Research On Finger Rehabilitation System Based On Imagery Motor EEG

At present,as China's population aging,stroke patients increased year by year.The stroke is accompanied by high incidence,high mortality,high morbidity.Even if the patients survive,they will be left with a degree of disability,especially the hand disability.Clinical experiments show that early rehabilitation training can help the patients restore or improve the function of the hand,but for most patients,rely on the treatment of few experts to obtain long-term rehabilitation training is not practical.For that reason,this paper designed a hand function rehabilitation training system based on motor imagination EEG to assist patients with hemiplegia in finger rehabilitation training.The main contents include:(1)Do research on the application of motor imagination EEG signals in finger rehabilitation,and analyze the generation and collection process of EEG signals,then determine the research idea and implementation plan based on IM-BCI according to the research status.(2)Design a flexible and comfortable wearing exoskeleton glove which is based on linkage structure and processed by 3D printing.the glove possesses light weight,force transmission and good robustness.the glove is drived by the linear motor to meet the requirements of the finger rehabilitation training required action.(3)Complete the design of the control system of the finger rehabilitation training system,including the control circuit and the software of on computer.The control circuit uses the STM32 microprocessor as the control core to control the pressure sensor and the linear motor,and communicates with the computer software through the serial port.The computer control system based on QT framework for interface design,to help patients complete rehabilitation testing,rehabilitation mode selection,rehabilitation evaluation and other functions.(4)Design an off-line motion imagination EEG experiment.EEG signals of four subjects in two states of movement and rest of the right index finger were collected by the EEG cap.use Fisher's criterion to select the optimal lead and use linear discriminant analysis(linear discriminant analysis LDA),support vector machine(support vector those,SVM)and gradient promotion decision tree(gradient boosting decision tree,GBDT)three classifier for extraction of EEG feature vector cross validation,and analyze the experimental results.