but is it the dimensions of the cortex that you are trying to generate? Say, from frontal pole to occipital pole? That I would do by specifying those points in fsaverage as labels, then mapping them to individuals using mri_label2label and computing their distance in native coords.
cheers, Bruce
On Wed, 14 Mar 2007, roberto toro wrote:
On 14 Mar 2007, at 00:28, Bruce Fischl wrote:
Hi Roberto,
the tal xforms will have scaling normalization in them, which I assume you don't want?
Bruce
Hi Bruce,
you are right. I would like to know only the orientation of the head and then compute the hemisphere size from the reconstructions. If I compute the eigenvalues of the rotation part of the matrix (the first 3 columns) I can recover the scaling factors, and then I should be able to get the rotation matrix by dividing the rows by their corresponding eigenvalue, isn't it?
thanks for your help, roberto