Dear Guiiaume,
Thanks for the explanation - I think I understand, but really want to make
sure. Please tell me whether the statements below are true:
1. By selecting the F-contrast for all effects of interest (not(!) the individual
T-contrasts) and then selecting (adjust), I will obtain a time series where
variance was removed that was due to patient motion and scanner drift
(i.e. for all effects of no interest).
2. Consequently, this time series represents a time series that contains only
BOLD signal fluctuations that are due to the "effects of interest" and that
would have been observed if no motion and no scanner drift would have
been present during data acquisition.
Finally, I am not clear about how to remove the constant term you mention in your
e-mail. I certainly would like to extract a time series (using spm_regions.m) that is
mean-centered and “whitened” (to get rid of boundary artefacts).
What feature do I need to select when extracting the data?
I appreciate your patience!
Many thanks again
Otto
|