I actually have a question about bedpostx and the second fiber fraction.
We formulated the hypothesis that patients would have a greater number
of crossing fibers in a certain tract. We tested this by finding how
many voxels in a specified region of interest (the mask for a tract) had
a value above 0.05 or 0.1 in mean_f2_samples. We found that while our
controls had a bigger tract (indicating more fanning out of the fibers)
there was no difference in the proportion of voxels that contained
crossing fibers. I found this not to be very satisfactory, because the
larger tract has to mean that the underlying directionality of the
fibers is different for the groups since we are starting the
tractography from an roi that is quite small and whose size is larger in
the patient group.
Is there a better way of doing this?
Another related question regards the threshold for crossing fibers. Is
there a proportionality between the threshold we use for identifying
voxels in mean_f2_samples and the angle that fibers cross at? i.e. does
a voxel that has a value of 0.05 in mean_f2_samples have a crossing
fiber that is likely to be more parallel to the first fiber than a voxel
where mean_f2_samples=0.1?
Is there a way to know what the angle between two fibers in a voxel is?
Stefano Marenco
-----Original Message-----
From: Tim Behrens [mailto:[log in to unmask]]
Sent: Wednesday, May 28, 2008 7:42 AM
To: [log in to unmask]
Subject: Re: [FSL] using local diffusion estimates
I don't think they have, but I don't see why you should not,
particularly if you are interested, for example, in the second fibre
fraction.
Cheers
T
On 28 May 2008, at 02:11, Scott Kolbe wrote:
> Just wondering if anyone has done any case/control comparisons on
> the local diffusion parameter estimates from BEDPOST/BEDPOSTX?
> cheers
> Scott
>
> --
> ========================
> Scott Kolbe
> Postgraduate Student
> Neuroimaging Group
> Howard Florey Institute &
> Centre for Neuroscience
> University of Melbourne
> VIC, Australia, 3010.
>
> ph: +61 3 8344 1887
> email: [log in to unmask]
> website: http://www.neuroimaging.org.au/index.php?id=383
>
|