Neural Network Connection Characteristics Between Motor Cortex and Spinal Cord During Rat Bipedal Walking.

OBJECT: Bipedal locomotion requires coordinated corticospinal interactions, yet how these networks dynamically reorganize across gait phases remains poorly understood. This study aimed to characterize the phase-dependent restructuring of corticospinal functional networks during rest, stance, and swing in healthy rats. METHODS: Single-unit spiking activity was simultaneously recorded from the hindlimb motor cortex and lumbar spinal cord of healthy rats across three behavioral states: Rest, stance, and swing. Granger causality analysis was applied to local field potentials (LFPs) to assess directed functional connectivity. Total Spiking Probability Edges (TSPE) algorithm was used to estimate spike-train functional connectivity. Network topology was characterized by global efficiency and rich-club coefficient. RESULTS: Granger causality revealed that corticospinal connectivity was most prominent in the beta band (13-35 Hz) during active locomotion. Global efficiency rose from rest (0.33) to stance (0.53) and swing (0.51). TSPE confirmed rich-club topology in all conditions. Spinal hub neurons predominated during locomotion, comprising 81.13% ± 15.5% of rich-club hubs at stance and 60.48% ± 10.21% at swing, while cortical and spinal hubs were balanced at rest. Mean latency decreased from 29.60 ± 20.66 ms at rest to 9.67 ± 1.65 ms during stance. CONCLUSIONS: The central nervous system dynamically restructures corticospinal functional networks across gait phases, with spinal hub neurons assuming a dominant role during active locomotion. This framework provides a mechanistic reference for studying corticospinal disruptions in spinal cord injury, Parkinson's disease, and stroke, and identifies phase-specific beta-band coupling and rich-club hub dynamics as candidate targets for closed-loop neural interfaces in gait rehabilitation.