NITRC SIGMA Rat Brain Templates and Atlases Forum: help

RE: Functional Nodes Coordinates

Liam Locke — Fri, 12 Aug 2022 21:53:32 GMT

Hi Partick,

I too am hoping to plot graphs of fMRI data in BrainNetViewer using the SIGMA functional atlas and am looking for a coordinates file or a way to easily generate one. Were you able to solve this issue?

Thanks,
Liam

Functional Nodes Coordinates

Patrick McCunn — Fri, 30 Jul 2021 2:50:35 GMT

Hello

First off thank you for this excellent tool. I'm currently using it for rodent analysis of fMRI and dMRI. I'm looking to present results using BrainNetViewer, the functional Atlas and the supplied Brain Mesh. My one question is if there is a node file for the functional networks which I can load up that has the COG coordinates that correspond to the functional ROI's. Or if there is a software that you would recommend to visualize the networks with the supplied label file, I would be happy to hear the recommendations.

Thank You

RE: VBM analysis

David Andre Barriere — Wed, 21 Oct 2020 13:57:04 GMT

Hello,
This is a SPM12 related issue.
We are working on providing soon a short tutorial and some updated TPM's to address this issue.
We will get back to you soon,
Kind regards,
The Sigma team

RE: VBM analysis

Nan Xu — Mon, 29 Jun 2020 4:48:46 GMT

Hi,

I have been running into the same errors while doing the T2 segment (SPM12). Followed your suggestions, I have increased the voxel size of the InVivo template and rat priors by 10, as well as set "no bias regularization"+"No Affine Registration". However, I am still running into the following segment failures:

SPM12: spm_preproc_run (v7670) 00:37:32 - 29/06/2020
========================================================================
Segment D:\t2.nii,1
29-Jun-2020 00:37:36 - Failed 'Segment'
Error using chol
Matrix must be positive definite.
In file "D:\spm12\spm_preproc8.m" (v7593), function "log_likelihoods" at line 929.
In file "D:\spm12\spm12\spm_preproc8.m" (v7593), function "latent" at line 969.
In file "D:\spm12\spm12\spm_preproc8.m" (v7593), function "spm_preproc8" at line 410.
In file "D:\spm12\spm12\spm_preproc_run.m" (v7670), function "run_job" at line 156.
In file "D:\spm12\spm12\spm_preproc_run.m" (v7670), function "spm_preproc_run" at line 41.
In file "D:\spm12\spm12\config\spm_cfg_preproc8.m" (v7629), function "spm_local_preproc_run" at line 474.

The following modules did not run:
Failed: Segment

Thanks,
hn

RE: VBM analysis

David Andre Barriere — Sat, 23 May 2020 14:10:59 GMT

Hello JW,
I suppose that your AD model induces brain atrophy instead of hypertrophy.
The main missing point that I see from here is that you missed to run a DARTEL normalization step on your data.

Indeed, the normalization step provided in the previous batch is a linear (X/Y/Z + shearing) and you have to perform a non-linear normalization before compare your dataset.
This step is quite long but it permit to calculate the jacobian fields (u_files, please see SPM doc) to warp your GM correctly.
The warpped GM images should be used to perform VBM.

Please have a look on the following papers for details and examples of VBM procedure in rats:

Regional gray matter volume increases following 7days of voluntary wheel running exercise: a longitudinal VBM study in rats.
Sumiyoshi A, Taki Y, Nonaka H, Takeuchi H, Kawashima R.

Voxel-based morphometry and histological analysis for evaluating hippocampal damage in a rat model of cardiopulmonary resuscitation.
Suzuki H, Sumiyoshi A, Taki Y, Matsumoto Y, Fukumoto Y, Kawashima R, Shimokawa H.

Structural and functional alterations in the retrosplenial cortex following neuropathic pain.
Barrière DA, Hamieh AM, Magalhães R, Traoré A, Barbier J, Bonny JM, Ardid D, Busserolles J, Mériaux S, Marchand F.

The SIGMA rat brain templates and atlases for multimodal MRI data analysis and visualization.
Barrière DA, Magalhães R, Novais A, Marques P, Selingue E, Geffroy F, Marques F, Cerqueira J, Sousa JC, Boumezbeur F, Bottlaender M, Jay TM, Cachia A, Sousa N, Mériaux S.

Otherwize, your comparison sounds good, nevertheless I recommand you to calculate the volume of each brain. To do that you could binarized and sum your GM and WM maps (released during the segmentation step(in native space)) in order to calculate the volume of brain parenchyma. Then, you could used the brain volume as regressor to limit the group effect in your analyse.

Finally, you should keep as default the Global calculation/Global normalization and Grand mean scaled value. I think is recommanded for PET analysis but not for MRI.

Good luck.

David from the SIGMA Team

RE: VBM analysis

jw p — Tue, 12 May 2020 11:54:28 GMT

Hellow,

For VBM analysis, I performed the normalization, the last step of the batch file sent last time, and then smoothed.
And statistical analysis was performed between the normal and patient mice by using swc1mr~~~.nii
The statistical analysis options were conducted as follows.

[b]Factorial design specification[/b]
Global calculation
. mean
Global normalization
. Overall grand mean scaling
.. Yes
... Grand mean scaled value
. Normalization: Proportional

F-test showed statistical difference in the hippocampus area.
So, I made ROI of the area showing statistical differences and compared it by extracting the value from the smoothed image using REX of CONN toolbox.
As a result, the Alzheimer's disease model had a larger value than normal mouse.

Does a high value mean that the volume is reduced?
Otherwise, is there any missing part in the analysis?

Thank you so much for your help.
Best.
JW Park

RE: VBM analysis

David Andre Barriere — Wed, 08 Apr 2020 18:06:11 GMT

In your data, the resolution matrix is not identical (1,0,0 ; 0,1,0 ; 0,0,1) and I found massive rotations within the header.

I recommand you to carefully rotate T2 images using dedicated function in FSL, AFNI or anatomist to align them onto the template before starting segmentation.

Usually, I use both anatomist and AIMS functions to do that.
For information see:
http://brainvisa.info/web/download.html
http://brainvisa.info/aimsdata-4.5/user_doc/tutorial.html

I successfully segmented and normalized your T2 image using SPM8. As previously discussed you should increase 10 times the resolution of your data and both priors and template to use SPM for segmentation

I also recommand you to use the brain mask during these processes.

See the batch file in attach.

Best.

David from the SIGMA Team.

RE: VBM analysis

David Andre Barriere — Sat, 04 Apr 2020 12:09:21 GMT

Hello,

We could propose to you to test directly. Could you send us one of your animal ?
You could send it to the following address : sigma.preclinical.resources@gmail.com.
Our data will be tested and removed after.
Sigma team.

RE: VBM analysis

jw p — Tue, 31 Mar 2020 7:54:47 GMT

Hello,

In vivo template was used for spm12 segmentation.
I did the analysis according to your suggestion.

- increase by 10 times the voxel size of the rat priors
- change affine regularisation in your batch as "No affine Registration"
- change biais regularisation in your batch as "no regularisation"

However, the following problems occurred.

SPM12: spm_preproc_run (v7670) 16:34:11 - 31/03/2020
========================================================================
Segment E:\MRI\Complex_trauma_MR\Complex_trauma_MR_\01_WT_01.dcm\o20190801_092845s123a1000.nii,1
76 wp_reg = 100; % Bias wp towards 1/Kb
31-Mar-2020 16:34:18 - Failed 'Segment'
Error using chol
Matrix must be positive definite.
In file "C:\Users\use\Documents\spm12\spm_preproc8.m" (v7593), function "log_likelihoods" at line 929.
In file "C:\Users\use\Documents\spm12\spm_preproc8.m" (v7593), function "latent" at line 969.
In file "C:\Users\use\Documents\spm12\spm_preproc8.m" (v7593), function "spm_preproc8" at line 410.
In file "C:\Users\use\Documents\spm12\spm_preproc_run.m" (v7670), function "run_job" at line 156.
In file "C:\Users\use\Documents\spm12\spm_preproc_run.m" (v7670), function "spm_preproc_run" at line 41.
In file "C:\Users\use\Documents\spm12\config\spm_cfg_preproc8.m" (v7629), function "spm_local_preproc_run" at line 474.

The following modules did not run:
Failed: Segment

RE: VBM analysis

David Andre Barriere — Mon, 30 Mar 2020 9:09:26 GMT

[i]Originally posted by jw p:[/i][quote]Hello.

I want to do VBM analysis using SPM12.
However, when performing segmentation analysis in SPM, the following problems occur:
"Warning: Matrix is singular, close to singular or badly scaled. Results may be
inaccurate. RCOND = NaN. "
The analysis stops with an error message.
Please see the attached segmentation batch file for advice.
If you are analyzing with another program, please recommend the program you used for analysis.

Thank you.[/quote][color=#000000]Hello jw p,[/color]
[color=#000000]I regularly used SPM to do VBM so it should works.
[/color]
[color=#000000]The error message return that there is NaN in priors, nevertheless, I removed them before to upload the dataset. Please could you tell me which priors (In vivo or ex vivo) you used ?
[/color]
[color=#000000]I also investigated your batch and I am not sure that you replaced the human GM, WM and CSF files by the rat ones (the name files are similar to human ones). [/color]
[color=#000000]If the priors are correct, I suggest to you to :[/color]
[color=#000000]- increase by 10 times the voxel size of the rat priors [/color]
[color=#000000]- change affine regularisation in your batch as "No affine Registration"[/color]
[color=#000000]- change biais regularisation in your batch as "no regularisation"[/color]

You could also used SPMmouse to use our tools (http://www.spmmouse.org/ or https://github.com/neurospin/spmmouse/blob/master/spmmouse.m)
[color=#000000]Do not hesitate to contact use again.[/color]

[color=#000000]David
[/color]