Design set-up
Posted on 03/12/10 06:38:05
Number of posts:
Dear all,
I would need some explanation on how group comparison has been done in PLS? For example I have a Block fMRI data with three groups (A,B,C) and three with-subject conditions (c1,c2,c3) and I have done mean-centering PLS specifying three different groups. I've got two significant LVs ,both of them reflecting commonalities between groups. For example LV1 for group 1 reflect condition c1 vs.c2 & c3 as this is the same for group 2&3. What I am interested in is just the differencies between group A versus B&C? I guess, the best approach of doing so can be provided utilizing non-rotated PLS. However how one can specified a certain contrast (e.g. [2 -1-1] accounting for group A vs.group B&C) while I have three condition?should it be specified within condition? for example if one is interested in the differcies between groups A vs. group B & C for condition c1 shall we set the contrast like:
group A: [2 0 0]
group B[-1 0 0]
group C[-1 0 0]
or shall one make a new detamatrix for each subject with only that specific condition and then make three groups and set the contras like:
group A: [2]
group B[-1]
group C[-1]
Can anybody give me an example of how to do this?
Thanks,
/Kambiz