基于独立元分析的脑磁图数据分析和处理(英文)  

Processing and analysis of magnetoencephalographic data based on independent component analysis

作　　者：王斌[1,2] 马洁铭[1] 张立明[1,2] 

机构地区：[1]复旦大学电子工程系 [2]复旦大学脑科学研究中心,上海市200433

出　　处：《中国临床康复》2005年第28期254-256,共3页Chinese Journal of Clinical Rehabilitation

基　　金：国家自然科学基金资助(30370392)~~

摘　　要：背景:诱发响应信号是由刺激的时间锁定的,对于一些特定的刺激呈现小的个人差距,脑磁图数据中诱发响应的提取对人脑功能的认识很重要。目的:将独立元分析应用于分离混迭的脑磁图多通道信号中的信号源,提出一个简单有效的基于独立元分析的脑磁图数据分析和处理方法。设计:单一样本分析。单位:复旦大学电子工程系和复旦大学脑科学研究中心。对象:实验于2002-09在日本通信综合研究所关西先端研究中心完成,选择日本东京药科大学的健康志愿者1例,男性;年龄23岁。受试者自愿参加。方法:①对脑磁图进行必要的预处理,如低通滤波和主成分分解。②采用独立元分析的方法对取自148个通道的脑磁图的数据进行分析和处理,尤其是诱发反应的提取。③对提取的各独立成分进行周期平均。主要观察指标:应用独立元分析方法对脑磁图数据分析。结果:①脑磁图信号有较高的冗余度,信号能量的绝大部分集中在前30个主成分中,从前30个主成分中抽取干扰源和诱发响应活动源。②眼动干扰源仍被清楚地检测和分离在第1个独立元中,心电干扰被分离在第20个独立元中。③α波呈现在第2,3,7和9等独立元中。波(13～30Hz)呈现在第11和第12独立元中。④诱发响应是响应于刺激的周期性波形,集中在第5独立元中。结论:利用独立元分析,可从混迭的脑磁图数据中分离这些干扰源,更进一步,消除这些干扰成分,可得到净化的脑磁图数据。借助独立元分析,有效的分离α波、β波以及眼动、眨眼等神经活动源,有可能为它们的脑神经活动研究提供新的方法和途径。利用独立元分析方法成功的进行了听觉诱发反应的分离和提取。BACKGROUND： Induced response signal is blocked by the time of stimulation, showing some individual differences by special stimulation. Extracting induced response from magnetoencephalographic （MEG） data is important for understanding the function of human brain. OBJECTIVE： To apply independent component analysis （ICA） for overlapping multi-channel MEG signals so as to put forward a simple and effective method to analyze MEG data. DESIGN： A single sample analysis. SETTING;： Electronic Engineering Department and Brain Scientific Research Center, Fudan University. PARTICIPANTS： The experiment was completed at Kansai Advanced Research Center of Japanese Communications Research Laboratory（CRL） in September 2002. One male healthy volunteer aged 23 years was selected from Tokyo Medical University of Japan, and other testees participated voluntarily. METHODS： ① The multi-channel MEG data were preprocessed, such as lowpass filtering and principal component decomposing. ② ICA was adopted to process the 148-channel MEG data, especially for the extraction of evoked response. ③ The average computation was performed for the extracted independent components. MAIN OUTCOME MEASURES： Results of MEG data with ICA method. RESULTS： ① MEG demonstrated high redundancy in the original signal set. Most of the signal energy could be compressed in the first 30 principal components. In other words, the artifaets and evoked activations were extracted only from the first 30 principal components. ② The eye movement artifacts were detected and isolated to independent component 1, and the cardiac interference was clearly concentrated in component 20. ③ Alpha bursts should be detected in components 2, 3, 7 and 9. Beta bursts （13-30 Hz） should be separated in independent components 11 and 12. ④ The evoked activation was obviously concentrated in component 5, which appeared a periodical waveform in response to the auditory stimulus. CONCLUSION： Interference source is separated from multi-channel MEG s

关 键 词：脑磁图描记术 算法 信号处理 计算机辅助 诱发电位 听觉 

分 类 号：R749.05[医药卫生—神经病学与精神病学]