Hi Jon,

Who said we don't? It is possible to do partial volume correction to the images that go into randomise. In fact, Doug Greve has recently published a paper on this topic, that revises and compares various methods. Most are compatible with randomise.

All the best,


On 11 August 2017 at 04:30, John anderson <[log in to unmask]> wrote:
Dear FSL experts,
I have two groups of PET images. I registered these images to MNI152 then I merged these images together using "fslmerge". I created a design matrix for a group comparison between the groups. Then I ran randomise with 5000 permutations and TFCE.

My questions is about partial volume effects PVE.

In this type of analysis do you correct for PVE? Specifically how this can be handled in Randomise?

Recently, I read about PVE in PET Surfer https://surfer.nmr.mgh.harvard.edu/fswiki/PetSurfer and I am a bit confused about the fact why in surface based analyses we need to correct for PVE and while in a group comparison analysis using randomise we don't do it.

I appreciate any clarification.
Looking forward to learn from you