Springer-Verlag 1998
Exp Brain Res (1998) 120:193±201
RESEARCH ARTICLE
U. SøawinÂska ´ F. TycÆ ´ S. Kasicki R. Navarrete ´ G. Vrbovµ
Time course of changes in EMG activity of fast muscles after partial denervation
Received: 17 September 1996 / Accepted: 3 November 1997
Abstract After partial denervation, the remaining motor units (MUs) of adult fast extensor digitorum longus muscle (EDL) expand their peripheral field. The time course of this event was studied using tension measurement and recordings of electromyographic (EMG) activity. The results show that after section of the L4 spinal nerve, when only 5.3 0.63 of the 40 MUs normally supplying EDL muscle remain, the force of individual motor units starts to increase between the 1st and 2nd week after the operation and continues to do so for a further week. The drastic reduction of the number of motoneurones supplying the fast EDL leads to an increase in activity of the remaining MUs. In the 1st week after partial denervation, there was a sharp increase in the EMG activity of remaining motor units. During the next 12 days, this increase became less marked, but EMG activity remained nevertheless significantly higher than that of the unoperated EDL muscle. Many MUs became tonically active during posture. The EMG activity pattern during locomotion was also altered, so that the burst duration was positively correlated with the step cycle duration. Moreover, shortly after partial U. SøawinÂska ´ F. TycÆ1 ´ S. Kasicki2 ´ R. Navarrete3 Nencki Institute of Experimental Biology and Institute of Biocybernetics and Biomedical Engineering, 3 Pasteura Str., 02-093 Warsaw, Poland G. Vrbovµ Department of Anatomy and Developmental Biology, University College London, Gower Str., London WC1E 6BT, UK
)
U. SøawinÂska ( ) Department of Neurophysiology, Nencki Institute of Experimental Biology, 3 Pasteura Str., 02-093 Warsaw, Poland e-mail:
[email protected], Fax: +48-22-82 53 42 Present addresses: Laboratory of Neuro Muscular Physiology, University Littoral, 17, ave Bleriot-BP 699, F-62228 Calais Cedex, France 2 Department of Neurophysiology, Nencki Institute of Experimental Biology, 3 Pasteura Str., 02-093 Warsaw, Poland 3 Department of Anatomy and Cell Biology, Charing Cross and Westminster Medical School, University of London, Fulham Palace Road, London W6 8RF, UK 1
denervation, the interlimb coordination was disturbed but returned to its original symmetrical use 1±2 weeks later. Key words EMG ´ Motor unit activity ´ Partial denervation ´ Interlimb coordination ´ Rat
Introduction After removal of a portion of the motor supply to a skeletal muscle, the remaining motoneurones enlarge their peripheral field by sprouting and reinnervation of the denervated muscle fibers. This process of sprouting involves the growth of branches from the remaining motor axons and the establishment of new neuromuscular contacts (Brown et al. 1981). Most studies that have investigated changes after partial denervation of adult skeletal muscles have dealt with alterations of the motor unit (MU) territory and histochemical properties of the muscle fibres (Kugelberg et al. 1970; Rafuse and Gordon 1996a; Vrbovµ et al. 1995). Expansion of MU territory after partial denervation has been found to occur in both slow and fast muscles in adult and neonate rats and cats (Connold et al. 1992; Fisher et al. 1989; Rafuse and Gordon 1996b; TycÆ and Vrbovµ 1995). In addition to the change in MU territory, the muscle fibre characteristics in partially denervated rat skeletal muscle are changed so that these muscles contain a larger proportion of slow muscle fibres (Connold et al. 1992; Fisher et al. 1989; TycÆ and Vrbovµ 1995). Since it is well established that increased activity leads to a transformation of type II to type I muscle fibres (Pette and Vrbovµ 1985), it might be that the transformation of muscle fibers from fast to slow type seen after partial denervation is induced by increased activity of the remaining MUs. Indeed, after partial denervation of the neonatal rat soleus muscle, the remaining MUs of the soleus muscle are more active (SøawinÂska et al. 1995). A much greater increase in MU activity is seen when the fast extensor digitorum longus (EDL) muscle is partially denervated shortly after birth (TycÆ and Vrbovµ 1995). Although in-
194
creased electromyographic (EMG) activity in muscles from patients that suffered loss of motoneurones due to disease has been reported (Dietz et al. 1986), it is not clear whether a similar increase in EMG activity occurs in normal individuals that lose part of the muscles innervation as a result of loss of axons. This question is addressed in the present study. Another problem that can be investigated using the model of partial denervation concerns the effects of perturbing the function of the muscle on the locomotor pattern. The coordination of the activity of individual muscles into a locomotor synergy is executed in the spinal cord via the central pattern generators (CPGs), assisted by spinal reflex mechanisms (Armstrong 1986; Grillner 1975, 1981). CPGs are responsible for interlimb coordination and sequential activation of limb muscles. Thus, individual muscles have their own characteristic duty cycles relative to the movement sequence. EDL muscle is one of the ankle dorsiflexors that is activated during the swing phase of the step cycle. In this muscle EMG activity is present near the end of the swing phase (Goudard et al. 1992; NicolopoulosStournaras and Iles 1984). The investigation of the phase relationships between the activity of the EDL muscles of both hindlimbs allows us to determine the details of the hindlimb coordination during movements generated by the CPG (Cazalets at al. 1990; Cohen and Gans 1975; Hruska et al. 1979). Whether this coordination is perturbed by partial denervation of one hindlimb is studied here. Thus, in this study, the effects of partial denervation on the activity of the remaining motor units to the EDL muscle of adult animals is investigated by assessing: (a) the time course of MU expansion, (b) the changes of the activity pattern of the remaining EDL MUs during spontaneous exploratory behaviour as well as during locomotor activity, and (c) the time course of the changes in the EMG activity patterns of the EDL muscle of both hindlimbs.
Fig. 1 Records of single twitches of extensor digitorum longus muscle (3 weeks after partial denervation) in response to stimulation of its motor nerve with increasing stimulus intensity. The stepwise increments of tension indicate the recruitment of additional motor axons. Calibration 20 ms, 10 g
Recordings of contraction Isometric tension recordings were carried out at intervals varying from 1 to 3 weeks after partial denervation. A total of 14 animals were studied at: (1) 1 week (n = 5), (2) 2 weeks (n = 5), and (3) 3 weeks after partial denervation (n = 4). With the animals under chloral hydrate anaesthesia (4.5%, 1 ml/100 g body weight), the distal tendons of the EDL muscles of both legs were dissected free and attached to strain gauges (Dynamometer UFI, Devices). Isometric contractions were elicited by electrical stimulation of the cut end of the motor nerve supplying the muscle via bipolar silver electrodes. The muscle tension was displayed on an oscilloscope screen (Tektronix R5113). The muscle length was adjusted to obtain maximum twitch tension at the supramaximal stimulus intensity. To estimate the number of MUs in the partially denervated and the contralateral EDL muscles, the motor nerve was stimulated by single pulses of 20 ms duration every 4 s. The stimulus strength was gradually increased to obtain stepwise increments of twitch tension, as individual motor axons became recruited. An example from such a recording is shown in Fig. 1. The number of stepwise increments was an indicator of the number of motor axons to the EDL muscle present in the nerve. The mean MU tension was obtained by dividing the maximum tetanic tension of the muscle by the number of MUs identified in it. The mean value calculated for the partially denervated muscle was expressed as a percentage of the mean value obtained for the contralateral control muscle. EMG recordings
Materials and methods Surgical procedures Fourteen male Wistar rats (3±4 months old) were used in these experiments. The EMG activity of EDL muscles in both hindlimbs was recorded using bipolar electrodes. The electrodes were made of multistranded, Teflon-coated stainless steel wire (7SS-2T; Clark Electromedical Instruments, UK) enclosed in silicone tubing and provided with a multipin connector reinforced with dental cement (Austenal Dental Products) and covered with a silastic, adhesive silicone (3140 RTV, Dow Corning; Hnik et al. 1978; SøawinÂska et al. 1995). Implantation of electrodes for chronic EMG recordings into EDL muscles was performed under halothane (2% in O2) anaesthesia using sterile precautions. The bared loops of the electrodes (with about 100 mm of the insulation removed as a recording surface) were led under the skin and sutured to the belly of the muscle, and the connector was secured to the back of the animal. Three to four days after the implantation of EMG electrodes, the animals were again anaesthetized with halothane. Using sterile precautions, the L4 spinal nerve on the left side of the animal was carefully exposed and a 1- to 2-mm section of the nerve was removed. Care was taken to avoid damaging the L5 spinal nerve. In all experiments the contralateral side was used as a control.
EMG signals were amplified using a Tektronix 2A61 differential amplifier, filtered (band pass 0.1±1 kHz) and recorded on a tape recorder (RACAL 4DS). The EMG activity was recorded in awake, freely moving rats during: (1) spontaneous exploratory activity in a cage, (2) spontaneous locomotion on a straight runway (1.5 m long and 5 cm wide) leading to a goal box, and (3) reflex responses to passive ankle dorsiflexion and plantar flexion. The recording of the EMG activity of EDL muscles in both hindlimbs began 2 days before and lasted until 3 weeks after partial denervation. Analysis of EMG activity The EMG activity of the same portions of EDL muscles were recorded before and after injury. Aggregate activity In order to provide a quantitative estimation of overall muscle activity, the aggregate EMG activity of EDL muscles of both hindlimbs was recorded during spontaneous exploratory behaviour. The aggregate EMG activity of each muscle was defined as a number of po-
195
Fig. 2 Rectified, integrated and averaged EMG activity of normal extensor digitorum longus (EDL) muscles of both hindlimbs (l left; r right). The lines show how the step cycle duration (CD), burst duration (BD), and onset time of burst activity (OT) were estimated tentials higher than the threshold close to the noise level (usually less than 50 mV). In each recording session, the aggregate EMG activity was measured during 10-min intervals. The results were expressed as mean counts per minute. Detection of crossings of the threshold by EMG signal was done using a spike trigger (NL 200, Neurolog system; Digitimer, UK). The analysed signal and triggered pulses were monitored on an oscilloscope and counted by a computer using a software package (MRATE program; Cambridge Electronic Design, UK). The threshold for the spike trigger was kept at the same level for the analysis of all recordings from the same animal before and after partial denervation. EMG pattern during locomotion In order to determine interlimb coordination, the EMG activity of EDL muscle of both hindlimbs was recorded during locomotion. The EMG recordings were rectified and integrated off-line (time constant 5 ms) and, after digitization at a sampling frequency of 400 Hz, portions of the recording were stored in the PC computer. Only these parts of the EMG activity were chosen for further analysis that were recorded during regular locomotion. The following parameters describing locomotor movements were assessed semiautomatically using software written in our laboratory (SøawinÂska et al. 1995): (1) the burst duration (BD) ± defined as the time between the beginning and the end of an EMG burst, (2) the step cycle duration (CD) ± defined as the time elapsed between consecutive EMG bursts of the same muscle, (3) the onset time of the EMG burst activity (OT) ± measured as the occurrence of the beginning of the EMG burst of one EDL muscle in respect to the beginning of the step cycle of the contralateral muscle; the onset time indicates the phase shift between two EDL muscles of both hindlimbs. Figure 2 shows how the OT, BD and CD parameters were defined. The relations between these parameters and the step cycle duration (i.e. the relation BD vs CD of the same hindlimb, and the phase shift reflected as OT of burst activity of one limb vs CD of the contralateral hindlimb) were examined using the method of linear regression (least-square method). The coefficients of the regression function were compared using the analysis of variance. The mean durations were compared using Students t-test.
Results Time course of changes of motor unit territory Maximal tetanic tension developed by the partially denervated and control EDL muscles was recorded at weekly intervals for 3 weeks after operation. In addition the numbers of MUs for each muscle was also determined. Figure 3A shows that there was no significant difference in the number of MUs found in EDL muscles 1, 2 or 3 weeks after partial denervation. The mean number of MUs left
Fig. 3A±C The motor unit numbers (A), tetanic tension (B) and tetanic tension per motor unit (C) of partially denervated EDL muscle expressed as a percentage of the contralateral muscle. The values were obtained at weekly intervals after partial denervation. Note that the force per motor unit is greater only 2 and 3 weeks after the operation (C). Bars SEM
after partial denervation was 5.3 0.63, 5.4 1.2 and 5.5 1.2, respectively, (mean SEM), showing that the number of MUs did not change during the first 3 weeks after partial denervation. The numbers of MUs found here are consistent with those present in partially denervated EDL muscles 2±9 months after the operation (Connold et al. 1992; TycÆ and Vrbovµ 1995). Since the normal EDL muscle has 40 MUs (Albani et al. 1988; Balice-Gordon and Thompson 1988; Close 1967) it appears that after section of the L4 spinal nerve only 13% of motor axons innervate the EDL muscle. One week after partial denervation, the maximal tetanic tension, expressed as a percentage of the maximal tetanic tension of the contralateral muscle, was 9.82 2.09% ( SEM), at 2 weeks this value was 19.75 5.12% and, by the third week, it reached 25.2 8.45% (Fig. 3B). Since the number of MUs in each group was the same, the increase in force production of the whole muscle is due to the greater force developed by each individual MU within the partially denervated muscles. One week after partial denervation, the mean MU tension (i.e. the maximal tetanic tension divided by the num-
196
Fig. 4A, B An example of raw EMG activity recorded during plantar flexion (P) of the foot. A Before partial denervation; B 2 weeks after partial denervation. Note atypical tonic activity of the motor unit in the partially denervated EDL muscle, turned off by dorsiflexion. Calibration 1 s, 0.5 mV (EDLl/r EDL muscle of left/right hindlimb, op partially denervated)
ber of MUs identified in the same muscle) was slightly less than that in the control EDL (84.8 18.6%). By 2 weeks after partial denervation, the mean MU tension was significantly greater (P < 0.01) in the partially denervated EDL than that of the control muscle (143 14.1%) and, by 3 weeks, it reached 194.4 67.1% of the control (Fig. 3C). This final value is close to that obtained in animals partially denervated at 18 days of age and investigated 2±9 months later (Connold et al. 1992; TycÆ and Vrbovµ 1995). Thus, by the 3rd week after partial denervation, the MUs reached their maximum expansion. General observations on EMG activity After partial denervation, the activity pattern of EDL muscle was altered. During quiet standing, the normal EDL muscle remained silent, while tonic firing of several MUs was often seen in the partially denervated EDL muscle. When the rats were held suspended in the air, the partially denervated muscles were active, while the control muscle under such condition was only sporadically active. In normal rats, passive plantar flexion elicits a short phasic burst of EMG activity in EDL muscle, whereas, in the partially denervated EDL muscles, plantar flexion elicited tonic firing (Fig. 4). Partial denervation induced transient changes in the pattern of EMG activity of the EDL muscles during locomotion. On the 1st day after spinal nerve section, the animals avoided using their injured hindlimb. In general, the EMG burst duration was longer than that seen in normal animals. In some cases the muscle was activated twice during one step cycle. Both phenomena could be observed in the same animal (even during the same run), and there was a continuum between these two patterns of muscle activity. Time course of changes of EMG activity in partially denervated EDL muscles Aggregate activity Figure 5A summarizes the results of aggregate EMG activity recorded from the control and partially denervated
Fig. 5 A Total aggregate EMG activity for the partially denervated (op) and contralateral unoperated (con) EDL muscles before and at different intervals after partial denervation. B Mean aggregate EMG activity calculated per motor unit for the partially denervated (op) and contralateral unoperated (con) EDL muscles before and at different intervals after partial denervation. Each point shows the mean number of muscle potentials above the threshold level, expressed as a number of counts measured per minute. When not shown, the SEM was smaller than the size of the corresponding symbol
EDL muscles during spontaneous activity in the home cage before and at different intervals after partial denervation. A sudden sharp increase in aggregate activity of partially denervated EDL muscles was observed 2±6 days after the operation (t-test, P < 0.001). This very high activity decreased by 7 days after partial denervation but was then followed by a gradual increase in aggregate activity, which peaked at 10±12 days and remained significantly higher than that of the unoperated EDL muscle (Fig. 5A). It is interesting that this phase of increased activity just preceded the first signs of increase in muscle force and MU expansion illustrated in Fig. 3C. In order to obtain an index of the overall aggregate activity per MU in the partially denervated or unoperated EDL muscles, the mean aggregate EMG activity established for each muscle was divided by the number of MUs counted in the terminal experiment. Figure 5B summarizes the values of the mean aggregate activity per MU. At all times studied, these values were significantly higher for the partially denervated EDL muscle (t-test, P < 0.0001). The time course of the changes illustrated in Fig. 5B follows closely that seen in Fig. 5A, but the extent of the increased activity is greater when expressed per MU. Thus, MUs of the partially denervated EDL muscle develop an increase in aggregate activity.
197 Fig. 6A±D Records of EMG activity from EDL muscles obtained during locomotion in a control rat (A) and at various postoperative intervals: B 1 day after and C 1 week after partial denervation. The plots on the right show the relation between the burst duration (BD) and step cycle duration (CD) for both EDL muscles during locomotion in a control animal and at various postoperative intervals. Calibration 1 s, 0.5 mV (EDLl/r EDL muscle of left/right hindlimb, op partially denervated, con contralateral control)
Activity during locomotion The EDL muscle belongs to the group of dorsiflexor muscles of the ankle joint. During standing, the EDL muscle is seldom activated and its EMG signal is almost flat. When the rat is walking, the rhythmic EMG activity related to the swing phase of the step cycle can be seen (Fig. 6A). Analysis of the relationship between the BD and the CD in normal adult animals confirmed that the BD of the normal EDL muscle is not correlated to the CD, i.e. no matter how fast the animal walks, the BD remains constant (Fig. 6A, right). The EDL muscles from both hindlimbs have similar BD, and the estimated regression lines of the relation of BD against CD were very similar. The analysis of variance did not show a significant difference between the regression coefficients. The 1st day after partial denervation the rat avoided the use of the partially denervated hindlimb. Nevertheless, rhythmic EMG activity was observed during locomotion in both EDL muscles (Fig. 6B). The BD in the partially denervated EDL was longer than that in the contralateral unoperated muscle (t-test, P < 0.001), and this difference was more pronounced during longer step cycles, i.e. during slow locomotion (Fig. 6B, right). Five days after partial denervation, the rats used their partially denervated leg in quadrupedal locomotion. The EMG activity of the EDL muscles during locomotion is shown in Fig. 6C. The analysis of EMG recordings showed that the BD of the partially denervated EDL muscle was longer than in the contralateral unoperated muscle. Moreover, unlike in the normal EDL muscle, the BD of the partially denervated EDL muscle was correlated with CD (r = 0.66), while BD of the unoperated EDL muscle was constant for various CD (see Fig. 6C, right). The analysis of variance of estimated regression lines
showed that the relationships between BD and CD, established for partially denervated and contralateral muscles, were significantly different (P < 0.001). Phase relationship between EDL muscle activity from both hindlimbs Normal animals The phase relationship between EDL muscle activity in both hindlimbs was obtained by investigating the onset time (OT) of the EMG burst of one EDL muscle with respect to the CD of the contralateral hindlimb during regular locomotion (see Fig. 2 in Materials and methods section). The block diagram in Fig. 7 shows the distribution of the phase shifts (the OT of one EDL muscle expressed as a percentage of the CD of the contralateral hindlimb) in two normal rats. In the majority of the animals studied (four of six rats), the EDL burst started approximately midway through the step cycle of the contralateral hindlimb, such that most values of OT were distributed around the midpoint. The histograms showing the results of this analysis on EDL muscles of both hindlimbs are very similar (Fig. 7A). This shows that in normal animals there is a phase shift of about 0.5 between EDL burst activity in both hindlimbs. The mean OT expressed as a percentage of CD varied from 54.6 1.3% to 55.4 1.9% in different animals. Therefore, in most animals the EDL muscles were activated in a bilaterally symmetrical way. However, in some animals (two of six), analysis of the phase relationships showed that the left and right EDL muscles were not activated in a fully symmetrical way (Fig. 7B). The mean OT expressed as a percentage of CD varied from 44.2 0.9% to 62.7 0.8% in both legs. This difference is also seen in Fig. 7B, right, where OT is
198
plotted against CD. When the OT was plotted against the CD, the estimated regression lines were similar for both EDL muscles in four of the six animals studied (see example of one experiment in Fig. 7A, right), and analysis of variance did not show a significant difference between the regression coefficients. In two other animals, the coefficients of estimated regression lines of OT vs CD relation were significantly different (P < 0.01). Thus, in most normal animals the activity pattern of EDL shows bilateral symmetry, while in a few rats a slightly asymmetrical pattern can be observed during locomotion throughout the whole spectrum of different speeds. Partial denervation
Fig. 7A, B Phase relationships of EDL burst activity of both hindlimbs. Left The distribution of the time when EMG bursts in one hindlimb occur in relation to the step cycle of contralateral hindlimb in two normal rats. Empty circles and filled circles below each histogram indicate the mean value. Right The relationships between onset time of the EMG burst and the step cycle of the contralateral hindlimb. The numbers indicate the number of step cycles of the left hindlimb (nl) and of the right hindlimb (nr) taken for the analysis
Fig. 8A±C Left The distribution of the time when EMG burst in one hindlimb occurs in relation to the step cycle of contralateral hindlimb at various postoperative intervals (A 1 day; B 1 week; C 2 weeks after partial denervation). Empty circles and filled circles below each histogram indicate the mean value of the presented data. Right The relationship between the onset time of the EMG burst and the step cycle of the contralateral hindlimb (A 1 day; B 1 week; C 2 weeks after partial denervation). The numbers (n) indicate the number of step cycles taken for the analysis (op partially denervated, con contralateral control)
One day after the operation, distinct difference in the phase relationship between the partially denervated and contralateral EDL muscle was observed. Figure 8A±C shows histograms of the phase shifts in EDL muscles. In the diagrams (on the right of Fig. 8A±C), the relationships between OT and CD are plotted. These examples are taken from one animal at various postoperative intervals. Figure 8A shows that, 1 day after partial denervation, the OT of the control EDL muscle was shortened (for different animals within the range 39.65 2.3% to
199
29.87 2.1%). This shortening is due to the fact that the EMG burst of the control EDL occurred soon after the beginning of the step cycle of the contralateral partially denervated EDL muscle. In the partially denervated leg, the activity of EDL muscle started later with respect to the beginning of the step cycle of the control, so that the OT was delayed (from 64.31 3.5% to 72.29 1.97% in individual rats). The regression lines in Fig. 8A, right, illustrate the asymmetry between both hindlimbs. The coefficients of estimated regression lines of the phase relationships for the partially denervated and control EDL muscles were significantly different (analysis of variance, P < 0.001). Thus, the symmetrically alternating pattern of EDL activity typical for locomotion in normal animals was altered 1 day after partial denervation. Interestingly, in spite of the later onset of the burst activity, the partially denervated EDL muscle was activated for a longer time (t-test, P < 0.001; Fig. 6B, right). This may prolong the swing phase of the partially denervated hindlimb. These results are consistent with our observation that 1 day after partial denervation the animals avoid using the denervated hindlimb during locomotion. One to three weeks after partial denervation the situation changed. The histograms of phase relationships of both EDL muscle taken 1 week after partial denervation, in most of the animals (four of six rats), were only slightly different (Fig. 8B). Between the 1st and 2nd week after partial denervation, the phase relationships in two EDL muscles became fully symmetrical, and histograms from both EDL muscles were very similar (Fig. 8C). The analysis of variance did not show a significant difference between the regression coefficients of the phase relationships for EDL muscle activity in both hindlimbs. Thus during this time the coordination between the two hindlimbs regained its symmetrical pattern.
Discussion Following partial denervation of adult skeletal muscles, the remaining motor axons increase their territory by sprouting, and compensate in this way for the loss of innervation (Brown et al. 1981; Connold et al. 1992; Thompson and Jansen 1977; TycÆ and Vrbovµ 1995). The present results show that the process of recovery of muscle force achieved by nerve sprouts begins 1 week after injury and continues for a further 2 weeks. By 3 weeks the mean tetanic tension per MU is considerably increased and is comparable with that seen after longer periods of recovery (Connold et al. 1992; TycÆ and Vrbovµ 1995). These findings confirm observation that the remaining MUs after partial denervation become larger within a relatively short period of time (Brown and Ironton 1978). Although there are several reports on functional changes of partially denervated muscles in rats (Brown et al. 1981; Gardiner et al. 1987; Kugelberg et al. 1970), none of these studied changes in EMG activity in freely moving animals. Our present results show that during the early period of time the activity
of MUs changed after partial denervation, and persistent EMG activity of partially denervated EDL muscle was recorded during standing or when the rat was suspended in the air. It could be argued that the L5 spinal nerve (which was now the only nerve supply to EDL muscle) contains mainly axons to slow MUs and that the continuous activity in the partially denervated EDL muscle reflects this situation. However, this is unlikely, for the normal EDL, which also contains these axons, is silent during rest or when the animal is suspended. Moreover, the time course of the twitch contraction shortly after section of the L4 spinal nerve is fast (see Fig. 1). The possibility that slow units sprout more than fast ones and that the increased activity is caused by this is unlikely, for the present result, which shows that 1 week after the operation, when there is no functional evidence of sprouting, the EMG activity is already increased. It is possible that transient polyneuronal innervation could contribute to the changes of EMG activity. However, even if this was the case, it is difficult to understand why this should be apparent mainly during rest. The early increase in activity seen here could be caused by postoperative pain. This is difficult to assess in our experimental conditions, but the operation that removed the L4 spinal nerve also reduced segmental sensory input, and the rats showed no obvious signs of discomfort. While unlikely, it is not impossible that some of the increase in EMG activity at rest is caused by inappropriate afferent inputs. The present results confirmed previous observations (Goudard et al. 1992; Vejsada et al. 1991) that flexor burst duration of dorsiflexor muscles of the ankle in intact rats does not vary with the CD. Our results are also consistent with the findings that the alternating hindlimb movement is a typical locomotor pattern observed in overground locomotion in rats (Goudard et al. 1992; Hruska et al. 1979; Nicolopoulos-Stournaras and Iles 1984). Nevertheless, some normal rats (two of six) exhibited a slightly asymmetrical pattern (about 10%) when walking or trotting. It is interesting that a similar ratio of rats with such a pattern of limb movements during locomotion was observed when using another method, which used contact electrodes placed on paws to record the stance duration (Zmysøowski et al. 1995). However, one cannot exclude the possibility that the slight asymmetry in interlimb coordination, observed in some intact rats might be caused by indwelling electrodes. Immediately after partial denervation, the animals avoided the use of their operated leg during locomotion, but the EMG recordings in the partially denervated EDL muscle nevertheless showed bursts of activity. The phase relationship between both EDL bursts was altered and the coordination of both hindlimbs became asymmetrical. It is known that locomotion and closely related rhythmic patterns are generated centrally in the spinal cord even in the absence of afferent feedback (Delcomyn 1980; Grillner 1975; Grillner and Zangger 1984; Lundberg 1981), although they are influenced by such feedback (Abraham and Loeb 1985; Forssberg 1979; Forss-
200
berg et al. 1977; Guertin et al. 1995; Perreault et al. 1995; Prochazka et al. 1978; Rossignol et al. 1988). Thus, the asymmetrical pattern of EMG activity of EDL muscles observed in our experiments shortly after partial denervation might be related to the reduction of segmental afferents in addition to the loss of efferent axons caused by section of the L4 spinal nerve. The asymmetries in hindlimb movements observed during locomotion might be associated with decreased or abnormal sensations (including pain). However, no signs of such sensation were noticed when handling the animals. Even during the 1st day after partial denervation, the rats did not show any signs of distress such as licking, biting or screaming, which are considered as painrelated behaviours (Detweiler et al. 1995; Kim at al. 1995; Mosconi and Kruger 1996). Seven days later, the hindlimb coordination gains almost its original symmetrical pattern. This implies that the network of neurones together with the remaining afferent and efferent systems now can control appropriately the activity bursts of EDL muscles. Thus, in our experiments the organization of the locomotor pattern after a 2-week period of disturbance caused by partial denervation returned to its original, strict symmetrical regime. Nevertheless, the altered temporal pattern of EDL muscle activity persisted so that the burst duration became positively correlated with the step cycle. Thus, in addition to the tonic component seen in partially denervated phasic muscles during standing, the phasic pattern of activity was modified during locomotion. In comparison with the partially denervated slow muscles (SøawinÂska et al. 1995), the activity of motoneurones to the fast EDL muscle was more affected by partial denervation. This indicates that phasic activity is more perturbed by partial denervation than postural activity and seems to be consistent with the nature of postural function, which, although less precise, is involved in all motor functions. Therefore, it could be that the postural system is less sensitive to perturbance, whereas the system of more precise phasic activity can be readily modified. Acknowledgements We are grateful to the Wellcome Trust and the MNDA for supporting this work. Part of the EMG analysis was carried out using equipment supplied by the Foundation For Polish Science ± Brain/31/94. It is a pleasure to thank Debbie Bartram and Jim Dick for their help.
References Abraham LD, Loeb GE (1985) The distal hindlimb musculature of the cat: patterns of normal use. Exp Brain Res 58:580±593 Albani M, Lowrie MB, Vrbovµ G (1988) Reorganization of motor units in reinnervated muscles of the rat. J Neurol Sci 88:195± 206 Armstrong DM (1986) Supraspinal contribution to the initiation and control of locomotion in the cat. Prog Neurobiol 26:273±361 Balice-Gordon J, Thompson WJ (1988) Synaptic rearrangements and alterations in motor unit properties in neonatal rat extensor digitorum longus muscle. J Physiol (Lond) 398:191±210
Brown MC, Ironton R (1978) Sprouting and regression of neuromuscular synapses in partially denervated mammalian muscles. J Physiol (Lond) 278:325±348 Brown MC, Holland RL, Hopkins WG (1981) Motor nerve sprouting. Annu Rev Neurosci 4:17±42 Cazalets JR, Menard J, CrØmieux J, Clarac F (1990) Variability as a characteristic of immature motor systems: an electromyographic study of swimming in the newborn rat. Behav Brain Res 40:215±225 Close R (1967) Properties of motor units in fast and slow skeletal muscles of the rat. J Physiol (Lond) 193:45±55 Cohen AH, Gans C (1975) Muscle activity in rat locomotion: movement analysis and electromyography of the flexors and extensors of the elbow. J Morphol 146:177±196 Connold AL, Fisher TJ, Maudarbocus S, Vrbovµ G (1992) Response of developing rat fast muscles to partial denervation. Neuroscience 46:981±988 Delcomyn F (1980) Neural basis of rhythmic behavior in animals. Science 210:492±498 Detweiler DJ, Rohde DS, Basbaum AI (1995) The development of opioid tolerance in the formalin test in the rat. Pain 63:251±254 Dietz V, Hollerbach S, Berger W (1986) Functional stretch reflex in neuromuscular disease. In: Dimitrijevic MR, Kabulas BA, Vrbovµ G (eds) Recent achievements in restorative neurology. Progressive neuromuscular diseases. Karger, Basel, pp 53±64 Fisher TJ, Vrbovµ G, Wijetunge A (1989) Partial denervation of the rat soleus muscle at two different developmental stages. Neuroscience 28:755±763 Forssberg H (1979) Stumbling corrective reaction: a phase-dependent compensatory reaction during locomotion. J Neurophysiol 42:936±953 Forssberg H, Grillner S, Rossignol S (1977) Phasic gain control of reflexes from the dorsum of the paw during spinal locomotion. Brain Res 132:121±139 Gardiner P, Michel R, Olha A, Pettigrew (1987) Force and fatiguability of sprouting motor units in partially denervated rat plantaris. Exp Brain Res 66:597±606 Goudard I, Orsal D, Cabelguen JM (1992) An electromyographic study of the hindlimb locomotor movements in the acute thalamic rat. Eur J Neurosci 4:1130±1139 Grillner S (1975) Locomotion in vertebrates: central mechanism and reflex interaction. Physiol Rev 55:247±304 Grillner S (1981) Control of locomotion in bipeds, tetrapods and fish. In: Brookhart JM, Mountcastle VB (eds) Motor control. (Handbook of physiology, sect 1, The nervous system, vol II) Am Physiol Soc, Bethesda, pp 1179±1236 Grillner S, Zangger P (1984) The effect of dorsal root transection on the efferent motor pattern in the cats hindlimb during locomotion. Acta Physiol Scand 120:393±405 Guertin P, Angel MJ, Perreault M-C, McCrea DA (1995) Ankle extensor group I afferents excite extensors throughout the hindlimb during fictive locomotion in the cat. J Physiol (Lond) 487:197± 209 Hnik P, Kasicki S, Afelt Z, Vejsada R (1978) Chronic polyelectromyography in awake, unrestrained animals. Physiol Bohemoslov 27:485±492 Hruska RE, Kennedy S, Silbergeld EK (1979) Quantitative aspects of normal locomotion in rats. Life Sci 25:171±180 Kim YI, Na HS, Yoon YW, Nahm SH, Ko KH, Hong SK (1995) Mechanical allodynia is more strongly manifested in older rats in an experimental model of peripheral neuropathy. Neurosci Lett 199:158±160 Kugelberg E, Edström L, Abbruzzese M (1970) Mapping of motor units in experimentally reinnervated rat muscles. J Neurol Neurosurg Psychiatry 33:319±329 Lundberg A (1981) Half-centers revisited. In: Szentagothai J, Palkovits M, Hamori J (eds) Regulatory functions of the CNS. Principles of motion and organization. (Adv Physiol Sci, vol 1) Pergamon Press, Budapest, pp 155±167 Mosconi T, Kruger L (1996) Fixed-diameter polyethylene cuffs applied to the rat sciatic nerve induce a painful neuropathy; ultrastructural morphometric analysis of axonal alterations. Pain 64:37±57
201 Nicolopoulos-Stournaras S, Iles JF (1984) Hindlimb muscle activity during locomotion in the rat (Rattus norvegicus) (Rodentia: Muridae). J Zool (Lond) 203:427±440 Perreault M-C, Angel MJ, Guertin P, McCrea DA (1995) Effects of stimulation of hindlimb flexor group II afferents during fictive locomotion in the cat. J Physiol (Lond) 487:211±220 Pette D, Vrbovµ G (1985) Invited review: neural control of phenotypic expression in mammalian muscle fibres. Muscle Nerve 8:676±689 Prochazka A, Sontag KH, Wand P (1978) Motor reactions to perturbations of gait: proprioceptive and somesthic involvement. Neurosci Lett 7:35±39 Rafuse VF, Gordon T (1996a) Self-reinnervated cat medial gastrocnemius muscles. I. Comparisons of the capacity for regenerating nerves to form enlarged motor units after extensive peripheral nerve injuries. J Neurophysiol 75:268±281 Rafuse VF, Gordon T (1996b) Self-reinnervated cat medial gastrocnemius muscles. II. Analysis of the mechanisms and significance of fiber type grouping in reinnervated muscles. J Neurophysiol 75:282±297
Rossignol S, Lund JP, Drew T (1988) The role of sensory inputs in regulating patterns of rhythmical movements in higher vertebrates. A comparison between locomotion, respiration and mastication. In: Cohen A, Rossignol S, Grillner S (eds) Neural control of rhythmic movements in vertebrates. Wiley, New York, pp 201±283 SøawinÂska U, Navarrete R, Kasicki S, Vrbovµ G (1995) Motor activity in rat soleus muscle after neonatal partial denervation. Neuromuscul Disord 5:179±186 Thompson W, Jansen JKS (1977) The extent of sprouting of remaining motor units in partially denervated immature and adult rat soleus muscle. Neuroscience 2:523±535 TycÆ F, Vrbovµ G (1995) The effect of partial denervation of developing rat fast muscles on their motor unit properties. J Physiol (Lond) 482:651±660 Vejsada R, Hnik P, Navarrete R, Palecek J, Soukup T, Borecka U, Payne R (1991) Motor functions in rat hindlimb muscles following neonatal sciatic nerve crush. Neuroscience 40:267±275 Vrbovµ G, Gordon T, Jones R (1995) Nerve-muscle interaction, 2nd edn. Chapman and Hall, London Zmysøowski W, Górska T, MajczynÂski H, Bem T (1995) Interlimb coordination in intact rats (abstract). Acta Neurobiol Exp 55:14