We use epi_reg for registration of the example_func (a single EPI from the timeseries that is used as the reference in motion correction) with the structural image and for determining how to perform distortion correction. You are right that the application of the distortion correction is then done after motion correction. Note that this ignores the second-order effect of how the distortion varies with motion, since this is much smaller than the main effect of the static distortion.
All the best,
On 17 Sep 2014, at 19:35, Ajay Kurani <[log in to unmask]> wrote:
> Hi Mark,
> Just as a follow up do you recommend running epi_reg (or fugue) after you perform all of the standard EPI preprecessing (ie despike, slice time correction, motion correction) or just run motion correction first (so that all timepoints are aligned), apply the field map correction and then continue with preprocessing. I would assume that you would need to at minimum run the motion correction in order to ensure the the field map correction is applied to the appropriate voxels and that none are shifted due to head motion.