Research On Hand Rehabilitation Based On BCI

Along with the global aging of population, the number of patients with hemiplegia, which caused by cerebrovascular disease or spinal cord injury, continues to increase. At the same time, the number of spinal cord injury due to traffic accidents is also more and more. These patients, who lost limb motor function, not only can’t take care of themselves, but also bring great burden to their families. The traditional ways of rehabilitation not only very expensive, and because lack of active participation of patients, the recovery effect is poor. How to be more effective in rehabilitation for these patients is the realistic demand of global aging of population, but also the information science, neuroscience, rehabilitation engineering and other cross discipline research hot spot.Brain computer interface(BCI) is a system that achieve the communication or control between human brain and computer or other electronic equipment through electroencephalograme(EEG). It can establish the brain output channel which does not depend on the peripheral nerve and muscle. By using BCI can make the patients with hemiplegia take participation of rehabilitation training, so speed up the rehabilitation of their hemiplegia part. In this paper, the Ocular Artifact(OA) automatic removal, adaptive feature extraction and EEG signals acquisition system for hand motor imagery are finish. The main results are as follows:(1) The ocular artifact automatic removal method based on FastKICA and DWTOcular movements are inevitable in EEG collection, and the resulting OA becomes one of the main interferences of EEG due to its great amplitude. Many methods have been proposed to remove OA from EEG recordings based on Blind Source Separation(BSS) algorithm. Often regression is performed in time or frequency domain by completely deleting the OA components. This can cause the overestimation of OA and the information loss of EEG, because EEG and electrooculogram(EOG) mix or spread bidirectionally. Furthermore, there exists a variety of noises, except for OA, and interference coupling in EEG, this also affects the OA removal performance, such as the robustness and anti-interference ability. Here, we propose a novel and generally applicable method, denoted as FKD, for removing OA from mixed EEG signals with the Fast Kernel Independent Component Analysis(FastKICA) and Discrete Wavelet Transform(DWT). In two cases of linear and nonlinear mixed models, many experiments are conducted with Brain Computer Interface(BCI) data set. The experiment results show that FKD has good performance comparing with other BBSbased OA removal methods, and it is more acceptable in actual BCI system.(2) The adaptive feature extraction method of EEG based on OEMD and CSPAccording to the characteristics of the MI-EEG time variant and individual differences, a novel adaptive feature extraction method called OEFCSP which based on the Orthogonal Empirical Mode Decomposition(OEMD), FIR filter and CSP algorithm is proposed. A channel selection algorithm is applied to the bandpass filtered EEG signals to reduce the number of channels. Then, each remained channel of EEG signal is adaptively decomposed into multiple orthogonal Intrinsic Mode Functions(IMFs) by OEMD, and each IMF is further bandpass filtered into multiple sub-band signals in division. Subsequently, the CSP features are extracted from each sub-band signal, and a feature ranking algorithm is employed to reorder the CSP features. At last, feature selection algorithm and classifiers are optimized together to classify the selected CSP features. Experiments are conducted on a publicly available dataset, and the experiment results show that OEFCSP yields relatively higher classification accuracies, better stability and adaptability compared to existing approaches.(3) The experiment design and analysis of signals acquisition for MI-EEGIn order to put the MI-EEG rehabilitation to practical application, we designed the acquisition experiment for MI-EEG of hand open/close, and the above research were applied to actual acquisition MI-EEG offline analysis to verify the effectiveness of the method. By using the Java SWT to design the experimental interface, and using the JMF to call the hand open/close action cue video. And the 64 channels EEG acquisition equipment is the product of Brain Products Company, collected 5 subjects EEG of real hand action and hand MI. Then FKD and OEFCSP were applied to ocular artifacts, feature extraction and classification for acquisition EEG and EOG. The results reflected the individual characteristics of different subject’s hand open/close MI-EEG. It provides a fundamental basis for future research and implementation of online hand MI-EEG.(4) The online rehabilitation system for hand movement based on MI-BCIOn the basis of the above research, the online rehabilitation system for hand movement based on MI-BCI was designed. Based on the MI movement experimental interface, the EEG synchronization acquisition was achieved through real-time control the g.MOBIlab+ collection instrument of g.tec company by using JNI and multithreading technology. Then, the online EOG artifact removal and pattern recognition were realized by using C++, matlab mixed programming and JNI technology. Finally, the classification result was input to the manipulator control module based on DSP through serial communication protocol, so we could control the open/close online of manipulator by hand MI-EEG. The Experimental results show that the system has good real-time, and played a positive role for MI-BCI technology to practice application and enhance patient rehabilitation initiative to improve the effectiveness of rehabilitation.