多体素模式分析在fMRI数据分析中的应用

用多体素模式分析（multivariatepatternanalysis，MVPA）方法分析功能磁共振成像（functionalmagneticresonanceimaging，fMRI）数据是近年来认知神经科学领域的研究热点。MVPA可分析大脑的解剖信息、大脑的功能活动以及人们的行为表现三者之间的关系，从而对人们某个认知行为背后的神经机制进行探讨。本文主要介绍以下内容：①如何用MVPA方法对fMRI数据进行处理和分析，主要包括数据预处理、特征提取和样本创建、分类器的选择与测试。②如何对MVPA的分析结果进行解释：结合认知心理学的相关知识对MVPA的分析结果进行解释。通过本文，可以对MVPA的基本理论，应用特点和发展趋势，以及如何用MVPA方法进行fMRI数据处理和分析有一个较为全面的认识。

Application of multivariate pattern analysis in fMRI data analysis

Multivariate pattern analysis （MVPA） in functional magnetic resonance imaging （fMRI） data is a popular research hot point in cognitive neuroscience in recent years. MVPA can investigate the relationships among brain anatomical information, brain functional activity and behavior performance of people. Therefore,we can explore the neural mechanism underlying certain particular cognitive behavior. This paper introduces the following contents：（ 1 ） How to use MVPA to process and analyze fMRI data,mainly including data preprocessing,feature extractions ,example creations, and choices and tests of classifier. （2） How to explain the analysis results produced by MVPA, such as explaining the analysis results by taking the knowledge of cognitive psychology into consideration. By reading this paper, readers might have a comprehensive understanding of the basic theory, application features and development trends of MVPA, furthermore, a thorough understanding of how to use MVPA to analyze fMRI data.