Personalized gait rehabilitation with spinal cord stimulation and machine learning: Recent advances and promising applications

Abstract

Lumbosacral spinal cord stimulation shows promise in restoring walking after spinal cord injury. This review discusses recently developed machine learning approaches to provide customized stimulation patterns and parameters according to the extent of injury to achieve community ambulation. Key challenges include the need for control strategies that enhance residual limb function and adapt to variable motor impairments across individuals. Efficient identification of optimal stimulation parameters and the ability to adapt parameters over time without manual tuning is essential for long-term use upon clinical translation of spinal cord stimulation therapies for rehabilitation. Machine learning provides the necessary framework for personalized rehabilitation treatment by offering a flexible architecture that evolves and adapts automatically to suit individual patient rehabilitation needs and preferences.