Dear all,
I think the worry for FDR is that it can have odd effects when you
have large areas of activation. For example, imagine the following
two situations 1) the whole of the right hemisphere is strongly
activated by the task, but the left is not activated atall 2) none of
the brain is activated by the task. In case 1), because of the way FDR
works, 5 percent of the voxels on the right side of the brain will be
labelled as activated. For 2) no voxels in the R (or left) should be
labelled as activated.
So, this can be a significant problem if you find a map with lots of
activation, and want to making a big point about a small area.
Personally, because of these worries, I feel much more comfortable
with the FWE in SnPM - which seems to me to be the-right-thing-to-do
(TM),
Best,
Matthew
|