Currently, the protocol does two different things for sub-cortical GM structures:
5ttgen hsvs: Will by default run FSL FIRST
Connectome construction: Given labelsgmfix is not run, the FreeSurfer volumetric segmentations will be used.
The tck2connectome 4mm radial search might be adequate to correct any assignment errors arising from this. But could think about whether we want to use the same method for both.
(Better again would be to run FSL FIRST once, and then provide that as input to both of these commands; but that would require upstream MRtrix3 changes)
Currently, the protocol does two different things for sub-cortical GM structures:
5ttgen hsvs: Will by default run FSL FIRSTlabelsgmfixis not run, the FreeSurfer volumetric segmentations will be used.The
tck2connectome4mm radial search might be adequate to correct any assignment errors arising from this. But could think about whether we want to use the same method for both.(Better again would be to run FSL FIRST once, and then provide that as input to both of these commands; but that would require upstream MRtrix3 changes)