Force generation performance and motor unit recruitment strategy in muscles of contralateral limbs

Journal of Electromyography and Kinesiology 9 (1999) 121–130

Force generation performance and motor unit recruitment strategy in muscles of contralateral limbs M. Bernardi

a,b,*

, F. Felici a, M. Marchetti a, F. Montellanico a, M.F. Piacentini a, M. Solomonow c

a

c

Istituto di Fisiologia Umana, Universita` degli Studi di Roma “La Sapienza”, Roma, Italy b IRCCS Santa Lucia, Roma, Italy Bioengineering Laboratory, Louisiana State University Medical Center, Department of Orthopaedic Surgery, New Orleans, LA, USA

Abstract The purpose of the present study was to determine whether the motor unit (MU) recruitment strategy of the agonist and antagonist muscles in the dominant arm differs from that in the non-dominant arm. The median frequency (MF) of the power density spectrum (PDS) of the electromyogram (EMG) was used as a tracking parameter to describe the MU recruitment. In 8 subjects the EMG was recorded from the biceps brachii and triceps brachii of each limb during isometric elbow flexion performed in a ramp fashion. Force was increased from 0 to 100% of the maximum voluntary contraction (MVC) in 3 s following a track displayed on an oscilloscope. When comparing the dominant versus non-dominant arm we found no statistical difference in the MU recruitment pattern of the biceps brachii and the triceps. Because the dominant arm was not always the better performing arm, we grouped the data according to the ability of the subjects to track the ramp signal. In this case we found a statistically significant difference between the better and worse performing arm in the full MU recruitment of the biceps. A more precise and accurate control of the increase in force was obtained when the central nervous system selected a slower and prolonged recruitment of MUs in the agonist muscle.  1999 Published by Elsevier Science Ltd. All rights reserved. Keywords: Muscle; Motor unit; Dominance; Electromyography; Skill

1. Introduction A preferential use of muscle groups in the dominant arm with respect to the non-dominant may induce different biochemical and structural modifications to the muscle fibers. Fugl-Meyer et al. [9] did actually demonstrate a greater percentage of type I fibers in the right exstensor carpi radialis brevi with respect to the contralateral muscle. The authors suggested that this difference may reflect a different training of the muscles and may explain the different fatiguability of the right hand compared with the left. Lower fatiguability of the right arm compared to the left arm was demonstrated by De Luca et al. [5] when recording the decrease in the median frequency of the power density spectrum of the electromyo-

* Corresponding author. Tel.: ⫹ 39-06-499-10816; fax: ⫹ 39-06445-2303; e-mail: [email protected].

graphic signals. Also these authors attributed the difference to the training effect due to a preferential use. The general hypothesis of the present investigation is that the different use of the dominant and the non-dominant arm may induce a different motor control strategy. This difference can be related to a different muscle fiber type composition. Based on Henneman’s size principle [11], Solomonow et al. [15] were able to experimentally reproduce the natural recruitment process in the cat’s gastrocnemius muscle using sophisticated electrical stimulation apparatus. During these experiments the electromyograms (EMGs) of the active muscle fibers were recorded. The results demonstrated a linear increase in the median frequency (MF) of the power density spectrum (PDS) of the EMG when motor units were progressively recruited. The maximum MF value was obtained when recruitment had been completed. Additional force above this level could be obtained however by increasing the motor unit

1050-6411/99/$ - see front matter  1999 Published by Elsevier Science Ltd. All rights reserved. PII: S 1 0 5 0 - 6 4 1 1 ( 9 8 ) 0 0 0 4 3 - 1

122

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

firing rate. Indeed, the authors demonstrated that an increase in the MU firing rate did not affect the MF. On the basis of these results it has been shown that the increasing average conduction velocity of the MU action potentials during recruitment is the major contributor to variations in MF. The MU firing rate may have only a limited effect on the MF and only at low levels of recruitment [6]. Therefore, the MF of the PDS has been used as a tracking parameter to describe motor unit recruitment strategy in humans during isometric voluntary contractions [2–4,14]. Based on Solomonow et al. [15], the force level at which the highest MF value occurred corresponded to the force level at the full MU recruitment. The quoted studies on humans have shown that the recruitment strategy varies in different muscles. Within the same muscle the recruitment strategy varies depending on the motor task requested of the muscle (increasing force either in a step-wise or in a ramp-wise fashion) and whether the muscle is acting as agonist or antagonist [2,4,14]. Recently [3], MF was used to study possible changes in motor unit recruitment strategy due to motor skill acquisition. The authors found that, after 6 weeks of specific training, the subjects demonstrated an improved ability to track the reference target. In other words, with training, the subjects were able to increase their force up to 100% maximum voluntary contraction (MVC) both more linearly and with the right timing (i.e. the slope necessary to conclude the performance in 3 s). This improvement was accompanied by a more uniform MU recruitment than before training, meaning that the recruitment was completed at a higher percentage of the MVC. It was concluded that skill acquisition has an impact on recruitment strategy. The purpose of this research was to evaluate the motor performance of the dominant and the non-dominant arm during the execution of the same motor task as the latter study (linearly increasing isometric force from 0 to 100% MVC in 3 s). Motor performance was assessed on the basis of the ability to track the reference target. Based on the results of the previously mentioned study [3], we hypothesized that the different motor performance (if any difference exists) between the two arms might be due to a different recruitment pattern. That is to say, evaluating the relationship between force and MF, we expected to find full MU recruitment at a higher force level in the dominant arm which was assumed to be the arm better able to perform the task. A more graduated motor unit recruitment can explain a more fine force regulation, and therefore a greater motor ability.

from 21 to 36), volunteered to participate in this study. They were informed of the purpose and the protocol of the experiment and signed an informed written consent form. All of them were right handed. 2.2. Instrumentation Surface EMG from the medial head of the biceps brachii and the lateral head of the triceps brachii were detected with two sets of bipolar silver/silver chloride electrodes (Blue Sensor Medicotest, Olstykke, Denmark), with a diameter of 7 mm. Before the electrodes were placed on the arm, the skin was slightly abraded and cleaned with an alcohol pad in order to reduce the skin impedance. Electrodes were gelled and placed on the medial head of the biceps brachii and the lateral head of the triceps brachii, halfway between the motor point and the distal tendon with a center-to-center distance of 30 mm. The distances of the electrodes from anatomical reference points such as olecranon and acromion, were measured in order to have a relatively comparable situation in subjects with different anatomical characteristics. A reference electrode was attached to unrelated tissue on the arm. Subjects were seated comfortably. Each subject’s arm was positioned on an inclined plane in order to obtain a fixed elbow angle of 1.57 radians (90°) (see Fig. 1). The height of the chair on which the subject was seated was regulated in order to allow the subject’s arm to rest comfortably on the inclined plane. Further details of this set-up are described in Felici et

2. Methods 2.1. Subjects Eight healthy sedentary subjects, 5 males and 3 females, with an average age of 28.4 ⫾ 5 years (range

Fig. 1. Force measurement set-up: The elbow angle is equal to 1.57 radians (90°). The wrist strap is positioned at a fixed distance (0.20 m) from the elbow center of rotation. The experimental set-up allows the forearm to be orthogonal to the force transducer cable.

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

al. [7]. A wrist strap was positioned on the forearm at a fixed distance (0.20 m) from the elbow center of rotation, and connected to the force transducer (Kistler, type 9311A; range from ⫺ 5000 N to 5000 N; sensitivity ⫺ 3.93 and 3.93 pC/N; linearly ⱕ 0.3% FSO). An angle of 1.57 radians between the subject’s forearm and the force transducer cable was strictly observed. This set-up allowed the subject to produce an isometric contraction of the biceps. The flexion torque collected through the force transducer was transmitted to a Kistler 5001 amplifier and then to an oscilloscope. The EMG from each muscle was differentially amplified (Grass RPS 107 amplifier), with a band-pass filter from 3 to 1000 Hz. The gain for the agonist muscle was up to 1000 times and for the antagonist up to 2000. A data acquisition card (National Instruments ATMIO 16E10) digitized the EMG and force signals. The signals were stored in an IBM compatible computer at a sampling rate of 2048 samples per second. One dual beam storage oscilloscope (Tektronix 5113) displayed the raw EMG signals from each muscle. A second oscilloscope displayed two lines, one showing the force exerted by the subject, the other representing the requested target force. The latter consisted of a line rising from the bottom to the top of the monitor in 3 s simulating a ramp of increasing force (from now on, this line will be called target line). A function generator (Schlumberger GBM 661A) provided the target line. The subjects tried to follow the ramp applying a linearly increasing contraction. 2.3. Protocol The experimental method consisted of a preparatory phase and an experimental phase. During the first phase general information about each subject was obtained. Then, electrodes were positioned and the subject was comfortably seated as previously described (Fig. 1). To assert that cross-talk between the medial head of the biceps and the lateral head of the triceps brachii did not contaminate the data [3,16], the EMGs of the two muscles were plotted against each other on the oscilloscope during each trial. If the x–y plot of the EMGs had formed an ellipse, it would have been concluded that the two signals were identical but with a phase shift. In such a case the configuration of the EMGs would have been discarded. Only trials with confirmed independent myolectrical signals were used. During the experimental phase, subjects were asked to exert their MVC, with both the dominant and the nondominant arm. Verbal encouragement and visual feedback helped subjects to exert their MVC. The highest force of three attempts was taken as representative of subjects’ MVC. A resting period of at least 5 min between attempts was strictly observed to minimize the effect of fatigue. Adjustments to the oscilloscope were made so that 0

123

and 100% MVC corresponded to the bottom and the top of the screen. Subjects were then asked to perform a linearly increasing isometric contraction with each arm from 0 to 100% MVC watching the screen. After a warning signal, the subjects had to follow the target line. Five-minute rest periods were observed between trials to avoid fatigue. For each arm the actual force of the subject and EMGs from the electrodes placed on the biceps and the triceps were recorded. On the basis of the following criteria some trials were rejected. According to Bernardi et al. [3], trials where there was not a correspondence between the maximum EMG root mean square (RMS) and maximum strength, were not accepted. Trials were rejected either when the motor performance was too different from the target (see the ability evaluation in the force analysis paragraph) or when the subjects’ force value did not reach 90% MVC. Another criterion of rejection was the duration of the ramp which had to last 3 s. Trials in which the subjects increased the force too quickly or too slowly were rejected. The criterion of rejection was a slope higher than 41% MVC per second or lower than 25% MVC per second. These limits corresponded to a percentage error of the slope equal to 25% (see the following paragraph: force analysis). Recording of the EMG and force signal started 1 s before the beginning of the execution of the ramp. The total duration of the record was 5 s. 2.4. EMG analysis The EMG signals were divided into segments of 512 points (0.25 s), each overlapping the previous by 50%, for a total of 31 epochs per trial. The epoch length was chosen in order to obtain an EMG signal that could be considered stationary with acceptable approximation. The 3 s ramp duration allowed a convenient amount of data to be processed, minimizing fatigue. The EMGs were analyzed both in the time domain and in the frequency domain. In the time domain the root mean square (RMS) of the signal was considered. For each epoch of duration T the RMS was calculated as:

冪冕 T

EMGRMS ⫽

1 2 x (t)dt T 0

where x(t) is the xEMG signal at the time t. Epochs were windowed with a Hanning window and zero padded by 1536 points to obtain a frequency resolution of 1 Hz. The evaluation of the MF is performed by means of a standard Fast Fourier Transform (FFT) on 2048 samples. The MF, which is the frequency value

124

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

that divides the spectrum into two regions of equal power, was defined as:

冕

fmed

MF:

冕 ⬁

P(f)df ⫽

fmed

0

冕 ⬁

1 P(f)df ⫽ P(f)df 2 0

A procedure to cancel the system noise [3] was used in processing the data. The first epoch in a trial, corresponding to 0% MVC (or no muscle activity), was used to assess the system noise level. We assumed the noise to be uncorrelated with the signal and the PDS was subtracted from all subsequent epochs. The noise level was in the range of about 40–45 dB. The signal noise ratio ranged from 35 to 65 dB in the force levels between 10 and 100% MVC. 2.5. Maximal torque analysis values Absolute force values obtained during MVC measurements were tested to evaluate for possible differences between arms. The force values were multiplied by the length of the lever arm to obtain the torque values. 2.6. Force analysis The signal tracking ability in performing the ramp was evaluated on the basis of two parameters named standard error of estimate (SEE) [13] and percentage error of the slope (PES). Both were calculated using the first order polynomial regression curve of the normalized force values. The regression curve was obtained with the least squared method. As assessed in a previous study [3], the force values below 10% MVC and above 90% MVC were not used in the regression analysis of the force. The SEE was calculated according to the following formula:

SEE ⫽

冪

冘 i

(Ei ⫺ r)2

lated. Whether the ramp performed was slower or faster than the ideal ramp was not relevant to the study. Therefore, the difference was converted into positive values with a square and a successive square root. This difference was expressed as percentage error with respect to the ideal slope and was named PES. As previously stated a ramp with a PES greater than 25% was rejected. 2.7. Statistical analysis All the statistical analyses were performed according to Sachs [13]. The SEE and PES data of each subject were evaluated comparing the results of the right and left arms. Each set of data was evaluated through paired Student’s t-tests (P ⬍ 0.05) to determine if a statistically significant difference existed between the tracking ability of the two arms. Differences in full MU recruitment were evaluated taking the force value (%MVC) at the highest MF value. Using the same grouping as above, possible differences between the two arms were evaluated through a paired Student’s t-test (P ⬍ 0.05).

3. Results All subjects declared that their dominant arm was the right one. 3.1. Maximal torque analysis Women showed an average maximal elbow torque equal to 53 ⫾ 4 Nm for the dominant (right) arm and 48 ⫾ 6 Nm for the non-dominant arm. Men showed an average maximal elbow torque equal to 87 ⫾ 10 Nm for the dominant (right) arm and 86 ⫾ 12 Nm for the nondominant arm. Maximal elbow torque differences between the dominant and non-dominant arms were not statistically significant for both genders.

0

n⫺2

where Ei is the actual normalized force value and r is the corresponding value of the linear regression curve. Therefore, SEE is expressed as %MVC being the average difference between the actual force and the linear regression curve, both expressed as a percentage of the MVC. The lower the SEE the better the ability in tracking the ramp. Indeed this parameter was actually used as the first indicator to assess the ability of the subject in performing the ramp. Ramps in which the SEE was greater than 4% MVC were rejected. The ideal slope for a 3 s ramp was an increment of 33.3% MVC per second. The difference between the ideal slope and the linear regression slope was calcu-

3.2. Force analysis Force and torque were used in the following part of this paper interchangeably. In fact both data were normalized with respect to the maximum and the moment arm was constant for all subjects. In accordance with our criteria of rejection, an average of four ramps for each arm were discarded for each subject. For each arm, the number of trials performed by the subjects ranged from 7 to 13 in order to obtain 5 good ramps. In the 5 “good” ramps: 1. the the 2. the 3. the

EMG RMS always reached its highest value at highest force value; SEE was lower than 4% MVC; PES was lower than 25%;

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

4. the maximal force value of the ramp was at least 90% MVC. Fig. 2 shows a typical trial of the test with part of the processing procedure regarding the force analysis. The actual force applied by the subject (continuous line) and the ideal ramp the subject was requested to follow (dashed line) are shown in the top graph. The actual ramp curve (dotted line) and its best fitting curve (first order polynomial regression) (continuous line) are shown in the bottom graph. All the curves are expressed in percentage values, normalizing the force data of the ramp in accordance with the MVC force measured during the test. The section between 10 and 90% MVC was chosen in the evaluation of ramp tracking ability (linear regression curve for the SEE and PES measurements), to avoid the inclusion of tracking errors due to the ramp transients at the beginning and at the end of the performance. In fact, at the beginning of the trials, the subjects found it particularly hard to increase the force linearly, the first 10% MVC of the force increment being either too fast or too slow. The force between 90% and 100% MVC was excluded because practically all the subjects slowed the force exertion in the attempt to reach the MVC. In other words the slope of this part of the ramp

Fig. 2. A typical trial. The top figure shows the actual ramp performed by the subject (continuous line) and the ideal ramp (dashed line) the subjects were requested to follow (33% MVC per second). The bottom figure shows the actual ramp (dotted line) and the best fitting curve of the actual force (first order polynomial regression). The actual regression (continuous line) is calculated between the 10% and the 90% MVC (dotted vertical lines). The values of the standard error of estimate (SEE) and the percentage error of the slope (PES) are also shown.

125

was reduced compared to the slope of the force increment from 10 to 90% MVC. The SEE represents the capability of the subject to track the ramp-signal as linearly as possible. The results showed that the SEE ranged between 0.89 and 3.78% MVC for the right arm and ranged between 0.5 and 3.7% MVC for the left arm. Table 1 shows, for each subject, the mean value of the SEE and the PES of the 5 good ramps of each arm. Comparing the results of the two arms (Table 1) with a paired Student’s t-test, no statistically significant difference was found. The analysis of the slope showed that the subjects performed the ramps either slower or faster than the ideal ramp (33.3% MVC per second slope). There was a consistency in the two arms in performing all ramps at approximately the same speed: too slow or too fast, i.e., when a subject performed too slowly with the right arm, he/she performed too slowly with the left arm also, and vice versa. Only subject number 7 performed the ramp slower than 33.3% MVC with the right arm and faster than 33.3% MVC with left arm. The mean actual slopes were 31.6 ⫾ 3.4% MVC per second for the dominant and 31.9 ⫾ 4.64% MVC for the non-dominant arm, ranging from 25% to 40% MVC per second for the dominant and 25% and 41% MVC per second for the non-dominant arm. The PES ranged from 1.21% to 19.09% for the dominant arm and from 1.12% to 21.21% for the non-dominant arm. The PES values (Table 1) showed no statistically significant difference when comparing the right arm and the left arm. The intra subjects evaluation (comparison of the subjects’ 5 ramps per arm with a paired Student’s t-test) showed that four out of eight subjects had a statistically significant lower SEE in the right arm, and three subjects showed a statistically significant lower SEE in the left arm. Only one subject did not show any statistical significance between the two arms. In Table 2 the same data as Table 1 are rearranged. The first column shows the mean SEE values of the arm with the lower SEE. The second column shows the mean SEE values of the arm with the higher SEE. We then performed a further paired Student’s t-test and there was a significant difference between the two arms. From then on the arm with the lower SEE was referred to as the better performing arm (BPA), being considered the arm with the better tracking ability. In the BPA, the SEE ranged between 1.06% MVC and 1.74% MVC and in the worse performing arm between 1.32% MVC and 3.4% MVC. The third and fourth columns of Table 2 show the mean PES values rearranged according to the arm with the lower and the higher SEE respectively. There was a statistically significant difference between the BPA and the worse performing arm (WPA).

126

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

Table 1 Ramp accuracy evaluation in accordance with the standard error estimate (SEE) and the slope error in percentage (PES) with respect to the ideal 33.3% MVC per second, in the right and in the left arm. The SEE and the PES are based on the analysis of the first order polynomial regression curve of the actual ramp performed by the subject. For each parameter the differences between arms are not statistically significant Subjects

SEE Right arm (% MVC)

SEE Left arm (% MVC)

PES Right arm (%)

PES Left arm (%)

1 2 3 4 5 6 7 8 Mean ⫾ SD

1.1 1.7 2.07 1.5 1.4 2.4 1.9 1.32 1.67 ⫾ 0.44

1.62 3.4 1.74 2.4 1.9 1.06 1.4 1.32 1.8 ⫾ 0.7

6.06 2.42 12.12 12 1.21 1.21 19.09 14.24 8.5 ⫾ 6.7

21.21 21.21 6.36 9 14.24 5.15 4.24 14.24 11.9 ⫾ 6.8

Table 2 Ramp accuracy evaluation in accordance with the standard error estimate (SEE) and the slope error in percentage (PES) with respect to the ideal 33.3% MVC per second, in the better (BPA) and worse (WPA) performing arm. The BPA was established on the basis of the lower SEE Subjects

SEE BPA (% MVC)

SEE WPA (% MVC)

PES BPA (%)

PES WPA (%)

1 2 3 4 5 6 7 8 Mean ⫾ SD

1.1 1.7 1.74 1.5 1.4 1.06 1.4 1.32 1.4 ⫾ 0.25

1.62 3.4 2.07 2.4 1.9 2.4 1.9 1.32 2.21 ⫾ 0.66

6.06 2.42 6.36 12 1.21 5.15

21.21 21.21 12.12 9 14.24 1.21 19.09 14.24 14.04 ⫾ 6.78

3.3. Neural activation The biceps brachii’s EMG RMS was used only to test whether the force increments were actually paralleled by increments in the biceps activation. Indeed an increment in force which was not accompanied by an RMS increment was possibly due to the activation of muscles other than the biceps. For this reason, as previously described, all trials in which no perfect parallelism between force and RMS increment was found, were discarded. 3.4. Biceps The biceps MF results paralleled the results obtained in tracking ability. The force value at the highest MF (%MVC) will be considered full MU recruitment following Solomonow et al. [15] and in accordance with previous papers on humans [2–4,14]. Table 3 shows the mean force values at full MU recruitment. Column 1 shows the results of the right biceps, column 2 shows the results of the left arm, column 3 shows the results

14.24 6.46 ⫾ 4.5

of the BPA and the column 4 the results of the WPA. When comparing right and left arm with a paired Student’s t-test, no statistically significant difference was obtained in full MU recruitment. Conversely, when comparing the data according to tracking ability, the values showed a statistically significant difference in the recruitment pattern. On average the BPA had full MU recruitment at about 80% ⫾ 8% MVC and the WPA at about 73% ⫾ 7% MVC. In each subject, the arm that had the lower SEE (i.e., more able to track the ramp) showed full MU recruitment at a higher percentage of the MVC. In the only subject in which the tracking ability of the two arms was exactly the same, also the full MU recruitment occurred at the same percentage of force. 3.5. Triceps The triceps EMG was recorded in order to detect the antagonist participation. Increasing antagonist activity during the flexion effort was demonstrated by the fact that the triceps EMG RMS increased paralleling the

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

127

Table 3 Force percentage of the biceps MVC at the highest median frequency value (MF) for the right and left biceps muscles (columns 1 and 2). Force percentage of the biceps MVC at the highest median frequency value (MF) for the better (BPA) and worse (WPA) performing arm (columns 3 and 4) Subjects

Force value at the highest MF (% MVC) Right biceps

Force value at the highest MF (% MVC) Left biceps

Force value at the highest MF (% MVC) BPA

Force value at the highest MF (% MVC) WPA

1 2 3 4 5 6 7 8 Mean ⫾ SD

81 82 87 86 84 72 64 68 78 ⫾ 8.7

69 73 91 80 72 74 71 68 75 ⫾ 7.5

81 82 91 86 84 74 71 68 80 ⫾ 7.9

69 73 87 80 72 72 64 68 73 ⫾ 7.3

increment of force up to a certain value. For each subject the mean force values where the RMS reached the maximum are shown in the first and second columns of Table 4 for the right and the left arm respectively. The third and fourth columns show the data rearranged according to tracking ability. The maximum triceps RMS recorded during flexion effort did not show any statistical difference between right and left arm and BPA and WPA. Columns 1 and 2 of Table 5 show the force values at the highest MF (elbow flexion percentage MVC) for the right and the left arm and columns 3 and 4 show the force values at the highest MF for the BPA and the WPA. The maximum MF in the triceps EMG signal was obtained at 71% ⫾ 6.5% of the flexor MVC in the right arm. In the left arm the corresponding value was 66% ⫾ 12% of the flexor MVC. No statistically significant difference was noted between right and left arms or between BPA or WPA (69% ⫾ 10% of the flexor MVC versus 68% ⫾ 11% of the flexor MVC).

4. Discussion The main finding of the present study was that the skill in performing a simple motor task is related to a more uniform distribution of motor unit (MU) recruitment. In the arm in which the requested motor task was better performed, the full MU recruitment occurred at about 80% MVC, while in the other it occurred at about 73% MVC, the difference being statistically significant at P ⬍ 0.05. These results confirm and expand on the results of a previous study on the effects of skill acquisition on motor unit recruitment strategy [3]. As previously mentioned, force and torque have been used in this paper interchangeably. In fact both data were normalized with respect to the maximum and the moment arm was constant for all subjects. No difference was measured in the MVC of right dominant versus left non-dominant arm in our subjects. Anthropometric measurements (arm circumference and skinfold measurements) also failed to detect any substantial dif-

Table 4 Force percentage of the biceps MVC at the highest root mean square (RMS) for the right and left triceps muscles, and for the better (BPA) and worse (WPA) performing arm. For each parameter the differences between arms are not statistically significant Subjects

Force at highest RMS (% MVC) Right triceps

Force at highest RMS (% MVC) Left triceps

Force at highest RMS (% MVC) BPA

Force at highest RMS (% MVC) WPA

1 2 3 4 5 6 7 8 Mean ⫾ SD

88 83.6 100 99 90 100 89 98 93.5 ⫾ 6.4

94 100 97 83 93 97 98 99 95.1 ⫾ 5.4

88 83.6 97 99 90 97 98 98 93.8 ⫾ 5.8

94 100 100 83 93 100 89 99 95.7 ⫾ 6.3

128

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

Table 5 Force percentage of the biceps MVC at the highest median frequency value (MF) for the right and left triceps muscles (columns 1 and 2). Force percentage of the biceps MVC at the highest median frequency value (MF) for the better (BPA) and worse (WPA) performing arm (columns 3 and 4) Subjects

Force value at the highest MF (% MVC) Right triceps

Force value at the highest MF (% MVC) Left triceps

Force value at the highest MF (% MVC) BPA

Force value at the highest MF (% MVC) WPA

1 2 3 4 5 6 7 8 Mean ⫾ SD

72 76 80 69 61 75 63 72 71 ⫾ 6.5

73 76 69 47 61 51 84 68 66 ⫾ 12.5

72 76 69 69 61 51 84 72 69 ⫾ 9.8

73 76 80 47 61 75 63 68 68 ⫾ 10.7

ference between the two arms. Therefore, we assume that about the same muscular mass can be assigned to the two biceps muscles. Although the dominant arm is obviously the more often used particularly in skill activities, we hypothesize that the conditioning activity (resistance exercise) is typically as equally demanding on both arms and its trophic effects are thus bilaterally distributed [8,12]. In conclusion, these results would confirm the assumption that a possible difference in generating force between the two arms, while performing complex motor activities, would only be explained by better muscle co-ordination in one arm. No difference in the maximal force of the two arms should be present when considering the performance of a single muscle [10]. In order to measure human motor ability, a usual procedure is to ask the subject to follow a track with some kind of pointer. The estimate of the ability should result by evaluating the errors. The procedure we applied in estimating the deviation of force actually developed from the target ramp took into account two quantities, the SEE and the PES. The linearity of the force increment is measured by the SEE of the linear regression curve of the actual normalized data of force compared to their linear regression curve. The PES of the actual normalized data of the ramp is measured calculating the difference between the regression curve slope (of the subject’s actual ramp) and the slope of the target ramp (33.3% MVC per second) expressing this difference as a percentage error. When considering both parameters together (SEE and the PES), all our subjects but three performed the task better with one arm than with the contra-lateral. One subject presented the same result for both parameters in both arms. Therefore, in 6 subjects we did not have any problem in judging the BPA in this specific task, because of the results of SEE and PES being consistent. In two subjects the two errors were differently distributed in the two arms (i.e., one

arm presented the lower SEE, while the other arm had the lower PES). In this case we chose the arm where the force was increased with the lower SEE as better performing arm (BPA). This decision was made on the basis of a previous study [3] in which SEE resulted as the most reliable parameter in demonstrating an improvement after specific training in tracking the ramp. It was very surprising to find that the BPA was not always the right one in subjects who declared themselves to be right-handed. The very relevant difference in the EMG tracings observed between the BPA and the other arm named worse performing arm (WPA) is the different level (percentage) of MVC at which the MF reached its maximum value. Following the Solomomow et al. interpretation [15], the maximum MF indicated the full MU recruitment point. These authors showed that motor unit recruitment has the major impact on the shift of the EMG power density spectrum towards the highest frequencies. The MF rises linearly when the motor units are recruited in an orderly manner. The changes in the firing rate do not affect MF and when MU recruitment has been completed the MF either stays stable or decreases. The main possible limits of the application of these procedures to detect MU recruitment in humans are MU synchronization and fatigue development [2]. The 3 s duration of the ramp was selected to avoid the effects of these two events [3,4,14]. The results of the present study show that the BPA was the arm in which the MU recruitment was completed later, that is at a higher level of force (higher %MVC). In other words, the BPA is the arm in which the MU recruitment is more uniformly used to increase the force. In previous work [3] some of the present authors observed that the skill acquisition of a motor task, similar to this research, was accompanied by a modification of the motor unit recruitment strategy. This modification consisted of a more uniform recruitment (full MU

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

recruitment at a higher percentage of the MVC) after a training protocol during which the subjects practised the same task numerous times a week. In discussing those results the authors had hypothesized that skill acquisition, that is a more precise control of force generation, was dependent on the fact that a smaller number of motor units were recruited at lower force levels. In controlling the force output of a single muscle the nervous system has only two possibilities: the recruitment and the firing frequency of each MU. In obedience to Henneman’s size principle the force will increase exponentially with respect to the recruitment. Therefore, in order to obtain a linear increase in force versus time the recruitment rate must be slower as the force becomes greater. The other determinant of the force output, the firing frequency, might contribute to smoothing the force increase, in the interval between the recruitment of “one” motor unit and the “next”. It could be interesting to discuss the fact that the level of full MU recruitment was slightly different comparing the results of the present research with those of the previous one [3]. This could be explained by the fact that in the present research only the cases in which the subject performed a good ramp were included in the evaluation of the EMG record, while the worse ramps (the ramps in which our strict criteria were not accomplished) were rejected for each arm. In this way, the subjects practised the motor performance during the test, repeating about ten ramps for each arm. The results of the SEE of the two studies are not comparable because, for the wide range of maximal force of the subjects included in the present research, the SEE analysis was performed on the normalized data. In the previous data this analysis was performed on the actual force data because there was no need to compare subjects and because there were no criteria to reject the ramps. The theory regarding the relationship between recruitment and linearly increasing force generation described above is our interpretation of the past and the present results. Obviously other mechanisms could be involved in explaining the present results. We can suppose, for example, that a different fiber composition could be present in the biceps of the two arms, as shown in other muscles by other authors [9]. A distribution of the motor units of the various types, more uniform in one arm than in the other, could play a role in the recruitment strategy. The force contributors at the elbow joint are two agonist muscles, biceps and brachialis and one antagonist, the triceps. We must exclude that another mechanism, different from recruitment and firing regulation of the quoted muscles, can play a role in controlling the force output. According to Solomonow et al. [15] we recorded the EMG from the triceps considering the RMS as the muscle activation index that includes both the motor unit recruitment and their increase in firing frequency. This parameter did not show any statistical difference

129

between the two arms, either when grouping the results according to dominance or according to the better and worse performing arm. According to the MF results, also the recruitment showed the same pattern of the RMS. No difference was observed either by grouping for side (left or right) or by the BPA. In particular, the EMG records show that in both arms the triceps is active practically up to the maximal force exerted by the biceps and the recruitment of the involved MUs occurs up to a range of force in average between 65 and 71% flexion MVC. Therefore, we assume that the differences observed in the BPA versus the contra-lateral are not due to a different use of the antagonist muscle. It must be stressed that in the case of the triceps the highest value of MF does not indicate the full MU recruitment, but the level of force where all the MU necessary for that motor task are recruited. The roles of the antagonist in assuring the joint stability, regulating stiffness and reducing the laxity [1] seem to be performed in a similar way in both arms. We can conclude that the control strategy of the activation of the antagonist muscle cannot be taken into account to explain the result in the ability to increase the force. On the basis of these observations we think that the data presented provide consistent evidence that a more precise and accurate control of the force increment is obtained when the central nervous system selects a slower and prolonged recruitment of motor units to perform the planned motor task. Further studies on this topic could be useful to find other parameters which can affect the motor unit recruitment strategy.

References [1] Baratta R, Solomonow M, Zhou BH, Letson D, Chuinard R, D’Ambrosia R. Muscular coactivation. The role of the antagonist musculature in maintaining knee stability. Am J Sports Med 1988;16:113–22. [2] Bernardi M, Solomonow M, Sanchez JH, Baratta RV, Nguyen G. Motor unit recruitment strategy of knee antagonist muscles in a step-wise, increasing isometric contraction. Eur J Appl Physiol 1995;70:493–501. [3] Bernardi M, Solomonow M, Nguyen G, Smith A, Baratta R. Motor unit recruitment strategy changes with skill acquisition. Eur J Appl Physiol 1996;74:52–9. [4] Bernardi M, Solomonow M, Baratta RV. Motor unit recruitment strategy of antagonist muscle pair during linearly increasing contraction. Electromyogr clin Neurophysiol 1997;37:3–12. [5] De Luca CJ, Sabbahi MA, Roy SH. Median frequency of the myoelectric signal effects of hand dominance. Eur J Appl Physiol 1986;55:457–64. [6] De Luca CJ, Basmajan JV. Muscles alive, their functions revealed by electromyography, 5th ed. Baltimore: Williams and Wilkins 1985. [7] Felici F, Colace L, Sbriccoli P. Surface EMG modifications after eccentric exercise. J Electromyogr Kinesiol 1997;7:193–202. [8] Fox EL, Bowers RW, Foss ML. The physiological basis of physical education, 4th ed. Dubuque: Wm C. Brown Publishers, 1989. [9] Fugl-Meyer AR, Eriksson A, Sjo¨stro¨m M, So¨derstro¨m G. Is mus-

130

[10]

[11] [12] [13] [14]

[15]

[16]

M. Bernardi et al. / Journal of Electromyography and Kinesiology 9 (1999) 121–130

cle structure influenced by genetical or functional factors? Acta Physiol Scand 1982;114:277–81. Jones DA, Round JM. Histochemistry and contractility. In: Skeletal muscle in health and disease. Manchester and New York: Manchester University Press, 1995:68. Henneman E, Somjen G, Carpenter D. Functional significance of cell size in spinal motoneurons. J Neurophysiol 1965;28:560–80. McArdle WD, Katch FI, Katch VL. Essentials of exercise physiology. Malvern: Lea and Febiger, 1994. Sachs L. Applied statistics. New York, Berlin, Heidelberg, Tokio: Springer-Verlag, 1982. Sanchez JH, Solomonow M, Baratta RV, D’Ambrosia R. Control strategies of the elbow antagonist muscle pair during two types of increasing isometric contractions. J Electromyogr Kinesiol 1993;3(1):33–40. Solomonow M, Baten C, Smit J, Baratta RV, Hermens H, D’Ambrosia R, Shoji H. Electromyogram power spectra frequencies associated with motor unit recruitment strategies. J Appl Physiol 1990;68(3):1177–85. Solomonow M, Baratta R, Bernardi M, Zhou BH, Lu Y, Zhu M, Acierno S. Surface and wire EMG crosstalk in neighbouring muscles. J Electromyogr Kinesiol 1994;4(3):131–42.

Marco Bernardi was born in Rome (Italy), on 23 April 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 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 Sports Science, the American College of Sports Medicine and the International Society for Electrophysiology and Kinesiology. He is a scientific consultant for the Rehabilitation Hospital IRCCS Santa Lucia (Rome, Italy). He is President of the Italian Medical Committee of the Italian Sports Federation for the Disabled. Francesco Felici was born in 1957. He graduated in Medicine in 1982. From 1983 to the present he was appointed as researcher with the University of Rome “La Sapienza”. He co-authored many papers on exercise physiology and motor control. He is also an external consultant of the department of Physiology and Biomechanics of the Italian National Olympic Committee (CONI).

Marco Marchetti was born in Carrara (Italy) on 10 February 1931. He received his Medical Doctor Degree from the University of Rome “La Sapienza” in 1955. Since 1955 he has worked, first as a Medical student and then as Assistant, at the Institute of Human Physiology of the same University, where he currently is Full Professor and Director. He is also Professor and Director of the Specialization School of Sports Medicine of the University of Rome “La Sapienza”. His main areas of scientific interest are Sport Physiology both in healthy (particularly in

sailing) and in physically handicapped subjects, Biomechanics of walking both in normal subjects and disabled patients and Muscle Physiology. He is a scientific consultant for the Rehabilitation Hospital IRCCS Santa Lucia (Rome, Italy). He is a member of the Italian Medical Committee of the Italian Federation of Sailing (FIV) and International Society for Electrophysiology and Kinesiology. Fabio Montellanico was born in Roccasecca (Frosinone, Italy) on 19 April 1971. He received his Medical Doctor Degree from the University of Rome “La Sapienza” in 1997. His major research interests are neuromuscular and cardiovascular adaptations to exercise and functional evaluation of able-bodied and disabled athletes.

Maria Francesca Piacentini was born in Rome on 7 September 1971. In December 1992 she graduated “cum laude” at the ISEF of Rome, the Superior Institute of Physical Education. In May 1996 she completed a two-year Master Degree in Exercise Physiology at the University of California at Berkeley, USA, under the supervision of Professor G.A. Brooks. Her main interests in research are metabolic adaptations to exercise and training (effects of training on the female sex-steroid hormones and substrate utilization), neuromuscular adaptations and strength training, and metabolic adaptations in paraplegic athletes. M. Solomonow received the B.Sc., and M.Sc. degrees in Engineering from California State University, Los Angeles, and the Ph.D. degree in Engineering and Neuroscience from University of California, Los Angeles. He is Professor and Director of Bioengineering, in the Department of Orthopedic Surgery, Louisiana State University, New Orleans, where he has been since 1983, following a faculty appointment at University of California, Los Angeles and Tulane University. He is a consultant to the National Science Foundation, National Institutes of Health, Veterans Administration, and various industrial firms, as well as several European, Asiatic and Canadian scientific agencies, serves on the Editorial Board of several scientific journals and is the Founding Editor-in-Chief of the Journal of Electromyography and Kinesiology. His research interests focus on basic and applied kinesiology in health and disease and in development of technology for rehabilitation of musculoskeletal deficits. Dr. Solomonow was awarded the I. Cahen, M.D. Professorship in Orthopedic Bioengineering (1997), the Doctor Honoris Causa (Brussels, 1997), the Distinguished Contribution in Orthopedics Award (Paris, 1990), and the R. Crump Award for Excellence in Medical Engineering Research (UCLA, 1977).