% above the recruitment threshold of the motor unit, whereas the target force in this study was on average 5.2% greater than the recruitment threshold of the motor unit. Moreover, the forearm was placed in a neutral posture (midway between supination and pronation) by Mottram et al. and was supinated in this study. The significance of forearm posture is that it influences the modulation of the inhibitory feedback received by the biceps brachii motor neurons from muscle afferents in the brachioradialis (Barry et al. 2008). The strength of this inhibitory pathway is stronger when the forearm is in a neutral posture compared with a supinated posture (Barry et al. 2008) and may have reduced the specific influence of the load type on the motor unit activity during the position task.

Despite the difference between the two studies in the change in recruitment threshold during the two tasks, there were some similarities in the results. Mottram and colleagues reported a greater number of newly recruited motor units during the position task compared with the force task, which is consistent with our observation of a greater decrease in the recruitment threshold of motor units not activated during at the start of the position task (Fig. 4B). Furthermore, the magnitude of the change in motor unit discharge rate is similar for the two studies.

Motor unit discharge rate

The decrease in recruitment threshold of motor units during the position task was associated with a decrease in mean discharge rate, an increase in discharge variability, and a prolongation of the first two ISIs at recruitment (Table 2; Fig. 6), whereas there were no significant changes during the force task. There are at least three potential explanations for the change in motor unit discharge during the position task. First, the repetitive discharge of action potentials by a motor neuron can induce a late spike frequency adaptation in motor neurons (Kernell and Monster 1982a,b). However, it is unlikely that late adaptation can explain the decline in discharge rate for motor units activated during the position task but not for those activated during the force because the initial discharge rate was similar between the two tasks during the ramp contractions that preceded the sustained contraction (Table 1) as well as the discharge rate at the beginning of the force and position tasks (Mottram et al. 2005). Moreover, late spike frequency adaptation cannot account for changes in motor units that were not activated during the task. In addition, late adaptation is more pronounced at high discharge rates (Kernell and Monster 1982a), and no association was observed between the initial discharge rate of a motor unit and the subsequent decline during either sustained contraction.

Second, an increase in synaptic activity may reduce the input resistance of the motor neurons and as a consequence decrease discharge rate (Kernell 2006). However, only motor neurons activated during the sustained contraction would experience such an adjustment. Because motor units that were not activated during the sustained contraction also exhibited a decrease in the discharge rate at recruitment, it seems unlikely that the decrease in discharge rate was caused by a change in intrinsic properties, at least for the high-threshold motor units. Rather, these results are consistent with the third possibility that the change in motor unit discharge rate during the position task was caused by a differential synaptic input to the motor neuron pool during the two tasks rather than a change in the intrinsic properties of the motor units.

The increase in the CV for ISI at recruitment during the position task suggests a more rapid increase in synaptic noise in this task (Stein et al. 2005). The variability in motor unit discharge times is caused by greater fluctuations in membrane potential because of an increase in excitatory and inhibitory input received by the motor neurons during the contraction (Berg et al. 2007). The synaptic inputs that can influence a motor neuron pool during a fatiguing contraction includes a reduction in input from Ia afferents (Macefield et al. 1991) and an increase in excitation of group III–IV afferents (Windhorst et al. 1997). Consistent with a differential synaptic input during the two fatiguing contractions, Klass et al. (2008b) found that the Hoffmann reflex (H reflex) exhibited a greater and faster decrease in amplitude during the position task than during the force task, whereas no difference was observed between the two tasks in the rate of increase in the amplitude of the motor-evoked potential elicited by transcranial magnetic stimulation. These results argue for a modulation of the H reflex at a presynaptic level, likely involving presynaptic inhibition of Ia afferent input onto motor neurons (Baudry et al. 2008a; Duchateau et al. 2002; Klass et al. 2008b). The faster motor unit recruitment during the position task may involve a more rapid accumulation of metabolites resulting in a greater increase in feedback transmitted by group III and IV afferents during the position task (Kaufman and Rybicki 1987; Kaufman et al. 1983). Because group III-IV muscle afferents can act through the Ia presynaptic inhibitory pathway (Brunetti et al. 2003; Pettorossi et al. 1999), this may have contributed to an increase in presynaptic inhibition of Ia afferents during the position task. This interpretation is consistent with the observed decrease in discharge rate during the ramp contractions that were performed ∼30 s after the end of the task, which is an insufficient duration for a complete removal of muscle metabolites (Bogdanis et al. 1995, 1996).

Alternatively, homosynaptic depression at the presynaptic terminal has also been reported to reduce the amplitude of the H reflex after previous activation of the monosynaptic reflex pathway (Hultborn and Nielsen 1998). There may have been more activity in the Ia afferent pathway during the position task as associated reflex responses are augmented when an individual maintains limb position against a compliant load compared with pushing a rigid restraint (Akazawa et al. 1983; Baudry et al. 2009; Maluf et al. 2007). As a consequence, significantly larger amounts of homosynaptic depression may have also contributed to the decrease in motor unit discharge rate during the position task. Regardless of the exact mechanism, the greater change in discharge rate coupled with the decrease in recruitment threshold during the position task underscores significant differences in the synaptic input received by the motor neuron pool during the two tasks, and this contributed to the briefer time to failure for the position task. Moreover, the decrease in discharge rate of the motor units that were not activated during the position task indicates that the alteration in the synaptic input during this task (e.g., decrease in Ia afferent input) was not limited to motor units that experienced repetitive activity but rather was distributed to much of the motor neuron pool, suggesting a nonselectivity of such changes to active motor units.

In conclusion, results obtained in this study are consistent with a differential modulation of motor unit recruitment and rate coding during sustained, submaximal isometric contractions that differ in the type of load encountered by the elbow flexor muscles. Moreover, the observation that changes occurred in much of the motor unit pool (recruitment threshold ranged from 6 to 56% MVC) and for motor units that were not activated during the task argues for a common modulation of the motor unit pool during the sustained contractions, with a faster time course of change during the position task. By identifying the physiological adjustments that cause a faster decrease in net torque during the position task, these results have implications for the ergonomic design of workstations and in the prescription of exercises for rehabilitation programs.

GRANTS

This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-043275 to R. M. Enoka.

REFERENCES

  1. Akazawa et al. 1983.
  2. Barry et al. 2008.
  3. Baudry et al. 2008a.
  4. Baudry et al. 2009.
  5. Baudry et al. 2008b.
  6. Berg et al. 2007.
  7. Bogdanis et al. 1996.
  8. Bogdanis et al. 1995.
  9. Brunetti et al. 2003.
  10. Carpentier et al. 2001.
  11. Christova and Kossev 2001.
  12. Desmedt and Godaux 1978.
  13. Duchateau et al. 2002.
  14. Ebenbichler et al. 2000.
  15. Enoka and Duchateau 2008.
  16. Enoka et al. 1988.
  17. Enoka et al. 1989.
  18. Enoka and Stuart 1992.
  19. Farina et al. 2009.
  20. Gandevia 2001.
  21. Gandevia et al. 1996.
  22. Garland et al. 1994.
  23. Henneman 1957.
  24. Hultborn and Nielsen 1998.
  25. Hunter et al. 2002.
  26. Kaufman et al. 1983.
  27. Kaufman and Rybicki 1987.
  28. Kernell 1965.
  29. Kernell 2006.
  30. Kernell and Monster 1982a.
  31. Kernell and Monster 1982b.
  32. Klass et al. 2008a.
  33. Klass et al. 2008b.
  34. Kukulka and Clamann 1981.
  35. Löscher et al. 1996.
  36. Macefield et al. 1991.
  37. Maluf et al. 2007.
  38. Maluf and Enoka 2005.
  39. Maluf et al. 2005.
  40. Mottram et al. 2005.
  41. Pettorossi et al. 1999.
  42. Rudroff et al. 2007.
  43. Rudroff et al. 2005.
  44. Schwindt 1973.
  45. Søgaard et al. 2006.
  46. Spiegel et al. 1996.
  47. Stein et al. 2005.
  48. Van Cutsem and Duchateau 2005.
  49. Van Cutsem et al. 1998.
  50. Windhorst et al. 1997.
  51. Zijdewind et al. 1998.
View Abstract