基于独立元分析的MEG(脑磁图)数据分析和处理

独立元分析(independent component analysis，ICA)可用于分离混迭的MEG(Magnetoencephalography)多通道信号中的信号源。从ICA分解的结果中消除干扰源和噪声，并将剩余分量投影回MEG多通道数据空间，可得到净化的MEG信号，表示各个信号源的各独立元分别投影回多通道，可对各活动源进行空间定位。特别是，响应于外界刺激的诱发活动源亦可从重叠的MEG多通道信号中得到分离，这对脑功能研究及脑医学临床应用极有吸引力。提出了一个简单有效的基于ICA的MEG数据分析和处理方法，研究和分析了一些实际应用问题，特别是给出了听觉诱发响应的一些有意义的分析结果。

ANALYSIS AND PROCESSING OF MAGNATO ENCEPHATO GRAPHY(MEG) DATA BASED ON INDEPENDENT COMPONENT ANALYSIS (ICA)

It has been verified that the ICA can isolate artifacts from overlapping MEG signals. Removing these artifacts from ICA representation, and then projecting the remaining components back into the data space, cleaned MEG data can be obtained. Other independent components can be separately projected back onto the scalp to show the patterns of activities related to the human brain. That the evoked signals in response to stim-ulation can also be separated from the overlapping MEG data is extremely attractive for the study of brain functions. Thus, a simple and efficient method for the analysis of MEG data has been proposed. Some practical problems for application were been studied and discussed, and some interesting results for auditory evoked ex-periments were been shown in the paper. The results demonstrate that ICA is an extremely promising tool for the processing and analysis of MEG data.