Neural Mechanism Exploration Of Adaptive Motor Control And Movement Identification Based On Primary Motor Cortex Neural Interface

With the continuous development of microelectrode manufacturing and multi-channel neural signal recording techniques, the neural interface(NI) system, extracting neuronal activity from the organism and communicating with computers or other external devices, has been paid a lot of attention. Based on the analysis of neural activity of the primary motor cortex(M1) in a primate NI system, the neural mechanism of primates' adaptive motor control and movement identification have been deeply explored in this thesis.Firstly, the neural mechanism of adaptive motor control when monkeys achieved the motor tasks with different perturbations has been explored. A dynamical economic cell allocation mechanism(DECAM) has been proposed, achieving different motor tasks by altering the number of recruited cells modulating different movement factors. Combining it with the behavioral performance of monkeys in the experiment, a hypothetical adaptive motor control framework which might be applicable to similar tasks has been put forward. The framework reveals the economy of adaptive motor control of organism at both of the neurophysiological cellular level and behavioral level, which is in accordance with the general rule of evolution.Secondly, using different pattern recognition algorithms, the motion intention of the monkey has been identified based on the neuronal discharge of M1. The choosing of appropriate neurons has been confirmed very important for the performance of neural decoding. Meanwhile, the decoding discrepancy of different movement factors based on different categories of cells has also implied the validity of cells classification based on 2-way ANOVA. On the other hand, the decoding performance has been shown good enough based on appropriate cells during movement preparation phase before the movement was actually executed, which is of great significance to the neural control of intelligent prostheses.In addition, the Kalman filter has been utilized to track the monkey's hand trajectories. Least squares(LS) and maximum likelihood(ML) methods have been used to estimate the key parameters in the model of Kalman filter, based on the discharge sequences of appropriate neurons and proper movement trajectory. The results demonstrate that the Kalman filter is effective for the prediction and tracking of the monkey's hand trajectory. The establishment of Kalman filter model is actually an adaptation to the monkey's movement habit, which is of great significance to the bionic development of future prosthetic system.Finally, a summary of the research has been given. And according to the insufficient of current research, the future work to be investigated has been indicated.