联合独立成分分析和典型相关分析盲分析fMRI数据

Blind analysis of functional MRI Data using independent component analysis and canonical correlation analysis

为了自动识别功能信号成分,通过对灰质数据和脑脊液数据独立成分的空间相关性进行典型相关分析,有效地解决了独立成分的排序问题.提出的方法不需要任何先验信息,能够稳健地识别与实验设计相关的功能信号成分,实现了对fMRI数据的盲分析.通过对临床真实fMRI数据的分析,阐明了提出方法的有效性及可靠性.

In order to identify functional component automatically, the spatial correlation of independent components of gray matter data and that of cerebrospinal fluid data is analyzed using canonical correlation analysis, and the ordering of independent components is resolved effectively. The proposed method can robustly recognize the component related the functional activation paradigm without any prior information, and achieve blind analysis of functional magnetic resonance imaging （fMRI） data. The experimental results of analyzing real fMRI data show the validity and the reliability of the presented method.