| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Motor Research Group, Institute of Pathophysiology, Friedrich-Schiller-University, D-07740 Jena, Germany; 2 Department of Clinical Neurophysiology, Institute of Neurology, University Medical Centre, NL-6500HB Nijmegen, The Netherlands; 3 Department of Neurology, University of Tübingen, D-72076 Tübingen, Germany
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
The frequency content of the surface electromyography (SEMG) signal, expressed as median frequency (MF), is often assumed to reflect the decline of muscle fiber conduction velocity in fatigue. MF also decreases when motor unit firings synchronize, and we hypothesized that this effect can explain the electrode-dependent pattern in our previous recordings from the trapezius muscle. An existing motoneuron (MN) model describes the afterhyperpolarization following a spike as an exponential function on which membrane noise is superimposed. Splitting the noise into a common and an individual component extended the model to a MN pool with a tunable level of firing synchrony. An analytical volume conduction model was used to generate motor unit action potentials to simulate SEMG. A realistic level of synchrony decreased the MF of the simulated bipolar SEMG by ~30% midway between endplate position and tendon but not above the endplate. This is in accordance with experimental data from the biceps brachii muscle. It was concluded that the pattern of decrease of MF during sustained contractions indeed reflects MN synchronization.
surface electromyography; short-term synchrony; fatigue; electromyogram modeling; biceps brachii muscle
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
SURFACE ELECTROMYOGRAPHY (SEMG) is often used to study muscle electrophysiology and central motor drive in fatiguing exercises (10, 43). A fatigued motor unit (MU) generates less force per motoneuron (MN) firing event. This decline must be compensated for by higher firing rates and/or recruitment of other MUs. Therefore, the amplitude (root mean square, RMS) of the SEMG will increase with fatigue if the force is kept constant. With fatigue at high isometric forces, the concentration of extracellular ions changes (potassium in particular; Ref. 39), and the muscle fiber conduction velocity decreases. This results in a compression of the spectrum of the SEMG and a reduction of its median frequency (MF) (27, 45). The increased spatial length of the intracellular action potential with decreasing velocity will lead to increased amplitudes. Although this is partially counteracted by an increased spatial dispersion, it could contribute to the increasing RMS with fatigue (42). Although the MF depends on the position of the recording electrode relative to the fibers (22, 36), conduction velocity and MF decrease by almost the same percentage in electrically evoked responses (26), a relationship that is confirmed by simulation studies (42). However, in voluntary contractions, the MF drops more than the muscle fiber conduction velocity (11, 21, 31). In the biceps muscle, a decrease of MF is also found at forces as low as 15% of maximum voluntary contraction (MVC), although the conduction velocity is unchanged during fatigue (20). These authors concluded that changes in the firing pattern, especially synchronization, of the MNs must be responsible for the spectral shift to lower frequencies. Short-term synchrony between two MNs, i.e., more firings at (almost) the same time than expected by chance, is caused by synaptic input that is common to both MNs. The synchronizing input could be assumed to be common to all MNs of a pool. Corticospinal cells have widespread projections to the pool of MNs supplying a muscle (29), and short-term synchrony is absent in patients after a contralateral stroke (7, 8). Synchrony also increases with force (16) and in an attention-demanding task (37). To overcome fatigue, the central drive to a muscle has to increase, and as fatigue progresses more synchrony due to increased corticospinal activity could be expected. Furthermore, during fatigue the input that a MN receives from muscle afferents decreases (9), requiring even more supraspinal input.
An increase of firing synchrony with fatigue has not been directly proven because of several methodological limitations. Measurement of synchrony requires the stable and accurate recording of single MU activities. At low forces, the recording could be done with high-density SEMG (18). However, this method cannot be applied with high forces to decompose the complex interference pattern. Needle electrodes can be employed to record MU firing patterns during fatigue, but they tend to get out of place with long and forceful contractions.
Indirect evidence for synchronization comes from simulations of MN firing patterns and the corresponding SEMG (12, 44). In these simulations, synchrony was induced by shifting a certain percentage of the discharges to be simultaneous with a reference MN. Although these models allow useful predictions, they do not attempt to describe the physiological mechanism of short-term synchrony.
An SEMG model connecting a pool of MNs to MU action potentials (MUAPs) was developed to predict the effect of synchrony on MF at different electrode positions. The present study extended the model MN of Matthews (23) to synchronize the firings of a pool. Matthews described the membrane potential of a MN after the discharge of an action potential by an exponential function. The spike phase and the early part of afterhyperpolarization (AHP) are not considered, because no new spikes can be generated in this refractory phase, and the course of the early part of the AHP is not influencing the occurrence of the next spike. During a steady isometric contraction, each spinal MN in the active pool receives an inflow of predominantly excitatory postsynaptic currents. The finite number of neurons innervating the MN and their discrete discharges produce a fluctuation of the net input. Matthews described the depolarization caused by the constant inflow as an offset of the membrane potential on which fluctuations were superimposed. The fluctuations were described as white noise that was filtered by using the time constant of the membrane. This approach separates the net excitatory drive from its variability, although they cannot be separated experimentally.
The synaptic input to a particular MN comes from many different sources that could have different distributions within the MN pool. One extreme would be input that is unique for a particular MN and does not reach the other MNs. The other extreme corresponds to a branching of the axons to form an evenly distributed projection to the whole MN pool. These two sorts of input lead to the extension of Matthews' (23) model in the present study: Synchrony was induced by splitting the noisy input into a "common" and an "individual" component. The common noise represents the input that is shared by all MNs of the pool. Peaks in the common noise discharge all MNs of the pool that are close to threshold and therefore tend to synchronize the firings. The individual component of noise varies between MNs and represents the input that is unique to a particular MN. The firing sequence of each MN of the pool of 300 MNs was used to generate a MUAP train in which the specific MUAP was generated with an analytical volume conduction model (1) that accurately predicts MUAPs of the biceps muscle (32). The MUAP trains of all MUs were added to simulate a SEMG signal.
In a recent study of the trapezius muscle (19), our laboratory described the topography of RMS and MF during fatiguing shoulder elevation with 50% MVC. Apart from the expected dependence of MF on electrode position, the normalized change of MF was position dependent as well. Above the endplate zone, the monopolar MF decreased more than it did closer to the fiber end. By contrast, the bipolar MF fell less at the endplate but decreased more midway between endplate and muscle fiber end.
It was hypothesized that synchronization, simulated by increasing the percentage of common input to the MN pool, can explain the electrode-dependent changes of SEMG characteristics during fatigue.
To confirm the above findings, a multichannel SEMG study on the biceps muscle contracting at 20% MVC was conducted. This particular muscle was selected because the volume conduction model was constructed and validated for this muscle (32). A moderate force level was measured to minimize the effect of conduction slowing. Central fatigue, which was expected, should synchronize firing.
| |
METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
SEMG recordings. Ten healthy volunteers (5 female, 5 male), mean age 24 yr (range 22-28), gave their informed consent to participate in the study. The subjects sat in front of a table. The position of the arm was such that the angle in the elbow joint was 90°. In this way, isometric elbow flexion force was measured by a wrist belt attached to a strain gauge sensor. The forearm was supinated to reduce the influence of the brachioradialis muscle. Three trials of MVC were performed, and the best attempt was accepted as MVC if the force differed by <10% between trials. After application of SEMG electrodes, the subject sustained a force of 20% MVC for 15 min. All subjects were able to maintain the target force for that period.
Thirty-two channels of SEMG were recorded from a rectangular 4 × 8 electrode grid with an interelectrode distance (IED) of 1 cm. The Ag/AgCl electrodes (diameter 5 mm) were placed in an elastic cuff around the upper arm. Care was taken to align the columns of eight electrodes parallel to the muscle belly. The distal electrode row was placed at the transition between muscle and tendon. With the elbow at 90°, the electrode grid covers about the distal two-thirds of the muscle belly. The endplate zone is therefore in the proximal third of the grid, as was confirmed by the resulting RMS profiles. The monopolar signals (referenced to the olecranon) were amplified, filtered at 10-700 Hz, and analog-to-digital converted with a resolution of 2.44 µV/bit (range 12 bit) and a rate of 3,906 s
1 · channel1 (Biovision,
Wehrheim, Germany). The signals were stored on hard disk for
subsequent off-line processing.
SEMG processing.
The SEMG processing steps were essentially the same as in a previous
experiment in the trapezius muscle (19). SEMG intervals of
1,048 ms (4,096 samples) were selected every 30 s during the measurement of 15 min. For each channel the calculation included the
following steps. After baseline shifts were removed by
subtracting the signal mean and digital low-pass filtering (400 Hz,
second-order Butterworth), the mean amplitude (RMS) of each interval
was calculated. After the linear trend was removed from the signal, the
power spectrum was estimated by using the fast Fourier transform
algorithm. The MF was calculated. By use of a linear regression on all
the data points, the intercept of the RMS (RMS0) and the
RMS after 15 min of fatiguing exercise at 20% MVC (RMSF)
were calculated. The normalized change of RMS was then calculated as
(RMSF 
RMS0)/RMS0. Similarly, MF0, MFF, and the normalized change
of MF [(MFF MF0)/MF0] were calculated from linear regression. Bipolar montages were constructed by subtracting the signals recorded monopolarly from adjacent electrodes in the fiber direction (distal-proximal). In
addition to bipolar montages with an IED of 1 cm, montages with an IED
of 2 cm were calculated by skipping one electrode. The same procedure
as above was applied to the bipolar SEMG signals. RMS, MF, and their
normalized change were averaged across subjects for each monopolar or
bipolar channel. The standard error of mean was calculated.
Simulation of motoneuron firing behavior.
In the MN model introduced by Matthews (23), the time
course of the membrane potential contains an AHP starting after each spike and approaching the threshold exponentially (Fig.
1, gray line). Noise is superimposed on
this average membrane potential trajectory. White noise with a sampling
interval of 1 ms (Fig. 1, top traces) is generated from a
Gaussian distribution with zero mean and a standard deviation (SD) of
one and is fed into an impulse response of a simple first-order
passive low-pass filter with a time constant of 4 ms, the
latter to simulate the decay of postsynaptic potentials. After such
filtering, performed as described by Matthews, the high-frequency
content of signal decreased (Fig. 1, third trace) and the SD
increased to 1.58.
|
20 NU and decaying with a time
constant of 30 ms was used (Fig. 1, bottom trace). These
values were found to describe the firing pattern of biceps MUs
accurately (23). The net input to a MN is introduced by shifting the whole membrane trajectory by a constant voltage. The
firing rate of MN was calculated as 1/(mean interspike interval). As
done by Matthews, intervals longer than 300 ms were excluded because
slow and irregular firing can be a problem in experimental recordings.
The relation between input voltage (in NU) and firing rate as estimated
from simulated spike trains of 30 s duration is given in Fig.
2A. The firing rate cannot be
lower than 3.3 Hz and was set to zero when no interspike intervals
shorter than 300 ms were present. This causes a sudden increase of the
firing rate in Fig. 2A. A negative input voltage indicates
that the average AHP will not reach the threshold. Firings can still be
induced by noise peaks. For the estimation of the topographical pattern of SEMG characteristics, the model was run with mean inputs of 6 and 10 NU. An input of 6 NU induces a firing rate of 14.8 Hz, which is within
normal values for the biceps contracting with 10 and 30% MVC
(5). Between the MNs of the pool, the input was varied
randomly according to a Gaussian distribution with a SD of 0 (no
variation), 1, or 2 NU but was constant within each MN over the
simulated time. Adding input voltage to AHP of the model MN is equal to
lowering the threshold. From a physiological point of view, the
distribution of input voltages simulates a MN pool with different
thresholds.
|
Synchrony between motoneuron firings. Short-term synchrony is generated by synaptic input that is common to a pair or a group of MNs (8). The above model MN was therefore slightly modified to allow for two sorts of noisy input. One part of the membrane noise is common to all MNs of the pool (simulating branched input), and another part was unique for each MN. A weighted average of the individual noise and the common noise was calculated with the required percentage of common input as weighting factor. Thus, for the condition "25% common input" to the MN pool, the noise signal was composed of 25% of the common noise signal and 75% of the individual noise signal for each MN. The combined noise was then filtered with a time constant of 4 ms as in the nonsynchronization case. The SD of the filtered noise was kept at 1 NU.
The amount of synchronous firing of a pair of MNs can be estimated from their cross-correlation histogram (8). Short-term synchrony refers to a higher number of nearly simultaneous firings than expected by chance. Nordstrom et al. (28) define the common input strength (CIS) as the number of extra simultaneous firings per second, i.e., the area of the peak in the cross-correlation histogram above baseline, normalized to the duration of the recording. The baseline was estimated as the mean number of spikes within a latency window of 10-50 ms before the reference spike (37). The number of extra simultaneous spikes was the total number of spikes within 10 ms before or after the reference spike minus the number of spikes expected according to the baseline count. CIS is not dependent on firing rate and is therefore superior to normalization to the number of firings (28). Because CIS can be estimated for any pair of spike trains and the percentage of common input is a property of the particular MN-pool model presented here, the relationship between these measures is not straightforward. The dependence of CIS on the percentage of common noisy input is given in Fig. 2B. At low forces, a CIS of ~1 s1 is
expected, but it can increase to 2.5 or 3.8 s1 at forces
of 50 and 100% MVC, respectively (16). CIS can only be
used to measure the synchrony between pairs of MNs. To measure synchrony in a MN pool, Yao et al. (44) introduced the
population synchrony index (PSI). First, the number of coincident
impulses, i.e., how often two, three, or more MNs fire within the same
millisecond, is calculated and weighted with a binomial factor. Then,
the PSI is calculated by subtracting the number of coincident impulses expected due to chance with normalization to this number. Therefore, PSI provides a normalized value of the total number of coincident impulses for all MNs in excess of that expected due to chance. (For
more details on the derivation of PSI, see Ref. 44.) The PSI for a pool of 300 MUs is given in Fig. 2C. In their
simulation, Yao et al. considered a PSI of 1 as moderate synchrony and
a PSI of 3 as high synchrony. An example of a firing pattern of 50 MNs receiving an input of 6 NU without and with synchrony induced by 50%
common noise is given in Fig. 2, D and E, respectively.
Simulation of MUAPs. The analytical volume conduction model of Blok (1) was used to simulate the potential distribution generated by an MU at the skin. In this model, the volume conductor is described by three cylindrical layers representing an inner muscle compartment (radius 3.6 cm), a subcutaneous fat layer (2 mm), and a skin layer (2 mm). The length of the cylinder was set to 25.6 cm. The conductivity of skin was 0.75 S/m and that of fat was 0.05 S/m. The radial conductivity of the anisotropic muscle tissue was 0.1 S/m, and the axial conductivity was 0.5 S/m.
The bioelectric source was simulated as a muscle fiber aligned parallel to the axis of the cylinder. The intracellular action potential was calculated according to Rosenfalck (35). It propagated with a velocity of 4 m/s toward both tendons. With histological methods, the average innervation ratio of the biceps muscle was estimated to be 750 muscle fibers per unit (6). Because of the skewed distribution of MU sizes within a muscle, most MUs are smaller than average (17) and all MUs were simulated as if consisting of 500 muscle fibers. The motor endplate of each fiber was at the axial middle of the cylinder and was allowed to vary with an SD of 3 mm between fibers. The muscle fibers had a length of 10 cm (5 cm in both directions from the endplate), and the position of the fiber end at the tendon varied with an SD of 4 mm. The onset of the action potential varied by an SD of 0.2 ms between fibers. The MUAP output of the volume conduction model had a sampling rate of 2,000 s1. The model parameters used here are taken from
Roeleveld et al. (32) or from Blok (1).
The model calculated MUAPs for MU depths of 6-25 mm below cylinder
surface in steps of 1 mm. The MUAP was stored with a spatial resolution
of 1 mm in the longitudinal direction and ~2 mm along the
circumference of the cylinder. To simulate the whole muscle, 300 MUs
were randomly positioned within the muscle compartment at a depth of at
least 6 mm from the surface and not deeper than 2.5 cm from the
electrode position (Fig. 3). To simulate
some physiological variability in the position of endplate regions and
muscle fiber ends, the MUAPs were also randomly shifted longitudinally with an SD of 5 mm. The MUAP at the recording electrode was taken from
the model run with the appropriate depth and relative electrode position along the circumference. An electrode diameter of 5 mm was
simulated by averaging all simulated positions closer than 2.5 mm. For
each of the 300 MUs, the MUAP at positions between 7 cm and 7 cm
(resolution 2 mm) from the middle of the cylinder (the average endplate
position) was stored. Similarly, bipolar MUAPs were generated by
subtracting the monopolar MUAPs from positions with an IED of 1 or 2 cm.
|
Simulated surface EMG interference pattern. From the firing pattern of 300 MNs and their corresponding 300 MUAPs, an SEMG interference pattern was generated by adding the MUAPs at the appropriate points in time generated by the MN model. With histological methods, the number of MUs in the biceps muscle was estimated to be 774 (6). This is in sharp contrast with a MU number estimate of 109 when using electrophysiological methods (25). The difference could be explained by the fact that the MUAPs from deep and distant MUs are small and could be missed during the incremental stimulation. Therefore, the electrophysiological method is more likely to underestimate the true number of MUs. Because not all MUs will be recruited at 20% MVC, a number of 300 MUs was used in the model. The number of MUs was not a critical parameter because a model-run with 100 MUs showed almost equal results. To estimate the MF and the RMS, 15 s of SEMG were generated. After 4 s were skipped from the beginning, 10 epochs of 2,048 samples were used to calculate MF and RMS, as was done for the measured data. The means of MF and RMS were plotted as a function of recording electrode position. The influence of synchrony was analyzed by plotting the difference in MF between the synchrony and the no-synchrony conditions normalized to the no-synchrony condition (normalized change). For the accurate calculation of the power spectrum at some selected electrode positions, 60 s of SEMG were generated and the power spectrum was estimated by averaging over 50 epochs with 2,048 samples.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Experiment.
The RMS of the biceps muscle SEMG recorded in an elbow flexion of 20%
MVC is given in Fig. 4. The highest
monopolar RMS was measured just proximal from the middle part of each
electrode row (Fig. 4A). The bipolar RMS was relatively low
in that region (Fig. 4, B and C). The region of
low bipolar RMS corresponds to the endplate position of the MUs, where
action potentials, propagating in opposite directions, partially
cancel. For the medial electrode rows, the endplate region is less
clearly distinguished. During 15 min of fatiguing contraction, the RMS
increased at all electrode positions. The monopolar RMS increased by
87-128%, whereas the bipolar RMS increased even by ~200% at
some electrodes.
|
|
Simulation.
With a common input of 50%, a CIS in the physiological range was
obtained (Fig. 2B). By using these model parameters, a PSI of ~3 was reached, a value that was considered to be high synchrony by Yao et al. (44). When the firing patterns in Fig. 2,
D and E, are compared, the synchrony of firing is
obvious. Synchronization increased the amplitude substantially (Fig.
6, A, B,
D, and E). In the power spectrum, synchronous
firing increased the power at low frequencies (Fig. 6, F,
G, I, and J). When the range of firing
rates was narrow (the same threshold for all MNs), there was a
prominent peak in the spectrum at the firing rate and its first
harmonic (Fig. 6J). The increase in the low frequencies depends on electrode position and montage. It is almost absent for
bipolar signals from the endplate zone (Fig. 6, C and
H).
|
|
|
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
An analytical volume conduction model (1) was connected to a MN model (23) to simulate the SEMG pattern over a whole muscle. The firings of the MN pool were partially synchronized by noisy input that was common to all MNs. The profile of RMS and MF that was predicted by the simulations was in accordance with experimental results from the trapezius muscle (19). In the biceps muscle, the highest bipolar MF was found above the endplate region (Fig. 5, B and C), a finding that also confirms results from vastus lateralis and tibialis anterior muscles (15, 36).
In the trapezius muscle (19) the pattern of monopolar MF was qualitatively inverse to that of the bipolar MF, which is similar to the model output presented here. The monopolar MF measured in the biceps muscle, however, does not exhibit any clear-cut pattern. No relation to the endplate zone, as defined by a low bipolar RMS, was found. This unexpected finding could be explained by electrical activity from deep or distant sources that only affect the monopolar signal. Whereas bipolar signals are mainly influenced by sources close to the electrodes, monopolar signals reflect deep sources and far-field potentials as well (41). Roeleveld et al. (33) compared the monopolar and bipolar RMS of the biceps muscle. At low to moderate forces the highest monopolar RMS was more distal than the lowest bipolar RMS. This finding was explained by a more distal position of the motor endplates of deep MUs. In the simulation presented here, the endplate position was allowed to vary between MUs, but not systematically with depth. However, with a difference of the endplate position of ~2 cm, which would be consistent with the data of Roeleveld et al. (33) and Scholle et al. (38), still some systematic variation of the MF had to be expected. Another possible explanation could be activity from other muscles than the biceps. Agonists of the biceps in elbow flexion are the brachialis and brachioradialis muscles. Although measurements were performed in a supinated hand position to minimize the contribution of the brachioradialis, this cannot be excluded. The proximal muscle-tendon transition of the brachioradialis is between the electrode grid and the reference electrode. The extinction of propagating action potentials at the muscle fiber end will generate far-field potentials (34) that are almost equal at all electrodes of the present montage. Synchrony is not restricted to MUs of the same muscle. MUs of agonistic muscles tend to fire in synchrony as well (30). Far-field potentials from brachioradialis MUs synchronized with other brachioradialis MUs and also with biceps MUs could therefore contribute to the strong and topographically uniform decrease of monopolar MF.
Synchronization produced a substantial increase of RMS for both the monopolar and the bipolar signal. Synchrony aligns the MUAPs, and they will add to produce high amplitudes. The increase in RMS that is found in fatigue has therefore been attributed not only to an increased firing rate and/or recruitment, but also to synchronization.
In the fatiguing contraction at 20% MVC, the change of MF depends on electrode position. With fatigue at high forces, the muscle fiber conduction velocity decreases and MF declines. When MU firings are synchronized, the bipolar MF is unaffected close to the endplate zone but decreased by ~30% midway between endplate and tendon. To achieve this in the simulations, only one model parameter, the percentage of common noisy input to the MN pool, had to be changed. In a fatiguing exercise, new MUs are recruited and the firing rate changes. The recruitment of new MUs is not expected to change the pattern, as long as the average endplate positions and fiber end positions are the same. This assumption appears to be true for the trapezius muscle but not for the biceps muscle (33, 38). As discussed above, the unequal distribution of endplate zones in the biceps muscle could contribute to the unexpected pattern of changes of monopolar MF.
In the present simulations, the firing rate between the 300 MNs of the pool could be varied. When the firing frequency of all MNs was set to the same value, there was a sharp peak in the spectrum at this frequency (Fig. 6J, black curve). As can be expected, this peak was even more prominent with synchronous firing (Fig. 6J, gray curve). With a large firing rate variation between MNs, the peak in the spectrum becomes smaller and broader, but the amount of variation did not influence the spatial distribution of MF over the simulated muscle. The dependency of MF on electrode position and synchrony remains. In a fatiguing contraction at moderate force the MU firing rate increases. To exclude the possibility that this influences the topographical distribution of MF, the model was run with an input of 10 NU and common inputs of 0 and 50%. When compared with an input of 6 NU, the increased firing rate had almost no influence on the MF and the change in MF induced by 50% common input. Taken together, changes in firing rate and firing rate distribution across MNs have no influence on MF.
Under physiological conditions, firing rate modulation and recruitment are linked. Because of the higher muscle fiber conduction velocity of high-threshold MUs, MF increases as force rises from low to moderate levels. Another change that could occur with fatigue is a derecruitment of fatigued MUs with simultaneous recruitment of fresh MUs ("alternation" or "rotation"). This possibility cannot be excluded for fatigue at 20% MVC. However, in the previous experiment on the trapezius muscle (19) the force was 50% MVC and a substantial role of recruitment is excluded when almost all MUs are active already from the beginning.
The same pattern was found for bipolar MF in the fatigued biceps muscle and in the trapezius muscle. These data suggest that in a fatiguing contraction the slower conduction decreases the MF at all electrode positions, whereas synchronization decreases MF dependent on electrode position. Both effects together can readily explain the experimental findings. With respect to electrode locations in bipolar kinesiological SEMG, it can be recommended from these results as well to place electrodes midway between endplate region and muscle-tendon transition (13, 14), where the influence of synchrony on the spectrum is strongest. In situations in which a separation between influences from muscle fiber conduction velocity and from synchrony is wanted, multichannel recordings are expected to give relevant clues.
In the present model, synchrony was induced by one common input signal to the whole MN pool, whereas the individual noisy input was unique to each MN. An assumption that seems more realistic would be the existence of many parallel input channels that send branches of their axons to a subgroup of MNs. Each MN then receives a weighted average of all input channels, the weighting factors representing the relative strength of the connections. Despite the simplified assumptions, the model presented here can predict the experimental data. The concept of one common input can be supported by experimental data. The corticospinal projection seems to be the major source of short-term synchrony, because in patients with a relatively selective lesion of the corticospinal tract after a stroke, short-term synchrony of MN firing is reduced or absent (7, 8). Recordings from corticospinal neurons in monkeys show that these neurons innervate a large part or the whole MN pool (29). A set of corticospinal neurons that all send projections to all MNs is well described by one channel of common input. Most studies of short-term synchrony recorded from hand muscles. A lower but still significant degree of synchrony was also found in the biceps brachii muscle (7). Probably, MN firings are already weakly correlated from the beginning but synchronize progressively as the corticospinal drive increases to compensate for peripheral fatigue. As to the presence of an individual input that is unique to a MN, one could think of spinal interneurons that project more focally (although not likely to only one MN). Their high number and different weighting would produce an input that does not correlate (or only weakly) between MNs.
Although the input to the pool of MNs by itself was always white random noise, i.e., did not prefer low frequencies, with short-term synchrony the power of SEMG increased at frequencies around the firing rate. When motor cortex activity is recorded simultaneously with either SEMG or single MU activity, significant corticomuscular coherence is found at specific frequencies, indicating that the inflow to the MN pool is modulated by cortical rhythms (2, 3, 24). The common drive behavior of MUs described by DeLuca et al. (4) appears to be another rhythmic input to the MN pool that is unrelated to short-term synchrony (40). In general, a change in corticomuscular coherence with the motor task is attributed to a changing descending motor command. The present study suggests that processes at both spinal and muscular levels might influence coherence spectra as well. Therefore, these coherence studies should carefully take electrode placement into account.
In conclusion, the model simulation of the SEMG interference pattern correctly predicted the experimental findings in the trapezius muscle and the bipolar SEMG of the biceps muscle. The effect of synchrony on the normalized change of MF was found to depend on electrode position and montage. The unexpected pattern of monopolar MF in the biceps muscle could be explained by activity from the co-contracting brachioradialis muscle. The results confirm and stress the importance of the influence of synchronization of MN firing on the SEMG signal during fatigue.
| |
ACKNOWLEDGEMENTS |
|---|
We thank Joleen Blok for help with the volume conduction model. We also thank Hans-Christoph Scholle and Nikolaus-Peter Schumann for comments on a draft of the manuscript and Marcie Matthews for linguistic correction.
| |
FOOTNOTES |
|---|
This research was supported by Deutsche Forschungsgemeinschaft (DFG Innovationskolleg Bewegungssysteme, Projekt A2).
Address for reprint requests and other correspondence: B. U. Kleine, Dept. of Neurology, Univ. of Tübingen, Hoppe-Seyler-Str. 3, D-72076 Tübingen, Germany (E-mail: bert.kleine{at}uni-tuebingen.de).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 26 December 2000; accepted in final form 29 May 2001.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
1.
Blok, JH.
AnVoLCon: analytical volume conduction model.
In: European Recommendations for Surface Electromyography. Results of the SENIAM Project (CD-ROM), edited by Freriks B,
and Hermens HJ.. Enschede, The Netherlands: Roessingh Research and Development, 2000.
2.
Brown, P.
Cortical drives to human muscle: the Piper and related rhythms.
Prog Neurobiol
60:
97-108,
2000[Web of Science][Medline].
3.
Conway, BA,
Halliday DM,
Farmer SF,
Shahani U,
Maas P,
Weir AI,
and
Rosenberg JR.
Synchronization between motor cortex and spinal motoneuronal pool during the performance of a maintained motor task in man.
J Physiol (Lond)
489:
917-924,
1995
4.
DeLuca, CJ,
LeFever RS,
McCue MP,
and
Xenakis AP.
Control scheme governing concurrently active human motor units during voluntary contractions.
J Physiol (Lond)
329:
129-142,
1982
5.
Dorfman, LJ,
Howard JE,
and
McGill KC.
Motor unit firing rates and firing rate variability in the detection of neuromuscular disorders.
Electroencephalogr Clin Neurophysiol
73:
215-224,
1989[Web of Science][Medline].
6.
Enoka, RM.
Morphological features and activation patterns of motor units.
J Clin Neurophysiol
12:
538-559,
1995[Web of Science][Medline].
7.
Farmer, SF,
Bremner FD,
Halliday DM,
Rosenberg JR,
and
Stephens JA.
The frequency content of common synaptic inputs to motoneurones studied during voluntary isometric contraction in man.
J Physiol (Lond)
470:
127-55,
1993
8.
Farmer, SF,
Halliday DM,
Conway BA,
Stephens JA,
and
Rosenberg JR.
A review of recent applications of cross-correlation methodologies to human motor unit recording.
J Neurosci Methods
74:
175-187,
1997[Web of Science][Medline].
9.
Gandevia, SC.
Neural control in human muscle fatigue: changes in muscle afferents, motoneurones and motocortical drive.
Acta Physiol Scand
162:
275-283,
1998[Web of Science][Medline].
10.
Gerdle, B,
Karlsson S,
Day S,
and
Djupsjöbacka M.
Acquisition, processing and analysis of the surface electromyogram.
In: Modern Techniques in Neuroscience Research, edited by Windhorst U,
and Johansson H.. Berlin: Springer, 2000, p. 705-755.
11.
Hägg, GM.
Interpretation of EMG spectral alterations and alteration indexes at sustained contraction.
J Appl Physiol
73:
1211-1217,
1992
12.
Hermens, HJ,
Bruggen TAM,
Baten CTM,
Rutten WLC,
and
Boom HBK
The median frequency of the surface EMG power spectrum in relation to motor unit firing and action potential properties.
J Electromyogr Kinesiol
2:
15-25,
1992.
13.
Hermens, HJ,
Freriks B,
Disselhorst-Klug C,
and
Rau G.
Development of recommendations for SEMG sensors and sensor placement procedures.
J Electromyogr Kinesiol
10:
361-374,
2000[Web of Science][Medline].
14.
Hermens, HJ,
Freriks B,
Merletti R,
Stegeman DF,
Blok JH,
Rau G,
Disselhorst-Klug C,
and
Hägg GM.
European Recommendations for Surface Electromyography. Results of the SENIAM Project. Enschede, The Netherlands: Roessingh Research and Development, 1999.
15.
Hogrel, JY,
Duchene J,
and
Marini JF.
Variability of some SEMG parameter estimates with electrode location.
J Electromyogr Kinesiol
8:
305-315,
1998[Web of Science][Medline].
16.
Kamen, G,
and
Roy A.
Motor unit synchronization in young and elderly adults.
Eur J Appl Physiol
81:
403-410,
2000[Web of Science][Medline].
17.
Kernell, D.
Organized variability in the neuromuscular system: a survey of task-related adaptations.
Arch Ital Biol
130:
19-66,
1992[Web of Science][Medline].
18.
Kleine, BU,
Blok JH,
Oostenveld R,
Praamstra P,
and
Stegeman DF.
Magnetic stimulation induced modulations of motor unit firings extracted from multi-channel surface EMG.
Muscle Nerve
23:
1005-1015,
2000[Web of Science][Medline].
19.
Kleine, BU,
Schumann NP,
Stegeman DF,
and
Scholle HC.
Surface EMG mapping of the human trapezius muscle: the topography of monopolar and bipolar surface EMG amplitude and spectrum parameters at varied forces and in fatigue.
Clin Neurophysiol
111:
686-693,
2000[Web of Science][Medline].
20.
Krogh-Lund, C,
and
Jorgensen K.
Modification of myo-electric power spectrum in fatigue from 15% maximal voluntary contraction of human elbow flexor muscles, to limit of endurance: reflection of conduction velocity variation and/or centrally mediated mechanisms?
Eur J Appl Physiol
64:
359-370,
1992[Web of Science].
21.
Krogh-Lund, C,
and
Jorgensen K.
Myo-electric fatigue manifestations revisited: power spectrum, conduction velocity, and amplitude of human elbow flexor muscles during isolated and repetitive endurance contractions at 30% maximal voluntary contraction.
Eur J Appl Physiol
66:
161-173,
1993[Web of Science].
22.
Lateva, ZC,
Dimitrova NA,
and
Dimitrov GV.
Effect of recording electrode position along a muscle fibre on surface potential power spectrum.
J Electromyogr Kinesiol
3:
195-204,
1993.
23.
Matthews, PB.
Relationship of firing intervals of human motor units to the trajectory of post-spike after-hyperpolarization and synaptic noise.
J Physiol (Lond)
492:
597-628,
1996
24.
McAuley, JH,
and
Marsden CD.
Physiological and pathological tremors and rhythmic central motor control.
Brain
123:
1545-1567,
2000
25.
McComas, AJ.
Motor units: how many, how large, what kind?
J Electromyogr Kinesiol
8:
391-402,
1998[Web of Science][Medline].
26.
Merletti, R,
Knaflitz M,
and
De Luca CJ.
Myoelectric manifestations of fatigue in voluntary and electrically elicited contractions.
J Appl Physiol
69:
1810-1820,
1990
27.
Merletti, R,
and
Lo Conte LR.
Surface EMG signal processing during isometric contractions.
J Electromyogr Kinesiol
7:
241-250,
1997[Web of Science][Medline].
28.
Nordstrom, MA,
Fuglevand AJ,
and
Enoka RM.
Estimating the strength of common input to human motoneurons from the cross-correlogram.
J Physiol (Lond)
453:
547-574,
1992
29.
Porter, R,
and
Lemon RN.
Corticospinal Function and Voluntary Movement. Oxford, UK: Oxford University Press, 1993.
30.
Powers, RK,
Van den Noven S,
and
Rymer WZ.
Evidence of shared, direct input to motoneurons supplying synergist muscles in humans.
Neurosci Lett
102:
76-81,
1989[Web of Science][Medline].
31.
Rainoldi, A,
Galardi G,
Maderna L,
Comi G,
Lo Conte L,
and
Merletti R.
Repeatability of surface EMG variables during voluntary isometric contractions of the biceps brachii muscle.
J Electromyogr Kinesiol
9:
105-119,
1999[Web of Science][Medline].
32.
Roeleveld, K,
Blok JH,
Stegeman DF,
and
van Oosterom A.
Volume conduction models for surface EMG: confrontation with measurements.
J Electromyogr Kinesiol
7:
221-232,
1997[Web of Science][Medline].
33.
Roeleveld, K,
Drost G,
Stegeman DF,
Blok JH,
and
Scholle HC.
Estimation of motor unit position from surface EMG topographical maps.
In: The State of the Art on Modelling Methods for Surface Electromyography, edited by Hermens HJ,
Stegeman DF,
Blok JH,
and Freriks B.. Enschede, The Netherlands: Roessingh Research and Development, 1998, p. 76-81.
34.
Roeleveld, K,
Stegeman DF,
Vingerhoets HM,
and
Van Oosterom A.
Motor unit potential contribution to surface electromyography.
Acta Physiol Scand
160:
175-183,
1997[Web of Science][Medline].
35.
Rosenfalck, P.
Intra- and extracellular potential fields of active nerve and muscle fibres. A physico-mathematical analysis of different models.
Acta Physiol Scand Suppl
321:
1-168,
1969.
36.
Roy, SH,
De Luca CJ,
and
Schneider J.
Effects of electrode location on myoelectric conduction velocity and median frequency estimates.
J Appl Physiol
61:
1510-1517,
1986
37.
Schmied, A,
Pagni S,
Sturm H,
and
Vedel JP.
Selective enhancement of motoneurone short-term synchrony during an attention-demanding task.
Exp Brain Res
133:
377-390,
2000[Web of Science][Medline].
38.
Scholle, HC,
Struppler A,
Anders C,
and
Schumann NP.
Kontrolle der isometrischen Muskelkontraktion bei muskulärer Hypotonie zentralen Ursprungs: EMG-Mapping-Analyse.
Z EEG EMG
23:
178-183,
1992.
39.
Sejersted, OM,
and
Sjogaard G.
Dynamics and consequences of potassium shifts in skeletal muscle and heart during exercise.
Physiol Rev
80:
1411-1481,
2000
40.
Semmler, JD,
Nordstrom MA,
and
Wallace CJ.
Relationship between motor unit short-term synchronisation and common drive in human first dorsal interosseus muscle.
Brain Res
767:
314-320,
1997[Web of Science][Medline].
41.
Stegeman, DF,
Dumitru D,
King JC,
and
Roeleveld K.
Near- and far-fields: source characteristics and the conducting medium in neurophysiology.
J Clin Neurophysiol
14:
429-442,
1997[Web of Science][Medline].
42.
Stegeman, DF,
and
Linssen WHJP
Muscle fibre membrane electrophysiology and surface EMG: a simulation study.
J Electromyogr Kinesiol
2:
130-140,
1992.
43.
Vollestad, NK.
Measurement of human muscle fatigue.
J Neurosci Methods
74:
219-227,
1997[Web of Science][Medline].
44.
Yao, W,
Fuglevand RJ,
and
Enoka RM.
Motor-unit synchronization increases EMG amplitude and decreases force steadiness of simulated contractions.
J Neurophysiol
83:
441-452,
2000
45.
Zwarts, MJ.
Evaluation of the estimation of muscle fiber conduction velocity. Surface versus needle method.
Electroencephalogr Clin Neurophysiol
73:
544-548,
1989[Web of Science][Medline].
This article has been cited by other articles:
|
|
 |
|
  J. S. Karlsson, K. Roeleveld, C. Gronlund, A. Holtermann, and N. Ostlund Signal processing of the surface electromyogram to gain insight into neuromuscular physiology Phil Trans R Soc A, January 28, 2009; 367(1887): 337 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J. Dartnall, M. A. Nordstrom, and J. G. Semmler Motor Unit Synchronization Is Increased in Biceps Brachii After Exercise-Induced Damage to Elbow Flexor Muscles J Neurophysiol, February 1, 2008; 99(2): 1008 - 1019. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. G. Keenan, D. Farina, F. G. Meyer, R. Merletti, and R. M. Enoka Sensitivity of the cross-correlaton between simulated surface EMGs for two muscles to detect motor unit synchronization J Appl Physiol, March 1, 2007; 102(3): 1193 - 1201. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. G. Keenan, D. Farina, R. Merletti, and R. M. Enoka Amplitude cancellation reduces the size of motor unit potentials averaged from the surface EMG J Appl Physiol, June 1, 2006; 100(6): 1928 - 1937. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Farina, L. Arendt-Nielsen, and T. Graven-Nielsen Spike-triggered average torque and muscle fiber conduction velocity of low-threshold motor units following submaximal endurance contractions J Appl Physiol, April 1, 2005; 98(4): 1495 - 1502. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. G. Keenan, D. Farina, K. S. Maluf, R. Merletti, and R. M. Enoka Influence of amplitude cancellation on the simulated surface electromyogram J Appl Physiol, January 1, 2005; 98(1): 120 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Farina, R. Merletti, and R. M. Enoka The extraction of neural strategies from the surface EMG J Appl Physiol, April 1, 2004; 96(4): 1486 - 1495. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L Rochette, S. K. Hunter, N Place, and R Lepers Activation varies among the knee extensor muscles during a submaximal fatiguing contraction in the seated and supine postures J Appl Physiol, October 1, 2003; 95(4): 1515 - 1522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Lowery, N. S. Stoykov, and T. A. Kuiken A simulation study to examine the use of cross-correlation as an estimate of surface EMG cross talk J Appl Physiol, April 1, 2003; 94(4): 1324 - 1334. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Farina, L. Fattorini, F. Felici, and G. Filligoi Nonlinear surface EMG analysis to detect changes of motor unit conduction velocity and synchronization J Appl Physiol, November 1, 2002; 93(5): 1753 - 1763. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
) is reached, a spike is fired
(dotted line) and the cycle restarts. Note that the common input
(top traces) is identical for both MNs (A and
B), whereas the individual input (second trace)
differs. The first firing of both MNs occurs simultaneously because
the depolarization was caused by a peak in the common input signal. All
traces are calibrated in noise units (NUs), with 1 NU being the
standard deviation of the membrane noise (third trace).
). D:
2 s of the firing pattern of a pool of 50 MNs firing independently
at 14.8 Hz (input 6 NU). E: synchronized firing of the same
50 MNs with 50% common noisy input.
),
bipolar with 1 cm IED (
) and bipolar SEMG with 2 cm IED
(
). RMS is given as a function of electrode position
with 1 referring to the most proximal and lateral electrode and 32 referring to the most distal and medial electrode. Values from the same
row of electrodes (8 electrodes in proximal-distal direction) are
connected by a solid line.
