Dear Teri,
Thank you – I have read the
paper (
HBM 5(2) 133-6 "Testing for anatomically specified
regional effects") and your technique makes much more
sense to me than using the SVC tool. With your method, a priori regions
can be examined, even with the confounds of limited spatial
resolution, imperfect normalization, and group data issues. Using
the SVC tool, you can just about pick your own p value. But, I
can’t quite figure out how to use the equation. Has there been
discussion about this previously? I realize this is an older
paper... I did look at the archives but nothing really popped
out.
To use this approach in a practical sense, you simply use the uncorrected
cluster-level P-value in the results table. If this P-value is less
than 0.05,
for the cluster nearest your a priori location, the cluster can be
declared significant.
The inference is very simple and can be useful when you know roughly
where
your activation of interest lies.
I hope this helps - Karl