[FieldTrip] One-sample t-test with cluster-based permutation test

Seung Goo Kim, Ph.D. seunggoo.kim at duke.edu
Thu Oct 4 22:44:03 CEST 2018


Dear FT-ers,

I (also) want to know how we could perform one-sample T-test using the cluster-based permutation test to compute and correct p-values.

I noticed that many people asked about this for a long time (https://mailman.science.ru.nl/pipermail/fieldtrip/2018-August/012314.html). I think it's because the one-sample T-test after baseline correction has been really a common way to analyze ERP/F data. Furthermore, in many functional studies (including fMRI studies), the first question at the second-level is if the effect of a certain contrast is non-zero over multiple subjects, thus one-sample T-test is a really natural thing to do.

It is possible to compare baseline-vs-activation trials at the first level, although it could be a problem if the baseline period is too short compared to the activation period, which is not really uncommon (for example, it would be rather common to have short inter-stimulus-interval than the stimulus length itself).

But at the second level, I have no idea how we could do that.

I know that a permutation distribution can be generated by randomly flipping signs for one-sample T-test in permutation test, so I added another resampling method to flip signs of the half of trials in resampledesign() but the result was quite different from the two-sample t-test comparing baseline vs activation trials. Could be there any specific reason for flipping signs would not work for cluster-based correction? One possible reason I could think of would be that the noise distribution is not actually symmetric, thus permuting labels and flipping signs create different permutation distributions.

Is there really no way to use cluster-based permutation test for one-sample T-test at the second level?

Best,
--
Seung-Goo ("SG") Kim, PhD

Postdoctoral Research Associate
O-Lab, Department of Psychology & Neuroscience, Duke University
Postal: 308 Research Drive, Durham, NC 27708, USA
Email: solleo at gmail.com<mailto:solleo at gmail.com>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20181004/36d773c3/attachment-0001.html>


More information about the fieldtrip mailing list