Hi Nicolas,
> I'm currently working with DTI images and wonder wether I could apply
> the field map corrections to DTI images.
>
> More specifically:
>
> 1/ we acquire our images with spin-echo EPI: is the Fieldmap method,
> firstly designed to handle gradient-echo EPIs, able to handle this kind
> of SE-sequence?
There should in principle be no problem to apply fieldmap corrections to
DTI data. For SE-EPI data I would strongly recommend checking the
"Jacobian modulation" box.
If you use non-EPI gradient-echo images to calculate the fieldmap things
should be straightforward. If you use EPI gradient-echo images you would
have to make sure that the readout time is the same for your SE-EPI and
your GE-EPI images. This is because we assume the same readout time when
inverting the fieldmap as when applying it to the image we want to
undistort. In principle there is nothing to prevent the use of two
different times, it just hasn't been implemented.
> 2/ Does it sound really bad to apply the unwarping *without* the
> estimated movement-by-distorsion implemented in the Realign and Unwrap
> toolbox? This is because the estimated motion is strongly correlated to
> the diffusion direction we applied for each image, perhaps even after a
> Eddy-current distorsions correction (that is correction of the
> distorsions due to the diffusion gradients with a method like
> http://brainvisa.info/pdf/mangin-MIA02.pdf).
No I don't think that is a problem. In fMRI our "signal" is tiny (~1% of
baseline) and therefore even such relatively small effect as
movement-by-distortion interaction is large in comparison. For DTI our
signal is ~50% (for b-value ~1000) and I think it is OK to ignore the
movement-by-distortion interaction.
Good luck Jesper
|