Introduction

Surface electromyography (EMG) is widely used to quantify the level of activation of working muscles and to detect the prodromic signs of muscle fatigue. For instance, an increase in the root mean square (RMS), an index of global EMG activity, is thought to reflect the recruitment of additional motor units as well as an increase in motor unit rate coding to compensate for the deficit of contractility resulting from the impairment of fatigued motor units (Moritani and DeVries 1978). In incremental cycling exercise, several EMG studies have reported for the quadriceps muscles one (Nagata et al. 1981; Chwalbinska-Moneta et al. 1998) or two (Lucia et al. 1999) EMG thresholds (EMGTh) according to the status of the subjects. These EMG thresholds are explained by an increased contribution of fast twitch motor units to maintain the required energy supply for contraction (Moritani and deVries 1978). However, others studies failed to report EMG thresholds in a similar testing situation (Taylor and Bronks 1995).

In order to correlate neuromuscular, metabolic and ventilatory exercise-induced changes, some studies have investigated the EMG responses to a cycling incremental test performed until exhaustion, and the metabolic and ventilatory changes associated with the increase in power output (i.e. lactate and ventilatory thresholds). Indeed, one would expect the metabolic and ventilatory changes associated with the aerobic–anaerobic transition phase to be associated with muscular changes reflected by non-linear increases in RMS. Following this line, Lucia et al. (1999) showed that the first and the second EMG thresholds were correlated with the first and the second ventilatory thresholds.

However, Jorge and Hull (1986) recorded the EMG activity of eight lower limb muscles during cycling exercise. Each muscle seems to accomplish a specific task according to the subject's capability to push and to pull up the pedal. In studies reporting EMG thresholds, these non-linear changes were not reported for all muscles involved in pedalling but rather for the vastus lateralis (Nagata et al. 1981; Lucia et al. 1999), rectus femoris (Chwalbinska-Moneta et al. 1998; Lucia et al. 1999) and soleus (Chwalbinska-Moneta et al. 1998).

The aim of the present study was first to verify the existence of two EMG thresholds in eight lower limb muscles in a homogenous population of professional road cyclists and, second, to examine the possibility of a simultaneous occurrence of the EMG and ventilatory thresholds.

Methods

Eight professional road cyclists participated in this study [age: 24 (1) years, height: 182 (3) cm, body mass: 74 (4) kg]. They had 11 (3) years experience of competition and had covered an average of 30,000 km (including training and competitions) in the previous season. All subjects volunteered to enter this experimental design approved by the local Ethics Committee and gave their informed consent. They had no recent or old pathology of limb muscles and joints and they were instructed to refrain from intense training during the 2 days before testing. Three of them were members of the French team in the World Championships for professional cyclists in 2002.

The protocol was performed on an electrically braked cycloergometer (Excalibur sport, Lode, The Netherlands). Subjects realized a progressive test during which all respiratory and ventilatory parameters [oxygen uptake (V̇O2), carbon dioxide output (V̇CO2), respiratory exchange ratio (RER), and minute ventilation (V̇ E)] were measured. Starting at 100 W, the workload was increased by 26 W·min−1. Subjects chose their pedalling rate (rpm) freely and wore sport shoes with clipless pedals. This incremental exercise was stopped when the power output could not be maintained.

Throughout the incremental exercise trial, software (Oxycon Beta, Hellige, Germany) computed breath-by-breath data of V̇ E, V̇O2, V̇CO2, and the ventilatory equivalents for O2 (V̇ E·V̇O2 −1) and CO2 (V̇ E·V̇CO2 −1). The first ventilatory threshold (VT1) corresponded to the power output value at which V̇ E·V̇O2 −1 exhibited a systematic increase without a concomitant increase in V̇ E·V̇CO2 −1 and the second ventilatory threshold (VT2) was determined by using the criterion of an increase in both V̇ E·V̇O2 −1 and V̇ E·V̇CO2 −1. Two independent observers detected VT1 and VT2 following the criteria previously described.

During the test, electromyographic activity was continuously recorded from eight lower limb muscles: vastus lateralis (VL), rectus femoris (RF), vastus medialis (VM), semimembranosus (SM), biceps femoris (BF), gastrocnemius lateralis (GL), gastrocnemius medialis (GM) and tibialis anterior (TA). A pair of surface electrodes (Universal Ag/AgCl electrodes, Contrôle Graphique Medical, France) was attached to the skin with a 2 cm inter-electrode distance. The electrodes were placed longitudinally with respect to the underlying muscle fibre arrangement. Prior to electrode application, the skin was shaved and abraded using sandpaper and cleaned with alcohol in order to minimize the impedance. The wires connected to the electrodes were well secured with tape to avoid artefacts from leg movements.

EMG was recorded with a ME3000P8 amplifier (Mega Electronics, Finland). The raw EMG signals were band pass filtered between 20 and 480 Hz, amplified, an analog-to-digital converted at a sampling rate of 1 kHz. An EMG power spectral density was computed for 2-s sampling periods at fixed intervals throughout the tests, using a Fast Fourier Transform algorithm and the RMS contained in each 2-s spectrum (corresponding to three revolutions at 90 rpm) was calculated in µV. The power output value at the first and the second EMG thresholds (EMGTh1 and EMGTh2) were determined as follows: EMGTh1 and EMGTh2 as the first and second non-linear increases in RMS respectively. Two independent observers detected EMGTh1 and EMGTh2.

Statistical analysis

Results are expressed as mean (standard error, SEM). The level of significance was set at p<0.05. A Fisher's exact test was made to compare the percentage of EMG threshold occurrence between the muscles. A repeated measures analysis of variance was used to compare the power output (watts) of occurrence of EMG thresholds for all the recorded muscles. Moreover, for the eight recorded muscles, repeated measures of analysis of variance were used to determine if there was a significant difference between the EMG and ventilatory thresholds expressed in power output (watts) and as a percentage of maximum power output (W max).

Results

Professional cyclists achieved a maximal oxygen consumption, i.e. V̇O2max of 5.4 (0.5) l·min−1 [i.e. 74.6 (2.5) ml·min−1·kg−1, range: 67.8–82.4 ml·min−1·kg−1] and a W max of 475 (30) W (range: 438–516 W).

For each muscle, the percentage of EMGTh1 and EMGTh2 occurrence is depicted in Table 1. The Fisher's exact test revealed no significant difference in the percentage of occurrence of the EMG thresholds. There is no significant difference between the muscles for the power output at which the two EMG thresholds occurred. EMGTh1 and EMGTh2 occurred at 52 (2)% and 86 (1)% of W max, respectively, while VT1 and VT2 occurred at 62 (9)% and 89 (7)% of W max. Thus, a significant difference existed between VT1 and EMGTh1 for all recorded muscles, EMGTh1 occurring before VT1. In contrast, no significant difference existed between EMGTh2 and VT2. An example of EMG and ventilatory threshold determination is depicted in Fig. 1.

Table 1 Percentage of occurrence of EMG thresholds in the eight recorded muscles
Full size table
Fig. 1A, B
figure 1

Individual example of evolutions in ventilatory equivalents (V̇ E·V̇O2 −1, V̇ E·V̇CO2 −1) (A) and RMS (B) during incremental pedalling exercise as a percentage of maximum power output (W max ). Note that VT2 and EMGTh2 are synchronized while VT1 and EMGTh1 are not. (EMG Th2 Second EMG threshold, RMS root mean square, V̇E minute ventilation, V̇CO2 carbon dioxide production, V̇O2 oxygen uptake)

Full size image

Discussion

To our knowledge this is the first study to examine, in a homogeneous population of professional road cyclists, the occurrence of EMGTh1 and EMGTh2 for eight lower limb muscles. Our results showed at least the occurrence of EMGTh1 in 50% (GL) of the subjects and in 63% (GM) for EMGTh2 (Table 1). EMGTh1 occurred significantly before VT1. In contrast, EMGTh2 and VT2 occurred at the same time.

Based on the assumption of an EMGTh, Lucia et al.(1999) recorded in an elite cyclist population the EMG response of VL and RF, i.e. muscles that are most implicated in pedalling activity. In addition to the classic ventilatory (VT1 and VT2) and lactic (LT1 and LT2) thresholds, they reported non-linear increases of RMS representing two breakpoints named EMG thresholds 1 and 2. They showed no significant difference between EMGTh1, VT1 and LT1 (the first lactic threshold) and between EMGTh2, VT2 and OBLA ([La]=4 mmol·l−1). In the present study, two EMG breakpoints named EMGTh1 and EMGTh2 were found during incremental exercise, and occurred at an exercise intensity of 52 (2)% and 86 (1)% of W max, respectively. In agreement with the results of Lucia et al. (1999), no difference was found between EMGTh2 and VT2. However, differing from the results of Lucia et al. (1999), EMGTh1 occurred significantly before VT1 in the eight recorded muscles. This may be explained by the fact that VT1 occurrence is the result of the onset of lactate production/utilization imbalance in the blood gathering from the involved contracting muscles among which some may produce lactate earlier than some others. On the other hand, EMG signals are a more direct reflection of the muscles involved. In other words, there is a possibility of detecting each individual muscle's EMG thresholds. This cannot be achieved when using a gas exchange method, which only reflects the averaged or overall metabolic consequences of all the muscles involved. Moreover, we showed in a previous study (Hug et al. 2003) that a EMG sign of neuromuscular fatigue occurred significantly before VT1 suggesting that the activation of muscle metaboceptors (i.e. type III and IV muscular afferents) might make a major contribution to the mechanisms responsible for the ventilatory changes leading to VT1.

In this present study, two EMG thresholds occurred simultaneously in muscles that were clearly different in their fibre type composition (there was no significant difference between the eight muscles in the occurrence of EMGTh1 and EMGTh2). Thus, in accordance with Lucia et al. (1999), we may conclude that, in professional cyclists, changes in the pattern of motor unit recruitment are not affected by the muscle fibre type composition. Statistical analysis revealed no significant inter-muscle differences concerning the percentage of EMG threshold occurrence. However, as shown in Table 1, EMGTh1 was determined in 100% of the subjects for VL but only in 50% for GL. This lack of significance is likely due to the small sample of the tested population (i.e. eight professional road cyclists). The measurement of the EMG response in VL and BF permitted the determination of two EMG thresholds in 100% of the subjects. In agreement with the results of Jorge and Hull (1986), these muscles are some of those most highly involved in cycling. Moreover, Takaishi et al. (1998) showed a particular pedalling skill adopted by trained cyclists that required a high level of involvement of the BF muscle. This result might explain the high percentage of occurrence of the EMG thresholds in these two muscles.

Conclusion

In professional road cyclists, the EMG activities of eight lower limb muscles involved in pedalling showed two non-linear increases (i.e. EMGTh1 and EMGTh2). The first EMG threshold preceded the first ventilatory threshold. Inversely, the second EMG and ventilatory thresholds occurred at the same percentage of power output.

These results suggest that the use of EMG may be a useful non-invasive method for detecting the second ventilatory threshold in most of the muscles involved in cycling exercise.