ACT tractography and fiber endpoint location

Hi everyone,

i was wondering where the fibers in ACT tractography should have their endpoint. Is the last fiber point supposed to be inside of the 5ttgen GM compartment or directly in front of it in the white matter?
I just filtered an ACT tractogram with as GM mask (5ttgen volume 0 thresholded at 0.5), by discarding all fibers not starting and ending inside of this mask. Many fibers are dsicarded by this filtering, so I suppose ACT forces the fibers to end BEFORE the gray matter?

Thanks for your help!


That wouldn’t exactly be the same as what ACT does. In your case, you’re facing the discretisation of the voxel grid, whereas ACT (if I’m not mistaken) performs interpolation of the 5TT image at each point of the streamline (and only after that interpolation it does the check; thresholding after interpolation is not the same as the other way around). So it’s possible both ways that an end point may then be in a voxel that is inside or right outside the (discrete) thresholded mask you defined. @rsmith can without a doubt elaborate on this. :wink:

By default, if the ACT-check succeeds (i.e. point in GM), it’ll just stop there. If you need it to be more accurately on the boundary that ACT implicitly uses, there’s also the -crop_at_gmwmi option. But as the name suggests, that would make the streamline strictly shorter (I think), by cutting short the last segment so the endpoint is on the boundary, and not “beyond”. Again, @rsmith can elaborate or correct me if I’m wrong here somewhere. :slightly_smiling_face:

Thanks Thijs, sounds plausible :slight_smile:
I will check if this causes my issues.