Muscle coordination adaptations to muscle fatigue have been extensively studied during intensity-matched submaximal tasks (1,2). By contrast, motor adaptations in response to muscle fatigue during maximal multijoint tasks have received much less attention. During such maximal, all-out exercises, decreased performance is classically observed early as fatigue occurs (3,4). To identify the muscles most affected by fatigue during a fatiguing sprint cycling exercise, studies have described either changes in EMG amplitude of the main lower limb muscles or changes in the power production at each joint (4,5). Although these studies observed a comparable decrease in muscle activation or power in specific joints, they have interpreted these adaptations differently. For Martin and Brown (3), the larger decrease in extension power at the ankle joint reported at the end of a 30-s sprint cycling exercise would be explained by a greater reduction in the force-generating capacity of the plantar flexors compared with the other muscles. For others (4,6), the related decreased activation of both the plantar flexors and the hamstrings with fatigue would also be a consequence of the decreased force-generating capacity of the quadriceps muscles. This interpretation is supported by the assumed primary roles of the biarticular muscles making (i) the observed decreased activation of gastrocnemii a possible consequence of a lower mechanical energy to be transferred from the fatigued quadriceps to the ankle joint (between-joint energy transfer principle) and (ii) the observed decreased activation of hamstring, a strategy to maintain an effective orientation of the pedal force (directional constraint principle) (7,8).
In the aforementioned studies (3,4,6), fatigue was not controlled in the sense that it was unknown which muscles were affected by fatigue and to what extent. In addition, both peripheral and central components might have been involved (9). These are critical issues because it is not possible to determine whether a change in the activation of a specific muscle occurred as a result of an actual decrease in its force-generating capacity and/or to maintain or optimize some biomechanical features of the task. Although the two control principles of biarticular muscle coordination (i.e., between-joint energy transfer and directional constraint of force application) are supported by simulation studies (7,10), there is a lack of experimental evidence. In particular, it remains unclear whether an energy transfer principle exists such that the central nervous system would decrease the drive to biarticular muscles as a result of the decreased force-generating capacity of associated main power–producer muscles. One original approach to address these issues would consist of inducing selective fatigue in one muscle group (11,12). Furthermore, no previous studies have quantified the effect of motor adaptations on pedal force orientation, making it difficult to experimentally test the directional constraint principle.
Assuming that the main agonist muscles are maximally activated during a maximal multijoint task, it is classically thought that the fatigue-induced decrease in force cannot be theoretically compensated by an increase in neural drive to these muscles or adjustments in motor coordination, leading to decreased performance. However, as mentioned above, the decreased activation of some biarticular muscles may not necessarily be considered detrimental. Moreover, although some extensor muscles (quadriceps, gastrocnemii) are maximally activated during an all-out sprint cycling task, other muscles, such as the hamstrings and hip flexors, are not (13). These muscles are not considered to be the main contributors to cycling power, but they do contribute significantly, notably to actively pull on the pedal in the upstroke phase (3,14). It is unclear whether they can maintain their activity and exhibit compensatory strategies to partly counteract the decreased force produced by the main agonists.
In the present study, we induced selective fatigue of the quadriceps muscle group. A neuromechanical approach (EMG, pedal force, and joint power) was used to describe motor coordination during a sprint cycling task. We aimed to investigate whether a reorganization of the neural command to fatigue-free muscles would occur and whether such a reorganization may act to maintain some mechanical characteristics of the task such as mechanical effectiveness. In accordance with the assumed roles of biarticular muscles in energy transfer between joints and directional constraints of force application (4,6,8,10), we hypothesized that a significant decrease in the activation level of biarticular muscles would occur in response to decreased power at the knee. We also predicted that flexor muscles, which are not initially activated at their maximal level, would exhibit an increased activation, which could be considered as a positive adaptation to limit the decrease in performance.
Fifteen active and healthy male volunteers participated in this study (mean ± SD: age 23.7 ± 3.3 yr, height 180.1 ± 8.1 cm, body mass 72.7 ± 8.9 kg). They had no prior cycling training experience and no history of lower limb injury. The experiment was approved by the local ethics committee (CPP Ouest V: n°2013-A01714-41) and was conducted according to the Declaration of Helsinki. The participants provided their written informed consent.
Participants performed a series of submaximal and all-out sprint pedaling tasks before (control condition) and after (fatigue condition) selective fatigue was induced in the knee extensors of one leg using an electromyostimulation protocol (Fig. 1).
As previously described in detail (15), participants sat on an isokinetic dynamometer (Biodex System 3 research; Biodex Medical, Shirley, NY) with their hip and knee flexed at 90° and 80°, respectively (0° being full extension of the hip and knee). The torso and the waist were strapped to the chair. Transcutaneous electromyostimulation was applied to the quadriceps muscle group of one leg (side randomized; right side n = 8 and left side n = 7) with the intent of fatiguing the whole muscle group. Electrodes (Stimex 50 × 90 mm, Monath Electronic, Rouffach, France) were placed over the motor points of the vastus lateralis (VL), vastus medialis (VM), and rectus femoris (RF). A fourth electrode (Stimex 80 × 130 mm, Monath Electronic) was placed over their proximal insertion (anode). A constant current stimulator (DS7A; Digitimer, Letchworth Garden City, UK) coupled with a train/delay generator (DG2A, Digitimer) was used to deliver a train of rectangular pulses (pulse duration = 450 μs) at 70 Hz within 3 s. Each stimulation was followed by a 3-s rest. After 15 min of stimulation (i.e., 150 contractions), a single isometric maximum voluntary contraction (MVC) was performed to assess loss of strength in the quadriceps. If the maximal torque during MVC did not decrease by ≥30% of the MVC performed before the fatigue protocol, a final set of 50 contractions was induced. The stimulation intensity was adjusted throughout the protocol to match the maximal tolerable level. Fatigue was characterized through the assessment of MVC torque, voluntary activation (VA) level, resting twitch (doublet stimulus), and M-wave before and after this protocol. To confirm that fatigue persisted in the leg throughout the experiment, two additional MVC were performed at the end of all procedures. To check that fatigue was absent in the contralateral leg, participants performed two maximal knee extensions before the stimulation protocol and at the end of all procedures.
The pedaling exercises were performed on an electronically braked cycle ergometer (Excalibur Sport; Lode, Groningen, the Netherlands). Before data collection, the positions of the handlebar and saddle were adjusted to fit the preferences of the participants. After a standardized warm-up (10 min pedaling at 100 W followed by 1 min 30 s at 250 W), participants performed two submaximal tasks of 1 min duration at 350 W and two 5-s all-out sprints at a fixed cadence of 90 rpm (isokinetic mode) while in a seated position. The order of the cycling bouts was randomized to avoid any confounding effects related to the repetition of the exercises. In addition, rest periods were provided after both submaximal pedaling (3 min) and all-out sprints (4 min). Note that the two 350-W bouts were performed to address a different research question (15) and were therefore not considered for the present study. The same pedaling tasks were then repeated within a period <3 min after the fatigue protocol. This period corresponds to the time needed to perform the neuromuscular test and to change position from the single-joint to the cycle ergometer. Only one all-out sprint was performed during the fatigue condition to avoid additional fatigue. The 2D pedal forces, lower limb kinematics, and myoelectrical activity of 11 muscles (bilaterally) were recorded during the pedaling tasks.
Materials and Procedures
A single stimulus was delivered to the femoral nerve to elicit the maximal M-wave at rest, and a doublet stimulus (100 Hz) was delivered during the plateau of MVC and within 5 s in the following rest period to elicit a superimposed and resting twitch, respectively. MVC, VA, M-wave, and resting twitch were measured using the torque and the EMG signal of VL and VM. Because these data have been reported in a previous article (15), they are not presented in detail herein, but they are discussed in the Discussion section.
The ergometer was equipped with instrumented pedals to measure 2D forces (VélUS group, Canada; for details, see Dorel et al., 2009). Briefly, Cartesian components of the force on the pedal corresponding to the horizontal and vertical components were measured during the cycling trials. To create a solid shoe–pedal interface, the pedals had a clipless configuration compatible with LOOK KEO cleats. All data from the instrumented pedals were digitized at 1000 Hz with a Mega data logger (ME6000; Mega Electronics Ltd., Linton, Cambridge, UK).
Three-dimensional kinematic data were recorded using an optoelectronic motion capture system composed of nine cameras (Flex13, 1.3 Mpx, OptiTrack, Natural Point). Twenty-eight retroreflective markers (diameter, 10 mm) were attached to the skin with double-sided tape on relevant anatomical landmarks bilaterally (i.e., the anterior–superior iliac spine, posterosuperior iliac spine, greater trochanter region, thigh [5 cm above the center of the patella and located with the knee extended], lateral and medial femoral condyles, tibial tuberosity and distal tibia [20 cm below the center of the patella], lateral and medial malleoli, posterior facet of the calcaneus, first and fifth metatarsophalangeal joints, and front tip of the cycling shoe). Additional markers were placed on the frame of the ergometer and on both pedals.
The kinematic data were recorded at 100 Hz. Body fat was estimated and segment length/circumference was measured (16), and both were included in the inverse dynamics model. The estimation of body fat was based on skinfold measurements from the biceps, triceps, subscapular, and suprailiac regions (17). At the beginning of the experimental session, three kinematic setup bouts were recorded. First, a static bout was recorded while the participants maintained an upright standing position. The participants then performed four hip extensions/flexions, four hip abductions/adductions, and four circumductions to locate both the right and left hip joint centers (18). Afterward, six reflective markers were removed bilaterally (medial malleolus, medial femoral condyles, and first metatarsophalangeal joint) to enable the participants to pedal on the ergometer without discomfort. EMG, force, and kinematic signals were synchronized using a trigger signal from the optoelectronic motion capture system.
Myoelectrical activity of 11 muscles was recorded bilaterally using surface EMG electrodes (diameter of the recording area, 5 mm; Covidien, Kendall; interelectrode distance of 20 mm) placed over the tibialis anterior (TA), soleus (SOL), gastrocnemius lateralis (GL), gastrocnemius medialis (GM), VM, VL, RF, biceps femoris (BF), semitendinosus (ST), gluteus maximus (GMAX), and tensor fascia latae (TFL). Before electrode application, the skin was shaved, abraded, and cleaned to reduce impedance. All cables and electrodes were well secured to the skin using adhesive tape. The EMG signals were preamplified close to the electrodes (×1000) and digitized at 1000 Hz using two synchronized EMG amplifier units (ME6000, Mega Electronics Ltd.).
All data were processed in MatLab R2013a (The Mathworks Inc.) using custom-written scripts. The force and kinematic signals were low-pass filtered at 10 Hz with a second-order Butterworth filter. The effective force (the force applied perpendicular to the crank) was derived using trigonometry using 2D pedal forces and pedal angle in the sagittal plane measured by reflective markers placed on the pedal and a potentiometer. The index of mechanical effectiveness (IE) was calculated as the ratio between the effective force and the norm of the total force applied to the pedal. The external power generated at each pedal (pedal power) was calculated as the product of effective force and the norm of the velocity of the pedal. A biomechanical model consisting of 24° of freedom was created. Joint centers were determined using the SCoRE algorithm (19), and joint angles were described using an X–Y–Z Cardan sequence, following the recommendations of the International Society of Biomechanics. The local 3D marker coordinates in the biomechanical model were used to estimate marker positions in the global bike frame by way of a forward kinematic function and an extended Kalman filter (20). The 2D joint reaction forces and net torque for the ankle, knee, and hip were derived using a conventional inverse dynamics method (for details, see ). The specific joint power at each joint was calculated as the dot product of the joint moment force and joint angular velocity.
The raw EMG data were band-pass filtered (bandwidth 10–500 Hz), root mean squared (RMS, 25-ms window), and low-pass filtered at 24 Hz (second-order Butterworth filter). All EMG data and data issued from the inverse kinematics and dynamics methods were resampled to obtain one value for every 5° of crank displacement. Pedaling cycles were detected using transistor–transistor logic rectangular pulses at the highest position of the right pedal (top dead center). The five most powerful pedaling cycles (excluding the first one) then were extracted from each sprint trial and averaged to obtain representative profiles of (i) pedal power, (ii) index of effectiveness, (iii) muscle EMG amplitude, and (iv) specific joint power. To provide a deeper understanding of the overall coordination strategy, data were also averaged in each phase (i.e., extension and flexion of the lower limb). On the basis of classical data reported in the literature (13,21), extension was defined as a crank angle between 340° and 160° and flexion as between 160° and 340°; with 0° indicating the vertical position of the crank. For the sake of clarity, EMG data are only reported for the phase in which the muscle under consideration was active (one phase for monoarticular muscles, and two phases for biarticular muscles).
Statistical tests were performed using STATISTICA (V8; StatSoft, Inc., Tulsa, OK). The Shapiro–Wilk test was used to test for a normal distribution. Data violating this criterion were transformed depending on the skew (logarithmic, square root, or reciprocal transformation). A repeated-measures ANOVA was used to test the effect of the fatiguing protocol on the MVC torque of the fatigued leg (within-subject factor: time [control, fatigue, and end of the protocol]). A paired samples t-test was used to compare the control and end of protocol MVC torque values of the nonfatigued leg. A paired samples t-test was also used to compare the control and immediately postfatigue protocol M-wave, resting twitch, and VA level of the fatigued leg.
For the pedaling task, a paired samples t-test was used to determine whether the total pedal power output (sum of both legs) was affected by condition (control and fatigue). A two-way repeated-measures ANOVA was used to determine the effects of legs and conditions on pedal power and IE averaged over the full pedaling cycle (within-subject factors: leg [fatigued and nonfatigued] and condition [control and fatigue]). Post hoc analyses were performed using the Bonferroni adjustment for multiple comparisons. To further determine the effect of fatigue on muscle coordination and where the alterations occurred during the crank cycle, paired samples t-tests were used to compare data averaged over the extension and flexion phases (pedal power, IE, EMG activity, and specific joint power) between control and fatigue conditions. An additional paired sampled t-test was performed specifically for the average positive and negative knee joint power over the extension phase between control and fatigue conditions. The significance level was set to P < 0.05. All results presented below are presented as mean ± SD.
Data have been reported in detail in a previous article (15). Briefly, and in relation to the specific aim of this study, knee isometric extension MVC torque produced by the fatigued leg decreased by 28.0% ± 6.8% (P < 0.001) immediately after the fatigue protocol and remained at this low level at the end of the entire protocol (−21.1% ± 10.5%; P = 0.006 compared with control). At the end of the fatigue protocol, resting twitch amplitude decreased by 24.8% ± 23.4% (P = 0.032), but VA was unchanged (P = 0.869). The MVC torque produced by the nonfatigued leg was slightly decreased at the end of the protocol when compared with control (−5.9% ± 8.2%; P = 0.020).
Pedal Power and Index of Effectiveness
The total pedal power output (sum of both legs) during the cycling sprint was significantly reduced after the fatigue protocol (control, 877.2 ± 132.5 W; fatigue, 851.9 ± 115.9 W; P = 0.033). We observed a significant leg–condition interaction (P < 0.001) on the power output averaged over the whole crank cycle (Table 1; Fig. 2). The power produced by the fatigued leg was lower during the fatigue condition than control (−25.5 ± 22.5 W; P < 0.001), whereas the power output produced by the nonfatigued leg did not change significantly (P = 1.0). Further analysis showed that the power produced by the fatigued leg during the extension phase significantly decreased (−66.9 ± 34.3 W; P < 0.001) during the fatigue condition, whereas the power produced during the flexion phase increased (+19.5 ± 21.9 W; P = 0.004). The power output produced by the nonfatigued leg during its extension phase was not significantly altered (P = 0.195) during fatigue compared with control, and the pedal power during flexion significantly increased during fatigue (+17.9 ± 28.3 W; P = 0.026).
There was neither a main effect of leg (P = 0.400) nor a leg–condition interaction (P = 0.07) on IE averaged over the full pedaling cycle. However, a main effect of condition (P < 0.001) was observed, showing a higher IE (fatigued leg, +6.0% ± 5.7%; nonfatigued leg, +10.2% ± 9.7%) during fatigue than during control. Inspection of each phase of the pedaling cycle revealed that the IE of the fatigued leg did not change (P = 0.643) during the extension phase, but it significantly increased during flexion in fatigue compared with control (+12.7% ± 10.4%; P < 0.001). For the nonfatigued leg, IE significantly increased during both the extension phase (+2.9% ± 3.3%; P = 0.004) and the flexion phase (+17.9% ± 18.9%; P = 0.003).
When considering the extension phase, the hip joint power was lower during fatigue than control (−30.1 ± 37.8 W; P = 0.008). No change was observed during the flexion phase (P = 0.068) (Table 1; Fig. 3). The mean power produced by the knee joint during the extension phase was not affected by fatigue (P = 0.236). Of note, during this phase, the knee joint power exhibited both positive and negative components (Fig. 3); the positive component significantly decreased (−34.4 ± 30.6 W; P < 0.001), whereas the power in the negative component increased (i.e., became less negative, +19.3 ± 23.8 W; P = 0.007). The knee joint power produced during the flexion phase remained unaltered by fatigue (P = 0.546) (Table 1; Fig. 3). The ankle joint power produced during the extension phase significantly decreased during fatigue (−20.8 ± 18.7 W; P = 0.001), but no changes were observed during the flexion phase (P = 0.469).
When considering the extension phase, the hip joint power was lower during fatigue than control (−19 ± 27.2 W; P = 0.017; Table 1; Fig. 3). By contrast, during the flexion phase, the hip joint power was higher during fatigue than control (+31 ± 25.2 W; P < 0.001). The knee joint power did not change during the extension phase in the fatigue condition (P = 0.158). The positive component of the knee extension power also did not change, but the power in the negative component increased significantly (i.e., became less negative, +9.9 ± 17.8 W; P = 0.048). When considering the flexion phase, the knee joint power did not change significantly during fatigue despite a tendency to decrease (−15 ± 27 W; P = 0.051). The ankle joint power produced during fatigue was lower during the extension phase (−11.8 ± 15.5 W; P = 0.011) and remained unchanged during the flexion phase (P = 0.737).
When considering the extension phase of the fatigued leg, we observed a decrease in EMG amplitude during fatigue compared with control for all muscles (range, −5.8% to −19.2%; all P values < 0.046), except for BF (P = 0.090; Table 2; Fig. 4). During the flexion phase, only the RF muscle exhibited a decrease in EMG amplitude during fatigue (P < 0.001).
When considering the extension phase of the nonfatigued leg, a decrease in EMG amplitude was observed during fatigue for some muscles (range, −6.9% to −19.5%; VM, GMAX, BF, ST, GL, and SOL [all P values < 0.027]), but not all (VL, RF, and GM; all P values >0.102; Table 2 and Fig. 4). During the flexion phase, the EMG amplitude of both BF (P = 0.004) and ST (P = 0.038) was lower during fatigue than control. The EMG amplitude of RF, GM, GL, and TA remained unchanged (all P values > 0.103). Only the activity level of TFL increased significantly during this phase (+27.5% ± 37.6%; P = 0.015).
The present results highlight two major adjustments of motor coordination that occurred during an all-out sprint pedaling task when the force-generating capacity of the quadriceps of one leg was experimentally decreased. First, in the fatigued leg, the reduction in positive knee extension power during the extension phase was accompanied by decreased activation of the biarticular hamstrings and gastrocnemii as well as the monoarticular GMAX and SOL muscles. These adaptations led to a decrease in hip and ankle extension power. Although they lead to the observed decrease in pedal power, these adaptations help in maintaining the net knee joint power and mechanical effectiveness at the pedal during this powerful phase. Second, a significant increase in both muscle activation and pedal power was observed during the flexion phase of both legs. Interestingly, the adjustments observed in the nonfatigued muscles of both legs were associated with improved mechanical effectiveness.
Adaptations of synergist/antagonist muscles within the fatigued leg
As detailed elsewhere (15), the electromyostimulation protocol was effective in inducing substantial peripheral impairment in the force-generating capacity of the quadriceps muscles. This led to a decrease in knee extension positive power (−34.4 ± 30.6 W) during the subsequent sprint pedaling. The unchanged maximal VA of the quadriceps during MVC confirmed the absence of central fatigue. Because the presence of fatigue makes EMG amplitude difficult to interpret (22,23), VL and VM EMG amplitudes were normalized to the maximal M-wave amplitude (Mmax) to better represent the neural drive to these muscles (24). Interestingly, the EMG/Mmax did not significantly change during the sprint performed in the fatigue condition (P = 0.31 and P = 0.82 for VL and VM, respectively). These findings suggest that peripheral quadriceps fatigue was responsible for the observed reduction in knee extension power during the sprint pedaling task, preventing the muscle from responding to an identical central motor drive by the same extent.
Along with the decrease in power produced by the knee extensor muscles, muscle activity decreased for most of the muscles involved in this extension phase. This included both the monoarticular (GMAX and SOL) and the biarticular muscles (ST, GM, and GL) and logically resulted in decreased hip and ankle power. In a recent study, O’Bryan et al. (25) reported similar results during an all-out sprint cycling performed immediately after a high-intensity cycling exercise inducing a knee-extensor fatigue. However, fatigue was not controlled in the latter study, making it complicated to explain the origin of these changes, i.e., increased activity of the group III and IV afferences from the fatigued quadriceps muscles or the fatigue directly induced on these muscles themselves. By inducing local fatigue in the knee extensors, the present study represents the first experimental evidence that the drive to some unfatigued synergist and antagonist muscles may be reduced in response to the decreased force-generating capacity of distant muscles. In the absence of global neural inhibition due to central fatigue mechanisms, these results are in accordance with our first hypothesis that selective peripheral fatigue would result in a reorganization of motor control strategy through reduced recruitment of unfatigued muscles.
Consequent to the decreased knee extension power, less mechanical energy needed to be transferred from the knee toward the ankle joint and from the limb to the crank. In line with previous works suggesting that calf muscles play a key role in this energy transfer, this might explain the decreased activation of gastrocnemii and SOL (7,8,26). Similarly, decreased activity of the monoarticular GMAX muscle and decreased hip extension power could be explained by the altered capacity of fatigued RF muscle to transfer hip extension torque to the knee. As a whole, these adaptations provide experimental evidence of the role of the mechanical energy transfer principle in the control of muscle coordination during a multijoint task (8,10). Moreover, during a cycling task, joint torque produced by each muscle group not only participates in the magnitude of total pedal force but also contributes to the ability to effectively orientate this force perpendicularly to the crank. Interestingly, decreased activation of the hamstrings was observed in the phase where they were coactivated with the quadriceps (i.e., 0°–120° in the extension phase) but not during the beginning of the flexion phase (Table 2; Fig. 4). This finding confirms previous studies suggesting that a decrease in hamstrings activity during a fatiguing repeated sprint cycling task would result from a loss of force produced by the quadriceps (6). We believe that this strategy of decreasing synergist and antagonist activity (particularly the biarticular muscles) is needed to maintain efficient force orientation on the pedal, as previously suggested by others (10,27). Our hypothesis is supported by an absence of change in the index of effectiveness, and therefore in the orientation of the total force on the pedal, between control and fatigue conditions.
Adaptations in other muscles in both legs
No change was detected in mean pedal power for the nonfatigued leg over the full pedaling cycle. This is in accordance with previous results showing an unchanged maximal pedal power for the contralateral leg after a fatiguing single-leg pedaling task, which suggests that no crossover fatigue effect occurred between legs (28). Nevertheless, we observed a significant increase in power during the flexion phase of this nonfatigued leg (Table 1). This was mainly explained by an increase of about 31 W in the hip-specific joint power (associated with increased TFL activity), resulting in a larger relative contribution of the hip to the total pedal power produced during this flexion phase (from 4.8% during control to 20.1% during fatigue). Likewise, a significant increase in pedal power was observed during the flexion phase for the fatigued leg (~20 W). Similar compensations were reported during a submaximal cycling exercise (15,29). Interestingly, this increase in activation confirms that even during a brief all-out cycling exercise, not all muscles are systematically maximally activated (13). Thus, the current finding supports our second hypothesis that, despite an observable global decrease in performance, the nervous system is able to increase the neural drive to some flexor muscles, which can partly compensate for the loss of power produced by a fatigued muscle group.
Previous studies suggest that between-leg adaptations could be mediated via interlimb neural pathways and/or could originate from mechanical coupling at the crank level (15,30,31). In the present study, because of the use of an isokinetic cycling mode, mechanical coupling at the crank level certainly would have been limited, which means that the decrease in force produced by the fatigued leg at the pedal might not have been directly detected by the nonfatigued leg at the contralateral pedal. Thus, it is reasonable to argue that, owing to interlimb neural pathways, the sensorimotor activity of the fatigued quadriceps would have directly affected the contralateral mechanics by increasing activity in the nonfatigued leg muscles involved in the flexion phase. By contrast, the slight decreased activity of some muscles in the nonfatigued leg (especially the hamstrings) further confirms that interlimb neural coupling is a complex mechanism and may act in a different way (31,32). In response to the fatigue state of the biarticular RF muscle involved in the posteroanterior transition of the pedal, the interlimb neural coupling would participate in decreasing the activation of the nonfatigued BF and ST muscles involved in the concomitant anteroposterior transition of the opposite pedal (31,32). Moreover, previous studies on pedaling tasks showed that the central pattern generators of each leg may share a common generator or at least may be modulated via interneuronal connections (33,34). Although unexpected, all these neural adaptations observed in the nonfatigued leg provide further evidence that complex neural connections between legs are involved and allow for transferring information to coordinate between lower-limb movements during the pedaling task (31,35).
By showing that the activation of a muscle depends not only on its sensorimotor status but also on the status of distant muscles, our results indicate that it is questionable to interpret a decrease in muscle activity during a maximal fatiguing multijoint task as evidence of a reduction in force-generating capacity. Moreover, the question remains as to whether the adaptations observed in response to muscle fatigue should be considered detrimental or beneficial for the overall performance. Contrary to the submaximal exercise, for which several optimization criteria of muscle coordination have been proposed (i.e., minimization of cost functions such as effort, neural activity, and/or error variability [10,36]), this question is far less complicated during a brief all-out sprint. In this case, the motor control “challenge” can be reduced to the primary criterion of maximizing total pedal power. On one hand, the decreased activation of gastocnemii, SOL, and GMAX of the affected leg led to an inability to maintain extension power at the ankle and hip. It therefore participated in decreasing the pedal power (i.e., −50.9 W). Furthermore, the small decrease in ankle and hip joint power also observed in the nonfatigued leg may be considered unfavorable by making it critical to maintain pedal power during the extension. On the other hand, several motor adaptations are beneficial for overall power output and, thus, based on this criterion, can be interpreted as a partial “optimization” of muscle coordination. First, the aforementioned global decreased activation of antagonist muscles during extension reduces the negative (i.e., flexion) knee joint power (+19.3 ± 23.8 W) and, thus, is an appropriate neural strategy for both the maintaining net knee power and the mechanical effectiveness of the fatigued leg and increasing the net knee power (tendency) and pedaling effectiveness of the nonfatigued leg (+3%, Table 1). This result is important because the ability to effectively orientate force is known to be a determinant of power production during an all-out sprint (14,37).
In addition, the increase in activation of hip flexor muscles led to an improvement of pedaling effectiveness and to increased power production in the flexion phase for both legs, suggesting that this upstroke phase was initially suboptimal. Interestingly, muscle fatigue was often demonstrated to impair intersegmental or interlimb coordination (38), which is traditionally considered to be a drawback to learning a new skill or working to improve the effectiveness of a movement. The present results might have practical applications. They suggest that prefatiguing extensor muscles could be used as a strategy to increase recruitment of flexor muscles, and therefore to partially adjust muscle coordination to improve the power output in a specific phase of the movement. Further studies are needed to determine (i) if other positive adaptations would be induced by fatiguing other muscle groups and (ii) if these adaptations would also occur in a population of well-trained sprint cyclists. To support the notion that local fatigue may represent an interesting paradigm for enhancing performance during such a task, a promising approach would be to determine whether these locomotor adaptations would be transferred to normal condition after a training period.
Finally, it is interesting to note that although a global decrease in activation was observed for almost all muscles (significant for 14 of 18 muscles involved in the extension phase in both legs; Table 2), only a small decrease in total cycling power was observed (−2.8%; P = 0.03). By considering total EMG intensity as a proxy for the metabolic power required for pedaling, some authors recently proposed to estimate a relative efficiency, calculated as the ratio of mechanical power output measured at the pedals to the sum of the total EMG activity levels across all muscles (39,40). In the current study, this ratio was clearly improved (i.e., an approximately 6% increase in both legs between control and fatigue conditions), which ultimately suggests better “neuromuscular efficiency” in the fatigue condition.
In conclusion, the current study shows that in response to local fatigue of one muscle group, the activity of all the coactivated synergist and antagonist muscles and the joint-specific power at adjacent joints are decreased. A decrease in EMG activity or joint-specific torque produced by one particular muscle group is then not necessarily related to a reduction in its force-generating capacity during a global fatiguing multijoint task. Alternatively, despite a slight decrease in activity for some muscles in the nonfatigued leg, positive adaptations occur by increasing the neural drive to flexor muscles and increasing the power produced by both legs in the flexion phase. These findings provide evidence that in response to peripheral fatigue, the nervous system modulates the coordination strategy (i) by decreasing the activation of nonfatigued synergist and antagonist muscles to ensure an effective orientation of pedal force and (ii) by increasing the neural drive in the direction of muscles involved in other specific parts of the task to maximize the power produced and then limit the decrease of total power at the crank level.
Project support was provided by the Region Pays de la Loire (ANOPACy project) and the French Ministry of Sport (14-R-23). Francois Hug was supported by a fellowship from the Institut Universiatire de France (IUF). The authors are grateful to the subjects for having agreed to participate in this study.
The authors report no conflict of interest. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
1. Enoka RM, Duchateau J. Muscle fatigue: what, why and how it influences muscle function. J Physiol
2. Bigland-Ritchie B, Woods JJ. Changes in muscle contractile properties and neural control during human muscular fatigue. Muscle Nerve
3. Martin JC, Brown NA. Joint-specific power
production and fatigue during maximal cycling. J Biomech
4. O’Bryan SJ, Brown NA, Billaut F, Rouffet DM. Changes in muscle coordination and power output during sprint cycling. Neurosci Lett
5. Hunter AM, St Clair Gibson A, Lambert MI, Nobbs L, Noakes TD. Effects of supramaximal exercise on the electromyographic signal. Br J Sports Med
6. Hautier CA, Arsac LM, Deghdegh K, Souquet J, Belli A, Lacour JR. Influence of fatigue on EMG/force ratio and cocontraction in cycling. Med Sci Sports Exerc
7. Raasch CC, Zajac FE. Locomotor strategy for pedaling: muscle groups and biomechanical functions. J Neurophysiol
8. van Ingen Schenau GJ, Boots PJ, de Groot G, Snackers RJ, van Woensel WW. The constrained control of force and position in multi-joint movements. Neuroscience
9. Fernandez-del-Olmo M, Rodriguez FA, Marquez G. Isometric knee extensor fatigue following a Wingate test: peripheral and central mechanisms. Scand J Med Sci Sports
10. Prilutsky BI. Coordination of two- and one-joint muscles: functional consequences and implications for motor control
. Motor Control
11. Bouillard K, Jubeau M, Nordez A, Hug F. Effect of vastus lateralis fatigue on load sharing between quadriceps femoris muscles during isometric knee extensions. J Neurophysiol
12. de Rugy A, Loeb GE, Carroll TJ. Muscle coordination is habitual rather than optimal. J Neurosci
13. Dorel S, Guilhem G, Couturier A, Hug F. Adjustment of muscle coordination during an all-out sprint cycling task. Med Sci Sports Exerc
14. Dorel S, Couturier A, Lacour JR, Vandewalle H, Hautier C, Hug F. Force-velocity relationship in cycling revisited: benefit of two-dimensional pedal forces analysis. Med Sci Sports Exerc
15. Brochner Nielsen NP, Hug F, Guevel A, Fohanno V, Lardy J, Dorel S. Motor adaptations to unilateral quadriceps fatigue during a bilateral pedaling task. Scand J Med Sci Sports
16. Zatsiorsky V, Seluyanov V. Estimation of the mass and inertia characteristics of the human body by means of the best predictive regression equations. Biomechanics IX-B
17. Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr
18. Begon M, Monnet T, Lacouture P. Effects of movement for estimating the hip joint centre. Gait Posture
19. Ehrig RM, Taylor WR, Duda GN, Heller MO. A survey of formal methods for determining the centre of rotation of ball joints. J Biomech
20. Fohanno V, Begon M, Lacouture P, Colloud F. Estimating joint kinematics of a whole body chain model with closed-loop constraints. Multibody Syst Dyn
21. Elmer SJ, Barratt PR, Korff T, Martin JC. Joint-specific power
production during submaximal and maximal cycling. Med Sci Sports Exerc
22. Keenan KG, Farina D, Merletti R, Enoka RM. Influence of motor unit properties on the size of the simulated evoked surface EMG potential. Exp Brain Res
23. Farina D, Merletti R, Enoka RM. The extraction of neural strategies from the surface EMG. J Appl Physiol (1985)
24. Hureau TJ, Ducrocq GP, Blain GM. Peripheral and central fatigue development during all-out repeated cycling sprints. Med Sci Sports Exerc
25. O’Bryan SJ, Billaut F, Taylor JL, Rouffet DM. Knee extensor fatigue developed during high-intensity exercise limits lower-limb power production. J Sports Sci
26. Fregly BJ, Zajac FE. A state-space analysis of mechanical energy generation, absorption, and transfer during pedaling. J Biomech
27. Neptune RR, Kautz SA, Zajac FE. Muscle contributions to specific biomechanical functions do not change in forward versus backward pedaling. J Biomech
28. Elmer SJ, Amann M, McDaniel J, Martin DT, Martin JC. Fatigue is specific to working muscles: no cross-over with single-leg cycling in trained cyclists. Eur J Appl Physiol
29. Hansen EA, Voigt M, Kersting UG, Madeleine P. Frequency and pattern of rhythmic leg movement in humans after fatiguing exercises. Motor Control
30. Alibiglou L, Lopez-Ortiz C, Walter CB, Brown DA. Bilateral limb phase relationship and its potential to alter muscle activity phasing during locomotion. J Neurophysiol
31. Ting LH, Kautz SA, Brown DA, Zajac FE. Contralateral movement and extensor force generation alter flexion phase muscle coordination in pedaling. J Neurophysiol
32. Kautz SA, Brown DA, Van der Loos HF, Zajac FE. Mutability of bifunctional thigh muscle activity in pedaling due to contralateral leg force generation. J Neurophysiol
33. Balter JE, Zehr EP. Neural coupling between the arms and legs during rhythmic locomotor-like cycling movement. J Neurophysiol
34. Stang J, Wiig H, Hermansen M, Hansen EA. Voluntary movement frequencies in submaximal one- and two-legged knee extension exercise and pedaling. Front Hum Neurosci
35. Zehr EP, Collins DF, Chua R. Human interlimb reflexes evoked by electrical stimulation of cutaneous nerves innervating the hand and foot. Exp Brain Res
36. Kautz SA, Neptune RR, Zajac FE. General coordination principles elucidated by forward dynamics: minimum fatigue does not explain muscle excitation in dynamic tasks. Motor Control
. 2000;4(1):75–80, discussion 97–116.
37. Dorel S, Drouet JM, Hug F, Lepretre PM, Champoux Y. New instrumented pedals to quantify 2D forces at the shoe-pedal interface in ecological conditions: preliminary study in elite track cyclists. Comput Methods Biomech Biomed Engin
38. Samaan MA, Hoch MC, Ringleb SI, Bawab S, Weinhandl JT. Isolated hamstrings fatigue alters hip and knee joint coordination during a cutting maneuver. J Appl Biomech
39. Blake OM, Wakeling JM. Muscle coordination limits efficiency and power output of human limb movement under a wide range of mechanical demands. J Neurophysiol
40. Wakeling JM, Blake OM, Wong I, Rana M, Lee SS. Movement mechanics as a determinate of muscle structure, recruitment and coordination. Philos Trans R Soc Lond B Biol Sci
Keywords:© 2018 American College of Sports Medicine
MULTIJOINT TASK; JOINT-SPECIFIC POWER; MOTOR CONTROL; MECHANICAL ENERGY TRANSFER; MECHANICAL EFFECTIVENESS