 |
 |
Hi, These matrices are used to do post-hoc calculations and do not affect the ICA decomposition at all. They therefore do not play any role in denoising and so no advantage in having them. All the best, Mark > On 26 Sep 2017, at 13:22, Alexandre Obert <[log in to unmask]> wrote: > > Dear all, > > I'm trying to use MELODIC and I saw that it's possible to enter a design.mat and a design.con in the analysis. I read that these options are implemented for dual regression analysis. > However, I wonder if there is any advantage to use these options in a denoising process, as it gives statistical significance of each IC for each contrast set up in a low level analysis (that could be an interesting cue). > > Regards, > > Alexandre