Dear Anastasia,
We have deliberately excluded the ventricles from the new version
of FIRST as the performance is not comparable with the other
structures. There have always been problems with them since the
original manual segmentations did not fully extend in the posterior
direction and so required an arbitrary cutoff to be applied. I
wouldn't trust results generated using the ventricle models in
previous versions. You would be better off using FNIRT, FAST
or FreeSurfer to segment the ventricles.
All the best,
Mark
On 1 Apr 2010, at 19:50, Anastasia A Ford wrote:
> Hi,
>
> Im running run_first_all script on different groups of subjects, and
> I noticed that the right and left lateral ventricles are not
> generated during the segmentation. I run the same command a few
> months ago with a similar dataset (same acquisition parameters) and
> all of the files, including the ventricles were generated. Is there
> a reason why the ventricles are missing? The version of FSL Im using
> is 4.1.5.
>
> Sincerely,
>
> Anastasia
>
|