Au travers de cette revue de la littérature, nous avons montré que la coordination AP était présente chez l’animal et chez l’humain que ce soit lors d’une locomotion quadrupède ou bipède. Nous avons également vu que le tapis roulant partitionné transverse était un outil pertinent pour d’étudier cette coordination AP en dissociant les vitesses des pattes antérieures (ou bras) de celles des pattes postérieures (ou jambes). Finalement, nous avons vu que la coordination AP était perturbée suite à des lésions de la moelle épinière et qu’elle pouvait être utilisée comme stratégie thérapeutique pour récupérer une activité locomotrice. Malgré les avancées des connaissances sur la coordination AP au cours des dernières années, de nombreuses questions neuro-anatomo-fonctionnelles persistent. D’un point de vue locomoteur, quel est le rôle d’un couplage asymétrique entre les CPGs cervicaux et lombaires? Comment la coordination AP peut impacter les autres paramètres cinématiques (durée du cycle et des phases, coordination gauche-droite) et EMGs (durée et amplitude des bouffées) souvent analysés lors d’étude de la locomotion? L’utilisation du tapis-roulant partitionné chez un animal intact est-elle possible? Dans ce cas, la coordination AP est-elle similaire à celle observée chez le chat décérébré? Quels sont les déficits de la coordination AP suite à une hémisection de la moelle épinière et comment ceux-ci peuvent impacter les autres aspects de la locomotion? Quelle coordination AP peut être induite chez un chat hémisectionné par l’utilisation d’un tapis roulant partitionné transverse et dans quelle mesure ce tapis peut être employé pour simuler une lésion ou entraîner une récupération suite à une lésion?
Les chapitres de mon travail de doctorat présentent les résultats obtenus lors d’expériences de locomotion partitionnée transverse chez le chat adulte. Les deux premières études ont été réalisées chez le chat intact et permettent une meilleure caractérisation de la coordination AP employé chez un animal quadrupède sur un tapis roulant partitionné transverse. La troisième étude présente une expérience de locomotion partitionnée transverse chez le chat ayant une hémisection de la moelle épinière en T10 et permet de mettre en avant une schématisation théorique du contrôle fonctionnel de la coordination AP.
4.1. Objectif 1
Caractériser la coordination AP lors de la locomotion partitionnée transverse chez le chat adulte intact. La coordination AP lors de la locomotion partitionnée transverse a été évaluée chez le chaton à l’état intact à faible différence de vitesses (Cruse and Warnecke, 1992) et chez le chat adulte décérébéré (Akay et al., 2006). Dans le but de proposer une meilleure compréhension des mécanismes du contrôle neurophysiologique de la coordination AP en se basant sur ce type d’expérimentation, il est nécessaire de connaitre les résultats lors d’une situation contrôle où toutes les voies pouvant influer sur la coordination AP sont toujours présentes. Cependant, une caractérisation de la coordination AP lors de locomotion partitionnée transverse chez le chat adulte intact n’a encore jamais été réalisée. Nos hypothèses sont qu’une adaptation asymétrique va avoir lieu avec une dissociation du rythme lorsque la vitesse des pattes antérieures est plus élevée. Qui plus est, malgré une dissociation du rythme, une certaine forme de coordination antéropostérieure va être préservée.
4.2. Objectif 2
Caractériser la coordination AP lors de la locomotion partitionnée transverse chez le chat adulte après une hémisection de la moelle épinière en T10 et comparer les résultats à ceux obtenus à l’état intact. Plusieurs études ont observé des déficits de la coordination AP suite à des lésions partielles de la moelle épinière se caractérisant principalement par une dissociation du rythme entre les pattes postérieures et antérieures (Barrière et al., 2010, 2008; Bem et al., 1995; Gorska et al., 2013, 1996; Kato et al., 1984). Cependant, ces déficits ont souvent été caractérisés lors de tâche de locomotion basique et leurs impacts dans des situations demandant un contrôle plus important de la coordination AP n’a pas encore été étudiés. Une caractérisation de ces déficits face à une tâche de locomotion partitionnée transverse pourrait permettre de mieux comprendre les mécanismes de contrôle neurophysiologique de la coordination AP à l’état intact et pathologique. Nos hypothèses sont que les lésions de la moelle épinière vont perturber la coordination AP sans pour autant entrainer une perte de celle-ci. De plus, la locomotion partitionnée transverse va pouvoir moduler la coordination AP.
2A
RTICLE1
Coordination between the fore- and hindlimbs is bidirectional, asymmetrically organized, and flexible during quadrupedal locomotion in the intact adult cat
Auteurs de l’article: Yann Thibaudier, Marie-France Hurteau, Alessandro Telonio, and Alain Frigon
Statut de l’article: publié dans Neuroscience 2013 Jun 14; 240:13-26. doi: 10.1016/j.neuroscience.2013.02.028.
Avant-propos: Pr. Alain Frigon et moi-même avons mis au point le protocole consistant à évaluer la locomotion partitionnée transverse chez le chat adulte intact. Avec l’aide de mes collègues, j’ai réalisé l’ensemble des expériences. Par la suite j’ai réalisé l’analyse, les figures ainsi que rédigé une première version de l’article. Puis nous avons amélioré cette première version à l’aide des nombreux commentaires et suggestions du Pr. Frigon et des coauteurs de l’article.
Résumé : Malgré l’importance évidente de la coordination entre les membres pour la locomotion quadrupède chez le mammifère terrestre, son organisation demeure peu comprise. Dans cet article, nous avons évalué la durée du cycle et des phases, ainsi que le patron locomoteur, chez quatre chats adultes intacts entraînés à marcher sur un tapis roulant partitionnée transverse permettant de contrôler indépendamment la vitesse des pattes antérieures et postérieures. Lorsque les pattes postérieures marchaient à une vitesse plus importante que les pattes antérieures, un rythme égal était tout le temps maintenu entre les pattes antérieures et postérieures, même à des différences de vitesses de ratio 1 :3 (0.4 :1.2m/s). Le patron locomoteur s’ajustait grâce à des changements de la durée de balancement et d’appui pour les pattes postérieures, alors que pour les pattes antérieures, uniquement la durée de la phase d’appui était affectée. Dans de telles conditions, quand les valeurs pour les pattes postérieures et antérieures étaient comparées à celles obtenus pour des vitesses équivalentes en locomotion non-partitionnée (valeurs prédites basées sur la vitesse de locomotion), la durée du cycle et des phases d’appui et de balancement des pattes postérieures étaient tout le temps plus longues que les valeurs prédites. À l’inverse, La durée du cycle et de la phase d’appui des pattes antérieures étaient plus courtes mais uniquement pour de grands écarts de vitesses alors que la durée de la phase de balancement était similaire aux valeurs prédites. La séquence des contacts quand la vitesse des pattes postérieures était plus importante était identique à celle obtenue en locomotion non- partitionnée. À l’inverse, quand les pattes antérieures marchaient à une vitesse un petit peu plus rapide que les pattes postérieures, une dissociation du rythme entre les pattes antérieures et postérieures était observée. Lors de telles dissociations, la durée du cycle et
des phases des pattes postérieures étaient similaires aux valeurs prédites par la locomotion non-partitionnée alors que les valeurs pour les pattes antérieures étaient réduites. De plus, de nombreuses séquences de contacts pouvaient être observées. Ainsi, les résultats de cette étude démontrent clairement l’existence d’un contrôle bidirectionnel, asymétrique et flexible de la coordination antéropostérieure lors de la locomotion quadrupède chez le chat adulte intact.
Abstract
Despite the obvious importance of inter-girdle coordination for quadrupedal locomotion in terrestrial mammals, its organization remains poorly understood. Here, we evaluated cycle and phase durations, as well as footfall patterns of 4 intact adult cats trained to walk on a transverse split-belt treadmill that could independently control fore- and hindlimb speed. When the hindlimbs walked at faster speeds than the forelimbs, an equal rhythm was always maintained between the fore- and hindlimbs, even at the highest fore- hindlimb speed ratio of 1:3 (0.4 : 1.2 m/s). The locomotor pattern adjusted through changes in both hindlimb stance and swing phase durations, whereas only the forelimb stance phase was affected. In such conditions, when fore- and hindlimb values were compared to those obtained at matched speeds during tied-belt walking (i.e. predicted values based on treadmill speed), hindlimb cycle, stance and swing durations were consistently longer than predicted. On the other hand, forelimb cycle and stance durations were shorter than predicted but only at the highest split-belt speed ratios. Forelimb swing durations were as predicted based on front belt speed. The sequence of footfall pattern when hindlimb speed was faster was identical to tied-belt walking. In stark contrast, when the forelimbs walked at slightly faster speeds than the hindlimbs, the rhythm between the fore- and hindlimbs broke down. In such conditions, the locomotor pattern was adjusted trough changes in stance and swing phase durations in both the fore- and hindlimbs. When the rhythm between the fore- and hindlimbs broke down, hindlimb cycle and phase durations were similar to predicted values, whereas forelimb values were shorter than predicted. Moreover, several additional sequences of footfall patterns were observed. Therefore, the results clearly demonstrate the existence of a bidirectional, asymmetric, and flexible control of inter-girdle coordination during quadrupedal locomotion in the intact adult cat.
Key words: Quadrupedal locomotion; inter-girdle coordination; transverse split-belt treadmill
Introduction
Proper coordination of the fore- and hindlimbs (i.e. inter-girdle coordination) in terrestrial mammals is essential to maintain stability during the forward progression of quadrupedal locomotion. Human infants also use quadrupedal forms of locomotion (Yang et al., 2004) and despite its bipedal nature, human adults are thought to have conserved a quadrupedal form of coordination during walking (Dietz, 2011; Dietz et al., 1994; Zehr and Duysens, 2004; Zehr et al., 2009). However, the general framework of inter-girdle coordination during quadrupedal locomotion in terrestrial mammals remains poorly understood and the results often seem to conflict [recently discussed in (Thibaudier and Hurteau, 2012)]. For instance, some have suggested that forelimb activity exerts greater influence on the hindlimbs (Akay et al., 2006; Miller et al., 1977), while other studies have proposed the opposite (Juvin et al., 2012; Juvin et al., 2005).
Although there is disagreement regarding the influence of ascending and descending inhibitory or excitatory pathways on inter-girdle coordination, most studies agree that spinal locomotor networks controlling the fore- and hindlimbs are asymmetrically coupled. Flexible patterns of inter-girdle coordination are undoubtedly required to orchestrate different rhythmic tasks, such as walking, running, and swimming. This, in turn, would require functional changes in the organization of pathways coupling the cervical and lumbosacral locomotor networks. As such, the task being performed must be considered when evaluating the organization of inter-girdle coordination. Specific patterns of inter- girdle coordination might also appear only under certain experimental conditions, such as following decerebration (Akay et al., 2006) or in isolated spinal cord preparations (Juvin et al., 2012; Juvin et al., 2005). To evaluate the organization of fore- and hindlimb coordination during quadrupedal locomotion it is important to study it during that very same task. Surprisingly, few data are available with such a model.
Here, we evaluated some features of the locomotor pattern in intact adult cats trained to walk on a transverse split-belt treadmill that independently controlled the speed of the fore- and hindlimbs. Such an approach was used before to study fore- and hindlimb coordination in intact juvenile cats (Cruse and Warnecke, 1992) and in decerebrate adult
cats (Akay et al., 2006). Cruse and Warnecke (1992) used only two slow treadmill speeds (0.3 m/s and 0.46 m/s) and concluded that ascending and descending influences were largely symmetric. Akay et al. (2006) used a wider range of speeds (i.e. 0.3-0.6 m/s) and concluded that forelimb activity had greater influence on hindlimb activity in decerebrate cats. However, they only detailed the effect of increasing the speed of the front treadmill while keeping hindlimb speed constant.
In the present study, we used a wide range of speeds (0.4-1.2 m/s) for the fore- and hindlimbs in freely behaving intact adult cats to probe the coordination between locomotor networks located at cervical and lumbar girdles. By changing the speed of the fore- or hindlimbs, while maintaining the speed of the limbs located at the other girdle constant, we provide the first evidence that inter-girdle coordination in quadrupedal intact adult mammals is bi-directional and asymmetrically organized.
We also evaluated patterns of footfall during transverse split-belt walking. In general, the vast majority of tetrapods, including cats, employ what is term a lateral sequence (LS) during terrestrial locomotion, whereby contact of a hindlimb is directly followed by contact of the ipsilateral forelimb [e.g. (Hildebrand, 1967; Lemelin et al., 2003; Muybridge, 1957; Stevens, 2006; Wetzel et al., 1976)]. In contrast, a diagonal sequence (DS) during terrestrial locomotion is characterized by contact of a hindlimb followed directly by contact of the contralateral forelimb, a pattern mostly found in non- human primates [e.g. (Hildebrand, 1967; Larson et al., 2000; Lemelin et al., 2003; Muybridge, 1957; Prost, 1965; Schmitt, 2003; Stevens, 2006; Stevens, 2008; Vilensky and Larson, 1989; Young et al., 2007)]. In the present study, we show that cats can adopt DS gait patterns when forelimb speed exceeds hindlimb speed.
Therefore, we present transverse split-treadmill in freely behaving cats as a novel technique to probe the functional organization and flexibility of inter-girdle coordination in an intact adult mammalian system.
Experimental Procedures
Animals and ethical considerations
All procedures were approved by the Animal Care Committee of the Université de Sherbrooke and were in accordance with policies and directives of the Canadian Council on Animal Care. Before the experiments, animals were housed and fed within designated areas. Four adult cats weighing between 4.0 and 9.0 kg were selected based on their ability to walk for prolonged period (10-15 minutes) on an animal treadmill with a single running surface (Fit-Fur-Life Ltd, UK). Cats were then trained to walk on a transverse split-belt treadmill composed of four independently controlled running surfaces 120 cm long and 30 cm wide (Bertec Corporation, Columbus, Ohio). In the present study, only the front and back surfaces on the left side of the treadmill were used. A Plexiglas box (120 cm long, 50 cm high) constrained the animals to walk with the fore and hindlimbs separately on the two belts. The box was open at the top and at the front. Food and affection were given as reward and the cats did not contact the Plexiglas box following training. Recording sessions started once the animals could walk comfortably with the fore- and hindlimbs on the front and rear belts, respectively, which required approximately 4 additional weeks of training on the transverse split-belt treadmill.
Experimental protocol
The experimental set-up and conditions are shown in Figure 1A. Each cat performed 5 sessions of several locomotor episodes of 10-15 cycles in 5 conditions: 1) fore- and hindlimbs walking at equal speeds (TIED) from 0.4 to 0.8 m/s; 2-3) forelimbs walking at a constant speed of 0.4 m/s (FORE.4) or 0.8 m/s (FORE.8) with hindlimb speed increasing from 0.4 to 0.8 m/s, and 4-5) hindlimbs walking at a constant speed of 0.4 m/s (HIND.4) or 0.8 m/s (HIND.8) with forelimb speed increasing from 0.4 to 0.8 m/s. In a few sessions, faster speeds (up to 1.2 m/s) were used to evaluate the effects of increasing the ratio of speed between the front and rear belts. Only episodes where the animal had its forelimbs and hindlimbs on their respective belts were retained for analysis.
Figure 1 Experimental set-up and conditions.
A) Transverse split-belt treadmill was used to evaluate inter-girdle coordination using 5 conditions: 1) fore- and hindlimbs walking at equal speeds (TIED) from 0.4 to 0.8 m/s; 2-3) forelimbs walking at a constant speed of 0.4 m/s (FORE.4) or 0.8 m/s (FORE.8) with hindlimb speed increasing from 0.4 to 0.8 m/s, and 4-5) hindlimbs walking at a constant speed of 0.4 m/s (HIND.4) or 0.8 m/s (HIND.8) with forelimb speed increasing from 0.4 to 0.8 m/s. Black boxes represent conditions when the hindlimbs walked faster than the forelimbs, whereas gray boxes represent conditions when the forelimbs walked faster than
the hindlimbs. Figure 1B shows an example of a footfall pattern that followed a sequence of right hindlimb contact (RHC) → right forelimb contact (RFC) → left hindlimb contact (LHC) → left forelimb contact (LFC). In the footfall patterns shown in Figs. 4, 8, and 12, RHC, RFC, LHC, and LFC are located in the same positions. F = forelimbs; H = hindlimbs.
Data acquisition and analysis
Videos were captured with two cameras (Basler AG) located on the left and right sides of the treadmill at 60 frames per second and analyzed off-line. A custom-made Labview program was used to acquire the images and synchronize the cameras. In the present study, we focused on cycle and phase durations to evaluate changes in the overall locomotor pattern during transverse split-belt walking. Cycle duration was measured from successive foot contacts. Stance duration corresponded to the interval of time from foot contact to foot lift-off while swing duration was measured from foot lift-off to foot contact. Durations for stance, swing, and cycle were averaged for each locomotor episode. For each cat, 5 sessions recorded on separate days were averaged.
In every condition, the sequential order of limb contacts (i.e. footfall pattern) was analyzed (see Fig. 1B). Each footfall pattern began and finished with right hindlimb contact. For simplicity, footfall patterns that occurred in less than 10% of trials for a given condition (i.e. TIED, F < H, F > H) were excluded. The phase relative to right hindlimb contact was also calculated. For this, the cycle was normalized from 0 to 1, or successive right hindlimb contacts. Thus, an event (e.g. right forelimb contact) that occurred at 0.2 indicates a contact at 20% of the cycle relative to right hindlimb contact.
Statistical analysis
For tied and split-belt conditions, a two-factor (girdle, speed) repeated measures analysis of variance (ANOVA) was used to compare cycle, stance and swing durations at different speeds in the fore- and hindlimbs. Pairwise comparisons were then performed when significant differences were found. In addition, paired t-tests were used to compare cycle, stance, and swing durations during tied- and split-belt conditions at similar matched
speed. Significance level was set at p < 0.05. All values in the histograms are the mean values obtained from the 4 cats ± standard deviation.
Results
The aim of this study was to investigate some features of inter-girdle coordination in the intact adult cat when the fore- and hindlimbs walked at different speeds using a transverse split-belt treadmill. The results show a clear bidirectional and asymmetric adjustment of the fore- and hindlimb locomotor patterns during transverse split-belt walking. We also show that cats can use a variety of footfall patterns when forelimb speed is greater than hindlimb speed.
Figure 2 Cycle and phase durations during tied-belt walking in a single cat. Stance and swing phases are shown during 5 s of tied-belt walking at speeds ranging from 0.4 m/s to 0.8 m/s. Each episode started at forelimb paw contact. FL = forelimbs; HL = hindlimbs.
As a basis for comparison with transverse split-belt walking, cycle, stance, and swing durations during tied-belt walking (i.e. equal speed of the fore- and hindlimbs) are first presented. In all experimental conditions, including tied-belt walking, the fore- and hindlimbs walked on separate belts. Figure 2 shows an example of tied-belt walking from 0.4 m/s to 0.8 m/s in 0.1 m/s increments in a single cat. As can be seen, cycle and stance durations in the fore- and hindlimbs decreased with increasing speed, whereas swing duration was relatively unchanged.
Figure 3 Cycle and phases durations during tied-belt conditions across cats. A) Cycle, B) stance, and C) swing durations are plotted from 0.4 m/s to 0.8 m/s in tied-belt conditions. Five sessions were averaged per cat. Each bar is the mean ± standard deviation of the 4 cats. A two-factor (limbs, speed) repeated measures analysis of variance (ANOVA) was used to compare cycle, stance and swing durations at different speeds in the fore- and hindlimbs. Pairwise comparisons were then performed when significant differences were found. Asterisks represent significant differences across speeds (pairwise comparisons, p ≤ 0.05). F = forelimbs; H = hindlimbs.
As expected, for the group, forelimb cycle durations were not significantly different from hindlimb cycle durations during tied-belt walking (Fig. 3), indicating that an equal rhythm, or cadence, was conserved between girdles (i.e. a 1:1 ratio). An increase in speed produced