Hi Peggy - If it looks like a single path, it definitely is an initialization error. If you reinitialize repeatedly and it doesn't get fixed, it might be that something went wrong earlier in the processing (for example some part of brain that the tract goes through is missing from the brain mask).
Best, a.y
On Mon, 25 Apr 2016, Peggy Skelly wrote:
Hi Tracula Experts, How do you determine if a pathway reconstructed well? I've been examining the merged pathways (<subj>/dpath/merged_avg33_mni_bbr.mgz) in freeview, and if the pathway looked too small when viewing with the default threshold, I would reinitialize those pathways. Most of the time, the new pathway would look ok upon re-inspection. A few pathways, however, still appear too small when viewed at the default threshold in Freeview, but at lower thresholds, they begin to look ok.
I've encountered 1 pathway that still appears too small in Freeview, but the output in pathstats.overall.txt and pathstats.byvoxel.txt seem reasonable. On the other hand, its path distribution (path.pd.nii.gz) does not look like a distribution, but a single path -- there are a 38 voxels with values of 4000, and all others are 0. This doesn't seem right. What could cause this to happen?
What objective measure can I use to determine if a pathway is reconstructed well enough?
Thanks, Peggy