[FieldTrip] Source statistics after applying a priori mask on gridpoints (or atlas?)

[Cécilia Mazzetti]

Thanks Tzvetan!

Yes, applying boferrroni to the extracted masks would actually be good enough I think..

Out of curiosity, do you think it would be ok to set all the power values outside the mask(defined grid of interest) to nan and perform source statistics as common with the manually masked power data for all conditions?


Thanks again for the help!!


Cecilia


Cecilia Mazzetti, Postdoc Researcher

Faculté de médecine de l'Université de Genève

Chemin des Mines 9, 1202 Genève

+41 2237 90992

________________________________
From: fieldtrip <fieldtrip-bounces at science.ru.nl> on behalf of Tzvetan Popov <tzvetan.popov at uni-konstanz.de>
Sent: Friday, November 13, 2020 12:43:53 PM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] Source statistics after applying a priori mask on gridpoints (or atlas?)

Dear Cecilia,

Hello fieldtrippers,
Have a question regarding ft_sourcestatistics.
I have a priori defined mask around central gridpoints which i would like to apply before running source statistics over two groups/conditions.
Is there a way/configuration option to do so in fieldtrip?
I am not aware of it.
Does it make sense or that is actually not commonly done?
Whether or not it makes sense is up to you to narrate. Indeed it is not commonly done.
(if so why?)
The goal of the statistic machinery as implemented in FieldTrip (both sensor and source level) is to provide you with a principled decision whether or not the data from 2 or more conditions is exchangeable. If you happen to find a cluster of sensors/voxels regardless of where in the brain, you reject this H0. Thats it. The machinery does not tell you much about where in the brain, time and freq etc.

In your case, I would simply extract the values from your mask of interest and apply Bonferroni correction. Wouldn’t that suffice?


another option could be to use an atlas (although it'd be more precise to be able to apply a mask based on gridpoints in my case)
As explained here https://www.fieldtriptoolbox.org/tutorial/salzburg/#compute-the-leadfield you can use your binary mask to constrain your source model. Then you use grand averaging and source statistic as you deem appropriate.

Good luck,
Tzvetan



Thanks in advance for sharing your wisdom!

Cecilia

Cecilia Mazzetti, Postdoc Researcher
Faculté de médecine de l'Université de Genève
Chemin des Mines 9, 1202 Genève
+41 2237 90992
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20201114/625d49b2/attachment.htm>