Joint-specific changes in locomotor complexity in the absence of muscle atrophy following incomplete spinal cord injury
© Hillen et al.; licensee BioMed Central Ltd. 2013
Received: 26 September 2012
Accepted: 26 July 2013
Published: 15 August 2013
Following incomplete spinal cord injury (iSCI), descending drive is impaired, possibly leading to a decrease in the complexity of gait. To test the hypothesis that iSCI impairs gait coordination and decreases locomotor complexity, we collected 3D joint angle kinematics and muscle parameters of rats with a sham or an incomplete spinal cord injury.
12 adult, female, Long-Evans rats, 6 sham and 6 mild-moderate T8 iSCI, were tested 4 weeks following injury. The Basso Beattie Bresnahan locomotor score was used to verify injury severity. Animals had reflective markers placed on the bony prominences of their limb joints and were filmed in 3D while walking on a treadmill. Joint angles and segment motion were analyzed quantitatively, and complexity of joint angle trajectory and overall gait were calculated using permutation entropy and principal component analysis, respectively. Following treadmill testing, the animals were euthanized and hindlimb muscles removed. Excised muscles were tested for mass, density, fiber length, pennation angle, and relaxed sarcomere length.
Muscle parameters were similar between groups with no evidence of muscle atrophy. The animals showed overextension of the ankle, which was compensated for by a decreased range of motion at the knee. Left-right coordination was altered, leading to left and right knee movements that are entirely out of phase, with one joint moving while the other is stationary. Movement patterns remained symmetric. Permutation entropy measures indicated changes in complexity on a joint specific basis, with the largest changes at the ankle. No significant difference was seen using principal component analysis. Rats were able to achieve stable weight bearing locomotion at reasonable speeds on the treadmill despite these deficiencies.
Decrease in supraspinal control following iSCI causes a loss of complexity of ankle kinematics. This loss can be entirely due to loss of supraspinal control in the absence of muscle atrophy and may be quantified using permutation entropy. Joint-specific differences in kinematic complexity may be attributed to different sources of motor control. This work indicates the importance of the ankle for rehabilitation interventions following spinal cord injury.
In order to target and quantify the overall pattern of changes in locomotion (particularly the results of decrease in supraspinal control), we assessed 3D locomotor kinematics and kinematic complexity in rats following mild-moderate incomplete thoracic spinal contusion. As a result of spinal cord injury, the connections between the brain and the spinal circuitry below the injury are disrupted. This leads to adaptations in the neurons of the brain and spinal cord as well as changes to the sensory afferents and motoneurons [1–3]. Along with these neural changes, the muscles in the distal limb undergo changes similar to those seen in many disuse paradigms. For example, the muscles tend to atrophy and muscle fibers shift towards faster twitch, more fatigueable ones [4–7]. These effects on muscle properties have been well studied in humans and other animals for a variety of disuse paradigms including weightlessness, bed rest, stroke, partial body support and constrained limbs [8–10]. As spinal cord injury related muscular and neural impairments affect the legs, locomotion is often used as a measure of impairment and recovery.
Both musculoskeletal and nervous system impairments may contribute to the behavioral impairments seen following incomplete SCI (iSCI). While one might expect aberrant neural control to dominate locomotor impairments following iSCI, data in humans has suggested that timing of voluntary initiation of ankle movement (control) is unaltered and loss of muscle strength (specifically dynamic muscle strength) may be responsible for maladaptive changes in ankle gait patterns [11, 12]. However, dynamic muscle strength (for instance, time to maximal contraction) may rely on neural control, not solely muscle strength. Additionally, neural control of muscle in an isolated, voluntary, movement is significantly different than movement during gait .
Following stroke, the overarching control system for gait is simplified. Even in normal individuals, synergies in muscle activations during locomotion are common . For instance, knee extensors and hip abductors activate at the same time to provide body support during stance . While these synergies exist in normal subjects, they are numerous and varied. Following stroke, these synergies collapse to just one pair: one for flexion and one for extension of the whole limb, indicating a decrease in motor control complexity . This decrease in complexity may be due to unmasking of the primitive gait controller in the spinal cord when there is reduced input from supraspinal centers . This idea is one of some debate however, with an alternative hypothesis that these synergies are purely due to constraints in the task being performed .
As iSCI also results in disruption of supraspinal motor control of the limbs, a similar muscle synergy and loss of complexity effect could be seen following injury. There is evidence that much of the coordination associated with locomotion resides in the spinal cord from studies in deafferented and decerebrate preparations , but no current study has assessed this effect in SCI subjects during locomotion. Existing reviews on the subject of motor synergy following spinal cord injury also use the stroke framework  due to the sparsity of information regarding spainal cord injury. Initial reports on muscle synergies following spinal cord injury have assessed hand movements for activities of daily living and, while differences are noted, change in quantity has not been reported .
There are a number of measures that can be used to analyze time series data such as joint angle locomotor data for synergy and synchronization. Synchronization of cyclic joint movement can be measured with principal components analysis (PCA) and alterations in phase delay between the joint motions. PCA assesses the quantity of information in the signal and how similar time series are to one another . As the joints (hip, knee, and ankle) become synchronized, fewer and fewer components would be required to describe the motion of all three joints. Joint synchronization can also be measured by examining the phase delays between the activation of joints within a limb. As the phase delay becomes closer to 0 (exactly in phase), the joint motions become more synchronized. Kinematic complexity can be measured on a joint by joint basis by looking at permutation entropy of the joint angle time series. This measure assesses the quantity of the information in the signal by identifying how likely it is for the joint angle to continue in the same direction from one time point to the next .
The rat model is currently one of the most studied and best understood models of SCI being used [23, 24]. Many researchers have studied the effects of SCI in the rat on a number of functional impairments and measured the effects of injury at multiple levels of physiology  and behavior . The general time course of rat hindlimb locomotor recovery following SCI has been characterized using the Basso, Beattie, Bresnahan locomotor recovery score . However, there has been relatively little research into specific hindlimb kinematics, with some notable exceptions [27–30]. Also, little information has been collected on the muscles in the rat hindlimb, usually only a few muscles per study [4, 31, 32].
This study addresses the hypothesis that iSCI impairs gait coordination and decreases locomotor complexity at each hindlimb joint and the entire hindlimb. This is accomplished by assessing complexity of 3D joint kinematics during treadmill locomotion and measuring muscle parameters in rats four weeks following either a sham injury or a mild-moderate spinal cord contusion injury. To our knowledge, no previous study has analyzed changes in kinematics complexity along with known levels of muscle atrophy. The analyzed data indicate that spinal contusion injury leads to impaired coordination and decreases in locomotor complexity on a joint-specific basis. The absence of changes in muscle parameters indicates that these locomotor changes can be due purely to losses in supraspinal drive.
Experimental data were collected from 12 female, adult, Long-Evans rats (270–300 g). Six rats received sham injuries and six received a mild-moderate spinal cord contusion injury. All rats were kept on a 12 hour light/dark cycle with ad libitum food and water. The study was approved by the Institutional Animal Care and Use committee (IACUC) of Arizona State University and complies with the Guide for the Care and Use of Laboratory animals.
Spinal cord injury surgery
Rats were randomly selected to undergo T8 vertebral (T9 spinal) sham or mild-moderate incomplete contusion injury (iSCI). Surgery was performed under aseptic conditions similar to Scheff et al. . Rats were anesthetized under 1-2% isoflourane and given an injection of the analgesic buprenorphine (0.02–0.05 mg/kg). The vertebral process of T8 was removed and a U-shaped laminectomy was performed. For animals in the contusion (iSCI) group, T7 and T9 were clamped in place and a mild-moderate contusion (154 ± 3 kdynes SD) was performed with an IH Instruments force controlled impactor for a T9 spinal level contusion. Following contusion or laminectomy, the muscles were closed in layers using resorbable sutures and the skin closed with wound clips (which were removed 1–2 weeks following surgery). The animal was then given injections of 5 cc of sterile saline, and 33.3 mg/kg of Cefazolin antibiotic and allowed to wake slowly on a heated pad.
Animal care post-surgery consisted of twice daily bladder expression until the animal was able to void on its own along with twice daily injections of saline, buprenorphine (as above) and antibiotic (as above) for the first 7 days following injury. Saline administration was continued if the animal remained dehydrated and antibiotic administration continued an additional week if the animals showed signs of urinary tract infection from a urinalysis test strip. Following surgery, animals were allowed to move freely in their cages for 4 weeks. Behavioral analysis consisting of the Basso, Beattie, and Bresnahan (BBB) 21 point locomotor score  was collected on the animals every day for the first week and each week thereafter in order to verify injury severity. Rats were observed for 4 min by two experimenters in order to score hindlimb function. The scale takes into account milestones representative of locomotor recovery including motion in the leg, paw position, coordination and weight support. Experimenters were not blinded to the group of the animal as the differences were clearly evident to trained individuals.
Hindlimb treadmill kinematics
Prior to the spinal cord surgery, animals were allowed to acclimatize to their new environment for one week and trained to walk on a treadmill (Columbus Instruments) for 2 days for 10 min each at a progression of speeds from 0–21 meters/min. 3D kinematic data was collected 4 weeks post injury on both sets of animals as described previously . Briefly, animals were anesthetized under 1-2% isoflourane, shaved as necessary, and cone-shaped reflective markers were attached bilaterally to the bony prominences of the ankle (lateral malleolus), knee (femoral condoyle), hip (greater trochanter and iliac crest), shoulder (greater tubercle), elbow (lateral epicondyle), and wrist (ulnar head). In addition, a strip of reflective tape was placed around the fifth metatarsal of the hindlimb, close to its attachment. The markers were not spherical, as is desirable for centroid calculations, but their small size (approximately 5 mm) minimized any error. Markers were placed while the animal lay on its side. Video was recorded from 4 infrared-sensitive cameras with co-localized infrared light sources (2 on each side of the animal) at 60 Hz. A 36 point static calibration object was first recorded. Error in the system was less than 0.5% in position and less than 1.5° in angle . Following calibration, the animal was placed on the treadmill. Kinematic data was recorded at treadmill speeds starting at 11 m/min, increasing to 21 m/min in 2 m/min increments. The speed was increased each time the animals completed at least 5 continuous stable step cycles (in the middle of the treadmill at a constant speed). The entire process took approximately 5–10 min per animal. If an animal was unable to complete all speeds on the first day, the task was repeated the next day. No animal took more than 2 days to complete all speeds.
Video was imported into Vicon Peak-Motus® software for analysis. The calibration object was digitized for odd and even fields to account for the de-interlacing of the frames done in Peak-Motus. A segment containing 5 stable step-cycles at 21 m/min was identified and imported into Peak-Motus. Video was manually synchronized using timestamps on the video. All reflective markers were digitized and the data was processed into joint angle versus time and marker 3D position data versus time. Gait cycles, swing, and stance phases were identified by marking lift-off (the first frame where the toe was not contacting the treadmill) and touchdown (the first frame where the toe contacted the treadmill) events for each limb. Joint angles and hindlimb coordination were assessed for changes with time. Data was scaled to the hindlimb cycle (with 200 points per cycle) and averaged across all cycles per animal. Specific joint angle values were also collected for swing and stance maxima and minima, the highest and lowest (respectively) values for joint angles during those gait cycle phases. Subcycle values are useful as neural control varies between stance and swing for reasons including loading and other afferent signals. Data from two of the animals (one sham and one iSCI) proved to be incompatible with the kinematic analysis software (most likely due to footfall event marking) and were excluded from most analyses. Unless specifically mentioned, n=5 animals per group were used.
In addition to displaying each joint independently, angle-angle plots were created to display information about coordination. Intralimb coordination was displayed as hip vs. knee, hip vs. ankle, knee vs. ankle, and shoulder vs. elbow. Interlimb coordination was displayed for each joint, left vs. right. These plots can be used for qualitative assessment of a number of features of locomotion [27, 34]. A vertical or horizontal line represents movement in one joint while the other is held constant. A continuously changing phase relationship between the two joints is indicated by diagonal segments, with negative slopes indicating an out-of-phase relationship and positive slopes indicating an in-phase relationship. Finally, in the interlimb coordination plots only, symmetry around the y=x axis can be used to assess symmetry in the joint kinematics.
θR i is the angle of the right side joint at point i, θL( i+hp ) is the angle of the left side joint at the point one half the cycle period ahead of the right side, and n is the number of points in the cycle.
As mentioned, PE was calculated for both averaged trajectories (a compilation trajectory of both left and right sides for each joint) and raw, unaveraged trajectories for each joint separately. Unlike Olofsen, we only used Tau=1 (three point motifs), not a combination with Tau=2 (six point motifs) as there are not multiple wave frequencies in gait as there are in EEG.
In the context of time series data, principal component analysis (PCA) is used to assess the similarity of waveforms with one another. It is used to reduce the dimensionality of data or to determine the highest sources of variation within the data . PCA was used to determine the loss in complexity in overall hindlimb gait following injury. If more variance in the gait data is accounted for by fewer principal components, then this may indicate that gait complexity has reduced, and the joints are moving more in synchronization with one another. When joints move in synchronization, the independence of the neural signals is reduced, thus fewer control signals are required to produce the movement. The reduction in needed control signals is a reduction in complexity. Thus, as the value for the first principal component (PC1, the percentage of variance account for by that component) increases, the complexity of the system decreases. As PC1 increases, less variance is available for remaining components. In the case of hindlimb angle trajectories, as PC1 approaches 100%, the three joints of the hindlimb become in phase with one another such that all three joints are flexing and extending together as opposed to each joint being controlled independently. This would be a much simpler pattern than is seen in normal subjects. Analysis was performed using the PRINCOMP function in MATLAB on the set of three hindlimb angle trajectories for both hindlimbs for each animal then averaged across groups (sham and iSCI). When using the three hindlimb joint trajectories instead of multiple subjects, the maximum number of the principal components is 3, so data reduction is limited. Instead of counting the number of principal components required to reach a specified variance, amount of variance accounted for by the first principal component was assessed. The method used was similar to those performed on a single joint angle across subjects [42–44], but the three hindlimb joint angles were compared as opposed to multiple subjects. The different time points represent the different “variables” in a standard PCA analysis. PCA was performed on each hindlimb of each animal and the proportion of variance accounted for by each of the three principal components was obtained.
where d is the sarcomere length, θ is the angle of diffraction, n is the diffraction order, and λ is the laser wavelength.
Differences in parameters between sham and iSCI groups were analyzed using a Wilcoxon-Mann–Whitney non-parametric test with p = 0.05 due to the small sample size. PE and PCA measures were conducted using standard t-tests, because in cases where both hindlimbs were measured independently, left and right limb values were considered repeated measures for data analysis. Variability has been reported as standard error (SEM) for data in which multiple cycles have been averaged together and standard deviations (SD) in all other cases. All statistical analyses were run in SAS 9.2 (SAS institute Inc, Cary, NC).
Open field locomotion indicates mild-moderate injury
Passive muscle parameters indicate a lack of muscle atrophy
Hindlimb muscle properties in sham and iSCI rats
Muscle weight (g)
Pennation angle (deg)
Fiber length (mm)
Sarcomere length (mm)
0.80 ± 0.10
0.82 ± 0.10
29.7 ± 9.0
23.3 ± 5.6
10.63 ± 4.53
7.77 ± 0.62
1.99 ± 0.15
2.12 ± 0.23
1.24 ± 0.20
1.18 ± 0.11
31.0 ± 7.4
27.3 ± 7.6
10.92 ± 6.57
9.93 ± 2.36
2.03 ± 0.07
2.10 ± 0.19
0.18 ± 0.03
0.18 ± 0.02
24.3 ± 11.2
21.0 ± 1.4
9.81 ± 3.04
8.69 ± 3.70
2.23 ± 0.11
2.14 ± 0.23
0.69 ± 0.06
0.66 ± 0.03
32.3 ± 11.6
26.3 ± 9.0
13.32 ± 3.16
11.72 ± 3.98
2.24 ± 0.32
2.07 ± 0.25
1.13 ± 0.09
1.22 ± 0.13
28.9 ± 4.6
27.3 ± 6.6
12.78 ± 1.08
13.67 ± 2.21
2.09 ± 0.06
2.09 ± 0.39
1.60 ± 0.08
1.51 ± 0.17
21.8 ± 17.9
22.5 ± 3.5
11.96 ± 4.21
14.76 ± 2.63
2.24 ± 0.40
2.18 ± 0.06
0.88 ± 0.15
0.86 ± 0.07
27.3 ± 2.1
34.2 ± 4.3
8.25 ± 1.59
8.19 ± 1.87
2.02 ± 0.19
1.95 ± 0.15
1.28 ± 0.23
1.22 ± 0.27
0.0 ± 0.0
0.0 ± 0.0
29.96 ± 3.49
26.18 ± 5.15
2.12 ± 0.18
2.05 ± 0.14
2.09 ± 0.33
2.12 ± 0.27
0.0 ± 0.0
0.0 ± 0.0
26.92 ± 1.70
22.45 ± 3.69
1.99 ± 0.17
2.03 ± 0.06
1.02 ± 0.24
0.90 ± 0.06
0.0 ± 0.0
0.0 ± 0.0
25.87 ± 9.85
22.70 ± 5.77
2.01 ± 0.09
1.94 ± 0.17
3D locomotor kinematics indicate decrease in complexity
PCA analysis quantified the complexity of movement for the entire hindlimb by assessing the three hindlimb angle trajectories for each individual animal’s kinematics. Analysis was performed on the unaveraged joint kinematics of each leg of each animal and averaged together for each group (sham and iSCI). Each component’s contribution to the overall variance of the hindlimb kinematics was compared between groups in Figure 6B. While no significant results are noted, there is a trend (p=0.1083) toward an increase in PC1, which would indicate a decrease in complexity in the iSCI group. The lack of significance may be due to the small sample size and minimal impact to gait by the 4th week following a mild-moderate injury. The first components of the PCA analysis most likely account for the gait cycle and the coordination between the flexors within the limb (and likewise the extensors) while the second and third components represent the intricacies of the joint trajectories. As the first component accounts for more and more of the variance, the three joints of the limb become more in phase with one another such that they could be controlled by a single neural signal.
In this study, data were collected from a set of mild to moderately injured rats in the chronic phase of recovery (4 weeks post-injury) where muscle atrophy was negligible, but significant locomotor impairments were still seen during walking. These impairments resulted in changes in kinematic complexity due to changes in neural control. The ankle specifically was particularly sensitive to loss of supraspinal control. No study to date has shown a decrease in kinematic complexity following incomplete spinal cord injury in rat, particularly when overall gait recovery was as complete as shown here. We also showed that changes in kinematic complexity were joint specific, indicating that different joints are under differing control in locomotion. Finally we collected muscle parameters from iSCI rats showing no indication of atrophy.
Lack of change in muscle parameters
Muscle parameters were collected to determine presence or lack of atrophy for iSCI rats four weeks post injury. Gregory et al.  looked at muscle disuse following complete transection injury in order to assess interspecies differences between rats and humans, but only assessed fiber type changes, not atrophy. Liu et al.  compared cross sectional area of hindlimb muscles for 18 weeks following a contusion injury (BBB: 11.1). They showed muscle atrophy in all muscles for the first 3 weeks following injury, with a gradual return to pre-injury values starting at week four, with all muscles recovering by 12 weeks. As their injury level was slightly higher than in this study and a different strain was used (Sprague–Dawley), the minor difference in results is consistent with our findings.
In a study of contusion injured rats, Hutchinson et al.  measured changes in muscle masses following injury. They used a higher level of injury (approximately 4–5 points lower in the BBB) and observed atrophy (~19%) in the gastrocnemius even 10 weeks post injury. In the current study, higher BBB scores were observed, indicating a lower level of injury. This led to weight bearing occurring a few days earlier, thus ameliorating the effects of the injury on muscle atrophy . Also of note, the rats in the current study gained weight following injury, while the rats in Hutchinson et al.’s study lost weight, possibly leading to an increase in atrophy rate. When compared to a weight-matched set of sham controls, the muscle masses for rats in the current study are approximately 1% (TA) to 23% (soleus) greater than those reported by Hutchinson. This could be due to differences in strain.
Changes in locomotor coordination in absence of atrophy
While similar data reduction techniques applied to a large number of EMG signals following stroke have been successful , PCA has not been applied to rat kinematic data in this fashion before. The analysis technique could be an important tool to assess gait with higher levels of injury or larger sample sizes. The lack of significant results could be complicated by the fact that the 3 joint trajectories are from a single animal so the results may appear different than those used in other types of PCA analysis of joint kinematics where each joint (hip, etc.) is assessed independent from one another .
In the intra-limb coordination plots (with the exception of the Knee-Hip plot), the waveforms appear to collapse to a diagonal in the injured animals, implying that the joint motions for injured animals become coordinated around just two control signals (flexion and extension) as seen in stroke . This is most noticeable in the Ankle-Hip plot. Increased joint synchronicity in injured animals is also supported by the hindlimb phase delay data, which showed a decrease in the Hip-Ankle phase delay in the iSCI animals; the peaks of the two angle trajectories came closer to one another in time. Ankle and knee joints, however, became more out of phase with one another following injury. Thus the decrease in phase delay between the hip and ankle (which would indicate a decrease in complexity in the PCA) may have been counteracted by the increase in phase delay between the ankle and knee (which would indicate an increase in complexity in the PCA).
As iSCI animals in this study did lose the second ankle angle peak but did not have muscle atrophy, this loss of kinematic complexity is most likely due to loss of supraspinal control (a delay in gastrocnemius EMG burst initiation ) and not loss of muscle strength. Finally, as indicated in Figure 4, the overextension in the ankle (likely due to the loss of supraspinal control ) and lower range of motion in the knee tended to counteract each other, indicating that the remaining impaired control signals to the hindlimb can be used to establish correct foot position .
The cruciform pattern seen in knee-knee intralimb coordination (Figure 8) is not unique to this study. In Jung et al. , rats were given the same level of injury as reported in this study. One week following the injury they were given 5 days of patterned electrical stimulation to the hip muscles in order to produce a locomotor pattern at the hip. 3D kinematics were then collected 14 days post injury (dpi). In both the current study and Jung, BBB scores indicated that the animals were nearly recovered; however, the animals permanently altered their gait as indicated by kinematics. The post-therapy animals developed a cruciform pattern in the Knee-Knee angle plots  (14 dpi, 7 days post therapy) similar to that observed in the iSCI animals in the current study at 28 dpi, suggesting that the stimulation therapy accelerated recovery following injury. While this pattern was different than that seen in normal animals, the animals were able to reach the targeted treadmill speeds with ease. Along with this similarity was one in forelimb-hindlimb coordination. In both Jung and the current study, forelimb-hindlimb coordination shifted from approximately 50% out-of-phase to 100% out-of-phase. As this gait is stable and effective, it is a successful endpoint for locomotor recovery despite not matching pre-injury patterns.
Joint-specific changes in locomotor complexity
While PCA analysis did not prove a decrease in complexity in the whole limb kinematics, PE analysis of individual joint kinematics showed joint-specific changes in complexity. PE also showed itself to be a quantitative measure of changes in joint angle trajectory. Changes in PE at the ankle were accompanied by changes in ankle angles. While analysis of control of locomotion tends to focus on the joints as similar structures (i.e., receiving similar input from sensors, supraspinal sources, and the central pattern generator (CPG) ), the results of this study emphasize the importance of also considering specific roles for individual joints [54–56]. In general, all joints do receive similar input from all sources of locomotor control. A tonic drive from the brain initiates gait and provides feed-forward adaptations to perturbations [57, 58]. The CPG and pattern shapers produce the basic profile for gait . Sensory information from muscles spindles, Golgi tendon organs, and cutaneous receptors modify the gait pattern [53, 60]. A number of specific aspects of gait are less general than this however, with each joint being affected differently by feed-forward supraspinal control or reflex sensory input during different phases of the gait cycle. Data from guinea fowl show a proximal to distal gradient in neuromechanical control; proximal muscles use more feed-forward control and the distal muscles less .
Ankle muscle amplitude is as much as 70% controlled by local sensors  which may be even greater when supraspinal control is decreased . This may be due to a very high proprioceptive gain in the ankle muscles [63, 64]. Due to the high reliance of the ankle on sensory input for motor control, the lack of significant muscle forces or cutaneous input during swing may contribute to the ankle flexor muscles (Tibialis Anterior) continued activation until the end of stance. This would lead to the lack of the normal end swing ankle extension that was noted in the iSCI animals in this study. Studies in humans indicate that the ankle has little to no feed-forward control following incomplete injury while the hip still maintains a large feed-forward component . In the absence of planned foot placement, the ankle stays flexed for the duration of swing and loses its second minima/maxima. This hypothesis is further supported by the type of injury sustained. The contusion injury model used strikes the dorsal surface of the cord leading to significant damage to the dorsal tracts . In the rat, the primary dorsal tracts are the dorsal corticospinal tract, the fasciculus gracilis (trunk and hindlimbs) and the fasciculus cuneatus (forelimbs) [65, 66]. In rats, the corticospinal tract serves a very limited purpose, only directly controlling individual digit movements [67, 68]. However, it also serves to modulate lumbar stretch reflexes . As the ankle is primarily controlled by local reflexes, damage to the dorsal corticospinal tract would lead to significant changes in ankle control.
As with the initiation of foot swing, the ankle tends to act as a controller for the hip, with ankle-foot loads modulating hip torques [54, 69]. The lack of feed-forward control at the ankle may serve a positive purpose. In guinea fowl, feed-forward planning for perturbations during overground locomotion has a more negative effect on gait than an unplanned perturbation .
One of the major functions of the knee is to couple the hip to the ankle. It contains many multiarticular muscles, coupling it with the proximal and distal joints. Knee extensors are synergistic with ankle extensors and inhibit ankle flexors . The extensors in the hip are coupled to the knee via the stretch reflex . Changes in the knee joint angle excursions are then likely due to the changes in control of the other joints . Further investigations will thus be necessary to determine the actual control mechanism at play at each joint and the overall limb.
There are three major conclusions in the paper. 1.) Locomotor changes (particularly decrease in locomotor “complexity”) following mild contusion injuries in rats happen in the absence of muscle changes allowing us to minimize the contributions of muscle changes in interpretation of the results. 2.) Changes in gait parameters are joint-specific, with more changes occurring the more distal the joint. 3.) Permutation entropy may be used to quantify differences in joint angle trajectories in a way that may be relevant to motor control (complexity and related quantity of neural control). This is the most evident when looking at the loss of the second local maxima in the ankle trajectory, but the measure can detect smaller changes as well.
In this study, treadmill walking was used to assess locomotor capabilities. The data showed that changes in kinematic complexity were joint-specific, indicating that different joints are under differing control in locomotion. Specifically, the ankle showed a decrease in complexity of movement, likely due to its unique role in locomotion. The ankle showed a decrease in kinematic complexity accompanied by changes in other ankle kinematic measures, supporting the use of the PE measure for gait analysis. Animals in this study showed a complete lack of changes in measured muscle parameters. This may indicate that weight bearing within 6 days following injury is sufficient to ameliorate the disuse-induced changes in the muscles. To our knowledge, no previous study has analyzed changes in kinematic complexity along with known levels of muscle atrophy.
Despite the injury, the animals do achieve stable and efficient gait, able to reach the desired 21 m/min speed used for data collection. Following the injury, communication between the hindlimbs and the brain is impaired and locomotor control shifts predominantly towards that of the local circuitry of the spinal cord and sensors. As such, a more primitive gait is unmasked. While rats may regain effective locomotion following injury, brain-spinal communication is never fully restored and thus the animal must adapt to use the local hindlimb circuitry for locomotion to become proficient in gait. This data suggests that evaluation of hindlimb EMG, specifically timing of bursts, may further demonstrate the loss of complexity in movement seen following injury, particularly at the ankle. This work may provide the basis for new strategies for rehabilitation interventions following spinal injury or modify existing ones by changing the rehabilitation parameters on a joint-specific basis, for instance, adding active ankle control to existing robotic interventions which currently use passive ankle control. Additionally, the study provides a novel set of muscle data that could be useful in the development of an empirically derived neuromusculoskeletal computational model for the rodent hindlimb.
The authors thank Tsukasa Kanchiku for training in the surgical methods. This work was supported by NIH R01-NS054282.
- Bareyre FM, Schwab ME: Inflammation, degeneration and regeneration in the injured spinal cord: insights from DNA microarrays. Trends Neurosci 2003, 26: 555-563. 10.1016/j.tins.2003.08.004PubMedGoogle Scholar
- Bradbury EJ, McMahon SB: Spinal cord repair strategies: why do they work? Nat Rev Neurosci 2006, 7: 644-653.PubMedGoogle Scholar
- Thuret S, Moon LD, Gage FH: Therapeutic interventions after spinal cord injury. Nat Rev Neurosci 2006, 7: 628-643.PubMedGoogle Scholar
- Hutchinson KJ, Linderman JK, Basso DM: Skeletal muscle adaptations following spinal cord contusion injury in rat and the relationship to locomotor function: a time course study. J Neurotrauma 2001, 18: 1075-1089. 10.1089/08977150152693764PubMedGoogle Scholar
- Talmadge RJ, Castro MJ, Apple DF Jr, Dudley GA: Phenotypic adaptations in human muscle fibers 6 and 24 wk after spinal cord injury. J Appl Physiol 2002, 92: 147-154.PubMedGoogle Scholar
- Stein RB, Chong SL, James KB, Kido A, Bell GJ, Tubman LA, Belanger M: Electrical stimulation for therapy and mobility after spinal cord injury. Prog Brain Res 2002, 137: 27-34.PubMedGoogle Scholar
- Shields RK: Muscular, skeletal, and neural adaptations following spinal cord injury. J Orthop Sports Phys Ther 2002, 32: 65-74.PubMedGoogle Scholar
- Bloomfield SA: Changes in musculoskeletal structure and function with prolonged bed rest. Med Sci Sports Exerc 1997, 29: 197-206.PubMedGoogle Scholar
- Bajotto G, Shimomura Y: Determinants of disuse-induced skeletal muscle atrophy: exercise and nutrition countermeasures to prevent protein loss. J Nutr Sci Vitaminol (Tokyo) 2006, 52: 233-247. 10.3177/jnsv.52.233Google Scholar
- Cornachione A, Cacao-Benedini LO, Shimano MM, Volpon JB, Martinez EZ, Mattiello-Sverzut AC: Morphological comparison of different protocols of skeletal muscle remobilization in rats after hindlimb suspension. Scand J Med Sci Sports 2008, 18: 453-461.PubMedGoogle Scholar
- Wirth B, van Hedel HJ, Curt A: Ankle dexterity is less impaired than muscle strength in incomplete spinal cord lesion. J Neurol 2008, 255: 273-279. 10.1007/s00415-008-0724-yPubMedGoogle Scholar
- Wirth B, Van Hedel HJ, Curt A: Changes in corticospinal function and ankle motor control during recovery from incomplete spinal cord injury. J Neurotrauma 2008, 25: 467-478. 10.1089/neu.2007.0472PubMedGoogle Scholar
- Maegele M, Muller S, Wernig A, Edgerton VR, Harkema SJ: Recruitment of spinal motor pools during voluntary movements versus stepping after human spinal cord injury. J Neurotrauma 2002, 19: 1217-1229. 10.1089/08977150260338010PubMedGoogle Scholar
- Ranganathan R, Krishnan C: Extracting synergies in gait: using EMG variability to evaluate control strategies. J Neurophysiol 2012, 108: 1537-1544. 10.1152/jn.01112.2011PubMed CentralPubMedGoogle Scholar
- Perry J: Gait analysis : normal and pathological function. Thorofare, NJ: SLACK; 1992.Google Scholar
- Clark DJ, Ting LH, Zajac FE, Neptune RR, Kautz SA: Merging of healthy motor modules predicts reduced locomotor performance and muscle coordination complexity post-stroke. J Neurophysiol 2010, 103: 844-857. 10.1152/jn.00825.2009PubMed CentralPubMedGoogle Scholar
- Yang JF, Gorassini M: Spinal and brain control of human walking: implications for retraining of walking. Neuroscientist 2006, 12: 379-389. 10.1177/1073858406292151PubMedGoogle Scholar
- Tresch MC, Jarc A: The case for and against muscle synergies. Curr Opin Neurobiol 2009, 19: 601-607. 10.1016/j.conb.2009.09.002PubMed CentralPubMedGoogle Scholar
- Markin SN, Lemay MA, Prilutsky BI, Rybak IA: Motoneuronal and muscle synergies involved in cat hindlimb control during fictive and real locomotion: a comparison study. J Neurophysiol 2012, 107: 2057-2071. 10.1152/jn.00865.2011PubMed CentralPubMedGoogle Scholar
- Giszter SF, Hart CB: Motor primitives and synergies in the spinal cord and after injury–the current state of play. Ann N Y Acad Sci 2013, 1279: 114-126. 10.1111/nyas.12065PubMed CentralPubMedGoogle Scholar
- Zariffa J, Steeves J, Pai DK: Changes in hand muscle synergies in subjects with spinal cord injury: characterization and functional implications. J Spinal Cord Med 2012, 35: 310-318. 10.1179/2045772312Y.0000000037PubMed CentralPubMedGoogle Scholar
- Shemmell J, Johansson J, Portra V, Gottlieb GL, Thomas JS, Corcos DM: Control of interjoint coordination during the swing phase of normal gait at different speeds. J Neuroeng Rehabil 2007, 4: 10. 10.1186/1743-0003-4-10PubMed CentralPubMedGoogle Scholar
- Onifer SM, Rabchevsky AG, Scheff SW: Rat models of traumatic spinal cord injury to assess motor recovery. Ilar J 2007, 48: 385-395. 10.1093/ilar.48.4.385PubMedGoogle Scholar
- Kwon BK, Oxland TR, Tetzlaff W: Animal models used in spinal cord regeneration research. Spine (Phila Pa 1976) 2002, 27: 1504-1510. 10.1097/00007632-200207150-00005Google Scholar
- Sedy J, Urdzikova L, Jendelova P, Sykova E: Methods for behavioral testing of spinal cord injured rats. Neurosci Biobehav Rev 2008, 32: 550-580. 10.1016/j.neubiorev.2007.10.001PubMedGoogle Scholar
- Basso DM, Beattie MS, Bresnahan JC: A sensitive and reliable locomotor rating scale for open field testing in rats. J Neurotrauma 1995, 12: 1-21. 10.1089/neu.1995.12.1PubMedGoogle Scholar
- Thota AK, Watson SC, Knapp E, Thompson B, Jung R: Neuromechanical control of locomotion in the rat. J Neurotrauma 2005, 22: 442-465. 10.1089/neu.2005.22.442PubMedGoogle Scholar
- Couto PA, Filipe VM, Magalhaes LG, Pereira JE, Costa LM, Melo-Pinto P, Bulas-Cruz J, Mauricio AC, Geuna S, Varejao AS: A comparison of two-dimensional and three-dimensional techniques for the determination of hindlimb kinematics during treadmill locomotion in rats following spinal cord injury. J Neurosci Methods 2008, 173: 193-200. 10.1016/j.jneumeth.2008.06.006PubMedGoogle Scholar
- Courtine G, Gerasimenko Y, van den Brand R, Yew A, Musienko P, Zhong H, Song B, Ao Y, Ichiyama RM, Lavrov I, et al.: Transformation of nonfunctional spinal circuits into functional states after the loss of brain input. Nat Neurosci 2009, 12: 1333-1342. 10.1038/nn.2401PubMed CentralPubMedGoogle Scholar
- Johnson WL, Jindrich DL, Roy RR, Edgerton VR: Quantitative metrics of spinal cord injury recovery in the rat using motion capture, electromyography and ground reaction force measurement. J Neurosci Methods 2012, 206: 65-72. 10.1016/j.jneumeth.2012.02.008PubMedGoogle Scholar
- Negredo P, Rivero JL, Gonzalez B, Ramon-Cueto A, Manso R: Slow- and fast-twitch rat hind limb skeletal muscle phenotypes 8 months after spinal cord transection and olfactory ensheathing glia transplantation. J Physiol 2008, 586: 2593-2610. 10.1113/jphysiol.2007.149120PubMed CentralPubMedGoogle Scholar
- Johnson WL, Jindrich DL, Zhong H, Roy RR, Edgerton VR: Application of a rat hindlimb model: a prediction of force spaces reachable through stimulation of nerve fascicles. IEEE Trans Biomed Eng 2011, 58: 3328-3338.PubMed CentralPubMedGoogle Scholar
- Scheff SW, Rabchevsky AG, Fugaccia I, Main JA, Lumpp JE Jr: Experimental modeling of spinal cord injury: characterization of a force-defined injury device. J Neurotrauma 2003, 20: 179-193. 10.1089/08977150360547099PubMedGoogle Scholar
- Winstein CJ, Garfinkel A: Qualitative dynamics of disordered human locomotion: a preliminary investigation. J Mot Behav 1989, 21: 373-391. 10.1080/00222895.1989.10735490PubMedGoogle Scholar
- Jung R, Belanger A, Kanchiku T, Fairchild M, Abbas JJ: Neuromuscular stimulation therapy after incomplete spinal cord injury promotes recovery of interlimb coordination during locomotion. J Neural Eng 2009, 6: 55010. 10.1088/1741-2560/6/5/055010Google Scholar
- Jung R, Brauer EJ, Abbas JJ: Real-time interaction between a neuromorphic electronic circuit and the spinal cord. IEEE Trans Neural Syst Rehabil Eng 2001, 9: 319-326. 10.1109/7333.948461PubMedGoogle Scholar
- Bandt C, Pompe B: Permutation entropy: a natural complexity measure for time series. Phys Rev Lett 2002, 88: 174102.PubMedGoogle Scholar
- Han TS, Kobayashi K: Mathematics of information and coding. Providence, R.I.: American Mathematical Society; 2002.Google Scholar
- Rathleff MS, Samani A, Olesen CG, Kersting UG, Madeleine P: Inverse relationship between the complexity of midfoot kinematics and muscle activation in patients with medial tibial stress syndrome. J Electromyogr Kinesiol 2011, 21: 638-644. 10.1016/j.jelekin.2011.03.001PubMedGoogle Scholar
- Olofsen E, Sleigh JW, Dahan A: Permutation entropy of the electroencephalogram: a measure of anaesthetic drug effect. Br J Anaesth 2008, 101: 810-821. 10.1093/bja/aen290PubMedGoogle Scholar
- Cao Y, Tung WW, Gao JB, Protopopescu VA, Hively LM: Detecting dynamical changes in time series using the permutation entropy. Phys Rev E Stat Nonlin Soft Matter Phys 2004, 70: 046217.PubMedGoogle Scholar
- Lee M, Roan M, Smith B: An application of principal component analysis for lower body kinematics between loaded and unloaded walking. J Biomech 2009, 42: 2226-2230. 10.1016/j.jbiomech.2009.06.052PubMedGoogle Scholar
- Deluzio KJ, Astephen JL: Biomechanical features of gait waveform data associated with knee osteoarthritis: an application of principal component analysis. Gait Posture 2007, 25: 86-93. 10.1016/j.gaitpost.2006.01.007PubMedGoogle Scholar
- Sadeghi H, Prince F, Sadeghi S, Labelle H: Principal component analysis of the power developed in the flexion/extension muscles of the hip in able-bodied gait. Med Eng Phys 2000, 22: 703-710. 10.1016/S1350-4533(01)00010-8PubMedGoogle Scholar
- Lieber RL, Yeh Y, Baskin RJ: Sarcomere length determination using laser diffraction. Effect of beam and fiber diameter. Biophys J 1984, 45: 1007-1016. 10.1016/S0006-3495(84)84246-0PubMed CentralPubMedGoogle Scholar
- Pribe C, Grossberg S, Cohen MA: Neural control of interlimb oscillations. II. Biped and quadruped gaits and bifurcations. Biol Cybern 1997, 77: 141-152. 10.1007/s004220050375PubMedGoogle Scholar
- Gregory CM, Vandenborne K, Castro MJ, Dudley GA: Human and rat skeletal muscle adaptations to spinal cord injury. Can J Appl Physiol 2003, 28: 491-500. 10.1139/h03-036PubMedGoogle Scholar
- Liu M, Bose P, Walter GA, Thompson FJ, Vandenborne K: A longitudinal study of skeletal muscle following spinal cord injury and locomotor training. Spinal Cord 2008, 46: 488-493. 10.1038/sj.sc.3102169PubMedGoogle Scholar
- Thomason DB, Biggs RB, Booth FW: Protein metabolism and beta-myosin heavy-chain mRNA in unweighted soleus muscle. Am J Physiol 1989, 257: R300-305.PubMedGoogle Scholar
- Thota A, Carlson S, Jung R: Recovery of locomotor function after treadmill training of incomplete spinal cord injured rats. Biomed Sci Instrum 2001, 37: 63-67.PubMedGoogle Scholar
- Ivanenko YP, Grasso R, Zago M, Molinari M, Scivoletto G, Castellano V, Macellari V, Lacquaniti F: Temporal components of the motor patterns expressed by the human spinal cord reflect foot kinematics. J Neurophysiol 2003, 90: 3555-3565. 10.1152/jn.00223.2003PubMedGoogle Scholar
- Jung R, Ichihara K, Venkatasubramanian G, Abbas JJ: Chronic neuromuscular electrical stimulation of paralyzed hindlimbs in a rodent model. J Neurosci Methods 2009, 183: 241-254. 10.1016/j.jneumeth.2009.06.043PubMed CentralPubMedGoogle Scholar
- Yakovenko S, Gritsenko V, Prochazka A: Contribution of stretch reflexes to locomotor control: a modeling study. Biol Cybern 2004, 90: 146-155. 10.1007/s00422-003-0449-zPubMedGoogle Scholar
- Gordon KE, Wu M, Kahn JH, Schmit BD: Feedback and feedforward locomotor adaptations to ankle-foot load in people with incomplete spinal cord injury. J Neurophysiol 2010, 104: 1325-1338. 10.1152/jn.00604.2009PubMedGoogle Scholar
- Daley MA, Felix G, Biewener AA: Running stability is enhanced by a proximo-distal gradient in joint neuromechanical control. J Exp Biol 2007, 210: 383-394. 10.1242/jeb.02668PubMed CentralPubMedGoogle Scholar
- Pearson KG: Role of sensory feedback in the control of stance duration in walking cats. Brain Res Rev 2008, 57: 222-227. 10.1016/j.brainresrev.2007.06.014PubMedGoogle Scholar
- Drew T, Jiang W, Widajewicz W: Contributions of the motor cortex to the control of the hindlimbs during locomotion in the cat. Brain Res Brain Res Rev 2002, 40: 178-191. 10.1016/S0165-0173(02)00200-XPubMedGoogle Scholar
- Drew T, Kalaska J, Krouchev N: Muscle synergies during locomotion in the cat: a model for motor cortex control. J Physiol 2008, 586: 1239-1245. 10.1113/jphysiol.2007.146605PubMed CentralPubMedGoogle Scholar
- McCrea DA, Rybak IA: Organization of mammalian locomotor rhythm and pattern generation. Brain Res Rev 2008, 57: 134-146. 10.1016/j.brainresrev.2007.08.006PubMed CentralPubMedGoogle Scholar
- Lam T, Pearson KG: The role of proprioceptive feedback in the regulation and adaptation of locomotor activity. Adv Exp Med Biol 2002, 508: 343-355. 10.1007/978-1-4615-0713-0_40PubMedGoogle Scholar
- Hiebert GW, Pearson KG: Contribution of sensory feedback to the generation of extensor activity during walking in the decerebrate Cat. J Neurophysiol 1999, 81: 758-770.PubMedGoogle Scholar
- Prochazka A, Gritsenko V, Yakovenko S: Sensory control of locomotion: reflexes versus higher-level control. Adv Exp Med Biol 2002, 508: 357-367. 10.1007/978-1-4615-0713-0_41PubMedGoogle Scholar
- Pearson KG, Misiaszek JE: Use-dependent gain change in the reflex contribution to extensor activity in walking cats. Brain Res 2000, 883: 131-134. 10.1016/S0006-8993(00)02880-8PubMedGoogle Scholar
- Donelan JM, Pearson KG: Contribution of force feedback to ankle extensor activity in decerebrate walking cats. J Neurophysiol 2004, 92: 2093-2104. 10.1152/jn.00325.2004PubMedGoogle Scholar
- Schwartz ED, Cooper ET, Chin CL, Wehrli S, Tessler A, Hackney DB: Ex vivo evaluation of ADC values within spinal cord white matter tracts. AJNR Am J Neuroradiol 2005, 26: 390-397.PubMedGoogle Scholar
- Gullapalli J, Krejza J, Schwartz ED: In vivo DTI evaluation of white matter tracts in rat spinal cord. J Magn Reson Imaging 2006, 24: 231-234.PubMedGoogle Scholar
- Chen XY, Wolpaw JR: Probable corticospinal tract control of spinal cord plasticity in the rat. J Neurophysiol 2002, 87: 645-652.PubMedGoogle Scholar
- Cheney PD, Fetz EE, Mewes K: Neural mechanisms underlying corticospinal and rubrospinal control of limb movements. Prog Brain Res 1991, 87: 213-252.PubMedGoogle Scholar
- Gordon KE, Wu M, Kahn JH, Dhaher YY, Schmit BD: Ankle load modulates hip kinetics and EMG during human locomotion. J Neurophysiol 2009, 101: 2062-2076. 10.1152/jn.90949.2008PubMed CentralPubMedGoogle Scholar
- Daley MA, Usherwood JR, Felix G, Biewener AA: Running over rough terrain: guinea fowl maintain dynamic stability despite a large unexpected change in substrate height. J Exp Biol 2006, 209: 171-187. 10.1242/jeb.01986PubMedGoogle Scholar
- Lamy JC, Iglesias C, Lackmy A, Nielsen JB, Katz R, Marchand-Pauvert V: Modulation of recurrent inhibition from knee extensors to ankle motoneurones during human walking. J Physiol 2008, 586: 5931-5946. 10.1113/jphysiol.2008.160630PubMed CentralPubMedGoogle Scholar
- Finley JM, Perreault EJ, Dhaher YY: Stretch reflex coupling between the hip and knee: implications for impaired gait following stroke. Exp Brain Res 2008, 188: 529-540. 10.1007/s00221-008-1383-zPubMed CentralPubMedGoogle Scholar
- Daley MA, Biewener AA: Running over rough terrain reveals limb control for intrinsic stability. Proc Natl Acad Sci USA 2006, 103: 15681-15686. 10.1073/pnas.0601473103PubMed CentralPubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.