Sorry, message cut off mid-write;
We're having some trouble with the pial surface for the temporal lobes. All four datasets we've processed so far have temporal lobe pial surfaces that are significantly smaller than they should be (attached)
I know there are control points for white matter - is there an equivalent for gray matter?
On Wed, 2006-08-16 at 14:44 -0400, Bruce Fischl wrote:
no - we don't really use the cortex labels from the aseg for this reason (and others). Instead we use the surfaces (?h.white and ?h.pial) for computing cortical properties, so you should be all set.
cheers, Bruce
On Wed, 16 Aug 2006, Bo Shi wrote:
Hi Bruce,
The pial surface for the top regions of the brain do fairly well in ignoring the marrow, however, the unwanted regions still get classified as gray matter (image attached).
This will adversely affect the reported cortex volume no?
Thanks, Bo
On Wed, 2006-08-16 at 14:14 -0400, Bruce Fischl wrote:
Hi Bo,
does it affect your pial surface? You may be fine with just leaving it in.
cheers, Bruce On Wed, 16 Aug 2006, Bo Shi wrote:
Howdy -
It seems that the skullstripping step is having some trouble removing bone marrow (see attached image).
Adjusting watershed values (I've tried as low as 5) does not make any significant improvement.
Does anyone else have this issue? Can anyone provide any hints as to how I might get a better automated skull-strip? I've been manually removing the marrow but it's getting real old.
Any thoughts would be much appreciated.
Bo
