Unable to create connectometry database

[Hannah Lindsey]

Hey Frank,

I'm trying to create a connectometry database with some longitudinal gqi data. I've never had any issue creating one in the past with a different dataset (also longitudinal), but I'm having some trouble with this particular dataset. I originally tried to do this on the Dec 6 2023 version (both Mac and PC), and I was able to create the database, but when I tried to compute the longitudinal change in Step C2, I get an error message that says "cannot compute differences in longitudinal data". I updated the software to the March 29 version (qt5 on Mac and GPU on PC), but I'm not even able to create the connectometry database from these data on the newer version. I'm able to add the subjects to Step C1a, but when I click "Create database", I get an error popup window, but there is no message (see attached screenshot).

I tried the same steps with a different dataset on both version, and everything worked great. 

Any ideas on what the problem could be?

 

[Frank Yeh]

Sorry about the problem. I will review code to see if I can find a clue.

I will get back to you soon.

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 on the web visit https://groups.google.com/d/msgid/dsi-studio/0131f5d1-b9e0-4e8e-b1b1-45c5cb1f5dd8n%40googlegroups.com.

[Frank Yeh]

I'm trying to create a connectometry database with some longitudinal gqi data. I've never had any issue creating one in the past with a different dataset (also longitudinal), but I'm having some trouble with this particular dataset. I originally tried to do this on the Dec 6 2023 version (both Mac and PC), and I was able to create the database, but when I tried to compute the longitudinal change in Step C2, I get an error message that says "cannot compute differences in longitudinal data".

DSI Studio found negative values in the FIB file and thus cannot compute differences.

It seems that the FIB files have a problem (should not contain negative values)

 

I updated the software to the March 29 version (qt5 on Mac and GPU on PC), but I'm not even able to create the connectometry database from these data on the newer version. I'm able to add the subjects to Step C1a, but when I click "Create database", I get an error popup window, but there is no message (see attached screenshot).

Could you post the console output? (click on the console button on the right upper corner of the main window)

It seems to me that there is some problem with one or more FIB files used here.

Likely you may need to recreate them.

Frank

 

I tried the same steps with a different dataset on both version, and everything worked great. 

Any ideas on what the problem could be?

 

--

[Hannah Lindsey]

I can try recreating the fib files.. I'll let you know if this fixes the issue or not.

In the meantime, here's the console output when I try to create the database with the Mar 29 version on a Mac:

DSI Studio version: Chen"陳" Mar 29 2024

| DSI Studio version: Chen"陳"

|_9 ms

open FIB file PBM-802-1.fib.gz

| loading fiber and image data

|-loading image volumes

| |_8 ms

| initiating data

| default template set to young adult

| FIB file loaded

|_112 ms

open FIB file ICBM152_adult.fib.gz

| loading fiber and image data

|-loading image volumes

| |_7 ms

| initiating data

| matched template (by file name): ICBM152_adult

| FIB file loaded

|_46 ms

reading data

|_0 ms

regrid

|_31 ms

[Hannah Lindsey]

Alrighty, I just recreated both pre and post src and fib files for one of the subjects, and I'm having the same issues as before, where I can't compute longitudinal differences on Dec 6 version or create a database at all on the Mar 29 version. 

If the issue is negative values, like you suggested, do you think this error would have occurred somewhere during the preprocessing (these data went through topup, eddy, and a bias correction), or is it an acquisition issue? Any ideas on how I can localize the source of the negative values and correct them?

On Tuesday, April 9, 2024 at 1:27:52 PM UTC-4 Frank Yeh wrote:

[Frank Yeh]

If the issue is negative values, like you suggested, do you think this error would have occurred somewhere during the preprocessing (these data went through topup, eddy, and a bias correction), or is it an acquisition issue? Any ideas on how I can localize the source of the negative values and correct them?

I never met this problem before, and there is a good chance that there is a bug that only happens under some special conditions (my gut feeling).

Hope to fix this asap.

If not too much hassle, I can create connectometry database from your data on my side and see if I can identify the problem.

SRC files are de-identified if there is no identifiable information on the file name.

Frank

[Erin Bondy]

Hi Frank,

I've actually run into the same issue when trying to create a connectometry database running DSI Studio (March 29 version) on linux. My console output looked the same as what Hannah provided earlier. 

Best,

Erin

[Frank Yeh]

Thank you for reporting! I am currently looking into this and will see if I can identify the problem in the first shot.

Best,

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 on the web visit https://groups.google.com/d/msgid/dsi-studio/dbfd8c58-3488-4e13-8561-6bd3e11d60d5n%40googlegroups.com.

[Frank Yeh]

I quickly created a connectometry database on my side, but it did not generate the same error.

Likely the problem could be either (1) platform-dependent (I tested on Windows) or (2) data-dependent.

If you may share just one SRC file, I will see if I can reproduce the error and fix it. 

Sorry for all the hassle/trouble, and hope to fix this soon.

Frank

[Frank Yeh]

Hi All,

    I finally reproduced the error and fixed it. The updated version is now building at https://github.com/frankyeh/DSI-Studio/actions/runs/8648152500 It should be ready in one hour.

    The cause of the problem is data-dependent. It happens when the data's spatial resolution is lower than the template (e.g. 2.1mm). DSI Studio will regrid and upsample the data to make it the same as the template. However, the code mistakenly reported a failure even though the regrid was done successfully.

    The updated version fixed this problem. The previous FIB files should be good to go.

    Nonetheless, I still have no clue about the negative value problem that stopped us from computing the longitudinal change. Hope to find the cause soon.

    Sorry for all the trouble!

Best,

Frank

[Hannah Lindsey]

Frank,

Thanks so much for fixing the connectometry database creation issue! As for the negative values problem, I can share the src files with you. I identified which subjects have the problem. Do you want me to send you src files for just the problematic scans or for the entire sample?

[Frank Yeh]

Just the one with the issue (SRC and the FIB).

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/d871e9df-120c-49fc-a404-0c01f2f3dbebn%40googlegroups.com.

[Hannah Lindsey]

Alrighty, I'm uploading them now.

[Frank Yeh]

Thank you for your help. I identified the problem and fixed it. Another update is now under build at https://github.com/frankyeh/DSI-Studio/actions/runs/8652048041

The cause of this bug is that the FIB was created using GQI, and DSI Studio needs a spatial warp to the MNI space. The cubic spline interpolation used in wrapping may introduce a negative value in the background, causing DSI Studio to think it is a longitudinal dataset.

We usually use QSDR FIB, and there is no need for additional spatial warping. Thus I never met this problem before (sorry about that).

The fix was done by adding an extra filter to remove negative values. Thus your GQI FIB files are good, and they can be directly used to create a new connectometry database.

BTW, in the future, the GQI FIB does not need the ODF value. The warping only needs the QA value, and ODF information is not used.

Thank you for the help, and sorry again for all the trouble.

Best regards,

Frank

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/a3107bba-7833-40af-be67-46c620dcef94n%40googlegroups.com.

[Hannah Lindsey]

Amazing! Thanks so much. I will let you know if I run into any more issues

[Erin Bondy]

Wonderful! Thank you so much, Frank!