I have a program that we use for fieldmap correction with AFNI. It started out as a wrapper for FSLs prelude and fugue. Over the past couple of years I have added a couple of minor changes in the calculation and the ability to reslice the fieldmap data collected in one plane to correct EPIs in another plane. This is useful if DTI and fMRI data are collected in the same session. (We usually collect DTI images in the axial plane and functional data in the sagittal plane, and fieldmaps collected in the sagittal plane with the frequency encode axis in the A-P direction appear to work best).
The program is written in python, reads the fieldmap from a dicom directory containing two gradient echo volumes collected at two TEs, and reads the EPI data from either a dicom directory, AFNI brik, or nifti file. It assumes that the coordinate transformation is stored in the qform fields of the nifti header.
I have been debating whether or not to distribute it on sourceforge. The state the code is such that I wouldn't want to refactor it before distributing it. This would be a Sunday afternoon endeavor but then what the heck, the packers aren't looking so good this year. If there is appreciable interest I would go ahead and do it.
I also disagree with the statement that the difference it makes is minor. For many individuals it IS minor, especially for women since their sinuses are smaller. For a fraction of subjects however, the correction is significant and probably essential if you are looking at frontal areas.
John