新的独立成分分析算法实现功能磁共振成像信号的盲分离

BLIND SOURCE SEPARATION FOR FMRI SIGNALS USING A NEW INDEPENDENT COMPONENT ANALYSIS ALGORITHM

采用独立成分分析(independentcomponentanalysis,ICA)的一种新的牛顿型算法来提取功能磁共振成像(functionalmagneticreasonanceimaging,fMRI)信号中的各种独立成分(包括与实验设计相关的成分以及各种噪声)。与fastICA相比,该算法减少了运算量,提高了运算速度,而且能够很好地分离出各个独立成分。结果表明该算法是一种有效的fMRI信号分析手段。

In order to separate independent components (task-related signal and other noises) from functional magnetic reasonance imaging(fMRI)signals, a new independent component analysis algorithm was used. In contrast to fastICA, the algorithm reduced computation and raised speed of operation. It also separated independent components from fMRI signals very well.