So I would do it like this: make a binary mask of your tract region of
interest. Detect its extent in the z axis. You can do this by making a
temporary folder and using fslsplit into slices and testing each in order to
see if it is non-zero (you can use fslstats for this). Once you have the
slices, you can use fslmaths -roi to make individual files that have full 3D
dimensions but contain the only slice of interest in another temporary
folder. Then loop through these slices with fslstats using them to mask the
FA (-k option) and then take the mean value at each level). This is how you
could do it with shell script and loops. If you are good with matlab, you
might be able to do this much more simply.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Martin Kavec
Sent: Wednesday, January 27, 2010 1:34 PM
To: [log in to unmask]
Subject: Re: [FSL] FA along a track
Hi Paul,
On Wednesday 27 January 2010 15:56:43 Paul Robinson wrote:
> Hi, Martin.
>
> Are you looking for something like this:
>
> fslstats subj_FA.nii.gz -k cortico-spinal.nii.gz -M
Not really, because this would give you a mean FA of the corticospinal
track,
which may vary substantially with anatomical location, which is what I am
interested in. Please read on the thread, Matt is pretty close to what I
want.
Thanks,
Martin
|