Hello,
I am running a newly installed version of eddy_cuda8.0 and I am using standard field maps (2 magnitude, 1 phase) so I am unable to use topup. I tried using the --field option to put my preprocessed field maps into eddy to correct for field inhomogeneities, motion etc... in one step, but got this error when I ran the command:
eddy: msg=Error when attempting to read --field file
terminate called after throwing an instance of 'EDDY::EddyException'
what(): eddy: msg=Error when attempting to read --field file
Aborted (core dumped)
I was able to run the eddy command fine without the --field option and applied my preprocessed field map to the eddy output using fugue. There was no error, but the two maps look pretty similar, although the voxel-wise values are slightly different.
Below are the preprocessing steps for my field maps (fm=magnitude, fp =phase):
# brain extract one of the magnitude images
bet ${SUBJ}_fm1.nii.gz ${SUBJ}_fm1_brain.nii.gz -f 0.3 -m
# prepare fieldmap
fsl_prepare_fieldmap SIEMENS ${SUBJ}_fmp.nii.gz ${SUBJ}_fm1_brain ${SUBJ}_fieldmap 2.46
# apply brain mask to fieldmap
fslmaths ${SUBJ}_fieldmap.nii.gz -mas ${SUBJ}_fm1_brain_mask.nii.gz ${SUBJ}_fieldmap_masked.nii.gz
### # smooth fieldmap
fugue --loadfmap=${SUBJ}_fieldmap_masked.nii.gz -s 4 --savefmap=${SUBJ}_fieldmap_masked_s4
# register magnitude image to dtidata and apply transform to fieldmap
flirt -in ${SUBJ}_fm1_brain.nii.gz -ref ../dti/${SUBJ}_dti_brain.nii.gz -out ${SUBJ}_fm1_brain_flirted -omat #fieldmap2diff.mat
flirt -in ${SUBJ}_fieldmap_masked_s4.nii.gz -ref ../dti/${SUBJ}_dti_brain.nii.gz -applyxfm -init fieldmap2diff.mat -out ${SUBJ}_fieldmap_flirted
#Output is in rad/s, change to HZ
3dcalc -a ${SUBJ}_fieldmap_flirted.nii.gz -expr 'a*0.159154943091895000' -prefix ${SUBJ}_fieldmap_flirted_HZ.nii.gz
Any help would be appreciated!
Amanda
|