| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Journal of Neurophysiology Vol. 79 No. 3 March 1998, pp. 1329-1340
Copyright ©1998 by the American Physiological Society
1 Department of Physiological Science and 2 Brain Research Institute, UCLA, Los Angeles, California 90095
 |
ABSTRACT |
|---|
 Abstract
 Introduction
Methods
Results
Discussion
References
|
|---|
de Leon, R. D., J. A. Hodgson, R. R. Roy, and V. R. Edgerton. Locomotor capacity attributable to step training versus spontaneous recovery after spinalization in adult cats. J. Neurophysiol. 79: 1329-1340, 1998. Locomotor performance, hindlimb muscle activity and gait patterns during stepping were studied in step-trained and non-trained female, adult spinal cats. Changes in locomotor characteristics relative to prespinalization bipedal and quadrupedal stepping patterns were used to evaluate the effects of step training on the capacity to execute full weight-bearing stepping after spinalization. Step training consisted of full weight-bearing stepping of the hindlimbs at the greatest range of treadmill speeds possible at any given stage of locomotor recovery. In the initial stages of training the limbs were assisted as needed to execute successful steps. On the basis of two behavioral criteria, the maximum speed of treadmill stepping and the number of successful steps per unit time, the ability to step was at least 3 times greater in animals trained to step versus those allowed to recover spontaneously, i.e., the non-trained. The greater success in stepping was reflected in several physiological and kinematic properties. For example, the amplitude of electromyograph (EMG) bursts in the tibialis anterior (an ankle dorsiflexor), the amount of extension at the end of both the stance (E3) and swing (E1) phases of the step cycle, and the amount of lift of the hindlimb during swing were greater in step-trained than in non-trained spinal cats. The changes that occurred in response to training reflected functional adaptations at specific phases of the step cycle, e.g., enhanced flexor and extensor function. The improved stepping capacity attributable to step training is interpreted as a change in the probability of the appropriate neurons being activated in a temporally appropriate manner. This interpretation, in turn, suggests that step training facilitated or reinforced the function of extant sensorimotor pathways rather than promoting the generation of additional pathways. These results show that the capacity of the adult lumbar spinal cord to generate full weight-bearing stepping over a range of speeds is defined, in large part, by the functional experience of the spinal cord after supraspinal connectivity has been eliminated. These results have obvious implications with regards to 1) the possibility of motor learning occurring in the spinal cord; 2) the importance of considering "motor experience" in assessing the effect of any postspinalization intervention; and 3) the utilization of use-dependent interventions in facilitating and enhancing motor recovery.
There have been several detailed descriptions of the locomotor capacity of adult spinalized cats that have been trained to step (Barbeau et al. 1987 Experimental design
COMPARISON OF STEPPING IN TRAINED AND NON-TRAINED SPINAL CATS.
Twelve adult (2-3 yr of age) female cats were trained to perform bipedal treadmill locomotion, i.e., only the hindlimbs stepped while the forelimbs rested on a platform. Age-related factors could not have explained the differences in recovery between the trained and non-trained spinal cats because all of the surgeries were performed in adult animals, i.e., after the period in which the "infant lesion effect" has been suggested to influence locomotor recovery (Robinson and Goldberger 1986a
COMPARISON OF PRESPINAL QUADRUPEDAL STEPPING TO PRE- AND POSTSPINAL BIPEDAL STEPPING.
Five cats (Step-trained cats 3-5 in Table 1 and 2 additional cats) were trained to perform quadrupedal and bipedal treadmill stepping before spinalization. After EMG and kinematic data were collected during normal stepping, the spinal cords were transected. All of the cats received bipedal step training, which began either 1 wk (n = 3) or 12 wk (n = 2) after spinalization. After training was initiated, tests of bipedal treadmill stepping were performed weekly for 12 wk.
Surgical procedures
During all surgical procedures sodium pentobarbital (35 mg/kg ip) was administered to each cat after pretreatment with atropine (ip) and acepromazine (im). Supplemental doses of anesthesia were administered as needed during surgery to maintain a low level of arousal.
Animal care procedures
Postspinalization management of the spinal cats has been detailed elsewhere (Roy et al. 1992 Hindlimb training and testing procedures
PRESPINALIZATION.
Before spinalization, bipedal hindlimb step training on a motorized treadmill was performed for 15-30 min/day, 5 days/week for 4-8 wk. The forelimbs of the cats rested on a platform raised ~2.5 cm above the treadmill belt. A cloth harness was fitted over the shoulders, between the forelimbs and around the upper trunk of all trained cats. To accustom the cats to stepping while supported in the harness, a food reward (Gerbers Baby Food) was provided. After ~2 wk of training, the cats performed consistent bipedal locomotion while in the harness at treadmill speeds of 0.2-1.0 m/s. In some animals, quadrupedal step training also was performed on the same treadmill as part of their daily training regimen. During quadrupedal stepping the animals were not restrained and food rewards were used to encourage the animals to maintain a steady gait at treadmill speeds of 0.2-1.0 m/s. Prespinalization stepping was tested at speeds up to 1.0 m/s, usually in increments of 0.2 m/s. Stepping at slower speeds was tested first and a minimum of 10 consecutive steps per speed was required before continuing testing at the next higher speed.
POSTSPINALIZATION.
After spinalization, training of bipedal hindlimb stepping was performed for 30 min/day, 5 days/week on the same treadmill that was used to train normal stepping. As during training of prespinal bipedal stepping, the forelimbs of the spinal cats rested on a platform above the treadmill belt and a cloth harness was used to support the upper trunk. In addition, a thin sheet of plexiglass, ~7 cm in height, was placed longitudinally to separate the hindpaws during stepping. The same platform, harness and divider were used during locomotor testing.
Assessment of locomotor ability
Stepping ability during a locomotor test was measured by the maximum speed of stepping and by the number of steps performed at a given speed. The maximum speed was defined as the fastest speed at which full weight-bearing steps were performed on the plantar surfaces of the paw in one hindlimb. To ensure that the animals could step reliably at the maximum speed, it was required that the cats execute at least 20 consecutive steps at the peak speed and at all speeds tested (in increments of 0.2 m/s) below the fastest speed. Thus, a maximum speed of 0.6 m/s indicates that the cat exhibited consistent stepping at 0.2, 0.4, and 0.6 m/s. In cases of inconsistent stepping performance, i.e., the inability to execute 20 consecutive steps at any speed, a maximum speed of 0 m/s was recorded. In these instances, the number of steps performed at slow speeds (0.2-0.4 m/s) provided a more useful indicator of stepping ability (e.g., compare maximum speed and step number for Non-trained cat 3, Tables 1 and 2). Step number was determined by counting the number of full weight-bearing plantar surface steps performed by the two hindlimbs at a given speed during a 45-s trial.
Data recording and analysis
EMG and kinematic data during stepping were recorded as described in detail previously (de Leon et al. 1994 Statistics
Comparisons of group mean values (trained versus non-trained; pre- versus postspinal) were statistically analyzed using computer-based resampling ("bootstrap") software (Resampling Stats 4.0.2, Arlington, VA), which is better suited for analyzing nonnormally distributed data than the more traditional population techniques (Efron and Tibshirani 1991 Locomotor performance after spinalization
On the basis of two behavioral criteria, the ability to step bipedally on a treadmill after spinalization improved more in cats that were trained to step than in those that were not trained. By week five, the step-trained cats could step at speeds that were on average ~6 times faster than the speeds attained by the non-trained cats (0.6 ± 0.2 vs. 0.1 ± 0.1 m/s; Fig. 1A). A second behavioral measure of performance, the number of successful steps occurring in a 45 s period at a speed of 0.4 m/s, also demonstrated the beneficial effects of step training. The step-trained cats performed ~30-50 more steps than the non-trained cats (Fig. 1B, weeks 6-12).
EMG burst characteristics
Several changes in the activation patterns of selected muscles reflect greater improvements in locomotion in the trained than in the non-trained cats. Valid comparisons of EMG patterns from prespinal through the 12 wk of training could be made by using the chronically implanted intramuscular electrodes. Examples of the EMG patterns during bipedal stepping before and after spinal transection are shown for one step-trained and one non-trained cat in Fig. 2. Before spinalization, clear consistent bursts of activity were recorded from each hindlimb muscle. When the cat was positioned over the treadmill during the first week after spinalization, little EMG activity was evident in either the step-trained or non-trained cats; none could execute weight-supported stepping. Four to six weeks after spinalization, rhythmic EMG activity was recorded from the hindlimb muscles in all step-trained and some non-trained cats, but the episodes of rhythmic activity were shorter in the non-trained cats. A typical non-trained cat could execute a few weight-supported steps on the dorsal or plantar surfaces of the paws before the hindlimbs stumbled (Fig. 2, d, c), usually followed by a series of unsuccessful attempts at stepping. In most muscles the EMG amplitudes after 4-6 wk approximated those recorded before spinalization in both trained and non-trained cats. The EMG patterns 12 wk after spinalization generally did not change in the trained and non-trained cats from that observed at 4 wk, although changes in the waveforms of flexor EMG bursts were apparent, e.g., TA EMG amplitudes in the trained but not in the non-trained animals were higher after than before spinalization.
CYCLE PERIOD.
The mean cycle periods between weeks four to eight after spinalization were 82 and 80% of the mean prespinalization cycle periods in step-trained and non-trained cats, respectively (Table 3). By week 12, the cycle periods in each step-trained cat had recovered to at least 90% of prespinalization values (Table 3). In contrast, the non-trained cats (except Non-trained cat 3) showed no further recovery in cycle period after week eight. Further, the mean cycle periods were significantly less in non-trained than step-trained cats at week 12 (Table 3).
MEAN EMG AMPLITUDE AND BURST DURATION.
During bipedal stepping, one EMG burst per step cycle was observed consistently in all muscles. The IP had two distinct EMG bursts per cycle, i.e., one burst during swing (IPsw) and one during stance (IPst) (Fig. 2). In some cats two bursts of activity also occurred per step cycle in the TA, both before and after spinalization. All TA EMG activity, however, was analyzed as a single burst for all cats.
Hindlimb excursion during the step cycle
Excursions of the hip, knee, ankle, and toes during bipedal stepping before spinalization and 12 wk afterward for one step-trained and one non-trained cat are shown in Fig. 5. Although there were some similarities in the steps performed by step-trained cats and the successful steps of non-trained cats, e.g., dragging of the toes at the initiation of the flexion phase followed by lifting the paw, three principal differences were evident. First, the ankle was lifted higher during swing of a successful step in step-trained compared with non-trained cats. Thus the vertical displacement of the ankle was significantly higher after spinalization in the step-trained cats (240% of prespinal values) than in the non-trained cats (114% of prespinal values). Second, the hindlimbs extended further at the end of stance (E3) in step-trained than in non-trained cats. For example, when the amount of backward movement was measured by the displacement of the toe at the beginning of swing (toe off) relative to the position of the hip, a significantly shorter distance was found in non-trained cats (79% of prespinal values) than was measured in the step-trained cats (108% of prespinal values; compare TO in Fig. 5, B and D). Third, the swing phase tended to project further forward at placement (beginning of stance) in step-trained than in non-trained cats. For example, relative to the hip position at touchdown, the forward placement of the toe was significantly greater after spinalization in step-trained cats (124% of prespinal values) than in the non-trained cats (110% of prespinal values; compare TD in Fig. 5, B and D).
Comparison of quadrupedal and bipedal stepping
Differences in mean EMG amplitudes and burst durations for the St, TA, VL, Sol, and MG were found when prespinal, quadrupedal stepping was compared with bipedal stepping after spinalization (Fig. 7). Furthermore, these differences were similar to those found when bipedal stepping before and after spinalization were compared (compare dark and light bars in Fig. 7). Therefore, based on burst amplitude and duration, similar conclusions could be drawn when either quadrupedal or bipedal stepping was used to evaluate the changes in EMG activity after spinalization. There were, however, some clear distinctions in the EMG waveforms between bipedal and quadrupedal stepping. For example, during quadrupedal stepping, IP activity consisted of a single burst that occurred during swing (Fig. 8, IPsw). In contrast, two bursts of activity, IPsw and IPst (Figs. 2 and 8), were observed in 85% (117/137) of the step cycles analyzed from the spinal cats. A similar incidence of double bursts in the IP, 92% (143/156 cycles), were observed during prespinal bipedal stepping. The IPst burst, on the other hand, was absent during quadrupedal stepping (0/83 cycles). Thus IPst burst data obtained after spinalization could not be compared with normal IP activity recorded during prespinalization quadrupedal stepping (Fig. 7).
Spontaneous versus training-induced recovery of stepping after spinalization
The present data differentiate the magnitude of the recovery of stepping after complete low thoracic spinal cord transection that can be attributed to routine daily care, i.e., spontaneous recovery, versus that which can be associated with full weight-bearing stepping for 30 min/day, 5 days/week. Multiple measures of stepping performance reflect a clear difference between the stepping capability that is attributable to spontaneous recovery versus that which occurs as a result of the practice of the motor task that is being tested, i.e., stepping. On the basis of the criteria of the maximum speed of full weight-bearing stepping, about four times more recovery occurred in the spinal cats that were trained to step. On the basis of the number of successful steps there was a threefold greater improvement if trained. In a previous study, we found an approximate threefold increase in the maximum speed that adult spinal cats could step (Lovely et al. 1986 Does training reinforce existing functional pathways or generate new pathways?
The acquisition of stepping reflects a change in the probability of the execution of a stepping event rather than the induction or absence of a pathway, which enables a new event (stepping) to occur. For example, training a spinal cord transected animal to stand, but not step, results in the inability to step (Edgerton et al. 1997b Role of muscle adaptations in improving stepping in trained spinal animals
Another factor that could have contributed to the greater postspinal acquisition of stepping in trained compared with non-trained cats could have been the recovery of some of the muscle mass that was lost after spinal injury. We have examined the potential role of muscle plasticity in explaining the differences in stepping performance of spinal animals in a wide variety of experiments (Roy and Acosta 1986 Specificity of training
These present data alone do not provide evidence that the enhanced recovery in step-trained animals resulted specifically from the enhancement of only those neural pathways that contribute to stepping. For example, it is possible that the step training had a generalized or systemic effect on the recovery of ancillary neural functions as well as those that generate stepping.
Significance of different plateau levels for the trained and non-trained spinal animals
Both the maximum speed of stepping and the number of consecutive successful steps suggest that a maximum level of performance for each group was reached within seven weeks after surgery. Thus, the rate of recovery, as well as the plateau level reached, was markedly enhanced by step training. These experiments, however, cannot determine whether or not the higher plateau reached in the training group was due to the fact that the training occurred during the first seven weeks after surgery. There are several suggestions that the plateau level could be affected by subsequent modification of the postsurgical treatment. If animals are trained to stand or not trained at all for several months and then are trained to step, there is significant improvement in their locomotor capability (Edgerton et al. 1997b Conclusions and perspective
The recovery of stepping after spinal transection could have resulted from changes in the efficacy of the existing neural networks that normally generate stepping. On the other hand it seems unlikely that the neural pathways in the spinal cord that generated or contributed to stepping while the CNS was intact were the same when supraspinal control was lost. Although the original lumbar intraspinal pathways probably remain intact after spinalization, they would be expected to function differently in a widely different intercellular milieux, because of the marked changes in synaptic densities and the type of synapses that are left intact. For example, the relative effect of any given intraspinally derived synaptic input will be higher given the absence of supraspinally derived synapses. The net effect is that the postspinal reacquisition of stepping occurred with a different spinal cord, because of the combination of neural pathways and synaptic efficacies that had generated stepping before spinal transection were unlikely to be the same combination as when stepping was executed after spinalization.
INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References
, 1993; Barbeau and Rossignol 1987, 1991; Edgerton et al. 1991, 1997a,b; Hodgson et al. 1994; Lovely et al. 1986, 1990). These studies show that spinal cats can execute full weight-bearing hindlimb stepping on a treadmill while restrained in a harness with their forelimbs resting on a platform, within 1-3 wk of spinalization when some level of step training is imposed (Barbeau and Rossignol 1987; Hodgson et al. 1994; Lovely et al. 1986, 1990) or when specific pharmacological agents, e.g., clonidine or strychnine, are administered to the trained spinal cats (Barbeau et al. 1993; Barbeau and Rossignol 1991; Edgerton et al. 1997a,b). As the postspinal time increases and the animals continue to be trained, hindlimb stepping ability progresses, i.e., increases in maximum treadmill speed and number of plantar surface steps performed, and usually reaches a plateau after 2-3 mo of training (Barbeau and Rossignol 1987; Edgerton et al. 1997a; Lovely et al. 1986). In addition, many features of the activation patterns (EMG) of the muscles and of the hindlimb kinematics of stepping are reestablished in trained adult spinal cats (Belanger et al. 1996; de Guzman et al. 1991; Forssberg et al. 1980; Lovely et al. 1990) and kittens (Grillner 1973, 1981; Smith et al. 1982).
) has used a non-trained group of spinal animals to differentiate the amount of recovery attributable to spontaneous, i.e., normally occurring recovery from spinal surgery, versus training interventions. Although Lovely et al. (1986) found that the maximum treadmill speed in step-trained cats was significantly greater than in non-trained cats 6 mo after spinalization (and 5 mo of training), there were no comparisons of the electromyograph (EMG) patterns or the kinematics of stepping between the groups.
METHODS
Abstract
Introduction
Methods
Results
Discussion
References
). EMG electrodes were implanted in selected hindlimb muscles and tests of bipedal stepping began 1 wk later. After sufficient EMG and kinematic data were collected, the spinal cords of the cats were completely transected (T12-T13). Beginning one week after spinalization, six of the spinal cats received bipedal step training (see Step-trained cats 1-6, Table 1) and six of the cats were not trained (see Non-trained cats 1-6, Table 1). Weekly tests of bipedal stepping on a treadmill were performed for 12 wk starting 1 wk after spinalization.
View this table:
TABLE 1.
Maximum treadmill speed* in step-trained and non-trained cats before and after spinalization
; Pierotti et al. 1989). After each muscle was back-stimulated to ensure proper electrode placement, each wire was secured in the muscle with a suture at its entry and exit from the muscle.
). Briefly, a skin incision was made on the back to expose the vertebral processes between ~T10 to L1. A partial laminectomy was performed to expose the spinal cord at the T12-T13 junction. Fine scissors and forceps were used to cut the dura and to perform the transection beginning on the dorsal surface of the cord between T12-T13. After the transection, no spinal cord matter was visible between the two cut ends of the cord. The ends of the cord retracted leaving a clear space between the two cut ends. This procedure allowed for the preservation of the large ventral artery of the spinal cord. Gelfoam was inserted in the space and the muscle and skin above the lesion site were closed with sutures.
). Cats were housed in spacious cages, 2-4 cats per cage, with the cage floors covered with shredded newspaper. The bladders and colons of the cats were expressed twice daily for the duration of the experiment. Dry kibble and water were given ad libitum and wet food was given once daily. All procedures were performed in accordance with the American Physiological Society Animal Care Guidelines and were approved by the Animal Use Committee at the University of California, Los Angeles.
View this table:
TABLE 2.
Number of successful steps* in step-trained and non-trained cats before and after spinalization
). Briefly, raw EMG signals were amplified and recorded on an FM tape recorder (TEAC Model XR-510, TEAC Corporation, Montebello, CA), while a camera and video cassette recorder (Panasonic System Camera, WV D5100 and Panasonic AG1280P; Cypress, CA) were used to record the video signals. An SMPTE time code generator (Model F30, Fast Forward Video; Irvine, CA) was used to synchronize video frames with the EMG signals recorded on FM tape.
). Briefly, the EMG signals from each muscle during 10-20 s bouts of stepping were sampled into an AMIGA computer at 2 kHz and calibrated. The EMG signals were rectified and smoothed by using a moving average (9-point moving average, i.e., 110 Hz low-pass filter). Computer software designed in-house was used to detect and display the start and end of each burst based on a given threshold level above the baseline noise for a channel. The starting and ending points of bursts were used to determine the relative timing of EMG activity recorded from different muscles. Cycle periods were calculated as the time between the starting points of successive bursts of EMG activity. Burst durations were calculated as the time between the start and the end of each burst. Mean EMG amplitude was calculated by dividing the integrated area of each unsmoothed burst by the burst duration. Averaged EMG waveforms were generated from unsmoothed bursts triggered to the start of TA activity in each recording channel. EMG activity corresponding to poorly executed steps, i.e., dorsal-surface steps or steps that were not full weight-bearing, were excluded from analyses (with the exception of Non-trained cat 6 in Table 2, which failed to exhibit any weight-bearing thus, rhythmic EMG activity during unsuccessful attempts at treadmill stepping was analyzed).
). Briefly, the bony landmarks on the hindlimb were digitized and calibrated and the stick figure representations were plotted. The knee position was triangulated as described previously (Goslow et al. 1973). Hip, knee, and ankle joint angles were generated by triangulating the x and y coordinates. Specifically, the greatest amount of flexion during the F and E2 phases of the step cycle (Philippson 1905) and the greatest amount of extension during the E1 and E3 phases were measured. The maximum amount of forward and backward hindlimb movement was measured by the displacement of the x coordinate of the paw marker (head of the 5th metatarsal) relative to the x coordinate of the hip marker (greater trochanter). The amount of paw lift during swing was measured by the displacement of the y coordinate of the ankle marker (lateral maleolus). Swing was defined as the period starting with forward movement of the paw (toe off) and ending with, but not including, paw contact (touchdown). Stance was defined as the period beginning with paw contact and ending before forward paw movement. These definitions were based on the reported characteristics of the swing phase of stepping in spinal cats, i.e., the paw drags before it is lifted (Belanger et al. 1996).
and 
). Most EMG and kinematic data for spinal cats were from single tests performed between 4-8 wk after spinalization, i.e., when full weight-bearing stepping was initially observed, and/or 12 wk after spinalization, i.e., when a plateau in locomotor performance was attained. When a sufficient number of step cycles from one cat was not available from a single test (<8 cycles), data from multiple tests were pooled for analysis (see Non-trained cats 4, 5, Table 2). Comparisons of the EMG characteristics of quadrupedal and bipedal locomotion were based on analyses of stepping in 5 step-trained cats. Data from 3 of the 5 cats (Step-trained cats 3, 4, 5, Table 2) were from tests performed before and 12 wk after spinalization. Data from the remaining 2 cats (not shown in Table 2) were from tests performed before and 24 wk after spinalization. EMG activity during prespinal quadrupedal and bipedal stepping and postspinal bipedal stepping after 12 wk of step training was analyzed in these cats.
). In the resampling approach, significance is judged by comparing statistics calculated from actual data to analogous statistics calculated from computer simulations that realize the null hypothesis of the experiment. In this case, the actual difference between group means was compared with the differences between means of groups generated from the original data by randomizing group membership. If the actual difference was >95 or 99 of 100 of the differences in the randomized data, the actual result was considered significant at the P < 0.05 or P < 0.01 level, respectively. Paired t-tests were used to compare mean cycle period before and 12 wk after spinalization in individual cats.
RESULTS
Abstract
Introduction
Methods
Results
Discussion
References
View larger version (25K):
[in a new window]
FIG. 1.
Performance of bipedal hindlimb stepping on a treadmill after spinalization in step-trained and non-trained cats. Average maximum speeds (A) and number of steps performed during 45 s of testing (0.4 m/s) (B) from 6 step-trained ( 
) and 6 non-trained (
) cats are shown 1-12 wk after spinalization and 3 non-trained cats 10 wk after spinalization (
) (see text for definition of maximum speed). The horizontal line indicates average maximum speed or number of steps observed during prespinalization tests of stepping. Data from individual cats are shown in Tables 1 and 2. Bars, standard errors of mean. * Significantly different from non-trained(P < 0.05).
View larger version (50K):
[in a new window]
FIG. 2.
Electromyograph (EMG) activity during bipedal hindlimb stepping in a step-trained and a non-trained cat before ( 
2 or 5 wk) and 1, 4, and 12 wk after spinalization. Raw EMG was recorded from selected hindlimb muscles (St, semitendinosus; VL, vastus lateralis; IP, iliopsoas; Sol, soleus; MG, medial gastrocnemius; TA, tibialis anterior) during stepping at a treadmill speed of 0.4 m/s. Stance phase during full weight-bearing step cycles is indicated by horizontal lines below EMG records. c, a step failure (collapse); d, contact of dorsal surface of paw on treadmill belt (toes curled). , IP bursts during swing (IPsw) , stance (IPst) in one step cycle. Horizontal calibration, 1 s, and vertical calibration, 1.0 mV, for all muscles except for Sol (2.0 mV). Data are from Step-trained cat 4 and Non-trained cat 3 (see Tables 1 and 2).
View this table:
TABLE 3.
Cycle period during bipedal hindlimb locomotion before and after spinalization
View larger version (25K):
[in a new window]
FIG. 3.
Amplitude (A) and duration (B) of EMG activity during bipedal locomotion in 6 step-trained and 6 non-trained spinal cats. Mean EMG amplitude and burst duration of 8-50 EMG bursts from selected hindlimb muscles (see Fig. 2 for muscle abbreviations) during full weight-bearing bipedal steps (0.4 m/s) before and 12 wk after spinalization were measured. Postspinal values were expressed as a percentage of prespinalization values in each cat. Data shown are mean ± SE of normalized values. Complete analyses were not performed for every cat (see Fig. 4 for number of cats). *, , and , significantly different from non-trained at P < 0.01, non-trained at P < 0.05, and prespinalization (100%) at P < 0.01.
View larger version (19K):
[in a new window]
FIG. 4.
Changes in EMG amplitude in step-trained and non-trained cats after spinalization. Mean EMG amplitude of 8-50 EMG bursts from selected hindlimb muscles (see Fig. 2 for muscle abbreviations) during full weight-bearing bipedal steps (0.4 m/s) before (Pre) and 12 wk after (Post) spinalization were measured. Data shown are changes from prespinal values (100%) for individual cats (average values of mean EMG amplitude across cats are shown in Fig. 3). Each symbol represents a different cat in each group. Note, scale of vertical axes for each plot is the same.
View larger version (32K):
[in a new window]
FIG. 5.
Stick figure representations of hindlimb from swing and stance phases of one step cycle (0.4 m/s) in a step-trained cat before (A) and after spinalization (B) and from a non-trained before (C) and after spinalization (D). 
, progression of hindlimb movement. After spinalization, swing was composed of a drag and lift component (see text) indicated by arrows. Touchdown at beginning of stance and toe off at beginning of swing are denoted by TD and TO, respectively. Horizontal position of each hindlimb representation was normalized to position of hip (greater trochanter). Horizontal and vertical calibration, 5 cm. Data are from Non-trained cat 1 and Step-trained cat 4 (see Tables 1 and 2).
).
View larger version (17K):
[in a new window]
FIG. 6.
Ankle, knee, and hip angles in step-trained ( 
) and non-trained () spinal cats during phases of step cycle. Maximum amount of joint flexion, i.e., smallest joint angles, at F and E2 and maximum amount of joint extension, i.e., greatest joint angles, at E1 and E3 were measured from 8-10 cycles of full weight-bearing bipedal hindlimb stepping (0.4 m/s) in 4 step-trained and 5 non-trained cats before and 12 wk after spinalization. Postspinal values were expressed as a percentage of prespinal values (- - -). Mean ± SE of normalized values are shown. E1 and E2 (yield) phases are absent at hip. *, , and , significantly different from non-trained at P < 0.01, non-trained at P < 0.05, and prespinalization (100%) at P < 0.01.
View larger version (22K):
[in a new window]
FIG. 7.
Amplitude and duration of EMG activity during bipedal hindlimb stepping in step-trained spinal cats relative to prespinal quadrupedal stepping. Mean EMG amplitude and burst duration of 8-50 EMG bursts from selected hindlimb muscles (see Fig. 2 for muscle abbreviations) during prespinal quadrupedal stepping and postspinal bipedal stepping were measured in 5 cats. Data shown are mean ± SE of normalized values. Postspinal values were expressed as percentage of prespinalization values for individual cats ( 
). For comparison, postspinal data normalized to bipedal stepping before spinalization are also shown (; see Fig. 3). NA, no IPst burst observed. * and , significantly different from prespinalization (100%) and data normalized to prespinal bipedal values at P < 0.01, respectively.
View larger version (29K):
[in a new window]
FIG. 8.
Average EMG waveforms during quadrupedal and bipedal stepping before spinalization and bipedal stepping after spinalization. Ten to 14 cycles of rectified EMG activity from Sol and IP muscles triggered from start of TA activity were averaged for each locomotor condition. IP bursts during swing and stance are denoted by sw and st, respectively. Note, average waveforms are shown relative to one cycle of TA activity (vertical lines denote onset of TA activity); thus, Sol and IPst activity from 2 cycles (bipedal stepping) are depicted. Horizontal calibration, 1 s, and vertical calibration, 0.1 mV for TA and IP and 0.2 mV for Sol. Data are from 1 step-trained cat (see METHODS).
View larger version (16K):
[in a new window]
FIG. 9.
Comparison of timing of EMG activity from selected hindlimb muscles during a single cycle of bipedal (dark bars; B) and quadrupedal (hatched bars; Q) stepping. Onset and offset for TA, IPsw, IPst, and Sol EMG bursts were normalized to onset of toe off at beginning of swing (see up arrow and vertical, dashed line) for bipedal and quadrupedal stepping. Eight to 50 cycles (0.4 m/s) of prespinal quadrupedal and pre- and postspinal bipedal stepping in 5 step-trained cats were analyzed to generate average onset and offset values shown. Average onsets of touchdown for 2 locomotor conditions are denoted by down arrows. IPst burst was absent during quadrupedal stepping, denoted by NA. Bars, SE of mean onset and offset.
DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References
). This similarity in results between studies was observed despite two notable differences between the experiments of Lovely et al. (1986) and the present study. First, in the earlier study the cats were tested 6 mo after spinalization and step training was initiated one month after spinalization compared with the present study where 3 mo of training were performed beginning 1 wk after spinalization. Secondly, the criteria for reaching the maximum speed were based on five consecutive steps in the earlier study and 20 consecutive steps in the present study.
; Forssberg et al. 1980; Lovely et al. 1990). In the present study, similar conclusions could be made for some, but not all, measures of locomotor performance whether bipedal or quadrupedal stepping was used as a criteria. For example, the appearance of double bursting in the IP and the earlier onset of TA activity relative to IP activity occurred during bipedal stepping before and after spinalization, but not during quadrupedal stepping. Other differences between quadrupedal and bipedal stepping were found by de Guzman et al. (1991) who reported a decrease in MG activity after spinalization when compared with prespinal quadrupedal, but not to prespinal bipedal, stepping. These data illustrate that, at least in some cases, the more appropriate prespinal control is bipedal rather than quadrupedal stepping. The explanation for the differences in the EMG patterns during bipedal and quadrupedal stepping cannot be determined from the present study. It is possible, for instance, that propriospinal and/or mechanical coupling between the forelimbs and hindlimbs might be the bases for these differences. Differences in the forelimb-hindlimb coordination during different locomotor gaits in cats have been associated with unique temporal patterns of muscle activity, i.e., the relative timing of forelimb flexor and hindlimb extensor EMG activity differs during pacing versus trotting on a treadmill (Miller et al. 1975). Furthermore there is evidence that these recruitment strategies are programmed at the spinal level (Miller and Van Der Meche 1976).
). Although these stand-trained spinal animals had not executed successful stepping for ~3 mo (none since spinal transection), they could execute full weight-bearing stepping within minutes after administration of strychnine, i.e., blocking glycine receptors. This finding suggests that the neural pathways needed to execute stepping were present even in those animals that could not step, but that these pathways were nonfunctional. It appears that strychnine modulated the balance of excitatory and inhibitory inputs to the interneurons and motor pools such that the probability of the appropriate sequence of synaptic events in the appropriate pathways increased dramatically and rapidly. Similar observations were observed when bicuculline was administered to non-trained spinal cats (Robinson and Goldberger 1986b).
).
; Roy et al. 1991, 1998). The overall conclusion is that the recovery in muscle mass that can occur as a result of motor training cannot account for the recovery in motor performance. Neither the magnitude nor the time course of these changes adequately paralleled the changes in stepping ability. Thus the larger changes in motor capacity observed in the trained cats were attributable primarily to changes in the probability of a given combination of neural pathways being activated in a temporally appropriate manner.
; Hodgson et al. 1994). There is also evidence that the specific source of afferent information used to initiate and maintain stepping during training becomes essential for the animal to execute the trained task. For example, if the tail is "crimped" to increase afferent stimulation during step training, then the ability to step becomes highly dependent on the presence of this source of stimulation (Edgerton et al. 1997a).
; Lovely al. 1986; Robinson and Goldberger 1986a). One possibility is that the animals that regained some stepping ability spontaneously executed weight-supporting stepping during routine movements in their cages. Brief, incidental experiences of weight-supporting stepping could have occurred each day and this could have had an effect. This raises an issue regarding how much training or experience is needed to trigger an improvement in the probability of successfully generating stepping. Long-term continuous activity recordings are needed to determine the degree to which the motor tasks that are practiced after spinal transection are the motor tasks that are acquired, either coincidentally (spontaneously) or purposely as part of a training regimen.
; Hodgson et al. 1994). In addition, if an animal that is trained to step is subsequently placed on a regimen of stand training, that animal loses its ability to step effectively. Therefore, on the basis of the present results and the observations that the behavioral potential of spinal cats can be modified by repetitive training for periods of up two years, there appears to be a long-term persistence of spinal plasticity after spinalization. Further experiments are needed, however, to determine whether or not early training can result in greater improvements when imposed immediately after spinal transection.
| |
ACKNOWLEDGEMENTS |
|---|
The authors thank S. Lauretz for excellent care of the animals and M. Recktenwald for technical assistance.
This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-16333.
| |
FOOTNOTES |
|---|
Address for reprint requests: R. D. de Leon, c/o V. R. Edgerton, Dept. of Physiological Science, UCLA, PO Box 951527, Los Angeles, CA 90095-1527.
Received 14 August 1997; accepted in final form 4 November 1997.
| |
REFERENCES |
|---|
|
Abstract
Introduction
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Barriere, H. Leblond, J. Provencher, and S. Rossignol Prominent Role of the Spinal Central Pattern Generator in the Recovery of Locomotion after Partial Spinal Cord Injuries J. Neurosci., April 9, 2008; 28(15): 3976 - 3987. [Abstract] [Full Text] [PDF]
|
|
