External Email - Use Caution

Hi,

Upon further examination of pathstats.byvoxel.txt and pathstats.overall.txt, I have 2 more questions. Since my previous 3 questions still have not been answered, I'm including them here so that they are all in one place. The 2 new questions are 4) and 5), my previous 3 questions are appended after.

4) I have noticed that the averages of AD, RD, MD and FA in pathstats.byvoxel.txt are exactly the same as the *_Avg_Center in pathstats.overall.txt. According to the tutorial, *_Avg_Center in pathstats.overall.txt is the "average over highest-probability path only" and the AD, RD, MD and FA in pathstats.byvoxel.txt is the value at the "position on the maximum a posteriori path". Does it mean that "highest-probability path only" and "position on the maximum a posteriori path" are the same or one of them should be changed to the other one?

5) Are the averages of AD_Avg, RD_Avg, MD_Avg and FA_Avg in pathstats.byvoxel.txt suppose to match *_Avg(average over the entire support of the path distribution) in pathstats.overall.txt? When I removed the duplicate points in pathstats.byvoxel.txt, the averages of AD_Avg, RD_Avg, MD_Avg and FA_Avg come very close to the value of *_Avg in pathstats.overall.txt

I don't think I can calculate the averages by including the duplicate points in 4) but removing them in 5). Either including them in both or not at all.

Questions from my previous email.

1) There are duplicate points in pathstats.byvoxel.txt, is it normal? I didn't try it on all 42 tractcs. For the few I've looked at, they all have duplicate points. I used the default setting in the config file to run trac-all.

2) When I do along-tract statistical analysis, sig.mgh and z.mgh are 1D files. When embed the sig.mgh and z.mgh statistical maps from the analysis into the template .trk (*.display.trk, which is 3D), it seems to me that sig.mgh and z.mgh don't have enough information to give *.display.trk the voxel values for the *.overlay.trk. There is no one-to-one voxel corresponding between sig.mgh and *.display.trk. What am I missing? Or what's the embedding process do?

3) When I convert .trk(*.overlay.trk, for example) to nii.gz or ascii using dmri_trk2trk or mri_convert the voxel values of the .trk file don't get converted into the .nii.gz or ascii file, only the x,y,z coordinates are converted. Is there a way to get the voxels values of the .trk file when converting to .nii.gz or ascii format?

Thanks,

King-Wai