Thanks.
My goal is to calculate the functional connectivity of several seeds that are close to each other. Since they are close to each other, there will be a lot of common variance between them and I wanted to just see the unique variance for each of them. I thought I could do that (getting just the unique correlations for each seed) by orthogonalizing each seed's time series with respect to the other seeds' time series. I've been looking at a way of doing that with 3dDeconvolve, but I'm not sure that it's possible. I was also checking today the 1dsvd program that you suggested, but I don't think this will help me either.
Maybe I'm wrong and there's an easier way of seeing the unique connectivity of each of these seeds.