Issue with tfce

External Email - Use Caution

Dear mne community,

I'm doing 1 sample permutation cluster t-tests (two tailed) between 2 conditions.

Most of the values in my array data (difference between conditions) are negative, hence I'm expecting negative T values (which is what I get when I'm not using tfce). However, when using tfce, the test returns only positive T values. Is there a way to overcome this issue?

Thanks in advance,

Fleur
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200424/caab048c/attachment.html

External Email - Use Caution

Hi Fleur,
Quoting from Smith & Nichols 2009 (https://doi.org/10.1016/j.neuroimage.2008.03.061):

"By default, voxels that were originally negative will be given zero TFCE score. However, if it is desired to apply the enhancement to both negative and positive values, the original image can be simply negated, passed through the enhancement, re-negated and combined (via addition) with the positive enhancement."

As far as I know, MNE-Python's implementation does *not* do this for you automatically.

-- dan
Daniel McCloy
https://dan.mccloy.info
Research Scientist
Institute for Learning and Brain Sciences
University of Washington

??? Original Message ???

External Email - Use Caution

Hello,
I am actually wondering if something might be wrong with the way the tfce is performed.
Almost all values in my array are negative, and I get only positive t values. I get exactly the same result when performing a test with tfce on X and on -X. Which means that the sign of the t values is wrong (for some of them at least), unless I'm missing something.

(I'm using 2-tailed cluster permutation test with tfce = dict(start=0., step=0.5)).
Kind regards,
Fleur

External Email - Use Caution

It is possible we don't do the re-negation step. Can you open an issue on
GitHub? If you can show the problem on some tiny toy data (3 "subjects", 2
"vertices", 2 "time points"?) that would help.

Eric