Orientation of along trat data after export

15 views
Skip to first unread message

Christian Gewinn (Christent)

unread,
Nov 11, 2025, 9:57:05 AMNov 11
to DSI Studio
Hello,

After having extracted the fornix succesfully(thanks for your help again) I have now encountered an important problem during analysis. While I do have data for all 100 segments extracted, and it looks great, I cant tell which side the tractogrpahy started from. Can I find this out somewhere or do I have to interpolate by coordinates.Ie does the fornix tract data start from the hippocampal or mamilarry body side.(datapoint 1)

Frank Yeh

unread,
Nov 11, 2025, 10:07:51 AMNov 11
to christia...@gmail.com, DSI Studio
How did you get the 100 segment data?

On Tue, Nov 11, 2025 at 9:57 AM Christian Gewinn (Christent) <christia...@gmail.com> wrote:
Hello,

After having extracted the fornix succesfully(thanks for your help again) I have now encountered an important problem during analysis. While I do have data for all 100 segments extracted, and it looks great, I cant tell which side the tractogrpahy started from. Can I find this out somewhere or do I have to interpolate by coordinates.Ie does the fornix tract data start from the hippocampal or mamilarry body side.(datapoint 1)

--
You received this message because you are subscribed to the Google Groups "DSI Studio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/1c3ea116-f3a6-47da-8116-9865c1acc07fn%40googlegroups.com.

Christian Gewinn (Christent)

unread,
Nov 12, 2025, 8:44:39 AMNov 12
to DSI Studio
I did what is described in this post here: https://groups.google.com/g/dsi-studio/c/ln8YTMe6OoQ?pli=1. It returned to me data in txt format with 100 datapoints per parameter(ie. fa,ad,md,rd) in a subject wise manner.

Frank Yeh

unread,
Nov 12, 2025, 8:22:43 PMNov 12
to christia...@gmail.com, DSI Studio
I revised the profile function to handle this issue, and please update DSI Studio and try the following:

To get profile from left fornix:

dsi_studio --action=ana --source=*.gqi.fz --tract=Z:\Christian\Tracto\N\ProjectionBasalGanglia_FornixL\*.ProjectionBasalGanglia_FornixL.tt.gz --profile=dti_fa:3:1

Here '3' indicates fiber orientation will be sorted from left to right, and the values will be arranged from hippocampus to mammillary body.

And for the right fornix:

dsi_studio --action=ana --source=*.gqi.fz --tract=Z:\Christian\Tracto\N\ProjectionBasalGanglia_FornixR\*.ProjectionBasalGanglia_FornixR.tt.gz --profile=dti_fa:6:1

Here '6' indicates fiber orientation will be sorted from right to left, and the values will also be arranged from hippocampus to mammillary body.

For other orientations, you may check out command line documentation on the website: https://dsi-studio.labsolver.org/doc/cli_t3.html

Best regards,
Frank

Christian Gewinn (Christent)

unread,
Nov 13, 2025, 7:46:24 AMNov 13
to DSI Studio
I sit possible to infer what direction the extraction happened in the prior extraction, I would assume there would be a starting roi for fornix defined in automated fiber tracking?
I have already done quite some work on the data, so I would prefer to use what I got now, or is that data unreliable in some way?

Cheers,

Christian

Frank Yeh

unread,
Nov 13, 2025, 8:35:08 AMNov 13
to christia...@gmail.com, DSI Studio
I sit possible to infer what direction the extraction happened in the prior extraction, I would assume there would be a starting roi for fornix defined in automated fiber tracking?

Not really.
Fiber tracking goes bidirectionally (as covered by workshop videos at practicum.labsolver.org), and the final coordinates will be listed by either way, depending on the local fiber orientations at the mid portion of the tract:

DSI Studio will list the tract coordinate in a way that makes the" largest absolute value of the mid direction vector" positive at LPS (left-posterior-superior) orientation.
Thus, depending on the brain geometry, the mid portion of fornix can be either the descending part to the hippocampus or the horizontal part at the top of the thalamus. If the mid portion is descending, then the profile will be listed from hippocampus to mammillary body (+ for superior direction). If the mid portion is at the top of the thalamus, the profile will be listed from the mammillary body to hippocampus (+ for posterior).

Trying to figure the direction for one scan could be easy, but how to do it automatically for a large number of scans is not possible, unless using the update done yesterday.

 
I have already done quite some work on the data, so I would prefer to use what I got now, or is that data unreliable in some way?


It is "reliable", but you would need to figure out the direction.

Best,
Frank
 

Christian Gewinn (Christent)

unread,
Nov 18, 2025, 8:16:29 AM (11 days ago) Nov 18
to DSI Studio
Hello Frank,

After trying out above, I encountered the error, which seems to be one that was introduced with this version, since the command worked for the jul 7th version when I tried it in an identical setup. This goes for both my old version of the command and the one you posted above.

action=ana

│ ├──thread_count=20

│ ├──source=*.gqi.fz

│ ├──searching "*.gqi.fz" in "Z:\\Christian\\Tracto\\Reconstructed\\reco"

│ ├──a total of 104 files found

│ ├──*.gqi.fz->//daten.uni-oldenburg.de/neurogroup$/Projects/NBACCHI/Christian/Tracto/Reconstructed/reco/sub-K001_ses-S001_dwi.gqi.fz

│ ├──Z:\Christian\Tracto\N\ProjectionBasalGanglia_FornixL\*.ProjectionBasalGanglia_FornixL.tt.gz->sub-K001_ses-S001_dwi.ProjectionBasalGanglia_FornixL.tt.gz

│ ├──┬──📟run ana

│ │ ├──source=//daten.uni-oldenburg.de/neurogroup$/Projects/NBACCHI/Christian/Tracto/Reconstructed/reco/sub-K001_ses-S001_dwi.gqi.fz

│ │ ├──┬──📂opening //daten.uni-oldenburg.de/neurogroup$/Projects/NBACCHI/Christian/Tracto/Reconstructed/reco/sub-K001_ses-S001_dwi.gqi.fz

│ │ │ ├──fib_ver: 202504

│ │ │ ├──dim: 129 158 137 vs: 1.25 1.25 1.25

│ │ │ ├──trans: -1.25 0 0 80; 0 -1.25 0 81.125; 0 0 1.25 -80; 0 0 0 1;

│ │ │ ├──is qsdr: no

│ │ │ ├──mask voxels: 802767

│ │ │ ├──┬──📟loading image volumes

│ │ │ │ └──⏱31ms

│ │ │ ├──matched template: "human"

│ │ │ └──⏱171ms

│ │ ├──tract=sub-K001_ses-S001_dwi.ProjectionBasalGanglia_FornixL.tt.gz

│ │ ├──❌sub-K001_ses-S001_dwi.ProjectionBasalGanglia_FornixL.tt.gz does not exist. terminating...

│ │ └──⏱180ms

│ └──⏱375ms

Frank Yeh

unread,
Nov 18, 2025, 8:44:55 AM (11 days ago) Nov 18
to christia...@gmail.com, DSI Studio
In this case, *.gqi.fz is pointing to
//daten.uni-oldenburg.de/neurogroup$/Projects/NBACCHI/Christian/Tracto/Reconstructed/reco/sub-K001_ses-S001_dwi.gqi.fz

where *=//daten.uni-oldenburg.de/neurogroup$/Projects/NBACCHI/Christian/Tracto/Reconstructed/reco/sub-K001_ses-S001_dwi

and thus Z:\Christian\Tracto\N\ProjectionBasalGanglia_FornixL\*.ProjectionBasalGanglia_FornixL.tt.gz
will become

Z:\Christian\Tracto\N\ProjectionBasalGanglia_FornixL\//daten.uni-oldenburg.de/neurogroup$/Projects/NBACCHI/Christian/Tracto/Reconstructed/reco/sub-K001_ses-S001_dwi.ProjectionBasalGanglia_FornixL.tt.gz

which does not work out.

I would try using the relative directories to avoid this issue.

Best regards,
Frank

On Tue, Nov 18, 2025 at 8:16 AM Christian Gewinn (Christent)
> To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/1d0ea837-6247-426f-af4e-2e6a3649ed08n%40googlegroups.com.

Frank Yeh

unread,
Nov 18, 2025, 2:25:39 PM (11 days ago) Nov 18
to christia...@gmail.com, DSI Studio
I will fix this issue and let you know once there is an update.

Frank Yeh

unread,
Nov 18, 2025, 8:25:08 PM (11 days ago) Nov 18
to christia...@gmail.com, DSI Studio
I updated DSI Studio. Please try it again and see if it works.
Thanks a lot for reporting this issue,
Frank

Christian Gewinn (Christent)

unread,
Nov 19, 2025, 7:58:08 AM (10 days ago) Nov 19
to DSI Studio
Hello Frank,

It did work properly with the relative paths with the old version, thanks for picking that out. I hope the fix will help future users though.
Everything has worked properly and I should have everything I need for further analysis.Thanks again for the help!

Cheers
Christian
Reply all
Reply to author
Forward
0 new messages