Hi all,
I’ve been using fMRIPrep for a few months now, and usually it works fine without any problems. However, when I’ve tried running it on subjects who have a large amount of signal dropout in the orbitofrontal areas, the coregistration is either rotated more than it should be (using the --force-no-bbr option) or there is a wide gap between the anatomical and functional images (using the --force-bbr option). For a subject with five runs, for example, I sometimes see this error with only one or two of the runs; the other ones look fine, even though all of the functional images appear to be in the same space and same orientation. For this dataset there are no field maps, so we haven’t been able to do any unwarping to recover the signal in the OFC.
Here’s an example of a rotated functional image, generated with the following code (Note: the $HOME working directory for the fmriprep files was removed before I ran the script each time):
fmriprep-docker $bids_root_dir $bids_root_dir/derivatives
participant
–participant-label $subj
–skip-bids-validation
–md-only-boilerplate
–fs-license-file $HOME/Desktop/Test/derivatives/license.txt
–force-no-bbr
–use-syn-sdc
–output-spaces MNI152NLin2009cAsym:res-2
–nthreads $nthreads
–stop-on-first-crash
–mem_mb $mem_mb
-w $HOME
And here is an example of the coregistration when using --force-bbr:
fmriprep-docker $bids_root_dir $bids_root_dir/derivatives
participant
–participant-label $subj
–skip-bids-validation
–md-only-boilerplate
–fs-license-file $HOME/Desktop/Test/derivatives/license.txt
–force-bbr
–use-syn-sdc
–output-spaces MNI152NLin2009cAsym:res-2
–nthreads $nthreads
–stop-on-first-crash
–mem_mb $mem_mb
-w $HOME
The same issues are observed even when not using the “–use-syn-sdc” option. I’ve also tried the option “–bold2t1w-dof 12”, but the results do not change.
I did, however, find that the final coregistration was nearly perfect when I chose to run the script without the “–fs-no-reconall” option:
This seems to fix my issue, but I’m still wondering why the volumetric coregistration failed, even after trying the remedies that other people have posted in the forum; when I try running coregistration outside of fmriprep with AFNI’s align_epi_anat.py, for example, I don’t run into this issue. If I don’t want to use surface-driven coregistration, what other methods are there for troubleshooting this?
Thank you!
-Andy
[Edit: I should also mention that I am using fMRIPrep version 20.1.1]