Hello experts,
I have 3 separate ROI masks, each comprised of two (bilateral) 3 mm spheres, all made in MarsBar. For two of them (amygdala and FFA), 3D clust sim reports a total number of 350 voxels in the mask, but for the third one (MTG) 3D clust sim reports 558 voxels in the mask. Accordingly, it gives a much larger cluster size needed to reach significance with the third mask as compared to the first two. Using AlphaSim gives the same results. All masks were created in an identical manner, and the voxel size and smoothing kernel specified were the same.
This seems odd to me, since all three ROI masks are made up two 3mm spheres, and should be the same size. I re-made the odd one out in case I had made a silly mistake, but got exactly the same output.
Any thoughts about why this might be?
Thanks so much!
|