[lenia]

Hi everybody!

Me and my workgroup want to do (semi) automatic segmentation of the cerebellum and other areas. After this, we are interested in using these ROIs to do functional analysis. We have been exploring BrainSuite’ software and we are very excited for using it. However, there are a number of a questions that we would like to ask you:

1. I understand that the SVReg registers the surfaces and volumes to a brain atlas manually labeled by an neuroanatomist. So the ROIs defined after this are not in native space, right? Are they defined after the normalization based on this specific atlas?
2. We are still debating if we want fully automatic segmentation or if we want to do manual corrections. If we want to do that, is it possible to edit the ROIs given by the program? In this specific project our sample are elderly adults: do you think that factor could lead to major differences on the automatic segmentation given by the program?
3. After we having the ROIs, we would like to use them on brainvoyager. Is it possible to extract the regions to a mask? Since the cerebellum is extracted on the processing stage, is it possible to have this ROI also?

Thank you!

Lénia

[Anand Joshi]

Hi lena, Thank you for using BrainSuite.
1. The ROIs generated by BrainSuite are in the native coordinates of the subject. This is because BrainSuite and SVReg warp atlas to match the subject and transfer the ROIs from the atlas to subject.
2. It is possible to edit the ROIs generated by BrainSuite using the Painter tool
http://brainsuite.org/delineation/roi/label-painter-tool/
If you want to modify the atlas or draw your own ROIs and make your own atlas, you can do that using the custom atlas creation method:
http://brainsuite.org/processing/svreg/creating-custom-atlases/
BrainSuite should be able to process elderly adults.
3. The ROIs are stored as a separate volume in nifti file format. Most brain imaging softwares support nifti format.

[lenia]

Hi Anand!

Thank you for the quick reply.
Regarding the third point, I cannot find the files for each ROI on the folder after the processing. How can I have access to the files?

Thank you again

[lenia]

Hello again,

Sorry for insisting on this topic, but I can’t find the files for each ROI. I also looked on this page (http://brainsuite.org/processing/surfaceextraction/output-files/) and I still can’t find the output files corresponding to the ROIs.

Can someone help me?
Thank you!

[Anand Joshi]

Hi Lenia, Sorry for the delay in getting back to you on this. The ROIs are generated by SVReg module in BrainSuite. The output file names are defined here:
http://brainsuite.org/processing/svreg/output/

Anand

[lenia]

Hello Anand,

Thank you for the reply. Perhaps I’m not being clear: do we have access to a separate file for each ROI? For example, do we have a file for the cerebellum, another file for the hippocampus, etc.? Because I can’t find this. Which file give us the ROIs?

Thank you again,

[Anand Joshi]

All the ROIs are stored in
*.svreg.label.nii.gz file.
The labels are actually 16-bit volumes, meaning that each structure label is a number between 0 and 65536. To analyze the scans automatically, it’s necessary to know which number corresponds to the region you’re interested in. Open the folder where the scan (and all of the files generated by the CSE sequence and SVReg) are located and open brainsuite_labeldescription.xml

The statistics of cortical thickness, tissue volume etc are stored in
*roiwise.stats.txt

[naeem dad]

can anyone help me out with this……i have Matlab17 installed already on my computer, does that affect the installation of brainsuit and is it appropriate to add MCR 2015b while having a previously installed Matlab17?

[Anand Joshi]

You need to have MCR 2015b installed on the computer. A different version of MCR or a full Matlab installation won’t work. Please make sure to have the exact version of MCR for BrainSuite to work correctly.
However, having additional installations of Matlab or different version of MCR won’t affect BrainSuite performance.

• This reply was modified 5 years, 8 months ago by Anand Joshi.

[Author]

Posts