Hi Ben
> I have a couple of questions regarding coregistration between a T1 image and
> the EPI images with SPM5. Is it better to use the T1 as the reference image,
> the mean EPI as the source, and apply the parameters to all other images, or
> the other way around, the mean EPI as the reference and the T1 the source?
In coregistration the reference image is the one that remains
stationary, and the source is moved to match it. Following
realignment your mean EPI is aligned with all of your EPI images, so
you generally don't want to move this. Normally you would specify the
mean EPI as the reference, and the T1 as the source; I don't know of
an advantage to doing anything more complicated.
> Using the EPI as the reference, only the T1 image is moved to match the EPI
> image. It saves processing time, but the T1 loses its resolution after
> reslicing. Is it possible not to reslice the T1 image?
If you just select Coreg:Estimate the headers of the T1 image are
still updated to reflect the coregistration, even without reslicing.
You can check this by using the Check Reg facility to ensure things
are lined up properly after coregistration (i.e. select the T1 and
mean EPI). Or, by looking at the header information before and after
using something like fslhd (from FSL).
Hope this helps,
Jonathan
|