关键词: 功能磁共振成像, 共线性, 一般线性模型, 正交化, FSL软件包

Abstract: Concernning the collinearity between the regressors in functional Magnetic Resonance Imaging(fMRI) model, a method of orthogonalization was proposed. Firstly, the interest regressor and the regressor to be orthogonalised are determined. Then, the related part is removed from the latter, and the collinear regressors are decomposed into independent part, so as to eliminate the effect of collinearity. The influence of orthogonalization on GLM is also discussed and analysed. Finally, experiments are carried out through some synthetic data and a current popular fMRI data analysis software package, FSL. The experimental results show that orthogonalization can eliminate the collinearity in the model and improve the significance of the interest regressor to achieve accurate functional localization, which is of great significance for the basic research and clinical treatment of the brain.

Key words: functional Magnetic Resonance Imaging(fMRI), collinearity, General Linear Model(GLM), orthogonalization, FSL(FMRFMRIB Software Library)