Dear all,
I have a cohort of longitudinal MRI data from two time points with one-year interval, and I would like to do longitudinal VBM analysis. I tried the preprocessing procedure provided by SPM8, which seems to be quite simple, but I have several questions about it.
1) As the manual said (p.18), it would be necessary to change some settings, such as DARTEL or spm8 default normalization (Line 65, replace the default 1 to 0). But why and what is the difference?
2) which segmentation is used in longitudinal VBM analysis, DARTEL or unified segmentation? I noticed that the tissue probability maps (TMP.nii) seem to be the ones for DARTEL, so it uses DARTEL for segmentation, right?
3) There are two 'modulated normalized' options, Affine+non-linear and Non-linear only. It is recommended to use the latter, but I find many papers using Affine+non-linear (Jacobian determinant). I am puzzled by that, so which one is better?
4) Since the procedure is supposed to get modulated images (I guess), but I only get wp1/2* files (normalized after segmentation) instead of modulated normalized images after running the batch. Maybe there is something wrong? How to modify it?
Sorry for so many questions, but I am really eager to any help from you. Thanks so much.
Best,
Yaqiong
|