I read somewhere that if movement parameters are greater than 200 microns, they
are significant. Based on that, I am now always using unwarp, and always using
the movement parameters as multiple regressors. But is the unwarp step useful?
I did some tests, and it seems the unwarp step is helpful. But can it
be harmful?
How could I find out, if possible quantitatively?
Next, if I always use unwarp, do I need the realign step at all? My
impression is that I
do not need it, the realign calculations are done again in unwarp
anyway. Is that true?
If the decision to use realign or unwarp depends on the context, then
one study is
intrasubject multi-sessions of epileptic patients, and another is a
sleep study with
a standard 2-level group single-session factorial design.
Thanks,
Philippe
|