BrainGraphStudio

An AutoML ToolKit for Classification of Static Functional Brain Graphs

Developed for Atrium Health's Laboratory for Complex Brain Networks

Developer: Berk Yalcinkaya

Overview

BrainGraphStudio is a GUI-based tool for training, building, and optimizing BrainGNN[1] or BrainGB[2] graph neural networks.

Install Instructions

BrainGraphStudio can be installed for CPU or GPU usage as follow. To download:

1. Install an Anaconda distribution of Python. Note you might need to use an anaconda prompt if you did not add anaconda to the path.
2. Open an anaconda prompt/command prompt
3. If you have an older bgs environment you should remove it with conda env remove -n bgs before creating a new one.
4. Create a new environment with conda create --name bgs python=3.9.0.
5. Activate this new environment by running conda activate bgs
6. To download our package plus all dependencies, run python -m pip install BrainGraphStudio[gpu] on Windows and python3 -m pip install BrainGraphStudio[gpu] on Linux, Ubuntu, and Mac OS. Replace gpu with cpu if you intend to run BrainGraphStudio without GPU. Note, on terminals running zhs, you might need to include the \ escape char before the brackets, as follows: BrainGraphStudio\[gpu\] or BrainGraphStudio\[cpu\]

Next, run the following commands:

pip install torch-scatter==2.0.8
pip install torch-sparse==0.6.12
pip install torch-spline-conv==1.2.1
pip install torch-geometric==2.0.4

Running BrainGraphStudio

To run BrainGraphStudio, open the terminal, activate your bgs conda environment and run the following command

bgs

This should open the UI window and prompt you to load your data, configure the model architecture, and define hyperparameters

References

[1] Xiaoxiao Li, Yuan Zhou, Nicha Dvornek, Muhan Zhang, Siyuan Gao, Juntang Zhuang, Dustin Scheinost, Lawrence H. Staib, Pamela Ventola, James S. Duncan, BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis, Medical Image Analysis, Volume 74, 2021, 102233, ISSN 1361-8415, https://doi.org/10.1016/j.media.2021.102233.

[2] Cui, H., Dai, W., Zhu, Y., Kan, X., Chen Gu, A. A., Lukemire, J., Zhan, L., He, L., Guo, Y., & Yang, C. (2022). BrainGB: A Benchmark for Brain Network Analysis with Graph Neural Networks. IEEE Transactions on Medical Imaging (TMI).