Hi Susie - The initial placement of the control points plays a much bigger role than the end ROIs. The path being initialized in a bad location would be more of a problem than the end ROIs being too generous. So I'd check rh.unc_AS_avg32_mni_bbr_cpts_5.nii.gz before editing the end ROIs. Does that look like it's better placed on right than the left? I'd also check if the brain mask (aparc+aseg_mask) is any different in that area on one side.
Hope this helps, a.y
On Tue, 9 Jul 2013, Susan Kuo wrote:
Hi Anastasia, For one of my subjects, I am noticing that the right uncinate fasciculus is very diffuse and more importantly, erroneously crosses from the frontal to the temporal lobe. I looked at the original endpoint ROI generated by trac-all -prior (rh.unc_AS_avg33_mni_bbr_end1.nii.gz) and the dilated endpoint ROI in the temporal lobe (rh.unc_AS_avg33_mni_bbr_end1_dil.nii.gz) and noticed that the endpoints appeared to cross over into the frontal lobe.
Attached is a snapshot for your perusal. On the left side is a display of the <subject>/dmri/dtifit_FA.nii with the <subject>/dpath/rh.unc_AS_avg33_mni_bbr/path.pd.nii.gz tract superimposed over. As you can see, the right unc tract crosses from the frontal to the temporal lobe. On the right side is a display of the <subject>/dmri/brain_anat_mni.nii.gz with the endpoints <subject>/dlabel/mni/rh.unc_AS_avg33_mni_bbr_end2.nii.gz in green and <subject>/dlabel/mni/rh.unc_AS_avg33_mni_bbr_end1.nii.gz in blue (the red dots are control points). As you can see, the blue endpoints bleed over from the temporal to the frontal lobe.
I tried to eliminate these blue points with fslview and then writing over rh.unc_AS_avg33_mni_bbr_end1.nii.gz. I performed a similar function manually editing rh.unc_AS_avg33_mni_bbr_end1_dil.nii.gz to better reflect the boundaries . I subsequently tried to re-run trac-all -path, and unfortunately, all my tracts looked much worse.
Do you know if incorrectly placed endpoints impair the path reconstruction? Do you think you can help me? Any advice you can give would be much appreciated.
-- Susie Kuo NIH
Mediocrity knows nothing higher than itself, but talent instantly recognizes genius. - Sir Arthur Conan Doyle, Sherlock Holmes- Valley of Fear