partial correlation between pairs of seeds

Hi everybody,

I have a doubt about how to look at partial correlation between pairs of seeds.
So, I have set up my study using 5 ROIs. Two of them are "seeds", in the sense that I am interested in voxel-wise analysis of the connectivity between the rest of the brain and these two ROIs. The other three ROIs will be used in a ROI-based connectivity analysis, comparing the connectivity between each of these 3 ROIs and the 2 seeds.
Now, with the seeds, I would like to do 3 different analyses:
1) "main effect" of each seed
2) "contrast between seeds" (i.e. seed1 > seed2 and vice versa)
3) partial connectivity main effect (i.e. the voxel-wise connectivity of seed1 after having removed the variance related to seed2 and vice-versa).
Now, if i understand correctly, one can simply thick the "partial correlation" box, but in that case I would look at the partial correlation of all the regions included in the analysis (i.e. for each ROIs, the connectivity of that seed, while factoring out the connectivity of all the other ROIs). So I thought that I could enter one of the two seeds of interest in the denoising step, and look at the other. Would it be correct ?
Anyway, if I use this trick, I will have to perform the denoising at least three time (one for the main analysis and the other two for the partial ones), that is pretty time consuming. Is there a faster way to do what I need ?

Thank in advance!

Alain