Hi Glen,
the point is how good is the EPI-T1 coregistration. If EPI images have
geometrical distorsion (usually at the lower part of the frontal lobe
and temporal poles) and you don't have the field map to correct for
it, then the best option should be to normalize EPI to the EPI
template and T1 to the T1 template. In the other case, when EPI images
has a perfect matching with T1-structural image, the best option
should be to normalize T1 to T1 template (by segmentation) and then
apply the resulting parameters to registered EPI images.
Hope this helps
Carles
Glen Lee <[log in to unmask]> ha escrit:
> Dear SPM experts,
> I'm trying to normalize both structural and EPI images onto the MNI
> templates after the coregistration step using SPM8.
> Reading threads of emails in the SPM archive, it seems that there are two
> different sequences of mapping each of them onto the standard space:
>
> 1) normalize the structural image onto the MNI T1 then apply this
> parameters to the EPIs to map those to the MNI EPI.
> 2) (reverse order of the first one) normalize the EPI images first then
> structural after that
>
> Which of those above works better? Or there is essentially no difference
> between them?
> Any help would be appreciated.
>
> - Glen
>
|