For my study, I conducted a between group t-test in FSL and then did a
cluster analyses to identify clusters that were active for group A vs.
group B above a particular threshold. I used a matlab script to obtain the
mean parameter estimate for each individual for each cluster and
ultimately computed a mean PE for each group for each cluster identified
in the ttest maps. I obtained these means from the PE maps produced in the
lower level analyses. Are the means derived from PE maps the same as the
means that go into the higher level analyses (t-test in this case)? I know
that both the varcopes and pe's are used in the higher level analyses and
I was wondering if there might be any kind of weigthing of the means by
corresponding varcopes before the t-test is conducted.
thanks for your help.