Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 2003 Feb 28;548(Pt 2):615–629. doi: 10.1113/jphysiol.2002.033126

Pharmacologically induced enhancement of recurrent inhibition in humans: effects on motoneurone discharge patterns

Benjamin Mattei *, Annie Schmied *, Riccardo Mazzocchio *, Barbara Decchi *, Alessandro Rossi *, Jean-Pierre Vedel *
PMCID: PMC2342872  PMID: 12611926

Abstract

The aim of the present study was to investigate the effects of spinal recurrent inhibition on human motoneurone discharge patterns. The tonic discharge activity of motor unit pairs was recorded in the extensor carpi radialis (ECR) and abductor digiti minimi (ADM) muscles during voluntary isometric contraction. While undergoing continuous intravenous saline (NaCl 0.9 %) perfusion, the subjects were given a short lasting injection of l-acetylcarnitine (l-Ac), which has been found to potentiate recurrent inhibition in humans. The variability, synchronization and coherence of the motor unit discharges were analysed during four successive test periods (lasting 2–3 min each). A significant decrease in the inter-spike interval (ISI) coefficient of variation was observed in the discharge patterns of the motor units tested in the ECR and not in the ADM, which were not accompanied by any consistent changes in the mean ISIs of the motor unit activity in either muscle. The l-Ac injection also led to a significant increase in the synchronization in half of the motor unit pairs tested in the ECR muscle (n = 29), whereas no consistent changes were observed with the ADM motor units (n = 25). However, coherence analysis failed to reveal any consistent differences in the incidence of significant values of coherence spectrum between the pre-injection and injection periods among the motor unit pairs tested with either saline or l-Ac injections, in either the ECR or ADM muscles. The contrasting effects on the variability and the synchronization of the motor unit discharges observed with ECR motoneurones known to undergo recurrent inhibition and with ADM motoneurones known to lack recurrent inhibition suggest that the drug may have specific effects which are mediated by an enhancement of the Renshaw cell activity. The decrease in the ISI variability is in line with the hypothesis that recurrent inhibition may contribute along with the post-spike after-hyperpolarization to limiting the influence of the synaptic noise on the firing times of steadily discharging motoneurones. The present data, which suggest that recurrent inhibition plays a synchronizing rather than a desynchronizing role, are in keeping with the fact that the Renshaw cells may provide an important source of common inhibitory inputs.

Although a large amount of data has been published on the spinal recurrent inhibitory network in animals, their functional interpretation still continues to raise many problems. One of the reasons for this situation is that the recurrent inhibition of spinal motoneurones is not simply a negative feedback loop, but forms an intricate network which is subject to many modulating influences arising from segmental afferents and spinally descending neural systems (Haase et al. 1975; Windhorst, 1990, 1996). Studies carried out on reduced preparations therefore have considerable limitations, since the functional speculations to which they have given rise have been mostly based on data obtained on an isolated system. In humans, the recurrent inhibitory network has been studied under physiological and pathological conditions, generally using a complex procedure involving supramaximal stimulation of peripheral nerves (see Katz & Pierrot-Deseilligny, 1998). However, most of the methods used so far have been necessarily restricted to non-invasive techniques, which have obviously raised some problems and may have led to some mistakes being made in the interpretation of the data.

Some progress has been achieved in the study of ongoing spinal recurrent inhibition in humans during voluntary motor activity thanks to the use of substances with cholinomimetic activity but no systemic side effects, such as l-acetylcarnitine (l-Ac) (Mazzocchio & Rossi, 1989, 1997). The most significant outcome of these studies has probably been the possibility of using l-Ac as an independent means of identifying the changes in motoneuronal activity due to the potentiation of the recurrent inhibitory system induced by this drug via the enhancement of the cholinergic transmission from motoneurone recurrent collaterals to Renshaw cells. Based on these premises, the present study was designed to investigate the effects of the intravenous administration of l-Ac on the firing patterns of single human motor units recorded from two muscles, the extensor carpi radialis (ECR) and abductor digiti minimi (ADM), which differ markedly in terms of their recurrent inhibitory inputs (Rossi & Mazzocchio, 1992; Katz et al. 1993).

The main aim was to determine whether motoneurone synchronization increases or decreases in response to potentiation of the recurrent inhibition: both of these effects have been previously postulated (Gel'fand et al. 1963; Elble & Randall, 1976; Adam et al. 1978; Davey et al. 1993; Türker et al. 1996; Maltenfort et al. 1998). In addition, the idea that recurrent inhibition might be involved in the stabilization of the motoneurone discharge variability (Tokizane & Shimazu, 1964; Derfler & Goldberg, 1978) was also investigated by testing the changes in the coefficient of variation (CV) of the inter-spike interval during l-Ac injection. In order to avoid the problems associated with the effects of muscle fatigue during maintained, although fairly weak, voluntary contraction and to prevent the risk of losing the units if the subjects were allowed to rest during the recording session, the duration of the drug infusion was strictly controlled and drastically reduced with respect to the original procedure (Mazzocchio & Rossi, 1997). This modification had no apparent side-effects and most importantly, it made it possible to clearly detect the recovery of the changes which were found to be closely associated with the l-Ac injection. The ECR and ADM muscles were chosen to ensure the specificity of the effects induced by l-Ac, since any change observed in the motor unit firing patterns in response to drug administration will occur in the motoneurones receiving recurrent inhibitory inputs, such as those of the ECR muscle, but not in those where this inhibition is lacking, such as those of the ADM (Rossi & Mazzocchio, 1992; Katz et al. 1993; Illert & Kümmel, 1999).

METHODS

The experimental procedure was approved by the Ethics Committee of the local Medical University. It was carried out on 11 healthy human subjects (3 females and 8 males, aged 22–63 years) who gave their prior informed consent in writing to the experimental procedure, as required by the Declaration of Helsinki (1964). Motor unit activities were recorded in the ECR and the ADM muscles of the subjects' right arms, all the subjects being right-handed.

Instructions to the subjects

The subjects were seated in an armchair, with their right forearm placed in a device ensuring that a stereotyped position was maintained from one experiment to another. When testing the motor unit properties in the ECR muscle, the distal end of the forearm was immobilized in a device holding the hand in a semi-prone position, with the wrist flexed at an angle of 10 deg. The back of the hand was kept in contact with a rigid rod on which the subjects were asked to push by selectively contracting their wrist extensor muscles while keeping their fingers passively flexed. When recording motor units in the ADM muscle, the subject's hand was maintained in a prone position, the fingers being positioned in order to permit selective isometric contraction of the muscle tested by abduction of the fifth finger on the rod.

The activity of motor unit pairs was continuously monitored on an oscilloscope screen facing the subject. A loudspeaker provided the subject with auditory feedback information about the firing activity of any unit that appeared to be recruited at a higher force level than the other one. The only instructions given to the participants were to adjust the level of contraction of the muscles tested to the minimum level necessary to effortlessly maintain the two motor units firing steadily.

Experimental procedure

During the pharmacological experimental sessions, which lasted about 520 to 1200 s, a continuous intravenous perfusion of sterile saline (NaCl 0.9 %), starting at very slow rate, was performed on the left arm of the subjects. After about 3 min, the saline perfusion was replaced by a short lasting injection (154 ± 43 s) of l-Ac (30 mg kg−1 dissolved in sterile saline) performed by means of a two-way tap. At the end of the l-Ac injection, the saline perfusion was resumed. The effectiveness of the short-lasting l-Ac injection in potentiating recurrent inhibition as tested by the conditioned H reflex (Bussel & Pierrot-Deseilligny, 1977), was similar to that observed with the original protocol (Mazzocchio & Rossi, 1997). Nevertheless, the l-Ac effect duration and the recovery period were much shorter, as required in this study (A. Rossi & R. Mazzocchio, unpublished data).

In order to monitor the time course of the effects of the drug injection, analyses were performed off-line during four successive sequences (Fig. 1) together covering the whole l-Ac injection period, one before the injection (Pre l-Ac), one encompassing the whole injection period (l-Ac) and two successive periods following the injection (Post l-Ac1 and Post l-Ac2). Control experiments in which saline perfusion only was administered (520–800 s), i.e. in which l-Ac injection was replaced by a saline injection (Saline 2 in Figs 2, 3 and 5), were also performed. The motor unit discharge properties were analysed in the same way here as during the l-Ac sessions, during four successive test periods with similar durations (Saline 1, 2, 3, 4; 148 ± 37 s). The subjects had no way of knowing whether l-Ac or saline was being injected in the second test period, since the experiments were conducted using the single-blind procedure.

Figure 1. Experimental procedure and analysis of the effects of l-Ac intravenous injection on the discharge patterns of a motor unit pair recorded in the ECR muscle.

Figure 1

Open in a new tab

While undergoing a continuous saline (NaCl 0.9 %) perfusion, the subject had to use the auditory biofeedback information provided to keep two ECR motor units (MUs) recorded by means of two intramuscular microelectrodes discharging steadily. After about 3 min, a short lasting injection of l-Ac was administered instead of the saline perfusion. The ISI CV (CV = ISI s.d./mean ISI) of the firing rate of each motor unit (A) and the synchronization (B) were analysed during four successive test periods, before (Pre l-Ac), during (l-Ac) and after (Post l-Ac1 and Post l-Ac2) the l-Ac injection. The presence of synchronization peaks was determined by computing the cross-correlation histograms (± 100 ms) and cumulative sums (CUSUM; Ellaway, 1978). Only the central parts of the cross-correlation histograms (± 50 ms) are shown in B, in order to focus on the peak region. The peak significance was assessed by determining the z score (z > 3.27; Garnett & Stephens, 1980) and the synchronization strength was assessed in terms of the synchronous impulse probability (SIP). The motor unit macro-potential (macro-MUP) was periodically extracted by spike-triggered averaging the surface EMG activity to ensure that the same motor units were being tested throughout the experimental session. N, number of triggering impulses.

Figure 2. Effects of the l-Ac injection on the discharge variability of the motor units recorded in the ECR (A) and in the ADM (B) muscles.

Figure 2

Open in a new tab

The mean value of the ISI CV of the ECR motor units was significantly smaller (P = 0.02) during l-Ac injection (A, filled circles) than during the Pre l-Ac period. No such effect was observed on the firing pattern of the ADM motor units (B, filled triangles). In the control experiments in which l-Ac injection was replaced by saline injection (Saline 2; single-blind procedure), the variability of the firing rates of the motor units recorded in the ECR (A, open circles) and ADM (B, open triangles) muscles increased gradually in the course of the recording sessions.

Figure 3. Effects of the l-Ac injection on the synchronization of motor unit pairs tested in the ECR muscle.

Figure 3

Open in a new tab

With the χ2 test procedure described in the Methods, nine ECR motor unit pairs showed a significant increase in their synchronization during l-Ac injection (A, continuous lines). No significant changes were observed in the control sessions (B) in which l-Ac injection was replaced by a saline injection (Saline 2). In A and B, the open triangles and arrows indicate the motor unit pairs that were not significantly synchronized in the pre-injection (l-Ac or Saline) period. Upon pooling the data (C, filled circles), the mean strength of the motor unit pair synchronization was found to be significantly greater during l-Ac injection (P < 0.001) and during the Post l-Ac1 test period (P = 0.015) than during the Pre l-Ac period. Pooled data from control sessions with saline perfusion alone showed the occurrence of no significant changes in the synchronous motor unit pattern (C, open circles).

Figure 5. Relationships between the changes in synchronization and discharge variability of each motor unit pair tested in the ECR (A) and ADM (B) muscles during l-Ac and saline pre-injection periods.

Figure 5

Open in a new tab

The changes in the synchronization peak area and in the mean geometric CV were assessed by subtracting the values of each parameter assessed during the saline pre-injection period from those assessed during the l-Ac (filled symbols) or saline (open symbols) injection period. A total number of eight (A, filled stars) out of the nine ECR motor unit pairs which discharged with a significantly greater synchrony during l-Ac injection also discharged with a lower variability. No such effects were observed in the ADM muscle (B).

Data recording

The discharge patterns of the motor unit pairs and the surface EMG activity were recorded in the ECR and ADM muscles of the right arm. Both muscle activities were digitized and recorded on a computer using a Spike2 CED device (Cambridge Electronic Design, Cambridge, UK).

The overall amplified and filtered (band-pass: 30 Hz-10 kHz) EMG activity of the ECR or ADM muscles was recorded using pairs of non-polarizable single-use surface electrodes (Ag-AgCl) placed 2 cm apart over each of the muscle sets. The amplified and filtered (300–3000 Hz) activity of two single motor units was recorded in the muscles tested by means of two sterilized single-use tungsten microelectrodes (impedance 12 MΩ tested at 1 kHz; Frederick Haer & Co., Bowdoinham, ME, USA). The microelectrodes were inserted transcutaneously 1 to 3 cm apart and moved in minute steps until a stable recording of clearly identifiable single motor unit action potentials was obtained from each electrode. The surface electrodes and the microelectrodes were connected to amplifiers via probes with an isolated ground for optimum protection of the subject (current leakage less than 3 μA).

Firing pattern analysis

The action potentials fired by a single motor unit were discriminated and analysed using Spike 2 (Cambridge Electronic Design, Cambridge, UK) and Matlab (The MathWorks Inc., MA, USA) purpose written softwares. Auto-correlation histograms were systematically computed with each of the motor unit spike trains in order to detect any spurious firing of other motor units. The motor unit firing behaviour was plotted on an instantaneous frequency curve and characterized by the mean value and the standard deviation (s.d.) of the inter-spike intervals (ISI) calculated during each of the recording sequences. The ISI variability was evaluated by calculating the ISI coefficient of variation (CV = ISI s.d./mean ISI). The few ISIs that were longer than 500 ms (about 4 to 5 times the mean) were taken to result from pauses in the motor unit tonic activity and were not included in the calculation of the inter-spike interval statistics.

Motor units were characterized in terms of their overall electrical activity, i.e. the motor unit macro potential (macro-MUP), defined as the sum of the action potentials of the muscle fibre population innervated by a motoneurone. The macro-MUPs were extracted by spike-triggered averaging the surface EMG activity, and their area (mV ms−1) was taken to indicate their size. The reproducibility of the size and shape of the macro-MUPs extracted during muscle contraction was checked to ensure that the same motor units were being tested throughout each experimental session (Fig. 1A).

The net drive acting on the motoneurone pool of the muscle tested during each recording sequence was expressed in terms of the mean rectified EMG activity (μV).

Synchronization analysis

The motoneurones' synchronization was analysed by cross-correlating the spike trains recorded from a motor unit pair in the four recording sequences (Fig. 1B). The cross-correlation histograms thus computed yielded the distribution of the impulses produced by the one motor unit per 1 ms-bin, 100 ms before and after the impulses generated by the other motor unit. The triggering unit was that which generated the largest number of spikes. The central peak formed by the synchronization in the cross-correlation histogram (cross-correlogram) was delimited by means of the cumulative sum (CUSUM; Ellaway, 1978) of the changes in the bin counts with respect to the mean count in a baseline lasting 80 ms. The peak onset and offset were delimited visually on the basis of the rising inflections occurring in the central region of the CUSUMs computed from the left end of the cross-correlogram. In those motor unit pairs in which the CUSUMs did not show a clear-cut inflection around the time 0 ms, the strength of synchronization was calculated during an arbitrarily chosen period from −10 ms to +10 ms around time 0 ms.

The above-chance synchronization was evaluated in terms of synchronous impulse probability (SIP), corresponding to the number of impulses in the whole peak above the mean impulse count in the baseline (peak area) divided by the number of triggering action potentials. Whenever the synchronization is due to excitatory postsynaptic potentials (EPSPs) generated by common inputs, this index can be taken as an index to the EPSP amplitude (cf. Kirkwood & Sears, 1991). This parameter is only weakly if ever influenced by the motor unit firing rates (see Schmied et al. 1994).

The significance of the synchronization peak was evaluated at a significance level of P < 0.001 on the basis of the z score (z > 3.27; Garnett & Stephens, 1980).

Statistical assessment of the changes in synchronization

The effects of the l-Ac injection on a motoneurone's synchronization were characterized in terms of its temporal distribution with respect to the triggering events (the mean value and the dispersion) and its strength (the total number of extra counts above chance level). With each motor unit pair tested, the temporal distribution and the strength of the synchronization assessed before the injection were compared with those observed during and just after the injection.

The population of synchronized spikes was extracted from each cross-correlogram as follows. Let ts,k (time start) and te,k (time end) be the left and right limits, respectively, of the synchronization peak measured as explained above under conditions k, where k stands for the period before, during or just after the injection. Let Pk be the population of spikes which occurred between ts,k and te,k in the cross-correlogram drawn up under condition k, and Dk its discretised distribution with respect to the triggering events. In order to compare the peak distributions, the spikes synchronized by chance must not be taken into consideration. We therefore defined D′k as the distribution Dk from which the baseline was subtracted. Note that the negative values of the distribution D′k were rounded out to zero to rule out the effects of any occasional decreases in the counts of the synchronization peaks.

The main problem encountered here was due to the fact that many statistical tests require the distribution to be continuous. Although the discretized distribution of the spikes synchronized above chance level with respect to the triggering events, D′k, is known, there is no way of distinguishing between these spikes and those synchronized by chance. It is therefore not possible to determine the exact times of occurrence of the spikes included in D′k. To deal with this problem, we sub-sampled Pk randomly under each condition in order to obtain a sample P′k of exactly timed spikes distributed as in D′k. The analyses comparing the peak distributions were thus performed on these subsamples. We repeated this process 10 times in order to average the results and thus reduce the effects of the sampling variability.

The normality of the continuous distributions of P′k was tested with a statistical test based on the measures of symmetry and kurtosis (α= 5 %), as described by D'Agostino & Pearson (1973). Two cases were then dealt with separately. If both distributions were Gaussian, the dispersion was compared between the various periods with the F test (α= 1 %). If the dispersions were not significantly different, a simple Student's t test (α= 1 %) was used to compare the centres of the peaks. Otherwise, the Welch's approximation of the t test (α= 1 %), improved by Fenstad, which is also called the t′ test, was used. If at least one of the distributions was not Gaussian, the improved normal approximation of the Mann-Whitney test (α= 1 %) was used to compare the centres of the peaks. Whenever they were significantly different, the two samples were re-centred before performing the Ansari-Bradley test (α= 1 %) to compare the dispersions.

Finally, we defined the area Ak of a synchronization peak obtained under condition k as the area of D′k (i.e. the total number of extra counts above chance level), and we took Tk to denote the number of triggering events. A χ2 test (α= 1 %) was performed on the distributions (Ak, Tk−Ak) to compare the synchronization strength before, during and just after the drug (or saline) injection.

Coherence analysis

In order to further investigate the physiological changes induced by the l-Ac injection, we computed the coherence of each motor unit pair tested in ECR and ADM muscles under two conditions (before and during saline or l-Ac injection). The coherence is an index to the correlation between the frequency components of two spike trains, and this analysis is therefore liable to reveal the frequency contents of the motoneurone common inputs which fire rhythmically enough to generate significant coherence values (Rosenberg et al. 1989; Farmer et al. 1993). The method is illustrated by the example given in Fig. 6.

Figure 6. Example of the changes induced by l-Ac injection on the synchronization and coherence between the discharge patterns of two ECR motor units.

Figure 6

Open in a new tab

Cross-correlation histogram analysis on one ECR recording session showing changes in synchronization during the saline pre-injection (A) and l-Ac injection periods (C), and analysis of the coherence spectra in the frequency band 0–70 Hz (B and D, respectively) during the same test periods. The net differences between the coherence values measured during both periods are plotted in E after subtracting B from D. A significant (P < 0.05) decrease was observed at 2 Hz, whereas significant increases occurred at 16 and 47 Hz (asterisks).

The coherence was calculated with the spike train time-series, based on the procedure described by Rosenberg et al. (1989) and Farmer et al. (1992) using a Matlab routine written by D. M. Halliday (University of York, York, UK). For each motor unit pair, coherence spectra were computed separately, from 0 to 100 Hz with a resolution of 1 Hz in the pre-injection and injection periods (Fig. 6B and D). In each spectrum, a 95 % confidence level (dotted line, Fig. 6B and D) was computed under the assumption that the two spike trains were independent (Rosenberg et al. 1989). Any coherence value which reached this level was taken to reflect the existence of a significant correlation between the corresponding frequency components of the spike trains.

A subdivided χ2 analysis (P < 0.05) with correction for continuity (Zar, 1996) was used to compare the number of significant coherence values observed between the pre-injection and injection periods at each frequency in the spectra computed with all the motor unit pairs tested with either saline or l-Ac injection in each muscle.

With each motor unit pair, the coherence values were compared statistically between the pre-injection and injection periods, using a method proposed by Rosenberg et al. (1989). This is based on the fact that |tanh−1 (c11/2) – tanh−1(c21/2)| (Nseg)1/2 follows a normalized Gaussian distribution, c1 and c2 being the coherence values measured in two conditions and Nseg being the number of complete segments of time used to compute each coherence value. Whenever this difference reached the 95 % confidence level, the changes in coherence were said to be significant. With each pair, the frequencies showing significant changes were listed and classified within four frequency bands of 0–5 Hz, 5–10 Hz, 10–15 Hz and 15–40 Hz. In each band, the numbers of motor unit pairs that showed a significant increase or decrease in the coherence at least once or no change at all were counted. Within each band, a χ2 analysis (P < 0.05) of contingency tables was performed in each muscle (ECR and ADM) between the two types of injection (saline and l-Ac).

Pooled data statistical analysis

Upon pooling the data obtained with the whole population of motor units tested in all the subjects, the characteristics of the synchronization peaks (the latency, duration, amplitude) and the motor unit firing behaviour (the mean duration and variability of the inter-spike intervals) were compared between the four sequences of analysis by performing an ANOVA for repeated measures (StatView 4.1; Abacus Concepts, Berkeley, CA, USA). The significance level was set here at P < 0.05.

RESULTS

A total number of 54 motor unit pairs was tested in the right ECR (n = 29) and ADM muscles (n = 25) of 11 subjects.

Figure 1 illustrates the course of a typical session. In Fig. 1A, the firing behaviour of each member of the pair of units tested (MU1 and MU2) is shown in terms of its instantaneous frequency. The fact that the same motor units were being tested throughout the whole session was confirmed by the identical motor unit macro-potentials (macro-MUP) shown above the instantaneous frequency plot at the beginning and end of the recording session. The four test periods are indicated below and the values of the ISI CV are given in each period.

In Fig. 1B, the central part (± 50 ms) of the cross-correlation histograms computed in each period are indicated with their CUSUM superimposed, and the number of triggering impulses (N), the above chance synchronization per trigger (SIP) and the z value giving the significance of the synchronization peak are given in each period. No conspicuous peaks well above significance level (z > 3.27) ever occurred outside the l-Ac injection period with this motor unit pair.

Effects of l-acetylcarnitine on the ISI CV of the motoneurone discharge behaviour

In the ECR muscle, the effects of l-Ac injection on the ISI CV in the motoneurone discharge behaviour were assessed after pooling the data obtained with 32 motor units (7 subjects). In comparison with the mean value obtained during the Pre l-Ac period (0.173 ± 0.046), the ISI CV decreased significantly during l-Ac injection (0.160 ± 0.035, P = 0.02; Fig. 2A, filled circles). By contrast, during the Post l-Ac1 and 2 periods the ISI CV gradually became larger than during the Pre l-Ac period (0.179 ± 0.042 and 0.191 ± 0.046, respectively; Fig. 2A).

A total number of nine control sessions (21 motor units) was run with five subjects (Fig. 2A). Under these conditions, the ISI CV was found to increase continuously during the four successive test periods, from the beginning to the end of the saline perfusion (0.172 ± 0.051, 0.182 ± 0.055, 0.195 ± 0.059, 0.196 ± 0.062, respectively; Fig. 2A, open circles).

Similar analyses were performed on 28 motor units tested in the ADM muscle of nine subjects during saline/l-Ac/saline perfusion sessions. No decrease in the variability was observed during l-Ac injection, and the ISI CV increased gradually during the four test periods (0.224 ± 0.069, 0.227 ± 0.091, 0.244 ± 0.091, 0.252 ± 0.093, respectively; Fig. 2B, filled triangles). Similar results were obtained on 18 motor units recorded in seven subjects perfused with saline only (Fig. 2B; 0.212 ± 0.053, 0.213 ± 0.053, 0.221 ± 0.045, 0.240 ± 0.061, respectively; open triangles).

It is worth noting that in the Pre l-Ac test periods involving only saline perfusion, the mean values of the ISI CV were found to be significantly higher (P = 0.002) in the motor units recorded in the ADM muscle (0.224 ± 0.069, Fig. 2B) than in the ECR muscle (0.173 ± 0.046). Similar differences were observed during all the control sessions with saline perfusion alone (ADM: 0.222 ± 0.055, ECR: 0.186 ± 0.056; P < 0.001).

Effects of l-acetylcarnitine on the synchronization of motoneurone activity

During saline/l-Ac/saline sessions, among the 18 pairs of motor units tested in the ECR muscle, a total number of 15 pairs produced significantly large synchronization peaks during the Pre l-Ac period (z > 3.27, P < 0.001; synchronization probability: mean ±s.d.: 0.055 ± 0.037 impulses (imp) trigger−1, range: 0.018–0.148 imp trigger−1, peak duration: 10.67 ± 4.39 ms).

As in the example illustrated in Fig. 1B, 10 out of the 18 motor unit pairs tested in the ECR muscle showed greater synchronization during l-Ac injection than during the other periods tested. According to the χ2 test procedure described in the Methods, this increase was found to be significant in the case of nine pairs (Fig. 3A, continuous lines). This increase was significant at P < 0.01 in the case of five of them and at P < 0.05 in that of the other four. In the case of four pairs, this effect was found to be prolonged during the Post l-Ac1 period. The recovery was complete in all cases in the Post l-Ac2 period, during which the synchronization no longer differed significantly from that observed in the Pre l-Ac period. The increase in synchronization observed during the l-Ac injection was not accompanied by any consistent changes in the duration or the position of the synchronization peaks computed in the four test periods. This was confirmed by the lack of any significant changes observed with the χ2 procedures used to compare the distributions of the synchronous firing events.

The increase in synchronization which occurred in response to l-Ac injection was confirmed upon pooling the data obtained on the 18 motor unit pairs tested (Fig. 3C, filled circles). The ANOVA procedure for repeated measures showed that in comparison with the mean synchronization value obtained during the Pre l-Ac period (0.050 ± 0.036 imp trigger−1), the synchronization was significantly stronger during the l-Ac injection and during the Post l-Ac1 period (0.063 ± 0.041 imp trigger−1, P < 0.001; 0.057 ± 0.041 imp trigger−1, P = 0.015, respectively). During the Post l-Ac2 period, the synchronization probability was found to have returned to its pre-injection value (0.050 ± 0.034 imp trigger−1).

It is worth noting that the increase in the synchronization probability as well as the decrease in the ISI variability associated with the l-Ac injection were not accompanied by any significant changes in either the inter-spike interval in the motor unit discharge pattern, taken as an index to the motoneurone net excitatory drive throughout the four test periods (0.099 ± 0.012, 0.096 ± 0.012, 0.098 ± 0.014, 0.096 ± 0.013 s, respectively), or in the mean surface EMG activity taken as an index to the motoneurone pool net excitatory drive (13.56 ± 6.36, 13.54 ± 6.40, 13.29 ± 5.43, 13.66 ± 5.79 μV, respectively) during the four successive test periods.

In the nine control experimental sessions during which the subjects were perfused with saline only (Fig. 3B), the χ2 test failed to detect any significant changes in the synchronization probability between the four test periods, whether or not the motor unit pairs showed significant (z > 3.27) synchronization peaks during the pre-injection period (seven and four pairs, respectively). Upon pooling the data (Fig. 3C, open circles), the synchronization strength was again found to be very similar during all four test periods (0.050 ± 0.021, 0.048 ± 0.018, 0.051 ± 0.020, 0.049 ± 0.019 imp trigger−1, respectively).

The same procedure was applied to analyse the changes in the synchronization probability of 16 motor unit pairs tested in the ADM muscle of eight subjects who underwent l-Ac injections. In the case of the unit pairs which were significantly synchronized during the Pre l-Ac period (n = 15; peak duration: 12.36 ± 4.01 ms) as well as in the case of one non-synchronized pair, no significant changes in the synchronization probability were observed during the l-Ac injection or during the Post l-Ac periods (Fig. 4A). Upon pooling the data, the synchronization strengths in the Pre l-Ac, l-Ac, Post l-Ac1 and Post l-Ac2 periods were: 0.094 ± 0.052, 0.093 ± 0.051, 0.096 ± 0.051, 0.095 ± 0.050 imp trigger−1, respectively; (Fig. 4C, filled triangles).

Figure 4. Effects of the l-Ac injection on the synchronization of motor unit pairs tested in the ADM muscle.

Figure 4

Open in a new tab

No significant changes in the synchronization of ADM motor units were observed during l-Ac injection (A) or during the control saline injection (B, Saline 2). In A and B, the open triangles and arrows indicate the motor unit pairs that were not significantly synchronized in the pre-injection (l-Ac or Saline) period. This finding was confirmed upon pooling the data (C) obtained in the drug sessions (filled triangles) and control sessions (open triangles).

In the same way, no significant changes were observed between the four test periods in the synchronization of the nine ADM motor unit pairs tested in the control experiments with saline injection (Fig. 4B). Upon pooling the data, the mean synchronization values were: 0.102 ± 0.061, 0.103 ± 0.051, 0.103 ± 0.053, 0.107 ± 0.052 imp trigger−1, respectively (Fig. 4C, open triangles).

In the ADM muscle, the mean duration of the inter-spike intervals (0.124 ± 0.017, 0.122 ± 0.020, 0.119 ± 0.019, 0.129 ± 0.021 s, respectively) and the mean surface EMG activity (6.78 ± 1.69, 6.84 ± 1.61, 6.68 ± 1.71, 6.67 ± 1.69 μV, respectively) were found to be similar in all four successive test periods.

It is worth noting that in addition to the greater ISI variability, the motor units tested in the ADM fired with much greater synchrony than those tested in the ECR muscle in the control experiments performed on the same subjects (0.104 ± 0.052 vs. 0.049 ± 0.019 imp trigger−1, respectively).

Figure 5 illustrates the covariations observed between the changes in the synchronization peak area and the changes in the mean geometric firing CVs in each of the motor unit pairs tested in the ECR (Fig. 5A) and ADM (Fig. 5B) muscles. The changes were calculated by subtracting the values of the synchronization and the variability parameters assessed during the pre-injection period from that assessed during l-Ac or saline injection period. This analysis showed that eight out of the nine ECR motor unit pairs (Fig. 5A, filled stars) which showed a significant increase in their synchronization (ordinate) during l-Ac injection, discharged with a lower variability (abscissa).

Effects of l-acetylcarnitine on the coherence of the motoneurone activity

The cross-correlation histograms (Fig. 6A and C) and the coherence spectra (Fig. 6B and D) computed during the pre-injection and the l-Ac injection periods with the spike trains of a motor unit pair tested in the ECR muscle are shown in Fig. 6. The occurrence of a conspicuous increase in the rate of synchronous firing was shown by the presence of a central peak which area was about 2.5 times larger during the drug injection (Fig. 6B) than previously (Fig. 6A). In both conditions (Fig. 6B and D), the coherence reached highly significant values between 0 and 10 Hz, whereas, the 95 % confidence level (dashed lines Fig. 6B and D) was more rarely crossed between 10 and 40 Hz. The net differences between the actual coherence values measured during the l-Ac injection period (Fig. 6D) and those measured during the pre-injection period (Fig. 6B) are plotted in Fig. 6E. The asterisks mark the few frequency values at which the statistical comparisons based on the subtraction of the coherence tanh−1 function (see Methods) reached the significance level (P < 0.05). In this example, a significant decrease in the coherence was observed at 2 Hz, whereas significant increases were observed at 16 and 47 Hz, respectively.

The incidence of significant coherence values during the pre-injection (dashed line) and injection periods (continuous line) with either saline or l-Ac are given in terms of the percentage of the motor unit pairs tested in ECR and ADM muscles in Fig. 7A and B, and C and D, respectively. The first point worth noting is that 100 % of the motor unit pairs showed significant coherence values within the 0–10 Hz frequency band in every case. Besides, nearly 50 % of them also showed significant coherence values between frequencies of 10 to 40 Hz, especially in the ADM muscle. In each of the four graphs and at each frequency value, the numbers of motor unit pairs showing significant coherence values before and during the injection were compared with a χ2 test. No consistent changes were detected between the two periods within each group of ECR and ADM motor units tested with either saline or l-Ac injection. Whichever type of injection was administered, the coherence spectra computed during two successive recording periods with the same population of motor units were always quite similar. This contrasts with the noticeable differences which can be observed between the spectra of various populations of motor units tested in the same muscle (see Fig. 7C and D).

Figure 7. Population analysis of the incidence of significant coherence in ECR and ADM muscles during l-Ac or saline injection.

Figure 7

Open in a new tab

The percentage of the motor units tested in the ECR (A and B) and ADM (C and D) muscles showing significant coherence values are plotted in the 0–100 Hz frequency range (abscissa) during the pre-injection (dashed lines) and injection period (continuous lines) with either saline (A and C) or l-Ac (B and D) injections. All the motor unit pairs tested in the ECR and ADM muscles showed significant coherence values within 0–10 Hz. In the 10–40 Hz frequency band the incidence of significant values tended to be higher in the ADM muscle than in the ECR muscle. In the motor unit population analysis, the χ2 test did not show the occurrence of any significant changes during either saline or l-Ac injection as compared to the pre-injection period in either of the muscles tested.

We then focused on the statistical detection of the changes in coherence between the pre-injection and injection periods with each motor unit pair, as shown in Fig. 6E. Significant increases or decreases were detected in various regions of the coherence spectrum, without any consistent trend towards a specific frequency band with all the populations of motor units tested in both muscles. With a given motor unit pair, the few significant changes detected were distributed within the four arbitrarily defined bands (0–5, 5–10, 10–15 and 15–40 Hz). These data are summarized in Table 1. All in all, the distributions of the increases and decreases in the coherence within each of the bands were not found to differ at all consistently between the two types of injection (saline and l-Ac) in either muscle. Moreover, the significant increases in synchronization observed during the l-Ac injection with nine of the pairs tested in the ECR were not consistently associated with any specific pattern of change in the coherence.

Table 1.

Changes in motor unit coherence values in the frequency band 0–40 Hz

ECR ECR ADM ADM
Coherence l-Ac saline l-Ac saline
0–5 Hz Increase 3 4 3 1
Decrease 8 2 5 2
No change 7 5 8 6
5–10 Hz Increase 6 2 2 2
Decrease 0 0 4 1
No change 12 9 10 6
10–15 Hz Increase 2 1 1 0
Decrease 0 0 0 0
No change 16 10 15 9
15–40 Hz Increase 4 1 2 4
Decrease 3 1 4 1
No change 11 9 10 4
Open in a new tab

DISCUSSION

The main finding obtained in the present study was that the potentiation of the cholinergic transmission induced by administering an intravenous l-Ac injection was associated with a significant decrease in the variability of the inter-spike interval and with a significant increase (ranging from 24 to 158 %) in the motor unit synchronization in the ECR muscle. This increased synchronization was not associated with any specific change in the coherence, however. By contrast, the same l-Ac injection had no effect on motoneurones devoid of recurrent inhibition such as those innervating the ADM.

Specificity of l-acetylcarnitine effects on the spinal recurrent inhibition

Although in principle, intravenous l-Ac injection is liable to potentiate all cholinergic pathways, there exist several lines of evidence supporting the idea that the present results may have resulted from the potentiation of the spinal recurrent inhibition by the l-Ac injection.

The possibility that l-Ac may have altered the neuromuscular transmission process can be ruled out for the following reasons. In their extensive studies on the effects of l-Ac on various upper and lower limb muscles, Rossi & Mazzocchio (1992) never observed any significant changes in size in the direct motor responses or H-reflex waves during l-Ac administration. The effects of l-Ac on the motoneurone discharge pattern may have been due to the potentiation of supraspinal networks rather than to spinal recurrent inhibition. In fact, the authors of some recent reviews have suggested that cholinergic neurones may be present in numerous parts of the nervous system, especially in the motor and sensory areas of the human cortex (Jones et al. 1999; Oda & Nakanishi, 2000). This possibility seems to be ruled out, however, by the lack of effect of l-Ac on the firing patterns of the motor units tested here in the ADM muscle, the activity of which, like that of the other hand muscles, is particularly dependent on cortico-spinal inputs. Owing to their unique position and properties, Renshaw cells are good candidates for mediating the effects of l-Ac observed on the motoneurone firing pattern. During voluntary muscle contraction, l-Ac can be expected to enhance the efficiency of motor axon collaterals on Renshaw cells. This enhancement might lead Renshaw cells to fire bursts after motoneurone impulses or increase their ability to fire spontaneously, as observed in animals (Ross et al. 1975; Van Keulen, 1981). This interpretation has been corroborated by the absence of l-Ac effects on the firing pattern of the ADM motoneurones, which are not subjected to recurrent inhibition (e.g. Katz et al. 1993).

Inter-spike interval duration and variability

Assuming that the Renshaw cell activity was enhanced during the l-Ac injection, one of the consequences which might be expected to occur would be a decrease in the firing rate of the target motoneurones. This was not in fact observed. Several conjectures can be made in this connection.

Firstly, on the basis of animal experiments, the effectiveness of recurrent inhibition at reducing the motoneurone firing rate is generally thought to be quite limited, in keeping with the small amplitude of the postsynaptic inhibitory potential or current generated by the Renshaw cells (cf. Windhorst, 1996).

Secondly, it must be kept in mind that the recording procedure used here relied entirely on the subject's ability to maintain the motor unit firing steadily. This is an obvious bias which limits the possibility of detecting any changes in the discharge frequency as long as they can be compensated for by the subject's voluntary drive. Nevertheless, it should be noted that if a surplus descending command was actually added to the motoneurone inputs to compensate for the surplus recurrent inhibition, this did not show up in the coherence spectra in either the 0–5 Hz band, thought to reflect modulations in the voluntary drive (De Luca et al. 1982), or the 19–40 Hz band, thought to reflect the activity of supraspinal inputs (Farmer et al. 1993).

In spite of the absence of any consistent changes in the mean ISI duration, we observed a significant decrease in the inter-spike interval variability in the ECR muscle during the administration of l-Ac injections. This is quite striking, since the ISI duration and variability are known to be correlated in the tonic discharges of human motor units (cf. Matthews, 1996).

Synaptic noise is known to be the main source of the discharge variability of spinal motoneurones (Calvin & Stevens, 1968). The after hyperpolarization (AHP) which follows each spike plays a major role in limiting the fluctuation of the motoneurone firing time due to synaptic noise (Person & Kudina, 1972; Matthews, 1996). It has been proposed, however, that the recurrent inhibition which occurs concurrently with AHP when it is generated by the discharge of the motoneurone itself and/or by the synchronized discharges of other motoneurones, might also minimize the fluctuations of the inter-spike intervals (Tokizane & Shimazu, 1964; Derfler & Goldberg, 1978). The decrease in the ISI variability observed during the administration of l-Ac is in good agreement with this hypothesis. Moreover, in the present control sessions with saline perfusion only, the ISI variability was found to be significantly lower in the ECR muscle, where recurrent inhibition is operative, than in the ADM muscle, where it is absent. This is in line with the proximo-distal increase observed in the ISI variability of arm muscle motor units (Kukulka & Clamann, 1981) and the proximo-distal decrease in the strength of recurrent inhibition observed in the same arm motor nuclei (Katz et al. 1993).

Interestingly, in the absence of l-Ac, the firing variability was found to increase steadily throughout the recording session. This may well reflect some fatigue-related effect on the motoneurone firing pattern (Nordstrom & Miles, 1991). In the ECR but not in the ADM muscles, this effect was blocked during l-Ac injection and resumed its course immediately afterwards, during the first post-injection period.

All in all, the present data point to the possibility that recurrent inhibition may be specifically involved in limiting the discharge variability of the motoneurones. The question remains to be elucidated, however, as to whether recurrent inhibition acts by locally shunting a specific component of the synaptic noise or by simply damping the overall impact of the synaptic noise together with AHP.

Increase in synchronization

In the present study, an enhancement of the synchronization of the voluntarily activated motor units tested in the ECR muscle was observed during the pharmacologically induced potentiation of recurrent inhibition. However, before addressing the contributory role of recurrent inhibition, it is necessary to take into account the possible existence of synaptic-like contacts between motoneurones and recurrent collaterals (cf. Cullheim et al. 1984). These contacts might mediate a cholinergic recurrent excitation which might be expected to be l-Ac-sensitive and able to synchronize the motoneurone discharges most effectively. Although there is little evidence that synaptic-like contacts of this kind might be actually functional (cf. Khatib et al. 1986), the possibility that they might be activated by l-Ac injection cannot be completely ruled out.

The contribution of recurrent inhibition to the coupling of motoneurone discharges has given rise to a long standing debate between those in favour of a desynchronizing or decorrelating role (Gel'fand et al. 1963; Adam et al. 1978; Maltenfort et al. 1998) and those in favour of a synchronizing or correlating role (Renshaw, 1946; Elble & Randall, 1976; Loeb et al. 1987; Davey et al. 1993; Türker et al. 1996).

In anaesthetized cats, the synchronous discharges of motoneurones activated by Ia inputs driven by sinusoidal muscle stretching was found to be enhanced by a pharmacological blockade of Renshaw cell activity (Adam et al. 1978). The differences between the experimental conditions used make it rather difficult to make comparisons with our data, however, and most importantly, Ia inputs are not generally held to be a major source of motoneurone synchronization during voluntary muscle contraction in humans (Farmer et al. 1993; Schmied et al. 1995). This question has also been addressed by Maltenfort et al. (1998) in a simulation study. The latter authors observed a tendency for motoneurones to be less synchronized than by chance in the absence of recurrent inhibition, which contrasted markedly with the strong synchronization observed in the motoneurone pools of the hand muscles devoid of recurrent inhibition. This pattern was partly reversed when recurrent inhibitory inputs were added when the simulation was performed with a broad band input (> 50 Hz) but not when narrower bandwidths were used. This was taken to reflect a decorrelating effect (Maltenfort et al. 1998).

Inhibitory processes have been thought to contribute to the synchronization of neuronal activity in several ways, which are not mutually exclusive. The first way involves common inhibitory synaptic inputs distributed to the neurones which discharge in synchrony (Moore et al. 1970). The second way involves oscillatory phenomena or slow waves simultaneously affecting the neurones' membrane potential, with a special contribution from inhibitory synaptic potentials (Jefferys et al. 1996; Desmaisons et al. 1999). The third way involves emergent network oscillations with a special contribution from reverberating inhibitory loops (for a review, see Whittington et al. 2000). Whether they affect the membrane potential or the neuronal network, these oscillatory phenomena would lead to rhythmic synchronization, which was not found to be a prominent feature in the coherence analysis performed in the present study. Nor were any consistent changes observed in the coherence spectra during the l-Ac injections, particularly in the 8–12 Hz band which has been thought to be possibly related to Renshaw cell activity (Elble & Randall, 1976). Hence, recurrent inhibition does not appear to be responsible for synchronized oscillatory activity.

The involvement of common excitatory inputs in the synchronization of motoneurone activity has been recognized for a long time, and is known as the short-term synchronization process (Sears & Stagg, 1976). There is some evidence that cortico-spinal inputs may make a preferential contribution to the synchronized motoneurone activity (Farmer et al. 1993; Schmied et al. 1995). In the present study as well as in an earlier report (Bremner et al. 1991), the degree of synchronization was found to be stronger in the motoneurone pools of forearm distal muscles, which are known to receive dense cortico-spinal projections but no recurrent inhibitory collaterals, than in the motoneurone pools of proximal muscles, where recurrent inhibition is present but the cortico-motoneuronal inputs are less powerful (for a review, see Lemon, 1993; Illert & Kümmel, 1999). In view of these data, the increase in the synchronization associated with the l-Ac-induced potentiation of recurrent inhibition is quite intriguing. The various hypotheses liable to account for this increase will now be discussed in turn.

The simplest hypothesis which can be advanced is that an increase may occur in the voluntary excitatory drive, including synchronizing cortico-motoneuronal inputs, which may compensate for the increase in the recurrent inhibitory inputs which can be expected to be induced by the l-Ac injections. This possibility was not supported, however, by the lack of consistent changes in coherence in the 15–40 Hz band, which is thought to reflect the activity of the motoneurones' cortical inputs (Farmer et al. 1993).

On the basis of data obtained in vitro (Türker & Powers, 2001), another hypothesis can be put forward, according to which recurrent inhibition may be one of the main sources of the common inhibitory inputs, and might take a direct part in the synchronizing processes at work in the proximal muscle motoneurone pool. A common input composed of relatively small inhibitory postsynaptic potentials (IPSPs), as might be expected from enhancement of recurrent inhibition, has been shown to give rise to relatively narrow peaks in motoneurone synchronization (Türker & Powers, 2001).

Finally, in keeping with the fact that common excitatory inputs are liable to contribute to the short-term synchronization process (Kirkwood & Sears, 1991), a third hypothesis can be proposed, involving the possibility of an enhancement of the common EPSPs' effectiveness by recurrent inhibition. In this context, facilitatory interactions between consecutive IPSPs and EPSPs have been recently reported to occur in several experimental studies showing how a hyperpolarizing event which is properly timed with respect to a depolarizing one may facilitate the crossing of the membrane threshold (Mainen & Sejnowsky, 1995; Desmaisons et al. 1999).

Further investigations are now obviously required to decipher how recurrent inhibition influences the firing times of tonically discharging motoneurones and facilitates their synchronization.

Conclusion

The present results suggest that Renshaw cells, besides their subtle stabilizing effects on the motoneurone discharge variability, may contribute to the synchronization of motor unit activity during voluntary muscle contraction. Even when it was enhanced by the pharmacologically induced potentiation of the recurrent inhibition, the degree of synchronization observed in the wrist extensor muscles was quite moderate. In addition, although the synchronization of cortico-spinal neurones is liable to enhance the force output of the innervated motoneurones (Baker et al. 1999), there is no evidence of any such effect when the motoneurones themselves discharge in synchrony (Rack & Westbury, 1969; Yao et al. 2000). No important consequences can therefore be expected to occur in terms of the force output with the moderate increase in synchrony observed under the present experimental conditions. An increase in the motoneurones' synchronization might, however, influence the synaptic effectiveness of the recurrent collaterals, which converge on Renshaw cells and, in turn, potentiate the impact of Renshaw cells on their multiple target neurones (e.g. heteronymous motoneurones, Ia interneurones, gamma motoneurones and ventral spino-cerebellar cells).

Acknowledgments

We are grateful to Dr Jessica Blanc for correcting the English manuscript. This research was supported by grants from the Association Française contre les Myopathies (AFM), the Fondation pour la Recherche Médicale (FRM), the Délégation Générale à l'Armement (DGA), and the Italian MIUR and Istituto Riabilitazione Fisiomedica Loretana (CB, Italy).

references

  1. Adam D, Windhorst U, Inbar GF. The effects of recurrent inhibition on the cross-correlated firing patterns of motoneurones (and their relation to signal transmission in the spinal cord-muscle channel) Biol Cyber. 1978;29:229–235. doi: 10.1007/BF00337280. [DOI] [PubMed] [Google Scholar]
  2. Baker SN, Kilner JM, Pinches EM, Lemon RM. The role of synchrony and oscillations in the motor output. Exp Brain Res. 1999;128:109–117. doi: 10.1007/s002210050825. [DOI] [PubMed] [Google Scholar]
  3. Bussel B, Pierrot-Deseilligny E. Inhibition of human motoneurones, probably of Renshaw origin, elicited by an orthodromic motor discharge. J Physiol. 1977;269:319–339. doi: 10.1113/jphysiol.1977.sp011904. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bremner FD, Barker JR, Stephens JA. Variation in the degree of synchronization exhibited by motor units lying in different finger muscles in man. J Physiol. 1991;432:381–399. doi: 10.1113/jphysiol.1991.sp018390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Calvin WH, Stevens CF. Synaptic noise and other sources of randomness in motoneuron interspike intervals. J Neurophysiol. 1968;31:574–587. doi: 10.1152/jn.1968.31.4.574. [DOI] [PubMed] [Google Scholar]
  6. Cullheim S, Lipsenthal L, Burke RE. Direct monosynaptic contacts between type-identified alpha-motoneurons in the cat. Brain Res. 1984;308:196–199. doi: 10.1016/0006-8993(84)90937-5. [DOI] [PubMed] [Google Scholar]
  7. D'Agostino RB, Pearson ES. Test of departure from normality. Empirical results for the distribution of b2 and √b1. Biometrika. 1973;60:613–622. [Google Scholar]
  8. Davey NJ, Ellaway PH, Baker JR, Friedland CL. Rhythmicity associated with high degree of short-term synchrony of motor unit discharge in man. Exp Physiol. 1993;78:649–661. doi: 10.1113/expphysiol.1993.sp003713. [DOI] [PubMed] [Google Scholar]
  9. De Luca CJ, LeFever RS, McCue MP, Xenakis AP. Control scheme governing concurrently active human motor units during voluntary contractions. J Physiol. 1982;329:129–142. doi: 10.1113/jphysiol.1982.sp014294. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Derfler B, Goldberg LJ. Spike train characteristics of single motor units in the human masseter muscle. Exp Neurol. 1978;61:592–608. doi: 10.1016/0014-4886(78)90026-2. [DOI] [PubMed] [Google Scholar]
  11. Desmaisons D, Vincent JD, Lledo PM. Control of action potential timing by intrinsic subthreshold oscillations in olfactory bulb output neurons. J Neurosci. 1999;19:10727–10737. doi: 10.1523/JNEUROSCI.19-24-10727.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Elble RJ, Randall JE. Motor unit activity responsible for 8 to 12 Hz component of human physiological finger tremor. J Neurophysiol. 1976;39:370–383. doi: 10.1152/jn.1976.39.2.370. [DOI] [PubMed] [Google Scholar]
  13. Ellaway PH. Cumulative sum technique and its application to the analysis of peristimulus time histograms. Electroencephalogr Clin Neurophysiol. 1978;45:302–304. doi: 10.1016/0013-4694(78)90017-2. [DOI] [PubMed] [Google Scholar]
  14. Farmer SF, Bremner FD, Halliday DM, Rosenberg JR, Stephens JA. The frequency content of common synaptic inputs to motoneurones studied during voluntary isometric contraction in man. J Physiol. 1993;470:127–155. doi: 10.1113/jphysiol.1993.sp019851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Garnett R, Stephens JA. The reflex response of single motor units in human first dorsal interosseus muscle following cutaneous afferent stimulation. J Physiol. 1980;303:351–364. doi: 10.1113/jphysiol.1980.sp013290. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gel'fand IM, Gurfinkel VS, Kotz YM, Testlin ML, Shik ML. Synchronization of motor units and associated model concept. Biofizika. 1963;8:475–486. [PubMed] [Google Scholar]
  17. Haase J, Cleveland S, Ross HG. Problems of postsynaptic autogenous and recurrent inhibition in the mammalian spinal cord. Rev Physiol Biochem Pharmacol. 1975;73:73–129. doi: 10.1007/BFb0034660. [DOI] [PubMed] [Google Scholar]
  18. Illert M, Kümmel H. Reflex pathways from large muscle spindle afferents and recurrent axon collaterals to motoneurones of wrist and digit muscles: a comparison in cats, monkeys and humans. Exp Brain Res. 1999;128:13–19. doi: 10.1007/s002210050812. [DOI] [PubMed] [Google Scholar]
  19. Jefferys JGR, Traub RD, Whittington MA. Neural networks for induced ‘40 Hz’ rhythms. Trends Neurosci. 1996;19:202–208. doi: 10.1016/s0166-2236(96)10023-0. [DOI] [PubMed] [Google Scholar]
  20. Jones S, Sudweeks S, Yakel JL. Nicotinic receptors in the brain: correlating physiology with function. Trends Neurosci. 1999;22:555–561. doi: 10.1016/s0166-2236(99)01471-x. [DOI] [PubMed] [Google Scholar]
  21. Katz R, Mazzocchio R, Pénicaud A, Rossi A. Distribution of recurrent inhibition in the human upper limb. Acta Physiol Scand. 1993;149:183–198. doi: 10.1111/j.1748-1716.1993.tb09611.x. [DOI] [PubMed] [Google Scholar]
  22. Katz R, Pierrot-Deseilligny E. Recurrent inhibition in humans. Prog Neurobiol. 1998;57:325–355. doi: 10.1016/s0301-0082(98)00056-2. [DOI] [PubMed] [Google Scholar]
  23. Khatib M, Hilaire G, Monteau R. Excitatory interactions between phrenic motoneurons in the cat. Exp Brain Res. 1986;62:273–280. doi: 10.1007/BF00238846. [DOI] [PubMed] [Google Scholar]
  24. Kirkwood PA, Sears TA. Cross-correlation analyses of motoneurone inputs in coordinated motor act. In: Kruger J, editor. Springer Verlag Series in Synergetics: Neuronal Comparativity. Berlin, Heidelberg: Blackwell Science Inc; 1991. pp. 225–248. [Google Scholar]
  25. Kukulka CG, Clamann HP. Comparison of the recruitment and discharge properties of motor units in human brachial biceps and adductor pollicis during isometric contractions. Brain Res. 1981;219:45–55. doi: 10.1016/0006-8993(81)90266-3. [DOI] [PubMed] [Google Scholar]
  26. Lemon RN. Cortical control of the primate hand. Exp Physiol. 1993;78:263–301. doi: 10.1113/expphysiol.1993.sp003686. [DOI] [PubMed] [Google Scholar]
  27. Loeb GE, Yee WJ, Pratt CA, Chanaud CM, Richmond FJR. Cross-correlation of EMG reveals widespread synchronization of motor units during some slow movements in intact cats. J Neurosci Meth. 1987;21:239–249. doi: 10.1016/0165-0270(87)90119-1. [DOI] [PubMed] [Google Scholar]
  28. Mainen ZF, Sejnowski TJ. Reliability of spike timing in neocortical neurons. Science. 1995;268:1503–1506. doi: 10.1126/science.7770778. [DOI] [PubMed] [Google Scholar]
  29. Maltenfort MG, Heckman CJ, Rymer WZ. Decorrelating actions of Renshaw interneurons on the firing of spinal motoneurons within a motor nucleus: a simulating study. J Neurophysiol. 1998;80:309–323. doi: 10.1152/jn.1998.80.1.309. [DOI] [PubMed] [Google Scholar]
  30. Matthews PBC. Relationship of firing intervals of human motor units to the trajectory of post-spike after-hyperpolarisation and synaptic noise. J Physiol. 1996;492:597–628. doi: 10.1113/jphysiol.1996.sp021332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mazzocchio R, Rossi A. Further evidence for Renshaw cell inhibition in man. A combined electrophysiological and pharmacological approach. Neurosci Lett. 1989;106:131–136. doi: 10.1016/0304-3940(89)90214-0. [DOI] [PubMed] [Google Scholar]
  32. Mazzocchio R, Rossi A. A method for potentiating Renshaw cell activity in human. Brain Res Prot. 1997;2:53–58. doi: 10.1016/s1385-299x(97)00028-7. [DOI] [PubMed] [Google Scholar]
  33. Moore GP, Segundo JP, Perkel DH, Levitan H. Statistical signs of synaptic interaction in neurons. Biophys J. 1970;10:876–900. doi: 10.1016/S0006-3495(70)86341-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nordstrom MA, Miles TS. Discharge variability and physiological properties of human masseter motor units. Brain Res. 1991;541:50–56. doi: 10.1016/0006-8993(91)91072-9. [DOI] [PubMed] [Google Scholar]
  35. Oda Y, Nakanishi I. The distribution of cholinergic neurons in the human central nervous system. Histol Histopathol. 2000;15:825–834. doi: 10.14670/HH-15.825. [DOI] [PubMed] [Google Scholar]
  36. Person RS, Kudina LP. Discharge frequency and discharge pattern of human motor units during voluntary contraction of muscle. Electroencephalogr Clin Neurophysiol. 1972;32:471–483. doi: 10.1016/0013-4694(72)90058-2. [DOI] [PubMed] [Google Scholar]
  37. Rack PMH, Westbury DR. The effect of length and stimulus rate on tension in the isometric cat soleus muscle. J Physiol. 1969;204:443–460. doi: 10.1113/jphysiol.1969.sp008923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Renshaw B. Central effects of centripetal impulses in axons of spinal ventral roots. J Neurophysiol. 1946;9:191–204. doi: 10.1152/jn.1946.9.3.191. [DOI] [PubMed] [Google Scholar]
  39. Rosenberg JR, Amjad AM, Breeze P, Brillinger DR, Halliday DM. The fourier approach to the identification of functional coupling between neuronal spike trains. Prog Biophys Mol Biol. 1989;53:1–31. doi: 10.1016/0079-6107(89)90004-7. [DOI] [PubMed] [Google Scholar]
  40. Ross HG, Cleveland S, Haase J. Contribution of single motoneurons to Renshaw cell activity. Neurosci Lett. 1975;1:105–108. doi: 10.1016/0304-3940(75)90053-1. [DOI] [PubMed] [Google Scholar]
  41. Rossi A, Mazzocchio R. Renshaw recurrent inhibition to motoneurones innervating proximal and distal muscles of the human upper and lower limbs. In: Jami L, Pierrot-Deseilligny E, Zytnicki D, editors. Muscle Afferents and Spinal Control of Movement. Oxford: Blackwell Science Inc; 1992. pp. 313–319. [Google Scholar]
  42. Schmied A, Vedel JP, Pagni S. Human spinal lateralization assessed from motoneurone synchronization: dependence on handedness and motor unit type. J Physiol. 1994;480:369–387. doi: 10.1113/jphysiol.1994.sp020367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Schmied A, Vedel JP, Pouget J, Forget R, Lamarre Y, Paillard J. Changes in motoneurone connectivity assessed from neuronal synchronization analysis. In: Taylor A, Gladden M, Durbaba R, editors. Alpha and Gamma Motor Systems. New York: Blackwell Science Inc; 1995. pp. 469–477. [Google Scholar]
  44. Sears TA, Stagg D. Short term synchronization of intercostal motoneurone activity. J Physiol. 1976;263:357–381. doi: 10.1113/jphysiol.1976.sp011635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tokizane T, Shimazu H. In: Functional Differentiation of Human Skeletal Muscle. Corticalization and Spinalization of Movement. Thomas CC, editor. Springfield IL USA: 1964. p. 66. [Google Scholar]
  46. Türker KS, Powers RK. Effects of common excitatory and inhibitory inputs on motoneuron synchronization. J Neurophysiol. 2001;86:2807–2822. doi: 10.1152/jn.2001.86.6.2807. [DOI] [PubMed] [Google Scholar]
  47. Türker KS, Schmied A, Gheng HB. Correlated changes in the firing rate of human motor units during voluntary contraction. Exp Brain Res. 1996;111:455–464. doi: 10.1007/BF00228735. [DOI] [PubMed] [Google Scholar]
  48. Van Keulen LCM. Autogenetic recurrent inhibition of individual spinal motoneurons of the cat. Neurosci Lett. 1981;21:297–300. doi: 10.1016/0304-3940(81)90220-2. [DOI] [PubMed] [Google Scholar]
  49. Whittington MA, Traub RD, Kopell N, Ermentrout B, Buhl EH. Inhibition-based rhythms: experimental and mathematical observations on network dynamics. Intern J Psychophysiol. 2000;38:315–336. doi: 10.1016/s0167-8760(00)00173-2. [DOI] [PubMed] [Google Scholar]
  50. Windhorst U. Activation of Renshaw cells. Prog Neurobiol. 1990;35:135–179. doi: 10.1016/0301-0082(90)90020-h. [DOI] [PubMed] [Google Scholar]
  51. Windhorst U. On the role of recurrent inhibitory feedback in motor control. Prog Neurobiol. 1996;49:517–587. doi: 10.1016/0301-0082(96)00023-8. [DOI] [PubMed] [Google Scholar]
  52. Yao W, Fuglevand AJ, Enoka RM. Motor unit synchronization increases EMG amplitude and decreases force steadiness of simulated contractions. J Neurophysiol. 2000;83:441–452. doi: 10.1152/jn.2000.83.1.441. [DOI] [PubMed] [Google Scholar]
  53. Zar JH. Biostatistical Analysis. 3. NJ, USA: Englewood Cliffs; 1996. p. 878. [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES