Dear Martin,
Step 3 three is wrong and because of that you can get into trouble during step 4. :)
As pTFCE already incorporates cluster information, it is not right to perform cluster-level thresholding on it.
Just apply the same voxel-level threshold you would apply for your original unenhanced zstat image. You can also get it from FSL's ptoz with the -g option or in R, from the fwer0.05.Z field of the object returned by pTFCE.
There's also a small command line utility which does everything, using FSL and the pTFCE R-package:
https://github.com/spisakt/pTFCE/wiki/3.-R-package#command-line-utility
I have updated the Wiki pages, see Step 5 here:
https://github.com/spisakt/pTFCE/wiki/3.-R-package#usage
Cheers,
Tamas
>Dear All
>
> In the spirit of being up to date with thresholding strategies, I was looking at pTFCE (see https://github.com/spisakt/pTFCE)
>
>perhaps a real basic question, but is the following flow correct (which I am trying to do in R)?
>
> 1. Get the un-thresholded zstat from the "stats" folder of a third level analysis (this is your "vanilla" FLAME1 3rd level analysis)
> 2. Run pTFCE and write out the enhanced zstat_pTFCE file
> 3. Apply cluster thresholding (e.g., with easythresh?) to the zstat_pTFCE file (at some conventional 3.09 0.05 level).
> 4. Publish in Science
>
> Is this the correct way of using it? (I could follow the instructions on the Wiki to get the files themselves, but I am just a tad unclear on how to properly threshold after that...)
>
> Thank you
>
> cheers
>
> Martin
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|