/Interpretable-Graph-Neural-Networks

Interpretable graph neural networks for fMRI data

Primary LanguagePython

Graph Neural Network for Brain Network Analysis

This is a customized implementation of of BrainGNN. We are using a private ABCD resting-state fMRI data with 53 features.

Usage

Setup

pip

See the requirements.txt for environment configuration.

pip install -r requirements.txt

PYG

To install pyg library, please refer to the document

Dataset

ABCD

Each subject is considered as a single graph, ROI's as nodes and partial correlation strength as edges.

Reference

This is a implementation of Interpretable Graph Neural Networks - https://www.sciencedirect.com/science/article/pii/S1361841521002784