External Email - Use Caution
Dear FreeSurfer team!
First of all, thanks for the great tools! I have recently started to work with DWI data and there are a few issues I was hoping you could help me with.
I have recently found a post on the FSL mailing list mentioning that the bvecs stored in the dicoms header on Siemens scanner are corrected for slice angulation (https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;5757c231.1005). I assume this means that, for Siemens dicoms, the bvecs are stored in image orientation? Indeed, the bvecs in the dicoms are different than the ones in the table provided by Siemens. When looking at the preprocessing performed by tracula, the flag --bvec-voxel is used when importing the data from dicoms; isn't this be conflicting with the fact that the bvecs are already in image orientation?
A second issue I have is when using the bvecs extracted using tracula to perform tractography in trackvis/diffusion toolkit. For some stange reason, I have to flip the z-axis of the bvec (obtained with either --bvec-voxel or --bvec-scanner) to obtain sensible tractography results. When comparing the bvecs from tracula to the table from Siemens the z-axis is indeed flipped. Unfortunately, this flip issue also occurs in a number of other conversion software (e.g. dcm2niix, mrtrix). Does anyone know where this could be coming from? Note, my data was acquired on a Siemens SkyraFit with software version VE11C.
Finally, does anyone know what orientation (image/scanner) is expected for the bvecs when using dti_tracker (in diffusion toolkit)? When looking at the default bvec file passed through the GUI interface to dti_tracker, it seems that it requires bvecs in scanner orientation, but I would like to confirm this. Unfortunately, I did not receive any reply to my emails to support@trackvis.org. Perhaps someone on this list knows?
Thanks in advance! Vincent.