mihkelwork / lg-gnn

This repo is the official implementation of Classification of Brain Disorders in rs-fMRI via Local-to-Global Graph Neural Networks

LG-GNN

This repo is the official implementation of Classification of Brain Disorders in rs-fMRI via Local-to-Global Graph Neural Networks

I. Usage:

The main contribution of our work is the proposed LG-GNN architecture, which enables crosstalk between local brain regions and the global population, and allows identification biomarkers.

The data used in our work are from ADNI and ABIDE. Please follow the relevant regulations to download from the websites.

Using the main.py to train and test the model on your own dataset. The proposed network LG-GNN is defined in the layer.py. It can be easily edited and embed in your own code.

II. Requirements:

• torch_geometric
• torch
• scipy
• visdom
• numpy
• os

III. Citation：

If our paper or code is helpful to you, please cite our paper. If you have any questions, please feel free to ask me.

@article{zhang2023classification,
  title={Classification of brain disorders in rs-fMRI via local-to-global graph neural networks},
  author={Zhang, Hao and Song, Ran and Wang, Liping and Zhang, Lin and Wang, Dawei and Wang, Cong and Zhang, Wei},
  journal={IEEE Transactions on Medical Imaging},
  volume={42},
  number={2},
  pages={444--455},
  year={2023},
  publisher={IEEE}
}