We introduce a novel GCN-based Conditional GAN model that generates representative and realistic synthetic FC, aiming to provide a more precise and robust MDD diagnosis. By integrating GCN architecture into our GAN model, our framework demonstrated the capability to accurately capture the complex topology characteristics and intricate patterns present in FC. Additionally, the class-aware discriminator further enhanced the quality and effectiveness of synthetic FC generation in GC-GAN framework. Furthermore, we introduced a topology refinement technique that optimizes the graph structure using the augmented synthetic FC, thereby improving the performance of MDD diagnosis.

• config.py
• load_data.py : load data
• model.py : gcn based model arichtecture
• gcgan_ae.py : gcn autoencoder for generator
• gnn_baseline.py : gcn classifier for baseline classifier and discriminator
• gcgan_train.py : gc-gan model
• gcgan_data_augmenation.py : data augmentation
• gcgan_topology_refinement.py : topology refinement
• mrmr_topology.py : mrmr feature selection algorithm based topology creation
• graph_utils.py : graph model function
• utils.py : acc,sen,spec function
• functional.py : mish activation
• etc

To run this project, you will need:

• Python 3.7 or higher
• torch 1.8.0+cu111
• torchvision 0.9.0+cu111
• torch-geometric 2.0.3
• torch-scatter 2.0.9
• torch-sparse 0.6.12
• numpy 1.22.2
• etc.

1. Make sure you have Python 3.7 or higher installed.

2. Install the required packages using pip:

pip install -r requirements.txt

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No. 2019-0-00079, Artificial Intelligence Graduate School Program(Korea University), No. 2022-0-00871, Development of AI Autonomy and Knowledge Enhancement for AI Agent Collaboration), and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS202300212498).