Dear FSL experts,
I am using fieldmap distortion correction in diffusion-weighted data through the FEAT pre-stats tab as recommended in a previous post <https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1204&L=FSL&D=0&1=FSL&9=A&I=-3&J=on&X=1958DF781C3F36D75A&Y=moana%40email.unc.edu&d=No+Match%3BMatch%3BMatches&z=4&P=221433>. My dataset was collected in a 3T Trio Siemens scanner, and I used the fsl_prepare_fieldmap tool (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FUGUE/Guide) to pre-process the fieldmap images.
I have a question regarding how to determine the unwarping direction to use. This is from the FEAT wiki page:
You also need to specify the Unwarp direction, which is the phase-encoding direction of your FMRI EPI data. 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. As it can be difficult to predict this sign when using a particular site/scanner/sequence for the first time, it is usual to try both positive and negative values in turn and see which gives better undistortion (the wrong sign will increase the amount of distortion rather than decrease it).
While it sounds clear cut how to decide the unwarping direction in a qualitative fashion, I noted that sometimes the temporal lobe look better using a particular direction, say minus Y, but the frontal lobe does not benefit as much and sometimes look a bit better in the distorted image. I am checking this via FSLview superimposing the distorted and undistorted DWI images over the structural T1 for each subject.
My questions are: 1. Is there a quantitative way to assess the best unwarping direction to be used for fieldmap distortion correction? 2. Does distortion correction affect the bvecs, and if so how do I compensate for it (I am already using the "fdt_rotate_bvecs" script following pre-processing with "eddy_correct")?
Estephan Moana-Filho, D.D.S., M.S.
Clinical Fellow, Oral & Maxillofacial Pain Program
Regional Center for Neurosensory Disorders
UNC School of Dentistry
[log in to unmask]