Hi Christian,
It all makes sense now :-) -- and in fact I had noticed before how the ICA decomposition changes in melodic depending on whether you apply global scaling to your data or not (I tried this once on some datasets with artifactual global signal shifts within a single EPI run), whereas this does not seem to make a difference in other sw packages (which, I assume, do remove the global signal).
thanks for clarifying -- and so quickly!
cheers,
giuseppe
|