JISCMail - FSL Archives

Hi,

we have been using Feat to unwarp our DWIs by fieldmaps with good success. Usually, we perform eddy_correct first and then unwarp the DWIs.

Now, here is the question, shouldn't the registration between the fieldmap magnitude and the DWIs example_func be based on the B0 image, i.e. the one without diffusion weighting? If so, this is usually the first image in the 4D DWI – so how would we tell the Feat gui to use this for registration? Or is that not usually important?

Second question: Why does http://www.fmrib.ox.ac.uk/fsl/feat5/detail.html state:

"The sign of this direction will depend on both the sign of the phase encode blips in the EPI sequence and on the sign of the fieldmap"

How does the phase encoding of the fieldmaps come in here (in case of a gradient echo fm)?

Can't answer the first question, I have no experience of that. For the second question, it isn't referring to the direction of phase-encoding in the field maps, but the sign of the phase values. There are two reasons this can vary. The sign convention used for phase on the scanner, and the order you do the subtraction, TE2-TE1 or TE1-TE2.

If you get the sign incorrect for the field map, you can either multiply it by –1, or use the --unwarpdir option in fugue to change the direction of the correction. I haven't used the GUI, presumably there's a similar option. If it's not correct it should be pretty obvious, since your images will end up more distorted!