- Email: [email protected]
Load-dependent muscle strategy during plantarflexion in humans
Journal of Electromyography and Kinesiology 9 (1999) 1–11
Load-dependent muscle strategy during plantarflexion in humans Alain Carpentier, Jacques Duchateau, Karl Hainaut
*
Laboratory of Biology, Universite´ Libre de Bruxelles, 28 avenue P. He´ger, CP 168, 10500 Brussels, Belgium Accepted 31 March 1998
Abstract This study analyses the relative contribution of the triceps surae and tibialis anterior (TA) muscles to tension development with reference to voluntary plantarflexion at two articular positions of the knee joint (extended and flexed at 90°) for various inertial loads. Subjects were instructed to perform plantarflexions at various sub-maximal and maximal velocities with no intention of stopping the movement. Whereas in one series of experiments the subjects were informed of the load countering the movement, in the other they were not. The average electromyographic (EMG) activity of the different muscles was recorded. The main results were that with loading: (a) greater maximal plantarflexion velocities were recorded in flexed as compared to extended-knee positions; (b) greater durations and amplitudes of agonist and antagonist EMG bursts were recorded; (c) the co-activation of the TA and triceps surae muscles was enhanced; (d) unexpected sub-maximal loads induced greater EMG activity and speed of movement. It is concluded that increasing the load during plantarflexion in humans brings about changes in neuromuscular strategies that contribute to the efficiency of contractile activity during rapid movements. The results also indicate that unexpected sub-maximal loading induces a potentiated neuromuscular activity which increases the speed of movement. 1999 Elsevier Science Ltd. All rights reserved. Keywords: Electromyography; Fast voluntary movement; Agonist; Antagonist co-activation; Mechanical loading
1. Introduction Triphasic electromyographic (EMG) activity has consistently been observed [1–5] in rapid goal-directed movements, with the sequence showing two discrete bursts of agonist EMG separated by an antagonist EMG pulse. In movements with no intention of stopping, only the first agonist EMG activity is present [6]. It is well known that the initial agonist EMG burst is closely related to muscle force during fast movements. However, it is not clear how external loading affects the duration and amplitude of this agonist activity at constant movement amplitude. It has been observed that the length of the agonist pulse increases with loading [1,7– 9]. In contrast, it has been reported that loading alters the size of the pulse, but not its duration [3,10]. More recently, Simmons and Richardson [11] supported and extended this last viewpoint with reference to move-
* Corresponding author. Tel.: ⫹ 32-2-650-21-45; fax: ⫹ 32-2-65039-66; e-mail: [email protected].
ments of constant amplitude perturbed by external loads with different mechanical properties. In a previous work [12], we examined the relative contributions of the triceps surae heads and the tibialis anterior (TA) to tension development with reference to voluntary plantarflexion at various velocities and two articular positions of the knee joint (extended and flexed at 90°). The results indicated that the larger maximal velocity of movement recorded in the flexed as compared to the extended knee was not primarily related to the neural command of the different triceps components, but rather to their mechanical properties. The load to be moved also affects the dynamic aspects of human movements, and it has been demonstrated that not only the intensity but also the timing of muscle activation adapts to loading conditions [13,10,14]. Another work on the relative contribution of the triceps surae heads at constant pedalling speed and against enhanced resistance [15] indicated that whereas the soleus (Sol) integrated EMG augmented linearly with force, the medial gastrocnemius (MG) EMG did not change significantly in the range of small mechanical resistances. Such results were considered to be consistent with the histochemical com-
1050-6411/99/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 1 0 5 0 - 6 4 1 1 ( 9 8 ) 0 0 0 2 2 - 4
2
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
position of Sol, which essentially contains slow muscle fibres in contrast to the more complex composition of the gastrocnemii (Gast) muscles [16]. An interesting question with respect to the effects of loading conditions during voluntary movements is whether the load is known or not. It has been shown that misjudgment of the inertial load induces compensation mechanisms by the nervous system [17,18], but it is not clear whether central and/or peripheral afferents are changed. The purpose of the present study is to examine the contribution of synergist muscles of different histochemical compositions and of an antagonist during plantarflexions performed with different loadings and at various speeds. The movements were performed in two knee positions (extended and flexed at 90°) in known and unexpected loading conditions.
2. Material and method 2.1. Subjects The subjects were physically active male volunteers ranging in age from 18 to 28 years. Eight subjects participated in the study of movement velocity against known loads and, of these, four took part in the maximal velocity experiments as well as in that performed against unknown loads. The subjects were all well accustomed to the experimental procedure and gave their informed consent prior to participating in this investigation. 2.2. Testing procedure and recordings During the experiment each subject lay on his back with the left lower limb flexed at 90° hip, knee and ankle angles. The foot was attached to a movable low friction support (inertia moment: 9.2 ⫻ 10−3 kg m2) and mechanically maintained at 90° dorsal flexion by a stop. The foot was fixed to the support by means of two straps and held in place by a heel block. One strap crossed the dorsum of the foot while the other was attached around the ankle and the calcaneum. The right leg was slightly fixed horizontally. First, the subjects were instructed to perform, at a start signal, random plantarflexions at various sub-maximal and maximal velocities against loads ranging from 0 to 6 kg in increments of 2 kg. For each load, the subjects performed 20–30 plantarflexions with no intention of stopping the movement. Secondly, the subjects performed plantarflexions at maximal velocity without knowing the load. In order to maintain the foot against the stop at 0 kg, a small initial tension (less than 1 kg) opposed the movement and was released simultaneously with the start signal. Each load (the same as in the first experiments) was presented to the subjects five times in random order. No visual feedback or infor-
mation concerning the movement velocity was given. Two articular knee positions (90° and 180°) were tested with the foot at 90° in both cases. The angular ankle motion was registered by means of a linear potentiometer fixed to the axis of rotation of the apparatus, and the angular velocity was calculated from the first derivative of its output after analogue-to-digital conversion. Electromyographic (EMG) activity was recorded from the soleus (Sol), lateral gastrocnemius (LG), medial gastrocnemius (MG) and tibialis anterior (TA) muscles by means of two silver disks (8 mm diameter) positioned 2 cm apart over the motor point of each muscle. The ground electrode was placed over the tibia. The motor point was located by means of electrical stimulation, and if the muscle had multiple motor points the one with the lowest threshold was chosen. The EMG signals were amplified (2000—5000 ⫻ ), filtered (10 Hz–5 kHz), fullwave rectified and displayed on a digital oscilloscope (Nicolet 4094C). Both the EMG and displacement signals were recorded on magnetic tape by an FM-8 channel recorder (Hewlett Packard, 3968a; bandwidth DC2500 Hz). The M (Mmax) wave was recorded under the same conditions as described above. The triceps surae was stimulated by rectangular pulses of 1 ms duration delivered to the tibial nerve via a custom-made stimulator triggered by a digital timer (Digitimer Ltd, 4030). A steel needle was used as the negative electrode (cathode) and a silver plate (2 ⫻ 3 cm) was chosen as the positive electrode (anode). The cathode was inserted subcutaneously over the nerve at the popliteal fossa, and the anode was positioned on the patella. The electrical stimulus used was progressively increased until the M wave amplitude reached a plateau. 2.3. Measurements and data analysis Analysis was performed off-line. The tapes were played back and both the EMG and displacement signals were submitted to digital conversion by a personal computer (sampling rate: 1000 Hz). The data were numerically smoothed by a 3-point moving averaging that produced a 100 Hz low-pass filtering (at ⫺ 3 dB). The angular velocity was computed by the software differentiation of the goniometer output and the peak velocity was determined for each trial. Only trials of comparable amplitude obtaining velocities of at least 150° s−1 were analysed due to the difficulty in determining the EMG onset accurately at a low velocity. The EMG activity and its duration were estimated from the onset of the main burst to its end (above 5% of the largest trace amplitude). The latency between movement and EMG activity was determined for each muscle as the difference between the onset of the velocity signal (the first derivative of displacement) and the EMG. The EMG activity was assessed by computing the area under the rectified and
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
smoothed EMG traces. Each EMG was divided by the burst duration in order to estimate an average EMG activity. Cross-talk was assessed in five of the subjects by stimulating each muscle individually at the motor point at near-maximal intensity and recording the EMG activity in the other synergist or antagonist muscles. The extent of the cross-talk was estimated from the relative amplitude of the corresponding M waves and the amplitude recorded during the maximal stimulation of the tibial (triceps surae) or peroneal nerve (TA). The muscles and nerves were stimulated by delivering rectangular pulses of 1 ms duration by means of a custom-made stimulator triggered by a digital timer (Digitimer Ltd, 4030). Under these conditions, cross-talk was found to be less than 8% of the maximal M wave. Since the EMG peak-to-peak amplitude in voluntary activation never exceeded 1.5 mV, which is roughly 10 ⫻ smaller than the M wave response, the extent of cross-talk was considered to be negligible both in our experiments (between the synergist or antagonist muscles) and in other studies [19,20].
3
2.4. Statistics Conventional statistical methods were used to calculate the means, standard deviations (SD) and coefficients of correlation. The best-fitting relationships were tested by linear and non-linear regressions on the basis of the leastsquares method. The analysis of covariance (ANCOVA) was used to calculate the regression coefficients of Y and X for each load, knee flexed vs. extended, and to test the slopes of the various regression lines. A repeatedmeasures analysis of variance (ANOVA) was used to compare the variations among the means, and the Tukey–Kramer multiple comparisons post-test was used when the value was less than 0.05. A probability of p ⬍ 0.05 was considered as significant for all conditions. 3. Results 3.1. Effects of load on movement and EMG patterns In the flexed-knee position the plantarflexion movements were larger (p ⬍ 0.01) in comparison with the
Fig. 1. Typical example of the EMG (rectified) activity patterns of the medial (MG) and lateral gastrocnemius (LG), soleus (Sol) and tibialis anterior (TA) together with the displacement and velocity traces recorded in one subject during plantarflexions at maximal velocities with no load (a–c) and with a 6 kg load (b–d). These recordings were obtained with the knee flexed at 90° (a, b) and in an extended position (c, d). The arrows indicate the onset and end of EMG measurements.
4
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
Fig. 1.
extended position, and ranged from 34.2 ⫾ 7.4° to 39.1 ⫾ 7.6° and 28.4 ⫾ 8.9° to 30.6 ⫾ 5.7°, respectively. With the increase of the load from 0 to 6 kg the maximal movement velocity recorded in the flexed knee dropped from 765 ⫾ 128° s−1 to 515 ⫾ 84° s−1 (means ⫾ SD, Fig. 1a, b). In extended-knee positions, the velocities decreased from 551 ⫾ 124° s−1 to 423 ⫾ 74° s−1 as the load increased (Fig. 1c, d). As the load grew, the EMG burst duration increased at all velocities of movement, and in our experimental conditions attained at maximal velocity 124.3 ⫾ 24.5 ms and 207.2 ⫾ 41.3 ms respectively at 0 kg and 6 kg in the flexed knee. In the extended knee, the values were 151.2 ⫾ 40.1 ms and 222.4 ⫾ 59.6 ms respectively. 3.2. Effects of loading on EMG activity With increasing velocity of movement, enhanced (p ⬍ 0.01) absolute EMG activity was recorded in all triceps surae muscles whatever the load, except for 6 kg in the flexed-knee position (Fig. 2). When the load was progressively enhanced from 0 kg to 6 kg at maximal velocity of movement, the normalized EMG activity of
Continued.
the triceps surae increased (p ⬍ 0.05; Fig. 3) in the flexed-knee position. However, when the leg was extended, the normalized EMG activity showed no significant change with loading (Fig. 3). Fig. 4 illustrates the relative contribution of the Gast and Sol EMG activities when the loading was enhanced during slow and fast movements performed in the two knee positions. It appears that in all loading conditions at maximal velocity the Sol/MG and Sol/LG EMG ratios were not significantly different when the movements were performed with the knee flexed or extended. For smaller angular velocities as in all loading conditions the Sol contribution appeared to be larger in the flexed-knee position as compared to the extended leg. This figure also shows that in both knee positions the contribution of Gast EMG augmented progressively with increasing loads: this increase was not significant (p > 0.05), however. Since the Gast length in the extended knee was larger when compared to the flexed position, a possible effect on EMG activity was tested by recording the Mmax in both positions. No significant (p > 0.05) difference was observed in the two knee positions but Mmax amplitude
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
5
movement velocity. The burst increase at maximal velocity, from 0 kg to 6 kg loading, was 76.3 ⫾ 27.0% in the flexed-knee position, and in the extended-knee position the increase was 53.9 ⫾ 37.9%. 3.4. Unknown loading condition
Fig. 2. Relationship between angular velocity and EMG average in the MG (a–c) and Sol (b–d) both for no load and for a 6 kg load. These graphs compare the data recorded in the flexed and extendedknee position for all of our subjects. The linear equation in the flexed and extended knee positions with no load are respectively: y ⫽ 0.14x ⫺ 12.4 (r ⫽ 0.87; p ⬍ 0.001) and y ⫽ 0.17x ⫹ 13.8 (r ⫽ 0.59; p ⬍ 0.001) for the MG and y ⫽ 0.03x ⫹ 9.6 (r ⫽ 0.31, p ⬍ 0.01) and y ⫽ 0.10x ⫺ 9.7 (r ⫽ 0.61; p ⬍ 0.001) for the Sol. For 6 kg load, the values are respectively: y ⫽ 0.22x ⫹ 9.8 (r ⫽ 0.51; p ⬍ 0.001) and y ⫽ 0.20x ⫹ 32.8 (r ⫽ 0.38; p ⬍ 0.001) for the MG and y ⫽ 0.002x ⫹ 32.0 (r ⫽ 0.01; p ⫽ 0.90) and y ⫽ 0.07x ⫹ 13.4 (r ⫽ 0.21; p ⬍ 0.01).
were slightly greater in the flexed knee by 10.3 ⫾ 6.8%, 8.1 ⫾ 9.9% and 0.7 ⫾ 2.3% in LG, MG and Sol, respectively. 3.3. Effect of load on antagonist EMG activity As loading increased, the TA showed an EMG activity which remained nearly concurrent with that of the triceps surae muscles (Fig. 1), and increased (p ⬍ 0.01) from 0 kg to 6 kg countering plantarflexion. It should be pointed out, however, that this activity modulation of TA with increasing loads remained in a rather limited range of its maximal activation capacity, and only increased from 19.5 ⫾ 4.1% to 24.7 ⫾ 5.7% of the EMG recorded during maximal contraction. With increased loading, the TA EMG activity changes were larger in the flexed-knee positions as compared to the extended leg, and at maximal velocity this increase was 121.7 ⫾ 39.9% and 20.7 ⫾ 12.8%, respectively (Fig. 3). Under all loading conditions, TA behaved similarly to the agonist muscles (increased burst) during increasing
No significant difference in movement amplitude was observed when the sujects were either aware or unaware of the load opposing plantarflexion; however, in the two knee positions greater maximal velocities were recorded in unknown loading conditions for all loads except 6 kg (Figs. 5 and 6). At 0 kg the mean maximal velocity was increased (p ⬍ 0.05) from 767 ⫾ 62° s−1 to 1071 ⫾ 238° s−1 and from 558 ⫾ 76° s−1 to 654 ⫾ 108° s−1 in the flexed and extended-knee positions respectively. These results are illustrated in Fig. 5 in one subject who performed plantarflexions at maximal velocity. Whether the load was known or not did not affect the latency between EMG and movement onsets, but at 0 kg all agonist EMG burst durations were greater (p ⬍ 0.01) (142.7 ⫾ 26.4 ms vs. 125.1 ⫾ 16.8 ms and 176.7 ⫾ 30.2 ms vs. 150.4 ⫾ 17.2 ms in the flexed and extendedknee positions respectively) at unknown loading. On the other hand, at 6 kg these burst durations were shorter (p ⬍ 0.01) in unknown loading condition in the flexed-knee position (171.3 ⫾ 39.4 ms vs. 218.6 ⫾ 27.6 ms) but not significantly different in the extended-leg position (211.4 ⫾ 38.9 ms vs. 195.4 ⫾ 35.4 ms). Enhanced EMG activity was observed at unknown 0 kg in all the muscles, without any change in the Sol/Gast ratio. Under unknown loading conditions, a more detailed EMG analysis indicated that whatever the load or knee position, the EMG activities recorded from their onset through the first 25 ms of the movement were not significantly different from those observed in the known loading condition. However, at unknown 0 kg the EMG was increased (p ⬍ 0.05) in the flexed-knee position (62.0 ⫾ 17.5% and 391.8 ⫾ 97.5% in a time interval of 25–100 ms and after 100 ms, respectively) and also in the extended-leg position (57.5 ⫾ 39.7% after 100 ms). At a 6 kg load the only recorded EMG change in the unknown condition was a decrease (34.4 ⫾ 6.5%; p ⬍ 0.01) in flexed-knee position during the interval of 25– 100 ms. Since all triceps surae entities showed a similar type of behaviour, the data reported above are the means for the three heads.
4. Discussion The central nervous system strategy in generating fast voluntary movements in skeletal muscles consisting of different fibre types remains an exciting question in motor control. The triceps surae is an interesting model for such an experimental approach because it is com-
6
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
Fig. 3. Relationship between the angular velocity and the EMG in the different triceps surae heads and TA for the four different loads when the knee was in the flexed (left panels) or extended position (right panels). The angular velocity has been normalized with respect to the maximal value recorded during the experimental session, and each EMG is expressed as a percentage of the value measured in that condition.
posed of the Sol, which contains mainly type I fibres, and the Gast, which has nearly the same proportion of type I and type II fibres [16]. Moreover, the Sol is a monoarticular muscle, whereas the Gast is a biarticular one, and their strategy in voluntary movements may differ [12,21]. The main results from the present experiments are that with loading: (a) greater maximal plantarflexion velocities were recorded in the flexed as compared to the extended-knee position; (b) the longer duration and amplitude of agonist EMG bursts were recorded; (c) at maximal velocity, there was a decreased Sol/Gast ratio; (d) the co-activation of the TA and triceps surae muscles was enhanced; (e) unexpected sub-maximal loads induced a potentiation of neuromuscular activity. Because cross-talk between muscles was negligible in our experiments, we consider that the EMG behaviour observed was both load- and position-dependent. Compared to the extended position, the movements performed in the flexed-knee position were slightly larger (Section 3) and the EMG responses could therefore be greater [22,23,3], but it is generally accepted that EMG activities are comparable [6] for such small movement amplitude differences. The observation that, at constant load, the maximal velocity of movement was larger at a 90° knee flexion as compared to the extended-limb position was surprising, but this could be explained if the Gast activation was greater in the flexed-knee position. In fact, the results showed a decrease in the absolute average EMG
in all the triceps surae muscles in the flexed position as compared to the extended-knee position, an observation which is consistent with the work of Cresswell et al. [24]. As previously observed [12], the optimum position for the Gast with respect to the speed of the movement is reached when the knee is flexed to 90°. This finding suggested that the larger maximal plantarflexion velocity observed in the flexed-knee position should not be primarily related to a different neural command of the triceps surae entities, but more likely to mechanical muscle properties. This viewpoint is further supported by the present results, which indicate that with an increasing load the largest velocity of movement is consistently recorded at a 90° knee flexion. It could be argued that in the extended-knee position TA activation might be present to maintain the ankle position at 90°; this is not the case in the flexed-knee position since there is a resistive torque opposed to the plantarflexion torque. This viewpoint might be valid at 0 kg, but in fact similar results were observed with an increase in loading which kept the foot passive at 90°. It is interesting that the decrease in maximal movement velocity observed with an increasing load is associated with a tendency of the Sol/Gast EMG to decrease in the flexed knee. Such results agree with the finding that at 6 kg the Sol EMG does not change with increasing velocity (Fig. 2d) and suggest that motor unit activation is saturated in this muscle but not in the Gast, since EMG increases with speed (Fig. 2c). Such different behaviour of muscles
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
7
Fig. 4. Comparison of the mean ( ⫾ SE, n ⫽ 8) Sol/MG (a, b) and Sol/LG (c, d) EMG ratio of slow (200°/s) and maximal angular velocities in the flexed (a–c) and extended (b–d) positions. *p ⬍ 0.05; **p ⬍ 0.01.
with different fibre type composition is apparently consistent with the observation that, in Sol motor units, the activation threshold decrease during ballistic contractions is greater when compared to muscles which contain a smaller amount of slow motor units [25]. A triphasic EMG burst has been reported [2–5] in ballistic goal-directed movements. The first phase is the agonist initiation burst; the second is the antagonist EMG activity that stops the movement; and the third is an agonist response to the antagonist burst. Other studies do not report this neuromuscular strategy but refer to the co-activation of agonists and antagonists, e.g. in wrist flexion with no intention of stopping the movement [6] and elbow flexion or extension performed against gravity [26]. Our results also obviously show the co-activation
of TA and triceps surae in plantarflexion during movement with no intention of stopping, and furthermore indicate that this co-activation is reinforced by external loading. It has been suggested [27,28] that the conditions of our experiment may lead to the co-activation of TA and triceps surae, to stabilizing the complex ankle joint and thus contributing to the efficiency of the triceps surae with respect to increasing the load countering plantarflexion. Even if many interesting studies [8,29,30,17,13,10, 31,14] are taken into consideration, the effect of external loading on the first agonist burst time is not clear, possibly because of protocol differences. For example, the relationship between loading, burst duration and burst amplitude is rather confusing [32,29,13]. In our experi-
8
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
Fig. 5. Example of the EMG (rectified) activity patterns of the medial (MG) and lateral gastrocnemius (LG), soleus (Sol) and tibialis anterior (TA) together with the displacement and velocity traces recorded in one subject during plantarflexions at maximal velocities at zero load both with (a) and without knowledge (b) of the load. The arrows indicate the onset and end of EMG measurements.
mental conditions, increasing the load induces enhanced duration and amplitude of the first EMG pulse which is similar in all triceps surae muscles. These results indicate that during plantarflexion, loading: (1) is controlled by changes in muscle activation duration and intensity; (2) induces similar changes of contraction control strategy in muscle components with different fibre type compositions. It appears from Fig. 3 that in TA and all triceps surae components, the EMG changes in function of increasing load (and speed) are proportionally larger in the flexed as compared to extended-knee position and that this difference is obviously greater in Gast and TA as compared to Sol. In other words, the main loadingrelated muscle activation changes are those of Gast and TA in the flexed knee position. This observation also supports the viewpoint that the biarticular Gast is in a more efficient mechanical length with respect to velocity in the flexed as compared to the extended-knee position since loading is more closely related to a change in its neural activation when the leg is flexed. In the extended
position, the Gast activation changes in function of loading are more or less comparable to those of the monoarticular Sol which is, of course, very little affected by a difference in the articular knee position. It is interesting that in both knee positions the antagonist TA behaves nearly like the Gast and also shows a greater loadinginduced activation change in the flexed as compared to the extended knee. This observation suggests that a change in agonist activation also affects the antagonist [12]. Perhaps the most intriguing section of the results is the finding that in both knee angulations the maximal velocity of movement and the EMG activities were greater when the subjects were not informed of the load opposing plantarflexion. The observation that EMG activity onset and pattern over the first 25 ms do not change in agonist and antagonist muscles in unknown loading conditions suggests that central programming is not affected. Thus, increases in reflex and, later, voluntary activities should explain the larger EMG responses
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
9
Fig. 6. Histograms showing the mean ( ⫾ SE, n ⫽ 4) angular velocity, the EMG of the medial and lateral gastrocnemius, soleus and tibialis anterior for two different loads (0 and 6 kg) in the two knee positions (knee flexed, KF; and extended, KE) both with and without knowledge of the load. *p ⬍ 0.05; **p ⬍ 0.01, ***p ⬍ 0.001.
and speed of movement recorded in unexpected loading conditions. This viewpoint is consistent with the fact that the latter part (after 100 ms) of the EMG activity is significantly greater in unexpected loading. The interesting
difference between EMG activities during maximal movement velocity under unexpected 0 kg and 6 kg loading conditions is that the increase in EMG activity in the 25–100 ms interval is larger in the flexed as com-
10
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
pared to the extended knee and decreases with loading, so slowing the movement down. These results support the suggestion that with an unexpectedly reduced inertial load, spindle afferents are induced [17] which, in our experiments, should be greater in ballistic movements at 0 kg as compared to the slower tension development related to increased loading. This speculation is in line with the observation that in the TA antagonist the EMG recorded between 25 and 100 ms does not decrease with loading. In fact, in the antagonist muscle, spindle afferents should also be induced by ␣–␥ co-activation and stretching during plantarflexion. Another interesting difference of the antagonist EMG activity during unexpected loading (Fig. 5) is the presence of a large burst later than 100 ms. This burst precedes the one related to foot dorsiflexion. It is difficult to deduce from our experimental data whether this burst is related to the increased duration of the antagonist co-activation during plantarflexion and/or to the decreased latency of its activity during dorsiflexion. However, these experiments are in line with the “equilibrium-point model” of EMG patterns during movements, and the proposition that peripheral afferents contribute to modulate the co-contraction and adaptative stiffness of the ankle joint with respect to the control of movement [33]. It is concluded that increasing the load during plantarflexion in humans induces changes in neuromuscular strategies that contribute to the efficiency of contractile activity during rapid movements. The results also indicate that unexpected sub-maximal loading induces a potentiation of both the central and peripheral neural drives to the agonist and antagonist muscles which increase the speed of movement. Acknowledgements This work was supported by NATO CRG no. 930261, the Fonds National de la Recherche Scientifique of Belgium and the Conseil de la Recherche of the Universite´ Libre de Bruxelles. The authors are grateful to Miss Anne Deisser for her assistance in the preparation of the manuscript. References [1] Angel RW. Electromyography during voluntary movements: the two-burst pattern. Electroencephalogr Clin Neurophysiol 1974; 36:493–8. [2] Gottlieb GL, Latash ML, Corcos DM, Livbinskas TJ, Agarwal GC. Organizing principles for single joint movements. V. Agonist–antagonist interactions. J Neurophysiol 1992;67:1417–27. [3] Hallett M, Marsden CD. Ballistic flexion movements of the human thumb. J Physiol (Lond) 1979;294:33–50. [4] Marsden CD, Obeso JA, Rothwell JC. The function of the antagonist muscle during fast limb movements in man. J Physiol 1983;335:1–13.
[5] Meinck H, Benecke R, Meyer W, Hohne J, Conrad B. Human ballistic finger flexion: uncoupling of the three-burst pattern. Exp Brain Res 1984;55:127–33. [6] Mustard BE, Lee RG. Relationship between EMG patterns and kinematic properties for flexion movements at the human wrist. Exp Brain Res 1987;66:247–56. [7] Benecke R, Meinck HM, Conrad B. Rapid goal-directed elbow flexion movements: limitations of the speed control system due to neural constraints. Exp Brain Res 1985;59:470–7. [8] Berardelli A, Rothwell JC, Day BL, Kachi T, Marsden CD. Duration of the first agonist EMG burst in ballistic arm movements. Exp Brain Res 1984;304:183–7. [9] Smeets JBJ, Erkelens CJ, Denier Van Der Gon JJ. Adjustments of fast goal-directed movements in response to an unexpected inertial load. Exp Brain Res 1990;81:303–12. [10] Lestienne F. Effects of inertial load and velocity on the braking process of voluntary limb movements. Exp Brain Res 1979; 35:407–18. [11] Simmons RW, Richardson C. Effects of different types of mechanical load on the duration of the initial agonist pulse. Exp Brain Res 1993;92:524–7. [12] Carpentier A, Duchateau J, Hainaut K. Velocity-dependent muscle strategy during plantarflexion in humans. J Electromyogr Kinesiol 1996;6(4):225–33. [13] Le Bozec S, Maton B, Cnockaert JC. The synergy of elbow extensor muscles during dynamic work in man. I. Elbow extension. Eur J Appl Physiol 1980;44:225–69. [14] Wadman WJ, Denier Van Der Gon JJ, Geuze RH, Mol CR. Control of fast goal-directed arm movements. J Hum Mov Stud 1979;5:3–17. [15] Duchateau J, Le Bozec S, Hainaut K. Contributions of slow and fast muscles of triceps surae to a cyclic movement. Eur J Appl Physiol 1986;55:476–81. [16] Johnson MA, Polgar J, Weightman D, Appleton D. Data on the distribution of fibre types in thirty-six human muscles: an autopsy study. J Neurol Sci 1973;18:111–29. [17] Latash ML. Control of fast elbow movement: a study of electromyographic patterns during movements against unexpectedly decreased inertial load. Exp Brain Res 1994;98:145–52. [18] Schmidt RA, McGown CM. Terminal accuracy of unexpectedly loaded rapid movements: evidence for a mass-spring mechanism in programming. J Motor Behav 1980;12:149–61. [19] Moritani T, Oddson L, Thorstensson A. Electromyographic evidence of selective fatigue during the eccentric phase of stretch/shortening cycles in man. Eur J Appl Physiol 1990; 60:425–9. [20] Solomonow M, Baratta R, Bernardi M, Zhou B, Lu Y, Zhu M, Acierno S. Surface and wire EMG crosstalk in neighbouring muscles. J Electromyogr Kinesiol 1994;4:131–42. [21] Jacobs R, Bobbert MF, Van Ingen Schenau GJ. Function of mono-and biarticular muscles in running. Med Sci Sports Exerc 1993;25(10):1163–73. [22] Brown SH, Cooke JD. Initial agonist burst duration depends on movement amplitude. Exp Brain Res 1984;55:523–7. [23] Cheron G, Godaux E. Self-terminated fast movement of the forearm in man: amplitude dependence of the triple burst pattern. J Biophys Biomech 1986;10:109–17. [24] Cresswell AG, Lo¨scher WN, Thorstensson A. Influence of gastrocnemius muscle length on triceps surae torque development and electromyographic activity in man. Exp Brain Res 1995; 105:283–90. [25] Desmedt JE, Godaux E. Ballistic contractions in fast or slow human muscles: discharge patterns of single motor units. J Physiol 1978;285:185–96. [26] Virji-Babul N, Cooke JD, Brown SH. Effects of gravitational forces on single joint arm movements in humans. Exp Brain Res 1994;99:338–46.
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
[27] Baratta R, Solomonow M, Zhou B, Letson D, Chuinard R, D’Ambrosia R. Muscular activation: the role of the antagonist musculature in maintaining knee stability. Am J Sports Med 1988;16:113–22. [28] Solomonow M, Baratta R, Zhou BH, D’Ambrosia R. Electromyogram coactivation patterns of the elbow antagonist muscles during slow isokinetic movement. Exp Neurol 1988;100:470–7. [29] Gottlieb GL, Chen CH, Corcos DM. Relations between joint torque, motion and electromyographic patterns at the human elbow. Exp Brain Res 1995;103:164–7. [30] Hoffman DS, Strick PL. Step-tracking movements of the wrist. III. Influence of changes in load on patterns of muscle activity. J Neurosci 1993;13(12):5212–27. [31] Sherwood DE, Schmidt RA, Walter CB. Rapid movements with reversals in direction. II. Control of movement amplitude and inertial load. Exp Brain Res 1988;69:355–67. [32] Cooke JD, Brown SH. Movement-related phasic muscle activation. III. The duration of phasic agonist activity initiating movement. Exp Brain Res 1994;99:473–82. [33] Latash ML, Goodman SR. An equilibrium-point model of electromyographic patterns during single-joint movements based on experimentally reconstructed control signals. J Electromyogr Kinesiol 1994;4:230–41. Alain Carpentier received the master’s degree in physical education and sports sciences from the Universite´ Libre de Bruxelles, Belgium, in 1988. In 1990 he joined the Laboratory of Biology and Research Unit in Neurophysiology (Professor K. Hainaut) as a Ph.D. student, where he presently works on the neuromuscular strategies in humans during voluntary contractions and will present his Ph.D. thesis in 1998. His other main research interest is related to the study of EMG changes during fatigue, including the analysis of single motor unit behaviour. He received the scientific award of the “Socie´te´ de Biome´canique” (1997) for that work.
11
Jacques Duchateau received the master’s degree in physical education and sports sciences (1977) and a Ph.D. degree (1981) from the Universite´ Libre de Bruxelles, Belgium. Since 1979, he has been affiliated with the Laboratory of Biology and Research Unit in Neurophysiology at the Universite´ Libre de Bruxelles. He was assistant professor from 1978 to 1984 and became full professor in 1985. His major interests focus on neuromuscular physiology and muscle biomechanics. Dr Duchateau serves on the Editorial Board of the Journal of Electromyography and Kinesiology. Karl Hainaut received Ph.D. degrees in physical education (1968) and in physiological sciences (1975) from the Universite´ Libre de Bruxelles, Belgium. Full-time professor at the Universite´ Libre de Bruxelles and director of the Laboratory of Biology and Research Unit in Neurophysiology, his research in neuromuscular physiology has consistently been supported by the National Fund for Scientific Research, the Academies of Sciences (Acade´mie The´re´sienne) and of Medicine, and more recently by NATO for a collaborative project with North American laboratories.
Load-dependent muscle strategy during plantarflexion in humans Alain Carpentier, Jacques Duchateau, Karl Hainaut
*
Laboratory of Biology, Universite´ Libre de Bruxelles, 28 avenue P. He´ger, CP 168, 10500 Brussels, Belgium Accepted 31 March 1998
Abstract This study analyses the relative contribution of the triceps surae and tibialis anterior (TA) muscles to tension development with reference to voluntary plantarflexion at two articular positions of the knee joint (extended and flexed at 90°) for various inertial loads. Subjects were instructed to perform plantarflexions at various sub-maximal and maximal velocities with no intention of stopping the movement. Whereas in one series of experiments the subjects were informed of the load countering the movement, in the other they were not. The average electromyographic (EMG) activity of the different muscles was recorded. The main results were that with loading: (a) greater maximal plantarflexion velocities were recorded in flexed as compared to extended-knee positions; (b) greater durations and amplitudes of agonist and antagonist EMG bursts were recorded; (c) the co-activation of the TA and triceps surae muscles was enhanced; (d) unexpected sub-maximal loads induced greater EMG activity and speed of movement. It is concluded that increasing the load during plantarflexion in humans brings about changes in neuromuscular strategies that contribute to the efficiency of contractile activity during rapid movements. The results also indicate that unexpected sub-maximal loading induces a potentiated neuromuscular activity which increases the speed of movement. 1999 Elsevier Science Ltd. All rights reserved. Keywords: Electromyography; Fast voluntary movement; Agonist; Antagonist co-activation; Mechanical loading
1. Introduction Triphasic electromyographic (EMG) activity has consistently been observed [1–5] in rapid goal-directed movements, with the sequence showing two discrete bursts of agonist EMG separated by an antagonist EMG pulse. In movements with no intention of stopping, only the first agonist EMG activity is present [6]. It is well known that the initial agonist EMG burst is closely related to muscle force during fast movements. However, it is not clear how external loading affects the duration and amplitude of this agonist activity at constant movement amplitude. It has been observed that the length of the agonist pulse increases with loading [1,7– 9]. In contrast, it has been reported that loading alters the size of the pulse, but not its duration [3,10]. More recently, Simmons and Richardson [11] supported and extended this last viewpoint with reference to move-
* Corresponding author. Tel.: ⫹ 32-2-650-21-45; fax: ⫹ 32-2-65039-66; e-mail: [email protected].
ments of constant amplitude perturbed by external loads with different mechanical properties. In a previous work [12], we examined the relative contributions of the triceps surae heads and the tibialis anterior (TA) to tension development with reference to voluntary plantarflexion at various velocities and two articular positions of the knee joint (extended and flexed at 90°). The results indicated that the larger maximal velocity of movement recorded in the flexed as compared to the extended knee was not primarily related to the neural command of the different triceps components, but rather to their mechanical properties. The load to be moved also affects the dynamic aspects of human movements, and it has been demonstrated that not only the intensity but also the timing of muscle activation adapts to loading conditions [13,10,14]. Another work on the relative contribution of the triceps surae heads at constant pedalling speed and against enhanced resistance [15] indicated that whereas the soleus (Sol) integrated EMG augmented linearly with force, the medial gastrocnemius (MG) EMG did not change significantly in the range of small mechanical resistances. Such results were considered to be consistent with the histochemical com-
1050-6411/99/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 1 0 5 0 - 6 4 1 1 ( 9 8 ) 0 0 0 2 2 - 4
2
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
position of Sol, which essentially contains slow muscle fibres in contrast to the more complex composition of the gastrocnemii (Gast) muscles [16]. An interesting question with respect to the effects of loading conditions during voluntary movements is whether the load is known or not. It has been shown that misjudgment of the inertial load induces compensation mechanisms by the nervous system [17,18], but it is not clear whether central and/or peripheral afferents are changed. The purpose of the present study is to examine the contribution of synergist muscles of different histochemical compositions and of an antagonist during plantarflexions performed with different loadings and at various speeds. The movements were performed in two knee positions (extended and flexed at 90°) in known and unexpected loading conditions.
2. Material and method 2.1. Subjects The subjects were physically active male volunteers ranging in age from 18 to 28 years. Eight subjects participated in the study of movement velocity against known loads and, of these, four took part in the maximal velocity experiments as well as in that performed against unknown loads. The subjects were all well accustomed to the experimental procedure and gave their informed consent prior to participating in this investigation. 2.2. Testing procedure and recordings During the experiment each subject lay on his back with the left lower limb flexed at 90° hip, knee and ankle angles. The foot was attached to a movable low friction support (inertia moment: 9.2 ⫻ 10−3 kg m2) and mechanically maintained at 90° dorsal flexion by a stop. The foot was fixed to the support by means of two straps and held in place by a heel block. One strap crossed the dorsum of the foot while the other was attached around the ankle and the calcaneum. The right leg was slightly fixed horizontally. First, the subjects were instructed to perform, at a start signal, random plantarflexions at various sub-maximal and maximal velocities against loads ranging from 0 to 6 kg in increments of 2 kg. For each load, the subjects performed 20–30 plantarflexions with no intention of stopping the movement. Secondly, the subjects performed plantarflexions at maximal velocity without knowing the load. In order to maintain the foot against the stop at 0 kg, a small initial tension (less than 1 kg) opposed the movement and was released simultaneously with the start signal. Each load (the same as in the first experiments) was presented to the subjects five times in random order. No visual feedback or infor-
mation concerning the movement velocity was given. Two articular knee positions (90° and 180°) were tested with the foot at 90° in both cases. The angular ankle motion was registered by means of a linear potentiometer fixed to the axis of rotation of the apparatus, and the angular velocity was calculated from the first derivative of its output after analogue-to-digital conversion. Electromyographic (EMG) activity was recorded from the soleus (Sol), lateral gastrocnemius (LG), medial gastrocnemius (MG) and tibialis anterior (TA) muscles by means of two silver disks (8 mm diameter) positioned 2 cm apart over the motor point of each muscle. The ground electrode was placed over the tibia. The motor point was located by means of electrical stimulation, and if the muscle had multiple motor points the one with the lowest threshold was chosen. The EMG signals were amplified (2000—5000 ⫻ ), filtered (10 Hz–5 kHz), fullwave rectified and displayed on a digital oscilloscope (Nicolet 4094C). Both the EMG and displacement signals were recorded on magnetic tape by an FM-8 channel recorder (Hewlett Packard, 3968a; bandwidth DC2500 Hz). The M (Mmax) wave was recorded under the same conditions as described above. The triceps surae was stimulated by rectangular pulses of 1 ms duration delivered to the tibial nerve via a custom-made stimulator triggered by a digital timer (Digitimer Ltd, 4030). A steel needle was used as the negative electrode (cathode) and a silver plate (2 ⫻ 3 cm) was chosen as the positive electrode (anode). The cathode was inserted subcutaneously over the nerve at the popliteal fossa, and the anode was positioned on the patella. The electrical stimulus used was progressively increased until the M wave amplitude reached a plateau. 2.3. Measurements and data analysis Analysis was performed off-line. The tapes were played back and both the EMG and displacement signals were submitted to digital conversion by a personal computer (sampling rate: 1000 Hz). The data were numerically smoothed by a 3-point moving averaging that produced a 100 Hz low-pass filtering (at ⫺ 3 dB). The angular velocity was computed by the software differentiation of the goniometer output and the peak velocity was determined for each trial. Only trials of comparable amplitude obtaining velocities of at least 150° s−1 were analysed due to the difficulty in determining the EMG onset accurately at a low velocity. The EMG activity and its duration were estimated from the onset of the main burst to its end (above 5% of the largest trace amplitude). The latency between movement and EMG activity was determined for each muscle as the difference between the onset of the velocity signal (the first derivative of displacement) and the EMG. The EMG activity was assessed by computing the area under the rectified and
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
smoothed EMG traces. Each EMG was divided by the burst duration in order to estimate an average EMG activity. Cross-talk was assessed in five of the subjects by stimulating each muscle individually at the motor point at near-maximal intensity and recording the EMG activity in the other synergist or antagonist muscles. The extent of the cross-talk was estimated from the relative amplitude of the corresponding M waves and the amplitude recorded during the maximal stimulation of the tibial (triceps surae) or peroneal nerve (TA). The muscles and nerves were stimulated by delivering rectangular pulses of 1 ms duration by means of a custom-made stimulator triggered by a digital timer (Digitimer Ltd, 4030). Under these conditions, cross-talk was found to be less than 8% of the maximal M wave. Since the EMG peak-to-peak amplitude in voluntary activation never exceeded 1.5 mV, which is roughly 10 ⫻ smaller than the M wave response, the extent of cross-talk was considered to be negligible both in our experiments (between the synergist or antagonist muscles) and in other studies [19,20].
3
2.4. Statistics Conventional statistical methods were used to calculate the means, standard deviations (SD) and coefficients of correlation. The best-fitting relationships were tested by linear and non-linear regressions on the basis of the leastsquares method. The analysis of covariance (ANCOVA) was used to calculate the regression coefficients of Y and X for each load, knee flexed vs. extended, and to test the slopes of the various regression lines. A repeatedmeasures analysis of variance (ANOVA) was used to compare the variations among the means, and the Tukey–Kramer multiple comparisons post-test was used when the value was less than 0.05. A probability of p ⬍ 0.05 was considered as significant for all conditions. 3. Results 3.1. Effects of load on movement and EMG patterns In the flexed-knee position the plantarflexion movements were larger (p ⬍ 0.01) in comparison with the
Fig. 1. Typical example of the EMG (rectified) activity patterns of the medial (MG) and lateral gastrocnemius (LG), soleus (Sol) and tibialis anterior (TA) together with the displacement and velocity traces recorded in one subject during plantarflexions at maximal velocities with no load (a–c) and with a 6 kg load (b–d). These recordings were obtained with the knee flexed at 90° (a, b) and in an extended position (c, d). The arrows indicate the onset and end of EMG measurements.
4
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
Fig. 1.
extended position, and ranged from 34.2 ⫾ 7.4° to 39.1 ⫾ 7.6° and 28.4 ⫾ 8.9° to 30.6 ⫾ 5.7°, respectively. With the increase of the load from 0 to 6 kg the maximal movement velocity recorded in the flexed knee dropped from 765 ⫾ 128° s−1 to 515 ⫾ 84° s−1 (means ⫾ SD, Fig. 1a, b). In extended-knee positions, the velocities decreased from 551 ⫾ 124° s−1 to 423 ⫾ 74° s−1 as the load increased (Fig. 1c, d). As the load grew, the EMG burst duration increased at all velocities of movement, and in our experimental conditions attained at maximal velocity 124.3 ⫾ 24.5 ms and 207.2 ⫾ 41.3 ms respectively at 0 kg and 6 kg in the flexed knee. In the extended knee, the values were 151.2 ⫾ 40.1 ms and 222.4 ⫾ 59.6 ms respectively. 3.2. Effects of loading on EMG activity With increasing velocity of movement, enhanced (p ⬍ 0.01) absolute EMG activity was recorded in all triceps surae muscles whatever the load, except for 6 kg in the flexed-knee position (Fig. 2). When the load was progressively enhanced from 0 kg to 6 kg at maximal velocity of movement, the normalized EMG activity of
Continued.
the triceps surae increased (p ⬍ 0.05; Fig. 3) in the flexed-knee position. However, when the leg was extended, the normalized EMG activity showed no significant change with loading (Fig. 3). Fig. 4 illustrates the relative contribution of the Gast and Sol EMG activities when the loading was enhanced during slow and fast movements performed in the two knee positions. It appears that in all loading conditions at maximal velocity the Sol/MG and Sol/LG EMG ratios were not significantly different when the movements were performed with the knee flexed or extended. For smaller angular velocities as in all loading conditions the Sol contribution appeared to be larger in the flexed-knee position as compared to the extended leg. This figure also shows that in both knee positions the contribution of Gast EMG augmented progressively with increasing loads: this increase was not significant (p > 0.05), however. Since the Gast length in the extended knee was larger when compared to the flexed position, a possible effect on EMG activity was tested by recording the Mmax in both positions. No significant (p > 0.05) difference was observed in the two knee positions but Mmax amplitude
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
5
movement velocity. The burst increase at maximal velocity, from 0 kg to 6 kg loading, was 76.3 ⫾ 27.0% in the flexed-knee position, and in the extended-knee position the increase was 53.9 ⫾ 37.9%. 3.4. Unknown loading condition
Fig. 2. Relationship between angular velocity and EMG average in the MG (a–c) and Sol (b–d) both for no load and for a 6 kg load. These graphs compare the data recorded in the flexed and extendedknee position for all of our subjects. The linear equation in the flexed and extended knee positions with no load are respectively: y ⫽ 0.14x ⫺ 12.4 (r ⫽ 0.87; p ⬍ 0.001) and y ⫽ 0.17x ⫹ 13.8 (r ⫽ 0.59; p ⬍ 0.001) for the MG and y ⫽ 0.03x ⫹ 9.6 (r ⫽ 0.31, p ⬍ 0.01) and y ⫽ 0.10x ⫺ 9.7 (r ⫽ 0.61; p ⬍ 0.001) for the Sol. For 6 kg load, the values are respectively: y ⫽ 0.22x ⫹ 9.8 (r ⫽ 0.51; p ⬍ 0.001) and y ⫽ 0.20x ⫹ 32.8 (r ⫽ 0.38; p ⬍ 0.001) for the MG and y ⫽ 0.002x ⫹ 32.0 (r ⫽ 0.01; p ⫽ 0.90) and y ⫽ 0.07x ⫹ 13.4 (r ⫽ 0.21; p ⬍ 0.01).
were slightly greater in the flexed knee by 10.3 ⫾ 6.8%, 8.1 ⫾ 9.9% and 0.7 ⫾ 2.3% in LG, MG and Sol, respectively. 3.3. Effect of load on antagonist EMG activity As loading increased, the TA showed an EMG activity which remained nearly concurrent with that of the triceps surae muscles (Fig. 1), and increased (p ⬍ 0.01) from 0 kg to 6 kg countering plantarflexion. It should be pointed out, however, that this activity modulation of TA with increasing loads remained in a rather limited range of its maximal activation capacity, and only increased from 19.5 ⫾ 4.1% to 24.7 ⫾ 5.7% of the EMG recorded during maximal contraction. With increased loading, the TA EMG activity changes were larger in the flexed-knee positions as compared to the extended leg, and at maximal velocity this increase was 121.7 ⫾ 39.9% and 20.7 ⫾ 12.8%, respectively (Fig. 3). Under all loading conditions, TA behaved similarly to the agonist muscles (increased burst) during increasing
No significant difference in movement amplitude was observed when the sujects were either aware or unaware of the load opposing plantarflexion; however, in the two knee positions greater maximal velocities were recorded in unknown loading conditions for all loads except 6 kg (Figs. 5 and 6). At 0 kg the mean maximal velocity was increased (p ⬍ 0.05) from 767 ⫾ 62° s−1 to 1071 ⫾ 238° s−1 and from 558 ⫾ 76° s−1 to 654 ⫾ 108° s−1 in the flexed and extended-knee positions respectively. These results are illustrated in Fig. 5 in one subject who performed plantarflexions at maximal velocity. Whether the load was known or not did not affect the latency between EMG and movement onsets, but at 0 kg all agonist EMG burst durations were greater (p ⬍ 0.01) (142.7 ⫾ 26.4 ms vs. 125.1 ⫾ 16.8 ms and 176.7 ⫾ 30.2 ms vs. 150.4 ⫾ 17.2 ms in the flexed and extendedknee positions respectively) at unknown loading. On the other hand, at 6 kg these burst durations were shorter (p ⬍ 0.01) in unknown loading condition in the flexed-knee position (171.3 ⫾ 39.4 ms vs. 218.6 ⫾ 27.6 ms) but not significantly different in the extended-leg position (211.4 ⫾ 38.9 ms vs. 195.4 ⫾ 35.4 ms). Enhanced EMG activity was observed at unknown 0 kg in all the muscles, without any change in the Sol/Gast ratio. Under unknown loading conditions, a more detailed EMG analysis indicated that whatever the load or knee position, the EMG activities recorded from their onset through the first 25 ms of the movement were not significantly different from those observed in the known loading condition. However, at unknown 0 kg the EMG was increased (p ⬍ 0.05) in the flexed-knee position (62.0 ⫾ 17.5% and 391.8 ⫾ 97.5% in a time interval of 25–100 ms and after 100 ms, respectively) and also in the extended-leg position (57.5 ⫾ 39.7% after 100 ms). At a 6 kg load the only recorded EMG change in the unknown condition was a decrease (34.4 ⫾ 6.5%; p ⬍ 0.01) in flexed-knee position during the interval of 25– 100 ms. Since all triceps surae entities showed a similar type of behaviour, the data reported above are the means for the three heads.
4. Discussion The central nervous system strategy in generating fast voluntary movements in skeletal muscles consisting of different fibre types remains an exciting question in motor control. The triceps surae is an interesting model for such an experimental approach because it is com-
6
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
Fig. 3. Relationship between the angular velocity and the EMG in the different triceps surae heads and TA for the four different loads when the knee was in the flexed (left panels) or extended position (right panels). The angular velocity has been normalized with respect to the maximal value recorded during the experimental session, and each EMG is expressed as a percentage of the value measured in that condition.
posed of the Sol, which contains mainly type I fibres, and the Gast, which has nearly the same proportion of type I and type II fibres [16]. Moreover, the Sol is a monoarticular muscle, whereas the Gast is a biarticular one, and their strategy in voluntary movements may differ [12,21]. The main results from the present experiments are that with loading: (a) greater maximal plantarflexion velocities were recorded in the flexed as compared to the extended-knee position; (b) the longer duration and amplitude of agonist EMG bursts were recorded; (c) at maximal velocity, there was a decreased Sol/Gast ratio; (d) the co-activation of the TA and triceps surae muscles was enhanced; (e) unexpected sub-maximal loads induced a potentiation of neuromuscular activity. Because cross-talk between muscles was negligible in our experiments, we consider that the EMG behaviour observed was both load- and position-dependent. Compared to the extended position, the movements performed in the flexed-knee position were slightly larger (Section 3) and the EMG responses could therefore be greater [22,23,3], but it is generally accepted that EMG activities are comparable [6] for such small movement amplitude differences. The observation that, at constant load, the maximal velocity of movement was larger at a 90° knee flexion as compared to the extended-limb position was surprising, but this could be explained if the Gast activation was greater in the flexed-knee position. In fact, the results showed a decrease in the absolute average EMG
in all the triceps surae muscles in the flexed position as compared to the extended-knee position, an observation which is consistent with the work of Cresswell et al. [24]. As previously observed [12], the optimum position for the Gast with respect to the speed of the movement is reached when the knee is flexed to 90°. This finding suggested that the larger maximal plantarflexion velocity observed in the flexed-knee position should not be primarily related to a different neural command of the triceps surae entities, but more likely to mechanical muscle properties. This viewpoint is further supported by the present results, which indicate that with an increasing load the largest velocity of movement is consistently recorded at a 90° knee flexion. It could be argued that in the extended-knee position TA activation might be present to maintain the ankle position at 90°; this is not the case in the flexed-knee position since there is a resistive torque opposed to the plantarflexion torque. This viewpoint might be valid at 0 kg, but in fact similar results were observed with an increase in loading which kept the foot passive at 90°. It is interesting that the decrease in maximal movement velocity observed with an increasing load is associated with a tendency of the Sol/Gast EMG to decrease in the flexed knee. Such results agree with the finding that at 6 kg the Sol EMG does not change with increasing velocity (Fig. 2d) and suggest that motor unit activation is saturated in this muscle but not in the Gast, since EMG increases with speed (Fig. 2c). Such different behaviour of muscles
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
7
Fig. 4. Comparison of the mean ( ⫾ SE, n ⫽ 8) Sol/MG (a, b) and Sol/LG (c, d) EMG ratio of slow (200°/s) and maximal angular velocities in the flexed (a–c) and extended (b–d) positions. *p ⬍ 0.05; **p ⬍ 0.01.
with different fibre type composition is apparently consistent with the observation that, in Sol motor units, the activation threshold decrease during ballistic contractions is greater when compared to muscles which contain a smaller amount of slow motor units [25]. A triphasic EMG burst has been reported [2–5] in ballistic goal-directed movements. The first phase is the agonist initiation burst; the second is the antagonist EMG activity that stops the movement; and the third is an agonist response to the antagonist burst. Other studies do not report this neuromuscular strategy but refer to the co-activation of agonists and antagonists, e.g. in wrist flexion with no intention of stopping the movement [6] and elbow flexion or extension performed against gravity [26]. Our results also obviously show the co-activation
of TA and triceps surae in plantarflexion during movement with no intention of stopping, and furthermore indicate that this co-activation is reinforced by external loading. It has been suggested [27,28] that the conditions of our experiment may lead to the co-activation of TA and triceps surae, to stabilizing the complex ankle joint and thus contributing to the efficiency of the triceps surae with respect to increasing the load countering plantarflexion. Even if many interesting studies [8,29,30,17,13,10, 31,14] are taken into consideration, the effect of external loading on the first agonist burst time is not clear, possibly because of protocol differences. For example, the relationship between loading, burst duration and burst amplitude is rather confusing [32,29,13]. In our experi-
8
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
Fig. 5. Example of the EMG (rectified) activity patterns of the medial (MG) and lateral gastrocnemius (LG), soleus (Sol) and tibialis anterior (TA) together with the displacement and velocity traces recorded in one subject during plantarflexions at maximal velocities at zero load both with (a) and without knowledge (b) of the load. The arrows indicate the onset and end of EMG measurements.
mental conditions, increasing the load induces enhanced duration and amplitude of the first EMG pulse which is similar in all triceps surae muscles. These results indicate that during plantarflexion, loading: (1) is controlled by changes in muscle activation duration and intensity; (2) induces similar changes of contraction control strategy in muscle components with different fibre type compositions. It appears from Fig. 3 that in TA and all triceps surae components, the EMG changes in function of increasing load (and speed) are proportionally larger in the flexed as compared to extended-knee position and that this difference is obviously greater in Gast and TA as compared to Sol. In other words, the main loadingrelated muscle activation changes are those of Gast and TA in the flexed knee position. This observation also supports the viewpoint that the biarticular Gast is in a more efficient mechanical length with respect to velocity in the flexed as compared to the extended-knee position since loading is more closely related to a change in its neural activation when the leg is flexed. In the extended
position, the Gast activation changes in function of loading are more or less comparable to those of the monoarticular Sol which is, of course, very little affected by a difference in the articular knee position. It is interesting that in both knee positions the antagonist TA behaves nearly like the Gast and also shows a greater loadinginduced activation change in the flexed as compared to the extended knee. This observation suggests that a change in agonist activation also affects the antagonist [12]. Perhaps the most intriguing section of the results is the finding that in both knee angulations the maximal velocity of movement and the EMG activities were greater when the subjects were not informed of the load opposing plantarflexion. The observation that EMG activity onset and pattern over the first 25 ms do not change in agonist and antagonist muscles in unknown loading conditions suggests that central programming is not affected. Thus, increases in reflex and, later, voluntary activities should explain the larger EMG responses
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
9
Fig. 6. Histograms showing the mean ( ⫾ SE, n ⫽ 4) angular velocity, the EMG of the medial and lateral gastrocnemius, soleus and tibialis anterior for two different loads (0 and 6 kg) in the two knee positions (knee flexed, KF; and extended, KE) both with and without knowledge of the load. *p ⬍ 0.05; **p ⬍ 0.01, ***p ⬍ 0.001.
and speed of movement recorded in unexpected loading conditions. This viewpoint is consistent with the fact that the latter part (after 100 ms) of the EMG activity is significantly greater in unexpected loading. The interesting
difference between EMG activities during maximal movement velocity under unexpected 0 kg and 6 kg loading conditions is that the increase in EMG activity in the 25–100 ms interval is larger in the flexed as com-
10
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
pared to the extended knee and decreases with loading, so slowing the movement down. These results support the suggestion that with an unexpectedly reduced inertial load, spindle afferents are induced [17] which, in our experiments, should be greater in ballistic movements at 0 kg as compared to the slower tension development related to increased loading. This speculation is in line with the observation that in the TA antagonist the EMG recorded between 25 and 100 ms does not decrease with loading. In fact, in the antagonist muscle, spindle afferents should also be induced by ␣–␥ co-activation and stretching during plantarflexion. Another interesting difference of the antagonist EMG activity during unexpected loading (Fig. 5) is the presence of a large burst later than 100 ms. This burst precedes the one related to foot dorsiflexion. It is difficult to deduce from our experimental data whether this burst is related to the increased duration of the antagonist co-activation during plantarflexion and/or to the decreased latency of its activity during dorsiflexion. However, these experiments are in line with the “equilibrium-point model” of EMG patterns during movements, and the proposition that peripheral afferents contribute to modulate the co-contraction and adaptative stiffness of the ankle joint with respect to the control of movement [33]. It is concluded that increasing the load during plantarflexion in humans induces changes in neuromuscular strategies that contribute to the efficiency of contractile activity during rapid movements. The results also indicate that unexpected sub-maximal loading induces a potentiation of both the central and peripheral neural drives to the agonist and antagonist muscles which increase the speed of movement. Acknowledgements This work was supported by NATO CRG no. 930261, the Fonds National de la Recherche Scientifique of Belgium and the Conseil de la Recherche of the Universite´ Libre de Bruxelles. The authors are grateful to Miss Anne Deisser for her assistance in the preparation of the manuscript. References [1] Angel RW. Electromyography during voluntary movements: the two-burst pattern. Electroencephalogr Clin Neurophysiol 1974; 36:493–8. [2] Gottlieb GL, Latash ML, Corcos DM, Livbinskas TJ, Agarwal GC. Organizing principles for single joint movements. V. Agonist–antagonist interactions. J Neurophysiol 1992;67:1417–27. [3] Hallett M, Marsden CD. Ballistic flexion movements of the human thumb. J Physiol (Lond) 1979;294:33–50. [4] Marsden CD, Obeso JA, Rothwell JC. The function of the antagonist muscle during fast limb movements in man. J Physiol 1983;335:1–13.
[5] Meinck H, Benecke R, Meyer W, Hohne J, Conrad B. Human ballistic finger flexion: uncoupling of the three-burst pattern. Exp Brain Res 1984;55:127–33. [6] Mustard BE, Lee RG. Relationship between EMG patterns and kinematic properties for flexion movements at the human wrist. Exp Brain Res 1987;66:247–56. [7] Benecke R, Meinck HM, Conrad B. Rapid goal-directed elbow flexion movements: limitations of the speed control system due to neural constraints. Exp Brain Res 1985;59:470–7. [8] Berardelli A, Rothwell JC, Day BL, Kachi T, Marsden CD. Duration of the first agonist EMG burst in ballistic arm movements. Exp Brain Res 1984;304:183–7. [9] Smeets JBJ, Erkelens CJ, Denier Van Der Gon JJ. Adjustments of fast goal-directed movements in response to an unexpected inertial load. Exp Brain Res 1990;81:303–12. [10] Lestienne F. Effects of inertial load and velocity on the braking process of voluntary limb movements. Exp Brain Res 1979; 35:407–18. [11] Simmons RW, Richardson C. Effects of different types of mechanical load on the duration of the initial agonist pulse. Exp Brain Res 1993;92:524–7. [12] Carpentier A, Duchateau J, Hainaut K. Velocity-dependent muscle strategy during plantarflexion in humans. J Electromyogr Kinesiol 1996;6(4):225–33. [13] Le Bozec S, Maton B, Cnockaert JC. The synergy of elbow extensor muscles during dynamic work in man. I. Elbow extension. Eur J Appl Physiol 1980;44:225–69. [14] Wadman WJ, Denier Van Der Gon JJ, Geuze RH, Mol CR. Control of fast goal-directed arm movements. J Hum Mov Stud 1979;5:3–17. [15] Duchateau J, Le Bozec S, Hainaut K. Contributions of slow and fast muscles of triceps surae to a cyclic movement. Eur J Appl Physiol 1986;55:476–81. [16] Johnson MA, Polgar J, Weightman D, Appleton D. Data on the distribution of fibre types in thirty-six human muscles: an autopsy study. J Neurol Sci 1973;18:111–29. [17] Latash ML. Control of fast elbow movement: a study of electromyographic patterns during movements against unexpectedly decreased inertial load. Exp Brain Res 1994;98:145–52. [18] Schmidt RA, McGown CM. Terminal accuracy of unexpectedly loaded rapid movements: evidence for a mass-spring mechanism in programming. J Motor Behav 1980;12:149–61. [19] Moritani T, Oddson L, Thorstensson A. Electromyographic evidence of selective fatigue during the eccentric phase of stretch/shortening cycles in man. Eur J Appl Physiol 1990; 60:425–9. [20] Solomonow M, Baratta R, Bernardi M, Zhou B, Lu Y, Zhu M, Acierno S. Surface and wire EMG crosstalk in neighbouring muscles. J Electromyogr Kinesiol 1994;4:131–42. [21] Jacobs R, Bobbert MF, Van Ingen Schenau GJ. Function of mono-and biarticular muscles in running. Med Sci Sports Exerc 1993;25(10):1163–73. [22] Brown SH, Cooke JD. Initial agonist burst duration depends on movement amplitude. Exp Brain Res 1984;55:523–7. [23] Cheron G, Godaux E. Self-terminated fast movement of the forearm in man: amplitude dependence of the triple burst pattern. J Biophys Biomech 1986;10:109–17. [24] Cresswell AG, Lo¨scher WN, Thorstensson A. Influence of gastrocnemius muscle length on triceps surae torque development and electromyographic activity in man. Exp Brain Res 1995; 105:283–90. [25] Desmedt JE, Godaux E. Ballistic contractions in fast or slow human muscles: discharge patterns of single motor units. J Physiol 1978;285:185–96. [26] Virji-Babul N, Cooke JD, Brown SH. Effects of gravitational forces on single joint arm movements in humans. Exp Brain Res 1994;99:338–46.
A. Carpentier et al. / Journal of Electromyography and Kinesiology 9 (1999) 1–11
[27] Baratta R, Solomonow M, Zhou B, Letson D, Chuinard R, D’Ambrosia R. Muscular activation: the role of the antagonist musculature in maintaining knee stability. Am J Sports Med 1988;16:113–22. [28] Solomonow M, Baratta R, Zhou BH, D’Ambrosia R. Electromyogram coactivation patterns of the elbow antagonist muscles during slow isokinetic movement. Exp Neurol 1988;100:470–7. [29] Gottlieb GL, Chen CH, Corcos DM. Relations between joint torque, motion and electromyographic patterns at the human elbow. Exp Brain Res 1995;103:164–7. [30] Hoffman DS, Strick PL. Step-tracking movements of the wrist. III. Influence of changes in load on patterns of muscle activity. J Neurosci 1993;13(12):5212–27. [31] Sherwood DE, Schmidt RA, Walter CB. Rapid movements with reversals in direction. II. Control of movement amplitude and inertial load. Exp Brain Res 1988;69:355–67. [32] Cooke JD, Brown SH. Movement-related phasic muscle activation. III. The duration of phasic agonist activity initiating movement. Exp Brain Res 1994;99:473–82. [33] Latash ML, Goodman SR. An equilibrium-point model of electromyographic patterns during single-joint movements based on experimentally reconstructed control signals. J Electromyogr Kinesiol 1994;4:230–41. Alain Carpentier received the master’s degree in physical education and sports sciences from the Universite´ Libre de Bruxelles, Belgium, in 1988. In 1990 he joined the Laboratory of Biology and Research Unit in Neurophysiology (Professor K. Hainaut) as a Ph.D. student, where he presently works on the neuromuscular strategies in humans during voluntary contractions and will present his Ph.D. thesis in 1998. His other main research interest is related to the study of EMG changes during fatigue, including the analysis of single motor unit behaviour. He received the scientific award of the “Socie´te´ de Biome´canique” (1997) for that work.
11
Jacques Duchateau received the master’s degree in physical education and sports sciences (1977) and a Ph.D. degree (1981) from the Universite´ Libre de Bruxelles, Belgium. Since 1979, he has been affiliated with the Laboratory of Biology and Research Unit in Neurophysiology at the Universite´ Libre de Bruxelles. He was assistant professor from 1978 to 1984 and became full professor in 1985. His major interests focus on neuromuscular physiology and muscle biomechanics. Dr Duchateau serves on the Editorial Board of the Journal of Electromyography and Kinesiology. Karl Hainaut received Ph.D. degrees in physical education (1968) and in physiological sciences (1975) from the Universite´ Libre de Bruxelles, Belgium. Full-time professor at the Universite´ Libre de Bruxelles and director of the Laboratory of Biology and Research Unit in Neurophysiology, his research in neuromuscular physiology has consistently been supported by the National Fund for Scientific Research, the Academies of Sciences (Acade´mie The´re´sienne) and of Medicine, and more recently by NATO for a collaborative project with North American laboratories.