Hello everyone, I am having a question about a pre-processing step that I have never seen used in MVPA: Do we have to remove the serial/temporal correlations? This pre-processing procedure is also widely known as 'whitening'. Has anyone tried to use it?
As shown in the literature, these serial correlations are some form of noise that exist between two consecutive scans (I think whitening procedures do not go for more scans, due to complexity maybe) . I don't think this procedure would help in MVPA, although it is a must in GLM analysis. I wonder if anyone disagree with this opinion. Any hint on this issue is highly appreciated, Best, -Rawi _______________________________________________ Pkg-ExpPsy-PyMVPA mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
