Experimental Physiology

Transcription

1 Exp Physiol 99.8 (2014) pp Research Paper Research Paper Dynamics of corticospinal changes during and after high-intensity quadriceps exercise Mathieu Gruet 1,2,3,JohnTemesi 4,5, Thomas Rupp 1,2,PatrickLevy 1,2, Samuel Verges 1,2 and Guillaume Y. Millet 2,4,5 1 Université Grenoble-Alpes, Laboratoire HP2, F Grenoble, France 2 INSERM, U1042, F Grenoble, France 3 Laboratoire Motricité Humaine, Education, Sport, Santé, UniversitédeToulon,France 4 Université de Lyon, F-42023, Saint-Etienne, France 5 Human Performance Laboratory, Faculty of Kinesiology, University of Calgary, Calgary, Alberta, Canada Experimental Physiology New Findings What is the central question of this study? Progressive development of the supraspinal component of central fatigue and increases in corticospinal excitability and inhibition have been demonstrated during fatiguing contractions of the elbow flexors. However, the kinetics of mechanical and EMG responses induced by transcranial magnetic stimulation during and after single-joint fatiguing knee-extensor exercise remains unknown. What is the main finding and its importance? Our results show that single-joint knee-extensor isometric exercise induces late supraspinal fatigue with increased intracortical inhibition, both of which recover quickly after task failure, and unchanged corticospinal excitability. This indicates that fatigue-induced corticospinal changes are muscle and/or limb specific and reinforces the need to measure corticospinal changes within seconds after task failure to avoid their underestimation. This study tested the hypothesis that during fatiguing quadriceps exercise, supraspinal fatigue develops late, is associated with both increased corticospinal excitability and inhibition and recovers quickly. Eight subjects performed 20 s contractions [15 s at 50% maximal voluntary contraction (MVC) followed by 5 s MVC] separated by a 10 s rest period until task failure. Transcranial magnetic stimulation (TMS) and electrical femoral nerve stimulation (PNS) were delivered 2 s apart during 50% MVC, during MVC and after MVC in relaxed muscle. Voluntary activation was assessed by TMS (VA TMS ) immediately before and after exercise and then three times over a 6 min recovery period. During exercise, MVC and twitch force evoked by PNS in relaxed muscle decreased progressively to 48 ± 8and36± 16% of control values, respectively (both P < 0.01). Significant changes in voluntary activation assessed by PNS and twitch evoked by TMS during MVC were observed during the last quarter of exercise only (from 96.4 ± 1.7 to 86 ± 13%, P = 0.03 and from 0.76 ± 0.8 to 4.9 ± 4.7% MVC, P = 0.02, from baseline to task failure, respectively). The TMS-induced silent period increased linearly during both MVC (by 79 ms) and 50% MVC (by 63 ms; both P < 0.01). Motor-evoked potential amplitude did not change during the protocol at any force levels. Both silent period and VA TMS recovered within S. Verges and G. Y. Millet share senior authorship. DOI: /expphysiol

2 1054 M. Gruet and others Exp Physiol 99.8 (2014) pp min postexercise, whereas MVC and twitch force evoked by PNS in relaxed muscle recovered to only 84 ± 9and73± 17% of control values 6 min after exercise, respectively. In conclusion, high-intensity single-joint quadriceps exercise induces supraspinal fatigue near task failure, with increased intracortical inhibition and, in contrast to previous upper-limb results, unchanged corticospinal excitability. These changes recover rapidly after task failure, emphasizing the need to measure corticospinal adaptations immediately at task failure to avoid underestimation of exercise-induced corticospinal changes. (Received 26 February 2014; accepted after revision 29 May 2014; first published online 6 June 2014) Corresponding author S. Verges: Laboratoire HP2 (U1042 INSERM), UF Recherche sur l Exercice, Hôpital Sud, Avenue Kimberley, Echirolles, France. sverges@chu-grenoble.fr Introduction Humanmusclefatigueisdefinedasanexercise-induced decrease in maximal voluntary muscle force and can originate at peripheral and/or central levels (Gandevia, 2001). During an intermittent fatiguing contraction protocol, a progressive decline in twitch or tetanic force elicited by peripheral nerve stimulation (PNS) in the relaxed muscle indicates the development of peripheral fatigue (Froyd et al. 2013). Meanwhile, the progressive increase in force increment evoked by PNS during brief maximal voluntary contractions (MVCs) indicates that some motor units are either not recruited or not firing optimally despite maximal voluntary effort. This is referred to as central fatigue and can originate at spinal and/or supraspinal levels. In particular, the increment in force evoked by transcranial magnetic stimulation (TMS) during brief MVCs can increase during a sustained fatiguing contraction, indicating progressive suboptimal output from the motor cortex (i.e. supraspinal fatigue; Gandevia et al. 1996). The development of fatigue is often associated with changes in EMG responses to TMS. During fatiguing isometric contractions, the size of the motor-evoked potential (MEP), reflecting corticospinal excitability, usually increases (Søgaard et al. 2006; Smith et al. 2007). The duration of the period of near silence in the EMG signal (silent period; SP) after an MEP also increases during fatiguing exercise, indicating increased inhibition within the motor cortex (Gandevia et al. 1996). The progressive development of peripheral and central fatigue, including the supraspinal component of the latter, and increases in corticospinal excitability and inhibition have been demonstrated during sustained maximal and submaximal isometric contractions of the elbow flexors (Hunter et al. 2006; Søgaard et al. 2006). However, few data are available during fatiguing exercise for other muscle groups, especially lower-limb muscles. As differences in motor unit recruitment patterns, firing rates and the strength of corticospinal projections exist between lowerand upper-limb muscles (Brouwer & Ashby, 1990, 1992; de Noordhout et al. 1999; Enoka & Fuglevand, 2001), one may expect specific corticospinal adaptations to lower-limb muscle fatigue. An increase in corticospinal excitability has been reported during submaximal (Hoffman et al. 2009) and maximal isometric contractions of the plantar flexors (Iguchi & Shields, 2012). However, the absence of SP data (Hoffman et al. 2009) and concomitant indices of central and supraspinal fatigue in both studies limit their interpretation. The knee extensors play a critical role in locomotor tasks such as walking, running and cycling. In addition, walking and cycling are often used in rehabilitation programmes. Thus, studying quadriceps muscle fatigability is of importance in both clinical and research settings. It is well established that both peripheral and central factors determined by PNS evaluation can contribute to quadriceps muscle fatigability (Burnley et al. 2012; Froyd et al. 2013); their relative contributions depend on the characteristics of the motor task (e.g. intermittent versus continuous, maximal versus submaximal, single-joint versus whole-body exercise; Millet & Lepers, 2004; Place et al. 2009; Burnley et al. 2012). The contribution of supraspinal mechanisms to exercise-induced reductions in quadriceps strength remains to be clarified. A reduction in cortical voluntary activation (measured by TMS; VA TMS ) of the quadriceps has been observed following whole-body exercise (Goodall et al. 2012; Temesi et al. 2014b), indicating the development of supraspinal fatigue during the task. It is possible that supraspinal fatigue of the quadriceps occurs late, because central fatigue has been shown to develop close to task failure during both intermittent isometric contractions (Bachasson et al. 2013) and cycling exercise (Decorte et al. 2012). Indeed, a similar pattern can be expected because the changes of central and supraspinal fatigue have been found to be related (Temesi et al. 2014b). The absence of studies investigating MEP and SP kinetics during fatiguing contractions of the quadriceps does not permit determination of whether this hypothetical late development of supraspinal fatigue is associated with EMG changes elicited by TMS. Given that similar

3 Exp Physiol 99.8 (2014) pp Corticospinal responses to quadriceps fatigue 1055 kinetics of MEP changes have been observed in previous single-joint studies independent of the investigated muscle (e.g. elbow versus plantar flexors; Søgaard et al. 2006; Hoffman et al. 2009), an increase in corticospinal excitability and inhibition can be expected. In most quadriceps studies, neuromuscular responses are recorded within several minutes after exercise termination (Sidhu et al. 2009b; Goodall et al. 2010, 2012; Fernandez-Del-Olmo et al. 2013; Temesi et al. 2013, 2014b). Froyd et al. (2013) recently demonstrated that substantial recovery of peripheral fatigue occurs within the first minute after high-intensity dynamic quadriceps exercise. Rapid recovery of cortical drive has also been demonstrated several seconds after a sustained 90 s MVC of the first dorsal interosseus (Szubski et al. 2007). These results suggest that previous studies investigating lower-limb fatigability by TMS after a short postexercise delay may have underestimated the extent to which corticospinal changes develop during exercise. The present study was therefore designed to investigate whether the dynamics of corticospinal changes during and after a high-intensity single-joint quadriceps exercise bout differ compared with upper-limb muscles. We hypothesized that supraspinal fatigue of the quadriceps has the following characteristics: (i) it develops late during intense single-joint exercise; (ii) it is associated with increases in corticospinal excitability and inhibition; and (iii) it recovers quickly after exercise termination. Methods Ethical approval Written informed consent was obtained, and the study was conducted according to the latest revision of the Declaration of Helsinki and approved by the local ethics committee (Comité de Protection des Personnes Sud-Est 1, France). Participants Eight healthy, active men participated in the study (age, 30 ± 8 years; height, 181 ± 5 cm; body mass, 73 ± 4 kg). Participants were non-smokers, non-epileptic and free of cardiorespiratory disease and contraindications to TMS. All subjects had at least 3 years of endurance sport experience and trained 4 ± 3sessionsperweek.They were asked to avoid strenuous exercise for a minimum of 48 h preceding the trial and to refrain from caffeine for at least 12 h before the start of the experiment. Experimental set-up Subjects sat upright in a custom-built chair, with both hips and knees at 90 deg of flexion. Knee-extensor force was measured during voluntary and evoked contractions by a calibrated force transducer (Meiri F dan; Celians, Montauban, France) with an amplifier attached by a non-compliant strap to the right leg immediately proximal to the malleoli of the ankle joint. The force transducer was fixed to the chair such that force was measured in a direct line to the applied force. The subjects were secured to the chair with non-compliant straps to minimize body movement. Surface EMG signals were recorded from the right vastus lateralis, vastus medialis, rectus femoris and biceps femoris muscles (as a surrogate for antagonist hamstring muscles). The EMG was recorded with a pair of self-adhesive surface electrodes (Meditrace 100; Covidien, Mansfield, MA, USA) in bipolar configuration with a 30 mm interelectrode distance, and the reference electrode was attached on the patella. A low impedance (<5k ) between the electrodes was obtained by shaving, gently abrading the skin with sandpaper and then cleaning it with isopropyl alcohol. Signals were analog-to-digitally converted at a sampling rate of 2000 Hz by a PowerLab system (16/30 ML880/P; ADInstruments, Bella Vista, NSW, Australia) and octal bio-amplifier (ML138; ADInstruments) with a bandpass filter (5 500 Hz) and analysed offline using Labchart 7 software (ADInstruments). Peripheral nerve stimulation Single electrical stimuli of 1 ms duration were delivered via a high-voltage (maximal voltage, 400 V), constant-current stimulator (DS7A; Digitimer, Welwyn Garden City, UK) to the right femoral nerve via a 30-mm-diameter surface cathode located in the femoral triangle (Meditrace 100; Covidien) and 50 mm 90 mm rectangular anode (Durastick Plus; DJO Global, Vista, CA, USA) located in the gluteal fold. Single stimuli were delivered incrementally to the relaxed muscle until plateaus in both the resting M-wave and the twitch amplitude were reached. The stimulus intensity was set at 30% above the level required to produce maximal M-wave (M max ) and twitch amplitudes. Motor cortical stimulation A magnetic stimulator (Magstim ; The Magstim Company Ltd, Whitland, UK) was used to stimulate the motor cortex. Single TMS pulses of 1 ms duration were delivered via a concave double-cone coil (110 mm diameter, maximal output of 1.4 T) to stimulate the left motor cortex preferentially, contralateral to the right leg. The coil was manually controlled by an experienced investigator throughout the protocol. Subjects wore a cervical collar to stabilize the head and neck and a latex swimming cap, on which lines were drawn between the pre-auricular points and from nasion to inion to

4 1056 M. Gruet and others Exp Physiol 99.8 (2014) pp identify the vertex. Every centimetre, from 1 cm anterior to 3 cm posterior to the vertex, was demarcated along the nasal inion line and also to 2 cm over the left motor cortex. At each point, a stimulus was delivered at 70% maximal stimulator output during brief ( 2 3 s) voluntary contractions of the knee extensors at 10% MVC. The coil was positioned at the site evoking the largest rectus femoris MEP amplitude with minimal biceps femoris MEP amplitude (Rupp et al. 2012). This coil position was drawn onto the swimming cap and used throughout the experimental session. The coil position was also verified before the delivery of each TMS pulse. Real-time visual feedback of target force levels was provided to subjects on a computer screen throughout the experiment. During all voluntary contractions, TMS was delivered once the subject had contracted to the appropriate force level and the force had stabilized (Gruet et al. 2013b). A stimulus response curve at 20% MVC was used to determine optimal TMS intensity, as recently described by Temesi et al. (2014a). Subjects performed brief contractions ( 2 3 s) at 20% MVC of the knee extensors with TMS delivered four consecutive times at each of the following randomly ordered stimulus intensities: 20, 30, 40, 50, 60, 70 and 80% maximal stimulator output. Stimuli were delivered at 15 s intervals. Optimal stimulus intensity was selected by identifying the minimal stimulus intensity to evoke maximal MEP amplitude (i.e. the lowest intensity resulting in an increase of <5% MEP amplitude at higher stimulus intensities) in all three quadriceps muscles. Antagonist MEP amplitude was examined to verify that optimal stimulus intensity did not elicit increased TMS-induced coactivation (Todd et al. 2003). The biceps femoris MEP amplitude was 7.5 ± 6.0% of the rectus femoris MEP amplitude at the optimal stimulus intensity. The same stimulus intensity (i.e. 59 ± 11% maximal simulator output) was used throughout the protocol. Protocol Pre- and postfatiguing task. All subjects were familiarized with the procedures to be conducted in the experimental session, including the performance of reproducible MVCs with and without the delivery of PNS and TMS. They were also trained to keep the head as immobile as possible during voluntary contractions to avoid small head displacements relative to the coil before TMS delivery. The protocol consisted of baseline measurements (Pre; total duration, 6 min), an intermittent isometric fatiguing exercise of the quadriceps muscles performed to task failure (designed to last 3 5 min) and a recovery phase (6 min; Fig. 1). Four sets of contractions were performed at Pre. Each set comprised a brief MVC followed by contractions at 75, 50 and 25% MVC performed in a random order between subjects. The same order of contraction intensities was used during the recovery phase. A 10 s rest period was provided between submaximal contractions and a 25 s rest period after each MVC and between sets. All voluntary contractions at Pre and during recovery lasted 5 s. The PNS and TMS (fixed order) were delivered manually 2 s apart during each contraction and immediately after MVC in relaxed muscle (Fig. 1). The order of PNS and TMS was not randomized in order to ensure that stimuli were delivered at the same force level and to avoid overestimating the size of the superimposed twitch evoked by TMS (SIT). Four identical sets (i.e. same timing as Pre) were performed immediately after exercise termination[post1(p1),p2,p3andp4,starting6,96,186 and 276 s after task failure, respectively; Fig. 1]. The 75, 50 and 25% MVC were recalculated for each set based on the preceding MVC. Fatiguing task. The fatiguing exercise consisted of repeated intermittent isometric knee extensions performed until task failure. The subjects performed 20 s contractions (15 s at 50% Pre MVC, immediately followed by a 5 s MVC) separated by a 10 s rest period (Fig. 1). The fatiguing exercise ended when the subject was unable to reach 50% Pre MVC for >2 s (i.e. task failure). The PNS and TMS were delivered 2 s apart, during the last 3 s of the 15 s 50% Pre MVC contraction, during the 5 s MVC, and immediately after the MVC in relaxed muscle. The choice of an intermittent task was required to evaluate simultaneously the kinetics of peripheral, central (with PNS) and supraspinal (with TMS) fatigue and changes in corticospinal excitability and inhibition in different muscular states (i.e. in the relaxed muscle, at 50% Pre MVC and during MVC). As fatigue-induced changes in parameters elicited by TMS (e.g. MEP) may depend on muscular state (i.e. relaxed versus contracted muscle, low versus high contraction intensities; for review see Gruet et al. (2013a)), we assessed corticospinal changes at different force levels. The subjects were instructed to return as quickly as possible to the desired force level after each TMS pulse elicited during voluntary contractions to allow valid assessment of SP (Mathis et al. 1998). Strong verbal encouragement was given throughout the protocol. Data analysis Peak-to-peak MEP amplitudes in each muscle were normalized to the amplitude of M max elicited at the same time point during a contraction of similar voluntary force (i.e. during MVC, MEP 100 ; during 50% MVC, MEP 50 ;and in the relaxed muscle, MEP 0 ). The duration of the SP was visually determined and calculated as the interval from TMS delivery to the resumption of continuous

5 Exp Physiol 99.8 (2014) pp Corticospinal responses to quadriceps fatigue 1057 voluntary EMG (Damron et al. 2008). The silent period was determined during MVC (SP 100 ) and contractions at 50% MVC (SP 50 ). Given that similar results were obtained for all quadriceps muscles [i.e. no interaction effect (time muscle) was found for any MEP or SP parameters], MEP amplitudes and SP durations from the vastus lateralis, rectus femoris and vastus medialis were averaged and used for further analysis. Root mean square EMG (EMG RMS ) was measured over 500 ms prior to PNS at 50 and 100% MVC throughout the protocol. The voluntary activation measured by PNS (VA PNS ) was assessed by twitch interpolation (Merton, 1954). The amplitude of the superimposed twitch elicited by PNS during a brief MVC was compared with that of the subsequent twitch in relaxed muscle using the following equation: VA PNS (%) = [1 (superimposed twitch/twitch)] 100 The VA TMS was assessed using the method described by Todd et al. (2003) and validated for the knee extensors (Goodall et al. 2009; Sidhu et al. 2009a). The amplitude of the twitch elicited by TMS in relaxed muscle was estimated rather than measured directly because motor cortical and spinal excitability increase with increasing muscle activity (Rothwell et al. 1991). The estimated resting twitch (ERT) was calculated by extrapolation of the linear relationship between the SIT amplitude and voluntary force during brief maximal and submaximal contractions. Coefficients of determination (r 2 ) of the regression of voluntary force and the superimposed twitch were significantly higher when using four points (25, 50, 75 and 100% MVC) compared with three points (50, 75 and 100% MVC) postexercise (average of 32 sets: r 2 = 0.91 ± 0.06 versus r 2 = 0.87 ± 0.11, P = 0.02). Thus, VA TMS and ERT were calculated from the linear relationship derived from four points because this method has previously been shown to be reliable (Sidhu et al. 2009a). One regression analysis was performed for each set of brief contractions, and the y-interceptwastakenastheert.theva TMS was then calculated using the following equation: VA TMS (%) = [1 (SIT during MVC/ERT)] 100 The SIT during MVC was also normalized to the strength prior to the stimulus (SIT/MVC) to obtain an index of supraspinal fatigue during exercise (Taylor et al. 2000; Todd et al. 2003; Szubski et al. 2007). Owing to the between-subject variability in time to task failure, all data during the fatiguing exercise were normalized as a percentage of endurance time (Lévénez et al. 2008). Data were interpolated to obtain a value every 25% of endurance time. This timing is illustrated in Figs 2, 3, 5 and 6. For each parameter, the pre-exercise values represented the mean of the four sets performed before the fatiguing task. The sets performed immediately % MVC X 4 15 s 10 s 5 s X 4 (P1, P2, P3, P4) PNS TMS s 15 s 15 s 30 s 30 s 6 s 30 s 15 s 15 s 30 s PRE-EXERCISE FATIGUING EXERCISE POSTEXERCISE Figure 1. Schematic illustration of the protocol Pre-exercise (Pre) measurements consisted of four sets of brief voluntary contractions [maximal voluntary contraction (MVC), then randomly ordered 75, 50 and 25% MVC contractions]. The 75, 50 and 25% MVCs were recalculated for each set, based on the preceding MVC. Peripheral nerve stimulation (PNS) and transcranial magnetic stimulation (TMS) were delivered 2 s apart during each brief contraction and immediately after the MVC in relaxed muscle. Four identical sets [Post 1 (P1), P2, P3 and P4, starting 6, 96, 186 and 276 s after task failure, respectively] were performed immediately after the fatiguing exercise (i.e. recovery period). The fatiguing exercise consisted of 20 s contractions (15 s at 50% MVC Pre and a 5 s MVC) separated by a 10 s rest period performed until the subject was unable to perform >2 sator above the level of 50% MVC Pre. The PNS and TMS were delivered during the last 3 s of the 15 s 50% MVC Pre, during the 5 s MVC and immediately after the MVC in relaxed muscle.

6 1058 M. Gruet and others Exp Physiol 99.8 (2014) pp after task failure (P1, P2, P3 and P4) were reported separately to describe the recovery kinetics. Coefficients of variation for VA TMS and ERT between the four sets at Pre were 1.6 and 9.4%, respectively. Statistics Statistical analyses were performed with Statistica (version 7.0; StatSoft Inc., Tulsa, OK, USA). Normality of distribution and homogeneity of variances of the main variables were verified using a Shapiro Wilk normality test and Levene s test, respectively. Changes in mechanical and EMG parameters were analysed during either the fatigue protocol or recovery by one-way ANOVAs with repeated measures (Pre, percentage endurance time, P1, P2, P3 and P4). Tukey s post hoc test was performed if the ANOVA revealed a significant main effect. Statistical significance was set at P < Results are given in the text as means ± SDandinfiguresasmeans± SEM. Results Time to task failure, MVC and peripheral function The mean time to task failure was 243 ± 69 s. Maximal voluntary contraction at Pre was 631 ± 69 N and decreased 120 gradually during the fatiguing exercise to 48 ± 8% of Pre values at task failure [F(4,28) = 78.89; P < 0.01]. The MVC recovered progressively but was still significantly less than Pre at P4 (Fig. 2). The twitch evoked in relaxed muscle declined progressively to 36 ± 16% of Pre values at task failure [F(4,28) = 50.06; P < 0.01] and only partly recovered to 73 ± 17%ofPrevaluesatP4(P < 0.01; Fig. 2). The M max amplitude measured in the relaxed muscle and during MVC did not change during the protocol in any muscle (P > 0.05). The M max amplitude at 50% MVC decreased from Pre to task failure in both rectus femoris [9.8 ± 2.8 versus 8.2 ± 2.7 mv, F(4,28) = 8.28; P < 0.01] and vastus medialis [15.6 ± 5.5 versus 13.7 ± 6.7 mv, F(4,28) = 3.81; P = 0.01], with the decrease mainly occurring during the last quarter of the fatiguing exercise. In both muscles, M max recovered by P1 (rectus femoris, 9.5 ± 3.5 mv; vastus medialis, 15.3 ± 5.6 mv; both P > 0.05). Voluntary EMG and central fatigue The EMG RMS during contractions at 50% MVC progressively increased during the fatiguing exercise in vastus lateralis, rectus femoris and vastus medialis % Pre values MVC MVC Twitch 0 Pre TF P1 P2 P3 P4 Fatiguing exercise (% endurance time) Postexercise Figure 2. Neuromuscular and peripheral fatigue Changes in MVC and twitch force evoked by peripheral nerve stimulation in resting muscle at baseline (Pre), during the fatiguing exercise until task failure (TF) and after exercise (P1 P4), expressed as a percentage of Pre values. Values are means ± SEM. P < 0.05 from Pre values.

7 Exp Physiol 99.8 (2014) pp Corticospinal responses to quadriceps fatigue 1059 [F(4,28) = 13.05, and 22.04, respectively; all P < 0.01]. The VA PNS declined during the fatiguing exercise, and this reduction was significant compared with Pre at task failure only [F(4,28) = 3.01, P = 0.03; Fig. 3]. During recovery, VA PNS was significantly lower than Pre at P1 and had recovered by P2. The SIT/MVC increased significantly from Pre to task failure (from 0.76 ± 0.8to4.9± 4.7% MVC), with the increase mainly occurring during the last quarter of the fatiguing exercise [F(4,28) = 3.51, P = 0.02]; however, it had recovered by P1 (Fig. 3). The VA TMS was significantly lowerthanpreatp1only(fig.4). Corticospinal inhibition and excitability The silent period increased linearly during the fatiguing exercise at both force levels [F(4,28) = and at 50 and 100% MVC, respectively; both P < 0.01; Fig. 5]. During recovery, SP 50 returned to Pre values by P1, whereas SP 100 remained significantly elevated compared with Pre values at P1 only (Fig. 5). These changes in SP were observed consistently in all quadriceps muscles (all P < 0.01). The MEP amplitude did not change 100 during exercise [MEP 0, F(4,28) = 2.36, P = 0.08; MEP 50, F(4,28) = 0.19, P = 0.95; and MEP 100, F(4,28) = 2.18, P = 0.10] or recovery (Fig. 6). When each quadriceps muscle was analysed separately, only the vastus lateralis showed a significant increase in MEP 100 at task failure [F(4,28) = 4.3, P = 0.01]. Discussion This study described supraspinal fatigue kinetics and the associated changes in corticospinal excitability and inhibition during and immediately after a high-intensity bout of quadriceps exercise. Supraspinal fatigue developed by the end of the task, and reduced cortical voluntary drive was observed immediately after task failure only. Corticospinal inhibition increased progressively during the fatiguing exercise but, contrary to what is generally observed in upper-limb muscles, this was not accompanied by significant changes in corticospinal excitability, suggesting a knee-extensor specificity. Both supraspinal fatigue and changes in corticospinal inhibition recovered within 2 min after task failure VA PNS (%) VA PNS SIT/MVC SIT/ MVC Pre TF P1 P2 P3 P4 0 Fatiguing exercise (% endurance time) Postexercise Figure 3. Supraspinal and central fatigue Changes in voluntary activation determined by peripheral nerve stimulation (VA PNS ) and the ratio between superimposed twitch elicited by transcranial magnetic stimulation and maximal voluntary contraction (SIT/MVC) at baseline (Pre), during the fatiguing exercise until task failure (TF) and after exercise (P1 P4). Values are means ± SEM. P < 0.05 from Pre values.

8 1060 M. Gruet and others Exp Physiol 99.8 (2014) pp Maximal voluntary force and peripheral fatigue Maximal voluntary force decreased progressively during the fatiguing exercise, indicating the development of neuromuscular fatigue. A substantial part of this fatigue was peripheral, as the twitch evoked from relaxed muscle also decreased progressively, consistent with previous single-joint muscle fatigue studies (Froyd et al. 2013). We reported a larger exercise-induced decrease in PNS-evoked quadriceps twitch amplitude than MVC, as has been described previously (e.g. Decorte et al. 2012). One likely explanation is that the resting twitch is affected by impairment of excitation contraction coupling that is partly overcome by the high-frequency motor unit firing rates required for an MVC. The decrease in twitch amplitude was not associated with changes in M max recorded in the relaxed muscle, suggesting that neuromuscular propagation did not contribute to decreased resting twitch amplitude. Therefore, most of the peripheral impairment induced by the fatiguing exercise originated within the muscle fibres, probably because of impaired excitation contraction coupling. One surprising result was the decrease in M max at task failure in both rectus femoris and vastus medialis when measured during 50% MVC only. This may have occurred because of the likelihood that more motoneurons and muscles fibres were refractory at the end of the fatiguing exercise as their firing rates increased to maintain the 50% Pre MVC target force level (corresponding to 100% MVC at task failure; Matthews 1999). This is consistent with both the progressive increase of EMG RMS during contractions at 50% MVC and M max being significantly smaller in the rectus femoris and vastus medialis muscles at Pre during MVC compared with 50 and 75% MVC (data not shown). The sensitivity of M max kinetics during exercise to the level of force at which M max is recorded highlights the importance of normalizing corticospinal responses (i.e. MEP) to M max recorded at similar contraction intensity. Central and supraspinal fatigue The progressive inability to drive the muscle maximally during the fatiguing exercise was reflected by the gradual decrease in VA PNS, indicating progressive development of central fatigue. The increase in SIT/MVC during the exercise suggested a supraspinal component to the observed central fatigue (i.e. suboptimal output from the motor cortex), as previously reported during submaximal and maximal fatiguing contractions of the elbow flexors (Hunter et al. 2006; Søgaard et al. 2006) and the first dorsal interosseus (Szubski et al. 2007). While the development of supraspinal fatigue was progressive in these studies, most of the increase in SIT/MVC occurred during the last quarter of the fatiguing exercise in the present study, and this became significant only at task failure. Differences VA TMS (%) ERT (N) 84 VA TMS ERT Pre TF P1 P2 P3 P4 40 Fatiguing exercise Postexercise Figure 4. Cortical voluntary activation and estimated resting twitch Changes in cortical voluntary activation (VA TMS ) and estimated resting twitch (ERT) at baseline and postexercise. Values are means ± SEM. P < 0.05 from Pre values.

9 Exp Physiol 99.8 (2014) pp Corticospinal responses to quadriceps fatigue 1061 in corticospinal projections between upper and lower limbs (Brouwer & Ashby 1992; de Noordhout et al. 1999) and/or specific roles of group III IV muscle afferents in extensor muscles (Martin et al. 2006; Hilty et al. 2011) may explain the late development of supraspinal fatigue in the quadriceps muscle. It is also possible that the intermittent nature of the task mayhavecontributedtothelatedevelopmentofcentral and supraspinal fatigue by allowing recovery during the 10 s resting period. However, late development of central fatigue measured in quadriceps muscles has been reported during various protocols (Place et al. 2004; Decorte et al. 2012; Bachasson et al. 2013) and thus does not seem specific to our task. We cannot, however, exclude the possibility that continuous single-joint fatiguing tasks, such as sustained MVCs, may lead to different kinetics of supraspinal fatigue, and this issue will require further investigation. In order to minimize the potential effect of recovery periods on the kinetics of supraspinal fatigue, we did not calculate VA TMS during exercise. This would have required the performance of at least two additional brief contractions at different contraction intensities 50% MVC and separated by several seconds of rest (Todd et al. 2003), therefore allowing even more recovery between the sustained contractions. Consequently, VA TMS was measured before the fatiguing exercise and immediately after task failure. The relationship between voluntary force and VA TMS was almost linear before and after fatigue in our study, as previously demonstrated for the knee extensors (Goodall et al. 2009; Sidhu et al. 2009a). Thus, it was possible to estimate the contribution of supraspinal fatigue to the total force loss. For each subject, the linear force VA TMS relationship was determined immediately after task failure. Using the regression equation, the force corresponding to pre-exercise VA TMS was determined and compared with the real postexercise MVC. The additional force loss was interpreted to be due to supraspinal fatigue, which accounted for 17% of the total force loss. Based on the rapid recovery of SIT/MVC between task failure and P1, it is likely that the reduction in VA TMS at P1 led to underestimation of the amount of supraspinal fatigue at task failure. The rapid recovery of both SIT/MVC and VA TMS (not significantly lower at P2, i.e. 2 3 min after task failure) indicates that previous whole-body studies, SP SP 50 SP (ms) Pre TF P1 P2 P3 P4 Fatiguing exercise (% endurance time) Postexercise Figure 5. Transcranial magnetic stimulation-induced silent period Changes in silent period (SP) determined at 50% MVC (SP 50 )andatmvc(sp 100 ) at baseline (Pre), during the fatiguing exercise until task failure (TF) and after exercise (P1 P4). Values are means ± SEM. P < 0.05 from Pre values.

10 1062 M. Gruet and others Exp Physiol 99.8 (2014) pp in which these parameters were measured several minutes after exercise termination (Sidhu et al. 2009b; Goodall et al. 2012; Fernandez-Del-Olmo et al. 2013; Temesi et al. 2013), may have underestimated the extent to which supraspinal fatigue develops during exercise. Likewise, averaging the sets of contractions to assess supraspinal fatigue over several minutes after a fatiguing exercise bout (Goodall et al. 2009, 2010; Sidhu et al. 2009b; Temesiet al. 2013) probably also underestimates the decline of VA TMS. Corticospinal excitability and inhibition The motor-evoked potential amplitude did not change during exercise or recovery at any force level. The absence of changes in MEP during sustained contractions contrasts with previous upper-limb studies, which consistently reported an increase in MEP size during submaximal and maximal contractions (Taylor et al. 2000; Søgaard et al. 2006; Smith et al. 2007; Szubski et al. 2007; Hunter et al. 2008; Lévénez et al. 2008). Given that changes in MEP with fatigue may depend on the level of force at TMS delivery (Iguchi & Shields, 2012; Gruet et al b), MEP kinetics were assessed during maximal (MEP 100 ) and submaximal contractions (MEP 50 ) and in the relaxed muscle (MEP 0 ). The fact that the MEP was unchanged in all three conditions supports our finding that corticospinal excitability does not increase during single-joint quadriceps exercise. Slight differences may, however, exist between quadriceps muscles, as suggested by the significant increase in MEP 100 in the vastus lateralis only. Motor-evoked potential kinetics during quadriceps exercise have only recently been reported during whole-body cycling exercise (Sidhu et al. 2012). These authors observed unchanged MEP size. They attributed this lack of change in corticospinal excitability to the specific metabolic demands associated with whole-body exercise (e.g. increased cardiorespiratory demands and changes in core or brain temperature, blood glucose or brain catecholamines). Our results suggest that a difference between lower-limb extensor and upper-limb flexor muscles may also influence fatigue-induced corticospinal changes. It has been demonstrated in the elbow-flexor muscles that small-diameter afferents activated by fatigue inhibit extensor motoneurons but MEP 100 MEP MEP 0 MEP (% M max ) Pre TF P1 P2 P3 P4 Fatiguing exercise (% endurance time) Postexercise Figure 6. Transcranial magnetic stimulation-induced motor-evoked potentials Changes in motor-evoked potential amplitude [MEP, expressed relative to the maximal M-wave (M max ) recorded in the same conditions] determined during MVC (MEP 100 ) during 50% MVC (MEP 50 )and in relaxed muscle (MEP 0 ) at Pre and during both the fatiguing exercise until task failure (TF) and postexercise. Values are mean ± SEM.

11 Exp Physiol 99.8 (2014) pp Corticospinal responses to quadriceps fatigue 1063 facilitate flexor motoneurons (Martin et al. 2006). This effect may not be unique to elbow-flexor muscles and may also apply to thigh muscles, explaining, at least in part, the absence of MEP facilitation during knee-extensor exercise. This hypothesis is supported by the large increase in SP during exercise found in our study, an increase that has been demonstrated to be mediated by group III IV muscle afferents in quadriceps muscles (Hilty et al. 2011). Conversely, in elbow flexors, fatigue-induced changes in SP have been found to be independent of the increased firing of these afferents (Gandevia et al. 1996). Finally, Sidhu et al. (2013) recently observed an increase in intracortical inhibition (measured by EMG suppression induced by subthreshold TMS) of the knee extensors during cycling exercise, further suggesting similar corticospinal responses during whole-body exercise and isometric single-joint contractions for the quadriceps. Similar to VA TMS,therapidrecoveryofSPfollowing exercise suggests that measuring TMS-induced changes in EMG responses several seconds after completion of the task may not reflect corticospinal adaptations during fatiguing exercise. This is of particular importance for whole-body studies, where the time between exercise termination and postexercise measurements may be several minutes in duration and thus should be reduced as much as possible to assess the full magnitude of corticospinal changes. Conclusion The present results suggest that high-intensity single-joint quadriceps exercise is accompanied by a late development of supraspinal fatigue, with increased intracortical inhibition and unchanged corticospinal excitability. This pattern contrasts with previous upper-limb studies and suggests specific corticospinal adaptations to lower-limb extensor fatigue. Moreover, the rapid recovery of both mechanical and EMG TMS parameters after exercise emphasizes the need to measure changes in supraspinal fatigue and corticospinal excitability within seconds after exercise cessation in order to avoid substantial underestimation of corticospinal changes associated with fatigue. References Bachasson D, Millet GY, Decorte N, Wuyam B, Levy P & Verges S (2013). Quadriceps function assessment using an incremental test and magnetic neurostimulation: a reliability study. J Electromyogr Kinesiol 23, Brouwer B & Ashby P (1990). Corticospinal projections to upper and lower limb spinal motoneurons in man. Electroencephalogr Clin Neurophysiol 76, Brouwer B & Ashby P (1992). Corticospinal projections to lower limb motoneurons in man. Exp Brain Res 89, Burnley M, Vanhatalo A & Jones AM (2012). Distinct profiles of neuromuscular fatigue during muscle contractions below and above the critical torque in humans. JApplPhysiol113, Damron LA, Dearth DJ, Hoffman RL & Clark BC (2008). Quantification of the corticospinal silent period evoked via transcranial magnetic stimulation. JNeurosciMethods173, Decorte N, Lafaix PA, Millet GY, Wuyam B & Verges S (2012). Central and peripheral fatigue kinetics during exhaustive constant-load cycling. Scand J Med Sci Sports 22, de Noordhout AM, Rapisarda G, Bogacz D, Gerard P, De Pasqua V, Pennisi G & Delwaide PJ (1999). Corticomotoneuronal synaptic connections in normal man: an electrophysiological study. Brain 122, Enoka RM & Fuglevand AJ (2001). Motor unit physiology: some unresolved issues. Muscle Nerve 24, Fernandez-Del-Olmo M, Rodriguez FA, Marquez G, Iglesias X, Marina M, Benitez A, Vallejo L & Acero RM (2013). Isometric knee extensor fatigue following a Wingate test: peripheral and central mechanisms. Scand J Med Sci Sports 23, Froyd C, Millet GY & Noakes TD (2013). The development of peripheral fatigue and short-term recovery during self-paced high-intensity exercise. JPhysiol591, Gandevia SC (2001). Spinal and supraspinal factors in human muscle fatigue. Physiol Rev 81, Gandevia SC, Allen GM, Butler JE & Taylor JL (1996). Supraspinal factors in human muscle fatigue: evidence for suboptimal output from the motor cortex. JPhysiol490, Goodall S, González-Alonso J, Ali L, Ross EZ & Romer LM (2012). Supraspinal fatigue after normoxic and hypoxic exercise in humans. JPhysiol590, Goodall S, Romer LM & Ross EZ (2009). Voluntary activation of human knee extensors measured using transcranial magnetic stimulation. Exp Physiol 94, Goodall S, Ross EZ & Romer LM (2010). Effect of graded hypoxia on supraspinal contributions to fatigue with unilateral knee-extensor contractions. JApplPhysiol109, GruetM,TemesiJ,RuppT,LevyP,MilletGY&VergesS (2013a). Stimulation of the motor cortex and corticospinal tract to assess human muscle fatigue. Neuroscience 231, Gruet M, Temesi J, Rupp T, Millet GY & Verges S (2013b). Effect of different approaches to target force on transcranial magnetic stimulation responses. Muscle Nerve 48, HiltyL,LutzK,MaurerK,RodenkirchT,SpenglerCM, Boutellier U, Jäncke L & Amann M (2011). Spinal opioid receptor-sensitive muscle afferents contribute to the fatigue-induced increase in intracortical inhibition in healthy humans. Exp Physiol 96, Hoffman BW, Oya T, Carroll TJ & Cresswell AG (2009). Increases in corticospinal responsiveness during a sustained submaximal plantar flexion. JApplPhysiol107, Hunter SK, Butler JE, Todd G, Gandevia SC & Taylor JL (2006). Supraspinal fatigue does not explain the sex difference in muscle fatigue of maximal contractions. JApplPhysiol101,

12 1064 M. Gruet and others Exp Physiol 99.8 (2014) pp Hunter SK, Todd G, Butler JE, Gandevia SC & Taylor JL (2008). Recovery from supraspinal fatigue is slowed in old adults after fatiguing maximal isometric contractions. JAppl Physiol 105, Iguchi M & Shields RK (2012). Cortical and segmental excitability during fatiguing contractions of the soleus muscle in humans. Clin Neurophysiol 123, Lévénez M, Garland SJ, Klass M & Duchateau J (2008). Cortical and spinal modulation of antagonist coactivation during a submaximal fatiguing contraction in humans. J Neurophysiol 99, Martin PG, Smith JL, Butler JE, Gandevia SC & Taylor JL (2006). Fatigue-sensitive afferents inhibit extensor but not flexor motoneurons in humans. JNeurosci26, Mathis J, de Quervain D & Hess CW (1998). Dependence of the transcranially induced silent period on the instruction set and the individual reaction time. Electroencephalogr Clin Neurophysiol 109, Matthews PB (1999). The effect of firing on the excitability of a model motoneurone and its implications for cortical stimulation. JPhysiol518, Merton PA (1954). Voluntary strength and fatigue. JPhysiol 123, Millet GY & Lepers R (2004). Alterations of neuromuscular function after prolonged running, cycling and skiing exercises. Sports Med 34, Place N, Bruton JD & Westerblad H (2009). Mechanisms of fatigue induced by isometric contractions in exercising humans and in mouse isolated single muscle fibres. Clin Exp Pharmacol Physiol 36, Place N, Lepers R, Deley G & Millet GY (2004). Time course of neuromuscular alterations during a prolonged running exercise. Med Sci Sports Exerc 36, Rothwell JC, Thompson PD, Day BL, Boyd S & Marsden CD (1991). Stimulation of the human motor cortex through the scalp. Exp Physiol 76, RuppT,JubeauM,WuyamB,PerreyS,LevyP,MilletGY& Verges S (2012). Time-dependant effect of acute hypoxia on corticospinal excitability in healthy humans. J Neurophysiol 108, Sidhu SK, Bentley DJ & Carroll TJ (2009a). Cortical voluntary activation of the human knee extensors can be reliably estimated using transcranial magnetic stimulation. Muscle Nerve 39, Sidhu SK,Bentley DJ & Carroll TJ (2009b). Locomotor exercise induces long-lasting impairments in the capacity of the human motor cortex to voluntarily activate knee extensor muscles. JApplPhysiol106, Sidhu SK, Cresswell AG & Carroll TJ (2012). Motor cortex excitability does not increase during sustained cycling exercise to volitional exhaustion. JApplPhysiol113, Sidhu SK, Lauber B, Cresswell AG & Carroll TJ (2013). Sustained cycling exercise increases intracortical inhibition. Med Sci Sports Exerc 45, Smith JL, Martin PG, Gandevia SC & Taylor JL (2007). Sustained contraction at very low forces produces prominent supraspinal fatigue in human elbow flexor muscles. JAppl Physiol 103, Søgaard K, Gandevia SC, Todd G, Petersen NT & Taylor JL (2006). The effect of sustained low-intensity contractions on supraspinal fatigue in human elbow flexor muscles. JPhysiol 573, Szubski C, Burtscher M & Löscher WN (2007). Neuromuscular fatigue during sustained contractions performed in short-term hypoxia. Med Sci Sports Exerc 39, Taylor JL, Allen GM, Butler JE & Gandevia SC (2000). Supraspinal fatigue during intermittent maximal voluntary contractions of the human elbow flexors. JApplPhysiol89, TemesiJ,ArnalPJ,DavrancheK,BonnefoyR,LevyP,VergesS & Millet GY (2013). Does central fatigue explain reduced cycling after complete sleep deprivation? Med Sci Sports Exerc 45, Temesi J, Gruet M, Rupp T, Verges S & Millet GY (2014a). Resting and active motor thresholds versus stimulus response curves to determine transcranial magnetic stimulation intensity in quadriceps femoris. J Neuroeng Rehabil 21, 40. Temesi J, Rupp T, Martin V, Arnal PJ, Feasson L, Verges S & Millet GY (2014b). Central fatigue assessed by transcranial magnetic stimulation in ultratrail running. Med Sci Sports Exerc 46, Todd G, Taylor JL & Gandevia SC (2003). Measurement of voluntary activation of fresh and fatigued human muscles using transcranial magnetic stimulation. JPhysiol551, Additional Information Competing interests None declared. Author contributions All experiments were performed at the Exercise Physiology Laboratory, Université de Lyon, F-42023,Saint-Etienne,France. Each author contributed to all aspects of the study and all authors approved the final version. Funding M.G. was supported by a grant from the Fonds de Dotation Recherche en Santé Respiratoire, appel d offres 2011 for research on the brain during exercise. J.T. was supported by a doctoral research grant from the Rhône-Alpes Region. Acknowledgements We sincerely acknowledge the assistance of Dr Léonard Féasson and Dr Pascal Edouard for conducting the medical inclusions and Régis Bonnefoy for his technical expertise.