How to capture modularity and integrate a set of multi-graphs into a single graph?
Please contact [email protected] for inquiries. Thanks.
This work is accepted at the MICCAI 2022 conference, Singapore.
Dual Hierarchical Integration Network: Dual Hierarchical Integration Network of Multigraphs for Connectional Brain Template Learning
Fatih Said Duran, Abdurrahman Beyaz and Islem Rekik
BASIRA Lab, Faculty of Computer and Informatics, Istanbul Technical University, Istanbul, Turkey
Abstract: A connectional brain template (CBT) is a normalized representation of a population of brain multigraphs, where two anatomical regions of interests (ROIs) are connected by multiple edges. Each edge captures a particular type of interaction between pairs of ROIs (e.g., structural/functional). Learning a well-centered and representative CBT of a particular brain multigraph population (e.g., healthy or atypical) is a means of modeling complex and varying ROI interactions in a holistic manner. Existing methods generate CBTs by locally integrating heterogeneous multi-edge attributes (e.g., weights and features). However, such methods are agnostic to brain network modularity as they ignore the hierarchical structure of neural interactions. Furthermore, they only perform node-level integration at the individual level without learning the multigraph representation at the group level in a layer-wise manner. To address these limitations, we propose Dual Hierarchical Integration Network (Dual-HINet) for connectional brain template estimation, which simultaneously learns the node-level and cluster-level integration processes using a dual graph neural network architecture. We also propose a novel loss objective to jointly learn the clustering assignment across different edge types and the centered CBT representation of the population multigraphs. Our Dual-HINet significantly outperforms state-of-the-art methods for learning CBTs on a large-scale multigraph connectomic datasets.
This code was implemented using Python 3.8 (Anaconda) on Windows 10.
You can edit config.py file to configure our Dual-HINet method according to your needs.- Go to https://www.anaconda.com/products/individual
- Download version for your system (We used Python 3.8 on 64bit Windows 10)
- Install the platform
- Create a conda environment by typing:
conda create –n DualHINet pip python=3.8
Copy and paste following commands to install all packages (CPU version)
$ conda activate DualHINet
$ conda install pytorch==1.4.0 torchvision==0.5.0 cpuonly -c pytorch
$ pip install scikit-learn
$ pip install matplotlib
$ pip install torch-scatter==latest+cpu -f https://pytorch-geometric.com/whl/torch-1.4.0.html
$ pip install torch-sparse==latest+cpu -f https://pytorch-geometric.com/whl/torch-1.4.0.html
$ pip install torch-cluster==latest+cpu -f https://pytorch-geometric.com/whl/torch-1.4.0.html
$ pip install torch-spline-conv==latest+cpu -f https://pytorch-geometric.com/whl/torch-1.4.0.html
$ pip install torch-geometricThis is all for CPU installation, please visit (optional) PyTorch-Geometric’s web page (https://pytorch-geometric.readthedocs.io/en/latest/notes/installation.html) for description on installing GPU version. Code will check the version of dependencies and availability of GPU. If everything is configured correctly, it will utilize GPU automatically.
In case you want to use Dual-HINet on your multi-graphs, we represent each brain multi-graph with a stacked symmetrical connectivity matrices. Therefore our model expects you to provide a path of M array saved as a binary file in MAT .mat format with shape [#Subjects, #Nodes, #Nodes, #Views]. For our case network nodes were regions of interest in the brain, however, any type of node is a valid input. If your networks are single graph, you can simply set #Views to 1. Also, you can run DualHInet on a simulated dataset by setting Dataset = 'S' at config.py file.
To run our code, open up a terminal at Dual-HINet’s directory and type in
$ conda activate DualHINet & python demo.py| Component | Content |
|---|---|
| config.py | Includes hyperparameter and other options. You may modify it according to your needs. |
| GNN.py | Graph Neural Network part of the model. |
| Dual_HINet.py | Driver code that import variables from config.py and trains Dual-HINet (cross-validation). |
| helper.py | Includes some helper functions |
| cbts and model params/ | Includes model parameters and the estimated CBTs for the datasets used on the paper. |
| simulated dataset/example.py | A simulated dataset that shows the required data format. |
| res/"model name"/ | After the training, this directory includes model parameters, final CBT, and subject biased CBTs for each fold. |
| temp/ | Includes interim model parameters that are saved for each 10 epoch. Deleted after the training. |
The figure demonstrates an example of output for a population of 155 ASD subjects where each subject has 6 views (each represented by 35 by 35 matrix). Our code takes in a numpy array of size [155, 35, 35, 6] and outputs a 35 by 35 matrix.
To install and run Dual-HINet, check the following YouTube video:
https://www.youtube.com/watch?v=Q_hyobvxpDY
To learn about how Dual-HINet works, check the following YouTube video:
https://www.youtube.com/watch?v=xpK5FMFWrMc
Fey, M. & Lenssen, J. E. Fast graph representation learning with PyTorch Geometric. In ICLR Workshop on Representation Learning on Graphs and Manifolds (2019).
Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, Zachary DeVito, Zeming Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. Automatic differentiation in pytorch. In NIPS-W, 2017.
...
@InProceedings{10.1007/978-3-031-16431-6_29,
author="Duran, Fatih Said
and Beyaz, Abdurrahman
and Rekik, Islem",
editor="Wang, Linwei
and Dou, Qi
and Fletcher, P. Thomas
and Speidel, Stefanie
and Li, Shuo",
title="Dual-HINet: Dual Hierarchical Integration Network of Multigraphs for Connectional Brain Template Learning",
booktitle="Medical Image Computing and Computer Assisted Intervention -- MICCAI 2022",
year="2022",
publisher="Springer Nature Switzerland",
address="Cham",
pages="305--314",
abstract="A connectional brain template (CBT) is a normalized representation of a population of brain multigraphs, where two anatomical regions of interests (ROIs) are connected by multiple edges. Each edge captures a particular type of interaction between pairs of ROIs (e.g., structural/functional). Learning a well-centered and representative CBT of a particular brain multigraph population (e.g., healthy or atypical) is a means of modeling complex and varying ROI interactions in a holistic manner. Existing methods generate CBTs by locally integrating heterogeneous multi-edge attributes (e.g., weights and features). However, such methods are agnostic to brain network modularity as they ignore the hierarchical structure of neural interactions. Furthermore, they only perform node-level integration at the individual level without learning the multigraph representation at the group level in a layer-wise manner. To address these limitations, we propose Dual Hierarchical Integration Network (Dual-HINet) for connectional brain template estimation, which simultaneously learns the node-level and cluster-level integration processes using a dual graph neural network architecture. We also propose a novel loss objective to jointly learn the clustering assignment across different edge types and the centered CBT representation of the population multigraphs. Our Dual-HINet significantly outperforms state-of-the-art methods for learning CBTs on a large-scale multigraph connectomic datasets. Our source code can be found at https://github.com/basiralab/Dual-HINet.",
isbn="978-3-031-16431-6"
}