Hi I am trying to calculate tPAC but gettign this:
Thanks
Hi Niko,
The function hilbert is implemented in the signal processing toolbox in Matlab. You probably do not have this toolbox. The hilbert transform is not a difficult function to implement yourself though I do not know if the brainstorm team is writing their own implementation.
Kind regards,
Steven
Thanks Steven,
Can you let me know what is the name of this Matlab toolbox? I would prefer to install it.
Dear Niko,
It is the signal processing toolbox.
Kind regards,
Steven
Ok thanks, I will try.
I updated the tPAC code so that it uses oc_hilbert.m (from Octave) when hilbert.m is not available:
Please update Brainstorm and try again.
@Samiee FYI
I replaced this other missing function "findpeaks" with "peakseek", already included in Brainstorm:
Please try once more updating Brainstorm.
Sorry in advance if it takes a few extra iterations, but we'll get there.
Thanks a lot Francois, it is working. I went further and tried to run t-test. I am not able to open t-test file as DirectPAC map, it can be opened only as a topoplot. I wonder if it can be possible to get DirectPAC maps?
It is not possible to compute statistic on the PAC comodulograms: The test is done only on the field "TF", and the structure sPAC is not discarded:
https://neuroimage.usc.edu/brainstorm/Tutorials/TutPac#File_contents
I disabled the PAC display menus for the statistics results from the popup menu:
https://github.com/brainstorm-tools/brainstorm3/commit/639254a4c8c34a1d5ad86178a19fe4502fdf899b
The only statistics available for PAC results are listed in process menu Frequency > Time-resolve phase-amplitude coupling.
@Samiee @Sylvain
Can you confirm?
Should we disable explicitly the computation of statistics on the PAC files, or is still useful to run tests on the TF field?
@niko808080
Parametric statistics computed with N=2 are not very meaningful, but I guess this was only a proof of concept, or an illustration for the bug?
yes, I just tried to test if stats will work.