Hi Anderson,

Thanks for the help. When viewing my results they looked very strange. Upon further investigation it looks as though the mask I supplied to PALM was a white matter mask (mask.mgh from running qdec initially) created when I ran qdec. I assumed this would be the whole cortex but I was wrong. Therefore it seems to only run permutation testing on the surface of the white matter as seen in the attached photo. Due to the fact that it is unsmoothed white matter, I think this is why we see some speckling bleeding through near the boundaries

In order to do permutation testing accurately for surface based cortical thickness, would the mask need to be a volume file which is between the pial and white matter surfaces or would it just need to be the pial surface (lh.pial / rh.pial), or something else? Any suggestions on the best way to create this?

Thanks,

Ajay

On Thu, Sep 15, 2016 at 1:46 PM, Ajay Kurani <dr.ajay.kurani@gmail.com> wrote:

Hello Freesurfer Experts,
I was running permutation simulations on cortical thickness data and I had an issue with non-orthogonal covariates with mri_glmfit-sim -perm. I then tried FSL's PALM which is an extension of randomize to calculate threshold free stats. I saved the output as logp(which is similar to qdec I believe), however I have not been able to load the stats files correctly. The output of palm is lh.thickness_tfce.mgz for my various contrasts.

1) Is .mgz the proper format for the stats files or do I need to convert this to another type like .mgh etc?

2) Can I display this in freeview or is another program needed? I also tried tksurfer but when I loaded the stats file as an overlay nothing displayed. I want to make sure that the stats is loaded as an overlay in freeview/tksurfer and if so, do I need to select anything special so that it scales the logp values correctly?

Thanks,
Ajay