Hi,
I have some structural images from 3 groups that I would like to analyze in terms of proportions of white matter-grey matter. I don't want to perform a whole-brain morphometry, I just want to check the proportions in certain ROIs. In order to do that I used FAST and got partial volume maps, in which intensities of voxels are proportions of WM and GM in that voxels. Then, I used flsmeants to get the mean intensities from the areas using masks I created from MNI 3mm brain template. The problem is that in order to use the masks, the images have to be registered to the same space as the masks (in this situation to a 3mm MNI template). But when to perform registration? When I registered the images before using FAST (after running BET) I got totally diferrent intensities from fslmeants than those I got when I registered FAST-generated output images. Which approach is the right one, and what is the difference?
Thanks in advance,
Michael
|