> We tried to process our data in SPM5 as well, and realized
> that there is a difference in the origin of the functional images. When
> we display the pre-preprocessing images on SPM5, the origin of the image
> is as expected (we align the oblique images to the AC PC line). However,
> on SPM2 the origin is significantly shifted. Could this be the source of
> our problems?
I think so. It's not clear to me what you've done in SPM2 and what
you've done in SPM5, but if you do everything in a single version of
SPM, then I think it should work.
SPM5 reorientation will alter the header information in the NIfTI
file, which SPM2 would subsequently treat as an Analyze file,
incorrectly setting the orientation (origin in your case). I think you
might be able to use SPM5 to reorient the images and then store the
transformation in a .mat file as well as the NIfTI headers; then the
image should display correctly registered in SPM2 (baring any trouble
with Analyze_flip settings...) and then normalise-write should be fine.
Try out the following tool:
and let me know if it doesn't work. Be very careful with left-right
orientation -- I haven't tested this carefully -- i.e. do as I say,
not as I have done ;-)