Hi Sasha
Something that works reasonably well for me is this approach:
standard_space_roi target ssroi -b
cog=$(fslstats ssroi -C)
bet target bet1 -c $cog -f 0.3
bet ssroi bet2 -c $cog -f 0.2
Whether bet1 or bet2 is the better extraction depends on how well the
pre-extraction works. Occasionally, ssroi will be undersegmented, in
which case bet2 is hopeless. Also, you can get holes in the mask if
you pre-extract
(https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=fsl;ce7862f8.1201), so
in most cases bet1 will be better, especially if you can tolerate
oversegmentations (reduce -f further if this is the case).
For alternatives, see the papers by Eskildsen et al., Neuroimage 2011
(BEaST) and Leung et al., Neuroimage 2011 (Brain MAPS) -- one or the
other may be worth implementing if you have lots of data to process.
Hope this helps
Rolf
On 24 January 2012 19:40, Alejandro Vicente Grabovetsky
<[log in to unmask]> wrote:
> Hello FSLers,
>
> I'm attempting to brain extract some high-field MP2RAGEs, affected by both
> inhomogeneity (which I can bias correct to some extent), but sometimes other
> artefacts. I know there is an option in BET to include both T1 and T2
> images, is there something equivalent for MP2RAGE, or some other option
> (besides -R, which does not always work) to make things more robust?
>
> Thank you very much in advance!
> Sasha
> --
> Vicente Grabovetsky, Alejandro (Sasha)
> Postdoctoral researcher
> Donders Centre for Cognitive Neuroimaging
> http://www.doellerlab.com
>
>
>
--
Rolf A Heckemann, MD PhD
Médecin chercheur
Fondation Neurodis
CERMEP - Imagerie du Vivant
Hôpital Neurologique Pierre Wertheimer
59 Boulevard Pinel
69003 Lyon
France
|