项目名称: 独立成分分析法在功能磁共振成像数据分析中的若干问题研究
项目编号: No.60805040
项目类型: 青年科学基金项目
立项/批准年度: 2009
项目学科: 武器工业
项目作者: 龙志颖
作者单位: 北京师范大学
项目金额: 20万元
中文摘要: 独立成分分析法(Independent Component Analysis, ICA)是一种用于从数据中挖掘出内在独立信号源的探索性工具。虽然ICA已经成为越来越受欢迎的一种功能磁共振成像(functional magnetic resonance imaging, fMRI)数据分析方法,但是该方法自身还存在一些的问题。本项目从以下三个方面在方法学上对ICA进行了系统、深入的探讨:(1)ICA结合投影(ICA with projection, ICAp)方法的改进及在多任务多被试fMRI数据分析上的扩展。(2)fMRI数据的独立成分个数估计方法的研究。(3)实时ICA方法的研究。最后我们将ICA方法用于老年人情绪自传体记忆的脑功能网络机制的研究中,并且开发ICAp处理的软件平台,以方便成果的推广与应用。本项目的研究成果对ICA方法的改进以及在fMRI数据分析中的扩展应用都具有重要的意义。
中文关键词: 独立成分分析;功能磁共振成像;成分个数;实时;多任务
英文摘要: Independent Component Analysis (ICA) is an exploratory method that can extract the intrinsic independent signal sources from the data. Although ICA has become a popular method for functional magnetic resonance imaging (fMRI) data analysis, it still has some limitations. The current project deeply and systematically investigated the improvement of the ICA method focusing on the following three issues. (1) The improvement of the ICA with projection (ICAp) method and the generalization of ICAp to the multi-task and multi-subject fMRI data. (2) The investigation of estimating the number of independent components of fMRI data. (3) The investigation of real-time ICA. Finally, the ICA method was applied to explore the brain functional networks underlying the emotional autobiographic memory of older adults. Moreover, we developed the software platform of the ICAp method to extend the application of ICAp. The achievements of this project are considerably important for improving ICA and generalizing the application of ICA to fMRI data.
英文关键词: ICA; fMRI; number of components; real time; muli-task