A quadrupedal neurorobotic model to study the role of long propriospinal neurons before and after spinal cord injury

Long propriospinal neurons coordinate activities between limb-specific neural circuitry in the cervical and lumbar enlargement. Silencing of either ascending or descending long propriospinal neurons in intact rats results in a disruption of left–right coordination. Yet, their silencing after a moderate thoracic contusive injury significantly improves locomotion. The underlying mechanisms are unclear and further complicated by the context-dependency of the phenotypes, and likely involve afferent feedback and biomechanical constraints. Here, we present an integrated neurorobotic model of quadrupedal locomotion (Figure) to study the involvement of long propriospinal neurons in locomotor control in intact rats and following thoracic contusion. We have expanded our previous neural network model of the spinal locomotor circuitry to drive locomotion of a simulated Unitree A1go quadrupedal robot. The model includes four rhythm generators, one per limb, interconnected by commissural and long propriospinal neurons. Activity from each rhythm generator controls pattern formation networks that coordinate muscle activation in each limb. Hill-type muscle models convert this activation into torque to actuate the motors and allow for calculation of proprioceptive feedback, which interacts with all levels of the spinal circuitry. The intact model produces walk and trot in a speed-dependent manner and adapts to terrain with varying slopes and friction. Neuronal manipulations within the model can be causally linked to changes in locomotor behavior, which provides a possibility to investigate circuit reorganization and the role of long propriospinal neurons post spinal cord contusion.