Patients with SCI show varied symptoms and levels of severity. With regard to severity, there is (1) complete SCI, i.e., complete loss of sensory and motor functions below the lesion, and there is also (2) incomplete SCI, i.e., preservation of sensory and motor functions. In the case of complete SCI, especially in patients with severe dislocation of the spinal column, the connection between the brain and lower circuits seems to be completely lost. In such cases, it is important to restore some connection beyond the lesion before considering reprogramming the reorganized neural circuits. For this purpose, the molecular approach has an advantage over the electrophysiological approach.

As mentioned above, molecular biology has revealed much about the mechanisms governing axonal elongation. Especially, the intracellular mechanisms that are triggered by nerve growth factor stimuli have been investigated in detail. Nerve growth factors (NGFs), brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3) bind to their own receptors expressed on the cell surface of neurons and transduce their effects via phosphorylation cascades of signaling molecules within the cell, which finally induce gene expression for axonal elongation. Among several signaling cascades, activation of the Mek-Erk signals is said to be important, and it has been shown that Mek activation is sufficient to induce axonal growth in PC12 cell line—a model cell line for neurite growth—even in the absence of nerve growth factors. Miura et al. examined the application of this paradigm in a spinal cord injury model [2]. In their report, he transected rat spinal cord at the thoracic 10 level and injected an adenovirus gene transfer vector into the parenchyma of the proximal stump. This type of gene transfer delivered the gene to not only the segmental neural cells in the spinal cord but also the primary motor neurons in the brain, such as the red nucleus, by retrograde transport along the axons. After 6 weeks of spinal transection and simultaneous gene transfer of the control gene (LacZ) or the constitutively active Mek gene, the effect of gene transfer was examined by both behavioral and histological evaluation. Behavioral evaluation using the Basso, Beattie, and Bresnahan (BBB) scale [3], which has been a well-accepted hind limb motor scoring scale in a rat spinal cord injury model, showed better functional recovery in the active Mek-transferred group than in the control group (Fig. 8.2). Histological evaluation performed by injecting an anterograde neuronal tracer into the red nucleus showed marked regeneration of the rubrospinal tract beyond the complete transection site. Taken together, activation of intracellular signals within the primary motor neuron in the brain can be one of the approaches to retain functional recovery in cases in which the axonal connection is completely lost at the lesion site.

The therapeutic strategy for patients with incomplete SCI should be different from that for those with complete SCI. Because the symptoms for this condition are varied and the degree of severity and the segment injured varies among patients with incomplete SCI, the patient’s condition needs to be thoroughly investigated, and the aspect of neural function that requires treatment should be determined. For example, the treatment approach for a patient who cannot stand even with support should differ from that for one who can walk with support but with a spastic gait. Here, we will discuss gait rehabilitation in patients with incomplete SCI, especially those who can stand with assistive tools and walk a few steps with assistance. These patients are classified as Frankel C and are considered to probably have SCI of “mild” severity. However, it is not practical for these patients to perform locomotion activity.

With regard to the modification of neural functions, physiological functions should be strengthened and abnormal functions corrected. Besides facilitation of voluntary movement of the lower limbs, automated movement of limbs is also considered an important physiological neural function in those patients. While walking, individuals do not have to pay attention to how to move their hip, knee, and ankle joints. It has been shown that a certain “gait program” exists within the central nervous systems and that we utilize this program. The existence of the gait program is shown using an experimental model of decerebrate cats [4–6] and also human subjects [7–9] on a treadmill; it revealed that the program is located in the spinal cord, and this program is now called the “central pattern generator (CPG).” We also examined the functions of CPG in patients with SCI under our experimental settings. By using the training device EasyStand Glider (Altimate Medical Inc., USA), we produced passive lower limb motion in an alternating manner and recorded electromyographic (EMG) signals in the lower limbs. Figure 8.3b shows the activities of each muscle that is completely paralyzed in the subject. Because we observed rhythmic burst from the right and left legs as observed when the limbs make stepping movements, we assumed that the CPG in this subject was activated. With the same device, we can also investigate which components of passive leg motion are critical for CPG activation. If the observed EMG activities are induced by stretch reflex in each leg, the same EMG activities are expected when only one leg is passively moved or both legs are passively moved but in a synchronized manner and not in an alternating manner. Even though the kinematic parameters of the examined leg were exactly the same under the experimental conditions, we only observed a gait-like EMG pattern in those patients who showed passive leg motion in an alternating manner (Fig. 8.3c). These results indicate that CPG activation is specific to afferent stimuli resulting from alternate leg motion [10].