Dear Maria,
> I looked again at my eddy output. It turns out that the default output for eddy is quite clean of motion artifacts. However, I was initially looking at the <name>.eddy_outlier_free_data and this still contains movement. Am I interpreting wrongly this output? Is it that this just has "corrected" slices for the ones that are considered outliers but does not have actually eddy and motion correction, requiring one to run eddy on it once again?
yes, sorry that is a little bit confusing. I haven’t advertised the --repol option as I was still a little uncertain about it at the time of the last release. In general if you have problems with outliers you would do better to wait for the next eddy release, which I hope to make very soon. We are just waiting to hear back on a manuscript about the outlier rejection.
Having said that, the “main” output of eddy is that which is named “out”.nii.gz, where “out” is whatever you gave for the --out argument. The eddy_outlier_free output is there in case someone would like to use eddy for detecting and replacing outliers, but something else for correcting movement/distortions. One of these ideas that seemed good at the time, but which seems to have created mainly confusion. I think it might disappear for the next release.
>
> Furthermore, I would like to now plot the translation, rotation and mean displacement parameters based on the <name>.eddy_parameters file. I know that the first 6 columns represent the translation and rotation. How could one plot also the information regarding the eddy correction?
The script below should give you the plots you want. Not that it is quite Q&D and you should for example not run multiple instances of it at the same time.
Jesper
#!/bin/sh
#
# Adapted from ec_plot to produce similar plots for eddy.
# Pass it exactly the same string as you used for the --out
# argument when running eddy.
#
# Jesper
if [ $# -ne 1 ] ; then
echo "Usage: `basename $0` <eddy output basename>"
exit 1;
fi
parfile=${1}.eddy_parameters
if [ ! -f $parfile ]; then
echo "$parfile not found"
exit 1
fi
sed 's/[\t ][\t ]*/ /g' < $parfile | cut -d' ' -f 1-3 > eddy_trans.txt
sed 's/[\t ][\t ]*/ /g' < $parfile | cut -d' ' -f 4-6 > eddy_rot.txt
echo "x" > grot_labels.txt
echo "y" >> grot_labels.txt
echo "z" >> grot_labels.txt
$FSLDIR/bin/fsl_tsplot -i eddy_rot.txt -t 'Eddy Current estimated rotations (radians)' -l grot_labels.txt -o eddy_rot.png
$FSLDIR/bin/fsl_tsplot -i eddy_trans.txt -t 'Eddy Current estimated translations (mm)' -l grot_labels.txt -o eddy_trans.png
rmsfile=${1}.eddy_movement_rms
if [ -f $rmsfile ]; then
echo "absolute" > grot_labels.txt
echo "relative" >> grot_labels.txt
$FSLDIR/bin/fsl_tsplot -i $rmsfile -t 'Eddy Current estimated mean displacement (mm)' -l grot_labels.txt -o eddy_disp.png
fi
# clean up temp files
/bin/rm grot_labels.txt eddy_rot.txt eddy_trans.txt
|