In spm8, after realignment of functional data, I specified Reference Image as mean* image, and Source Image as T1 (This seems mean that the resulting mean image was coregistered to the corresponding T1 image). In normalization, I specified Source Image as mean* image, Images to Write as r* images (resliced images), and the Template Image as EPI.nii.
According to the manual, (I) If I set Reference Image as mean* image, and Source Image as T1, the structural image would be registered to the functional realigned images. (II) If I set Reference Image as T1, and Source Image as mean* image, the functional realigned images would be registered to the structural images.
What's the difference between (I) and (II) and, during coregistration, should I use (I) or (II) [my current example above is (I)].
I know that in fsl, functional images are first registered to T1 images, and then normalized to standard template, but I want to make sure if spm does the same processes.
Thanks in advance.