基于独立元分析的脑磁图源定位  

MEG SOURCES LOCALIZATION BASED ON INDEPENDENT COMPONENT ANALYSIS

作　　者：马洁铭[1] 王斌[1] 张立明[1] 

机构地区：[1]复旦大学信息学院电子工程系,上海200433

出　　处：《生物物理学报》2006年第5期389-396,共8页Acta Biophysica Sinica

基　　金：国家自然科学基金项目(30370392;60672116)~~

摘　　要：在脑磁图源定位问题中,通常感兴趣的是脑内众多神经活动源中的一个或几个,而传统源定位方法,如多信号分类方法,需要将所有源的位置都确定后,通过重组信号波形才能获得所感兴趣源的位置信息。为了提高定位效率,文章作者提出了一种结合独立元分析的脑磁图源定位方法。实验结果表明,该方法能加快定位的速度,同时能够在一定程度上克服噪声的影响,具有更强的抗噪声能力。In MEG （magnetoencephalographic） sources localization, usually only one or several neural activitation sources are concerned. But, according to the traditional methods such as multiple signal classification （MUSIC） algorithm, all the activitation sources must be localized to reform the signal waveforms to determine the locations of the interested sources. In order to obtain the better localization efficiency, a MEG sources localization method based on independent component analysis was proposed. Experimental results showed that this method could speed up the sources localization and was more robust to the additive noise and artifact interferences.

关 键 词：脑磁图 源定位 独立元分析 计算速度 鲁棒性 

分 类 号：R312[医药卫生—基础医学]