svreg_multiparc

This module generates multiple parcellations of a single subject that has been processed by BrainSuite and SVReg. The figure below shows multiple parcellations of a single subject shown on a smooth representation of the cortex.

System requirements:

Same as in BrainSuite system requirements. Make sure that you have BrainSuite installed and running on your platform.

Usage

To run the svreg_multiparc module,

1. First download the USCBrainMulti atlas here, and unzip it.

2. For matlab, clone the repository and see main_svreg_multiparc.m. For running using binaries, download the binaries for svreg_multiparc here.

3. Process your T1 MRI using BrainSuite and SVReg sequence. Please make sure to use BCI-DNI or USCBrain brain atlas.

4. Run the command. Usage is: svreg_multiparc.sh [subjectbase] [multiatlas dir] [BainSuitePath] [AtlasName] where subbasename: this is as in BrainSuite processed MRI. If the input t1 was path/inmri.nii.gz then subbasename is path/inmri.nii.gz multiatlas_dir: directory path where multi atlas was unzipped in step 1. BraiSuitePath: Path of BrainSuite installation AtlasName: 'all' for all atlases, or specify the one atlas that you want to use. Example usage is as follows:

   ./svreg_multiparc.sh /path/subjectmriname /path/USCBrainMulti /path/BrainSuite19b DKT40

Here are screenshots of smoothed representation of labelled subject cortex for one subject.

Visualize the labels in BrainSuite

You can visualize the generated parcellations in BrainSuite.

1. first open the original mri in BrainSuite.
2. open the multiparc subfolder in the subject folder that has been generated by svreg_multiparc
3. File->Open->Label file and load the .label.nii.gz files from the multiparc subdirectory.
4. drag and drop the two dfs files into brainsuite interface.
5. Load the label description by Tools->Painter tools-> Load Description. Select brainsuite_labeldescrition..xml

You will see the gui as below. You can click on surface or volume and see the label at the bottom left of the GUI.

Comparison

A quantitaive comparison and validation still needs to be performed. Here is one result

List of atlases supported

While it is easy to add any atlas to the svreg-multiparc module, the following atlases have already been imported and are ready to use: AAL(1,2,3), BCI-DNI, Brainnetome, Brodmann, Desikan-Killiany, Destrieux, DKT40, Economo, Entorhinal, Glasser2016, Hammers mith, Harvard Oxford, HCP-MMP1, HMAT, JHU, Juelich, Perirhinal, Schaefer 2018, Talairach, USCBrain, USCLobes, Yeo2011_7, Yeo2011_17

References

1. Joshi AA, Tadel F, Taylor KN, Medani T., Akrami H, Nair D, Shattuck DW, Leahy RM, SVReg-MultiParc: A Method for Automated Labeling of Brains with Multiple Parcellation Schemes, OHBM 2022. poster

2. Joshi AA, Tadel F, Taylor KN, Akrami H, Nair D, Leahy RM, Automatic Labeling of Individual Brains with Multiple Parcellation Schemes, AES 2021. poster