Spinal sensorimotor circuits play a prominent role in hindlimb locomotor recovery after staggered thoracic lateral hemisections but cannot restore posture and interlimb coordination during quadrupedal locomotion in adult cats

Abstract

Spinal sensorimotor circuits interact with supraspinal and peripheral inputs to generate quadrupedal locomotion. Ascending and descending spinal pathways ensure coordination between the fore-and hindlimbs. Spinal cord injury disrupts these pathways. To investigate the control of interlimb coordination and hindlimb locomotor recovery, we performed two lateral thoracic hemisections placed on opposite sides of the cord (right T5-T6 and left T10-T11) at an interval of approximately two months in eight adult cats. In three cats, we then made a complete spinal transection caudal to the second hemisection at T12-T13. We collected electromyography and kinematic data during quadrupedal and hindlimb-only locomotion before and after spinal lesions. We show that 1) cats spontaneously recover quadrupedal locomotion following staggered hemisections but require balance assistance after the second one, 2) coordination between the fore-and hindlimbs displays 2:1 patterns and becomes weaker and more variable after both hemisections, 3) left-right asymmetries in hindlimb stance and swing durations appear after the first hemisection and reverse after the second, and 4) support periods reorganize after staggered hemisections to favor support involving both forelimbs and diagonal limbs. Cats expressed hindlimb locomotion the day following spinal transection, indicating that lumbar sensorimotor circuits play a prominent role in hindlimb locomotor recovery after staggered hemisections. These results reflect a series of changes in spinal sensorimotor circuits that allow cats to maintain and recover some level of quadrupedal locomotor functionality with diminished motor commands from the brain and cervical cord, although the control of posture and interlimb coordination remains impaired.

Significance statement: Coordinating the limbs during locomotion depends on pathways in the spinal cord. We used a spinal cord injury model that disrupts communication between the brain and spinal cord by sectioning half of the spinal cord on one side and then about two months later, half the spinal cord on the other side at different levels of the thoracic cord in cats. We show that despite a strong contribution from neural circuits located below the second spinal cord injury in the recovery of hindlimb locomotion, the coordination between the forelimbs and hindlimbs weakens and postural control is impaired. We can use our model to test approaches to restore the control of interlimb coordination and posture during locomotion after spinal cord injury.

Figure 1.. Staggered hemisections paradigm and extent of lesions.

Schematic representation of the staggered hemisections and extent of the first and second spinal lesions on the right (T5-T6) and left (T10-T11) sides, respectively, for individual cats. The black area represents the lesioned region.

Figure 2.. Quadrupedal treadmill locomotion before and after staggered hemisections.

Activity from selected fore- (FL) and hindlimb (HL) muscles and stance phases (thick horizontal lines of the left (L) and right (R) limbs in Cat AR at 0.4 m/s. Grey stance phases indicate cycles with 2:1 fore-hind coordination. BB, Biceps brachii; TRI, Triceps brachii; ECU, Extensor carpi ulnaris; SRT, Sartorius; SOL; Soleus.

Figure 3.. Coordination between right homolateral limbs before and after staggered hemisection.

Distribution of right homolateral couplings for the group during 1:1 and 2:1 (first and second forelimb steps) fore-hind coordination. Each bar represents the number of right homolateral couplings found for all eight cats at phase intervals of ten degrees.

Figure 4.. Interlimb phasing and variations during quadrupedal treadmill locomotion before and after staggered hemisections for the group.

A) Phase intervals for forelimb and hindlimb couplings. B) Coefficients of variation for six limb pairs. We averaged 8–36 cycles per cat at each time point. The bars represent mean ± SD for the group (n = 8 cats) while grey circles represent individual data points (mean for each cat). The P values show the main effect of state (one-factor Friedman test). Asterisks indicate significant differences between time points from the Wilcoxon signed-rank test with Bonferroni’s correction.

Figure 5.. Temporal adjustments during quadrupedal treadmill locomotion before and after staggered hemisections for the group.

A and B) Cycle, stance and swing durations for the fore- and hindlimbs, respectively. C) Asymmetry indexes of temporal variables. We averaged 8–36 cycles per cat. The bars represent mean ± SD for the group (n = 8 cats) while grey circles represent individual data points (mean for each cat). The P values show the main effect of state (one-factor Friedman test). Asterisks indicate significant differences between time points from the Wilcoxon signed-rank test with Bonferroni’s correction.

Figure 6.. Support periods during quadrupedal treadmill locomotion before and after staggered hemisection for the group.

Individual periods of support normalized to right hindlimb cycle duration. We averaged 8–36 cycles per cat. The bars represent mean ± SD for the group (n = 8 cats) while grey circles represent individual data points (mean for each cat). The P values show the main effect of state (one-factor Friedman test). Asterisks indicate significant differences between time points from the Wilcoxon signed-rank test with Bonferroni’s correction.

Figure 7.. Spatial adjustments during quadrupedal treadmill locomotion before and after staggered hemisections for the group.

A and B) Stride length and distances at contact and liftoff for the fore- and hindlimbs, respectively. C) Asymmetry indexes of spatial variables. We averaged 8–36 cycles per cat. The bars represent mean ± SD for the group (n = 8 cats) while grey circles represent individual data points (mean for each cat). The P values show the main effect of state (one-factor Friedman test). Asterisks indicate significant differences between time points from the Wilcoxon signed-rank test with Bonferroni’s correction.

Figure 8.. Homolateral limb interference during quadrupedal treadmill locomotion before and after staggered hemisections for the group.

Each panel shown horizontal distances between homolateral hindlimbs (HL) and forelimbs (FL) at contact and liftoff of the left and right forelimb. We averaged 8–36 (17.94±7.08) cycles per cat. The bars represent mean ± SD for the group (n = 8 cats) while grey circles represent individual data points (mean for each cat). The P values show the main effect of state (one-factor Friedman test). Asterisks indicate significant differences between time points from the Wilcoxon signed-rank test with Bonferroni’s correction.

Figure 9.. Hindlimb-only and quadrupedal treadmill locomotion before and after complete spinal transection.

A) Activity from selected hindlimb muscles and stance phases (thick horizontal lines of the left (LHL) and right (RHL) hindlimbs in Cat JA at 0.4 m/s. B) Activity from selected hindlimb muscles and stance phases (thick horizontal lines of the left (L) and right (R) limbs in Cat HO at 0.4 m/s. Grey and blue stance phases indicate cycles with 2:1 and 1:2 fore-hind coordination, respectively. BB, Biceps brachii; SOL; Soleus; SRT, Sartorius; Triceps brachii.

Figure 10.. Potential changes in spinal sensorimotor circuits after staggered hemisections.

In the intact state, descending supraspinal and propriospinal pathways reach lumbar spinal interneurons that control spinal motoneurons. Pathways transmitting signals from proprioceptive and cutaneous afferents ascend to the brain and project locally to spinal interneurons. After the first hemisection performed on the right side, ipsilesional lumbar neurons have weaker activity and increased weight support of the contralesional hindlimb increases load feedback from extensors and cutaneous afferents. Thicker lines represent increase influence. The left spinal network increases its influence on the right spinal network. New descending and ascending pathways also form to facilitate communication between the brain and spinal cord. After the second hemisection performed on the left side, neurons of the right spinal network have recovered their activity. Direct ascending and descending pathways are disrupted but new pathways can form through short propriospinal relays. After spinal transection, both the left and right spinal network function without descending inputs and hindlimb locomotion is expressed, possibly via strengthened sensorimotor interactions bilaterally.