[Freesurfer] Cortical labels for WM seeding

Hello FS Gurus,

I would like to export FS labels as seed regions for deterministic tractography in TrackVis (probabilistic is not appropriate in this investigation). I run mri_label2vol with “--proj abs -4 0 .1” to try and grow the label into the WM and exclude GM. While I get a biologically plausible ROIs and plenty of good streamlines, I do get some stray streamlines that jump across the sulci into adjacent gyri which (I have read) is an artefact of seeding from GM. Have I got the right approach? Should I create an inverse intracranial volume mask as an exclusion mask? Something else?

Kind regards,

Sean Hatton Brain & Mind Research Institute University of Sydney, Australia.