基于空间约束的时间独立成分分析方法分析fMRI数据

Analysis of Functional MRI Data Using Temporal Independent Component Analysis with Spatial Restriction

首先采用相关分析初步检测可能的功能激活区域,并以初步检测的功能激活区域作为空间约束条件,对fMRI数据进行时间模式的独立成分分析,然后利用功能实验设计时序信息,通过典型相关分析方法对独立成分排序,自动识别与功能实验设计相关的功能信号成分,最后以识别的功能信号成分作为参考函数,重新利用相关分析自适应地分析fMRI数据。通过对实际的fMRI数据分析验证了提出方法的有效性及可靠性。

Functional MRI data is analyzed by using temporal independent component analysis confined to the potential activation extent which is firstly detected by using correlation analysis. Based on the prior information about the activation paradigm, the functional signal component is identified automatically by ordering the resultant independent components with canonical correlation analysis. Finally, the functional MRI data is reanalyzed adaptively by using correlation analysis with the recognized functional signal component as the reference function. The experimental results of analyzing actual functional MRI data show the validity and the reliability of the presented method.