Motor Cortical Representation And Decoding Of Monkey Reach And Grasp Movement For Brain Machine Interfaces

Reaching out to grasp different shaped objects is the most common and important function of the hands in our daily of life. The loss of this function, because of physical injuries or nervous system disease, will result in serious physiology and psychological barriers. Brain machine interfaces (BMIs) can restore the lost motor function through directly reading the motor intents from brain (bypassing damaged peripheral nerves) and translating them into the commands of artificial auxiliary equipment to complete alternative movement. Base on non-human primate BMIs, this work studies the motor cortical representation and decoding of reach and grasp movement, realizes a practical BMIs system that enabled the monkey control a robot hand to grasp different objects using neural signals. This result implies new rehabilitation for motor disabled people.Two monkeys were trained to complete different reach and grasp tasks. At the same time, we recorded the neural signals from dorsal premotor and primary motor cortices and inspected the single unit and neural ensemble activity tuning property during the time course of reaching and grasping movement. Using quantitative methods, we found the most informative period for grasp encoding. These periods were further used for reach direction and grasp type decoding, which resulted an above90%performance. By continuously predicting the monkey’s movement states, we realized an asynchronous BMI system, which can control a prosthetic hand online to grasp different objects using the same gesture monkey employed.This study realized online grasp decoding and prosthetic hand control for the first time. The innovation are,(ⅰ) verifying the neural pattern changes of PMd area during reach and grasp course;(ⅱ) decoding four different kinds of grasp gestures with an accuracy rate of85%;(ⅲ) realizing online BMI system for grasp prosthesis with an accuracy of75%. The results work out both neural mechanisms and real-time control problems. Overall, our results inspect the neural firing properties and enables asynchronous neural control of a prosthetic hand, which underline a feasible hand neural prosthesis in BMIs.