You would typically just use the rc1 and rc2 images in order to
simultaneously align GM with GM, and WM with WM. For more tissue
types, the procedure becomes very memory hungry, although may run OK
on a bigger (64 bit?) computer.
Note if you did want to align soft tissue, CSF, bone etc, then all six
scans would not actually be used. You'd just use the first five as
the algorithm works with an implicit background class, consisting of 1
minus the sum of the tissues. If all six rc*.nii files are used, then
this background class will be all zeros (because
GM+WM+CSF+bone+tissue+air=1), so is not needed.
On 8 March 2011 10:21, Vincent Koppelmans <[log in to unmask]> wrote:
> Dear SPM experts,
> Currently I am creating a DARTEL template of 368 subjects. For each subjects I have 6 RC files (RC1 - RC6).
> When I create a template based on all RC files for all subjects (6 * 368) I get an out of memory error.
> However, when I create a template based on only the first three (3 * 368) RC files (grey matter/white matter/csf) my computer manages to complete the template creation.
> I am wondering whether I need to use all the RC files, or if creating a template based on RC1, RC2 and RC3 also renders a valid template. If so, could I also just run 'create template' six times, each with a different set of RC images?
> Kind regards,