Interneuron-linked oscillators within the lumbar cord for flexion and extension of the hindlimbs, called central pattern generators (CPGs), have become a target for both locomotor training interventions and for neural repair (Chapter 13 of Volume I). This circuitry may be driven by the timing of cutaneous and weight-bearing inputs from the soles and by proprioceptive and load information from joints, especially the hips, during the practice of stepping. Following a low thoracic spinal cord transection, cats and rats have been trained to step with their hindlimbs on a moving treadmill belt with support for the sagging trunk. Pulling down on their tails or a noxious input enhances hindlimb loading in extension. The animals are not as successful walking over ground. Training-induced adaptations within the cord in these animal models point to the potential of plasticity induced by rehabilitation to lead to behavioral gains. It seems likely that a network of locomotor spinal motoneu-rons and interneurons has been conserved in humans (Dimitrijevic et al., 1998), along with other forms of spinal organization (Lemay and Grill, 2004) that increase the flexibility of supraspinal regions to control hindlimb and lower extremity movements.
Studies in patients with clinically complete SCI reveal similar responses to limb loading and hip inputs as were found in spinal transected cats (Chapter 30 Volume II). However, plasticity within the cord may be no more important than plasticity induced within cortical and subcortical nodes of the sensorimotor network for walking in patients who have residual descending supraspinal input to the lumbosacral motor pools. Review articles may overstate the evidence for the case that spinal networks, in concert with specific sensory information, are responsible for locomotion in man. Rationales for gait retraining in patients can draw from animal models about specific sensory inputs that aid the timing and efficacy of the step cycle. These sensory inputs, however, act at many levels of the neuroaxis, beyond the CPGs.
Figure 3.1 shows the extent of cortical reorganization induced by a conus/cauda equina SCI followed by gait training. This 52-year-old subject suffered a burst fracture of the first lumbar vertebral body 8 years prior to the functional magnetic resonance imaging (fMRI) study. He had no movement or sensation below L3 on the left and L4 on the right. He dorsiflexed his right ankle a maximum of 5° and could not offer resistance. He was able to extend each knee against gravity and minimal resistance. The subject walked modest distances with a reciprocal gait at 1.5 mph wearing bilateral solid ankle-foot orthoses with a cane in each hand for balance. The fMRI study of voluntary right ankle dorsiflexion (Fig. 3.1(a)) reveals significantly greater recruitment compared to control subjects, involving the contralateral thoracolumbar representation in primary sensori-motor cortex (M1S1) and in bilateral M1S1 that represents the whole leg, the supplementary motor area, basal ganglia, and cerebellar hemispheres. Contralateral cingulate motor cortex and premotor area activation and bilateral parietal, insula, and dorsolateral prefrontal cortex (dlpf) activity also significantly
& SMA g
Was this article helpful?