Hey,

I'm running a paired sample T-test with FSL randomise. The data I'm using is searchlight MVPA results. After viewing the output, however, the correction process does not seem to have worked exactly as hoped. We are seeing wildly spread out clusters across multiple regions, including white matter. When we use a gray matter mask, it looks better, but still worrisome. For example, there are a few single voxel activations, which to my understanding of TFCE, shouldn't survive post-correction. Obviously I would like to be confident that my correction process has worked but i have low confidence in these.

Please see the attached images. The first image is without the gray matter mask. The second is with a gray matter mask, displaying the single voxel activation I am worried about. Apologies for the small size, they have to be small for me to post them here.

This is the command I am running:

randomise -i allsubs.nii.gz -o pairedT -d ../../../scripts/design.mat -t ../../../scripts/design.con -e ../../../scripts/design.grp -m gm_mask_no_cerebellum_reslice.nii.gz -T


Has anyone experienced a similar problem before, or does anyone have an idea as to what the problem/solution might be?

Cheers



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

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