Hi Xiaofu - It's hard to tell but it looks like the brain mask (which comes from the structural data) is a bit too large in the temporal lobe, so it's possible that the structural-to-diffusion registration didn't go so well for this subject. You can check on this by overlaying the structural segmentation transformed into diffusion space (from dlabel/diff/aparc+aseg...) onto the FA map.

Best, a.y

On Sun, 1 May 2016, Xiaofu wrote:

Dear Anastasia or Tracula-group,    I have run Tracula and found there was only one single curve for some tracts. But others went well. I tried setting reinit parameter to 1 and rerunning "trac-all -prior" and "trac-all -path" but didn't get improved, e.g., the right UNC was a single curve (see attached). I also checked the two end point set of right UNC (i.e.,endpt1.pd.nii.gz and endpt2.pd.nii.gz) (see attached) and found those end points were visually not correct when loaded to FA image. I guess that may be the reason of the single curve problem. Is there anyway to fix this problem? Moreover, do you have any suggestions on the tract statistics for those single curve tracts. I'd appreciate it if could take a look at it for me. If you need more files for the diagnosis please let me know.    Looking forward to hearing from you. Thank you very much.

Best, Xiaofu