BrainNNExplainer: An Interpretable Graph Neural Network Framework for Brain Network based Disease Analysis

Interpretable brain network models for disease prediction are of great value for the advancement of neuroscience. GNNs are promising to model complicated network data, but they are prone to overfitting and suffer from poor interpretability, which prevents their usage in decision-critical scenarios like healthcare. To bridge this gap, we propose BrainNNExplainer, an interpretable GNN framework for brain network analysis. It is mainly composed of two jointly learned modules: a backbone prediction model that is specifically designed for brain networks and an explanation generator that highlights disease-specific prominent brain network connections. Extensive experimental results with visualizations on two challenging disease prediction datasets demonstrate the unique interpretability and outstanding performance of BrainNNExplainer.

[1]  Haiping Lu,et al.  MPCA: Multilinear Principal Component Analysis of Tensor Objects , 2008, IEEE Transactions on Neural Networks.

[2]  Ghassan Hamarneh,et al.  BrainNetCNN: Convolutional neural networks for brain networks; towards predicting neurodevelopment , 2017, NeuroImage.

[3]  My T. Thai,et al.  PGM-Explainer: Probabilistic Graphical Model Explanations for Graph Neural Networks , 2020, NeurIPS.

[4]  Philip S. Yu,et al.  Clustering on Multi-source Incomplete Data via Tensor Modeling and Factorization , 2015, PAKDD.

[5]  Philip S. Yu,et al.  Multi-view Graph Embedding with Hub Detection for Brain Network Analysis , 2017, 2017 IEEE International Conference on Data Mining (ICDM).

[6]  Jure Leskovec,et al.  Hierarchical Graph Representation Learning with Differentiable Pooling , 2018, NeurIPS.

[7]  Paige L. Williams,et al.  Default Mode Connectivity in Youth With Perinatally Acquired HIV , 2015, Medicine.

[8]  Wenhan Shi,et al.  Conditional Structure Generation through Graph Variational Generative Adversarial Nets , 2019, NeurIPS.

[9]  Jure Leskovec,et al.  node2vec: Scalable Feature Learning for Networks , 2016, KDD.

[10]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.

[11]  Chandan Ganesh,et al.  BrainNET: Inference of brain network topology using Machine Learning , 2019, bioRxiv.

[12]  Shuiwang Ji,et al.  Explainability in Graph Neural Networks: A Taxonomic Survey , 2020, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[13]  Jiayu Zhou,et al.  Boosted Sparse and Low-Rank Tensor Regression , 2018, NeurIPS.

[14]  Jure Leskovec,et al.  GNNExplainer: Generating Explanations for Graph Neural Networks , 2019, NeurIPS.

[15]  M. Ruggeri,et al.  Longitudinal investigation of the parietal lobe anatomy in bipolar disorder and its association with general functioning , 2017, Psychiatry Research: Neuroimaging.

[16]  Andreas Daffertshofer,et al.  Comparing Brain Networks of Different Size and Connectivity Density Using Graph Theory , 2010, PloS one.

[17]  Olaf Sporns,et al.  Complex network measures of brain connectivity: Uses and interpretations , 2010, NeuroImage.

[18]  Carl Yang,et al.  Joint Embedding of Structural and Functional Brain Networks with Graph Neural Networks for Mental Illness Diagnosis , 2021, ArXiv.

[19]  Pan Li,et al.  On Positional and Structural Node Features for Graph Neural Networks on Non-attributed Graphs , 2021, ArXiv.

[20]  Carl Yang,et al.  Co-Embedding Network Nodes and Hierarchical Labels with Taxonomy Based Generative Adversarial Networks , 2020, 2020 IEEE International Conference on Data Mining (ICDM).

[21]  Chen Cai,et al.  A simple yet effective baseline for non-attribute graph classification , 2018, ArXiv.

[22]  Daoqiang Zhang,et al.  Sub-network Based Kernels for Brain Network Classification , 2016, BCB.

[23]  Bo Zong,et al.  Parameterized Explainer for Graph Neural Network , 2020, NeurIPS.

[24]  S. Teipel,et al.  Multimodal analysis of functional and structural disconnection in Alzheimer's disease using multiple kernel SVM , 2015, Human brain mapping.

[25]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[26]  Raymond Chi-Wing Wong,et al.  When Do GNNs Work: Understanding and Improving Neighborhood Aggregation , 2020, IJCAI.

[27]  Yaron Lipman,et al.  Invariant and Equivariant Graph Networks , 2018, ICLR.

[28]  Juntang Zhuang,et al.  BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis , 2020, bioRxiv.

[29]  Jure Leskovec,et al.  How Powerful are Graph Neural Networks? , 2018, ICLR.

[30]  Susan M. Courtney,et al.  Probabilistic White Matter Atlases of Human Auditory, Basal Ganglia, Language, Precuneus, Sensorimotor, Visual and Visuospatial Networks , 2017, Frontiers in human neuroscience.

[31]  M. Greicius,et al.  Decoding subject-driven cognitive states with whole-brain connectivity patterns. , 2012, Cerebral cortex.

[32]  Pietro Liò,et al.  Graph Attention Networks , 2017, ICLR.

[33]  Z. Yao,et al.  Decreased dynamism of overlapping brain sub-networks in Major Depressive Disorder. , 2020, Journal of psychiatric research.

[34]  Jan Eric Lenssen,et al.  Fast Graph Representation Learning with PyTorch Geometric , 2019, ArXiv.

[35]  Pietro Liò,et al.  Deep Graph Infomax , 2018, ICLR.

[36]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[37]  S. Francis,et al.  Parietal lobe and disorganisation syndrome in schizophrenia and psychotic bipolar disorder: A bimodal connectivity study , 2020, Psychiatry Research: Neuroimaging.

[38]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[39]  Philip S. Yu,et al.  Multi-View Multi-Graph Embedding for Brain Network Clustering Analysis , 2018, AAAI.

[40]  Carl Yang,et al.  Heterogeneous Network Representation Learning: A Unified Framework With Survey and Benchmark , 2020, IEEE Transactions on Knowledge and Data Engineering.

[41]  A. Simmons,et al.  Stability of graph theoretical measures in structural brain networks in Alzheimer’s disease , 2018, Scientific Reports.

[42]  A. Laird,et al.  HIV infection is linked with reduced error-related default mode network suppression and poorer medication management abilities , 2021, Progress in Neuro-Psychopharmacology and Biological Psychiatry.

[43]  Dimitrios Pantazis,et al.  A new Graph Gaussian embedding method for analyzing the effects of cognitive training , 2020, PLoS Comput. Biol..

[44]  Michael Sejr Schlichtkrull,et al.  Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking , 2020, ArXiv.