Locomotor muscle fatigue modifies central motor drive in healthy humans and imposes a limitation to exercise performance


Corresponding author M. Amann: The John Rankin Laboratory of Pulmonary Medicine, 4245 Medical Science Center, 1300 University Avenue, Madison, WI 53706, USA. Email: amann@wisc.edu


We asked whether the central effects of fatiguing locomotor muscle fatigue exert an inhibitory influence on central motor drive to regulate the total degree of peripheral fatigue development. Eight cyclists performed constant-workload prefatigue trials (a) to exhaustion (83% of peak power output (Wpeak), 10 ± 1 min; PFT83%), and (b) for an identical duration but at 67%Wpeak (PFT67%). Exercise-induced peripheral quadriceps fatigue was assessed via changes in potentiated quadriceps twitch force (ΔQtw,pot) from pre- to post-exercise in response to supra-maximal femoral nerve stimulation (ΔQtw,pot). On different days, each subject randomly performed three 5 km time trials (TTs). First, subjects repeated PFT83% and the TT was started 4 min later with a known level of pre-existing locomotor muscle fatigue (ΔQtw,pot−36%) (PFT83%-TT). Second, subjects repeated PFT67% and the TT was started 4 min later with a known level of pre-existing locomotor muscle fatigue (ΔQtw,pot−20%) (PFT67%-TT). Finally, a control TT was performed without any pre-existing level of fatigue. Central neural drive during the three TTs was estimated via quadriceps EMG. Increases in pre-existing locomotor muscle fatigue from control TT to PFT83%-TT resulted in significant dose-dependent changes in central motor drive (−23%), power output (−14%), and performance time (+6%) during the TTs. However, the magnitude of locomotor muscle fatigue following various TTs was not different (ΔQtw,pot of −35 to −37%, P= 0.35). We suggest that feedback from fatiguing muscle plays an important role in the determination of central motor drive and force output, so that the development of peripheral muscle fatigue is confined to a certain level.

We have postulated, based on correlative evidence, that endurance exercise performance is determined to a significant extent by the feedback effects of exercise-induced peripheral muscle fatigue on central motor output (Amann et al. 2006a, 2007a,b; Romer et al. 2007). The key data supporting this postulate were obtained in studies that altered inspired O2 fraction inline image (and arterial O2 content) during endurance exercise and used supra-maximal stimulation of the femoral nerve to determine the amount of quadriceps fatigue incurred as a result of the exercise (Amann et al. 2006a,b, 2007b). The consistent finding was that alterations in arterial O2 content caused significant changes in central motor drive (i.e. quadriceps EMG) and power output during the exercise and consequently exercise performance time, but the amount of peripheral quadriceps fatigue incurred at exercise termination was identical. Accordingly, we postulated that peripheral locomotor muscle fatigue was a tightly regulated variable during exercise, such that central motor drive and therefore locomotor power output would be ‘adjusted’ by the performer so as to prevent peripheral muscle fatigue from rising above a critical threshold, beyond which the level of sensory input would not be tolerated (Gandevia, 2001).

Our present study used an interventional approach to provide a further test of this hypothesis. We prefatigued the locomotor muscles to varying degrees prior to the performance of a 5 km cycling time trial (TT) and quantified the magnitude of peripheral fatigue via supra-maximal magnetic stimulation of the motor nerve. We then observed if these varying levels of ‘prefatigue’ would influence the central motor output during exercise, the performance of a 5 km TT, and the eventual level of locomotor muscle fatigue achieved at end-exercise.



Eight competitive male cyclists volunteered to participate in the study (age 22.7 ± 1.9 years, body mass 71.1 ± 2.8 kg, stature 1.79 ± 0.07 m, maximal O2 consumption inline image 63.1 ± 3.5 ml kg −1min−1). Written informed consent was obtained from each participant. All procedures conformed to the standards set by the Declaration of Helsinki and the protocol was approved by the institution's human subjects committee.

Exercise responses

Ventilation and pulmonary gas exchange were measured breath-by-breath at rest and throughout exercise using an open circuit system including two pneumotachographs (model 3800; Hans Rudolph, Shawnee, KS, USA) (inspiration, expiration) and two Perkin-Elmer mass spectrometer (model 1100; Waltham, MA, USA) for the analysis of mixed expired and end-tidal gases (Harms et al. 1998). Arterial O2 saturation was estimated inline image using a pulse oximeter (N-595; Nellcor, Pleasanton, CA, USA) with adhesive forehead sensors. Heart rate was measured from the R-R interval of an electrocardiogram using a three-lead arrangement. Ratings of perceived exertion for dyspnoea and limb discomfort were obtained at rest and every minute during exercise using Borg's modified CR10 scale (Borg, 1998). Arterialized (Finalgon; Boehringer Ingelheim, Germany) capillary blood samples were collected from an earlobe at rest, every 3 min during the constant-workload trials and every kilometre during the TTs for determination of total whole blood lactate concentration ([La]B) using an electrochemical analyser (YSI 1500 Sport; YSI Life Sciences, Yellow Springs, OH, USA).

Neuromuscular function

Electromyography Quadriceps EMGs were recorded from the right vastus lateralis (VL), vastus medialis (VM) and rectus femoris (RF) using monitoring electrodes with full-surface solid adhesive hydrogel (H59P; Kendall, Mansfield, MA, USA), with on-site amplification. Electrodes were placed in a bipolar electrode configuration over the middle of the respective muscle belly. The active electrode was placed over the motor point of the muscle. The recording electrode was moved along the muscle until a good configuration – confirmed by a ‘maximal’ M-wave shape – was achieved. The reference electrode was placed over an electrically neutral site. The position of the EMG electrodes was marked with indelible ink to ensure that they were placed in the same location at subsequent visits. Proper electrode configuration was checked before the beginning of every experiment. To minimize movement artifacts, electrode cables were fastened to the subject's quadriceps using medical adhesive tape and wrapped in elastic bandage. The VL, VM and RF electrodes were used to record: (a) magnetically evoked compound muscle action potentials (M-waves) to evaluate changes in membrane excitability; and (b) EMG for VL throughout exercise to estimate fatigue and central neural drive. The M-wave properties included conduction time, peak amplitude and area (Caquelard et al. 2000; Sandiford et al. 2005; Katayama et al. 2007). Membrane excitability was maintained from pre- to post-exercise in all trials as indicated by unchanged M-wave characteristics. This suggests that the observed changes in potentiated twitch force (Qtw,pot) are mainly due to changes within the quadriceps and that peripheral failure of electrical transmission might be excluded.

Raw EMG signals from VL, VM and RF corresponding to each muscle contraction during the exercise trials and the pre- and post-exercise MVC manoeuvres were recorded for later analysis. The EMG signals were amplified and filtered by a Butterworth band pass filter (BMA −830; CWE, Ardmore, PA, USA) with a low pass cut-off frequency of 10 Hz and a high pass cut-off frequency of 1 kHz. The slope of the filters was −6 dB octave−1. The filtered EMG signals were sampled at 2 kHz by a 16-bit A/D converter (PCI-MIO-16XE-50; National Instruments, Austin, TX, USA) with custom software (Labview 6.0; National Instruments). A computer algorithm identified the onset of activity where the rectified EMG signals deviated by more than two standard deviations above the baselines for at least 100 ms. Each EMG burst was visually inspected to verify the timing identified by the computer. For data analysis, the integral of each burst (integrated EMG, iEMG) was calculated using the formula iEMG inline image where m is the raw EMG signal.

A 1024 point fast Fourier transform was used to compute a power spectrum periodogram. The mean power frequency (MPF) was calculated using the formula


where Sm(f) is the power density spectrum of the EMG signal.

Magnetic stimulation For a detailed description we refer the reader to previous studies from our laboratory (Amann et al. 2006b, 2007b; Romer et al. 2006). Briefly, subjects lay semirecumbent on a table with the right thigh resting in a preformed holder, the knee joint angle set at 1.57 rad (90 deg) of flexion and the arms folded across the chest. A magnetic stimulator (Magstim 200; The Magstim Company, Wales, UK) connected to a double 70 mm coil was used to stimulate the femoral nerve. The evoked quadriceps twitch force was obtained from a calibrated load cell (model SM 1000; Interface, Scottsdale, AZ, USA) connected to a noncompliant strap, which was placed around the subject's right leg just superior to the ankle malleoli. To determine whether nerve stimulation was supramaximal, three single twitches were obtained every 30 s at 50, 60, 70, 80, 85, 90, 95 and 100% of maximal stimulator power output. A near plateau in baseline Qtw and M-wave amplitudes with increasing stimulus intensities was observed in every subject, indicating maximal depolarization of the femoral nerve.

The Qtw,pot is more sensitive for detecting fatigue than the nonpotentiated twitch, particularly when the degree of fatigue is small (Kufel et al. 2002). Furthermore, Qtw,pot is more valid for comparing differences in fatigue when levels of post-activation potentiation are unequal, as was expected to occur in the present study where the exercise durations differed (Rassier & Macintosh, 2000). Accordingly, we measured quadriceps twitch force 5 s after a 5 s maximal voluntary contraction (MVC) of the quadriceps and repeated this procedure six times such that six Qtw,pot were obtained (Kufel et al. 2002). Like others (Kufel et al. 2002), we found that the degree of potentiation was slightly smaller after the first and, to a lesser extent, after the second MVC; therefore, we discarded the first two measurements. Activation of the quadriceps during the MVCs was assessed using a superimposed twitch technique (Merton, 1954; Strojnik & Komi, 1998). Briefly, the force produced during a superimposed single twitch on the MVC was compared with the force produced by the potentiated single twitch delivered 5 s afterward. The assessment procedure was performed before exercise (∼20 min) and at 4 min after exercise, which represented the duration of the break between the prefatigue trials and the subsequent TTs (see below). Peak force, contraction time (CT), maximal rate of force development (MRFD), one-half relaxation time (RT0.5), and maximal relaxation rate (MRR) were analysed for all Qtw,pot (Lepers et al. 2002; Sandiford et al. 2005).


At preliminary visits to the laboratory, subjects were thoroughly familiarized with the procedures used to assess neuromuscular functions. All participants performed a practice 5 km cycling TT, a practice constant-workload trial to the limit of exhaustion, and a maximal incremental exercise test (20 W + 25 W min−1; Amann et al. 2004) on a computer-controlled electromagnetically braked cycle ergometer (Velotron, Elite Model; Racer Mate, Seattle, WA, USA) for the determination of peak power output (Wpeak) and inline image. The following two visits were used to evaluate the degree of locomotor muscle fatigue induced via constant-load cycling at two different exercise intensities (prefatigue trials, PFTs). First, all subjects cycled to exhaustion (Amann et al. 2007b) at 83 ± 1% of Wpeak (PFT83%) (resulting in a pre- to post-exercise change in Qtw,potQtw,pot) of −36 ± 2%) (see Fig. 1). Second, exercise was repeated for the identical duration (10.1 ± 0.5 min) but at a lower exercise intensity (67 ± 1% of Wpeak; PFT67%) (resulting in ΔQtw,pot of −20 ± 1%) (Fig. 1). During these PFTs, subjects used visual and verbal feedback in order to maintain their self-selected pedal cadence (105 ± 2), as inferred from the practice trials. Next, on separate days and in random order, all participants performed three 5 km cycling TTs (performance trials) (Fig. 1). The first performance trial was carried out in a ‘fresh’ state without pre-existing locomotor muscle fatigue (Ctrl). The second performance trial (PFT83%-TT) was started following a 4 min rest period which was preceded by the PFT83% prefatigue trial; hence, the PFT83%-TT performance trial was launched with a known level of pre-existing locomotor muscle fatigue (ΔQtw,pot of −36%). The third performance trial (PFT67%-TT) was started following a 4 min rest period which was preceded by the PFT67% prefatigue trial; hence, the PFT67%-TT performance trial was launched with a known level of pre-existing locomotor muscle fatigue (ΔQtw,pot of −20%).

Figure 1.

Schematic illustration of the experimental design
As the first step, the prefatigue trials were carried out in the indicated order and separated by at least 48 h. For the second step various performance trials were conducted in random order and separated by at least 48 h.

All of the exercise trials were preceded by a 10 min warm-up at 1.5 W (kg body mass)−1 and the subjects remained seated throughout exercise. To avoid initial peak force outputs in the performance and the PFTs, subjects were instructed to slowly pick up their pace, the recording period started after the mean power output and pedal cadence, adopted from the practice trials, was reached (within 10 s). Neuromuscular functions were assessed before and 4 min after exercise (Fig. 1). The subjects were naive to the purpose of the study and the expected outcomes. Each exercise session was separated by at least 48 h and was completed at the same time of day. Subjects were instructed to refrain from caffeine for 12 h, and stressful exercise for 48 h, before each exercise trial. Ambient temperature and relative humidity were not different between conditions.

The exercise time in both PFTs was 10.1 ± 0.5 min. The group mean power output during PFT83% was 347 ± 14 W (equals 83 ± 1%Wpeak or 98 ± 1%inline image) which was individually chosen based on the mean power output from the practice 5 km TT. The power output during PFT67% was 276 ± 14 W (equals 67 ± 1%Wpeak or 75 ± 2%inline image) which was based on the intention to induce a lesser degree of end-exercise locomotor muscle fatigue compared with PFT83% (see below). The effects of various PFTs on physiological responses during the final minute of exercise are represented in Table 1 and Fig. 2. All variables indicate a significantly lower exertion at end-exercise in PFT67%versus PFT83%.

Table 1.  Physiological response to the final 30 s of constant work rate (prefatigue trials) and 5 km time trial (performance trials) exercise
 Prefatigue trial (PFT)5 km time trials (TT)
  1. Values are expressed as means ±s.e.m.aMean power output over entire time trial. Note that all given mean values for the time trials and prefatigue trials (except power output and exercise time) were taken over the final 30 s of exercise.*P < 0.05 versus PFT83%, †P < 0.05 versus Ctrl, ‡P < 0.05 versus PFT83%-TT; N= 8

Power output (W)347 ± 14276 ± 10* 347 ± 14a 298 ± 14a   332 ± 18a†,‡
Power output (% of Wpeak)83 ± 167 ± 1*83 ± 172 ± 2 †  80 ± 2†,‡
Exercise time (min)10.1 ± 0.510.1 ± 0.5  7.3 ± 0.1 7.8 ± 0.1 †    7.5 ± 0.1†,‡
r.p.m.105 ± 3 105 ± 2  105 ± 2 104 ± 2  107 ± 2 
inline image (%)92.8 ± 0.897.4 ± 0.4*92.1 ± 0.8 94.3 ± 0.8 †  95.1 ± 0.5†
HR (beats min−1)194 ± 2 175 ± 3* 197 ± 3 192 ± 3†  192 ± 3†
RPE (dyspnea) 8.8 ± 0.4 5.4 ± 0.5* 8.5 ± 0.48.7 ± 0.4 8.7 ± 0.2
RPE (limb discomfort) 8.7 ± 0.4 5.3 ± 0.6* 8.6 ± 0.38.8 ± 0.4 8.6 ± 0.2
fR (breaths min−1)61 ± 243 ± 2*62 ± 363 ± 3 63 ± 3
VT (l) 2.7 ± 0.12.5 ± 0.1 2.8 ± 0.2 2.5 ± 0.1 †  2.6 ± 0.2 †
inline image (l min−1)165 ± 8 104 ± 3* 174 ± 8 154 ± 8† 160 ± 9†
inline image (l min−1) 4.4 ± 0.2 3.4 ± 0.1* 4.4 ± 0.2 3.8 ± 0.2 †   4.0 ± 0.2†,‡
inline image (% of inline image)98 ± 175 ± 2*98 ± 186 ± 4 †  93 ± 2†,‡
inline image (l min−1) 4.3 ± 0.2 3.4 ± 0.1* 4.6 ± 0.3 3.7 ± 0.2†   4.1 ± 0.3†,‡
inline image42 ± 131 ± 1*41 ± 242 ± 1 41 ± 1
inline image41 ± 131 ± 2*38 ± 142 ± 2 40 ± 1
inline image (mmHg)112 ± 1 103 ± 1* 110 ± 2 113 ± 2  112 ± 2 
inline image (mmHg)31 ± 238 ± 1*30 ± 132 ± 1 32 ± 2
Capillary [La]B (mmol l−1)13.5 ± 1.5 4.7 ± 1.0*12.2 ± 0.5  7.6 ± 0.9†   7.5 ± 0.6 †
Figure 2.

Capillary blood lactate during the two constant-workload prefatiguing trials (PFT83%, 347 ± 14 W; PFT67%, 276 ± 10 W; 10.1 ± 0.5 min each) followed by the two experimental time trials (PFT83%-TT and PFT67%-TT)
The control time trial (Ctrl) was performed without pre-existing locomotor muscle fatigue. The two experimental time trials were started after a 4 min resting phase following the prefatiguing exercises which resulted in a reduction in potentiated twitch force (Qtw,pot) of about 36 and 20% for PFT83% and PFT67%, respectively. *P < 0.05, ‡P < 0.05 versus Ctrl.

Reliability, reproducibility and technical considerations

Magnetic femoral nerve stimulation For between-day reliability, the subjects repeated the magnetic stimulation protocol at rest on separate visits to the laboratory. There was no systematic bias in the baseline measurements between days. Mean between-day, within-subject coefficients of variation for Qtw,pot were 4.3 ± 0.8 (range 1.0–6.8), 2.1 ± 1.3 (range 0.6–5.9) for MVC and 1.2 ± 0.5 (range 0.0–3.4) for voluntary muscle activation. Additional reliability measures regarding magnetic nerve stimulation as well as technical considerations addressing the limitations of surface EMG and magnetic femoral nerve stimulation can be found in published reports (Enoka & Stuart, 1992; Keenan et al. 2005; Amann et al. 2006a,b, 2007b; Romer et al. 2006).

Exercise trials Following the practice trials, all eight subjects repeated Ctrl (TT) and PFT83% (constant workload trial to exhaustion) twice on different days for reproducibility measures. No systemic changes occurred in group mean values for performance time (Ctrl and PFT83%) and mean power output (Ctrl). Between-day coefficients of variation for performance time were 0.7 ± 0.1% (range 0.2–1.1) and 5.2 ± 1.3% (range 0.8–12.6) for Ctrl and PFT83%, respectively; and 1.3 ± 0.1% (range 1.0–1.8) for mean power output during Ctrl.

Statistical analyses

Repeated measures ANOVA was used to test for within-group effects across time. Following significant main effects, planned pairwise comparisons were made using the Holm's sequential Bonferroni procedure. Results are expressed as means ±s.e.m. Statistical significance was set at P < 0.05.


The effects of PFTs on locomotor muscle fatigue

Contractile functionQtw,pot following exercise was markedly decreased (P < 0.01) from pre-exercise baseline in both PFTs indicating substantial levels of locomotor muscle fatigue. However, the decrease in force output following PFT83%(59 ± 5 N) was substantially greater compared with PFT67% (35 ± 4 N) 4 min post exercise (P < 0.01) (Table 2, Fig. 3). Various within-twitch measurements, namely MRFD, MRR and RT0.5, as well as MVC force measures, complemented the findings reported for Qtw,pot. Significant differences for various within-twitch variables and MVC force were found between the two PFTs (Table 2).

Table 2.  Effects of constant work rate (prefatigue trials) and 5 km time trial exercise on quadriceps muscle functions
 Prefatigue trialTime trials
  1. Peripheral fatigue was assessed via supramaximal magnetic stimulation of the femoral nerve before and 4 min after exercise. Changes in fatigue variables are expressed as a percent change from pre-exercise baseline to 4 min after the termination of exercise Values are expressed as means ±s.e.m. MRFD, maximal rate of force development; MRR, maximal rate of relaxation; CT, contraction time; RT0.5, one-half relaxation time; MVC, maximal voluntary contraction. Percent muscle activation is based on superimposed twitch technique. Majority of variables changed significantly compared with baseline 4 min after exercise (P < 0.01). #Not significantly different from pre-exercise baseline. Pre-exercise, resting mean values for potentiated single twitch, MRFD, MRR, CT, RT0.5, MVC, and percentage muscle activation were 170 ± 2 N, 1553 ± 21 N s−1, 1060 ± 13 N s−1, 0.26 ± 0.00 s, 0.12 ± 0.00 s, 548 ± 7 N, and 95.6 ± 0.5%, respectively. N= 8;*P < 0.05 versus PFT83%, †P < 0.05 versus Ctrl, ‡P < 0.05 versus PFT83%-TT

Power output (W)  347 ± 14276 ± 10  347 ± 14  298 ± 14†   332 ± 18†,‡
Exercise time (s)  10.1 ± 0.510.1 ± 0.5   7.3 ± 0.1   7.8 ± 0.1†    7.5 ± 0.2†,‡
Percentage change from pre to 4 min postexercise
 Potentiated single twitch (N)−35.8 ± 2.3−20.1 ± 0.9* −35.5 ± 2.3−36.7 ± 2.8−35.1 ± 2.2 
 MRFD (N s−1)−33.2 ± 3.1−15.5 ± 1.1* −32.4 ± 2.5−34.5 ± 3.3−32.4 ± 3.0 
 MRR (N s−1)−34.0 ± 2.7−18.0 ± 1.0* −33.2 ± 2.5−35.6 ± 1.9−33.2 ± 2.5 
 CT (s) −6.5 ± 0.7− 5.2 ± 0.4* − 6.4 ± 0.7 −6.3 ± 0.6−6.1 ± 0.6
 RT0.5 (s)  6.8 ± 1.1  2.2 ± 0.7*   7.2 ± 1.5  7.2 ± 1.5  6.6 ± 0.9
 MVC peak force (N)−10.3 ± 1.6     0.7 ± 1.3 #* −9.6 ± 0.8−10.4 ± 1.9−9.3 ± 1.0
 % Muscle activation    0.4 ± 0.8 # −0.2 ± 0.4 #   −0.5 ± 0.3 #    0.0 ± 0.5 #   0.3 ± 0.4 #
Figure 3.

Peripheral quadriceps fatigue expressed as a percent change in Qtw,pot from pre- to 4 min post-exercise
The two constant-workload trials (PFT83%, 347 ± 14 W; and PFT67%, 276 ± 10 W; 10.1 ± 0.5 min each) were performed to induce a pre-existing level of peripheral fatigue. The control performance trial (Ctrl) was conducted without pre-existing locomotor muscle fatigue. The two experimental performance trials (PFT83%-TT and PFT67%-TT) were conducted with a pre-existing level of peripheral quadriceps fatigue. The time trials started after a 4 min resting phase following either PFT83% or PFT67%. Note that despite significantly different levels of pre-existing locomotor muscle fatigue, resulting in substantially different exercise performances, end-exercise locomotor muscle fatigue was almost identical between the three performance trials and the PFT83% prefatiguing trial (dashed line) supporting the hypothesis of an existing critical threshold of fatigue. n= 8; *P < 0.01.

EMG activity Integrated EMG of vastus lateralis rose significantly from the first minute of exercise to the point of exhaustion in all eight subjects at both PFTs (P < 0.01). The exercise-induced increase in iEMG was significantly greater in PFT83%versus PFT67% (15.8 ± 1.8 and 8.0 ± 1.9%, respectively) (Fig. 4). The continuous increase in iEMG was associated with a significant decrease in MPF from the first minute to the termination of exercise at both trials. At termination of exercise, MPF had fallen significantly more during PFT83%versus PFT67% (11.6 ± 1.1 and 6.7 ± 0.9%, respectively) (Fig. 4).

Figure 4.

Myoelectrical activity
A, integrated EMG (iEMG), and B, mean power frequency (MPF), of vastus lateralis used to illustrate the development of peripheral locomotor muscle fatigue during the two constant-workload prefatigue trials (PFT83%, 347 ± 14 W; and PFT67%, 276 ± 10 W). Values are normalized to the first minute of exercise. Mean values for iEMG and MPF during each muscle contraction (cycle revolution) were calculated and averaged over each 60 s period. *P < 0.01.

The effects of pre-existing fatigue on subsequent TT performance, iEMG and locomotor muscle fatigue

TT performance Exercise performance was significantly affected by the level of pre-existing locomotor muscle fatigue. Time to complete the 5 km performance task increased from Ctrl (no pre-existing level of fatigue) to PFT67%-TT (moderate level of pre-existing fatigue, Table 2) by 2.0 ± 0.8% (range 1–8%; P < 0.05) and from Ctrl to PFT83%-TT (severe level of pre-existing fatigue, Table 2) by 6.1 ± 0.7% (range 3–9%; P < 0.01) in all eight subjects (Table 1). These changes in time to completion are reflected in mean power output during the TTs which was 4.4 ± 1.8% (range 1–16%) lower in PFT67%-TT and 14.1 ± 1.6% (range 6–21%) lower in PFT83%-TT versus Ctrl (P < 0.05 and 0.01, respectively) (Table 1). The profile of power output during various performance trials are shown in Fig. 5.

Figure 5.

Effect of pre-existing locomotor muscle fatigue on neural drive and power output during a 5 km time trial
The control time trial was performed without pre-existing locomotor muscle fatigue. The two experimental time trials (PFT83%-TT and PFT67%-TT) were performed with different levels of pre-existing quadriceps fatigue (ΔQtw of about −36 and −20% for PFT83%-TT and PFT67%-TT, respectively). A, effects of pre-existing locomotor muscle fatigue on group mean iEMG of vastus lateralis normalized to the iEMG obtained during pre-exercise (unfatigued) maximal voluntary contractions (MVC) of the quadriceps. Each point represents the mean iEMG of the preceding 0.5 km section. Mean iEMG during the time trial was significantly reduced from Ctrl to PFT83%-TT. B, group mean variations in power output during the 5 km time trial with three different levels of pre-existing fatigue. Group mean power output was 347 ± 14 W, 298 ± 14 W and 332 ± 18 W (P < 0.05) for Ctrl, PFT83%-TT and PFT67%-TT, respectively; n= 8.

Central neural driveFigure 5 illustrates changes in iEMG during the three performance trials; each point represents a 500 m segment of the TT. For all TTs, mean iEMG of the vastus lateralis was normalized to the iEMG obtained from pre-exercise MVC manoeuvres (performed without pre-existing fatigue) on each day. The average iEMG ratio over the entire TT was reduced (P < 0.05) with each increment in pre-existing level of peripheral locomotor muscle fatigue, i.e. from Ctrl to PFT67%-TT (6.9 ± 0.7%) and PFT67%-TT to PFT83%-TT (17.5 ± 1.4%) (Fig. 6). A reduction in average iEMG from Ctrl to PFT83%-TT occurred in seven of the eight subjects (range 8–36%).

Figure 6.

Summary of effects of pre-existing level of locomotor muscle fatigue on group mean neural drive and mean power output during a 5 km time trial, and end-exercise locomotor muscle fatigue (percentage reduction in Qtw,pot)
The control 5 km time trial was started without pre-existing locomotor muscle fatigue. The remaining time trials were started with significant levels of locomotor muscle fatigue (ΔQtw,pot−20 and −36% for PFT67%-TT and PFT83%-TT, respectively, P < 0.05).

Contractile function Immediately after each TT, group mean Qtw,pot was reduced between 57 and 62 N from pre-exercise baseline (P < 0.01). These reductions in Qtw,pot 4 min following the TTs were not significantly different between conditions of varying pre-existing levels of locomotor muscle fatigue (−35 to −37%; P= 0.35) (Table 2, Fig. 6). All within-twitch measurements (MRFD, MRR, CT and RT0.5) were significantly altered from baseline immediately post-exercise (Table 2). No differences in these variables were observed 4 min post-exercise between the three performance trials (P= 0.18–0.91). Peak force during the 5 s MVC manoeuvres was significantly decreased from baseline after all trials (P < 0.01). These reductions in MVC force did not differ among the three TTs (P= 0.78) (Table 2).

Relationship between Qtw,pot and TT iEMGFigure 7 illustrates the relationship between pre-TT quadriceps strength (Qtw,pot) and average neural drive (iEMG) during subsequent TTs. In general, the results suggest that any exercise-induced reduction in Qtw,pot existing prior to the start of a 5 km TT causes a significant reduction in average iEMG during the TT and the greater the peripheral fatigue the greater the reduction in iEMG during subsequent exercise.

Figure 7.

Group mean data (n= 8) showing the relationship between Qtw,pot as measured prior to the start of a 5 km time trial, and mean iEMG of vastus lateralis during subsequent time trials
Pre-existing levels of fatigue (i.e. reductions in pre-time trial Qtw,pot from Ctrl) was induced via constant-workload exercise.

Contractile functions: PFT83%versus performance trials The reductions in Qtw,pot 4 min after exercise was nearly identical between PFT83% and the three performance trials (P= 0.44; Table 2, Fig. 3). Furthermore, no significant differences were observed among various within twitch measurements (P= 0.37–0.92) and reductions in peak forces during the MVC manoeuvres were similar following exercise (P= 0.78) (Table 2).


We asked if central motor drive – and therefore exercise performance – is regulated to avoid the development of peripheral locomotor muscle fatigue beyond an individual, highly task-specific, critical threshold. Using constant-workload trials we induced two significantly different levels of locomotor muscle fatigue immediately prior to the start of a 5 km cycling TT performance test during which the subjects were able to voluntarily choose their power output in order to finish the task as fast as possible. Pre-existing locomotor muscle fatigue had a substantial dose-dependent inverse effect on central motor output and power output during the 5 km TTs and a direct effect on performance time. However, despite the significant differences in quadriceps fatigue at the initiation of the TT, the magnitude of peripheral muscle fatigue developed at end-exercise was identical (Fig. 6). These data emphasize the crucial role of locomotor muscle fatigue on exercise performance via its inhibitory influence on central motor drive and, furthermore, confirm the status of peripheral fatigue as a carefully regulated variable.

Indices of reductions in central motor drive with locomotor muscle fatigue

Changes in central motor drive during various TTs generally followed those in power output, both with increasing distance covered and with alterations in the level of prefatigued locomotor muscle (Fig. 5). Accordingly, we believe that the observed changes in EMG of the vastus lateralis associated with alterations in the pre-existing level of locomotor muscle fatigue permit the qualitative conclusion that changes in locomotor muscle power output were in a similar direction to those in central neural locomotor output. However, we cannot quantify the proportion of the observed changes in power output attributable to changes in central neural drive. Instead, it is also plausible that the observed level of peripheral fatigue (i.e. biochemical changes within the muscle) prevented the muscle from responding to the same extent to an identical level of motor command (also see below).

The estimation of central motor drive will be affected by the accuracy of EMG measurements via surface electrodes (Keenan et al. 2005; Amann et al. 2006a,b) and potential changes could be masked by noise. Therefore, since we postulate that the limitation to exercise performance is due primarily to a consciously and/or subconsciously reduced central motor command rather than due to the inability of the locomotor musculature to respond to increases in neural drive, an additional and crucial observation indicating reduced central command during the TT tests needs to be emphasized. Namely, power output was identical, at least for brief time periods, at the beginning of each TT – despite different levels of pre-existing fatigue – and rose to almost identical levels in the final 200 m (Fig. 5). These findings indicate that the locomotor musculature maintained responsive to central motor drive throughout exercise even in conditions of severe peripheral fatigue. This maintained capability of the locomotor muscles to increase power output in response to increases in central neural drive emphasizes the fact that the subjects chose– again, consciously and/or subconsciously – to keep central motor output at a certain submaximal level presumably to avoid further accumulation of peripheral fatigue (i.e. further accumulation of nociceptor stimulating metabolic byproducts; see below) and consequently intolerable levels of sensory input. Accordingly, we suggest that the stimulation of intramuscular pain receptors via metabolic byproducts of muscular contractions and associated feedback to the CNS could be viewed as a dose-dependent trigger of central fatigue, i.e. the stronger the stimulus and the greater the rate of development of peripheral fatigue, the greater the inhibitory afferent feedback to the CNS which in turn affects the determination of central motor drive.

Effect of prior fatigue on exercise performance and the development of locomotor muscle fatigue

Central neural drive – and consequently power output – during various 5 km TTs was inversely related to the level of pre-existing locomotor muscle fatigue. In other words, the higher the level of pre-existing locomotor muscle fatigue (i.e. the lower the pre-existing Qtw,pot), the lower the average neural drive during the subsequent TT (Fig. 7). The striking finding was that at end-exercise the level of peripheral fatigue was identical between the three TTs – independent of the level of pre-existing fatigue and/or the marked differences in exercise performance (see summary in Fig. 6). For instance, one of the TTs was started with a severe level of pre-existing fatigue as induced via high-intensity constant-workload exercise to voluntary exhaustion (PFT83%). Hence, the individual critical threshold of peripheral fatigue, as proposed previously (Amann et al. 2006a), – or sensory tolerance limit (Gandevia, 2001) – had already been achieved when the TT started (see Fig. 3). Astonishingly, since the level of locomotor muscle fatigue at the end of the TT (PFT83%-TT) was identical compared with the pre-existing level at the start of the TT (i.e. at the critical threshold of fatigue) (see Fig. 3), the subjects, who were instructed to finish the TT as fast as possible, must have voluntarily chosen a power output throughout the race low enough to result in no further accumulation of peripheral fatigue. On the other hand, when the TT was started with no pre-existing fatigue (control) or a lower level of pre-existing locomotor muscle fatigue (PFT67%-TT, see Fig. 3), peripheral fatigue further accumulated throughout the subsequent TT to reach the critical threshold at end-exercise. Combined, these findings suggest that central neural drive during various TTs is modulated to result in a rate of development of peripheral fatigue that prohibits one from exceeding the critical threshold or sensory tolerance limit associated with a certain level of peripheral fatigue. Our present use of prior exercise as a means of varying the degree of peripheral fatigue has produced very similar effects on central motor output, performance, and end-exercise peripheral fatigue as we previously reported using varying inline image values to change arterial O2 content across the range of hyperoxia to moderate hypoxaemia (Amann et al. 2006a, 2007b).

We emphasize that the percentage reduction in Qtw,pot associated with the critical threshold of peripheral fatigue as observed in these and previous experiments is likely to be highly task specific (Enoka & Stuart, 1992) and conditions specific. For example, we have recently reported a much lower level of peripheral fatigue to develop following exhaustive exercise in very severe hypoxia inline image (Amann et al. 2007b) and others have suggested that brief all-out Wingate-type tests or small muscle group tests also do not provide large amounts of peripheral muscle fatigue (Calbet et al. 2003, 2006).

Prior exercise – varying in duration, intensity and muscle group utilized – has previously been used to induce different levels of metabolites in order to investigate their influence on subsequent metabolic and perception responses and exercise performance (Karlsson et al. 1975; Hogan & Welch, 1984; Bangsbo et al. 1996; Hargreaves et al. 1998; Burnley et al. 2002; Wilkerson et al. 2004; Eston et al. 2007; Raymer et al. 2007). Although no previous study has prefatigued the locomotor muscle, quantified the degree of peripheral fatigue and evaluated these effects on subsequent central motor drive and exercise performance, the approach by Hogan & Welch (1984) is relevant to our work. While working at the same absolute workload and for the same duration but breathing different inline image values (0.16 or 0.60) during exercise, the authors induced different levels of systemic [H+] and [La] and investigated these effects on subsequent (4 min break) performance trials in normoxia utilizing the same muscle group (cycling). Blood [H+] and [La] at the beginning of these constant-workload performance trials were significantly higher following the hypoxic versus hyperoxic ‘pre-treatment’ and resulted in a significantly shorter time to exhaustion for the subsequent constant-load exercise trial. Nevertheless, [H+] and [La] at exhaustion were not significantly different between trials despite the marked differences in exercise performance time. Their postulate of muscle intracellular [H+] as a regulated variable may be consistent with our present findings since [H+] accumulation might be considered an indicator of peripheral fatigue (Hogan & Welch, 1984). However, the view on intracellular [H+] is controversial inasmuch that some believe in its role as only a marker of muscle fatigue (Miller et al. 1988; Cady et al. 1989; Weiner et al. 1990; Kent-Braun, 1999; Kristensen et al. 2005; Messonnier et al. 2007), while more recent investigations cast doubt on the influence of in vivo muscle pH on fatigability (Pate et al. 1995; Bangsbo et al. 1996; Bruton et al. 1998; Dibold & Fitts, 2000; Pedersen et al. 2004). Our current findings do not support a role for blood [La] as a regulated variable at least during TT cycling performances. First, the level of blood [La] at the end of various TTs was different in the current investigation (see Fig. 2) despite nearly identical levels of locomotor muscle fatigue. Furthermore, in one of our earlier studies different exercise performances also yielded significantly different levels of blood [La] whereas end-exercise locomotor muscle fatigue was identical (Amann et al. 2006a).

Intermediate role of the somatosensory system in the regulation of central motor drive

How does the central nervous system know when and to what extent voluntary motor neuron activation (i.e. central motor drive) needs to be adjusted in order to avoid excessive development of peripheral fatigue beyond a sensory tolerance limit associated with potential locomotor muscle tissue damage and pain? Aδ and C-polymodal nociceptors are distributed throughout muscles and connected via myelinated and unmyelinated peripheral nerve fibres (group III and IV afferents, respectively) to the dorsal horn in the spinal cord from which the impulse ascends to the brain. Noxious stimuli of these free nerve endings include substances released and triggered by damaged cells (e.g. bradykinin, prostaglandins, histamine, etc.) as well as metabolic products of muscular contraction (e.g. H+) (Rotto & Kaufman, 1988; Mense, 1996) and fatigue (Darques et al. 1998). The afferent supraspinal projection of these nociceptive stimuli involves multiple ascending pathways feeding into subcortical and cortical structures including thalamus, limbic system and prefrontal cortex (Almeida et al. 2004). Therefore, by projecting into these areas of the brain, nociceptive afferent feedback is likely to be involved in the determination of goal-directed behaviour (limbic processing, i.e. motivation; Schultz et al. 1998) and motor processing (Chaudhuri & Behan, 2000), both of which are significant determinants of central motor drive. It has recently been suggested that the inhibitory effects of group III and IV afferents on voluntary motoneuron activation during local isometric exercise act ‘upstream’ of the motor cortex (Gandevia et al. 1996; Taylor et al. 1996, 2000).

Limitations: other potential mediators of central motor drive

The reduced central motor drive observed during the TTs subsequent to the PFTs is a distinct indication of central fatigue whose mechanisms are complex, interactive and not mutually exclusive (Gandevia, 2001). While we present simple interventional evidence supporting a significant contribution of peripheral locomotor muscle fatigue to central fatigue, other determinants have been suggested and might additionally and/or equally augment the reduction in central drive during the TTs with prior exercise versus the control trial without prior exercise. First, prior exercise in the current study might have caused a disturbance in brain neurotransmitter levels or turnover known to result in central fatigue (Newsholme et al. 1987; Blomstrand et al. 1988; Davis & Bailey, 1997; Meeusen et al. 2006; Newsholme & Blomstrand, 2006) and potential carry-over effects could have contributed to the reduced motor drive during subsequent TTs. Especially relevant in this context is the increase in brain serotonin (5-hydroxytryptamine), which is a result of exercise-induced increases in the plasma level of its precursor tryptophan (Chaouloff et al. 1985; Blomstrand, 2006), since the level of circulating plasma tryptophan might still have been significantly increased at the start of the TTs following the PFTs. Second, it has been shown that central command during high intensity exercise can lead to a local depletion of brain glycogen (Dalsgaard et al. 2002) causing central fatigue. Therefore, depletion or even only a partial reduction of brain glycogen during the PFTs might have affected central motor drive during subsequent TTs since the glycogen stores could most probably not have been replenished during the 4 min break between exercises. Third, increases in core and especially brain temperature of over 40°C have been shown to impair voluntary central motor drive and the ability to sustain maximal muscle activation (Nybo & Secher, 2004). Although oesophageal temperature during a laboratory 5 km TT (21°C ambient temperature) rises only to about 39°C (Amann et al. 2006a), a higher core temperature with prior exercise can not be excluded and might have contributed to the reduction in central motor drive. Finally, while our correlative findings support the concept that neural feedback from fatiguing locomotor muscle exerts an inhibitory effect on central motor drive, it has been suggested that inhibitory reflexes might also originate in fatiguing respiratory muscles (Bigland-Ritchie & Vollestad, 1988; Kayser, 2003).

Our correlative findings to date are fairly consistent in support of a significant role for peripheral muscle fatigue as a major determinant of central motor drive (Amann et al. 2006a, 2007b; Romer et al. 2007) and our present findings using an interventional approach are also consistent with this postulate. Nonetheless, this complex question now needs to be addressed using a more specific intervention capable of clearly differentiating the effects of peripheral muscle fatigue, per se, on central motor output in the exercising human. The use of spinal blockade of sensory input from fatiguing locomotor muscles (Rotto & Kaufman, 1988) should accomplish this goal.

Summary and future direction

By inducing various degrees of locomotor muscle fatigue prior to a performance task, we have shown a dose-dependent effect of peripheral fatigue on central neural drive during exercise. Central motor command, and consequently power output and exercise performance, was highest during the TT performed without pre-existing peripheral fatigue and lowest when the TT was started with a severe degree of locomotor muscle fatigue. Furthermore, exercise-induced locomotor muscle fatigue at the termination of the TTs was nearly identical despite significant differences in exercise performance and pre-existing levels of quadriceps fatigue confirming the status of locomotor muscle fatigue as a carefully regulated variable. Nonetheless, a limitation is imposed on our interpretation of our findings because the prefatiguing exercise might also bring into play other non-peripheral effectors of central fatigue. We now need further more specific intervention studies which block sensory neural feedback from fatiguing locomotor muscles to evaluate its specific effects on central motor command.



This research was supported by a National Heart, Lung and Blood Institute (NHLBI) RO1 grant (HL-15469). We thank Mrs Cynthia E. Bird and Mrs Rose E. Voelker for valuable assistance with lactate assessment and EMG data analyses, respectively. Mr David Pegelow's technical assistance with data collection was important to the conduct of these studies.