Exploiting Graphoelements and Convolutional Neural Networks with Long Short Term Memory for Classification of the Human Electroencephalogram

P. Nejedly, V. Kremen, V. Sladky, J. Cimbalnik, P. Klimes, F. Plesinger, I. Viscor, M. Pail, J. Halamek, Benjamin Brinkmann, M. Brazdil, P. Jurak, Gregory Alan Worrell

Research output: Contribution to journalArticle

Abstract

The electroencephalogram (EEG) is a cornerstone of neurophysiological research and clinical neurology. Historically, the classification of EEG as showing normal physiological or abnormal pathological activity has been performed by expert visual review. The potential value of unbiased, automated EEG classification has long been recognized, and in recent years the application of machine learning methods has received significant attention. A variety of solutions using convolutional neural networks (CNN) for EEG classification have emerged with impressive results. However, interpretation of CNN results and their connection with underlying basic electrophysiology has been unclear. This paper proposes a CNN architecture, which enables interpretation of intracranial EEG (iEEG) transients driving classification of brain activity as normal, pathological or artifactual. The goal is accomplished using CNN with long short-term memory (LSTM). We show that the method allows the visualization of iEEG graphoelements with the highest contribution to the final classification result using a classification heatmap and thus enables review of the raw iEEG data and interpret the decision of the model by electrophysiology means.

Original languageEnglish (US)
Article number11383
JournalScientific reports
Volume9
Issue number1
DOIs
StatePublished - Dec 1 2019

Fingerprint

Long-Term Memory
Short-Term Memory
Electroencephalography
Electrophysiology
Neurology
Brain
Research

ASJC Scopus subject areas

  • General

Cite this

Exploiting Graphoelements and Convolutional Neural Networks with Long Short Term Memory for Classification of the Human Electroencephalogram. / Nejedly, P.; Kremen, V.; Sladky, V.; Cimbalnik, J.; Klimes, P.; Plesinger, F.; Viscor, I.; Pail, M.; Halamek, J.; Brinkmann, Benjamin; Brazdil, M.; Jurak, P.; Worrell, Gregory Alan.

In: Scientific reports, Vol. 9, No. 1, 11383, 01.12.2019.

Research output: Contribution to journalArticle

Nejedly, P, Kremen, V, Sladky, V, Cimbalnik, J, Klimes, P, Plesinger, F, Viscor, I, Pail, M, Halamek, J, Brinkmann, B, Brazdil, M, Jurak, P & Worrell, GA 2019, 'Exploiting Graphoelements and Convolutional Neural Networks with Long Short Term Memory for Classification of the Human Electroencephalogram', Scientific reports, vol. 9, no. 1, 11383. https://doi.org/10.1038/s41598-019-47854-6
Nejedly, P. ; Kremen, V. ; Sladky, V. ; Cimbalnik, J. ; Klimes, P. ; Plesinger, F. ; Viscor, I. ; Pail, M. ; Halamek, J. ; Brinkmann, Benjamin ; Brazdil, M. ; Jurak, P. ; Worrell, Gregory Alan. / Exploiting Graphoelements and Convolutional Neural Networks with Long Short Term Memory for Classification of the Human Electroencephalogram. In: Scientific reports. 2019 ; Vol. 9, No. 1.
@article{da7261198890452b9865bde073862bf3,
title = "Exploiting Graphoelements and Convolutional Neural Networks with Long Short Term Memory for Classification of the Human Electroencephalogram",
abstract = "The electroencephalogram (EEG) is a cornerstone of neurophysiological research and clinical neurology. Historically, the classification of EEG as showing normal physiological or abnormal pathological activity has been performed by expert visual review. The potential value of unbiased, automated EEG classification has long been recognized, and in recent years the application of machine learning methods has received significant attention. A variety of solutions using convolutional neural networks (CNN) for EEG classification have emerged with impressive results. However, interpretation of CNN results and their connection with underlying basic electrophysiology has been unclear. This paper proposes a CNN architecture, which enables interpretation of intracranial EEG (iEEG) transients driving classification of brain activity as normal, pathological or artifactual. The goal is accomplished using CNN with long short-term memory (LSTM). We show that the method allows the visualization of iEEG graphoelements with the highest contribution to the final classification result using a classification heatmap and thus enables review of the raw iEEG data and interpret the decision of the model by electrophysiology means.",
author = "P. Nejedly and V. Kremen and V. Sladky and J. Cimbalnik and P. Klimes and F. Plesinger and I. Viscor and M. Pail and J. Halamek and Benjamin Brinkmann and M. Brazdil and P. Jurak and Worrell, {Gregory Alan}",
year = "2019",
month = "12",
day = "1",
doi = "10.1038/s41598-019-47854-6",
language = "English (US)",
volume = "9",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - Exploiting Graphoelements and Convolutional Neural Networks with Long Short Term Memory for Classification of the Human Electroencephalogram

AU - Nejedly, P.

AU - Kremen, V.

AU - Sladky, V.

AU - Cimbalnik, J.

AU - Klimes, P.

AU - Plesinger, F.

AU - Viscor, I.

AU - Pail, M.

AU - Halamek, J.

AU - Brinkmann, Benjamin

AU - Brazdil, M.

AU - Jurak, P.

AU - Worrell, Gregory Alan

PY - 2019/12/1

Y1 - 2019/12/1

N2 - The electroencephalogram (EEG) is a cornerstone of neurophysiological research and clinical neurology. Historically, the classification of EEG as showing normal physiological or abnormal pathological activity has been performed by expert visual review. The potential value of unbiased, automated EEG classification has long been recognized, and in recent years the application of machine learning methods has received significant attention. A variety of solutions using convolutional neural networks (CNN) for EEG classification have emerged with impressive results. However, interpretation of CNN results and their connection with underlying basic electrophysiology has been unclear. This paper proposes a CNN architecture, which enables interpretation of intracranial EEG (iEEG) transients driving classification of brain activity as normal, pathological or artifactual. The goal is accomplished using CNN with long short-term memory (LSTM). We show that the method allows the visualization of iEEG graphoelements with the highest contribution to the final classification result using a classification heatmap and thus enables review of the raw iEEG data and interpret the decision of the model by electrophysiology means.

AB - The electroencephalogram (EEG) is a cornerstone of neurophysiological research and clinical neurology. Historically, the classification of EEG as showing normal physiological or abnormal pathological activity has been performed by expert visual review. The potential value of unbiased, automated EEG classification has long been recognized, and in recent years the application of machine learning methods has received significant attention. A variety of solutions using convolutional neural networks (CNN) for EEG classification have emerged with impressive results. However, interpretation of CNN results and their connection with underlying basic electrophysiology has been unclear. This paper proposes a CNN architecture, which enables interpretation of intracranial EEG (iEEG) transients driving classification of brain activity as normal, pathological or artifactual. The goal is accomplished using CNN with long short-term memory (LSTM). We show that the method allows the visualization of iEEG graphoelements with the highest contribution to the final classification result using a classification heatmap and thus enables review of the raw iEEG data and interpret the decision of the model by electrophysiology means.

UR - http://www.scopus.com/inward/record.url?scp=85070218024&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070218024&partnerID=8YFLogxK

U2 - 10.1038/s41598-019-47854-6

DO - 10.1038/s41598-019-47854-6

M3 - Article

VL - 9

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

IS - 1

M1 - 11383

ER -