Amplitude and spectral characteristics of biceps Brachii sEMG depend upon speed of isometric force generation

Journal of Electromyography and Kinesiology 13 (2003) 139–147 www.elsevier.com/locate/jelekin

Amplitude and spectral characteristics of biceps Brachii sEMG depend upon speed of isometric force generation P. Sbriccoli a, I. Bazzucchi a, A. Rosponi d, M. Bernardi b, G. De Vito c, F. Felici a,∗ b

a University Institute of Motor Sciences, Piazza Lauro De Bosis 15, I-00194 Rome, Italy Post-graduate School of Specialization in Sport Medicine, Faculty of Medicine, University of Rome “La Sapienza”, P. le Aldo Moro 5, I00185 Rome, Italy c Scottish School of Sport Studies, Strathclyde University, Glasgow, UK d IRCCS Santa Lucia Foundation Rehabilitation Hospital, Via Ardeatina 306, I-00179 Rome, Italy

Received 4 October 2002; accepted 24 October 2002

Abstract In the present study the influence of speed of contraction on the interplay between recruitment and firing rate of motor units (MUs) was assessed. The surface electromyographic (sEMG) signal was recorded in nine healthy subjects from the right biceps brachii using a linear electrode array during ramp isometric contractions (from 0 to 100% of the maximal voluntary force, MVC) at 5, 10, and 20% MVC s-1 (ramp phase), followed by 10 s of sustained MVC (hold phase). The median frequency (MDF), Root Mean Square (RMS) and conduction velocity (CV) of sEMG, were computed on adjacent epochs covering a force range of 5% MVC each. Full motor unit recruitment (FMUR) point was assessed as the force level at which MDF reached its maximum value; the MDF decay during the hold phase was taken as an index of localized muscle fatigue. At 5% MVC s-1, FMUR was reached at 52.3% MVC. At 10%MVC s-1 FMUR was achieved at 58% MVC; while at 20% MVC s-1 FMUR point was located at 77% MVC, being statistically different from 5 and 10% MVCs-1 ramps (p⬍0.05). The MDF decay was steeper at higher speed. CV modifications mirrored those reported for MDF. The RMS increased in a curvilinear fashion and the maximum value was always attained during the hold phase. Our findings suggest that MU recruitment strategies are significantly related to the speed of contraction even in a single muscle.  2003 Elsevier Science Ltd. All rights reserved. Keywords: sEMG; Isometric contraction; Ramp; Motor control strategies

1. Introduction The term “motor unit activation” refers to the combination of recruitment and rate coding of motor units (MU) within muscles [25]. The relative role of these two variables in controlling force production has been debated since the early seventies [7,22]. In order to achieve a maximal voluntary contraction (MVC), all MUs must be recruited and then driven at the appropriate rate. Corresponding author. Tel.: +39-06-3673-3540; fax: +39-063673-3214. Abbreviations: sEMG, surface electromyography/gram; MU, motor unit; RMS, root mean square; MDF, median frequency; FMUR, full motor unit recruitment; CV, conduction velocity; BB, biceps brachii; MVC, maximal voluntary contraction E-mail address: [email protected] (F. Felici). ∗

It has been shown that full MU recruitment (FMUR) is obtained at different percentages of the MVC depending on factors such as muscle size, fiber type composition, and muscle function [25]. It is difficult to assess the precise association between muscle function and strategy of organization of MU activity [10,16] even using needle electromyography (EMG). However, modeling work [16] as well as experimental evidence from animal [28] and human studies [18,22] clarified some of the underlying rules. In feline muscle Solomonow et al. [28], experimentally reproduced MU recruitment according to the size principle proposed by Henneman et al. [17]. They showed that the median frequency (MDF) of the EMG power spectrum increased linearly with orderly recruitment of MUs until FMUR was obtained and that the MDF was not influenced by the firing rate of active MUs. Under this con-

1050-6411/03/$ - see front matter  2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S1050-6411(02)00098-6

140

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

trolled experimental situation, the conclusion holds that EMG frequency content is not significantly affected by MU firing rates and, therefore, the increase in the average muscle fiber conduction velocity (CV) during recruitment is the major contributor to variations in MDF. However, as also pointed out by the authors [28], the extrapolation of their results to surface EMG data (sEMG) can be questionable. To our knowledge, only one paper [12] studied the assumed linear correlation between CV and MDF during simulated and real voluntary ramp isometric contractions. These authors showed that the influence of volume conductor could be an important confounding factor when assessing FMUR on the basis of spectral variables alone. However, due to the limitation of the model adopted, the effects of the concomitant manifestations of myoelectric fatigue were not taken into account. Despite these limitations, the time and frequency domain analysis of sEMG is a common method used to assess non-invasively MU activation during increasing force voluntary isometric contractions, although this approach cannot distinguish the contribution of individual MUs. Several studies that have investigated sEMG changes during contractions performed in a ramp fashion [2,3,23,26] observed that during voluntary isometric rapid contractions of the elbow flexors (from 0 to 100% MVC in 3 s), FMUR ranged between 65 and 85% MVC, as revealed by the MDF, depending upon the individual skill in performing the exercise. Differences in MU recruitment in relation to different muscle actions (i.e. acting as agonists or antagonists) were also investigated on different muscle groups [3,4]. The main purpose of the present study was to investigate if the interplay between MU recruitment and firing rate can be affected by changes in the speed of contraction during ramp fashion muscle contractions. Since the assessment of FMUR by means of sEMG spectral data was questioned, we also tested our results by means of simultaneous computation of spectral variables and CV data. Besides, we investigated the influence of myoelectric fatigue on the behavior of time and spectral sEMG parameters and average muscle fibers CV during the sustained phase of isometric contractions.

fied chair (see Fig. 1, Panel A). The subject’s shoulder was secured through a strap to the modified chair back. Subjects performed isometric contractions using an anatomic device that allowed the elbow angle to be fixed at 90° as previously described [13]. The arm chosen for measurements was the right, which in all cases was the dominant one. The hand was maintained halfway between pronation and supination. Prior to the experimental session (see below), subjects familiarized themselves in performing MVC and ramp contractions of the elbow flexors from 0 to 100% MVC at different speeds (5, 10 and 20% MVC s-1). Ramp duration was, thus, different among speeds: 20, 10, and 5 s from slow to fast ramps. The desired force output and ramp trajectories were displayed on a computer screen along with the output of the force transducer, providing continous feedback for the subject. Subjects performed a different number of training trials in order to execute the exercise correctly at any given speed. 2.2. Experimental protocol Each subject was requested to attend the laboratory for five experimental sessions, each separated by at least two days. During each experimental session, subjects performed one attempt at any speed of contraction. Attempts were separated one another by 30 min. A control MVC was measured before each ramp and, when motor performance was different from the target, i.e. subjects were not able to reach the requested force level, the interval between trials was protracted until complete recovery was obtained. During the session the three speeds of contraction were administered in random order. This procedure was adopted in order to minimize the cumulative effect of fatigue. Therefore, five attempts

2. Materials and methods 2.1. Subjects and preliminary sessions Nine subjects (five males and four females) participated in this study (age: 29 ± 7 years; stature: 1.69 ± 0.88 m; body mass: 60 ± 15 kg). None had any previous history of neuromuscular disorder and each gave written informed consent prior to the experiment. Subjects were seated comfortably in a custom-modi-

Fig. 1. Panel A: Experimental set-up. Panel B: Four bars linear array. SD2: Central single differential channel; DD1 and DD2, double differential channels.

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

at each speed of contraction were accepted according to the criteria proposed by Bernardi et al. [2]. However, trials where there was no correspondence between the maximum sEMG RMS and maximum strength were not considered. Furthermore, trials were rejected either when motor performance was different from the target, or when the individual force did not reach the 100% MVC and therefore only in four cases (two subjects) it was necessary to repeat a single trial 2.3. MVC and sEMG measurements The MVC task consisted of rapidly increasing the force exerted and to follow the performance on the computer screen while being encouraged to achieve a maximum, and to maintain it for at least 5 s before relaxing. Three maximal attempts were performed, separated by 5 min to recover from fatigue, and the best performance was chosen for further analysis. A further check on MVC was done in accordance with the procedure indicated by Baratta et al. [1]. In short, after MVC determination, subjects were asked to reach and follow a line set at 10% higher than the measured MVC. In two cases, subjects exceeded the previous MVC level; in this case the MVC was updated. The sEMG was recorded from the long head of right biceps brachii (BB) during linearly increasing isometric contractions from 0 to 100% MVC (ramp phase) and during the subsequent 10 s in which the MVC was sustained (constant force phase). Before sEMG recording the BB motor point(s) was identified by means of electrical stimulation of the skin overlying the muscle according to the procedure described by Farina et al. [11]. After skin abrasion and cleaning with ethyl alcohol, a 4 silver bar linear array electrodes (5 mm long, 1 mm diameter, interelectrode distance: 10 mm) was placed by interposition of conductive gel over the longitudinal axis of the BB muscle between the motor point and the distal tendon. The contour of the electrode was drawn on the skin with dermographic ink to ensure correct repositioning of the electrode in the following sessions. The reference electrode was placed over the olecranon; sEMG signals were amplified using a portable electromyograph (SATEM, mod. VD 10/4, pass band 10–1000 Hz). The sEMG signal was detected in a single differential mode using the configuration described on Fig. 1B. Three single differential sEMG signals (SD1, SD2, SD3) were recorded. Two double differential sEMG signals (DD1 and DD2, interelectrode distance: 10 mm) were then computed and used for CV calculation as described in the following paragraphs. Force and sEMG signals were A/D sampled at 2048 points per second at 12-bit resolution (DAQ card AI-16XE-50, National Instruments, TX) and the raw signals were stored on a PC for further analyses.

141

2.4. sEMG data analysis Time and frequency domain analyses were performed over the central single differential channel (SD2 in Fig. 1B). The MDF and the RMS values were computed during the ramp and constant force phases. In order to cope with the non-stationarity introduced by the ramp phase, the length of sEMG signals epochs was adapted to the different speeds of contraction [5]. The length of sEMG epochs was chosen as such that the correspondent force variation was always equal to 5% MVC. Thus, window lengths were: 250 ms for 20% MVC s-1, 500 ms for 10% MVC s-1 and 1000 ms for 5% MVC s-1 ramps respectively. Subsequent epochs were overlapped one another by half their length, thus allowing a force window resolution of 2.5% MVC. This procedure also produced the same number of samples describing the ramp phase, thus allowing a convenient comparison of ramps of different duration. When appropriate (10–20% MVC s-1 ramps) epochs were zero padded up to 2048 points to obtain a frequency resolution of 1 Hz [2]. For each trial, MDF and RMS data were normalized with respect to the local maximum value obtained thus allowing appropriate grouping (grand averages) of contractions performed at different speeds and by different subjects. Finally, grand averages were normalized to their maximum. 2.5. Conduction velocity (CV) Muscle fiber action potential CV was assessed by means of cross-correlation function [14] between the two double differential channels, DD1 and DD2 in Fig. 1B, using the same segmentation as for sEMG analysis. This method assumes that the time delay between two similar but not identical signals is the amount of time shift that must be applied to one of the signals to minimize the mean square error with the other [27]. This time shift is the same that maximizes the cross-correlation between the two signals. The cross correlation Rxy(t) of the signals x(t) and y(t) is defined as:

冕

⫹⬁

Rxy(t) ⫽ x(t)丢y(t) ⫽

x(t)y(t ⫹ t)dt

⫺⬁

where the symbol 丢 denotes correlation. Estimates of CV were accepted only when the EMG cross-correlation function values were higher than 0.7. 2.6. Full MU recruitment (FMUR) and myoelectric fatigue FMUR was considered as the force value (%MVC) at which the highest MDF value was reached [28]. To test

142

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

the dependency of FMUR on the speed of contraction, a one-way analysis of variance [24] was used on FMUR points obtained from all subjects at 5, 10 and 20% MVCs-1 and the level of significance was set at p ⬍ 0.05. FMUR during ramp phase was also compared with the point at which average CV reached a maximum. The level of agreement between MDF and CV data was tested by means of a paired Student’s t-test [24]. A linear regression of MDF data was performed over the 10 seconds of the constant force phase. According to Felici et al. [15] , for each individual subject, the linearity of the regression computed on the set of five curves at any given speed, was tested by means of an appropriate F-test. The slopes of the regression lines thus obtained were normalized with respect to their intercepts and used as indicator of localized muscle fatigue [21] and expressed as percentage per second. To test the dependency of MDF slopes on the speed of contraction, a one-way analysis of variance [24] was used on data points obtained from all subjects at 5, 10 and 20% MVC s-1. An alpha level of p ⬍ 0.05 was adopted as significant for all computations.

3. Results Prior to the experimental session, subjects performed on average 6±2 training ramps at any given speed. Each subject reported the ramp at 5% MVC s-1 as the most demanding in terms of fatigue; however, it was the faster ramps that took more attempts (three on average) to be matched precisely. The achievement of familiarization was assessed having the subject performing the ramp at the desired speed (visual inspection) and using as primary mover the BB muscle. During the actual experiment (see below), in accordance with our criteria for rejection, only four trials in two subjects were discarded and the test repeated after a convenient recovery time. In three cases this was due to subjects’ fatigue (inability to reach the force target), and in one case because the biceps brachii was not the prime mover muscle. All data are grand averages ± SD of 45 tests at any given speed. 3.1. Force MVC force did not vary among different sessions. The time course of force at different speeds is reported in Figs. 2–4 for ramp speeds of 5% MVC s-1, 10% MVC s-1 and 20% MVC s-1 respectively. Data are reported as mean curves obtained on all trials in all subjects. It is evident that in all cases subjects were able to perform the ramps correctly at each speed of contraction. During the constant force phase, the MVC was maintained up

to the end of the exercise (10 s of attempted constant force) at 10 and 20% MVC s-1 contractions. Differently, at the slower speed (5% MVC s-1), subjects could just reach the MVC, but could not maintain the force level requested, as indicated by the standard deviation values. Notably, the MVC was reached with a certain delay with respect to the ideal force track, probably due to a “braking action” close to the point at which force must be kept constant. This is progressively more evident from lower to higher speed of contraction. However, even at 20% MVC s-1, at the expected time the force value was very close to 100% MVC (from 99 to 93%). 3.2. RMS The RMS increased in a curvilinear fashion following the force increment at each speed of contraction up to the MVC. It is to note that at any given force level during the ramp phase, RMS value did not change among the three speed (Fig. 5). Maximum RMS value was reached always after the MVC. During slower ramps (5% MVC s-1; Fig. 2) the RMS showed a first increase up to 50% MVC (10th second), thereafter the RMS increased with a different slope until the end of the ramp phase. This change in the rate of RMS increase was also present at the other speeds of contraction (see Figs. 3 and 4), although it was less evident. During the constant force phase, the RMS slightly continued to increase until the end of the contraction at 10% and 20% MVC s-1, whilst at 5% MVC s-1 the RMS reached its maximum after the beginning of the constant force phase and then started decreasing. 3.3. CV The average CV is presented in Figs. 2–4 for ramp speeds of 5, 10 and 20% MVC s-1, respectively. The BB CV increase during the ramp phase parallels that of MDF. During slower contractions CV maximum is reached at 52.3 %MVC ± 6.8, at intermediate speed (10% MVC s-1) maximum CV value is attained at 56.1 %MVC ± 5.4, while at 20% MVC maximum CV is located at 85% MVC ± 9.3. Normalized CV decay during the constant force phase was steeper at 20% MVC s-1 [–1.45% ± 0.2] than those obtained at 10% MVC s-1 [–1.08% ± 0.2], and 5% MVC s-1 [–1.1% ± 0.15] (p ⬍ 0.05). 3.4. MDF—FMUR During the ramp phase it is evident a first phase during which the MDF increases with force in a linear fashion and a second phase during which progressively slows down to flatten in a plateau and then starts to decrease. The presence of the plateau made difficult a precise identification of the FMUR point, especially at 5% MVC

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

143

Fig. 2. Time course of RMS (triangles), MDF (open circles), force (continuous line), and CV (filled circles) at 5% MVC s-1. Curves are averages (±SD) of all tests performed by all subjects and are normalized to their maximum value. Vertical line marks the end of ramp phase.

Fig. 3. Time course of RMS (triangles), MDF (open circles), force (continuous line), and CV (filled circles) at 10% MVC s-1. Curves are averages (±SD) of all tests performed by all subjects and are normalized to their maximum value. Vertical line marks the end of ramp phase.

s-1. In accordance with other authors [28] we choose FMUR point as the force level at which the highest MDF value occurred. At 5% MVC s-1 FMUR was reached on average at 52.3% ± 7.3%MVC. At intermediate speed (10% MVCs-1) FMUR was obtained at 58.63% ± 15.1%MVC, which was not statistically different with respect to 5% MVC s-1 ramps (p ⫽ 0.19). At 20% MVC s-1, FMUR was achieved on average at 77.6% ± 15.1%MVC that was statistically different from 5% and 10% MVC s-1 (p ⬍ 0.01). These results demonstrated a strong dependence of FMUR from the speed of force increase (p ⬍ 0.01). At FMUR point, RMS was equal to 38% of its

maximum value at 5% MVC s-1, 47% at 10% MVC s, and 58% at 20% MVC s-1. Indeed, CV at FMUR point was 100, 95.68 and 98% of its maximum value at 5, 10, and 20% MVC s-1, respectively. MDF decay obtained during the static phase of contraction increased from 5% MVC s-1 to 10% and 20% MVC s-1 contractions. Normalized slopes of the linear regressions computed over the steady state force phase, were equal to –1.64 ± 0.1 %s ⫺ 1 at 5% MVC s-1, –2.24 ± 0.5 % s ⫺ 1 at 10% MVC s-1 and –2.74% ± 0.6 at 20% MVC s-1. Values at 10 and 20% MVC s-1 were statistically different from that at 5% MVC s-1 (p ⬍ 0.05). The comparison between FMUR point, estimated as

1

144

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

Fig. 4. Time course of RMS (triangles), MDF (open circles), force (continuous line), and CV (filled circles) at 20% MVC s-1. Curves are averages (±SD) of all tests performed by all subjects and are normalized to their maximum value. Vertical line marks the end of ramp phase.

Fig. 5. RMS vs Force. All normalized RMS mean values (N=120) obtained on all subjects during the ramp phase at each speed of contraction are plotted versus normalized force data.

MDF maximum value and CV maximal values, failed to show significant difference at the considered speed of contraction. In addition, no difference was found, among the three contractions, in the rate of MDF decay and CV reduction measured during constant force phases.

4. Discussion The present study was designed to investigate the dependency of MU recruitment strategy and rate coding from the speed of force development during ramp isometric contractions. The main results we obtained seem to suggest that, in a single muscle, MU recruitment and

rate coding are engaged in different proportions as the speed of isometric force increase changes. No indications about modification of the size principle in the present experimental set-up have been provided. Finally, a curvilinear relationship between percentage of isometric force and percentage of RMS value of sEMG has been shown. Considering that results at 10% MVCs-1 during the ramp phase were, in no instance, statistically different from those obtained at 5% MVC s-1, the following discussion will refer only to data obtained at 5% and 20% MVC s-1. Based on the assumption of a linear correlation between MDF and CV [28], it was stated that MDF increase over contraction reflects MU recruitment. From this point of view, our results indicate that the behavior of MDF and average muscle CV considering both slow and fast ramp speeds can be very similar. This result is also supported by the observation that MDF and CV provide similar estimate of FMUR in both slow and fast ramp contractions. Thus it seems reasonable to infer that during ramp contractions spectral sEMG data provided a reliable estimate of FMUR. We obtained different values of FMUR at different speeds of contraction, being the FMUR point closer to the end of the ramp phase as the speed of contraction increases. How could the displacement of FMUR point toward lower force levels during slower ramps be explained? A recent study [6], investigated individual MU activation pattern in the first dorsal interosseus muscle during a fatigue test performed at 50% MVC. Although the order of MU recruitment was not modified during fatigue, the authors found that the force value at which high-threshold MUs were recruited was

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

decreased, and that high-threshold MUs were activated at lower forces by the end of the fatigue test. In accordance with these findings, if new high-threshold MUs are recruited earlier during slower ramps, this should appear also as earlier manifestation of sEMG spectral compression due to myoelectric fatigue. The rate of MDF and CV decay during the hold (constant mechanical output) phase of the exercise was slower for the slower ramps with respect to faster ones. As reported in Fig. 2, at 5% MVC s-1 FMUR was reached on average at 52.3% MVC, which means approximately 10 s after the beginning of the exercise. During the remaining 10 s of the ramp, then, within the MU pool that is detected by the electrodes, myoelectric manifestations of fatigue will show up and eventually prevail. When the target MVC is reached, the faster, late recruited MUs are already fatigued. This implies also that from FMUR point, the additional force gain to reach MVC can be accomplished only by increasing the firing rate of the active MUs, as expressed by the RMS increase from the FMUR point to the end of the ramp phase (MVC). The modulation of the MU rate coding seems thus to be variable and depending upon the duration and intensity of fatiguing contraction [6]. The relative weight of MU recruitment and rate coding along the force axis seems also dependent upon the functional specialization of a given muscle [10,16]. It seems that muscles involved in small and precise movements find in the firing rate control the major factor in force modulation. By contrast, larger muscles seem to rely more on the recruitment mechanism to produce smooth contractions since each unit would add a much smaller relative increment to the total force [8,18]. This is the first report, to our knowledge, showing that also within a single muscle, the upper limit of MU recruitment, which denotes the transition to reliance on modulation of firing rate [10,16], is not rigidly fixed but there is a possibility for distributing the relative intervention of MU recruitment and rate coding in order to achieve the requested motor task and that this possibility is strongly dependent upon the extent of muscle fatigue. The curvilinear relationship between %RMS and %MVC obtained in the present study is not a new finding [9,19,23]. The BB muscle (long head) is a bifunctional one, i.e. it may contribute also to the supination of the forearm. Dupont et al. [9] have shown that the shape of the sEMG-torque relation obtained during ramp contraction is mostly curvilinear when the BB acts in supination or in a combination of flexion + supination. In accordance with Dupont et al. [9] work, in the present experiment the relationship became curvilinear above or around 50% MVC. We cannot exclude a certain degree of supination had taken place when the intensity of effort approached the maximum values. In addition, we also confirmed here results obtained by Lawrence and De Luca [19] in humans and by Solomonow et al. [29] in

145

animal preparations showing the invariance of this relationship irrespective of contraction speed. This suggests that the rate of increase of sEMG amplitude, which is mainly affected by rate coding of active MU, is adjusted according to the rate of force increase. Besides, in our experiments the RMS maximum value was always attained during the hold phase, suggesting the presence of a functional reserve in the central control of the rate coding mechanism. From this point of view, the most demanding task is represented by the slowest ramps: in this condition, during the hold phase the RMS started to decline along with force, indicating a high level of either myoelectric and mechanical muscle fatigue. Besides, our results expand on those from previous studies on animals, confirming that the sEMG vs force response depend on the MU recruitment and discharge rate strategies employed by the muscle [29]. It is evident that for a correct interpretation of the results of the present study, the adequacy and limitations of sEMG to be used as a tool for the detection of MU recruitment must be acknowledged. From this point of view, the relative location of MU within the muscle is an important issue to be discussed. High threshold MUs have been shown to be superficially located in the human BB muscle [7,20]. These MUs will contribute with larger potential to the sEMG signal, thus explaining that the normalized RMS increases larger than the normalized force. Similarly, it can be speculated that when a high threshold MU is recruited, which is located deep in the muscle, its territory may be outside the detection area of the electrodes, thus preventing its contribution to sEMG to be accounted for. However, as stated in the previous sections, we carefully checked for a continuous RMS increase during ramp performance. We cannot exclude this possibility from our data; however, given the assessment of a constant RMS increase with force output, this error, if present, should have exerted a negligible influence. As a matter of fact, the recruitment of a new MU may play a major role on spectral sEMG data. Although the relationship between spectral sEMG parameters and force increase during ramp contractions has already been tested [2,3,28], this topic is still under debate. Farina et al. [12], raised some questions about the limitations of spectral sEMG parameters in determining MU recruitment strategies. Their simulated recruitment data indicated that, according to volume conductor theory, any relationship between MU recruitment and spectral parameters changes could be masked by muscle anatomical and geometrical factors. In the case of a late recruited deep MU, then, its contribution to sEMG spectrum can even be in the direction of a decrease in its frequency content. On the other hand, simulated data indicated that CV could increase also when only rate coding is present. In summary, the filtering properties of the interposed tissues may mask the recruitment of a high threshold MU (no or adverse effect on MDF), while global muscle

146

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147

CV can still continue to increase [12]. Even if the above considerations must be considered, the applicability of these simulation results to real experimental data could have limitations due to the lack of information about the rate of force development in simulation and about the force level at which simulated MU recruitment ended. Moreover, modeling/simulation studies do not entirely consider factors related to muscle fiber localization within the muscle [7,20] and, as also pointed out by Farina et al. [12], to fatigue. In Farina et al. [12] actual experiments, the ramp speed was around 27% MVC s-1. In that condition CV maximum followed the MDF plateau of about 40% MVC. In our study a not significant difference of about 10% MVC between FMUR and CV maximum was observed during the fastest ramps (20% MVCs-1). These data seem to indicate that there is a trend for the progressive uncoupling of these two sEMG parameters as the speed of force increases, possibly due to a volume conductor effect. In conclusion, the above discussed results point to the primary role of the modulation of interplay between MU full recruitment and MU rate coding control mechanism during different isometric motor tasks even in a single muscle group. This may have important implications in the field of rehabilitation and training. The present observation must be extended to special populations, i.e. well trained power athletes and healthy elderly people. The role of afferent feedback on the modulation of the central control mechanism of the individual MUs remains to be investigated in more detail.

References [1] R.V. Baratta, M. Solomonow, D.H. Zhou, M. Zhu, Methods to reduce the variability of EMG power spectrum estimates, J. Electromyogr. Kinesiol. 8 (1998) 279–285. [2] M. Bernardi, F. Felici, M. Marchetti, F. Montellanico, M.F. Piacentini, M. Solomonow, Force generation performance and motor unit recruitment strategy in muscles of contralateral limbs, J. Electromyogr. Kinesiol. 9 (1999) 121–130. [3] M. Bernardi, M. Solomonow, G. Nguyen, A. Smith, R. Baratta, Motor unit recruitment strategy changes with skill acquisition, Eur. J. Appl. Physiol. 74 (1996) 52–59. [4] M. Bilodeau, A. Bertrand Arsenault, D. Gravel, D. Bourbonnais, EMG power spectra of elbow extensors during ramp and step isometric contractions, Eur. J. Appl. Physiol. 63 (1991) 24–28. [5] M. Bilodeau, C. Goulet, S. Nadeau, A. Bertrand Arsenault, D. Gravel, Comparison of the EMG power spectrum of the human soleus and gastrocnemious muscles, Eur. J. Appl. Physiol. 68 (1994) 395–401. [6] A. Carpentier, J. Duchateau, K. Hainaut, Motor unit behaviour and contractile changes during fatigue in the human first dorsal interosseus, J. Physiol. (Lond) 534 (3) (2001) 903–912. [7] H.P. Clamann, Activity of single motor units during isometric tension, Neurology 20 (1970) 117–128. [8] C.J. De Luca, R.S. Le Fever, M.P. Mc Cue, A.P. Xenakis, Behaviour of human motor units in different muscles during linearly varying contractions, J. Physiol. (Lond) 329 (1982) 113–128.

[9] L. Dupont, D. Gamet, C. Perot, Motor unit recruitment and EMG power spectra during ramp contractions of a bifunctional muscle, J. Electromyogr. Kinesiol. 221 (2000) 7–2224. [10] R.M. Enoka, A.J. Fuglevand, Motor unit physiology: some unresolved issues, Muscle and Nerve 24 (2001) 4–17. [11] D. Farina, E. Fortunato, R. Merletti, Non invasive estimation of motor unit conduction velocity distribution using linear electrode arrays, IEEE Trans. Biomed. Eng 47 (3) (2000) 380–388. [12] D. Farina, M. Fosci, R. Merletti, Motor unit recruitment strategies investigated by surface EMG variables, J. Appl. Physiol. 92 (2002) 235–247. [13] F. Felici, A. Rosponi, P. Sbriccoli, M. Scarcia, I. Bazzucchi, M. Iannattone, Effect of exposure to altitude on isometric endurance, Eur. J. Appl. Physiol. 85 (2001) 507–512. [14] F. Felici, A. Rosponi, P. Sbriccoli, G.C. Filligoi, L. Fattorini, M. Marchetti, Linear and non-linear analysis of sEMG in weight lifters, Eur. J. Appl. Physiol. 84 (2001) 337–342. [15] F. Felici, L. Colace, P. Sbriccoli, Surface EMG modifications after eccentric exercise, J. Electromyogr. Kinesiol. 7 (1997) 193–202. [16] A.J. Fuglevand, D.A. Winter, A.E. Patla, Models of recruitment and rate coding organization in motor-unit pools, J. Neurophysiol. 70 (6) (1993) 2470–2488. [17] E. Henneman, G. Somjen, D.O. Carpenter, Functional significance of cell size in spinal motoneurons, J. Neurophysiol. 28 (1965) 560–580. [18] C.G. Kukulka, H.P. Clamann, Comparison of the recruitment and discharge properties of motor units in human brachial biceps and adductor pollicis during isometric contraction, Brain Res 219 (1981) 45–55. [19] J.H. Lawrence, C.J. De Luca, Myoelectric signal vs force relationship in different human muscles, J. Appl. Physiol. 54 (1983) 1653–1659. [20] P. Manta, N. Kalfakis, E. Kararizou, D. Vassilopoulos, C. Papagergiou, Size and proportion of fiber types in human muscle fascicles, Clin. Neuropathol 15 (2) (1996) 116–118. [21] R. Merletti, M. Knaflitz, C.J. De Luca, Myoelectric manifestations of fatigue in voluntary and electrically elicited muscle contractions, J. Appl. Physiol. 69 (1990) 1810–1820. [22] H.S. Milner-Brown, R.B. Stein, R. Yemm, Changes in firing rate of human motor units during linearly changing voluntary contractions, J. Physiol. (Lond) 230 (1973) 371–390. [23] T. Moritani, M. Muro, Motor unit activity and surface electromyogram power spectrum during increasing force of contraction, Eur. J. Appl. Physiol. 56 (1987) 260–265. [24] L. Sachs, Applied statistics. A Handbook of Techniques, second ed., Springer Verlag, Berlin, 1984. [25] D.G. Sale, Influence of exercise and training on motor unit activation, Exerc. Sport Sci. Rev 15 (1987) 95–151. [26] K. Seki, Y. Mijazaki, M. Watanabe, A. Nagata, M. Narusawa, Surface electromyogram spectral characterization and motor unit activity during voluntary ramp contractions in men, Eur. J. Appl. Physiol. 63 (1991) 165–172. [27] G. Sollie, H. Hermens, K.L. Boon, W. Wallinga-De-Longe, G. Zivold, The measurement of the conduction velocity of muscle fibers with surface EMG according to the cross correlation method, Electromyogr. Clin. Neurophysiol 25 (1985) 193–204. [28] M. Solomonow, C. Baten, J. Smit, R. Baratta, H. Hermens, R. D’Ambrosia, H. Shoji, Electromyogram power spectra frequencies associated with motor unit recruitment strategies, J. Appl. Physiol. 68 (3) (1990) 1177–1185. [29] M. Solomonow, R. Baratta, H. Shoji, R. D’Ambrosia, The EMGforce relationships of skeletal muscle; dependence on contraction rate, and motor units control strategy, Electromyogr. Clin. Neurophysiol 30 (1990) 141–152.

P. Sbriccoli et al. / Journal of Electromyography and Kinesiology 13 (2003) 139–147 Paola Sbriccoli, Ph.D. in Physiopatology of Movement, was born in 1964. In 1986 she graduated “cum laude” at the ISEF (Superior Institute of Physical Education) of Rome. In 1994 she received her Medical Doctor degree from the University of Rome “La Sapienza” and in 1998 she specialized “cum laude” in Sport Medicine. From 1991 to 2000 she has worked, first as a Medical student and then as a qualified Doctor, at the Institute of Human Physiology of the same University. Her present appointment is at University Institute of Motor Sciences of Rome, Faculty of Motor Sciences as researcher in Methods and Teaching of Sports Activities. She is ordinary member of Italian Physiological Society and of the European College of Sport Science. Her main interests in research are non-invasive assessment of muscle damage and repair, and linear and non-linear analysis of sEMG signals in healthy humans. Ilenia Bazzucchi was born in 1976. She received her Diploma in Physical Education in 2000 from the Institute of Physical Education of Rome Italy and in 2001 she graduated cum laude in Motor Sciences at the University Institute of Motor Sciences of Rome. In 2002 she obtained a student mobility grant and she was accepted as a visiting researcher by the Strathclyde Institute for Biomedical Sciences of Glasgow (UK). She is ordinary member of the European College of Sport Science. Her main research interests are non-invasive assessment of muscle damage and repair, age- related modifications of muscle strength and neuromuscular control, neuromuscular activation during isometric contractions. Alessandro Rosponi was born in 1967. He received his Medical Doctor degree in 1995 from University of Rome “La Sapienza” and he specialized “cum laude” in Sport Medicine in 2000. Since 1993 he has worked first as a Medical student and then as a doctor at the Institute of Human Physiology of the same University. He is a PhD student in Physiopathology of Movement at the University of Rome “La Sapienza”. His research interests are related in the field of Physiology applied to physical exercise, muscle physiology and muscle adaptations to altitude.

147

Marco Bernardi was born in 1961. He received his Medical Doctor degree in 1987 from the University of Rome “La Sapienza” and specialized in Sports Medicine in 1991. Since 1985 he has worked, first as a Medical student and then as a qualified Doctor, at the Institute of Human Physiology of the same University, where he currently works as a researcher and teaches Human Physiology at the Faculty of Medicine. From February 1993 to June 1994 he conducted research in motor unit recruitment strategy (a research topic in which he is still involved) at the Bioengineering Laboratory of the Louisiana State University Medical Center in New Orleans, Louisiana, USA. His other main areas of scientific interest are Exercise Physiology both in healthy and physically handicapped subjects and rehabilitation and locomotor recovery of disabled patients. He is a member of the Italian Society of Physiology, the European College of Sport Science, the American College of Sport Medicine and the International Society for Electromyography and Kinesiology. He is a scientific consultant for the S. Lucia Foundation Rehabilitation Hospital, Rome, Italy. He is President of the Italian Medical Committee of the Italian Sports Federation of Disabled. Giuseppe De Vito was born in 1958. He received his degree in medicine in 1986, specialised in Sports Medicine in 1989 and finally in 1994 completed his PhD in Exercise physiology. All the 3 courses were performed at the University La Sapienza of Rome Italy. From 1994 to 1996 he served as physician/physiologist of the Italian Olympic sailing team. In 1996 he moved to the University of Strathclyde in Glasgow (UK) where he is currently Reader in exercise Physiology at the Strathclyde Institute for Biomedical Sciences. His primary area of teaching is human physiology and exercise physiology with a special attention to ageing. His research interests involve mainly 2 areas: muscle function and ageing and autonomic cardiovascular control in health and disease. He is ordinary member of both British and Italian Physiological societies, the British association of Sport and Exercise science and the European college of Sport Science. He serves as a manuscript reviewer for several scientific journals and he is actually also a grant proposal reviewer for the British association Research into Ageing. Francesco Felici received his Medical Doctor Degree in Medicine in 1982 from Rome University “La Sapienza” and specialized “cum laude” in Sport Medicine in 1985. He was visiting researcher at the Neuromuscular Research Center of Boston University during 1985 and 1990. At present he is Associate Professor of Human and Exercise Physiology with the University Institute of Motor Sciences of Rome, Faculty of Motor Sciences. He is a member of the Italian Physiological Society, of the International Society of Electromyography and Kinesiology, of the European College of Sport Sciences. He is member of the Editorial Board of the Journal of Sports Medicine and Physical Fitness and of the Journal of Electromyography and Kinesiology. His main research interests are Exercise Physiology and biomechanics in healthy subjects with special emphasis on regulation of motor units recruitment and firing rate during voluntary contractions.