 |
 |
Hi all again, I have one more question regarding the use of FNIRT which is also related to the problem I've described in my previous post. I need the MNI152_T1_2mm_brain registered back to the space of my structural data (which is of resolution of 1mm). The proposed scheme in the fnirt user guide would be 1) fnirt the Struct_1mm.nii to the MNI152_T1_2mm, 2) calculate the invert transform , 3) apply the invert transform to the MNI152_T1_2mm_brain. Isn't it possible to use fnirt to directly register the MNI152_T1_2mm_brain to the native space of my subject without having to resort to the inverting process? Something like: flirt -in MNI152_T1_2mm_brain.nii -ref betted_struct_1mm.nii -omat aff_mni2struct.mat fnirt --in=MNI152_T1_2mm.nii --ref=struct_1mm.nii --aff=aff_mni2struc.mat ... --cout=warp_mni2struct.nii applywarp --in=MNI152_T1_2mm_brain.nii --ref=struct_1mm.nii --out=MNIbrainInStruct.nii --coef=warp_mni2struct.nii. If this option is possible what changes need to be done to the configuration file cnf? Correct me if I am wrong but when we fnirt an image to the MNI template there is some registration error in the process. When later on we use invwarp to reverse the my_struct->MNI mapping then some extra error related to the inverting process is adding up. So can't we avoid this extra error by just directly fnirt the MNI template to the Structural image space? Thank you again Antonios Konstantinos