滤除功能磁共振成像信号干扰成分的小波方法  

Removing Confounds from Functional Magnetic Resonance Imaging Signal by Wavelet Transform.

作　　者：支联合[1] 单保慈[1] 刘华[1] 李琳[1] 王鹏[1] 唐孝威[1] 

机构地区：[1]中国科学院高能物理研究所核分析技术重点实验室

出　　处：《航天医学与医学工程》2006年第4期306-308,共3页Space Medicine & Medical Engineering

基　　金：国家自然科学基金资助(30570508)

摘　　要：目的滤除fM R I时间序列信号里的干扰成分。方法对fM R I时间序列信号进行多级平稳小波分解,得到不同时频尺度内的小波系数。去掉干扰成分所在时频尺度内的小波系数,对保留的小波系数进行重建得到滤除干扰成分的时间序列信号。结果多名被试者实验数据的平均结果显示本文方法有效滤除了信号内的干扰成分。结论本文为fM R I时间序列信号预处理提供了一种新的高性能的方法。Objective To find a method for removing confounds from fMRI time course. Method Stationary wavelet transform was used to decompose the fMRI time course signal into wavelet coefficients of different time-frequency scales. Then the wavelet coefficients of time-frequency scales in which confounds existed was deleted and the signal without confounds was obtained after reconstructing the remaining wavelet coefficients. Result The mean result of the multisubject data demonstrated that the proposed method removed confounds in fMRI time course efficiently. Conclusion The article provides a new and high capability method for optimal fMRI time series preprocessing.

关 键 词：功能磁共振成像 平稳小波变换 随机噪声 低频漂移 数据处理 

分 类 号：R319[医药卫生—基础医学] TP391[自动化与计算机技术—计算机应用技术]