Dear Elisabeth,
On Tue, 14 Jun 2016 08:27:30 +0100, Elisabeth Kaminski <[log in to unmask]> wrote:
>Dear Christian, dear SPM users,
>
>I have a question regarding the expert user settings for the longitudinal version of the CAT12 toolbox.
>Is it possible (or easy to implement) to change the settings for the modulation step in the preprocessing pipeline to "non-linear only"?
>I found out that this is possible in the "non-longitudinal" segment data batch using the expert mode but I have not found a similar line using the longitudinal toolbox batch.
>I was also trying to copy the line of interest in my lngtd job file (matlabbatch{1}.spm.tools.cat.estwrite.output.GM.mod = 2;), however, this doesn´t seem to work!
you can change line 28 in cat_long_main.m to:
matlabbatch{4}.spm.tools.cat.tools.defs.modulate = 2;
An alternative would be to use TIV as global scaling to mimic the old modulation approach. The non-linear-only modulation is equivalent to a global scaling with the (inverse) scaling due to affine registration. Affine (inverse) scaling strongly correlates with TIV. However, "real" TIV is better than this approximation.
Please note that the switch in CAT12 from the old modulation to the SPM default modulation and the additional use of TIV as covariate was not made without any strong reasons. There are a couple of recent papers that demonstrate the advantage of using TIV rather as covariate over global scaling. However, global scaling should be used if TIV is showing a reasonable correlation with your covariate of interest. In that case not only the variance that can be explained by TIV will be removed from your data but also variance you are interested in. I have also updated the CAT12 manual regarding this issue...
Best,
Christian
>
>Thanks a lot in advance for your help!
>Elisabeth
|