Modeling Multivariate Biosignals With Graph Neural Networks and Structured State Space Models

Siyi Tang, Jared A Dunnmon, Qu Liangqiong, Khaled K Saab, Tina Baykaner, Christopher Lee-Messer, Daniel L Rubin
Proceedings of the Conference on Health, Inference, and Learning, PMLR 209:50-71, 2023.

Abstract

Multivariate biosignals are prevalent in many medical domains, such as electroencephalography, polysomnography, and electrocardiography. Modeling spatiotemporal dependencies in multivariate biosignals is challenging due to (1) long-range temporal dependencies and (2) complex spatial correlations between the electrodes. To address these challenges, we propose representing multivariate biosignals as time-dependent graphs and introduce \textsc{GraphS4mer}, a general graph neural network (GNN) architecture that improves performance on biosignal classification tasks by modeling spatiotemporal dependencies in biosignals. Specifically, (1) we leverage the Structured State Space architecture, a state-of-the-art deep sequence model, to capture long-range temporal dependencies in biosignals and (2) we propose a graph structure learning layer in \textsc{GraphS4mer} to learn dynamically evolving graph structures in the data. We evaluate our proposed model on three distinct biosignal classification tasks and show that \textsc{GraphS4mer} consistently improves over existing models, including (1) seizure detection from electroencephalographic signals, outperforming a previous GNN with self-supervised pre-training by 3.1 points in AUROC; (2) sleep staging from polysomnographic signals, a 4.1 points improvement in macro-F1 score compared to existing sleep staging models; and (3) 12-lead electrocardiogram classification, outperforming previous state-of-the-art models by 2.7 points in macro-F1 score.

Cite this Paper

BibTeX
@InProceedings{pmlr-v209-tang23a, title = {Modeling Multivariate Biosignals With Graph Neural Networks and Structured State Space Models}, author = {Tang, Siyi and Dunnmon, Jared A and Liangqiong, Qu and Saab, Khaled K and Baykaner, Tina and Lee-Messer, Christopher and Rubin, Daniel L}, booktitle = {Proceedings of the Conference on Health, Inference, and Learning}, pages = {50--71}, year = {2023}, editor = {Mortazavi, Bobak J. and Sarker, Tasmie and Beam, Andrew and Ho, Joyce C.}, volume = {209}, series = {Proceedings of Machine Learning Research}, month = {22 Jun--24 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v209/tang23a/tang23a.pdf}, url = {https://proceedings.mlr.press/v209/tang23a.html}, abstract = {Multivariate biosignals are prevalent in many medical domains, such as electroencephalography, polysomnography, and electrocardiography. Modeling spatiotemporal dependencies in multivariate biosignals is challenging due to (1) long-range temporal dependencies and (2) complex spatial correlations between the electrodes. To address these challenges, we propose representing multivariate biosignals as time-dependent graphs and introduce \textsc{GraphS4mer}, a general graph neural network (GNN) architecture that improves performance on biosignal classification tasks by modeling spatiotemporal dependencies in biosignals. Specifically, (1) we leverage the Structured State Space architecture, a state-of-the-art deep sequence model, to capture long-range temporal dependencies in biosignals and (2) we propose a graph structure learning layer in \textsc{GraphS4mer} to learn dynamically evolving graph structures in the data. We evaluate our proposed model on three distinct biosignal classification tasks and show that \textsc{GraphS4mer} consistently improves over existing models, including (1) seizure detection from electroencephalographic signals, outperforming a previous GNN with self-supervised pre-training by 3.1 points in AUROC; (2) sleep staging from polysomnographic signals, a 4.1 points improvement in macro-F1 score compared to existing sleep staging models; and (3) 12-lead electrocardiogram classification, outperforming previous state-of-the-art models by 2.7 points in macro-F1 score.} }
Endnote
%0 Conference Paper %T Modeling Multivariate Biosignals With Graph Neural Networks and Structured State Space Models %A Siyi Tang %A Jared A Dunnmon %A Qu Liangqiong %A Khaled K Saab %A Tina Baykaner %A Christopher Lee-Messer %A Daniel L Rubin %B Proceedings of the Conference on Health, Inference, and Learning %C Proceedings of Machine Learning Research %D 2023 %E Bobak J. Mortazavi %E Tasmie Sarker %E Andrew Beam %E Joyce C. Ho %F pmlr-v209-tang23a %I PMLR %P 50--71 %U https://proceedings.mlr.press/v209/tang23a.html %V 209 %X Multivariate biosignals are prevalent in many medical domains, such as electroencephalography, polysomnography, and electrocardiography. Modeling spatiotemporal dependencies in multivariate biosignals is challenging due to (1) long-range temporal dependencies and (2) complex spatial correlations between the electrodes. To address these challenges, we propose representing multivariate biosignals as time-dependent graphs and introduce \textsc{GraphS4mer}, a general graph neural network (GNN) architecture that improves performance on biosignal classification tasks by modeling spatiotemporal dependencies in biosignals. Specifically, (1) we leverage the Structured State Space architecture, a state-of-the-art deep sequence model, to capture long-range temporal dependencies in biosignals and (2) we propose a graph structure learning layer in \textsc{GraphS4mer} to learn dynamically evolving graph structures in the data. We evaluate our proposed model on three distinct biosignal classification tasks and show that \textsc{GraphS4mer} consistently improves over existing models, including (1) seizure detection from electroencephalographic signals, outperforming a previous GNN with self-supervised pre-training by 3.1 points in AUROC; (2) sleep staging from polysomnographic signals, a 4.1 points improvement in macro-F1 score compared to existing sleep staging models; and (3) 12-lead electrocardiogram classification, outperforming previous state-of-the-art models by 2.7 points in macro-F1 score.
APA
Tang, S., Dunnmon, J.A., Liangqiong, Q., Saab, K.K., Baykaner, T., Lee-Messer, C. & Rubin, D.L.. (2023). Modeling Multivariate Biosignals With Graph Neural Networks and Structured State Space Models. Proceedings of the Conference on Health, Inference, and Learning, in Proceedings of Machine Learning Research 209:50-71 Available from https://proceedings.mlr.press/v209/tang23a.html.

Related Material