Journal Logo


Respiratory Muscle Fatigue Alters Cycling Performance and Locomotor Muscle Fatigue


Author Information
Medicine & Science in Sports & Exercise: November 2020 - Volume 52 - Issue 11 - p 2380-2389
doi: 10.1249/MSS.0000000000002399
  • Free


Intense, repeated contractions of skeletal muscle can lead to a temporary reduction in the ability of the muscle to generate force (1). This condition, known as muscle fatigue, can occur during whole-body exercise such as cycling, as well as during isolated muscle tasks such as inspiratory breathing against a resistance (2,3). A task- and individual-specific magnitude of locomotor muscle fatigue has been demonstrated at the end of high-intensity exercise, suggesting that a “critical threshold” of fatigue exists and limits exercise performance (2,4,5). It seems that to prevent the locomotor muscles from exceeding this critical threshold, decreases in motoneuronal output (estimated through surface electromyography (EMG) of the locomotor muscles) and power output during exercise transpire (2,6,7). The central nervous system regulates motoneuronal output, in part, by processing feedback from group III/IV afferent neurons that are sensitive to mechanical and metabolic stimuli in the locomotor muscles, likely to constrain intramuscular metabolic perturbation to a tolerable limit (6,8). Indeed, specific alterations in intramuscular metabolites such as adenosine diphosphate, hydrogen ions, inorganic phosphate, and phosphocreatine are present at the end of high-intensity exercise (6,9,10).

Motoneuronal output, power output, and the magnitude of locomotor muscle fatigue can also be influenced by sensory feedback from previously exercised muscles. For example, time to exhaustion and the amount of peripheral fatigue were reduced after constant load cycling that was preceded by arm cranking exercise (11). Similarly, performing single-leg knee extensions in one leg after exercising the contralateral leg to task failure shortened time to exhaustion and reduced the magnitude of peripheral fatigue compared with no prior leg exercise (12). These findings lend support to the idea of a “sensory tolerance limit” whereby sensory feedback originating from the previously exercised muscles combines with the sensory feedback originating in the muscles executing the performance task, and this combination exceeds a global sensory threshold (1,13). Exceeding this sensory limit causes task failure to occur without the attainment of a critical level of locomotor muscle fatigue. This concept illustrates the integrative nature of limitations to exercise performance in that global feedback from multiple sources is combined to determine motoneuronal output rather than only feedback from the locomotor muscles.

An important source of sensory feedback during exercise is the respiratory system, which is known to contribute to exercise limitations and fatigue from both a physiological and perceptual standpoint (14). If sensory feedback from the respiratory system is mismatched with respiratory motorneuronal output, dyspnea can occur, which may contribute to the attainment of a sensory tolerance limit. One attractive approach to understand the respiratory system’s role in the regulation of exercise performance, as well as its influence on perceptions of dyspnea and locomotor muscle fatigue, is to induce respiratory muscle fatigue (RMF) before exercise. Although some studies have assessed the effect of prior respiratory muscle work on exercise performance, the data are equivocal (15–19). Differences in the methods used to induce RMF and/or differences in the protocol used to assess exercise performance may explain the discrepant results. Regarding the latter, some studies have used constant load cycling exercise to investigate the effects of preexisting RMF (19) or changes to the work of breathing (20) on peripheral fatigue and exercise tolerance. Although the use of constant load exercise is advantageous for the comparison of specific physiological variables, per se, it does not allow for an examination into how alterations to perception of effort (e.g., ratings of dyspnea) might alter the self-regulation of power output during an exercise performance task. Therefore, the purpose of this study was to determine if preexisting RMF alters exercise performance, motoneuronal output estimated through surface EMG, and the development of locomotor muscle fatigue during a self-paced cycling time trial where power output can be freely altered by the subject. We hypothesized that preexisting RMF would worsen cycling performance, reduce motoneuronal output, and reduce the magnitude of locomotor muscle fatigue after exercise.



Eight healthy, trained male cyclists (age, 22 ± 4 yr; height, 185 ± 6 cm; body mass, 78 ± 9 kg; forced vital capacity (FVC), 6.2 ± 1.0 L; forced expiratory volume in 1 s (FEV1), 5.0 ± 0.9 L; FEV1/FVC, 82% ± 4%; peak expiratory flow (PEF), 10.5 ± 1.5 L·s−1) volunteered to participate in the study. All subjects displayed normal pulmonary function based on prediction equations (21). Subjects were tested at the same time of day for each visit and were instructed to arrive at the laboratory having abstained from caffeine, alcohol, and strenuous exercise for the previous 24 h. All subjects provided written informed consent, and all experimental procedures were approved by the institutional review board of Indiana University.

Experimental design

Subjects were tested on 3 d, each separated by a minimum of 48 h and a maximum of 1 wk. Day 1 included pulmonary function testing, measurement of maximal inspiratory mouth pressure, and familiarization to the 5-km cycling time trial, neuromuscular function testing procedures, and the resistive breathing task. Days 2 and 3 were performed in random order and included a 5-km cycling time trial preceded by either seated rest (CON) or an inspiratory resistive breathing task (RMF). Days 2 and 3 also included measures of preexercise and postexercise locomotor (i.e., quadriceps) neuromuscular function. Measurement of maximal inspiratory mouth pressure after the resistive breathing task began 1 min after task failure. Four maximal efforts, each separated by 30 s, were obtained. Subjects were then transferred to the cycle ergometer (1.5 min) and completed a 3-min warm-up followed by 1 min of rest. Thus, the 5-km time trial started 8 min after the completion of the resistive breathing task. Measurements of quadriceps function began 2 min after the completion of the 5-km time trial.

Pulmonary function testing

Pulmonary function was assessed using independently calibrated inspired and expired pneumotachographs (Series 3813/4813; Hans Rudolph, Shawnee, KS) and done in accordance with standardized procedures (22). Measurements included FVC, FEV1, and PEF. Subjects were familiarized with the performance of FVC maneuvers and were provided with verbal feedback. Subjects performed three to five maximal maneuvers consisting of a maximal inspiration to total lung capacity and then a maximal expiration to residual volume. The largest FVC, FEV1, and PEF were used as a representation of the subject’s pulmonary function.

Maximal inspiratory mouth pressure

Mouth pressure (Pm) was measured through a side port in the mouthpiece connected to a pressure transducer (Hector Engineering, Ellettsville, IN) which was calibrated using a water manometer. The assessment of maximal mouth pressure was done in accordance with standardized procedures (23). Subjects performed four maximal inspiratory efforts initiated from residual volume, with the average of the highest three defined as maximal inspiratory pressure (Pm,max). The inspiratory efforts were performed against a customized device with a 2-mm orifice to prevent glottic closure. Maximal inspiratory mouth pressure was measured on day 1, as well as before and 1 min after the resistive breathing task.

Cycling time trial

Power output during the 5-km cycling time trials was measured using an electromagnetically braked cycle ergometer (Velotron Racermate, Seattle, WA). The only feedback given during the time trial was a continual display of elapsed distance. Seat and handlebar placements were set according to subject preference and recorded during the familiarization trial to ensure consistent placement during subsequent trials.

Exercise responses

Metabolic and ventilatory variables were continuously measured during exercise using open circuit, indirect calorimetry with a customized metabolic cart. Subjects breathed through a mouthpiece (9060; Hans Rudolph) connected to a low resistance, two-way nonrebreathing valve (2700; Hans Rudolph) from which expired gases were collected in a 5-L mixing chamber. Dried samples from this mixing chamber, sampled at 300 mL·min−1, were used to determine fractional concentrations of O2 and CO2 by separate O2 and CO2 gas analyzers (S-3A/CD-3A; Ametek Thermox Instruments, Pittsburgh, PA). Analyzers were calibrated with known gas concentrations within the physiological range. Air flow and gas concentrations were digitized and sampled at 50 Hz using a data acquisition system (DASYLab; National Instruments, Norton, MA). Inspiratory capacity maneuvers were performed at the last 0.1 km of each 1-km segment (i.e., 0.9, 1.9, etc.). End-expiratory lung volume (EELV) was calculated by subtracting inspiratory capacity volume from FVC and expressed as percent of FVC. End-inspiratory lung volume (EILV) was calculated by adding tidal volume (VT) to EELV and expressed as a percent of FVC. RPE (Borg original 6–20 scale) and dyspnea (modified 0–10 scale) (24) were obtained at the three-quarter point of each kilometer (i.e., 0.75, 1.75, etc.) during the time trial. Before the start of the time trial, a standard script was read to the subjects explaining the dyspnea scale: “This is a scale that asks you to rate the difficulty of your breathing. It starts at the number zero where your breathing is causing you no difficulty and progresses through to the number ten where your breathing difficulty is maximal. At the three-quarter point of each kilometer, you will be asked to rate the difficulty of your breathing by pointing to the corresponding the number on the laminated card.”

Inspiratory resistive breathing

A bout of inspiratory resistive breathing to task failure was used during the RMF visit. The resistive breathing task was similar in design to other tasks known to induce RMF (3,25). Subjects sat upright and breathed through a customized two-way nonrebreathing valve connected to a customized flow resistor on the inspiratory side. Target mouth pressure during inspiration was 60% of Pm,max. During the familiarization visit, subjects practiced breathing through the apparatus, and the flow resistor was adjusted for each subject to ensure the target pressure could be attained at the required ventilatory parameters. The flow resistor setting was recorded, and the same setting was used during the subsequent visit (RMF) to the laboratory. Subjects were required to maintain a specific breathing frequency (15 breaths per minute) and inspiratory duty cycle (TI/TTOT = 0.67). Ventilatory parameters were maintained through feedback provided by a visual tracing of mouth pressure and metronomes with distinct inspiratory and expiratory tones. Mouth pressures during each breath were not maintained in a square wave manner; therefore, peak mouth pressure during each breath (Pm) and mean mouth pressure during each breath (P¯m) are both reported. The tension time index was calculated for each minute as the product of the inspiratory duty cycle and the ratio of the mean mouth pressure to the maximal mouth pressure (TTI = TI/TTOT × P¯m/Pm,max). Task failure was defined as a failure to generate the target mouth pressure for four consecutive breaths. RMF was defined as a significant change in maximal mouth pressure after the resistive breathing task compared with resting measures. During the control visit, participants sat quietly for 20 min in the same chair used for the resistive breathing task.

Locomotor muscle fatigue

Locomotor muscle fatigue was quantified as the percent reduction in potentiated quadriceps twitch force (Qtw,pot) from preexercise to postexercise. Participants were seated on a customized chair with a trunk/thigh angle of 150° and a right knee joint angle of 90°. An inelastic strap connected to a calibrated load cell (Model Z Tension Load Cell; Dillon, Fairmont, MN) was placed around the right ankle just superior to the malleoli. A commercially available magnetic stimulator (Magstim 200-2; Jali Medical, Newton, MA) connected to a 70-mm coil was used to stimulate the femoral nerve, and all evoked forces were sampled at 2000 Hz using a data acquisition system (AcqKnowledge Software v 5.0; BIOPAC Systems, Inc., Goleta, CA). The area of stimulation was determined by the coil location within the femoral triangle that evoked the largest unpotentiated quadriceps twitch force (Qtw). This location was marked using indelible ink, and triangulated measurements from anatomical landmarks were recorded in order to ensure correct placement during all testing sessions. To confirm that stimulation was supramaximal, two twitches were obtained, each separated by 30 s, at 70%, 80%, 90%, 95%, and 100% of the stimulator’s output. Before and 2 min after exercise six maximal voluntary contractions (MVC) of the right quadriceps, separated by 30 s, were performed for 3 s each. The twitch force in response to a single supramaximal stimulation (Qtw,pot) was measured 2 s after the MVC. Twitch force amplitude was calculated as the difference between the baseline force and the peak force induced by the supramaximal stimulation and was reported as the average value of the last four twitches.

Motoneuronal output

Myoelectrical activity of the vastus lateralis was recorded by surface EMG using Ag/AgCl electrodes (EL501, BIOPAC Systems, Inc.). Electrodes were placed ~4 cm apart on the muscle belly of the vastus lateralis in parallel to the fiber orientation. The location of the electrodes was marked using indelible ink, and triangulated measurements from anatomical landmarks were recorded in order to ensure correct placement during all testing sessions. The electrodes were used to measure muscle action potentials (M-waves) evoked from magnetic stimulation and EMG throughout exercise to estimate motoneuronal output (2,12). The raw signals were digitized, sampled at 2000 Hz, band-pass filtered (20–500 Hz), and analyzed using commercially available software (AcqKnowledge Software v 5.0; BIOPAC Systems, Inc.). M-wave amplitude was calculated as the peak-to-peak amplitude and was reported as the average value of the last four M-waves. Muscle contractions were identified by locating when the EMG signal rose >2.5 SD above baseline noise. The root mean squared EMG (RMS EMG) for each contraction was calculated using a time constant of 100 ms and was averaged over each 1-km segment of the time trial before being normalized to the highest 0.25-s average RMS EMG recorded during the preexercise MVC.

Statistical analysis

Dependent variables were assessed for normality by visually inspecting QQ-norm plots and using the Shapiro–Wilk test, and for sphericity using the Mauchly test when necessary. If sphericity was violated, the departure from sphericity (ε) was calculated. If ε < 0.75, the Greenhouse–Geisser correction was applied, and if ε > 0.75, the Huynh–Feldt correction was applied. Paired t-tests were used to detect differences in quadriceps neuromuscular function, cycling performance, and maximal inspiratory mouth pressure. A 2 × 5 (condition by distance) repeated-measures ANOVA was used to detect differences in metabolic, ventilatory, and perceptual variables, as well as power output and RMS EMG, across the cycling time trial. Following significant F ratios, pairwise comparisons were made using the Holm–Bonferroni method. Statistical significance was set at P < 0.05.


Inspiratory resistive breathing

All subjects completed the inspiratory resistive breathing protocol until task failure (26 ± 11 min). Physiological responses to the breathing task are shown in Table 1. All subjects adequately maintained breathing frequency and inspiratory duty cycle parameters throughout the duration of the task. All subjects generated mouth pressures equal to the target pressures until task failure. As shown in Figure 1, maximal inspiratory mouth pressure was reduced in all subjects after the resistive breathing task compared with baseline pressure, with a group mean decrease of 28 ± 12 cm H2O (95% confidence interval (CI), 17–38 cm H2O; t7 = 6.412, P < 0.001), which corresponds to a change from baseline of 19% ± 9%.

TABLE 1 - Physiological responses during inspiratory resistive breathing task.
Time, min
1 2 10 Task Failure (26 ± 11 min)
P m, cm H2O −89 ± 19 −93 ± 25 −88 ± 25 −82 ± 25
P m/P m,max, % 62 ± 6 64 ± 10 62 ± 10 57 ± 8
P¯m, cm H2O −48 ± 11 −50 ± 13 −46 ± 14 −45 ± 14
P¯m/P m,max, % 34 ± 6 35 ± 7 32 ± 6 31 ± 6
T I/T TOT 0.65 ± 0.03 0.65 ± 0.04 0.67 ± 0.04 0.66 ± 0.04
TTIm 0.22 ± 0.04 0.23 ± 0.05 0.21 ± 0.04 0.21 ± 0.05
f b, breaths per minute 14 ± 1 15 ± 1 15 ± 1 15 ± 1
V T, L 1.8 ± 0.3 1.8 ± 0.4 1.7 ± 0.2 1.6 ± 0.3
V̇E, L·min−1 26 ± 4 27 ± 4 25 ± 3 24 ± 5
P ETCO2, mm Hg 34 ± 4 34 ± 4 29 ± 3* 28 ± 1*
Data are presented as means ± SD.
*Significantly different from first minute (P < 0.05).
TTIm, tension time index of mouth pressure; fb, breathing frequency; PETCO2, end tidal partial pressure of CO2.

Maximal inspiratory mouth pressure. Reductions in maximal inspiratory mouth pressure (P m,max) after the inspiratory resistive breathing task. Filled circles and solid lines are individual subjects. Open squares are group means ± SD (n = 8). #Significantly different from PRE (P < 0.001).

Motoneuronal output, power output, and cycling performance

RMS EMG and power output throughout the time trial are shown in Figure 2. RMS EMG analysis was conducted on seven subjects, as there were technical difficulties during one data collection visit for one subject. There was a condition–distance interaction for RMS EMG (F4,24 = 5.801, P = 0.002, ηp2 = 0.492). Pairwise comparisons revealed a reduction in RMS EMG (as a % of MVC) in RMF compared with CON during 1 km (mean difference, 8% ± 5%; 95% CI, 4%–13%; t6 = 4.569, P = 0.004) and 2 km (mean difference, 6% ± 4%; 95% CI, 2%–9%; t6 = 3.937, P = 0.008), but no differences thereafter. There was also a condition–distance interaction for power output (F4,28 = 6.428, P = 0.001, ηp2 = 0.479). Pairwise comparisons revealed a reduction in power output in RMF compared with CON during 1 km (mean difference, 62 ± 52 W; 95% CI, 18–105 W; t7 = 3.374, P = 0.012) and 2 km (mean difference, 21 ± 16 W; 95% CI, 7–35 W; t7 = 3.611, P = 0.009), but no differences thereafter. Time to complete the 5-km cycling time trial was increased (i.e., worse performance) in all subjects in RMF compared with CON (464 ± 24 vs 455 ± 26 s), with a group mean increase of 9 ± 4 s (95% CI, 5–12 s; t7 = 5.625, P = 0.001; Fig. 3). The increased time to completion corresponds to a 1.9% ± 0.9% (range, 0.4%–3.8%) change in performance.

Motoneuronal output, power output, ratings of dyspnea, and ventilation. Vastus lateralis RMS EMG (A) and power output (B) during the 5-km cycling time trial. RMS EMG was normalized to RMS EMG during preexercise MVC. RMS EMG and power output data points represent the group mean value over the preceding 1-km segment. Ratings of dyspnea (C) and ventilation (V˙E; D) during the 5-km cycling time trial. Dyspnea was measured within the preceding 0.25 km of each data point, and V˙E was averaged over the preceding 0.5-km segment. Filled circles and solid lines are CON. Open circles and dashed lines are RMF. Data are presented as means ± SEM for clarity (n = 7 for RMS EMG and n = 8 for other variables). *Significantly different from CON (P < 0.05).
Time trial performance. A, Increase in time to complete 5-km cycling time trial (i.e., worse performance) after the resistive breathing task. Filled circles and solid lines are individual subjects. Open squares are group means ± SD (n = 8). #Significantly different from CON (P < 0.001). B, Boxplot of the percent change in performance from CON to RMF. Higher percent change indicates longer time to complete 5-km cycling during RMF condition (i.e., worse performance). Filled circles are individual subjects.

Physiological responses to cycling time trial

Physiological responses during the 5-km cycling time trial are displayed in Table 2. Multiple variables had a condition and distance interaction, with pairwise differences between CON and RMF during 1 and 2 km only. No pairwise differences were detected from kilometers 3 through 5. There was a condition–distance interaction for minute ventilation (V˙E; F4,28 = 5.691, P = 0.002, ηp2 = 0.448). Pairwise comparisons revealed differences between CON and RMF during 1 km (mean difference, 21 ± 13 L·min−1; 95% CI, 9–32 L·min−1; t7 = 4.569, P = 0.003) and 2 km (mean difference, 16 ± 12 L·min−1; 95% CI, 5–26 L·min−1; t7 = 3.595, P = 0.009). There was also a condition–distance interaction for VT (F4,28 = 7.741, P < 0.001, ηp2 = 0.525). Pairwise comparisons revealed differences between CON and RMF during 1 km (mean difference, 0.4 ± 0.3 L; 95% CI, 0.2–0.7 L; t7 = 3.872, P = 0.006) and 2 km (mean difference, 0.3 ± 0.2 L; 95% CI, 0.1–0.5 L; t7 = 3.981, P = 0.005). For breathing frequency, no interaction or main effect was detected. In addition, there was no interaction or main effect for operating lung volumes or inspiratory duty cycle (TI/TTOT).

TABLE 2 - Physiological responses during 5-km cycling time trial.
Distance, km
1 2 3 4 5
V˙E, L·min−1 135 ± 24 152 ± 31 157 ± 24 160 ± 22 170 ± 19
f b, breaths per minute 44 ± 12 48 ± 15 53 ± 13 54 ± 11 59 ± 9
V T, L 3.3 ± 0.8 3.2 ± 0.8 3.1 ± 0.7 3.0 ± 0.6 2.8 ± 0.5
 EILV, % FVC 85 ± 7 84 ± 4 83 ± 6 83 ± 5 82 ± 6
 EELV, % FVC 27 ± 8 26 ± 9 27 ± 6 30 ± 10 31 ± 11
T I/T TOT 0.47 ± 0.03 0.46 ± 0.04 0.48 ± 0.03 0.47 ± 0.02 0.48 ± 0.01
 V˙O2, mL·kg−1·min−1 57 ± 5 58 ± 6 60 ± 6 60 ± 6 60 ± 5
 V˙O2, L·min−1 4.4 ± 0.7 4.5 ± 0.7 4.6 ± 0.7 4.6 ± 0.5 4.6 ± 0.5
V˙CO2, L·min−1 5.0 ± 0.7 5.2 ± 0.8 5.0 ± 0.8 4.8 ± 0.6 4.8 ± 0.6
V˙E/V˙O2 31 ± 6 34 ± 6 34 ± 4 35 ± 3 37 ± 4
V˙E/V˙CO2 27 ± 4 29 ± 4 32 ± 4 33 ± 3 36 ± 1
 RER 1.13 ± 0.07 1.17 ± 0.05 1.09 ± 0.02 1.05 ± 0.02 1.05 ± 0.03
 V˙E, L·min−1 114 ± 19* 135 ± 23* 147 ± 19 156 ± 19 169 ± 23
f b, breaths per minute 41 ± 11 48 ± 14 50 ± 12 54 ± 11 57 ± 6
V T, L 2.9 ± 0.6* 2.9 ± 0.7* 3.0 ± 0.7 3.0 ± 0.6 2.9 ± 0.5
 EILV, % FVC 80 ± 5 80 ± 7 82 ± 6 85 ± 8 84 ± 5
 EELV, % FVC 29 ± 7 30 ± 9 29 ± 10 31 ± 10 33 ± 7
T I/T TOT 0.47 ± 0.04 0.46 ± 0.03 0.47 ± 0.03 0.47 ± 0.02 0.47 ± 0.02
 V˙O2, mL·kg−1·min−1 53 ± 5* 55 ± 5 59 ± 6 60 ± 6 62 ± 6
 V˙O2, L·min−1 4.1 ± 0.6* 4.2 ± 0.5* 4.5 ± 0.5 4.6 ± 0.5 4.7 ± 0.5
 V˙CO2, L·min−1 3.9 ± 0.5* 4.5 ± 0.5* 4.7 ± 0.6 4.7 ± 0.6 4.9 ± 0.7
V˙E/V˙O2 28 ± 4 32 ± 4 33 ± 3 34 ± 3 36 ± 3
V˙E/V˙CO2 29 ± 3 30 ± 4 31 ± 4 34 ± 4 35 ± 4
 RER 0.98 ± 0.07* 1.06 ± 0.04* 1.04 ± 0.03 1.02 ± 0.04 1.03 ± 0.06
Data are presented as means ± SD.
*Significantly different from CON at same distance (P < 0.05).
fb, breathing frequency; V˙E/V˙O2, ventilatory equivalent for oxygen; V˙E/V˙CO2, ventilatory equivalent for carbon dioxide.

There was a condition–distance interaction for absolute oxygen uptake (V˙O2; F4,28 = 9.370, P < 0.001, ηp2 = 0.572). Pairwise comparisons revealed differences between CON and RMF during 1 km (mean difference, 0.4 ± 0.3 L·min−1; 95% CI, 0.1–0.6 L·min−1; t7 = 3.850, P = 0.006) and 2 km (mean difference, 0.2 ± 0.2 L·min−1; 95% CI, 0.1–0.4 L·min−1; t7 = 3.404, P = 0.011). In addition, there was a condition–distance interaction for absolute carbon dioxide output (V˙CO2; F4,28 = 19.173, P < 0.001, ηp2 = 0.733). Pairwise comparisons revealed differences between CON and RMF during 1 km (mean difference, 1.0 ± 0.3 L·min−1; 95% CI, 0.8–1.3 L·min−1; t7 = 9.376, P < 0.001) and 2 km (mean difference, 0.7 ± 0.3 L·min−1; 95% CI, 0.4–1.0 L·min−1; t7 = 5.723, P = 0.001). The differences observed between absolute V˙O2 and V˙CO2 were reflected in an RER condition–distance interaction (F4,28 = 22.997, P < 0.001, ηp2 = 0.767), with pairwise comparisons revealing differences between CON and RMF during 1 km (mean difference, 0.15 ± 0.06; 95% CI, 0.10–0.20 L; t7 = 7.673, P < 0.001) and 2 km (mean difference, 0.11 ± 0.03; 95% CI, 0.08–0.14; t7 = 8.804, P < 0.001). There was no interaction or main effect for ventilatory equivalents of O2 or CO2.

Sensory perception

There was no interaction (F4,28 = 2.456, P = 0.07, ηp2 = 0.260) or main effect of condition (F4,28 = 0.704, P = 0.429, ηp2 = 0.091) for ratings of dyspnea (Fig. 2). Similarly, there was no interaction (F4,28 = 2.579 P = 0.06, ηp2 = 0.269) or main effect of condition (F4,28 = 3.066, P = 0.123, ηp2 = 0.305) for RPE.

Quadriceps muscle function

A near plateau (<5% increase in Qtw) was observed in each subject as stimulation intensity increased from 90% to 100%, and group mean Qtw at 100% was not different from 90% (P = 0.15) or 95% (P = 0.26), indicating the 100% stimulation intensity was supramaximal.

There were no differences in preexercise Qtw,pot between CON (177 ± 39 N) and RMF (173 ± 36 N; mean difference, 4 ± 12 N; 95% CI, −5 to 15 N; t7 = 1.092, P = 0.31). A large reduction in Qtw,pot was observed in both CON and RMF after the 5-km time trial; however, the preexercise to postexercise percent change was greater in CON (28% ± 7%) compared with RMF (21% ± 6%; mean difference, 7% ± 7%; 95% CI, 1%–12%; t7 = 3.006, P = 0.02; Fig. 4).

Neuromuscular function. Reductions in magnetically evoked potentiated quadriceps twitch force (Q tw, pot; A) and peak force generated during MVC (B) after 5-km cycling time trial. Filled circles and solid lines are individual subjects. Open squares are group means ± SD (n = 8). *Significantly different from CON (P < 0.05).

There were no differences in preexercise MVC between CON (543 ± 112 N) and RMF (535 ± 113 N; mean difference, 8 ± 30 N; 95% CI, −17 to 33 N; t7 = 0.738, P = 0.49). The preexercise to postexercise percent change in MVC was greater in CON compared with RMF CON (17% ± 9%) compared with RMF (10% ± 4%; mean difference, 8% ± 7%; 95% CI, 2%–14%; t7 = 3.101, P = 0.02; Fig. 4).

No differences in M-wave amplitude were observed preexercise to postexercise in CON (9.4 ± 0.3 vs 9.3 ± 0.3 mV, P = 0.29) or RMF (9.4 ± 0.3 vs 9.4 ± 0.3 mV, P = 0.29), suggesting unchanged membrane excitability.


Mean between-day, within-subject coefficients of variation were calculated for cycling performance using FAM and CON visits, and for preexercise Qtw,pot and MVC using CON and RMF visits, according to the logarithmic method. Mean between-day, within-subject coefficients of variation were 2.1% for cycling performance, 5.1% for Qtw,pot, and 4.1% for MVC, which are similar to published values from our laboratory and others (26,27).


The main findings from the present study are threefold. First, preexisting RMF caused alterations to the sensory perception of breathing such that ratings of dyspnea during RMF were similar with CON, although ventilation was lower in RMF compared with CON. Second, concomitant decreases in estimated motoneuronal output (assessed with surface EMG) and power output resulted in a slower time to complete the 5-km time trial in RMF compared with CON. Third, the degree of exercise-induced locomotor muscle fatigue was attenuated after the time trial with preexisting RMF. These findings suggest that altered ratings of dyspnea for a given ventilation, secondary to RMF, constrained motoneuronal output and power output during exercise. Furthermore, the reduction in power output observed during the RMF trial likely resulted in less metabolic perturbation within the locomotor muscles, thereby limiting the overall magnitude of exercise-induced locomotor muscle fatigue. Collectively, these findings suggest that the respiratory system influences the sensory tolerance limit and is an integral component in the regulation of exercise performance and the development of locomotor muscle fatigue.

Sensory perception

We used an inspiratory resistive breathing task to induce RMF before a cycling time trial and observed a substantial decrease in V˙E during the first 2 km of the RMF time trial, yet ratings of dyspnea were similar to the CON time trial. Thus, it seems that RMF caused alterations in dyspnea for a given level of ventilation. The alteration of perception due to RMF is consistent with other studies (3,15,17,18) and is likely caused by increased afferent feedback from fatigued respiratory muscles and changes in corollary discharge. The respiratory musculature contains group IV neurons with free nerve endings that respond to noxious levels of various metabolites (e.g., protons and lactate) that accumulate during diaphragm ischemia and fatigue (8,28–30). These afferents project into laminae I and II in the spinal cord and provide feedback to the somatosensory cortex (31). The information provided by afferent feedback is compared with feedforward corollary discharge and, if mismatched, results in changes to sensory perceptions of breathing (14). During short-duration, high-intensity exercise such as that undertaken in our study, extensive RMF does not occur during the early stages of exercise but instead develops toward the mid and latter portions of exercise as the cumulative work of breathing becomes considerable (32). Therefore, starting the RMF time trial with a substantial degree of RMF likely caused a mismatch between the corollary discharge and the actual afferent feedback, leading to increased dyspnea for a given level of V˙E, at least early in the time trial (1–2 km). However, as the time trial continued into the mid and latter portions of the RMF time trial (3–5 km), during which RMF normally occurs in our subject population, the magnitude of the mismatch presumably decreased, resulting in similar ratings of dyspnea along with similar ventilation.

There was also a significantly lower power output during the first 2 km of the RMF time trial, although RPE was similar to CON. These alterations to perception of effort are similar to other studies investigating the effects of preexisting muscle fatigue on subsequent exercise performance (11,18,33). Two of these studies used constant load exercise (11,33), whereas one study used self-paced exercise (18) comparable in nature to our 5-km time trial, which allowed subjects to freely vary their power output. Interestingly, the latter study observed changes to perception of effort only during the early stages of exercise, comparable to our results. The increase in RPE for a given level of power output early in exercise may be explained by the integration of both respiratory and locomotor muscle sensory information before the somatosensory cortex (14). The input of the exacerbated respiratory sensations along with any ongoing locomotor muscle afferent feedback could collectively cause an increase in perceived exertion. Similar findings of ensemble input from various sources that alters sensory perception have been demonstrated (11,12). We interpret these findings to suggest that sensory information, whether originating within the respiratory system or the locomotor system, influences perception of effort and contributes to a global sensory tolerance limit (13).

Motoneuronal output and power output

There was a concomitant decrease in quadriceps RMS EMG and power output during the first 2 km of the cycling time trial with preexisting RMF; however, no differences thereafter. Although changes in motoneuronal output and the regulation of power output during exercise can be influenced by group III/IV afferents originating in locomotor muscles (2,6), two observations suggest the locomotor afferents were not the exclusive determining factor in alterations during the time trial with preexisting RMF. First, preexercise quadriceps neuromuscular function was similar between RMF and CON, but the decline in quadriceps function after exercise was attenuated in RMF compared with CON. This skeletal muscle “reserve” capacity suggests a source other than the locomotor afferents in the regulation of motoneuronal output. Second, the differences in RMS EMG and power output between CON and RMF only occurred when there were also alterations to ratings of dyspnea for a given ventilation. As the differences in ratings of dyspnea diminished, the differences in EMG and power output between CON and RMF correspondingly diminished. Accordingly, we interpret the changes in EMG and power output to be largely determined by the alterations to sensory perception associated with RMF, as opposed to solely changes in locomotor muscle function.

Pharmacologically attenuating sensory afferents in the locomotor muscles increases motoneuronal output and power output, and can improve exercise performance (6,34,35). Our study was designed, as others have done (2,11), to do the opposite—to fatigue a muscle group before exercise to presumably increase the activity of the sensory afferents. Therefore, the observation of decreases in EMG and power output in RMF provides further evidence that sensory information related to muscle fatigue, independent of its origin, contributes to the regulation of motoneuronal output. Moreover, because we prefatigued the respiratory muscles, as opposed to the locomotor muscles, it seems that the respiratory system contributes to a sensory tolerance limit and is involved in a global negative feedback loop that regulates motoneuronal output and exercise performance (1,13).

Exercise performance

The observed decreases in power output during the first 2 km of RMF resulted in a ~2% longer time to complete the 5 km (i.e., worse performance). The magnitude of the change in performance caused by differences in power output only early in exercise is comparable to other studies (36,37). In regard to prefatiguing exercise, the magnitude of the observed change in performance is in agreement with some studies (2,18) but is much less than others (11,17,19,33). These discrepancies may be due to utilization of different exercise modes. Studies that reported large reductions in exercise tolerance (14%–38%) (11,17,19,33) used constant load exercise, whereas studies that reported more modest reductions in performance of 2% to 6% (2,18) used a self-paced exercise modality in which power output was free to vary, which is similar in design to our study. Ultimately, the differential effects of preexisting fatigue on constant load versus self-paced exercise requires future inquiry.

Locomotor muscle fatigue

Large reductions in the force development capability of the quadriceps occurred after both cycling time trials; however, the reduction was greater in CON compared with RMF, suggesting that the locomotor muscles did not meet their critical threshold in RMF and possessed a skeletal muscle reserve. The etiology of peripheral fatigue is complex, but evidence from isolated muscle fiber studies demonstrates that the accumulation of metabolites (e.g., Pi, H+, and Mg2+) is one of the mechanisms that contributes to declines in force production (38,39). We did not make any direct measurements of metabolite accumulation during exercise, and therefore, it is difficult to estimate its influence on locomotor muscle fatigue in our study. However, the average RER during the first 2 km of CON was ~1.15, whereas the average RER during the first 2 km of RMF was ~1.02, indicating a larger production of carbon dioxide in CON compared with RMF. As power output was also less during the RMF compared with CON, we judiciously speculate that less intramuscular metabolic perturbation occurred early in the RMF time trial and was manifested in less locomotor muscle fatigue after the completion of exercise.

Limitations and technical considerations

We chose to use changes in maximal inspiratory mouth pressure to assess RMF; however, we acknowledge that the accurate measurement of maximal mouth pressures can be affected by subject familiarization and motivation. To mitigate these effects, subjects were thoroughly familiarized to the maximal pressure testing procedure during the initial visit and strong verbal encouragement along with visual feedback of pressure generation was provided during measurement. We observed a reduction in maximal mouth pressure in all subjects during the resistive breathing task, resulting in a robust group mean decline in maximal pressure generation of 19% ± 9%. The magnitude of this change is in agreement with (40), who reported a ~24% decline in maximal inspiratory mouth pressure along with a reduction in transdiaphragmatic twitch pressures. In addition, the resistive breathing task we utilized was chosen purposefully, as it has been shown to induce diaphragm fatigue (3,25). Our study design did not allow for the measurement of the time course of recovery from fatigue; however, a similar resistive breathing task that induced diaphragm fatigue caused decrements in pressure-generating capacity of the muscle to last up to 30 min (25). As the cycling task in our study began ~8 min after cessation of the breathing task, it is unlikely that substantial recovery from fatigue occurred. Taken together, we are confident that the observed changes in maximal inspiratory mouth pressure occurred because of a substantial degree of RMF, and this fatigue of the respiratory muscles was present at the start of the cycling task.

The exclusion of women from this initial examination of RMF on time trial performance also warrants comment. Some evidence suggests that differences exist between men and women in diaphragm fatigability during inspiratory resistive breathing (33) and also during high-intensity exercise (33,41). Furthermore, based on recent evidence, it seems men and women have different perceptions of effort (e.g., dyspnea, RPE) during exercise with preexisting diaphragm fatigue (33) as well as differences in recruitment patterns of accessory muscles during high-intensity exercise (42). Based on these possible confounding variables, we elected to test only men for this initial examination of RMF on time trial performance. Thus, generalizing our findings to all populations is limited. Future studies should investigate the mechanisms underlying the sex-based differences mentioned previously, as well as potential sex-based differences in the sensory tolerance limit.

Estimating motoneuronal output from surface EMG at the target muscle has inherent limitations that have been expertly review elsewhere (43) and is not an exact measure of the neural drive to the locomotor muscles. We attempted to minimize some of the limitations by carefully recording the location of the electrodes at the initial visit by using triangulated measurements from anatomical landmarks as well as marking the location with indelible ink. We also normalized the RMS EMG activity during the cycling exercise to the maximal RMS EMG observed during the preexercise MVC. The observed differences in RMS EMG activity between CON and RMF occurred during the same portions of the time trial as the observed differences in power output between CON and RMF. In addition, toward the end of the time trial in each condition, subjects increased both power output and RMS EMG. Therefore, we interpreted these changes to mean that the cyclists were self-regulating their power output as part of a global feedback loop.


In conclusion, preexisting RMF alters sensory perception during subsequent whole-body exercise, leading to decreases in power output. The reduction in power output during whole-body exercise likely diminishes metabolic perturbation and attenuates locomotor muscle fatigue. These findings indicate that the respiratory system influences the sensory tolerance limit and is an integral component in a global feedback loop that regulates exercise performance and the development of locomotor muscle fatigue.

The authors thank Ben Thompson and Allison Barnette for their assistance with data collection.

T. J. F., M. N. B., H. L. P., D. M. K., T. D. M., and R. F. C. conceived and designed the experiments; T. J. F. and M. N. B. performed the experiments; T. J. F. and R. F. C. analyzed data and interpreted the results of experiments; T. J. F. and R. F. C. prepared the figures; T. J. F. and R. F. C. drafted the manuscript; T. J. F., M. N. B., H. L. P., D. M. K., T. D. M., and R. F. C. edited, revised, and approved the final version of the manuscript.

This study was not funded. The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results of this study do not constitute endorsement by the American College of Sports Medicine.


1. Gandevia SC. Spinal and supraspinal factors in human muscle fatigue. Physiol Rev. 2001;81(4):1725–89.
2. Amann M, Dempsey JA. Locomotor muscle fatigue modifies central motor drive in healthy humans and imposes a limitation to exercise performance. J Physiol. 2008;586(1):161–73.
3. Welch JF, Archiza B, Guenette JA, West CR, Sheel AW. Sex differences in diaphragmatic fatigue: the cardiovascular response to inspiratory resistance. J Physiol. 2018;596(17):4017–32.
4. Hureau TJ, Olivier N, Millet GY, Meste O, Blain GM. Exercise performance is regulated during repeated sprints to limit the development of peripheral fatigue beyond a critical threshold. Exp Physiol. 2014;99(7):951–63.
5. Rossman MJ, Garten RS, Venturelli M, Amann M, Richardson RS. The role of active muscle mass in determining the magnitude of peripheral fatigue during dynamic exercise. Am J Physiol Regul Integr Comp Physiol. 2014;306(12):R934–40.
6. Blain GM, Mangum TS, Sidhu SK, et al. Group III/IV muscle afferents limit the intramuscular metabolic perturbation during whole body exercise in humans. J Physiol. 2016;594(18):5303–15.
7. Garland S, Kaufman M. Role of muscle afferents in the inhibition of motoneurons during fatigue. Adv Exp Med Bio. 1995;384:271–8.
8. Light AR, Hughen RW, Zhang J, Rainier J, Liu Z, Lee J. Dorsal root ganglion neurons innervating skeletal muscle respond to physiological combinations of protons, ATP, and lactate mediated by ASIC, P2X, and TRPV1. J Neurophysiol. 2008;100(3):1184–201.
9. Hogan MC, Richardson RS, Haseler LJ. Human muscle performance and PCr hydrolysis with varied inspired oxygen fractions: a31P-MRS study. J Appl Physiol. 1999;86(4):1367–73.
10. Burnley M, Vanhatalo A, Fulford J, Jones AM. Similar metabolic perturbations during all-out and constant force exhaustive exercise in humans: a 31P magnetic resonance spectroscopy study. Exp Physiol. 2010;95(7):798–807.
11. Johnson MA, Sharpe GR, Williams NC, Hannah R. Locomotor muscle fatigue is not critically regulated after prior upper body exercise. J Appl Physiol. 2015;119(7):840–50.
12. Amann M, Venturelli M, Ives SJ, et al. Peripheral fatigue limits endurance exercise via a sensory feedback-mediated reduction in spinal motoneuronal output. J Appl Physiol. 2013;115(3):355–64.
13. Hureau TJ, Romer LM, Amann M. The ‘sensory tolerance limit’: a hypothetical construct determining exercise performance?Eur J Sport Sci. 2018;18(1):13–24.
14. Sheel AW, Foster GE, Romer LM. Exercise and its impact on dyspnea. Curr Opin Pharmacol. 2011;11(3):195–203.
15. Martin B, Heintzelman M, Chen H. Exercise performance after ventilatory work. J Appl Physiol. 1982;52(6):1581–5.
16. Dodd S, Powers S, Thompson D, Landry G, Lawler J. Exercise performance following intense, short-term ventilatory work. Int J Sports Med. 1989;10(1):48–52.
17. Mador MJ, Acevedo FA. Effect of respiratory muscle fatigue on subsequent exercise performance. J Appl Physiol. 1991;70(5):2059–65.
18. Verges S, Sager Y, Erni C, Spengler C. Expiratory muscle fatigue impairs exercise performance. Eur J Appl Physiol. 2007;101(2):225–32.
19. Wüthrich TU, Notter DA, Spengler CM. Effect of inspiratory muscle fatigue on exercise performance taking into account the fatigue-induced excess respiratory drive. Exp Physiol. 2013;98(12):1705–17.
20. Romer LM, Lovering AT, Haverkamp HC, Pegelow DF, Dempsey JA. Effect of inspiratory muscle work on peripheral fatigue of locomotor muscles in healthy humans. J Physiol. 2006;571(2):425–39.
21. Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general US population. Am J Respir Crit Care Med. 1999;159(1):179–87.
22. Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Respir J. 2005;26(2):319–38.
23. American Thoracic Society/European Respiratory Society. ATS/ERS Statement on respiratory muscle testing. Am J Respir Crit Care Med. 2002;166(4):518.
24. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14(5):377–81.
25. Peters CM, Welch JF, Dominelli PB, et al. Influence of inspiratory resistive loading on expiratory muscle fatigue in healthy humans. Exp Physiol. 2017;102(9):1221–33.
26. Paris HL, Fulton TJ, Chapman RF, Fly AD, Koceja DM, Mickleborough TD. Effect of carbohydrate ingestion on central fatigue during prolonged running exercise in moderate hypoxia. J Appl Physiol. 2018;126(1):141–51.
27. Romer LM, Haverkamp HC, Lovering AT, Pegelow DF, Dempsey JA. Effect of exercise-induced arterial hypoxemia on quadriceps muscle fatigue in healthy humans. Am J Physiol Regul Integr Comp Physiol. 2006;290(2):R365–75.
28. Graham R, Jammes Y, Delpierre S, Grimaud C, Roussos C. The effects of ischemia, lactic acid and hypertonic sodium chloride on phrenic afferent discharge during spontaneous diaphragmatic contraction. Neurosci Lett. 1986;67(3):257–62.
29. Hill JM. Discharge of group IV phrenic afferent fibers increases during diaphragmatic fatigue. Brain Res. 2000;856(1–2):240–4.
30. Pollak KA, Swenson JD, Vanhaitsma TA, et al. Exogenously applied muscle metabolites synergistically evoke sensations of muscle fatigue and pain in human subjects. Exp Physiol. 2014;99(2):368–80.
31. Jankowski MP, Rau KK, Ekmann KM, Anderson CE, Koerber HR. Comprehensive phenotyping of group III and IV muscle afferents in mouse. J Neurophysiol. 2013;109(9):2374–81.
32. Archiza B, Welch JF, Geary CM, Allen GP, Borghi-Silva A, Sheel AW. Temporal characteristics of exercise-induced diaphragmatic fatigue. J Appl Physiol. 2017;124(4):906–14.
33. Welch JF, Archiza B, Guenette JA, West CR, Sheel AW. Effect of diaphragm fatigue on subsequent exercise tolerance in healthy men and women. J Appl Physiol. 2018;125(12):1987–96.
34. Broxterman RM, Hureau TJ, Layec G, et al. Influence of group III/IV muscle afferents on small muscle mass exercise performance: a bioenergetics perspective. J Physiol. 2018;596(12):2301–14.
35. Hureau TJ, Weavil JC, Thurston TS, et al. Pharmacological attenuation of group III/IV muscle afferents improves endurance performance when oxygen delivery to locomotor muscles is preserved. J Appl Physiol. 2019;127(5):1257–66.
36. Aisbett B, Le Rossignol P, McConell GK, Abbiss CR, Snow R. Effects of starting strategy on 5-min cycling time-trial performance. J Sports Sci. 2009;27(11):1201–9.
37. Felippe LC, Ferreira GA, Learsi SK, Boari D, Bertuzzi R, Lima-Silva AE. Caffeine increases both total work performed above critical power and peripheral fatigue during a 4-km cycling time trial. J Appl Physiol. 2018;124(6):1491–501.
38. Westerblad H, Allen D. Myoplasmic free Mg2+ concentration during repetitive stimulation of single fibres from mouse skeletal muscle. J Physiol. 1992;453(1):413–34.
39. Sundberg CW, Hunter SK, Trappe SW, Smith CS, Fitts RH. Effects of elevated H+ and Pi on the contractile mechanics of skeletal muscle fibres from young and old men: implications for muscle fatigue in humans. J Physiol. 2018;596(17):3993–4015.
40. Sliwinski P, Yan S, Gauthier AP, Macklem PT. Influence of global inspiratory muscle fatigue on breathing during exercise. J Appl Physiol. 1996;80:1270–8.
41. Guenette JA, Romer LM, Querido JS, et al. Sex differences in exercise-induced diaphragmatic fatigue in endurance-trained athletes. J Appl Physiol. 2010;109(1):35–46.
42. Mitchell RA, Schaeffer MR, Ramsook AH, Wilkie SS, Guenette JA. Sex differences in respiratory muscle activation patterns during high-intensity exercise in healthy humans. Respir Physiol Neurobiol. 2018;247:57–60.
43. Farina D, Holobar A, Merletti R, Enoka RM. Decoding the neural drive to muscles from the surface electromyogram. Clin Neurophysiol. 2010;121(10):1616–23.


Copyright © 2020 by the American College of Sports Medicine