CLI issue when calling an aparc+aseg region

[Hannah Lindsey]

Hi! I want to use the cerebellum-cortex regions from native aparc+aseg files as ROAs when performing differential tractography on the CLI, but because of the "+" in "aparc+aseg.nii.gz", it keeps thinking that I'm trying to call a region called "aparc" (even with quotation marks around the full filename) and then terminates with an error stating that it can't open the file. 

Here's my command:

dsi_studio \

--action=trk \

--source=CH01-1.fz \

--other_slices=CH01-2.fz.qa.nii.gz,brain.nii.gz \

--roa="aparc+aseg.nii.gz":Left-Cerebellum-Cortex+"aparc+aseg.nii.gz":Right-Cerebellum-Cortex \

--dt_metric1=qa \

--dt_metric2=CH01-2.fz.qa \

--min_length=20 \

--max_length=300 \

--dt_threshold=.20 \

--dt_threshold_type=3 \

--track_voxel_ratio=10 \

--seed_count=0 \

--turning_angle=60 \

--step_size=0 \

--smoothing=.20 \

--tip_iteration=4 \

--output=test.tt.gz

and here's the output in terminal:

DSI Studio version: Hou "侯" May 22 2025

┌──📟command line

├──action=trk

├──thread_count=10

├──source=CH01-1.fz

├──┬──📟run 

│  ├──┬──📂opening CH01-1.fz

│  │  ├──fib_ver: 202504

│  │  ├──dim: 80 100 80 vs: 2 2 2 

│  │  ├──trans: -1.99637 0.119541 -0.0148602 166.847; -0.120337 -1.96591 0.343052 118.413; 0.00604865 0.352122 1.96954 -55.3634; 0 0 0 1; 

│  │  ├──is qsdr: no

│  │  ├──mask voxels: 172024

│  │  ├──┬──📟loading image volumes

│  │  │  └──⏱2ms

│  │  ├──matched template: "human"

│  │  └──⏱9ms

│  ├──other_slices=CH01-2.fz.qa.nii.gz,brain.nii.gz

│  ├──add slice: CH01-2.fz.qa.nii.gz

│  ├──┬──📂open CH01-2.fz.qa.nii.gz

│  │  ├──┬──📂opening CH01-2.fz.qa.nii.gz

│  │  │  ├──slice dim: 80 100 80

│  │  │  ├──slice vs: 2 2 2 

│  │  │  ├──slice trans: -1.99801 0.0274871 0.083771 158.608; -0.00794973 -1.94995 0.438963 103.015; 0.0899554 0.449424 1.94812 -56.8086; 0 0 0 1; 

│  │  │  ├──registration will be applied even though the image size is identical. To disable registration, add 'reg' to the file name. 

│  │  │  └──⏱1ms

│  │  ├──add new slices: CH01-2.fz.qa

│  │  ├──dimension: 80 100 80

│  │  ├──The slices have the same dimension as DWI.

│  │  ├──running rigid body transformation to the slices. To disable it, add 'reg' to the file name.

│  │  ├──iso loaded

│  │  ├──registration started using rigid body with regular bound

│  │  ├──rigid body registration using mutual info on gpu

│  │  ├──registration at:2 2 2  size:80 100 80

│  │  ├──reg:1 search:0 cost:60.5005 t:0.12 -0.79 -0.13 r:0.000 0.000 0.000

│  │  ├──reg:3 search:0 cost:60.4831 t:0.11 -0.79 -0.13 r:0.004 0.000 -0.005

│  │  ├──result cost: 60.4831

│  │  ├──reg:1 cost:60.4804 t:0.13 -0.77 -0.11 r:0.004 0.000 -0.005

│  │  ├──reg:3 cost:60.4804 t:0.13 -0.77 -0.11 r:0.004 0.000 -0.005

│  │  ├──result cost: 60.4804

│  │  ├──reg:1 cost:60.4804 t:0.13 -0.77 -0.11 r:0.004 0.000 -0.005

│  │  ├──reg:3 cost:60.4804 t:0.13 -0.77 -0.11 r:0.004 0.000 -0.005

│  │  ├──result cost: 60.4804

│  │  ├──reg:1 cost:60.4759 t:0.15 -0.71 -0.04 r:0.004 0.000 -0.005

│  │  ├──reg:3 cost:60.4759 t:0.15 -0.71 -0.04 r:0.004 0.000 -0.005

│  │  ├──result: t:0.15 -0.71 -0.04 r:0.004 0.000 -0.005

│  │  ├──registration completed

│  │  └──⏱962ms

│  ├──add slice: brain.nii.gz

│  ├──┬──📂open brain.nii.gz

│  │  ├──┬──📂opening brain.nii.gz

│  │  │  ├──slice dim: 256 256 256

│  │  │  ├──slice vs: 1 1 1 

│  │  │  ├──slice trans: -1 -0 -0 128.5; 0 -1 -0 127.5; 0 -0 1 -126.5; 0 0 0 1; 

│  │  │  └──⏱102ms

│  │  ├──add new slices: brain

│  │  ├──dimension: 256 256 256

│  │  ├──running rigid body transformation to the slices. To disable it, add 'reg' to the file name.

│  │  ├──registration started using rigid body with regular bound

│  │  ├──rigid body registration using mutual info on gpu

│  │  ├──registration at:2 2 2  size:128 128 128

│  │  ├──reg:1 search:7 cost:227.312 t:-2.39 -8.77 -9.39 r:0.000 0.000 0.000

│  │  ├──reg:3 search:1 cost:227.061 t:-1.92 -8.81 -9.39 r:-0.081 -0.001 0.014

│  │  ├──registration at:1 1 1  size:256 256 256

│  │  ├──reg:1 search:0 cost:2085.18 t:-1.33 -8.87 -8.84 r:-0.081 -0.001 0.014

│  │  ├──reg:3 search:0 cost:2085.16 t:-1.35 -8.88 -8.83 r:-0.077 -0.004 0.017

│  │  ├──result cost: 2085.16

│  │  ├──reg:1 cost:2085.16 t:-1.35 -8.88 -8.83 r:-0.077 -0.004 0.017

│  │  ├──reg:3 cost:2085.16 t:-1.35 -8.88 -8.83 r:-0.077 -0.004 0.017

│  │  ├──result cost: 2085.16

│  │  ├──reg:1 cost:2085.15 t:-1.31 -8.80 -8.72 r:-0.077 -0.004 0.017

│  │  ├──reg:3 cost:2085.15 t:-1.31 -8.80 -8.72 r:-0.077 -0.004 0.017

│  │  ├──result: t:-1.31 -8.80 -8.72 r:-0.077 -0.004 0.017

│  │  ├──registration completed

│  │  └──⏱9.122s

│  ├──enable differential tracking. available metrics: qa dti_fa md ad rd iso gfa rdi CH01-2.fz.qa brain

│  ├──dt_metric1=qa

│  ├──dt_metric2=CH01-2.fz.qa

│  ├──dt_threshold_type=3

│  ├──dt metrics:(CH01-2.fz.qa-qa)/qa

│  ├──┬──📟tracking parameters:

│  │  ├──otsu_threshold=0.6

│  │  ├──fa_threshold=0

│  │  ├──dt_threshold=.20

│  │  ├──turning_angle=60

│  │  ├──step_size=0

│  │  ├──smoothing=.20

│  │  ├──min_length=20

│  │  ├──max_length=300

│  │  ├──track_voxel_ratio=10

│  │  ├──random_seed=0

│  │  ├──method=0

│  │  ├──check_ending=0

│  │  ├──tip_iteration=4

│  │  ├──seed_count=0

│  │  └──⏱0ms

│  ├──┬──📟setting up regions

│  │  ├──roa=aparc+aseg.nii.gz:Left-Cerebellum-Cortex+aparc+aseg.nii.gz:Right-Cerebellum-Cortex

│  │  ├──load volume from aparc

│  │  ├──❌cannot open file as a regionaparc

│  │  └──⏱0ms

│  └──⏱10.096s

└──⏱10.099s

I know that, in order for the region names to be listed in the interface, the file has to be called "aparc+aseg", so I'm not sure how to handle this without manually opening the aparc+aseg region set and saving them all with a different file name for each subject, which I don't really want to do lol. I was going to do this on the command line, but it seems that --action=vis now only works on through the GUI console, and I don't know how to write a for loop in the console. I have also used --action=vis many times in earlier Hou builds to efficiently perform visual QCs and take screenshots when I have a lot of results from various .fz files, so if it no longer can be used via the command line, that would be a real bummer.... is this something you can reinstate as a function?

As always, thanks for your help!!

[Frank Yeh]

Thanks for reporting this issue.
DSI Studio will be updated in two hours to use this new syntax:

--roa=aparc+aseg.nii.gz:Left-Cerebellum-Cortex,aparc+aseg.nii.gz:Right-Cerebellum-Cortex


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 visit https://groups.google.com/d/msgid/dsi-studio/f66dd2f5-4c70-4ac7-93ec-2c94ce44ec05n%40googlegroups.com.

[Hannah Lindsey]

That works! Thank you!

If using the --cmd="..." function, are plus signs still required when adding additional commands? If so, how would I handle something like this?

--cmd="add_slice,brain.nii.gz+set_slice,aparc+aseg.nii.gz" \

[Frank Yeh]

The--action=vis function can be totally replaced by the 'command
history' function.
You may check out recent workshop videos (week 1) at
https://practicum.labsolver.org/

> To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/37db3835-5dd3-4a5c-a0d6-c36ef8ae6addn%40googlegroups.com.

[Hannah Lindsey]

It can, but only if all the fibs and tract files are in the same directory, right? I can’t get it to let me select more than one file at a time when they’re organized within subject-specific subdirectories (e.g., BIDS)

_________________________________

Hannah M. Lindsey, PhD

Research Associate

TBI and Concussion Center

Department of Neurology

University of Utah

On May 27, 2025, at 11:39 am, Frank Yeh <fran...@gmail.com> wrote:

The--action=vis function can be totally replaced by the 'command

[Frank Yeh]

No worries.
It should be able to handle files in different directories.
If not, I will help fix it as soon as possible.
Frank

[Hannah Lindsey]

At least on a Mac (I haven't tested it on my PC), you are not able to run the command history function on files located within multiple subdirectories. The "finder" window that pops up in newer Hou versions only lets you select multiple files within a single directory. Any file selected becomes deselected whenever you navigate to a different directory. Fixing this would be wonderful (I honestly much prefer the old "Mac-esque" finder window, but that's just my opinion). However, I still would really, really, really LOVE to have the --action=vis command back, if that's possible. There have been a few things that didn't work when I tried to run them through the command history function (e.g., performing tractography with an ROI), and I also really prefer to loop through commands on the CLI rather than the GUI so I can do other things on my computer in the meantime. Is it possible to bring the --action=vis command back and fix the file system issue??? Please??

PS. I tried running --action=vis on an older Hou version (March 4, 2025) that I had used it in before, but I got an error stating "❌Incompatible FIB format. please update DSI Studio to open this new FIB file" when I tried to use it to open an .fz file that I created with Hou May 27 2025. It works on .fz files I created a few months ago. Not sure if that matters if you're going to make the command work with newer versions, but if not, do you know how I can work around that?

Thanks again for all your help!

Hannah

[Frank Yeh]

I will review the code again and see what I can do.

Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/ba5dbe68-e576-4272-88b9-e132986f9447n%40googlegroups.com.

[Frank Yeh]

We can use the command history .csv file directly from the command line interface (CLI) to achieve the same visualization as --action=vis.
Some examples are available here: https://dsi-studio.labsolver.org/doc/cli_vis.html

For instance, you can run it like this:
dsi_studio.app command_history.csv

Regarding the 'finder' window issue you mentioned, I understand it can be quite challenging to work with.
This difficulty stems from a known problem with Qt on macOS 15. We've heard from many users that the standard finder window doesn't appear, making it impossible to open any files.
Unfortunately, at the moment, the only workaround available on macOS for this specific issue is to use the default Qt File Open Dialog.

Maintaining support for macOS presents increasing challenges each year. Recent macOS versions require applications to be notarized, which involves a developer fee, to ensure they open correctly.
Additionally, the compiler on macOS is moving away from conventional OpenGL. I'm seeing compiler warnings about this, indicating that support for the  current 3D code might eventually cease.

Given my limited resources and development capacity, these evolving technical requirements mean there's a possibility that future versions of DSI Studio may not be compatible with the very latest macOS releases.

I will do my best to see how far I can go.

Best,

Frank

[Hannah Lindsey]

That all makes sense. I will try to utilize my PC more in the future (I literally only have it to run DSI Studio lol). Thanks so much for all of your help! I really appreciate all you do to keep DSI Studio working well for everyone! 

Thanks for pointing me in the direction of the .csv option for command history too. I'm sorry I missed that in the documentation and will see how that works for my purposes.

[Frank Yeh]

No problem at all. Also thanks for all the suggestions. 

Please feel free to let me know more. I will do my best to see if there is anything I can improve on.

Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/90412f9c-a288-4511-b56b-fc2fb02ce1ecn%40googlegroups.com.

[Hannah Lindsey]

I just tried to do the command history thing on my PC, but I am still unable to get it to work when data is organized within subject-specific subfolders (e.g., BIDS). The two issues I've run into are: (1) you still can't select more than one .fz when they are in different subfolders, and (2) when trying to open tt.gz files (for example), it will apply the full path from the original subject (e.g., /path/subs/101/tract.tt.gz) to the other subject (e.g., sub 102), so it opens the original subject's tracts in the other subjects dataset... I also tried using the command.csv file on the command line, but it doesn't work with for loops, and the same issue (#2 above) occurs due to the original subject-specific path recorded in the csv. Please let me know if I'm still missing something.

I'm currently performing differential tractography on two separate datasets, both with multiple follow-up scans, so there's a ton of data to sort through/look at/compare, and doing it all manually in the GUI is very tedious and time consuming. 

Any help is appreciated, but I am fully aware of your commitment to do everything you can to make DSI Studio work well for everyone, so I understand if this is just something that will have to be done manually, and I appreciate all that you do anyway :)

[Frank Yeh]

Thanks lot for reporting this issue.
I will implement a button in the GUI that selects a root folder, and DSI Studio will find all files under it.
Once available, I will let you know.
Frank

To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/3b242c0f-d685-4b5c-a74b-2a7173b7ac5en%40googlegroups.com.

[Frank Yeh]

I added a button and improved the support for BIDS. The new version is under build at https://github.com/frankyeh/DSI-Studio/actions/runs/15783102800
MacOS version may take long (few hours) to build. Once updated, please see if it works for you.

You may check the console output to see if it opens/saves the correct file.

If not, please let me know. I will see how I can fix it.

Best,

Frank

[Hannah Lindsey]

Thank you so much! I’ll let you know how it goes

_________________________________

Hannah M. Lindsey, PhD

Research Associate

TBI and Concussion Center

Department of Neurology

University of Utah

On Jun 20, 2025, at 12:16 pm, Frank Yeh <fran...@gmail.com> wrote:



[Hannah Lindsey]

Okay, so I've tested the command history function in the GUI on my PC. The new "Apply to folders" option works as expected, so thank you SO MUCH for that!

I tested a few different functions, and so far, simple things like, autotrack, saving, and exporting tract and region stats work well.

I wanted to try some more complicated things, so first I performed the following steps: (1) opened a baseline dataset for sub 102, (2) inserted the follow-up .nii.gz files extracted from each of the various metrics, (3) loaded the differential tractography results for each metric, (4) exported and saved the stats from all of them into a single .txt file. I know this is something that I know can be done on the CLI with --action=ana, but I'm not sure if there's a way to insert other images using that command (I tried using the --cmd= flag, but it told me that wasn't an option). I could export the tract stats from the follow-up .fz file, but because it would need to be rotated to the baseline dataset before loading the tract files, I thought doing it this way might be easier. When I applied the command history to a different subject (sub 112), the command executed successfully, but the stats that were saved seem to be partially from sub 102?? I've attached the two text files so you can see which rows are the same and which are different.

Then I tried one more set of functions that I don't know how to do on the CLI without the --action=vis function. I loaded all differential tractography result (one for each metric), unchecked them all, and then selected one at a time and saved the 3D screen. The first problem I ran into was that, when applying the command history to a different subject, it doesn't check/uncheck tracts to save individual screenshots of them, but instead keeps them all selected and just takes screenshots of that. Also, no matter which tract is selected, it always wants to save the filename with whatever the last opened tract name is (even if the last opened tract is deleted), and then would overwrite every previous screenshot. After some trial and error, I got it to do what I wanted it to do by only loading one tract, saving the 3D screen, deleting that tract, and then opening up the next one, saving the 3D screen, deleting it, and so on. So there was a work around, but I thought I'd let you know about the stuff I had to troubleshoot around.

I see that the Mac 15 update is ready, so I'll test everything out on that too and let you know if there are any additional issues I run into. Thanks again for everything!

Hannah

[Frank Yeh]

It seems to me that the metrics are complete.
'qir' is qa-to-iso ratio, which is additionally calculated if there is
qa and iso.

>
> Then I tried one more set of functions that I don't know how to do on the CLI without the --action=vis function. I loaded all differential tractography result (one for each metric), unchecked them all, and then selected one at a time and saved the 3D screen. The first problem I ran into was that, when applying the command history to a different subject, it doesn't check/uncheck tracts to save individual screenshots of them, but instead keeps them all selected and just takes screenshots of that.

I will see if I can fix this in the next release.

Best,
Frank

[Hannah Lindsey]

Sorry I was unclear about the metrics reported in the text files. 

I've attached the correct stats from sub-112's differential tractography (102-stats-actual.txt), along with the files I shared earlier and renamed them to avoid confusion. The one that was generated from the command history function for sub-112 (112-stats-command-history), reports values based on the sub-102's differential .tt.gz files. If you compare that report with 102-stats-actual.txt, you'll see that all shape metrics are the same between the two, and the first cell in the final rows indicates that those metrics were calculated from sub-102 in both files, and as such, are the also the same between the two files. The qa-qir rows indicate different values between the two files, but they are still incorrect, because they are stats for 102's tracts estimated from 112's images.

I also just took a closer look at the screenshots generated with command history, and it appears that sub 102's follow-up images and tract files were in fact inserted into 112. I've attached three screenshots demonstrating this: The first (102-rd-actual.jpg) is sub 102 with diff tractography results for RD overlaid on an inserted image from their follow-up scan (ignore the fact that's it's AD lol) and with a native-space isosurface generated from their QA map. The second (112-rd-actual.jpg) is from the exact same steps applied manually to sub-112 in the GUI, so it's their actual diff tractography for RD overlaid on their own follow-up data and their own isosurface. The third (112-rd-command-history.jpg) is the image generated for 112 when I applied the command history function to their data based on what I did to sub 102. You can see here that it's still 112's isosurface, but 102's follow-up image and tracts were loaded onto 112's dataset when command history was used.

I hope this helps clarify the issue!

[Frank Yeh]

Thanks a lot for providing these details

I will look into the issue and fix the bug once I spot it!

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/c8bbbe9c-be92-4ac2-abd2-67af9674900fn%40googlegroups.com.

[Hannah Lindsey]

Thanks so much!

FYI, the same kind of issue seems to apply to saving tract files. I used autotrack on sub-01/ses-01/dwi.fz, saved the tt.gz to that same directory, and exported tract stats to that directory too. I applied the command history to sub-01/ses-02/dwi.fz, and it overwrote the tt.gz file in sub-01/ses-01 with the new tract generated from sub-01/ses-02; however, it saved the correct ses-02 stats from the ses-02 tracts to the correct sub-01/ses-02 directory.

[Frank Yeh]

I created the same file structure, and DSI Studio seemed to save correctly:

First manually do it for ses-01's scan

1. opening E:/New folder/sub-01/ses-01/dwi.fz
2. after autotrack, save tracts to E:/New
folder/sub-01/ses-01/ArcuateFasciculusL.tt.gz

Apply to E:/New folder/sub-01/ses-02/dwi.fz

It saves tracts at E:/New folder/sub-01/ses-02/ArcuateFasciculusL.tt.gz

For the previous metrics issue, do you have a simple example that I
can use to reproduce the bug?

Best,
Frank

> To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/c8b4b1ba-1f74-492c-af8c-41a2c8e2dfdfn%40googlegroups.com.

[Hannah Lindsey]

When you ask for a simple example, would you like me to share the datasets and file structure with you?