Cross-validated MANOVA MVPA

Dear all

Does anybody has experience with the matlab toolbox for cross-validated ANOVA ?
It appears that this innovative method has some crisply advantages compared to standard decoding methods using classification accuracy measures.
Please see:

I would like to analyze my fMRI data using this multivariate form of the GLM. So I have two sessions, first session consists of 30 painful stimulations (PAIN) whereas the second session includes 30 non-painful stimulations (NOPAIN). In a first try, I would like to calculate the pattern distinctness D between PAIN and NOPAIN. In the first-level design (SPM),both sessions include the regressor of interest (painful stimulations, non-painful stimulations resp.) and the 6 motion parameters. Now I have to define the contrast (Cs) which would be Cs = { [ 1 0 0 0 0 0 0 -1]}.

Is this procedure correct ?