You are right that SIENAX bypasses the problem of differentiating CSF and
skull by normalising using head size (via registration to the MNI152)
instead od W+G+CSF.
You CAN find the estimated CSF volume by running avwstats on the pve_0
image but as you point out in cases of severe atrophy some peripheral CSF
will get disregarded this way.
Your other suggestion is also a possibility - if you divide the MNI152
"ICV" (on the basis of adding up the standard space apriori maps and
thresholding at totalP=0.5 I have just estimated this as 1847712 mm^3) by
the SIENAX volumetric scaling factor then you could use that.
Regards, Steve.
On Sat, 12 Oct 2002, Wei Wen wrote:
> Hi FSL team,
>
> To avoid the difficult task of separating CSF/skull, sienax normalizes brain
> volume with respect to a standard. This approach is good for many
> applications such as atrophy detection etc. However, if I do wish to have
> intracranial volume as a denominator to normalise some other measurements,
> what should I do then? Since CSF estimation may not be terribly robust, one
> may want to use some methods other than just the sum of CSF, GM and WM. With
> sienax, the scaling factor is worked out. If we have the ICV of the standard
> brain that sienax uses in generating the scaling factor with, then we should
> be able to work out ICV of an individual brain. Am I correct or not? Any
> better method? If it is possible, then where do I find ICV of the standard
> brain (MNI 152 brain?)? The mean of those 152 brains' ICV? Please give me
> some clues?
>
> Many thanks and regards,
>
> Wei Wen
> UNSW/Australia
>
Stephen M. Smith
Head of Image Analysis, FMRIB
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|