Hi
There are some instruction on http://www.fmrib.ox.ac.uk/fsl/fdt/fdt_biggest.html
and you may also read Behrens et al, Nat Neurosci 2003 to see
different options on how to report classification results.
find_the_biggest gives you a hard segmentation based on the max
connectivity score. It comes in the form of a spatial map rather than
%voxels as you suggest.
Cheers,
Saad.
On 18 Nov 2008, at 00:57, Jenifer Juranek wrote:
> We're interested in comparing subjects using outputs from running
> find_the_biggest on seed based classification (classification
> targets=3, seed
> mask =1). Any recommendations how to go about this? All voxels in
> the seed
> mask are accounted for by the 3 classification targets. Would a
> simple % of
> #voxels per classy target wrt #voxels in seed mask be reasonable?
>
> Thanks,
> Jenifer
>
Saad Jbabdi
Oxford University FMRIB Centre
JR Hospital, Headington, OX3 9DU, UK
+44 (0) 1865 222545 (fax 717)
www.fmrib.ox.ac.uk/~saad
|