Hi, yes if you feed in a binary 0/1 integer-format mask and use the
nearest-neighbour option then you expect to lose a lot of the voxels
due to the downsampling. This is why the default is trilinear.
On 8 Nov 2007, at 03:46, bs Jeong wrote:
> Dear FSL users,
>
> I get ROI of 58voxels as tbss result.
> After performing tbss_deproject with -n option, I can not find any ROI
> region on original and FAi space in some, not all, subjects who can
> be shown
> their ROIs on all_FA space. When trilinear option in stead of nearest
> neighbor was applied, all subjects had ROI regions on original spaced.
>
> The resolution of original space is 1.7*1.7*5mm.
> ROIs of 58 voxels transformed 3~5 voxels on original space which can
> be
> shown after tbss_deproject.
>
> 1. What is the best way for probabilistic tractograpy in subject who
> have no
> ROI after tbss_deproject with -n option?
Use trilinear, and re-binarise the final image using fslmaths. (using -
thr and -bin options).
>
>
> 2. Could the value of ROI affect the fdt result? Do I have to make
> ROI to
> binary mask before fdt?
As above.
>
>
> 3. How to compare the fdt results from between 3 voxels (subject 1)
> and 5
> voxels (subject 2)? Can I divide fdt results by 3 or 5, respectively?
Sorry - not sure exactly what you mean on this question.
Cheersy.
>
>
> Best,
>
> bsJeong
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
|