Hi,
> Thanks Mark for your quick reply.
>
> So, correct me if I am wrong, the fieldmap is applied at each step of the optimisation to undistort the epi image regarding it's new tentative alignment taking into account the epi slicing direction.
That's correct.
> I noticed the _warp file output in epi_reg which is the fieldmap flirted into the epi image space through the T1 image as reference.
The warp should take the input epi image to the T1 structural space. The last line of the epi_reg script (with fieldmaps) does:
applywarp -i ${vepi} -r ${vrefhead} -o ${vout} -w ${vout}_warp --interp=spline --abs
which shows how it can be applied.
> Having also observed some inconsistent behavior using weighting, I will consider that it is partially taken into account by the signal loss causing low gradient values accounting less into the cost function.
Sounds sensible to me.
All the best,
Mark
|