Hi,
> Lucy L. Brown wrote:
> If you don't have a B0 map, can you use some sort of template for
> correct for distortion? Does FUGUE provide a method for undistorting
> without the B0 map, too?
There is no method for undistorting without a fieldmap at the moment.
We are working on a method that will calculate the fieldmap from a
structural scan, but that is a little while off.
> Erik-Jan Vlieger wrote:
> I am unaware of such an option. If you want to unwarp some of the
distortion
> perhaps using AIR5 with lots of degrees of freedom may help?
Ideally you want a method that only undistorts/warps in the phase-encode
direction.
However, one problem common to both methods is that you cannot recover
the true intensity which is lost due to through-slice dephasing
(drop-out). In these regions the values that you are comparing
shouldn't really be trusted. On the plus side, this is really only
a problem near the air-filled sinuses in the head (lower slices in
the brain) so most of the brain should be OK, and hence undistorting
is only important insofar as it affects the global registration.
An alternative way to get good global registrations without a field map
is just to make a cost-function weighting image that has zero weights
in the suspected regions of distortion. This will then cause these
areas to be ignored in the registration, preventing them biassing
good fits in the other areas.
All the best,
Mark
|