Dear FreeSurfer users,

I've been using FreeSurfer and Tracula versions 6.0. I've run Tracula on DWI scans, and now I am seeking to run dt_recon on them as well. As a new user, it seems both programs follow similar initial pre-processing steps (eddy current & motion correction), which the documentation says can take 1-2 hours. I'd imagine that these steps not need be repeated for running dt_recon if Tracula has already been run.

I have a few questions:

Q1: I ran dt_recon per documentation and the full process finished without error within 20 minutes for several initial DWI scans I ran. This is concerning since the documentation estimates eddy current & motion correction by itself to take 1-2 hours. Is 20 minutes reasonable?

Q2: Is there a way to compare outputs from dt_recon & Tracula to ensure consistency in both processes? For example, can I compare stats in the fa.nii.gz file? I compared snapshots (attached) of wmparc (from recon-all of T1-weighted MRIscan), fa.nii.gz (from dt_recon), and dtifit_FA.nii (from Tracula) at the same slice. There are clear differences between dt_recon and Tracula that also hold at other slices.

I believe Tracula registers DWI data to subject's anatomical space from recon-all of T1-weighted scans. That explains why higher fractional anisotropy values (white) in dtifit_FA.nii lined up more nicely with white matter in wmparc.nii than that of fa.nii.gz from dt_recon. I want to use mri_segstats on the fa.nii from dt_recon (along lines of https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg37032.html). Does this require first registering fa.nii to subject's native space to create a "fa-anat.nii" file? If so, how can I do that? Do I apply an inverse transform from the talairach registration of this data (fa-tal.nii.gz) to native space via ${SUB}/mri/transforms/talairach.lta?

Q3: One post on the FreeSurfer Forum suggests that, if dt_recon is run first, then Tracula should still be run from the start because trac-all --preproc prepares data in an orientation that FSL's bedpostx is expecting. https://mail.nmr.mgh.harvard.edu/pipermail//freesurfer/2012-March/023081.htm... However, does the inverse also hold? If I've run TRACULA (all three steps), is there a way to run dt_recon in a way that can allow using the output TRACULA's initial steps?

Thank you.

Best, Chintan