Problem tracking Left vs Right corticospinal tract from Template

[Giorgio Marcello]

Hi everyone, does anyone know why, when I try to track the left cortico-spinal tract, I get a "gap" and some lateral fiber deviation? The right one seems fine. I'm using the "Hou Jan 23" version with a fiber template created from about 50 newborn subjects. The problem persists with every template metric used (MD, AD, FA, etc.).

Is it correct to use the statistical results from the right tract that I can visualize even if there is the reconstruction is "broken"?

[Frank Yeh]

Could you describe how to create the template?

[Giorgio Marcello]

Sorry I meant database, not template.

I've done as follow. For each subject: 

* Step1 DICOM to Nifti.

* Step2 Recostruction with Eddy and Moction Correction; then deleted the bad slices from "ROW DWI data". After this corrections I used GQI, template "dHCP_neonate" and "no high b for DTI" and run the reconstruction. I have also visually checked every fib files for artifacts.

* Then I made the database by aggregating the fib files with Step C1 using template "dHCP_neonate"

[Frank Yeh]

This is what I got tracking the neonate template. Mapping CST in the neonate template is not trivial.
The anisotropy is pretty low, and AutoTrack is not yet available for this population.
You may need to experience strategies introduce at practicum.labsolver.org to get it

Hope this helps,

Frank

--
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/411ce9af-171b-4622-8ad8-94be16b90c62n%40googlegroups.com.

[Giorgio Marcello]

I know is not difficult, in fact I can make it for the single subject and as I said for the right CST, but when I try to track the left CST from the tamplate I always get this "error". I placed a ROI over the projecting fibers of pons/cerebral peduncle, a ROA for middle cerebellar peduncle and a limiting volume for one side. Tracking threshold 0; Min lenght 30 mm; Max 200 mm; Angular threshold from 0 to 20 (15 seems the best). But while I can track the left, as other bundles as well, I cannot make it work for right CST

[Frank Yeh]

The current neonate and baby at HCP lifespan studies are not good enough. Thus the template is hard to track.

To my best knowledge, there is a new HCP-baby project, and they are going to release a much better set of data. Once the data are available, I will create a new template to replace the current one.

Sorry about the difficulty.

Best,

Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/c040da0e-4029-422a-b09f-aa56122064b3n%40googlegroups.com.

[tommaso casseri]

Hi.

I tried to track CST in the neonate template of DSI studio (not superimposing any database or single patient) and met the same issue. No problem with right CST while lef CST seems broken.

I used a single ROI at about the level of the PLIC (as shown in the image, same ROI bilaterally).
I used the last HOU version.

Thanks,

Tommaso

[Frank Yeh]

I am going to check again and update it.

See if it helps.

Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/2e7e2575-d5d4-4ddc-ac6d-1777bebdf13an%40googlegroups.com.

[Frank Yeh]

Hi All,

    I did a detailed check and confirmed the template fiber has a problem at the right around the internal capsule, as shown by the beginning of this post.

    The template was created a long time ago, and I am going to regenerate the template and see if it works. 

    It may take a few days. I will update the progress in this post.

    Thanks again for reporting all the issues.

Best,

Frank

[tommaso casseri]

Hi Frank,

thank you very much!

Best,

Tommaso

[Frank Yeh]

Hi All,

    I re-created the neonate template. Please check the updated version if it works for you. 

Best,

Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/d02357ae-c058-4744-a53f-4d42b544a4b8n%40googlegroups.com.

[Giorgio Marcello]

Hi All,

now it works fine, there isn't any "gap" in CST tracking. Thanks a lot Frank!

[tommaso casseri]

Hi all,

I also confirm, the new template works well.

Thank you Frank!

[Giorgio Marcello]

Hi, sorry everyone but I'm here to re-open the discussion. Now apparently there is no problem in tracts reconstruction but when I use ROIs and tracts to obtain stat results I can only get "0" value. If I use the database made from old DSI studio version I can get some results, but none with the database created with same fib files but with Hou Mar 1 version.

[Frank Yeh]

When you created the new db file, did you start from the .src.gz file to get the new .fz files and then use .fz to create .db.fz?

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/debe56b7-5227-40c3-b0ea-bbd300a29c82n%40googlegroups.com.

[Giorgio Marcello]

I've already tried different strategies and noticed the following points.

1. If I start from DICOM/NIfTI and create a database again (nii → sz → fz → db.fz), I obtain no result.
2. I get the same result using fib files created with an old DSI Studio version.
3. If I open with "Hou 1 Mar" a db made with an old DSI Studio version, it works fine—I can obtain stats from ROIs and tracts. Is it correct to do so? If I can visually check that ROIs and tracts created on the new template fit well on the old database based on the old template, can I extract statistical data with the new version of the software, or is there any problem?
4. Using a VOI almost as big as the entire volume of the template image, I can get some stats from single subjects. Is it possible that during coregistration occur some geometric error? (attached image)
5. There is no problem opening a new or old .fib file (not a db) and extracting statistical data or tracking.

 

[Frank Yeh]

Sorry about the problem and thank you for reporting.

I will check and fix it asap.

Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/370e31d1-60b3-429e-b512-9a3b75475b7fn%40googlegroups.com.

[Frank Yeh]

Thank you for your help. I found the bug and fixed it. Please update DSI Studio and see if it works for you.

The bug is located at fz → db.fz step, where the template resolution (0.5) is too high for the data.
The updated version will downsample the template to match that of .fz's resolution.

Thanks again for reporting the bug.

If there is any remaining issue, please let me know.

Best,

Frank

[Giorgio Marcello]

Hi Frank, after downloading the last DSI Studio version (GPU) the main .exe file is missing in the folder, so I cannot check if there is still some issue.

[Frank Yeh]

I fixed it. Please see the updated one works for you.
Thanks again!
Frank

> --
> 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/9199d05b-d98d-4372-a70d-a9089430c9bcn%40googlegroups.com.

[Giorgio Marcello]

It is still missing, sorry.

[Frank Yeh]

I am updating it again. My apology for the inconvenience. 

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/c234f3a8-cbfc-4874-9d47-8eb8edb8e722n%40googlegroups.com.

[Frank Yeh]

The update is ready.

[Giorgio Marcello]

Hi Frank, now the .exe file is in the folder but the problem with the database persist. I've already tried to make a new dti_fa database, starting from nifti files to db.fz, the statistical results is again a "0" value.
Has anyone else had the same problem?

Il giorno martedì 11 marzo 2025 alle 16:47:54 UTC+1 Frank Yeh ha scritto:

The update is ready.

[Frank Yeh]

Sorry about that. I tested on my side, and it works for neonate data.
Do you have one sample .sz file? I can use it to create a db and see
if it works.

Thank you for your patience.
Frank

> --
> 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/bb4d1e35-07e4-4ad4-8446-9ed43afe1126n%40googlegroups.com.

[Giorgio Marcello]

Hi Frank, sorry for the late response. I'm uploading right now a sample of 10 .sz files.

[Frank Yeh]

Thanks! I will check the data and update DSI Studio soon.
Frank

On Tue, Mar 18, 2025 at 2:31 PM Giorgio Marcello

<giorgioma...@gmail.com> wrote:
>
> Hi Frank, sorry for the late response. I'm uploading right now a sample of 10 .sz files.
>
> Il giorno martedì 11 marzo 2025 alle 21:25:29 UTC+1 Frank Yeh ha scritto:
>>
>> Sorry about that. I tested on my side, and it works for neonate data.
>> Do you have one sample .sz file? I can use it to create a db and see
>> if it works.
>>
>> Thank you for your patience.
>> Frank
>>
>> On Tue, Mar 11, 2025 at 4:10 PM Giorgio Marcello
>> <giorgioma...@gmail.com> wrote:
>> >
>> > Hi Frank, now the .exe file is in the folder but the problem with the database persist. I've already tried to make a new dti_fa database, starting from nifti files to db.fz, the statistical results is again a "0" value.
>> > Has anyone else had the same problem?
>> > Il giorno martedì 11 marzo 2025 alle 16:47:54 UTC+1 Frank Yeh ha scritto:
>> >>
>> >> The update is ready.
>> >
>> > --
>> > 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/bb4d1e35-07e4-4ad4-8446-9ed43afe1126n%40googlegroups.com.
>
> --
> 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/d62bc494-a788-4225-bbe2-a758ef803e04n%40googlegroups.com.

[Frank Yeh]

Everything went smoothly on my end. I used QSDR at 1mm to get .fz files
The connectometry database looks good.

Perhaps, you may post your exact step, and I will see if I can spot a problem.

Frank

[Giorgio Marcello]

 I did not use the QSDR. With the old version—the one before the right CST tracking problem—I used to generate .fz files using GQI, then obtain the .db file and extract statistics without any issues. The steps are, as I said before:

 For each subject:
* Step1 DICOM to Nifti.
* Step2 Recostruction with Eddy and Moction Correction; then deleted the bad slices from "ROW DWI data". After this corrections I used GQI, template "dHCP_neonate" and "no high b for DTI" and run the reconstruction. I have also visually checked every fib files for artifacts.
* Then I made the database by aggregating the fib files with Step C1 using template "dHCP_neonate"

[Frank Yeh]

Using QSDR with 1mm resolution would work.

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/36b864fd-f83c-40ab-9f4a-cc19346ed8c5n%40googlegroups.com.