Behavioral state classification in epileptic brain using intracranial electrophysiology

Vaclav Kremen, Juliano J. Duque, Benjamin Brinkmann, Brent M. Berry, Michal T. Kucewicz, Fatemeh Khadjevand, Jamie Van Gompel, Squire Matthew Stead, Erik K St Louis, Gregory Alan Worrell

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Objective. Automated behavioral state classification can benefit next generation implantable epilepsy devices. In this study we explored the feasibility of automated awake (AW) and slow wave sleep (SWS) classification using wide bandwidth intracranial EEG (iEEG) in patients undergoing evaluation for epilepsy surgery. Approach. Data from seven patients (age , 4 women) who underwent intracranial depth electrode implantation for iEEG monitoring were included. Spectral power features (0.1-600 Hz) spanning several frequency bands from a single electrode were used to train and test a support vector machine classifier. Main results. Classification accuracy of 97.8 ± 0.3% (normal tissue) and 89.4 ± 0.8% (epileptic tissue) across seven subjects using multiple spectral power features from a single electrode was achieved. Spectral power features from electrodes placed in normal temporal neocortex were found to be more useful (accuracy 90.8 ± 0.8%) for sleep-wake state classification than electrodes located in normal hippocampus (87.1 ± 1.6%). Spectral power in high frequency band features (Ripple (80-250 Hz), Fast Ripple (250-600 Hz)) showed comparable performance for AW and SWS classification as the best performing Berger bands (Alpha, Beta, low Gamma) with accuracy ≥ 90% using a single electrode contact and single spectral feature. Significance. Automated classification of wake and SWS should prove useful for future implantable epilepsy devices with limited computational power, memory, and number of electrodes. Applications include quantifying patient sleep patterns and behavioral state dependent detection, prediction, and electrical stimulation therapies.

Original languageEnglish (US)
Article number026001
JournalJournal of Neural Engineering
Volume14
Issue number2
DOIs
StatePublished - Jan 19 2017

Fingerprint

Electrophysiology
Brain
Electrodes
Sleep
Epilepsy
Electroencephalography
Frequency bands
Tissue
Electric Stimulation Therapy
Equipment and Supplies
Neocortex
Feasibility Studies
Surgery
Support vector machines
Hippocampus
Classifiers
Power (Psychology)
Bandwidth
Data storage equipment
Monitoring

Keywords

  • behavioral states
  • classification
  • intracranial EEG
  • machine learning
  • sleep staging

ASJC Scopus subject areas

  • Biomedical Engineering
  • Cellular and Molecular Neuroscience

Cite this

Behavioral state classification in epileptic brain using intracranial electrophysiology. / Kremen, Vaclav; Duque, Juliano J.; Brinkmann, Benjamin; Berry, Brent M.; Kucewicz, Michal T.; Khadjevand, Fatemeh; Van Gompel, Jamie; Stead, Squire Matthew; St Louis, Erik K; Worrell, Gregory Alan.

In: Journal of Neural Engineering, Vol. 14, No. 2, 026001, 19.01.2017.

Research output: Contribution to journalArticle

@article{b6c5c2070b4f43ab9064420033922095,
title = "Behavioral state classification in epileptic brain using intracranial electrophysiology",
abstract = "Objective. Automated behavioral state classification can benefit next generation implantable epilepsy devices. In this study we explored the feasibility of automated awake (AW) and slow wave sleep (SWS) classification using wide bandwidth intracranial EEG (iEEG) in patients undergoing evaluation for epilepsy surgery. Approach. Data from seven patients (age , 4 women) who underwent intracranial depth electrode implantation for iEEG monitoring were included. Spectral power features (0.1-600 Hz) spanning several frequency bands from a single electrode were used to train and test a support vector machine classifier. Main results. Classification accuracy of 97.8 ± 0.3{\%} (normal tissue) and 89.4 ± 0.8{\%} (epileptic tissue) across seven subjects using multiple spectral power features from a single electrode was achieved. Spectral power features from electrodes placed in normal temporal neocortex were found to be more useful (accuracy 90.8 ± 0.8{\%}) for sleep-wake state classification than electrodes located in normal hippocampus (87.1 ± 1.6{\%}). Spectral power in high frequency band features (Ripple (80-250 Hz), Fast Ripple (250-600 Hz)) showed comparable performance for AW and SWS classification as the best performing Berger bands (Alpha, Beta, low Gamma) with accuracy ≥ 90{\%} using a single electrode contact and single spectral feature. Significance. Automated classification of wake and SWS should prove useful for future implantable epilepsy devices with limited computational power, memory, and number of electrodes. Applications include quantifying patient sleep patterns and behavioral state dependent detection, prediction, and electrical stimulation therapies.",
keywords = "behavioral states, classification, intracranial EEG, machine learning, sleep staging",
author = "Vaclav Kremen and Duque, {Juliano J.} and Benjamin Brinkmann and Berry, {Brent M.} and Kucewicz, {Michal T.} and Fatemeh Khadjevand and {Van Gompel}, Jamie and Stead, {Squire Matthew} and {St Louis}, {Erik K} and Worrell, {Gregory Alan}",
year = "2017",
month = "1",
day = "19",
doi = "10.1088/1741-2552/aa5688",
language = "English (US)",
volume = "14",
journal = "Journal of Neural Engineering",
issn = "1741-2560",
publisher = "IOP Publishing Ltd.",
number = "2",

}

TY - JOUR

T1 - Behavioral state classification in epileptic brain using intracranial electrophysiology

AU - Kremen, Vaclav

AU - Duque, Juliano J.

AU - Brinkmann, Benjamin

AU - Berry, Brent M.

AU - Kucewicz, Michal T.

AU - Khadjevand, Fatemeh

AU - Van Gompel, Jamie

AU - Stead, Squire Matthew

AU - St Louis, Erik K

AU - Worrell, Gregory Alan

PY - 2017/1/19

Y1 - 2017/1/19

N2 - Objective. Automated behavioral state classification can benefit next generation implantable epilepsy devices. In this study we explored the feasibility of automated awake (AW) and slow wave sleep (SWS) classification using wide bandwidth intracranial EEG (iEEG) in patients undergoing evaluation for epilepsy surgery. Approach. Data from seven patients (age , 4 women) who underwent intracranial depth electrode implantation for iEEG monitoring were included. Spectral power features (0.1-600 Hz) spanning several frequency bands from a single electrode were used to train and test a support vector machine classifier. Main results. Classification accuracy of 97.8 ± 0.3% (normal tissue) and 89.4 ± 0.8% (epileptic tissue) across seven subjects using multiple spectral power features from a single electrode was achieved. Spectral power features from electrodes placed in normal temporal neocortex were found to be more useful (accuracy 90.8 ± 0.8%) for sleep-wake state classification than electrodes located in normal hippocampus (87.1 ± 1.6%). Spectral power in high frequency band features (Ripple (80-250 Hz), Fast Ripple (250-600 Hz)) showed comparable performance for AW and SWS classification as the best performing Berger bands (Alpha, Beta, low Gamma) with accuracy ≥ 90% using a single electrode contact and single spectral feature. Significance. Automated classification of wake and SWS should prove useful for future implantable epilepsy devices with limited computational power, memory, and number of electrodes. Applications include quantifying patient sleep patterns and behavioral state dependent detection, prediction, and electrical stimulation therapies.

AB - Objective. Automated behavioral state classification can benefit next generation implantable epilepsy devices. In this study we explored the feasibility of automated awake (AW) and slow wave sleep (SWS) classification using wide bandwidth intracranial EEG (iEEG) in patients undergoing evaluation for epilepsy surgery. Approach. Data from seven patients (age , 4 women) who underwent intracranial depth electrode implantation for iEEG monitoring were included. Spectral power features (0.1-600 Hz) spanning several frequency bands from a single electrode were used to train and test a support vector machine classifier. Main results. Classification accuracy of 97.8 ± 0.3% (normal tissue) and 89.4 ± 0.8% (epileptic tissue) across seven subjects using multiple spectral power features from a single electrode was achieved. Spectral power features from electrodes placed in normal temporal neocortex were found to be more useful (accuracy 90.8 ± 0.8%) for sleep-wake state classification than electrodes located in normal hippocampus (87.1 ± 1.6%). Spectral power in high frequency band features (Ripple (80-250 Hz), Fast Ripple (250-600 Hz)) showed comparable performance for AW and SWS classification as the best performing Berger bands (Alpha, Beta, low Gamma) with accuracy ≥ 90% using a single electrode contact and single spectral feature. Significance. Automated classification of wake and SWS should prove useful for future implantable epilepsy devices with limited computational power, memory, and number of electrodes. Applications include quantifying patient sleep patterns and behavioral state dependent detection, prediction, and electrical stimulation therapies.

KW - behavioral states

KW - classification

KW - intracranial EEG

KW - machine learning

KW - sleep staging

UR - http://www.scopus.com/inward/record.url?scp=85015716549&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015716549&partnerID=8YFLogxK

U2 - 10.1088/1741-2552/aa5688

DO - 10.1088/1741-2552/aa5688

M3 - Article

C2 - 28050973

AN - SCOPUS:85015716549

VL - 14

JO - Journal of Neural Engineering

JF - Journal of Neural Engineering

SN - 1741-2560

IS - 2

M1 - 026001

ER -