Hi,
You are right - there is no tool available currently to correct
for distortion without having acquired a fieldmap. Unfortunately
this means that you either try to get a good brainstem registration
at the expense of other areas, or live with the imperfect brainstem
registration.
If the purpose of the registration is to define seeds then I suggest
that you either manually define your seed masks in this area or
transfer a standard space mask to the area using fnirt and invwarp.
If you do this, make sure you crop the input images (with fslroi) to get
rid of all the space outside of the head, as this will speed up the
process considerably. You can try the registration either with the
b=0 images or with the FA image.
All the best,
Mark
On 3 Nov 2009, at 19:19, Anastasia A Ford wrote:
> Mark,
>
> Thank you for taking a look at the images. No distortion correction
> (besides
> the eddy currents) was applied to these images. I was under the
> impression
> that distortion algorithms need a fieldmap, which we did not acquire
> as part
> of this data acquisition. Are there any distortion corrections
> available
> within FSL that do not require a field map?
>
> We are interested in good alignment within the brainstem, as we are
> using it
> as the seed for tracing CST fibers. However, if good alignment of the
> brainstem will be at the cost of mis-registration of other regions,
> as the
> cortex for example, I agree that it will not be very beneficial in
> this case.
>
> Is it possible to use the b-zero image to perform any type of
> distortion
> correction?
>
> Sincerely,
>
> Anastasia
>
|