Integrating Convolutional Neural Networks and Probabilistic Graphical Modeling for Epileptic Seizure Detection in Multichannel EEG

Jeff Craley, Emily Johnson, Archana Venkataraman

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Manual seizure detection in clinical electroencephalography (EEG) is time consuming and requires extensive training. In addition, the seizure origin and spreading pattern is valuable for therapeutic planning but cannot always be manually disambiguated. Prior work in automated seizure detection has focused on engineering new features that better capture the seizure activity. However, these methods ignore crucial information in the data and are not sensitive enough to track the seizure propagation. In this work we introduce a hybrid Probabilistic Graphical Model-Convolutional Neural Network (PGM-CNN) for seizure tracking in multichannel EEG. Our model leverages the power of deep learning for data driven analysis of the raw EEG time series while retaining clinically relevant information through the latent PGM prior. We validate our hybrid model on clinical EEG data from two hospitals with distinct patient populations. Our system achieves better detection performance than baseline methods, which exclusively use PGMs or neural networks.

Original languageEnglish (US)
Title of host publicationInformation Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings
EditorsSiqi Bao, Albert C.S. Chung, James C. Gee, Paul A. Yushkevich
PublisherSpringer Verlag
Pages291-303
Number of pages13
ISBN (Print)9783030203504
DOIs
StatePublished - Jan 1 2019
Event26th International Conference on Information Processing in Medical Imaging, IPMI 2019 - Hong Kong, China
Duration: Jun 2 2019Jun 7 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11492 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference26th International Conference on Information Processing in Medical Imaging, IPMI 2019
CountryChina
CityHong Kong
Period6/2/196/7/19

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Integrating Convolutional Neural Networks and Probabilistic Graphical Modeling for Epileptic Seizure Detection in Multichannel EEG'. Together they form a unique fingerprint.

Cite this