Dear Ariadna,
Concerning steps 3 and 4, this is "too much". With Dartel, the segmented files are warped together, but the final output does not correspond to the MNI space. You would either go with
a) "Normalise to MNI space" - this is an affine registration of your final template into the space of the tissue priors, per default the TPM.nii in SPM, which correspond to the 6th gen asymmetric MNI space variant
b) "Population to ICBM registration" - this is a high-dimensional (non-linear) registration of your final template into the 2009 MNI space variant.
In any case, one of these two is sufficient.
Apart from that, your preprocessing pipeline seems to be overly complex to me. I also don't really understand why you would want to segment your T1 files twice.
With CerebroMatic (or the accompanying scripts available at the toolbox site), you can create age-appropriate Dartel templates. Thus it should be sufficient to adjust the settings in CAT12: Segmentation, i.e. switching to the other tissue probabiliy maps and Dartel templates. Assuming that the CerebroMatic Dartel template is in MNI space then there is no need for a) or b) at all.
If the CerebroMatic Dartel templates were not in MNI space you would still go with CAT12: Segmentation, but with either a) or b) in a following step and possibly the deformations utility.
Only if you were to create your own template it would be necessary to first write out affinely registered tissue files, warp them together with Dartel, and, as the final stage does not correspond to MNI space, go with either a) or b).
Anyway. As I'm not sure so far which of the options you want to choose I won't comment on the details for now (is there a need for the deformations utility yes/no and if yes, which files to select, ...).
Best regards
Helmut
|