[caret-users] Brain extraction

[Hamied Haroon]

Hi Donna,

Thanks so much for your kind reply.

I'm actually particularly interested in the F6 volume. I want to
non-linearly register the F6 MRI volume to a brain-only MRI volume I
have of an ex-vivo Macaca fascicularis. I then want to apply the same
non-linear warp to the cortical partitioning schemes available in Caret
to spatially match them to my dataset. Non-linear registration of a
brain-only volume to a whole-head volume doesn't work well, so that's
why I need to extract the brain in the F6 MRI volume. I'm guessing that
it's best to register the Caret atlas associated with the same species
(M. fascicularis) as the dataset I have. Would you agree? Or would it be
valid to use the F99UA12 dataset in this case?

I look forward to hearing from you.

Many thanks.

Best wishes,
 
Hamied
 

-----Original Message-----
From: caret-users-bounces at brainvis.wustl.edu
[mailto:caret-users-bounces at brainvis.wustl.edu] On Behalf Of Donna
Dierker
Sent: 15 July 2008 17:01
To: Caret, SureFit, and SuMS software users
Subject: Re: [caret-users] Brain extraction

Hi Hamied,

If you mean generate output similar to FSL's BET, then I'd say, no -- 
not automatically.

In the case of F99UA12, the AFNI volume 
Macaque.F99UA1.BOTH.SEGMENT.clean+orig.HEAD in the September 2006 
tutorial 
(http://sumsdb.wustl.edu/sums/archivelist.do?archive_id=6595030), 
MACAQUE subdirectory, is in the ballpark.  It is a segmentation of both 
hemispheres, but it aims for the cortical midthickness, rather than the 
pial surface.  Also, the sulci are not filled in, the way a BET-style 
mask would be.  Using utilities in Caret's Window: Caret Command Script 
Builder: VOLUME DILATE and ERODE you could fill in the sulci, and then 
zero out all voxels outside the resulting mask.

In the case of F6, I could find no segmentation volume, but this dataset

contains both fiducial coord files:

Macaque.F6.BOTH.VincentEtAl_Nature07_Fig2.73730.spec
http://sumsdb.wustl.edu/sums/archivelist.do?archive_id=6621813

You could project these surfaces to volumes; fill the cavity; and get a 
segmentation not unlike the F99UA12 volume described above.  You'd need 
to similarly dilate and erode to fill the sulci.

The situation looks similar for M3:  Fiducial surfaces exist and could 
be back projected to volumes; dilated; eroded; and extra-mask voxels set

to zero.

With the exception of F99UA12, these volumes are affine means of 
multiple subjects.  While you could segment them in Caret, I wouldn't 
describe this process as automatic.  While you'd probably have a handle 
or two to patch with F99UA12, your work would be far easier with that 
volume, not only because it is a single subject, but also because the 
quality is spectacular.  But why do it when it has been done, and the 
segmentation volume is available?

How were you planning to use the extracted brains?

Donna

On 07/15/2008 09:50 AM, Hamied Haroon wrote:
>
> Hi,
>
>  
>
> Is there any way in Caret (v5.502) to automatically extract the brain 
> in the MRI volumes of the F99UA12, F6 and M3 macaque atlases, please?
>
>  
>
> Many thanks.
>
>  
>
> Best wishes,
>
>  
>
> Hamied
>
>  
>
>  
>
>
------------------------------------------------------------------------
>
> _______________________________________________
> caret-users mailing list
> caret-users at brainvis.wustl.edu
> http://brainvis.wustl.edu/mailman/listinfo/caret-users
>   

_______________________________________________
caret-users mailing list
caret-users at brainvis.wustl.edu
http://brainvis.wustl.edu/mailman/listinfo/caret-users