Print

Print


Dear Alexandre,

I only have Siemens fieldmap (phase difference and two magnitude images) in a dataset and the fieldmaps are 4 mm isotopic.

I would like to combine PRELUDE/FUGUE with eddy tool using --field option in order to minimise the number of interpolations. I have a few questions regarding how to prepare the fieldmap for eddy.

I used to run the following operations from the fsl_prepare_fieldmap script on a masked fieldmap in order to prepare the fieldmap:
1) Convert the fieldmap into radians
2) Unwrap the fieldmap with PRELUDE
3) Convert the fieldmap to rad/s (division by the echo time difference)
4) Call FUGUE to extrapolate from mask (fill holes, etc)
5) Demean to avoid gross shifting
6) Clean up edge voxels
Then, I smoothed a little bit the fieldmap (2mm 3D Gaussian) prior to registering the fieldmap onto the b0 for the preprocessing with fugue.

I have two main questions:
1) In the documentation of FSL eddy, the --field flag requires the prepared fieldmap in Hz, however I am not sure when to do that. Can I simply replace the step 3 by the conversion into Hz?

yes.


2) Regarding the regularisation step of the fieldmap, I was advised to do a 3D smoothing of 2mm but I did not find any documentation about it. Does the fieldmap still need to be smoothed for eddy? In a former thread, I read that we may need to smooth the warpfield in order to be invertible. Should I first try with no smoothing then iteratively increase the smoothing of the fieldmap if eddy does not work?

It is unlikely that eddy will work with un-smoothed fieldmaps. It is not necessarily the field itself that is the problem, but rather the sharp edges of the field when preparing it using FSL tools. This tends to lead to a discontinuity that makes it non-invertible. Smoothing it with a few mm FWHM will have relatively little effect on the field itself (it being quite smooth already), but will remove the sharp edges.

My suggestion would be

7. Register the fieldmap to the first b0 volume using flirt
8. Smooth the fieldmap with a FWHM of a few mm.

It is still not ideal, and for the future I recommend acquiring topup-style fieldmaps. They are much faster to acquire and easier to work with.

Jesper

 

Thanks in advance for your answers.

Best,
Alexandre


To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1




To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1