Dear Antti,
> I am attempting to do higher level analysis but wantes to keep the resulting
> SPMs in functional image dimensions for the purposes of further processing
> done on these maps (thresholding).
>
> A simple way of accomplishing this would seem to be using one functional
> image from one subject as reference (standard). Could I simply select one
> example_func from the subject group instead of the standard avg152T1_brain
> in Feat? The registration seems to work, but are there potential pitfalls in
> this approach?
Well one potential pitfall is the issue of image distortion and signal
dropout that varies rather significantly from person to person in EPI
images. One of the advantages to registering to a T1 image and from T1 to
a "standard" anatomical image is that one minimises this problem.
If you want to simply keep the voxel dimensions the same after
registration you could simply create an X x Y x Zmm version of the
template brain and use that for regristration. Presumably this will lower
the quality of the registration somewhat because it increases the partial
volume effects but it might not be too bad given the image is fairly
smooth already.
Cheers,
Joe
|