Hi Mark,
thank you. Just to be clear, by "subtracting a proportion" you mean
something like subtracting a 70% (or so) version of the betted brain? As in:
fslmaths t1_brain.nii.gz -mul 0.7 t1_brain_70
fslmaths t1.nii.gz -sub t1_brain_70.nii.gz t1_4skull_estimation
thank you
martin
On 9/23/2011 5:05 AM, Mark Jenkinson wrote:
> Dear Martin,
>
> What you are suggesting all sounds sensible.
> The one thing that might prove tricky is the skull estimation.
> You need to avoid there being an "edge" of intensity change
> right at the brain edge, or otherwise it won't find the skull and
> will be normalising to the brain size, which is not what you
> want for determining atrophy.
>
> So try just subtracting a proportion of the brain from the
> original image, just to dark this boundary but leave the skull/scalp
> boundary as a stronger one. Try that and run BET on it and see
> if you are getting a decent result for the skull/scalp points. Note
> that they are never perfect but as long as over half are on the
> correct surface then things should work OK. The best way of
> telling whether it is working well is to look at the registration outputs
> and check that the skull/scalp is well aligned. If not, then it will
> still require some fixing.
>
> I hope this helps.
> All the best,
> Mark
>
>
>
> On 22 Sep 2011, at 22:28, Martin M Monti wrote:
>
>> Hi, I am working with sienax to get some volumetric info on a group of patients, but the brains are so pathological that automatic bet and registration are really thrown off. Indeed, for many of these patients I had to manually go in and delete bits of skull/eyeballs/neck that I just can't seem to remove with bet.
>>
>> Now, I do have (manually adjusted) brain extracted images for each of these patients, so I am trying to see if I can use them to work-aroud the sienax script. This would require 2 things:
>>
>> 1) generate a skull image (which unfortunately I do not have..). Is there a way I could generate them from the bet-ted brain (e.g. if I subtracted the bet-ted brain from the original image, would that be usable for the purposes of sienax as a *_skull.nii.gz image)?
>> 2) Once I have _brain and _skull images I should be able to comment out the bet command in the sienax script and run it from there?
>>
>> Does it sound possible? Is there a better/different way to work with these "unbettable" brains?
>>
>> cheers
>>
>> Martin
>>
>>
>>
>> --
>> Martin M Monti, PhD
>> Assistant Professor
>> 7461E Franz Hall
>> UCLA Department of Psychology
>> BOX 951563,
>> Los Angeles, CA 90095-1563
>> 310-825-8546
>> http://montilab.psych.ucla.edu
>> http://faculty.psych.ucla.edu/directory/faculty.php?id=187&area=3<http://faculty.psych.ucla.edu/directory/faculty.php?id=187&area=3%20>
>>
>> “Well! I’ve often seen a cat without a grin,” thought Alice,
>> “but a grin without a cat! It’s the most curious thing I ever
>> saw in my life!” [Lewis Carroll, Alice in Wonderland]
>>
--
Martin M Monti, PhD
Assistant Professor
7461E Franz Hall
UCLA Department of Psychology
BOX 951563,
Los Angeles, CA 90095-1563
310-825-8546
http://montilab.psych.ucla.edu
http://faculty.psych.ucla.edu/directory/faculty.php?id=187&area=3
<http://faculty.psych.ucla.edu/directory/faculty.php?id=187&area=3%20>
“Well! I’ve often seen a cat without a grin,” thought Alice,
“but a grin without a cat! It’s the most curious thing I ever
saw in my life!” [Lewis Carroll, Alice in Wonderland]
|