A Direct Sensorimotor Connection with the Spared Neural Code of Movement to Regain Motor Function

Descripción del proyecto

Restaurar las vías neurales de las motoneuronas espinales preservadas

Las interfaces neuronales actuales que asisten a pacientes con parálisis muscular suelen conllevar riesgos quirúrgicos y dan un control deficiente de los movimientos. En los casos de lesión medular, a menudo quedan neuronas motoras espinales activas. Se puede acceder a ellas con medidas mínimamente invasivas y aprovecharlas para restaurar las conexiones neuronales. Para ello, el equipo del proyecto GRASPAGAIN, financiado por el CEI, diseñará una interfaz neural bidireccional que proporcionará retroalimentación al cerebro a partir de la estimulación en tiempo real de la actividad eferente de las motoneuronas espinales. Utilizando métodos de aprendizaje profundo y sensores electromiográficos, desarrollará métodos de alta resolución para activar y mejorar la salida escatimada de la médula espinal mediante la identificación de potenciales de acción de unidades motoras para los músculos de la mano. Esta tecnología pretende potenciar el movimiento voluntario de los pacientes, lo que ofrece mayor independencia y una mejor calidad de vida.

Objetivo

Many individuals suffer partial or complete muscle paralysis with no available cures. Even though neural interfaces have the potential to restore motor function with assistive systems, their use is still very limited. Even in the case of state-of-the-art invasive neural implants, the control of the movements of the paralyzed limbs is highly unsatisfactory. These neural interfaces suffer high surgical risks, poor control of the activity of spinal motor neurons, and inaccurate mapping of the attempted movements. Spinal motor neurons are the last cells of the nervous system that convert motor commands into movement and their activity can be accessed with minimally invasive methods. In most neural lesions, such as spinal cord injury and stroke, there are functionally active spinal motor neurons projecting to paralyzed muscles that are modulated by brain input. In this project, I propose a bidirectional interface that is driven by the real-time identification of efferent spinal motor neuron activity. We will develop novel sensing, decoding, and feedback methods with precise cellular resolution. This neural interface will map, engage, and augment the spared output of the spinal cord through new deep learning methods and hundreds of fine-tuned electromyographic sensors recording action potentials of individual motor units for the muscles controlling the hand. The output of this interface will enable highly accurate temporal associations between efferent motor neuron activity and sensorimotor feedback by delivering multiple visual and somatosensory inputs. This bidirectional neural interface will entrain and monitor the spared neural pathways at the direct cellular level with the goal of transforming and augmenting the activity of the spared motor neurons into highly functional motor dimensions. Using these new technologies, we aim to answer open questions in movement neuroscience and spinal cord injury.