Hi Bruce,
Thanks for the quick reply. I've just finished rerunning it with spm, fsl and header init and the latter seemed to work if I write out a _bbreg.nii file. The former two failed, however. Apparently using the header seems to work for this one subject and one run. I'm testing the rest of the runs.
This is an example from the FSL init. Please note that I'm running it through nipype, hence the jumbled subdirectory strucutre.
bbregister --t2 --epi-mask --init-fsl --fslmat /data/pt_nmc002/TDMGB_FMRI_coreg/tmp/preproc/_subject_id_sub-01/_init_fsl/bbreg/mapflow/_bbreg0/meanusub-01_task-spespk_run-01_bold_bbreg_sub-01.mat --reg /data/pt_nmc002/TDMGB_FMRI_coreg/tmp/preproc/_subject_id_sub-01/_init_fsl/bbreg/mapflow/_bbreg0/meanusub-01_task-spespk_run-01_bold_bbreg_sub-01.dat --o /data/pt_nmc002/TDMGB_FMRI_coreg/tmp/preproc/_subject_id_sub-01/_init_fsl/bbreg/mapflow/_bbreg0/meanusub-01_task-spespk_run-01_bold_bbreg.nii --mov /data/pt_nmc002/TDMGB_FMRI_coreg/tmp/preproc/_subject_id_sub-01/_init_fsl/bbreg/mapflow/_bbreg0/meanusub-01_task-spespk_run-01_bold.nii --s sub-01
The output is attached. Details are under the heading 'Terminal output'. I'm also attaching a screenshot of the T2 with a white overlay for the header init. It seems to be alright. The output from the header init is also attached. The difference in the command line is --init-header instead of --init-fsl.
Downstream, however, the registration to MNI through ANTs seems to fail, even though I've converted the transformation matrices to ITK through C3D. Hmmmm... I just had a revelation. I may have used the wrong T1 to calculate the registration matrix through ANTs. Instead of using the T1 from the ReconAll output, I used the one from the fsl.bet. I'll make sure to use the ReconAll T1 and get back to you.
I would appreciate a comment, however, on how to get better registration of partial FoV T2 images to T1 images through BBRegister. It seems that the easiest option is a full head EPI at the time of the experiment.
Regards, Glad
On 07/11/2017 03:33 PM, Bruce Fischl wrote:
Hi Glad
can you send us the full command line you are using and screen output?
thanks
Bruce On Tue, 11 Jul 2017, Paul Glad Mihai wrote:
Dear List,
I've been quite frustrated with the BBRegister results when trying to register a partial field of view EPI mean functional image to a T1 scan. I've tried different init arguments (fsl, spm, header) and none really work. I've found that the initial registration is causing the problem, since the T1 and mean EPi are in reasonable alignment, but not perfect. As soon as it passes the initial alignment it's totally off and obviously fails in the BBR step. I have no full head EPI, unfortunately. I have a full head scout which seems to be a T1 image with really poor contrast. If I try to register this to the T1 through BBRegister it works. But if I use it as an intermediate file it still fails. What can I do? How does the intermediate file come into play? Will it get registered to the T1 initially and then the resulting matrix is used further? I also tried to feed the resulting registration matrix as init-reg but that didn't work either.
Is there a way to skip the initial registration?
Can I mask the T1 image with the FoV of the functional scan if they are initially in reasonable alignment? Would that work? Would that mean that I have to re-run ReconAll with the masked T1?
Would love to hear back from someone.
Regards, Glad
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
