Research On Finger Rehabilitation System Based On Brain-computer Interface

Brain-Computer Interface(BCI)is a novel interface between human and computers based on EEG signal.It does not depend on normal output channels of the peripheral nerve and muscles of the human body,and established the control channels between the brain and external communication equipment directly.An important application of BCI system is to provide a new way of rehabilitation for hemiplegic patients,which can help patients with hemiplegia to increase the plastic potential of axoneure,reconstruct the damaged cerebral cortex and repair the connection between the external limbs and the central nervous system of the brain,so the BCI system have great significance and wide application prospects for the field of paralytic rehabilitation.In the beginning,this paper introduces the background and research significance of the finger rehabilitation system based on the BCI,analyzes the generation?characteristics and collection method of Electroencephalography(EEG),and combines the research status and research methods at home and abroad of the BCI system,This paper establishes Motor Imagery-based(MI-based)BCI system.Then,according to the theory of hemiplegia rehabilitation therapy and the design principle of rehabilitation system,the whole design framework of the system is proposed.At the same time,the method of feature extraction and feature classification of EEG is analyzed in detail.Secondly,According to the biological characteristics of the human finger and the needs of the rehabilitation of fingers,the finger rehabilitation training mechanism for hemiplegia patients is designed.Then,the lower computer's hardware circuit and software system flow are analyzed in detail,USB communication protocol between the upper and lower computer is designed,and the upper computer's software interface based on C++ programming language and Qt development framework is realized according to the user's requirements.Finally,The experiment collected four subjects' EEG signals in two states of image right index finger movement and rest by EEG cap and related devices,established the optimization selection scheme based on Fisher linear discriminant criteria(FDC)and appropriate feature extraction method was selected.Then,three classifiers of linear discriminant analysis(LDA)? support vector machines(SVM)and gradient boosting decision tree(GBDT)are used to cross verify the EEG feature vectors extracted from different experimental subjects,the experimental achieved a good classification result,thereby,this article provides a good foundation for the research of online finger rehabilitation system based on MI-BCI.