Hello,
Occasionally I run into a problem with lower quality, noisy
T1 images (1.5mmx1mmx1mm) where almost all of the tissue is classified as white
matter. Upon inspecting these images (attached), I can understand why it would
make this mistake; the white-gray matter border in them is very difficult to
identify, and tissue intensity is almost homogenous, except for areas of
partial voluming at the very outer edge (which is the only tissue contained by
the pial surface – the white surface on these images).
Is there any way I can force the white matter segmentation
step to be more conservative with some of its settings? For example, could I modify
the intensity “threshold” for the wm segmentation step? Since this
isn’t an error described in the wiki, I’m guessing control points
won’t do me much good. Any suggestions would be great.
Victor Laluz : Imaging Coordinator : UCSF Memory and