作者: K.J. Worsley , M. Andermann , T. Koulis , D. MacDonald , A.C. Evans
DOI: 10.1002/(SICI)1097-0193(1999)8:2/3<98::AID-HBM5>3.0.CO;2-F
关键词: Isotropy 、 Point spread function 、 Mathematics 、 Parametric statistics 、 Maxima and minima 、 Image warping 、 Random field 、 Mathematical analysis 、 Smoothness (probability theory) 、 Surface (mathematics)
摘要: If the noise component of image data is nonisotropic, i.e., if it has nonconstant smoothness or effective point spread function, then theoretical results for P value local maxima and size suprathreshold clusters a statistical parametric map (SPM) based on random field theory are not valid. This assumption reasonable PET smoothed fMRI data, but these projected onto an unfolded, inflated, flattened 2D cortical surface. Anatomical such as structure masks, surface displacements, deformation vectors also highly nonisotropic. The solution offered here to suppose that can be warped (in sense) into space where isotropic. subsequent corrected values do depend finding this warping; sufficient only know warping exists. Hum. Brain Mapping 8:98–101, 1999. © 1999 Wiley-Liss, Inc.