[Freesurfer] Gray Matter probabilistic Segmentation

Dear FreeSurfer experts,

I have some experience with FSL and SPM, but not a lot with FreeSurfer, so please bear with me.

When I observe FreeSurfer segmentations, I see that images are binary (either a voxel belongs to a tissue class, or it does not) and not probabilistic. I want to obtain total gray matter volume from the FreeSurfer segmentation, so I searched the forum and came across the following post:

http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg11790.html

I followed directions which resulted in three whole-brain gray-matter probabilistic map:

https://dl.dropbox.com/u/6747155/FS_GM_prob.png

The three maps end in 000.mgz, 001.mgz, and 002.mgz.

...

Could Somebody explain to me the differences between these images?

Furthermore, the intensity range in the maps goes from 0 to 40 (as can be observed from the histogram). How do I convert this image to probabilities 0-1? I don't believe dividing by 40 is correct, because that would mean that the very few voxels will have a probability of being grey matter higher than 50% (see the histogram).

Thanks in advance,

Vincent