Peripheral fatigue limits endurance exercise via a sensory feedback-mediated reduction in spinal motoneuronal output

Abstract

This study sought to determine whether afferent feedback associated with peripheral muscle fatigue inhibits central motor drive (CMD) and thereby limits endurance exercise performance. On two separate days, eight men performed constant-load, single-leg knee extensor exercise to exhaustion (85% of peak power) with each leg (Leg₁ and Leg₂). On another day, the performance test was repeated with one leg (Leg₁) and consecutively (within 10 s) with the other/contralateral leg (Leg₂-post). Exercise-induced quadriceps fatigue was assessed by reductions in potentiated quadriceps twitch-force from pre- to postexercise (ΔQ_tw,pot) in response to supramaximal magnetic femoral nerve stimulation. The output from spinal motoneurons, estimated from quadriceps electromyography (iEMG), was used to reflect changes in CMD. Rating of perceived exertion (RPE) was recorded during exercise. Time to exhaustion (∼9.3 min) and exercise-induced ΔQ_tw,pot (∼51%) were similar in Leg₁ and Leg₂ (P > 0.5). In the consecutive leg trial, endurance performance of the first leg was similar to that observed during the initial trial (∼9.3 min; P = 0.8); however, time to exhaustion of the consecutively exercising contralateral leg (Leg₂-post) was shorter than the initial Leg₂ trial (4.7 ± 0.6 vs. 9.2 ± 0.4 min; P < 0.01). Additionally, ΔQ_tw,pot following Leg₂-post was less than Leg₂ (33 ± 3 vs 52 ± 3%; P < 0.01). Although the slope of iEMG was similar during Leg₂ and Leg₂-post, end-exercise iEMG following Leg₂-post was 26% lower compared with Leg₂ (P < 0.05). Despite a similar rate of rise, RPE was consistently ∼28% higher throughout Leg₂-post vs. Leg₂ (P < 0.05). In conclusion, this study provides evidence that peripheral fatigue and associated afferent feedback limits the development of peripheral fatigue and compromises endurance exercise performance by inhibiting CMD.

numerous studies have revealed that the inability to continue high-intensity, constant-load endurance exercise (i.e., exhaustion) coincides with a specific level of peripheral locomotor muscle fatigue (3, 4, 21, 22, 44–46). Based on the evidence that voluntarily exercising humans never exceed this specific and individually different degree of peripheral fatigue, the existence of a “critical threshold” of muscle fatigue was previously proposed (3, 4). Muscle afferent fibers, which relay fatigue-related metabolic perturbations within the working limb muscles to the central nervous system (CNS) (27, 30, 57), have been suggested to play an important role in determining this critical threshold of fatigue (5, 23), which likely coincides with an individual's “sensory tolerance limit” (23). Specifically, during constant-load endurance exercise, mechano- and metabo-sensitive group III/IV muscle afferents provide inhibitory input to the CNS. This feedback limits voluntary descending drive to the primary motor cortex [i.e., central motor drive (CMD)] (52) and exercise performance, restricting the development of locomotor muscle fatigue to a critical threshold that is associated with a certain level of intramuscular metabolic perturbation (4). It should be acknowledged that this paradigm includes only one of several potential mechanisms that may account for the inability to continue high-intensity, constant-load endurance exercise and requires further study. Psychological factors (18), extreme environmental influences (6, 39, 54), and exercise-induced alterations in CNS neurotransmitter systems (36) are examples of other potential mechanisms limiting CMD and/or the output of spinal motoneurons and exercise performance.

In recent experiments designed to examine the effect of locomotor muscle afferents on the development of peripheral fatigue, pharmacological blockade was used to reduce sensory feedback during high-intensity leg cycling (2). With greatly reduced inhibitory feedback from group III/IV leg muscle afferents to the CNS, output/drive from the spinal motoneurons to locomotor muscles (as estimated from surface EMG) was, compared with placebo conditions, substantially increased during exercise. The exercising humans (i.e., the CNS) in these experiments “ignored” the sensory tolerance limit and “tolerated” the development of peripheral muscle fatigue substantially beyond their critical threshold, which was determined in placebo conditions (i.e., with intact afferent feedback) (2). Although providing strong evidence in favor of an inhibitory influence of fatigue-related sensory feedback on the output from spinal motoneurons, a caveat of this approach is that blocking group III/IV muscle afferents also impairs the cardiovascular and ventilatory responses during exercise (1, 7), which, per se, limits exercise performance (2).

In addition to reducing the inhibitory neural feedback during exercise via spinal blockade, an alternative approach to discern the role of group III/IV muscle afferents in regulating the output of spinal motoneurons and limiting endurance performance has been to elevate the inhibitory feedback mediated by these sensory neurons. In previous investigations, peripheral locomotor muscle fatigue and associated metabolic disturbances were either voluntarily (3) or electrically (22) induced immediately before a cycling performance test to increase fatigue-related inhibitory neural feedback to the CNS during the subsequent performance trial. The outcome of these experiments indicates that the higher the level of preexisting muscle fatigue and associated inhibitory feedback, the lower the output of spinal motoneurons and cycling performance during a subsequent performance trial (3, 22). Again, although providing evidence for an inhibitory effect of peripheral muscle fatigue and associated neural feedback on the output of spinal motoneurons, the observed performance limitations might, at least in part, have been due to a compromised muscle response to a given neural input (i.e., peripheral fatigue).

To circumvent previous experimental shortcomings, we induced quadriceps fatigue in one leg using voluntary dynamic single-leg knee-extensor exercise to exhaustion with the goal to raise the ensemble muscle afferent feedback during a subsequent endurance performance test with the other leg. We hypothesized that 1) fatigue-related intramuscular metabolic disturbances and associated inhibitory feedback from one leg would compromise performance of the consecutively exercised contralateral leg and 2) peripheral quadriceps fatigue, at voluntary exhaustion, of the consecutively exercised leg would be substantially below the level of fatigue observed during control trials.

METHODS

Subjects

Eight recreationally active, healthy male volunteers participated in this study (age 24 ± 1 yr, body mass 83 ± 6 kg, height 178 ± 4 cm). Seven of the participants were determined to be right-leg dominant, whereas one participant was left-leg dominant (41). Quadriceps muscle mass (26) was similar in both limbs (2.4 ± 0.1 kg; P = 0.6). The study was approved by the University of Utah and the Salt Lake City VA Medical Center Institutional Review Boards, and written, informed consent was obtained from each subject before participation.

Experimental Protocol

During four 1-h practice sessions, all participants were thoroughly familiarized with various experimental procedures, including single-leg knee-extensor exercise (KE) (8, 43). On 2 additional days, all subjects performed incremental single-leg KE until they were unable to continue the prescribed work rate (15 W + 5 W/min) to determine peak power output (W_peak) of each leg.

On 2 separate days following these initial sessions, each subject performed, in random order, constant-load, single-leg KE (60 RPM, 85% of W_peak) to task failure (cadence below 50 RPM for ≥10 s) with both legs (days 1 and 2; Fig. 1). This RPM was chosen based on our previous work with this model, indicating that a failure to maintain 50 RPM for longer than 10 s leads to voluntary termination of exercise within 30 s. Neuromuscular function of the exercising leg was assessed before and 2 min after exercise (Fig. 1). To determine whether exercise-induced quadriceps fatigue in one leg affected neuromuscular function in the other leg, all subjects performed a third trial (day 3; Fig. 1). Here, neuromuscular function of one leg was assessed before and again 2 min after constant-load exercise to task failure with the other leg. On a fourth day, all participants carried out constant-load, single-leg KE to task failure with one leg (Leg₁) and consecutively (<10 s) performed constant-load, single-leg KE, again to task failure, with the other leg (Leg₂-post) (day 4; Fig. 1). In this trial, at task failure following exercise with the first leg (Leg₁), a cuff around the upper part of the thigh of Leg₁ was inflated to 250 mmHg until the Leg₂-post exercise trial was started (day 4; Fig. 1). Neuromuscular function of Leg₂-post was assessed before and 2 min after exercise (Fig. 1). Sessions (day) 3 and 4 were in random order. The warm-up performed on each of the test days consisted of 10 min of two-leg KE at 15 W. To avoid initial peak force outputs at the beginning of each performance trial, the ergometer was initially accelerated by one of the investigators (<2 s). Subjects were instructed to strictly maintain 60 RPM throughout KE. All performance trials were separated by 48–72 h and balanced with respect to leg dominance.

Fig. 1.

Experimental design. Days 1 and 2 and days 3 and 4 were carried out in random order and separated by at least 48 h. Leg order and dominance were balanced.

Exercise Responses

At rest and throughout exercise, pulmonary gas exchange and ventilation were measured continuously using an open-circuit calorimetry system (Parvo Medics, True Max 2400, Salt Lake City, UT). Femoral blood flow (FBF) was measured at rest and during exercise using ultrasound Doppler (Logic 7, General Electric Medical Systems). Simultaneous measurements of common femoral arterial blood velocity (V_mean) and vessel diameter were performed distal to the inguinal ligament and proximal to the bifurcation of the deep and superficial femoral arteries. Using arterial diameter and V_mean, FBF was calculated as: FBF = V_meanπ(vessel diameter/2)² × 60.

Cardiac output (CO) and mean arterial pressure (MAP) were determined using a Finometer (Finapres Medical Systems, Amsterdam, The Netherlands) and the Modelflow algorithm (Beatscope version 1.1a; Finapres Medical Systems) (50). Heart rate (HR) was measured from the R-R interval of an electrocardiogram (ECG) using a three-lead arrangement. Rating of perceived exertion (RPE) was obtained at rest and every minute during exercise using Borg's modified CR10 scale (14) following the recommended instructions/verbiage for using the scale (38). Venous blood samples from an antecubital catheter were collected in trial 4, as illustrated in Fig. 1. To evaluate the effects of severe quadriceps fatigue on remote rested muscles, we measured maximal handgrip forces of the dominant arm before the performance of Leg₁ and the consecutive performance test in Leg₂-post on day 4 (Fig. 1).

Neuromuscular Function

Electromyography.

As previously described (3), quadriceps EMG was recorded from the vastus lateralis using electrodes with full-surface solid adhesive hydrogel (H59P, Tyco Healthcare Group, Mansfield, MA). The position of the EMG electrodes was marked with indelible ink to ensure placement in the same location on subsequent visits. The vastus lateralis electrodes were used to record 1) magnetically evoked muscle action potentials (M-waves, peak-to-peak amplitude) to evaluate changes in membrane excitability and 2) EMG throughout exercise to estimate the output of spinal motoneurons and the development of peripheral quadriceps fatigue. Membrane excitability was maintained from pre- to postexercise in all trials as indicated by unchanged peak-to-peak M-wave amplitudes. This suggests that the observed changes in potentiated twitch force (Q_tw,pot; see below) were predominantly due to changes within the quadriceps and that electrical transmission failure or reduced sarcolemma excitability can be excluded. During KE performance trials, EMG signal corresponding to vastus lateralis muscle contractions were recorded for later analysis. The raw EMG signal was digitized and recorded at 1.5 kHz using adhesive silver electrodes placed over the muscle belly ∼4 cm apart. The raw EMG signal was filtered with a band-pass filter (low-pass cut-off frequency of 15 Hz, high-pass cut-off frequency of 650 Hz), and muscle activity was automatically located (AcqKnowledge, Biopac Systems, Goleta, CA). For this, the analysis program calculated the standard deviation (SD) of the baseline noise (i.e., electrical signal between the bursts). The onset of a burst was identified as the point at which the EMG signal rose to a value >2.5 SD above baseline. The same approach was utilized to identify the offset of a burst. Accurate identification of EMG activity was verified by visual inspection. Integrated EMG (iEMG) for each contraction throughout the protocol was calculated using the Acknoweldge software.

Magnetic nerve stimulation.

A detailed description of the assessment of muscle function with magnetic nerve stimulation can be found in prior publications by our group utilizing this technique (4). Briefly, subjects rested on an adjustable chair with the right or left thigh lying in a preformed holder, with knee joint angle set at 90° of flexion and arms folded across the chest. A magnetic stimulator (Magstim 200; The Magstim, Wales, UK) connected to a double 70-mm branding iron coil was used to stimulate the femoral nerve. The position of the coil was marked with indelible ink to ensure placement in the same location during all visits. The evoked quadriceps twitch force was obtained from a calibrated load cell (model MLP-300; Transducer Techniques, Temecula, CA) connected to a noncompliant strap, which was placed around the subject's right or left leg just superior to the ankle malleoli. Unpotentiated quadriceps single twitch forces (Q_tw) were obtained every 30 s at 50, 60, 70, 80, 85, 90, 95, and 100% of maximal stimulator output. The increment in Q_tw and M-wave amplitude from 90 to 95% of the stimulator output was 0.6 ± 0.4% (P = 0.14) and 1.4 ± 0.5% (P = 0.13), respectively; the increment in Q_tw and M-wave amplitude from 95 to 100% of the stimulator output was 0.4 ± 0.2% (P = 0.36) and −0.1 ± 0.6% (P = 0.96), respectively. A plateau in baseline Q_tw and M-wave amplitudes with increasing stimulus intensities was observed in every subject. The stimulator output that produced the first maximal response was 90%. The stimulator output was set to 100% in every subject and in all trials. Q_tw,pot has been documented to be more sensitive for detecting fatigue than the nonpotentiated twitch (28). Accordingly, we measured Q_tw,pot 5 s after a 5-s isometric maximal voluntary contraction (MVC) of the quadriceps and repeated this procedure six times. The interval between the MVCs was 35 s. As reported in a previous study (28), the degree of potentiation was slightly smaller after the first and, to a lesser extent, after the second MVC; therefore, the first two Q_tw,pot measurements were discarded. This assessment procedure was performed 20 min before exercise and 2 min after exercise. Peak force, maximal rate of force development (MRFD), and maximal relaxation rate (MRR) were analyzed for all Q_tw,pot (47). Voluntary activation of the quadriceps during the MVCs was assessed using a superimposed twitch technique (37). Briefly, the force produced during a single twitch superimposed on the MVC was compared with the force produced by the potentiated single twitch delivered 5 s afterward.

Blood Analysis

Venous blood samples from an antecubital catheter were drawn into glass tubes containing EDTA during trial 4 (Fig. 1). The concentration of branched-chain amino acids (BCAA; valine, isoleucine, and leucine) was determined utilizing an amino acid assay kit (ab83374, Abcam Cambridge, MA). The concentration of free tryptophan (TRP) was measured using a modification of the spectrofluorometric method (9).

Statistical Analysis

To determine the effects of increased afferent feedback (condition: two levels, with and without fatigue in the contralateral limb) on exercise-induced changes in various physiological parameters over time (time: five levels, five time points), we employed a two-way (2 × 5) repeated-measures ANOVA. To determine the effects of increased afferent feedback (condition: two levels, with and without fatigue in the contralateral limb) on exercise-induced changes in quadriceps muscle function (time: four time points, pre- and postexercise in each condition), we employed a two-way (2 × 4) repeated-measures ANOVA. Post hoc analysis was performed using Tukey's post hoc test of honestly significant difference. Mauchly's test of sphericity was run to determine whether the assumption for homogeneity of variance was met for each variable. If there was a significant difference (P < 0.05) in the variance of the differences across time in each condition for any measured variable and therefore the assumption of normal distribution was violated, the Greenhouse-Geisser Epsilon correction was applied to the degrees of freedom. Statistical significance was set at P < 0.05. Data are presented as means ± SE.

RESULTS

Dominant vs. Non-Dominant Leg: Endurance Capacity and Quadriceps Fatigability (Days 1 and 2)

Both whole body peak oxygen uptake and W_peak were similar for the dominant and the nondominant leg [1.7 ± 0.2 vs. 1.6 ± 0.2 l/min (P = 0.82) and 61 ± 6 vs. 58 ± 4 W (P = 0.23), respectively] (Fig. 1; days 1 and 2). Time to task failure at 85% of W_peak was not different between the two legs (P = 0.79; Table 1). Immediately after exercise with either leg, quadriceps muscle function was markedly decreased from preexercise baseline (P < 0.001). The degree of exercise-induced peripheral fatigue was similar in the dominant and the nondominant leg (P = 0.48; Table 1).

Table 1.

Effects of constant-load, single-leg knee-extensor exercise to exhaustion on quadriceps muscle function of the dominant and nondominant leg

	Dominant Leg	Nondominant Leg
Time to task failure, min	9.2 ± 0.4	9.4 ± 0.9
Workload, W	52 ± 5	49 ± 3
Percent change from preexercise to 2 min postexercise
MVC	−30 ± 3	−29 ± 2
Qtw,pot	−52 ± 2	−50 ± 6
MRFD	−55 ± 4	−50 ± 7
MRR	−51 ± 4	−53 ± 3
Muscle activation	1 ± 2*	−1 ± 2*

Values are means ± SD (n = 8). Seven subjects were right leg dominant; one subject was left leg dominant. MVC, maximal voluntary contraction; Q_tw,pot, potentiated single twitch; MRFD, maximal rate of force development; MRR, maximal rate of relaxation. Preexercise, resting mean values for MVC, Q_tw,pot, MRFD, MRR, and voluntary muscle activation in the dominant leg were 573 ± 52 N, 170 ± 20 N, 1.8 ± 0.3 N/ms, 0.7 ± 0.1 N/ms, and 94 ± 3%, respectively. Preexercise resting values were similar in the nondominant leg.

^*Not significantly altered from preexercise.

Effect of Quadriceps Fatigue in One Leg on Muscle Function of the Rested Contralateral Leg (Day 3)

Despite the development of severe exercise-induced quadriceps fatigue in one leg (Fig. 1, day 3), quadriceps muscle function in the contralateral leg was not affected. This was evident by similar pre- vs. postexercise values for Q_tw,pot (∼170 N; P = 0.44), MVC force (∼570 N; P = 0.57), and voluntary muscle activation (∼93%; P = 0.89).

Effect of Quadriceps Fatigue on Endurance Exercise Performance and Associated Development of Peripheral Fatigue in the Consecutively Exercised Contralateral Leg (Day 4)

Exercise performance.

Compared with control exercise (i.e., day 2; time to task failure: 9.2 ± 0.4 min), endurance time to task failure was −49 ± 6% shorter (P < 0.001) when the identical exercise was performed with pre-induced quadriceps fatigue in the contralateral leg (i.e., Leg₂post trial). This reduction in endurance time to task failure was consistent in all subjects (range −33 to −75%; Fig. 2).

Fig. 2.

Individual and group mean data illustrating endurance time to task failure and peripheral quadriceps fatigue under control conditions (Leg₂, i.e., exercise performed with rested contralateral leg) and under experimental conditions (Leg₂post, i.e., exercise performed with severe quadriceps fatigue in the contralateral leg).

Contractile function.

Immediately after both Leg₂ (day 2), and Leg₂-post (day 4) exercise trials, group mean Q_tw,pot was significantly reduced from preexercise baseline [F(1,16) = 17.8; P < 0.01]. However, compared with control exercise (i.e., Leg₂ trial), the exercise-induced reduction in Q_tw,pot in Leg₂-post was attenuated (Table 2; Fig. 2). All within-twitch measurements (MRFD, MRR) were also significantly altered from baseline immediately postexercise (Table 2). Each of these exercise-induced reductions was significantly greater during the Leg₂ compared with the Leg₂-post trial (Table 2). Peak force during the 5-s MVC maneuvers was significantly decreased from baseline after all trials [F(1,16) = 8.8; P < 0.01]. Again, consistent with the other fatigue-related results, reductions in MVC force were significantly greater following the Leg₂ trial (day 2) compared with the Leg₂-post trial (day 4) (Table 2; Fig. 2). Muscle activation remained unchanged from pre- to postexercise in both trials [∼94%; F(1,14) = 0.2; P > 0.3].

Table 2.

Effect of constant-load, single-leg knee-extensor exercise to exhaustion on quadriceps muscle function of the control trial (Leg₂, day 2) vs. the consecutively exercised leg trial (Leg₂-post, day 4)

	Leg2	Leg2-post
Time to task failure, min	9.2 ± 0.4	4.7 ± 0.6 *
Workload, W	51 ± 4	51 ± 4
Percent change from pre- to 2 min postexercise
MVC	−30 ± 3	−23 ± 3 *
Qtw,pot	−52 ± 3	−33 ± 3 *
MRFD	−55 ± 4	−19 ± 9 *
MRR	−52 ± 4	−36 ± 4 *
Muscle activation	−1 ± 1†	−2 ± 1†

Values are means ± SD (n = 8). Preexercise, resting mean values for MVC, Q_tw,pot, MRFD, MRR, and voluntary muscle activation were 559 ± 55 N; 165 ± 22 N; 3.2 ± 0.4 N/ms; 0.5 ± 0.1 N/ms, and 94 ± 3%, respectively.

^*Significant difference vs. Leg₂ (P < 0.05).

^†Not significantly altered from preexercise.

Cardiorespiratory and Hemodynamic Responses

Cardiorespiratory and hemodynamic responses to constant-load KE exercise trials with and without quadriceps fatigue in the contralateral leg are illustrated in Fig. 3. Quadriceps fatigue in the contralateral leg had a significant effect on oxygen uptake [V̇o₂; F(1,14) = 56.5; P < 0.01], CO₂ uptake [V̇co₂; F(1,14) = 56.5; P < 0.01], minute ventilation [V̇e; F(1,14) = 57.2; P < 0.01], V̇e/V̇co₂ [F(1,14) = 57.5; P < 0.01], HR [F(1,14) = 43.1; P < 0.01], and cardiac output [CO; F(1,14) = 57.0; P < 0.01] during the first 3 min of exercise. However, during the final minute of exercise, these variables were similar between the Leg₂ (day 2) and Leg₂-post (day 4) trials (all P > 0.1; Fig. 3). Although MAP [F(1,14) = 57.5; P < 0.01] was, compared with Leg₂, significantly higher during the first 2 min of the Leg₂-post trial, there was no significant difference between the trials throughout the remainder of the exercise (Fig. 3F). Femoral blood flow was similar at the onset of exercise in both trials (P = 0.47). However, at minute 1, femoral blood flow was significantly lower during Leg₂-post compared with Leg₂ exercise, but this difference progressively diminished, and blood flow was similar throughout the rest of exercise (Fig. 3E).

Fig. 3.

Physiological responses to constant-load single leg knee-extensor exercise without (Leg₂) and with preexisting quadriceps fatigue in the contralateral leg (Leg₂-post). V̇o₂, oxygen consumption; V̇co₂, carbon dioxide production; V̇e, minute ventilation; MAP, mean arterial pressure; HR, heart rate; CO, cardiac output. *Significant difference vs. Leg₂ (P < 0.05).

Ratings of Perceived Exertion

The rate-of-rise of RPE was similar during the Leg₂ trial (day 2) compared with the Leg₂-post trial (day 4) (P = 0.6). However, the RPE score was, compared with Leg₂, significantly higher at the onset of the Leg₂-post trial. Indeed, the prior development of quadriceps fatigue in the contralateral leg had a significant effect on RPE during the first 3 min of exercise, with a ∼28% elevation in RPE evident throughout [F(1,14) = 44.6; P < 0.01] (Fig. 4). End-exercise RPE was similar in both trials.

Fig. 4.

Ratings of perceived exertion (RPE) during constant-load, single-leg knee extensor exercise performed without (Leg₂) and with (Leg₂-post) a preexisting level of quadriceps fatigue in the contralateral leg.

Plasma Branched-Chain Amino Acids and Tryptophan

The concentration of plasma free tryptophan (Trp) significantly increased from rest to the transition between Leg₁ and Leg2-post [Fig. 1, day 4; F(2,14) = 9.7; P = 0.02]; post-exercise Trp concentration ([Trp]) was similar to the [Trp] observed during the transition between Leg₁ and Leg2-post (P = 0.25). Branched-chain amino acids (BCAA) were similar at the three time points [F(2,14) = 0.6; P = 0.25]. Therefore the [Trp]-to-[BCAA] ratio significantly increased from rest to the transition between Leg₁ and Leg2-post [F(2,14) = 6.3; P = 0.04]; postexercise [Trp]-to-[BCAA] ratio was not different from the ratio observed during the transition between Leg₁ and Leg2-post [F(2,14) = 6.3; P = 0.84].

iEMG

iEMG rose significantly from the first minute of exercise to task failure in each of the two conditions [Fig. 5; F(1,14) of >4.3; P < 0.05]. Given the similar rate of rise during Leg₂ (day 2) and Leg₂-post (day 4) (P = 0.13), the increase in iEMG during the first 3 min of exercise was very similar in both conditions (P > 0.17). However, at end-exercise, iEMG in Leg₂ had increased to a significant extent (347 ± 39%; range 198–395%) compared with Leg₂-post (Fig. 5).

Fig. 5.

Integrated EMG (iEMG) of vastus lateralis during constant-load, single-leg knee extensor exercise performed with the same leg without (Leg₂) and with a severe degree of preexisting quadriceps fatigue in the contralateral leg (Leg₂-post). Values are normalized to the first minute of exercise. Mean values for iEMG during each muscle contraction (knee extension) were calculated and averaged over each 60-s period. Data are from five subjects. *Significant difference vs. end-exercise Leg₂ (P < 0.05).

Handgrip Force

Handgrip MVC was unchanged from before the Leg₁ performance trial to immediately after the Leg₂-post performance trial (<2 s) on day 4 (31 ± 3 and 32 ± 2 kg, respectively; P = 0.22) (Fig. 1, day 4).

DISCUSSION

This study sought to determine whether afferent feedback associated with peripheral muscle fatigue and the affiliated intramuscular metabolic disturbance inhibits spinal motoneuronal output and thereby limits endurance exercise performance. Pre-induced quadriceps fatigue in one leg reduced endurance time to exhaustion of the consecutively exercised contralateral leg by ∼49%. Since the circulatory and ventilatory responses during such small muscle mass exercise are well within the respective maximal capacities, the impact of muscle afferents on endurance performance was likely independent of their well known role in regulating, and potentially limiting, peripheral hemodynamic responses during exercise. Therefore, the present data document the limiting effect of peripheral muscle fatigue and associated afferent feedback on endurance exercise performance, apparently achieved by restricting the output of spinal motoneurons to the working skeletal muscle.

Evidence of Increased Muscle Afferent Feedback in the Consecutive Leg Exercise Trial (Leg₂-Post)

An important premise of the present study was that the development of quadriceps fatigue in one leg would increase the ensemble input of fatigue-sensitive group III/IV muscle afferents to the CNS during the consecutive endurance test of the contralateral leg. This experimental design allowed the evaluation of exercise performance of a single leg, whereas afferent feedback to the CNS arose from two legs. Although the direct assessment of afferent feedback in humans is currently not possible, several lines of indirect evidence suggest that this was, in fact, likely achieved. First, direct nerve recordings from group IV muscle afferents following electrically induced muscle fatigue in anesthetized animals indicate that the sensory discharge remained elevated above baseline for up to 15 min after fatiguing muscle contractions had stopped (19, 20, 25). Second, it has previously been documented that experimentally augmenting lower limb muscle afferent activity during one-leg KE, achieved by postexercise cuff occlusion of the contralateral leg, causes a greater exercise pressor reflex compared with control conditions (no prior exercise and no cuff occlusion in other leg) (16, 49). The present findings, characterized by an elevated exercise pressor and ventilatory response during Leg₂-post (Fig. 3), conform to these earlier observations. Finally, RPE is modulated by afferent feedback from fatiguing muscles (1, 2). In the present study, RPE scores were consistently higher during the first 3 min of the Leg₂-post trial despite very similar levels of iEMG compared with those recorded in the Leg₂ trial (Figs. 4 and 5). Taken together, these prior studies and current observations support the premise that, due to the pre-induced fatigue in Leg₁, the ensemble lower limb muscle afferent feedback was greater during the Leg₂-post compared with the Leg₂ trial, even though the exercise challenge (work rate) was similar. The achievement of this somewhat unique experimental paradigm is an important foundation from which to interpret the results of the present study.

Muscle Afferent Feedback, Central Motor Drive, and Rate of Perceived Exertion

Although many recent studies have attempted to assess the inhibitory effect of muscle afferents on CMD/the output of spinal motoneurons (2, 24, 48, 51, 55), each is confounded, to some degree, by issues such as an attenuated cardiopulmonary response to exercise (afferent blocking studies) and changes in peripheral muscle function (prefatigue studies). In contrast, due to the somewhat innovative experimental design utilized here, the present study was not burdened by these caveats. Specifically, this study provides evidence of the inhibitory effect of muscle afferents on the output of spinal motoneurons, which resulted in an ∼49% reduction in endurance performance when consecutive leg exercise in the contralateral leg was superimposed upon the already fatigued leg.

This study is not the first to document a “cross-over” effect from an exhausted muscle on one side of the body to a rested homologous contralateral muscle. Indeed, utilizing short-duration maximal isometric muscle contractions, several studies (32, 42, 55, 58) have documented a significant reduction in CMD/the output of spinal motoneurons to the contralateral rested muscle. With the exception of one study (32), these changes were, in contrast to our findings, insufficient to impair exercise performance. This discrepancy between prior observations and the present results might be explained, at least in part, by gender differences, differences in the exercise modality, differences in the muscle mass utilized and thus the total amount of group III/IV afferent feedback, and/or the use of flexor vs. extensor muscles between which responsiveness has been documented to be differentially affected by muscle afferents (33).

The exact mechanism by which group III/IV muscle afferent feedback impairs the output of spinal motoneurons and exercise performance is unclear; however, at least two potential mechanisms have previously been suggested: first, a group III/IV-mediated inhibition of the motor pathway from the motor cortex to the contracting muscles; and second, a group III/IV-mediated impairment of voluntary descending drive occurring upstream from the motor cortex (53). Regarding the former, based on studies of ischemia, utilizing maximal isometric single-arm muscle exercise, group III/IV-mediated afferent feedback (evoked by postexercise muscle occlusion) has been documented to depress the responsiveness of the motoneuron pool innervating extensor muscles, which might, at least in part, account for impaired output of spinal motoneurons and performance (33). However, this hypothesis is somewhat contradicted by experiments from the same group when it was revealed that group III/IV afferent feedback evoked by hypertonic saline infusion facilitates the responsiveness of flexor and extensor motoneurons while inhibiting motor cortical cells. The net effect was unchanged excitability of the motor pathway from the motor cortex to the contracting muscles (34). Further conflict in this area is provided by the findings from postexercise muscle occlusion studies, which document a clear dissociation of exercise-induced alterations in the responsiveness of motoneurons and motor cortical cells from group III/IV muscle afferent firing and voluntary muscle activation, at least in flexor muscles (24, 52). Based on these opposing reports, it remains unclear whether group III/IV muscle afferent feedback depressed the pathway from the motor cortex to the knee extensors in the present study and thereby potentially contributed to the compromised output of spinal motoneurons and exercise performance in Leg₂-post. These uncertainties leave us with the previously proposed mechanism that the group III/IV-related impairment in the output of spinal motoneurons and performance was mediated upstream from the motor cortex (53).

It has been theorized, in a recent viewpoint article, that RPE during exercise is independent of muscle afferent feedback and is exclusively determined by the corollary discharge associated with CMD (31). Of note, the present data collected across the Leg₁ and Leg₂-post trials fail to support this controversial idea as RPE values were consistently higher when the same exercise was performed with indistinguishable levels of iEMG (Fig. 4). By experimental design, what was different between these trials was the cardiorespiratory response and the augmented afferent feedback from the lower limbs in the Leg₂-post trial, which likely resulted in the greater RPE. Therefore, the present findings provide further evidence in support of the concept that RPE is actually modulated, in part, by muscle afferent feedback.

A Potential Link Between Afferent Feedback, Peripheral Locomotor Muscle Fatigue, CMD, and Endurance Performance

Growing evidence suggests that humans voluntarily terminate high-intensity, constant-load endurance exercise once their individual sensory tolerance limit, a hypothetical construct that might coincide with a certain level of peripheral fatigue and associated intramuscular metabolic milieu, is reached (23). Group III/IV muscle afferents, which relate intramuscular metabolic changes to the CNS, have been suggested to play a critical role in determining the sensory tolerance limit (4, 23). The findings from the present study now suggest that the processes determining the sensory tolerance limit, and thus the termination of exercise, might include the magnitude of both muscle afferent feedback and CMD. A conceptual framework for the interactions between afferent feedback, peripheral fatigue, CMD, and endurance exercise performance during the events of the present study is illustrated in Fig. 6. Specifically, with the initial development of peripheral fatigue during exercise of Leg₁ (point A in Fig. 6), both muscle afferent feedback and CMD increased progressively until the sensory tolerance limit was reached and the subjects were no longer able to continue the task (point B in Fig. 6). At this point (cessation of exercise with Leg₁), CMD to Leg₁ ceased entirely while, due to cuff occlusion of this leg (from point B to C in Fig. 6), afferent firing remained high. Within 10 s of the cessation of the Leg₁ trial, the cuff was released and the elevated afferent feedback from Leg₁ began to recover (19, 20, 25). However, now, with the start of the consecutive leg trial, concomitantly the afferent feedback and CMD related to Leg₂-post started to increase (point D in Fig. 6). At this point, subjects presumably exercised below their sensory tolerance limit and were thus willing and able to attempt the requested task (Leg₂-post). However, during the subsequent minutes of Leg₂-post, afferent feedback from Leg₁ (although recovering) likely remained fairly high, adding to the continuously increasing afferent feedback and CMD associated with Leg₂-post (points D and E in Fig. 6). Consequently, the sensory tolerance limit was reached during Leg₂-post more rapidly (points D and E in Fig. 6) than during the same exercise performed without preexisting peripheral fatigue in the contralateral leg [i.e., Leg₁ (points A and B in Fig. 6) or Leg₂ trial (not shown)]. Once the sensory tolerance limit was reached in Leg₂-post (point E in Fig. 6), subjects were no longer able to continue the single-leg KE, and the outcome was a greatly attenuated exercise performance and a significantly lower level of end-exercise peripheral fatigue in Leg₂-post (Fig. 2).

Fig. 6.

Schematic illustration reflecting potential sensory alterations during the consecutive single-leg knee extensor performance tests. With the onset of exercise of the first leg (Leg₁), both muscle afferent feedback and central motor drive (CMD) started to progressively rise (points A and B) until the sensory tolerance limit (dashed line) was reached at exhaustion (point B). With the end of Leg₁ exercise, CMD to this leg ceased entirely (thin dotted line), whereas group III/IV afferent firing continued due to the cuff inflation at a high level. Within 10 s, the cuff was released (point C), afferent firing from Leg₁ began to decline (dotted line), and afferent feedback and CMD related to the now exercising second leg (Leg₂-post) started to increase. In addition, afferent feedback from Leg₁ (although recovering) likely remained fairly high, adding to the continuously increasing afferent feedback and CMD associated with the exercise of the second leg (Leg₂-post) (points D and E). Consequently, the tolerance limit for this Leg₂-post trial was reached relatively quickly, as indicated by the short time to exhaustion (point E).

Experimental Considerations

Although this study circumvented several common caveats of previous research in this area, there is a need to consider other factors that may have influenced this experiment. Specifically, other determinants of central fatigue, apart from muscle afferent feedback, could potentially have contributed to the reduction in spinal motoneuronal output and exercise performance observed during the Leg₂-post trial. For example, humoral factors associated with the severe quadriceps fatigue in Leg₁ might have resulted in peripheral fatigue in the contralateral quadriceps (Leg₂-post) or facilitated CMD inhibition by directly affecting the CNS. To evaluate the influence of such humoral factors in the periphery, we measured handgrip MVC before and immediately after the consecutive exhaustive exercise tests of both legs (Fig. 1, day 4), recorded iEMG during leg exercise, and evaluated quadriceps muscle function in one leg before and again after exhaustive exercise of the contralateral leg (Fig. 1, day 3). The conclusion from these assessments was that a humoral cross-over effect of exercise-induced peripheral quadriceps fatigue into other skeletal muscle can be excluded since neither handgrip MVC nor quadriceps muscle function of the rested contralateral leg were altered by exhaustive single-leg KE. Furthermore, peripheral fatigue in one leg did not affect the rate of development of quadriceps fatigue in the contralateral leg, as evidenced by the similar rate of rise in iEMG during Leg₂ and Leg₂-post (Fig. 5).

It has also been proposed that CMD might be compromised by increases in brain serotonin levels during exercise (12, 36). Brain serotonin synthesis depends, among others, on the plasma level of TRP and its transport across the blood-brain barrier (17). An increase in the plasma ratio of TRP/BCAA is indicative of an elevated transport of TRP into the brain, which may facilitate the development of central fatigue and associated attenuation in CMD (11). We, indeed, observed a small but significant increase in the plasma ratio of TRP/BCAA from rest to the transition phase from Leg₁ to Leg₂-post; however, this ratio remained unchanged throughout the remainder of the trial. The magnitude of the ∼20% increase in TRP/BCAA in the present study, utilizing a small muscle mass exercise paradigm, was minimal compared with the over 200% increase observed after exhaustive whole body exercise (12, 40). Additionally, given that handgrip MVC was unchanged from preexercise baseline after exhaustive exercise of both legs, it might be concluded that the small change in TRP/BCAA, or any other exercise- or fatigue-induced alterations in brain neurotransmitters and their interaction with specific receptors, had no significant influence on the magnitude of CMD.

Furthermore, a link, presumably mediated by transcallosal connections (15), exists between right and left motor cortices in humans, and various cross-over effects have been identified. One example of these interhemispheric interactions following unilateral fatiguing exercise is the reduction of intracortical facilitation in the motor cortex supplying the contralateral homologous resting muscle (10). Furthermore, unilateral fatiguing exercise can decrease the postexercise excitability of the motor pathway from the contralateral motor cortex to the homologous resting muscle (13). Such interhemispheric modulation of motor pathway excitability could theoretically diminish the output of the spinal motorneurons such that, to maintain a given workload, the CNS would need to increase voluntary drive to the motor cortex. This higher motor drive would, of course, not be reflected in an increased iEMG at the muscle, since the higher drive would simply compensate for the reduced motor pathway excitability with a net effect of an unchanged iEMG. However, this phenomenon could explain the higher cardioventilatory responses during the consecutive single-leg KE in the present study (Fig. 4), which are well known to be affected by CMD (56). Additionally, the higher motor drive necessary to work against a given load would require a greater effort, which the subjects might not be able to generate for long, and this could explain the reduced exercise performance in Leg₂-post. The strongest argument against a significant role for reduced intracortical facilitation or a decreased excitability of the motor pathway supplying contralateral homologous resting muscle is temporal in nature. Specifically, neither of these processes occur immediately after a fatiguing task and have actually only been observed >5 min after unilateral exercise ceased (10, 13, 35). Additionally, during intense unilateral muscle contractions, there is some evidence that corticospinal excitability of the target muscle as well as the homologous contralateral resting muscle may actually be increased, not decreased (29). Finally, in combination, these observations suggest that interhemispheric inhibition at the motor cortical level, or “downstream,” likely did not contribute to the decreased exercise performance in Leg₂-post.

Finally, cognitive demand as well as afferent feedback from the heart and pulmonary system were presumably higher during the Leg₂-post performance test (i.e., second performance test on that day) compared with that during the Leg₂ performance test (i.e., the only performance test on that day). Therefore, we cannot exclude these differences as potential contributors to the shorter endurance time in Leg₂-post.

In conclusion, this study provides evidence that peripheral fatigue and the associated intramuscular metabolic disturbances compromise high-intensity endurance performance independent of the well known fatigue-related changes distal to the neuromuscular junction that attenuate the response of muscle to neural activation. We conclude that group III/IV muscle afferent feedback associated with intramuscular metabolic perturbation has an inhibitory effect on the CNS, which limits the output of spinal motoneurons and therefore endurance exercise performance.

GRANTS

Research reported in this publication was supported by the National Heart, Lung, and Blood Institute (HL-103786 and HL-116579 to M. Amann; HL-09183 to R. S. Richardson) and a VA Merit Grant (E6910R).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: M.A. and M.V. conception and design of research; M.A., M.V., S.J.I., J.M., G.L., and M.J.R. performed experiments; M.A. and M.V. analyzed data; M.A. and M.V. interpreted results of experiments; M.A. and M.V. prepared figures; M.A. and M.V. drafted manuscript; M.A., M.V., S.J.I., J.M., G.L., M.J.R., and R.S.R. edited and revised manuscript; M.A., M.V., and R.S.R. approved final version of manuscript.