[pymvpa] search-light vs. ROI analysis: significance puzzle

[Vadim Axel]

Hi MVPA colleagues,

Absolutely naive question: suppose I have single a-priori defined ROI where
I get a modest group-level beyond chance prediction of p-value=0.01
(one-tail t-test vs. 0.5, across subjects). Now I run a group level
whole-brain search-light and I am expected to find at least one cluster of
beyond chance prediction in the environment of my ROI. Correct? But when I
run whole-brain search-light I have to make a multiple-comparison
correction (let's say FDR=0.05), which means that my p-value threshold is
something like 0.001 in best case. Clearly, my cluster cannot survive it.
Does it make sense to you? What is the solution?  I can make a small-volume
search-light (with corresponding small multiple-comparison correction) -
just to show that my ROI analysis is reproducible. Yet, people will say
that who knows, whether with such a liberal correction I can now find many
other clusters in search-light.

Thanks for help,
Vadim
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20130225/892b868b/attachment.html>