Hi,
The range represents the extreme values and could be driven here by a
few noisy voxels - which typically occur around the edge of the brain.
It isn't necessarily a big problem, and using a slightly eroded brain mask
can often fix it. However, you need to check that this is the case by looking
at the coloured image - if it is mainly spread over the colours near zero then
that's fine, but if you see a lot of colours spanning the full range within the
brain then that would be problematic. The majority of voxels should have
a shift of 2 voxels or less, and you can also load the appropriate volume in
fslview.
As for applying the shift map - I cannot tell without seeing your command
lines. Note that you can only apply *warps* with applywarp (they should
have dim4=3 if you look with fslhd) but you can convert a shiftmap to a
warp with convertwarp.
I hope this helps.
All the best,
Mark
On 7 Mar 2011, at 20:26, Tyler Blazey wrote:
> Hi,
>
> What is a normal range for the voxel shift map produced by Fugue? The shift map I am getting has a range of -18 to 30, which seems kinda high considering my voxels are 2mm isotropic. Also I have tried confirming the shift map by creating a mask on the distorted epi and applying the shift warp to the mask. When I count the number of voxels that mask was shifted by, it does not seem to match what is indicated by the shift map. I have tried using applywarp with both --interp=nn and --interp=trilinear, and I still can't get the shifts to match. Am I doing something wrong, or should I just not expect the results to match up so simply?
>
> Thanks,
>
> - Tyler
>
|