Hi Tabinda,
based on the number of fibres per voxel you have specified in your bedpostx call (--n option), you will get a mean_f<i>samples map containing the estimated anisotropic volume fraction of each fibre.
If i=3, you can find the voxels with three fibres doing something like:
fslmaths mean_f3samples -thr 0.05 -bin 3fib_mask
3fib_mask will be a binary mask showing those voxels where you have 3 fibres. Typically a threshold of 0.05 is fine.
Hope this helps, cheers,
Matteo
> On 30 Sep 2017, at 12:08, Tabinda Sarwar <[log in to unmask]> wrote:
>
> Dear FSL experts,
>
> For my study I am trying to calculate number of voxels having multiple
> fibers from BEDPOSTX with n=10. If I threshold on anisotropic fraction
> (using mean_f<i>samples ) at 0.01, still I could extract voxels having
> volume fraction greater than 0.01 which means that a voxel has 10
> fibers which is equivalent to the maximum number of fibers specified
> initially. This means that the results are dependent on the
> specification of number of fibers to resolve. What should be an
> appropriate threshold for extracting fibers if 0.01 is too low?
> Do you think this is the right way to calculate multiple fibers in
> voxel? Or there is something which I am doing wrong? Is there some
> other way to calculate the number of voxels having multiple fibers in
> FSL?
>
> I am using HCP data along with the application of white matter mask
>
> Thanks,
> Tabinda
|