Hi,
SIENA only calculates a global atrophy measure so isn't appropriate but
your approach seems fine to me, although it would be nicer to do
things in
the forward direction and then use the inverse warp. Note that you can
substantially speed up the inverse warp by initially cropping your
structural
image to exclude as much empty space as possible around the head (in all
directions). You need to do this before running FLIRT and FNIRT. In
fact,
if you do it before BET as well then it is likely that BET will work
more robustly
in general. However, if your registration of the standard to the
structural looks
like it is working well then there is no problem with your approach
either, we
just haven't tested/optimised FNIRT as much in this mode.
All the best,
Mark
On 30 Jun 2009, at 10:50, roberto toro wrote:
> Hello again,
>
> concerning my question yesterday about how to measure the volume of
> the grey/white matter in the frontal/parietal/occipital/temporal
> lobes; is SIENA able to do this kind of computation?
>
> thank you in advance for your help!
>
> roberto
>
> On Mon, Jun 29, 2009 at 2:10 PM, roberto toro <[log in to unmask]> wrote:
> Hello!
>
> I'm measuring the total volume of grey and white matter in the
> frontal, parietal, occipital and temporal lobes in a bunch of
> subjects (T1W). But I am not sure that my method is the most
> appropriate...
> Here's what I do:
>
> 1. BET the subject
> 2. FLIRT MNI152_T1_2mm_brain to subject_bet and save the
> transformation to affine.mat
> 3. FNIRT MNI152_T1_2mm (with skull) to subject (with skull) and save
> the warp
> 4. I have an atlas in MNI152 space with labels for the four lobe
> (left and right hemispheres), diencephalon, pons and cerebellum. I
> ApplyWarp the warp from step 3 to this atlas
> 5. FAST subject_bet
> 6. I wrote a command to count the number of grey and white matter
> voxels in each of the lobes (using the FASTed subject and the warped
> atlas).
>
> In particular, I'm a bit uncomfortable transforming the atlas into
> the subject instead of transforming the subject into the atlas, but
> inverting the warp was taking too much time.
>
> Thank you for your help!
>
> cheers,
> roberto
>
|