Dear Freesurfer team,

I have a more general question about how best to approach multiple comparison correction for a large dataset where permutations are not viable and where fsaverage resolution is too high. I've fitted a GLM with mri_fitglm in the fsaverage5 space, which results in smaller files and quicker compute times. Next, I've attempted to create MC simulations for fsaverage5 (following https://surfer.nmr.mgh.harvard.edu/fswiki/BuildYourOwnMonteCarlo), but am running into errors from mri_mcsim:

Segmentation Fault (core dumped)

mri_mcsim works fine on subject bert, however.

I am aware that others also had this problem with fsaverage5:

https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg70327.html

Could you please advise on the best way to run MC correction with large-scale data (30k+ subjects)?

Thank you for your help,

Best wishes,

Peter

PS Please find the full mri_mcsim output here:

 mri_mcsim --o ./mult-comp-cor/fsaverage5/lh/cortex --base mc-z --surface fsaverage5 lh --nreps 5000 --fwhm 10