Research On Visual Guidance And Pattern Classification Of Motor Imagery EEG

Brain computer interface(BCI),using modern information processing technology,has established a new information communication and control technology which does not rely on conventional brain information output path,and provides a new idea for human-computer interaction.Motor imagery(MI)is a common kind of BCI,and its spontaneous characteristics make its application more promising.However,how to effectively obtain and identify EEG signals is still a major obstacle to the development of BCI based on MI.Therefore,this paper carries out the research work on the MI EEG by two aspects of visual guidance and pattern classification:(1)To improve the training efficiency of MI under visual guidance and ameliorate the classification accuracy of BCI,the influence of virtual reality(VR)environment on MI training and the differences of EEG classification models under different visual guidance were studied.Firstly,three kinds of 3D hand interactive animation and EEG acquisition program were designed.Secondly,in the rendering environment of the head mounted helmet(HMD)and the planar liquid crystal display(LCD),the left hand and right hand MI training was conducted on 5 healthy subjects,including standard experiment(the single experiment lasted for 5 min)and long-time experiment(the single experiment lasted for 15min).Finally,through the pattern classification of EEG data,the influence of rendering environment and content form on classification accuracy was analyzed.The experimental results show that there is a significant difference in the presentation of HMD and LCD in visual guided MI training.The VR environment presented by HMD can improve the accuracy of MI classification and prolong the duration of single training.In addition,the classification model under different visual guidance content is also different.When the testing samples and training samples are the same visual guidance content,the average classification accuracy is16.34% higher than that of different samples.(2)In order to reduce the redundant information in feature extraction and improve the accuracy of EEG recognition,a pattern classification method based on wavelet packet decomposition(WPD)-common spatial pattern(CSP)-adaptive differential evolution(ADE)is proposed for feature extraction of EEG signals.Firstly,WPD-CSP is used to extract features of EEG and retain all the features after CSP.Then,ADEalgorithm is used to select feature and select the best feature subset for classification.Based on differential evolution algorithm,ADE algorithm dynamically balances the crossover probability and scaling factor to maintain the balance between global search and population diversity in evolutionary process.The experimental results show that the proposed WPD-CSP-ADE method can effectively improve the classification accuracy and significantly reduce the number of features used for classification.In addition,the ADE algorithm has better performance than genetic algorithm,particle swarm optimization and differential evolution algorithm.