Nearly 6 million people in the US are affected by some form of paralysis. For such patients, whose treatment options are limited, neural prosthetic systems using a brain machine interface (BMI) hold the potential to drastically improve quality of life. Recent research has demonstrated the promising clinical potential of neural prosthetics, but the performance and robustness of these systems could be improved significantly by incorporating deeper knowledge of motor cortex physiology. Existing BMI systems operate by decoding a patient's intention to move from a statistical model of the activity patterns of neurons in their motor cortex, which are recorded using an implanted electrode array. The overarching goal of this project is to better understand the neural substrate of movement generation, with an eye towards improving the performance and robustness of neural prosthetics and bringing this technology closer to clinical viability. Previous work in our lab has demonstrated that the pattern of neural activity in motor cortex prior to making a movement influences the kinematics of the subsequent movement. The work proposed here aims to extend these findings by determining how the the primate motor system limits variability of this neural preparatory activity for a given reaching movement to maximize the precision of the upcoming reach. The second major goal of this work is to gain insights into how the motor cortex uses knowledge of limb properties (e.g.: mass and limb dynamics) to A) appropriately respond to sensory feedback while actively preparing movements, and B) respond to errors during preparation and movement that result from the motor system's intrinsic variability (noise) or from external mechanical disturbances to the limb. Novel decoding algorithms, which incorporate knowledge of how the brain prepares upcoming movements and tunes its response to sensory feedback in a manner specific to the limb, may help patients with paralysis make faster and more natural movements using BMI devices. Incorporating this knowledge into BMI systems is likely to help improve their speed and robustness to neural and motor noise, especially as the complexity of these systems increases with the introduction of robotic arms and legs.
Nearly 6 million people in the US alone suffer from a neurological injury or disease that results in profound movement impairment . Brain machine interfaces (BMI) and neural prosthetic systems hold the potential to drastically improve quality of life for such patients, whose other treatment options remain limited. This work will explore the neural mechanisms underlying movement preparation and sensory feedback control in the motor cortex, with the goal of improving the speed and robustness of neural prosthetic systems in clinical applications.