Dear Antonis,
I would use 6 dof for the first flirt call.
As for the second, if it is failing then it might be due to the quality of the segmentation. How good is your segmentation?
And have to restricted it to white matter only?
The default for bbrtype and bbrslope will be appropriate for using either gradient-echo or spin-echo EPIs, so you shouldn't need to change these.
An alternative you might like to try is the epi_reg script, although it does make very similar calls. Still, you might find it useful to look through.
All the best,
Mark
On 11 Jan 2014, at 17:13, Antonis M <[log in to unmask]> wrote:
> Dear all,
>
> I am trying to register the b0 map to the T1 (or T2) image of a subject with BBR but I get weird results (out of plane).
> I am using the following commands, is it correct?
>
> flirt -ref $subj-T1.nii.gz -in $subj-b0.nii.gz -omat $subj-init.mat
> flirt -ref $subj-T1.nii.gz -in $subj-b0.nii.gz -omat $subj-bbr.mat -out $subj-bbr.nii.gz -init $subj-init.mat -wmseg $subj-wm.nii.gz -cost bbr
>
> Also could you provide me with some information on the usage of bbrtype and bbrslope?
> I suppose that according to "Accurate and robust brain image alignment using boundary-based registration" the bbrslope is set positive for gray matter brighter than white and the inverse. What would be the values that I should use according to your experience?
>
> Thank you very much,
> Antonis
|