[FieldTrip] FDR correction of multiple cluster-based permutation tests

[Blume Christine]

Dear Son,

I had a similar issue and had talked to our mathematicians/statisticians. They recommended to use Bonferroni...although that is not the answer you were looking for and other people might have better ideas, I though the info might still be valuable.

Best,
Christine

Von: fieldtrip-bounces at science.ru.nl [mailto:fieldtrip-bounces at science.ru.nl] Im Auftrag von Ta Dinh, Son
Gesendet: Dienstag, 03. April 2018 11:06
An: fieldtrip at science.ru.nl
Betreff: [FieldTrip] FDR correction of multiple cluster-based permutation tests

Dear FieldTrip list,

I would like to correct the results of multiple cluster-based permutation tests using FDR. This is obviously trivial when I have a single cluster for every single test. However, this does not happen in general. My biggest problem is when I have no cluster at all because I also don't have a p value then.
Is there a reasonable way to extract a p value from the cluster-statistics when no cluster can be found? I have thought about just using p = 1 in these cases, but that seems very arbitrary and intuitively wrong.
Can anybody think of a different way of FDR-correcting multiple cluster-based permutation tests?

Obviously, I could also just use Bonferroni to adjust the alpha level, but that is simply too restrictive in general. But maybe someone knows a different way of correcting for multiple comparisons for this case that is neither Bonferroni nor FDR?

Any help and/or references would be greatly appreciated!

Best,
Son

Son Ta Dinh, M.Sc.
PhD student in Human Pain Research
Klinikum rechts der Isar
Technische Universität München
Munich, Germany
Phone: +49 89 4140 7664<tel:%2B49%2089%204140%207664>
http://www.painlabmunich.de/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20180403/6e2f2b50/attachment-0002.html>