同步化分析在意识任务脑机接口的应用  

Application of synchronization analysis on mental tasks based brain-computer interfaces

作　　者：王力[1] 蒋本聪 董倩妍 胡晓[1] 

机构地区：[1]广州大学机械与电气工程学院,广东广州510006

出　　处：《广州大学学报（自然科学版）》2017年第6期63-68,共6页Journal of Guangzhou University:Natural Science Edition

基　　金：广州市信息处理与传输重点实验室资助项目(201605030014);广州市市属高校科研资助项目(1201630210);广州大学科技创新培育基金资助项目

摘　　要：为了分析大脑皮层之间在附加语言想象的意识任务时的功能整合和协作,并提升该任务的分类准确率,可计算脑电信号的互相关值以进行同步化分析.(1)计算所有电极两两之间的互相关值;(2)根据差异最大化的原则来选择2组电极对;(3)将这2对电极的互相关值与共空间模式的特征值组合后用支持向量机分类.对10位受试者实验结果分析表明,有7位受试者所选的电极处于大脑皮层的额叶区,说明额叶区在想象期与其他脑区进行了具有差异性的信息交换.2类意识任务之间的平均分类准确率在添加了同步化特征后提高了1.8%.同步化分析既能显示大脑皮层间的信息交流,又能改善意识任务脑机接口的性能.In order to analyze the functional integration and coordination of the cerebral cortex during the mental tasks with speech imagery, and enhance the classification accuracy of the tasks, the cross-correlation values of EEG signals can be calculated for synchronization analysis. Firstly, all electrodes are paired to calculate the cross correlation values. Secondly, two sets of electrode pairs are selected according to the principle of differ- ence maximization. Lastly, the cross correlation values of two pairs of electrodes are combined with the charac- teristic values of common spatial pattern and then classified by support vector machines. The results of 10 sub- jects show that selected electrodes of 7 subjects are in the frontal area of the cerebral cortex, and frontal area has different information exchanges with other brain regions during the imaginary phase. The average classification accuracy between two types of mental tasks is increased by 1.8% after adding synchronization features. Syn- chronization analysis can not only display the information communication between the cerebral cortex, but also improve the performance of the brain computer interface based on mental tasks.

关 键 词：意识任务 脑电信号 同步化分析 脑机接口 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]