|
Abstract:
|
Neural decoding has played a key role in recent advances in motor cortical Brain-Machine Interfaces. A number of practical algorithms have been proposed to predict continuous movement in an on-line circumstance using recordings from a chronically implanted multi-electrode micro-array in the motor cortex. Though effective, those models were based on a strong assumption that the neural signal sequences are a stationary process. Recent work, however, indicates that the motor system significantly varies over time. To characterize the dynamic relationship between neural signals and hand kinematics, here we develop an adaptive approach based on the well-known Kalman filter. This new approach recursively updates encoding model in an on-line fashion. Experimental results show that it decodes more accurately than a non-adaptive Kalman filter, while remains real-time efficiency.
|
