基于相关分析的fMRI时间序列图像处理  

fMRI Time Series Image Processing Based on Correlation Analysis

作　　者：普杰信[1,2] 刘森[3,2] 黄心汉[1] 赵力[2] 

机构地区：[1]华中科技大学控制科学与工程系,武汉430074 [2]河南科技大学电子信息工程学院,洛阳471003 [3]南京理工大学自动化学院,南京210094

出　　处：《中国生物医学工程学报》2008年第6期854-859,共6页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(60475021);河南省杰出青年基金资助项目(0412000400)

摘　　要：脑与认知科学研究中,对功能磁共振图像的分析,不仅要求脑功能激活区的准确定位,而且要得到脑激活区的动态变化。提出了通过对激活区体元时间序列的分解,构造序列参考波形,用相关分析法检测识别每个时间周期的激活类型,得到脑功能激活区的动态变化图像、生理信号变化时间及信号强度曲线。经视觉试验实测数据建模并检验,该方法可有效检测fMRI数据中与试验密切相关的脑激活区及其周期性变化。In the research of brain and cognitive science, analyzing fMRI data can not only detect and locate the functional activated signals accurately but also obtain the dynamic changes of activated areas. In this paper, a new method for fMRI data processing was proposed. With the decomposition for the time series data in the activated areas, the referenced waves of sequence were constructed, the respective activated types in each time period were identified using the correlation analysis. The images of reflecting the dynamic variety were obtained; the curves indicating time and intensity variety of activated physiological signal were obtained too. The experimental results indicated the better validity and reliability of the proposed method.

关 键 词：功能磁共振成像 时间序列 相关分析 参考波形 

分 类 号：R318[医药卫生—生物医学工程]