Problem: SD_STREAM method for fiber tracking

In the CSD-based tractography framework,I used the SD_STREAM method for fiber tracking, but the phenomenon of unexplained termination occurred. After adjusting the angle threshold and cutoff are not significantly improved, even the fiber results are highly similar. The first and second codes are as follow:
tckgen -algorithm SD_STREAM -angle 80 -maxlen 850 -minlen 10 wm_fod.mif -seed_image seed_ROI.mif -select 40000 -cutoff 0.01 seed_ROI_rectanfle.tck

tckgen -algorithm SD_STREAM -angle 90 -maxlen 850 -minlen 10 wm_fod.mif -seed_image seed_ROI.mif -select 40000 -cutoff 0.00000001 seed_ROI_rectanfle_mincut.tck

What can I do to solve such a problem? Thank you!

Welcome @pya!

It’s difficult to diagnose precisely from your description and images. E.g. What is it about the terminations that makes them “unexplained”? Indeed from that I would suggest running tckgen with the -info option, since that will actually provide a breakdown of the different reasons for which streamlines were terminated.

For SD_STREAM, the angle and amplitude cutoffs will not change the streamlines orientations in any way; they will only influence where streamlines terminate. But if using 90 degrees and a tiny cutoff threshold and the streamlines are terminating where you do not expect them to, there must be some other criterion causing them to terminate; hopefully -info will provide insight.

Your FOD image also looks quite noisy; it may be unrelated to the specific issue that you’re trying to communicate, but it may be worth checking your processing to see that you’re not chasing a red herring.