Hi,

This seems at least somewhat similar to an issue that I have run in to before. As I recall, my concern was that the COPE values in the unthresholded images using FLAME 1+2 differ based on the threshold chosen, while this is not the case if only FLAME1 is run. My understanding is that the FLAME2 process only runs on voxels that are close to the threshold, where it will actually "matter" to defining the clusters, otherwise just FLAME1 results are used. I'm not sure if that explains what you are seeing, but it seems plausible that with slightly different input data, a similar issue will come into play?

Anyway, hopefully you can get a response from somebody more expert than me, but this may help some.

Best,

Michael

On Sun, Jun 16, 2019 at 9:14 AM Anthony Sali <[log in to unmask]> wrote:

Hi,

I'm still stuck on this and I just wanted to see if anyone had some advice on what might be going on. To reiterate, the issue is that FLAME 1+2 gives slightly different results if I feed into it fixed effects results that were run with all subjects in a single model as on the FSL tutorial website or if I feed into it fixed effects results that were run individually for each subject. The subjects are of course the same in both cases.

I would be grateful for any input!

Thanks,
Anthony

########################################################################

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1