Hello all,
I am interested in using the tissue classifications
from SPM5 to calculate intra-cranial volumes. My first
approach to this was to segment an SPGR into c1, c2,
c3 (GM, WM, CSF) and than calculate the volume by
finding the sum of all voxels in the three partitions.
This works fine and gives me an estimate. I then
thought to threshold the segments based on what
probability I wanted to use in order to include them
in the calculation. I used multiple thresholds and,
not surprising, found a sensitivity of the total
intra-cranial volume to the threshold used. Is there
any convention (preferably published) for what
threshold to use for this approach? Or a more
sophisticated calculation of an automated
intra-cranial volume calculation? I have looked an not
found anything. (Sorry if I didn't look hard enough).
Furthermore, is there any way to segment CSF in the
ventricles from the CSF just inside the skull?
Thank you very much,
Jason.s
|