3dTrackID: tracks vs pairwise connections

AFNI Message Board
1

Hello,

I’m using 3dTrackID -mode MINIP to be able to visualize the tracts running through my ROIs. What does it mean that there are 0 tracks in net[0], but pairwise connections in netw[0]=1? If there are pairwise connections shouldn’t there be tracts to make those connections? What is the distinction between those values?

Thanks,

Carissa

2

Hi, Carissa-

That is, indeed, a bit of a puzzler.

Would you mind please posting the following:

  • the output of “afni -ver” in the terminal,
  • the full 3dTrackID command run,
  • and the text dumped into the terminal when running that 3dTrackID command?

Thanks,
pt

3

I’m not entirely sure why, but it did have to do with the AFNI version I was using. I was using the AFNI_2011_12_21_1014 version. However switching to the AFNI_17.2.09 version works just fine.

Perhaps other people will run into this problem so I’ll still give you the outputs.

The command I’m using is: 3dTrackID -mode MINIP -dti_in dt -netrois RH_Masks.nii -prefix RH_miniP5_uncut -uncert dt_UNC.nii.gz -nifti -mini_num 5 -logic AND -uncut_at_rois

Below is what is output on the terminal:

++ROI logic type is: AND
++Tracking mode: MINIP

++Number of ROIs in netw[0] = 2
++Have labeltable for naming things.
++ Running with 5 mini-prob iteraions
++
++ SEARCHING for files with prefix ‘dt*’
++ Obtained 14 prefix-matching files to sort

++SCALAR FINDINGS:
‘FA’ ‘L1’ ‘L2’ ‘L3’ ‘MD’ ‘RD’ ‘cwts’
++Done with scalar search, found: 8 parameters (well, including internal RD calc) → so will have 21 output data matrices.
++ Calculating RD

++ VECTOR FINDINGS:
‘V1’ ‘V2’ ‘V3’
++Minor note: there is a tiny (magn < 0.1) vector in the 0-th direction set. → Will exclude that voxel from walkable mask-- recommend checking model fit.
++Tracking progress count: start …
[ 1/5 ] → 0.65 min
[ 2/5 ] → 1.17 min
[ 3/5 ] → 1.62 min
[ 4/5 ] → 1.98 min
[ 5/5 ] → 2.30 min
++ Done tracking, tidying up outputs…
++ From tracking, net[0] has 0 tracks.
**ERROR: Failed to write the network.
++ Writing output (RPI, same as your input): RH_miniP5_uncut …
++Number of pairwise connections in netw[0] = 1

4

Hi, Carissa-

<< I’m not entirely sure why, but it did have to do with the AFNI version I was using. I was using the AFNI_2011_12_21_1014 version. However switching to the AFNI_17.2.09 version works just fine. >>
Cool, glad that was so easily sortable! The vagueries of the (much) earlier version will be relegated to the past…

Thanks for posting the prior output, though, for the benefits of other users who might have (very) old AFNIs-- hopefully this will help them to see the light for updating!

One comment on your 3dTrackID command that I might make-- I notice that you are using the “-uncut_at_rois” option, which means that the ends of tracts sticking out beyond the target ROIs are still kept as part of the “connection”. In most cases, I would think that would not be general behavior. To my mind, talking about the connections “between” targets would mean that hte tracked region should be restricted to: between+including the targets; between targets + target surface; or strictly between the targets. These behavioral options are described in the help docs here, with clarifying pictures:
https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/FATCAT/Tracking.html#using-target-surfaces-to-control-tract-trimming

There may very well be cases where “-uncut_at_rois” might be useful/preferable, but I just thought I would mention this point.

–pt