Hi Dorit, I'm suspecting that the initial registration is off. Can you upload the data to somewhere I can get at it?
doug
On 12/18/2017 03:08 PM, Dorit Kliemann wrote:
Hi,
I’m having some troubles with using bbregister on hemispherectomy brains (i.e only one hemisphere; data is successfully processed with recon-all dev version - thanks to Bruce).
I use bbregister to create a register.dat for the functional (preprocessed with fmriprep and registered -but not resampled- to standard space with fmriprep) to structural. Goal is to use the aparc+aseg with mri_label2vol to transform to functional volume space. Ultimate goal is then to extract time series with mrisegstats. And this seems to work fine for 3 out of 5 subjects with hemispherectomy.
However for two subjects the bbregister steps produced a bad output (as in the input tilts vertically, see screenshot overlaid on T1). One reason might be that for those subjects two heads are in a bid of a ‘funky tilted’ position, which maybe causes the automatic registration in bbregister to fail? Regardless of the reason, I was just wondering if any of you (Doug?) have further suggestions?
One issue I noticed in the log is its suggested to use the T1 flag (not the bold) but its a bold input, just weird intensities probably bc of the ‘empty’ hemisphere. I also tried to run it with the T1 flag..same issue
Thank you, Dorit
Here is the command I use for bbregister bbregister --s subjx --mov func.nii.gz --init-fsl --reg subjx_register.dat --bold --lh-only
Also happy to upload the relevant data or paste the whole log. I use Fs 6.0 to run bbregister,mri_label2vol and mri_segstats.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer