Hi Gerhard
I don't think anything has gone wrong, except perhaps your
expectation. If you take a look at HarvardOxford-sub-maxprob-thr25 in
fslview, the issue will become clearer. I'm picking a voxel that's
outside the mask you've generated, but has a putamen probability
higher than 25%, e.g. 61,135,66. The probability of finding putamen
here is 42%, but the probability of this voxel being cerebral white
matter here is even higher. So, since this is a maximum probability
map, the value assigned is 58, denoting white matter, rather than 51,
for which you are thresholding.
Another way to look at it is to say that your mask contains all voxels
where the probability of finding putamen is at least 25% AND where no
other structure has a higher probability of occurring.
Hope that helps
Rolf
On Thu, Jan 07, 2010 at 09:08:08AM +0000, Gerhard Jocham wrote:
> Hi all,
>
> I've been trying to construct a ROI of putamen that includes all voxels that
> have a 25% probability of belonging to the putamen according to the Harvard-
> Oxford subcortical atlas
>
> I've used the following command to get this ROI:
>
> fslmaths /fsl/data/atlases/HarvardOxford/HarvardOxford-sub-maxprob-thr25-
> 1mm.nii.gz -uthr 51 -thr 51-bin 25_prob_r_putamen
>
> This works fine, however, when overlaying the generated mask on the 1mm
> MNI Template brain in fslview and having the atlas switched on, I noticed that
> some voxels just 1mm outside of the mask (i.e. where the mask is zero) had a
> probability of belonging to the right putamen that was substantially higher
> than 25%, I've seen values up to 47%.
>
> I would have expected that everyting outside the mask should have values for
> right putamen that are below 25%. Does anyone have an idea what could
> have gone wrong here?
>
> Thank you very much,
>
> Gerhard
>
>
>
>
--
Rolf A Heckemann, MD PhD
Médecin chercheur
Fondation Neurodis
CERMEP - Imagerie du Vivant
Hôpital Neurologique Pierre Wertheimer
59 Boulevard Pinel
69003 Lyon
France
1262861686
|