> I read this thread concerning differences between SPM2 and SPM5 for realign
> and normalize fMRI data, and I still have some misunderstandings:
> With SPM2, following realignment step, *.mat files were attached to
> slice-timed (a*.img/hdr) images, so that , in the normalization step, we
> could apply normalization parameters to a*.img (and not to ra*.img ,
> minimizing the number of reslincing steps).
> With SPM5 (using *.nii files), there is no more *.mat files created
> following realignment.
> So my question is:
> -Do we have to apply normalization parameters to ra*.nii images if we want
> to keep the realignment parameters, or can we still apply these
> normalization parameters to the a*.nii images?
The contents of the .mat files are now saved in the NIfTI headers, so you
should be able to continue as previously.
>
> If it is the case, then another point raises: How can we be sure that a
> a*.nii dataset has previously been realigned? the previous versions of SPM
> had a very important feature preventing this: never modify a file without
> adding a prefix in the name.
I would suggest you keep a record of what has been done (either use a log-book
or save all your job files). Alternatively, you could look for the rp_*.txt
file or run the following:
spm_check_orientations(spm_vol(spm_select(Inf,'image')))
If it gives an error, then this is an indication that the images are realigned
but not resliced.
Best regards,
-John
|