Flip the images prior to spatially normalising them - and use a symmetric
template (or tissue probability maps).
The stats part of SPM probably objected to your data because the images were
not in voxel-for-voxel alignement (ie the matrices in the headers or mat
files were different). You could reslice the images to bring them into
alignment (eg by Coreg and reslice only).
On Thursday 22 January 2009 10:48, Alf Mante wrote:
> in a experiment we differentiated dominant and non-dominant hands in
> healthy subjects. One subject was left dominant and the other ones right
> dominant. For group analysis I flipped the volumes from the left hand
> person. In SPM5 model estimation there is an error, because of the other
> orientation from the flip images of right dominant person.
> Does someone have a suggeston?
> Thanks for help!