Impairment of Postural Control in Rabbits With Extensive Spinal Lesions

Abstract

Our previous studies on rabbits demonstrated that the ventral spinal pathways are of primary importance for postural control in the hindquarters. After ventral hemisection, postural control did not recover, whereas after dorsal or lateral hemisection it did. The aim of this study was to examine postural capacity of rabbits after more extensive lesion (3/4 section of the spinal cord at T₁₂ level), that is, with only one ventral quadrant spared (VQ animals). They were tested before (control) and after lesion on the platform periodically tilted in the frontal plane. In control animals, tilts of the platform regularly elicited coordinated electromyographic (EMG) responses in the hindlimbs, which resulted in generation of postural corrections and in maintenance of balance. In VQ rabbits, the EMG responses appeared only in a part of tilt cycles, and they could be either correctly or incorrectly phased in relation to tilts. Because of a reduced value and incorrect phasing of EMG responses on both sides, this muscle activity did not cause postural corrective movements in the majority of rabbits, and the body swayed together with the platform. In these rabbits, the ability to perform postural corrections did not recover during the whole period of observation (≤30 days). Low probability of correct EMG responses to tilts in most rabbits as well as an appearance of incorrect responses to tilts suggest that the spinal reflex chains, necessary for postural control, have not been specifically selected by a reduced supraspinal drive transmitted via a single ventral quadrant.

INTRODUCTION

Maintenance of the basic body posture, upright in humans and dorsal-side-up in quadrupeds, and keeping balance is a fundamental motor function. This nonvolitional form of motor activity in many species is based on inborn neural mechanisms (Deliagina et al. 2006; Massion 1998). An efficient control of the basic posture is equally important for standing and during walking (Horak and Macpherson 1996; Macpherson et al. 1997; Orlovsky et al. 1999) as well as for providing support of voluntary limb movements (Massion and Dufosse 1988).

Damage to the spinal communication system (descending and ascending spinal pathways), caused by, e.g., a spinal cord injury (SCI), can result in an impairment of the postural control. This affects equilibrium both during standing and during locomotion and can impair voluntary movements that need postural support (Horak and Macpherson 1996). Apparently, postural deficits depend on the location and extent of SCI. A complete lesion of the spinal cord in the thoracic region separates the spinal postural mechanisms of the hindquarters from the rest of the CNS. This has dramatic consequences. Studies on chronic spinal cats have shown that they exhibit very poor responses to disturbances of posture and are not able to maintain the dorsal-side-up orientation of the caudal part of their body (Barbeau et al. 2002; Macpherson and Fung 1999; Rossignol et al. 1999, 2002). However, a reduced postural control (brief standing episodes) may remain (Giuliani and Smith 1985; Kellog et al. 1946) and can be improved by training (see e.g., de Leon et al. 1998; Edgerton et al. 2001, 2004; Grillner 1973; Pratt et al. 1994).

Worsening of lateral stability of the hindquarters was observed during locomotion in cats with ventral and ventrolateral spinal lesions; a loss of balance depended on the extent of the lesion (Brustein and Rossignol 1998, 1999). By contrast, dorsal spinal lesions caused only transient deficits in postural stability during locomotion (Jiang and Drew 1996). Our recent studies on rabbits with different damages to the spinal cord in the thoracic region (Lyalka et al. 2005) demonstrated that ventral spinal pathways are of primary importance for postural control in the hindquarters in the task of standing and keeping balance on the tilting platform. When the dorsal pathways were transected but the ventral pathways were spared (dorsal hemisection, DHS), postural corrections recovered in 1–2 wk after lesion. By contrast, when the ventral pathways were transected (ventral hemisection, VHS), postural corrective movements were absent, though reduced EMG responses to tilts (either correctly or incorrectly phased) could be observed. After the lateral hemisection (LHS), which spared one dorsal and one ventral quadrant on the same side, postural corrections recovered in 1–3 wk. Thus these experiments have shown that integrity of at least a part of ventral pathways is critical for the ability to generate postural corrections.

In the present study, we continued the analysis of effects of SCI on postural performance in the rabbit. For this animal model, the functional organization of postural mechanisms has been characterized in our previous studies (Beloozerova et al. 2003; Deliagina et al. 2000). In particular, it was shown that postural mechanisms of the fore- and hindquarters are relatively independent of each other, suggesting that SCI at the thoracic level affects primarily the hindlimb postural mechanisms. The aim of this study was to examine the postural capacity of rabbits with more extensive lesions than used in the previous study (Lyalka et al. 2005). For this purpose, we performed the 3/4 section of the spinal cord at the T₁₂ level so that only one ventral quadrant was left intact (VQ animals). These studies have shown that the 3/4 section severely damaged the postural performance in the hindquarters, and the postural control in most animals did not recover during the whole period of observation (up to one month).

A brief account of part of this study has been published in abstract form (Lyalka et al. 2006).

METHODS

Experiments were carried out on five adult male New Zealand rabbits (weighing 2.5–3.5 kg). All experiments were conducted with the approval of the local ethical committee (Norra Djurförsöksetiska Nämnden) in Stockholm.

Surgical procedures

Each animal was subjected to two operations under Hypnorm-midazolam anesthesia, using aseptic procedures. During the first surgery, bipolar EMG electrodes (0.2 mm flexible stainless steel Teflon-insulated wires) were implanted bilaterally into m. gastrocnemius lateralis (Gast, ankle extensor), m. vastus lateralis (Vast, knee extensor), and m. biceps femoris (Bic, knee flexor). The wires were led subcutaneously toward the head and then through a small incision in the skin on the dorsal aspect of the neck. The wound was sutured so that the wires were fastened to the skin. A small connector was soldered to each wire at a distance of 2–3 cm from the skin.

In 3–4 days, when the animal had recovered completely from the first surgery, its postural responses to tilts were tested (see following text), and afterward a second surgery (spinal damage) was performed. An incision was made along the dorsal midline in the lower thoracic region. A laminectomy at T₁₂ level was done, the dura was opened on the dorsum of the cord, and the lesion (transection of 2 dorsal quadrants and the left ventral quadrant) was done under the dissecting microscope by means of spring scissors, microsurgery forceps, and small scalpel. Afterward the incision was closed in anatomical layers.

Experimental design

Postural tests on a tilting platform have been described earlier (Beloozerova et al. 2003; Deliagina et al. 2000; Lyalka et al. 2005). No special training of the rabbits was required prior to testing. For testing, an animal was positioned on the two platforms (P1 and P2 in Fig. 1A), so that P1 supported the forelimbs, and P2 supported the hindlimbs. The sagittal plane of the animal was aligned to the axis of the platform rotation (Fig. 1, B and C). The surface of the platforms was covered with sandpaper to prevent sliding of the animal during tilts.

FIG. 1.

Experimental design. A–C: testing of postural responses to tilts. The animal was standing on 2 platforms, 1 under the fore limbs (P1) and 1 under the hind limbs (P2). Platform P2 could be tilted in the transverse plane (α is the platform tilt angle). The sagittal plane of the animal was aligned to the axis of platform rotation. Mechanical sensor S, positioned at the half-height of the body, measured lateral displacements of the caudal part of the trunk in relation to the P2 platform. D: schematic representation of the trajectory of the platform angle (α) and corrective movements of the caudal trunk (S).

The platform supporting the hindquarters could be periodically tilted in the frontal (transverse) plane of the animal (angle α, Fig. 1, A and C), while P1 was kept horizontal. A trapezoidal tilt trajectory was used, with transitions between stationary (extreme) positions lasting for 0.5–0.7 s and with each position being maintained for 2–3 s (Fig. 1D). Tilts were symmetrical in relation to the horizontal position, with the peak-to-peak value of 40°. Smaller values (20 or 30°) were used only if an animal with a spinal lesion could not compensate for 40° peak-to-peak tilts. Usually, 50–60 tilt cycles were performed during a trial.

In a previous study (Beloozerova et al. 2003), it was shown that lateral displacements of the trunk in relation to the platform well characterized the efficacy of stabilization of the dorsal-side-up trunk position. In the present study, lateral displacements of the caudal part of the trunk in relation to the platform P2 (postural corrections) were monitored by the sensor S. The sensor consisted of a variable resistor whose axis was rotated by means of a long lever; the latter was touching the lateral aspect of the trunk at the half-height of the body (Fig. 1, A–C).

The signals from the EMG electrodes and from the position sensors were amplified, digitized with a sampling frequency of 5 kHz (EMGs) and 1 kHz (sensors), and recorded on a computer disk using the data acquisition and analysis software (Power-1401/Spike-2, Cambridge Electronic Design, Cambridge, UK). The EMG signals were rectified and smoothed (time constant, 50 ms).

To characterize the body configuration in the standing animal, video recordings (25 frame/s) from the side were performed on the horizontal platform (see Fig. 5H). In addition, a view from below was video recorded using a transparent horizontal platform and a mirror (see Fig. 5, A and B). The video camera was positioned at a distance of ∼2 m from the rabbit. The recordings were analyzed frame-by-frame. The resolution of video images was ∼2 mm.

FIG. 5.

Effect of the 3/4 section on the postural configuration. A and B: postural configuration of the rabbit before (A) and after 3/4 section of the spinal cord (B), view from below. To characterize the position of each hindlimb, the following values were measured: the foot angle (ɛ), as well as the coordinates (x and y) of the rostral point of the foot in relation to the mid-body axes. C: foot angles in the hindlimbs of rabbit 61 in control (day 0) and on different days after the 3/4 section. D: foot position of rabbit 61 in control and on different days after the 3/4 section. E–G: foot angles (E), lateral (F), and anterior-posterior (G) foot position (mean ± SE) averaged over 5 animals. H: the height (Ht) of the hindquarters was measured in the standing position. I: the height as a function of time after lesion in individual rabbits (0 is control).

All quantitative data in this study are presented as the means ± SD or SE. Student's t-test was used to characterize the statistical significance when comparing different means; the significance level was set at P = 0.05.

Histological procedures

At the termination of the experimental series, rabbits were deeply anesthetized with pentobarbital sodium and perfused with isotonic saline followed by a 10% formalin solution. Frozen sections of 30 μm thickness were cut in the region of spinal cord damage. The tissue was stained for Nissl substance with cresyl violet. The position and the extent of lesions were verified by observation of a series of magnified digital images of sections.

RESULTS

Postural performance before 3/4 section

All animals initially served as controls and were tested on the tilting platform before the lesions. The rabbits were easily engaged in the postural task, and maintained balance when the caudal platform (P2 in Fig. 1A) was periodically tilted. The animals exhibited stereotypic postural responses that included an extension of the hindlimb on the side moving down and a flexion of the hindlimb on the opposite side, as shown schematically in Fig. 1, B and C (see also Beloozerova et al. 2003; Deliagina et al. 2000; Lyalka et al. 2005). These flexion and extension limb movements made the trunk displace in the transverse plane in relation to the platform, in a direction opposite to the platform tilt. The corrective trunk movements reduced the deviation of the body from the dorsal-side-up position. Figure 2A shows lateral displacements of the trunk in relation to the platform (monitored by the sensor S, see Fig. 1, A–C) caused by tilts of the platform. The corrective trunk movements (S) were in anti-phase with the platform tilting movements (α). The peak-to-peak value of S in intact rabbits ranged from 2.5 to 4.5 cm (3.9 ± 0.5 cm, mean ± SE).

FIG. 2.

Examples of postural responses before and after 3/4 section. A and B: kinematical and electromyographic (EMG) responses to tilts in control (A) and 21 days after 3/4 section (B; rabbit 60). C: kinematical and EMG responses 20 days after 3/4 section (rabbit 45). α, tilt angle of the P2 platform; S, lateral displacement of the caudal part of the trunk in relation to the P2 platform. The EMGs of the following muscles are presented: left (L) and right (R) vastus (Vast), gastrocnemius (Gast) and biceps femoris (Bic). White arrows in B indicate correct responses; black arrows indicate incorrect responses. Shaded column highlights half of the tilt cycles to facilitate comparison between curves.

Compensatory movements of the hindlimbs were caused by a specific pattern of muscle activity (see also Beloozerova et al. 2003; Deliagina et al. 2000; Lyalka et al. 2005). When the ipsilateral side of the platform was moving downward, the limb was extending due to activation of extensor muscles. When the platform was moving upward, the limb was flexing due to a reduction of the extensor activity and activation of some flexor muscles. This EMG pattern is illustrated in Fig. 2A. Tilts of the platform caused a periodical modulation of the EMGs of the left and right m. vastus (Vast). The activity increased during ipsilateral tilts and decreased during contralateral tilts. The pattern similar to that in Vast was observed in m. gastrocnemius (Gast, not illustrated). A different pattern was observed in m. biceps femoris (Bic): this muscle was activated during contralateral tilt, that is, in anti-phase with Vast.

Postural performance after 3/4 section

POSTURAL REFLEXES.

In all five animals, the postural corrections and the ability to maintain equilibrium in the hindquarters were absent during at least a few days after the lesion. When the animal was positioned on the tilting platform, its hindquarters passively followed the platform movements and with larger tilts the rabbit could fall sideways. The rabbits also did not use their hindlimbs for locomotion and moved around using their forelimbs only. Sometimes forelimb stepping movements were accompanied by small jerk-like stepping movements of the hindlimbs. Later on, some elements of postural control re-appeared, but the degree of recovery was very small in all subjects except for rabbit 45. Locomotion with involvement of the hindlimbs also recovered in only this rabbit.

Figure 2B shows responses to tilts in one of the animals with bad recovery (rabbit 60, day 11 after lesion). In this rabbit before lesion (Fig. 2A), lateral displacements of the caudal trunk (S) occurred in anti-phase to tilts of the caudal platform (α), indicating the presence of efficient postural corrections. After lesion, caudal trunk displacements instead occurred in-phase with tilts, indicating that the hindquarters passively swayed toward the side tilting downward. Tilts of the platform, however, evoked EMG responses in a part of the hindlimb muscles. Both Vast(L) and Vast(R) were activated, but in wrong phases (with the contralateral tilt, these “incorrect” responses are indicated by black arrows in Fig. 2B). Besides that, Vast(L) responded not in each cycle. Bic(L) was activated in a correct phase (with the contralateral tilt, white arrows). However, this “correct” response fluctuated considerably in amplitude, and it was absent in a part of tilt cycles. Finally, Fig. 2B shows that Bic(R) did not respond to tilts. Spontaneous changes of the type of response in the same muscle (from correct to incorrect, and vice versa) were sometimes observed (not illustrated).

To characterize these irregular EMG responses to tilts, in each trial, we calculated the percent of correct responses, incorrect responses, and cases without response; this was done for each of the six recorded hindlimb muscles. Figure 3 shows these values for the trial, a part of which was presented in Fig. 2B (rabbit 60, 11 days after lesion). One can see that, on the average, the cycles with no response in a given muscle were much more numerous than the cycles with correct or incorrect responses. The data for individual muscles were then averaged over all muscles (Fig. 4D, day 11). On the average, correct and incorrect responses constituted ∼17 and 28%, respectively, whereas cycles without response constituted ∼55%. No significant difference was observed between the proportion of correct (or incorrect) responses on the left and right side.

FIG. 3.

Characterization of EMG responses to tilts in the badly compensated rabbit (60). All EMG responses in individual muscles (recorded in 1 trial) were classified into 3 categories—correct responses (extensor EMG is timed to the ipsilateral tilt, flexor EMG is timed to the contralateral tilt), incorrect responses (opposite phase relations), and no response. The relative number of responses in each category (%) is indicated.

FIG. 4.

Summary of responses to tilts in ventral quadrant (VQ) rabbits. The diagrams (A–E) show, for individual rabbits, the percent of correct responses, incorrect responses, and cycles with no responses (averaged over all recorded muscles of both limbs) as a function of postlesion time (day 0 is control). Days of testing when the postural motor responses were present or absent are indicated by black and white horizontal bars, respectively. Right: extent of the spinal cord damage in individual rabbits. The shape of the tissue that remained intact was determined from the photomicrograph taken at the center of the lesion (1st column) and projected on a scheme of the spinal cord section taken more rostrally (2nd column). In the photos, the lesioned area is delimited by a continuous white line, and the intact gray matter by a dotted line, the remainder corresponding to the intact white matter. Arrows indicate a position of the ventral fissure. In the schemes, the damaged area is shaded.

Figure 2C shows responses to tilts in the only animal exhibiting a good recovery (rabbit 45, day 20 after lesion). The timing of corrective movements was normal (in anti-phase to tilts), and the value of corrective movements (∼4 cm) was close to normal one (3.9 ± 0.5 cm). The phasing of EMG responses in the series of tilt cycles (shown in Fig. 2C) was also correct. However, when analyzing other trials, a few cycles with incorrect EMG responses or without any response could be found. This is reflected in the summary diagram (Fig. 4A, day 20). The number of correct responses was <100% even on days 24 and 29 after lesion (∼80 and 70%, respectively). No significant difference was found between the muscular responses in the left and right limbs.

The data on postural reflexes at different post-SCI stages are summarized in Fig. 4 for five individual VQ rabbits. The days of testing when corrective movements were present (as in Fig. 2, A and C) or absent (as in Fig. 2B) are indicated by black and white horizontal bars, respectively. Corrective movements were observed in control (day 0) in all rabbits. After the 3/4 section, corrective movements were observed only in rabbit 45 starting from day 9, in all other rabbits, corrective movements were absent during the whole period of observation. The diagrams also show, for individual rabbits, the percent of correct EMG responses, incorrect EMG responses, and cycles with no response, as a function of postlesion time.

There was no significant difference in the proportion of correct EMG responses observed in the limb ipsilateral to the intact VQ and in the opposite limb; this proportion was 69 ± 6 vs. 61 ± 10% in the recovered rabbit 45; and 13 ± 4 vs. 19 ± 5% in the group of rabbits 49, 50, 60, and 61 (means ± SE, P > 0.1, t-test).

Table 1 shows the main results obtained for five individual rabbits. In four of them, the postural control was severely damaged: the corrective movements were absent, and the proportion of correct EMG responses was small (<30%, averaging over all days of testing).

TABLE 1.

Postural performance in individual VQ rabbits

Rabbit No.	45	49	50	60	61
Period of observation, day	29	13	30	18	20
Corrective movement	Yes	No	No	No	No
Correct EMG responses, %	55 ± 22	30 ± 13	8 ± 7	13 ± 8	2 ± 1

In the bottom row, a proportion of correct responses in each animal, averaged over all tests, is given (means ± SD). VQ, ventral quadrant; EMG, electromyogram.

In Fig. 4 (right), there are shown the photomicrographs of lesions and the reconstructed lesion sites for all five rabbits. In all subjects, the most part of the right ventral quadrant was not damaged, whereas three other quadrants were destroyed to a large extent. After these lesions, the right reticulospinal tract (descending in the ventro-medial and ventro-lateral areas) (see Blessing et al. 1981) was not damaged as well as the right vestibulospinal tract descending in the ventro-medial area (Akaike and Westerman 1973; Blessing et al. 1981). In rabbits 45, 49, and 60 (and to a lesser extent in rabbit 50), the medial part of the left ventral quadrant remained undamaged, suggesting preservation of a small portion of the left vestibulospinal tract. In rabbit 61, a small ventral part of the right dorsal quadrant was left intact. In rabbit 60, a small area around the central canal remained intact. There was no evident correlation between the extent of the lesion and the postural deficit. For example, the lesion in rabbit 45 (with a good recovery of corrective movements) had no specific difference as compared with rabbit 49 in which corrective movements did not recover.

POSTURAL CONFIGURATION.

A 3/4 section caused changes in the position of the hindlimbs in relation to the trunk. In control, both hindlimbs were turned outward (Fig. 5A), that is, their angles were positive. After the 3/4 section, they were turned inward (Fig. 5B). Figure 5C illustrates changes of the foot angles for one of the rabbits. In the first test after the 3/4 section (day 3), the foot angle in both limbs became negative. Over time, the foot angle in both limbs increased and became positive but did not return to its normal value until the end of observations (day 26). The inward turn of the foot as well as a very slow compensation of this deficit was observed in all five VQ rabbits, including the well-compensated rabbit 45. Figure 5E shows the population average of the foot angles in control, in the first test, and in the last test.

After the 3/4 section, the distance between toes decreased considerably because they were positioned closer to the midline; then the toes gradually approached their initial lateral position (Fig. 5, D and F). Also, after the 3/4 section, the limbs were positioned more caudally as compared with control (Fig. 5, A and B). Over time, the anterior/posterior position of the feet returned to its control value (Fig. 5, D and G).

In VQ rabbits, the trunk was slightly twisted and/or bent laterally. Because these distortions of the trunk configuration were small, they were not documented. In previous studies it was shown that in the animals with both ventral quadrants transected (ventral hemisection), the tone in extensor muscles gradually increased with time, causing an increasing extension of both hindlimbs (Lyalka et al. 2005). As a result, the height of the hindquarters above the supporting surface in the standing animal increased considerably (almost twice during 1 mo after lesion). This occurred to a much lesser extent in VQ animals. Figure 5H shows how the height (Ht) of the hindquarters in the standing position was measured, and Fig. 5I shows the height as a function of time in individual VQ rabbits. One can see a slight increase of the height over time. These changes in limb configuration were often accompanied by gradual development of contractures (Moryama et al. 2004), which, by the end of observation period, could reduce the range of possible movements in the hindlimb joints.

DISCUSSION

Significance of ventral spinal pathways for postural control

In our previous study (Lyalka et al. 2005), it was shown that the dorsal and ventral spinal pathways are of different importance for postural control in the hindquarters of the rabbit. After the dorsal hemisection (DHS) or lateral hemisection (LHS), postural corrective responses to tilts recovered in 1–3 wk, whereas after the ventral hemisection (VHS) they disappeared and did not recover. These experiments demonstrated the key role of the ventral spinal pathways for the generation of postural corrections. They also showed that two undamaged ventral quadrants, or one ventral plus one dorsal quadrant, are sufficient for postural corrections to recover. By contrast, two dorsal quadrants are not sufficient for recuperation of postural functions.

A common part that was spared in both types of SCI (DHS and LHS), in which postural control recuperated, was one ventral quadrant. Is one ventral quadrant both necessary and sufficient for postural control to recover? To answer this question, in the present study we examined the postural capacity of rabbits after the 3/4 section of the cord that spared only one ventral quadrant (VQ animals). In all rabbits, the initial effects of lesion were similar. They included disappearance of corrective motor responses to tilts, a dramatic decrease of the proportion of correct EMG responses, and appearance of incorrect EMG responses, which were never observed in intact rabbits (Fig. 4). In four of five VQ rabbits, corrective motor responses did not re-appear during the whole period of observation (≤1 mo). In these animals, the proportion of correct EMG responses was very small (<30%, Table 1).

A different group of VQ rabbits (n = 6) was used in the studies of the effect on postural recovery produced by postural training as well as by electrical and pharmacological stimulation of the spinal cord (Lyalka et al. 2007, 2008). In all these animals, the 3/4 section also resulted in abolishment of corrective movements, which did not recover during the period of observation; the proportion of correct EMG responses was also very small. One can therefore conclude that in the overwhelming majority of cases, postural corrective movements disappeared after the 3/4 section and did not recover. These findings suggest that a single ventral quadrant contains the amount of fibers that is insufficient for activation of spinal postural mechanisms in the overwhelming majority of VQ rabbits.

However, in one of the rabbits (45), not only the EMG responses recovered considerably, but also the motor corrective responses became close to normal (Fig. 2C). On the other hand, no significant difference in the degree of lesion was found between rabbit 45 and other animals (Fig. 4). However, because exact quantification of the extent of the lesion was difficult to perform, one cannot exclude the possibility that, in rabbit 45, the lesion had damaged fewer descending fibers important for the postural control than in other rabbits. One can also suggest that because of unknown reasons, rabbit 45 had a higher efficacy of descending spinal pathways or a higher excitability of spinal reflexes. As a result, the unilateral supraspinal drive (coming through the remaining spinal pathways, most likely reticulo- and vestibulospinal) was sufficient for bilateral activation of postural mechanisms in that rabbit.

Bilateral action of individual reticulospinal neurons is well documented (Drew 1991; Drew and Rossignol 1990a,b). Their axons may cross the mid-line to innervate both sides of the lumbar enlargement (Matsuyama et al. 1993). It is also known that some reticulospinal neurons innervate commissural interneurons that provide an indirect pathway to affect contralateral activity (Jankowska et al. 2003; Matsuyama et al. 2004), which could explain the bilateral nature of postural deficit in VQ rabbits.

It seems likely that, when the other ventral quadrant (in DHS animals) or the dorsal quadrant (in LHS animals) is intact as well, the amount of posture-related fibers (including fibers of the main descending systems, i.e., vestibulo-, reticulo-, rubro- and corticospinal ones) substantially exceeds the critical value, and this is a reason for recovery of postural control in DHS and LHS subjects (Lyalka et al. 2005).

Ventral spinal pathways (most likely reticulo- and vestibulospinal) are also known to be of crucial importance for the initiation and control of locomotion (Afelt 1974; Brustein and Rossignol 1998; Eidelberg et al. 1981; Noga et al. 1991; Steeves and Jordan 1980). Ventral pathways are thus involved in the formation of gross motor synergies controlling two automatic motor behaviors, locomotion and posture.

Spinal versus supraspinal origin of postural responses

The posture of hindquarters is controlled by two closed-loop nervous mechanisms (Deliagina et al. 2006, 2008; Horak and Macpherson 1996; Lyalka et al. 2005). One mechanism resides in the spinal cord, and its functioning is based on the spinal postural reflexes. The other mechanism contains a long reflex loop involving brain centers. Both mechanisms are driven by input from limb mechanoreceptors and compensate for postural disturbances by generating corrective motor responses. The relative contribution of the spinal and supraspinal mechanisms to the generation of postural corrections is not clear, however (for discussion, see Deliagina et al. 2006, 2008; Lyalka et al. 2005).

Recovery of postural corrective motor and EMG responses was observed in rabbit 45 and, to some extent, in rabbit 49 (Fig. 4). The 3/4 section, which damages more than half of ascending and descending fibers, apparently causes a dramatic decrease, first, in the ascending posture-related somatosensory signals from the hindlimbs and, second, in the descending motor commands for postural corrections. It seems difficult to explain the persistence of postural responses in a part of VQ animals by the operation of this heavily damaged long-loop mechanism. It seems more likely that the spinal reflex machinery plays an important role in the recovery of postural control after the 3/4 section. We suggest that after the lesion, the spinal postural circuits could be activated bilaterally by a unilateral tonic drive through the pathways descending in the intact ventral quadrant. It seems likely that the significant recovery of postural reflexes in rabbit 45 (Fig. 2C) is associated with an increased efficacy of this activating drive. If so, attempts to substitute this supraspinal drive in spinal subjects by pharmacological or electrical stimulation of the cord below the lesion seem plausible. These methods appeared efficient for restoration of locomotor function in spinal animals (e.g., Gerasimenko et al. 2008; Musienko et al. 2007; Rossignol et al. 1998, 2001).

Postural mechanisms damaged by 3/4 section

In intact rabbits, EMG responses were uniform in all tilt cycles—the hindlimb extensors (Gast and Vast) were activated with the ipsilateral tilt, and the flexor (Bic) was activated with the contralateral tilt (Fig. 2A). These responses were termed the correct ones.

By contrast, in VQ rabbits, the EMG responses to tilts were not uniform: they were correctly phased in some tilt cycles and incorrectly phased or absent in other cycles. The VQ rabbits differed in the proportion of correct responses, which ranged from 55% in the best-compensated rabbit 45 to 5–30% in all other rabbits (see Fig. 4 and Table 1). In rabbits 50, 60, and 61, the correct EMG responses and the incorrect ones were equally frequent. In the badly compensated rabbits, the correlation between responses in different muscles was weak (Fig. 2B). These two phenomena, i.e., inconsistency of the response pattern with repetitive application of the same somatosensory stimulus in a given subject, and a great variety of response patterns elicited by the same stimulus over the subjects with similar spinal damage, is also characteristic for humans (Harkema 2008).

Postural responses to tilts in VQ-animals can be considered (at least partly) as spinal reflexes caused by afferent signals coming from limb mechanoreceptors (Deliagina et al. 2006). A low probability and small value of correct (properly phased) EMG responses in the majority of VQ rabbits as well as appearance of incorrect responses, and spontaneous switches between correct and incorrect responses suggest that the spinal reflex chains, generating postural corrections, have not been specifically selected by the reduced supraspinal drive. As a result, afferent input evokes not only correct postural reflexes but also incorrect ones, and their relationships depend on spontaneous changes in the excitability of spinal networks. For further discussion of the issue that SCI causes distortions of spinal functional connections, see Cai et al. (2006), Edgerton et al. (1997), and Frigon and Rossignol (2006).

Besides the reduction or abolishment of postural reflexes, the 3/4 section caused distortions in the postural body configuration. They included more medial and caudal position of the hindlimbs as compared with control as well as smaller feet angles. These distortions very slowly decreased with time (Fig. 5) as was also shown previously for the distortions caused by hemisections of the spinal cord (Lyalka et al. 2005). A clear-cut scoliosis (Herman et al. 1985) gradually developed in rabbits with the lateral hemisection (Lyalka et al. 2005), whereas in VQ animals it developed to a much lesser extent. Distortions of body and limbs configuration were observed in all rabbits, independently on the degree of impairment of postural responses. A possible explanation for this finding could be a different origin of the two types of distortions—the distortions of postural responses seem to be caused by the damage to spinal reflexes, whereas distortions in the body configuration seem to be caused by imbalance in supraspinal tonic drive to various components of the spinal postural network (Lyalka et al. 2005).

GRANTS

This work was supported by grants from National Institute of Neurological Disorders and Stroke Grant R01 NS-049884, the Swedish Research Council (11554), and Gösta Fraenckels Foundation to T. G. Deliagina.