Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness

K. Patlatzoglou; S. Chennu; M. Boly; Q. Noirhomme; V. Bonhomme; J.F. Brichant; O. Gosseries; S. Laureys; S. Wang; V. Yamamoto; J. Su; Y. Yang; E. Jones; L. Iasemidis; T. Mitchell

doi:10.1007/978-3-030-05587-5_21

Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness

K. Patlatzoglou^*, S. Chennu, M. Boly, Q. Noirhomme, V. Bonhomme, J.F. Brichant, O. Gosseries, S. Laureys, S. Wang, V. Yamamoto, J. Su, Y. Yang, E. Jones, L. Iasemidis, T. Mitchell

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

2 Citations (Web of Science)

Abstract

Despite the common use of anesthetics to modulate consciousness in the clinic, brain-based monitoring of consciousness is uncommon. We combined electroencephalographic measurement of brain activity with deep neural networks to automatically discriminate anesthetic states induced by propofol. Our results with leave-one-participant-out-cross-validation show that convolutional neural networks significantly outperform multilayer perceptrons in discrimination accuracy when working with raw time series. Perceptrons achieved comparable accuracy when provided with power spectral densities. These findings highlight the potential of deep convolutional networks for completely automatic extraction of useful spatio-temporo-spectral features from human EEG.

Original language	English
Title of host publication	Brain Informatics
Subtitle of host publication	International Conference, BI 2018, Arlington, TX, USA, December 7–9, 2018, Proceedings
Editors	Shouyi Wang, Vicky Yamamoto, Jianzhong Su, Yang Yang, Erick Jones, Leon Iasemidis, Tom Mitchell
Publisher	Springer
Pages	216-225
Number of pages	10
ISBN (Print)	9783030055868
DOIs	https://doi.org/10.1007/978-3-030-05587-5_21
Publication status	Published - 2018
Event	International Conference on Brain Informatics (BI) - Arlington, United States Duration: 7 Dec 2018 → 9 Dec 2018 http://www.wikicfp.com/cfp/servlet/event.showcfp?eventid=73741&copyownerid=94309

Publication series

Series	Lecture Notes in Artificial Intelligence
Volume	11309
ISSN	0302-9743

Conference

Conference	International Conference on Brain Informatics (BI)
Country/Territory	United States
City	Arlington
Period	7/12/18 → 9/12/18
Internet address	http://www.wikicfp.com/cfp/servlet/event.showcfp?eventid=73741&copyownerid=94309

Keywords

Consciousness
Anesthesia
EEG
Deep learning

Access to Document

10.1007/978-3-030-05587-5_21

Cite this

Patlatzoglou, K., Chennu, S., Boly, M., Noirhomme, Q., Bonhomme, V., Brichant, J. F., Gosseries, O., Laureys, S., Wang, S., Yamamoto, V., Su, J., Yang, Y., Jones, E., Iasemidis, L., & Mitchell, T. (2018). Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness. In S. Wang, V. Yamamoto, J. Su, Y. Yang, E. Jones, L. Iasemidis, & T. Mitchell (Eds.), Brain Informatics: International Conference, BI 2018, Arlington, TX, USA, December 7–9, 2018, Proceedings (pp. 216-225). Springer. Lecture Notes in Artificial Intelligence Vol. 11309 https://doi.org/10.1007/978-3-030-05587-5_21

Patlatzoglou, K. ; Chennu, S. ; Boly, M. et al. / Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness. Brain Informatics: International Conference, BI 2018, Arlington, TX, USA, December 7–9, 2018, Proceedings. editor / Shouyi Wang ; Vicky Yamamoto ; Jianzhong Su ; Yang Yang ; Erick Jones ; Leon Iasemidis ; Tom Mitchell. Springer, 2018. pp. 216-225 (Lecture Notes in Artificial Intelligence, Vol. 11309).

@inproceedings{b01fa4123fa14af1b4a446f668dbe4d5,

title = "Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness",

abstract = "Despite the common use of anesthetics to modulate consciousness in the clinic, brain-based monitoring of consciousness is uncommon. We combined electroencephalographic measurement of brain activity with deep neural networks to automatically discriminate anesthetic states induced by propofol. Our results with leave-one-participant-out-cross-validation show that convolutional neural networks significantly outperform multilayer perceptrons in discrimination accuracy when working with raw time series. Perceptrons achieved comparable accuracy when provided with power spectral densities. These findings highlight the potential of deep convolutional networks for completely automatic extraction of useful spatio-temporo-spectral features from human EEG.",

keywords = "Consciousness, Anesthesia, EEG, Deep learning",

author = "K. Patlatzoglou and S. Chennu and M. Boly and Q. Noirhomme and V. Bonhomme and J.F. Brichant and O. Gosseries and S. Laureys and S. Wang and V. Yamamoto and J. Su and Y. Yang and E. Jones and L. Iasemidis and T. Mitchell",

year = "2018",

doi = "10.1007/978-3-030-05587-5_21",

language = "English",

isbn = "9783030055868",

series = "Lecture Notes in Artificial Intelligence",

publisher = "Springer",

pages = "216--225",

editor = "Shouyi Wang and Vicky Yamamoto and Jianzhong Su and Yang Yang and Erick Jones and Leon Iasemidis and Tom Mitchell",

booktitle = "Brain Informatics",

address = "United States",

note = "International Conference on Brain Informatics (BI) ; Conference date: 07-12-2018 Through 09-12-2018",

url = "http://www.wikicfp.com/cfp/servlet/event.showcfp?eventid=73741&copyownerid=94309",

}

Patlatzoglou, K, Chennu, S, Boly, M, Noirhomme, Q, Bonhomme, V, Brichant, JF, Gosseries, O, Laureys, S, Wang, S, Yamamoto, V, Su, J, Yang, Y, Jones, E, Iasemidis, L & Mitchell, T 2018, Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness. in S Wang, V Yamamoto, J Su, Y Yang, E Jones, L Iasemidis & T Mitchell (eds), Brain Informatics: International Conference, BI 2018, Arlington, TX, USA, December 7–9, 2018, Proceedings. Springer, Lecture Notes in Artificial Intelligence, vol. 11309, pp. 216-225, International Conference on Brain Informatics (BI), Arlington, Texas, United States, 7/12/18. https://doi.org/10.1007/978-3-030-05587-5_21

Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness. / Patlatzoglou, K.; Chennu, S.; Boly, M. et al.

Brain Informatics: International Conference, BI 2018, Arlington, TX, USA, December 7–9, 2018, Proceedings. ed. / Shouyi Wang; Vicky Yamamoto; Jianzhong Su; Yang Yang; Erick Jones; Leon Iasemidis; Tom Mitchell. Springer, 2018. p. 216-225 (Lecture Notes in Artificial Intelligence, Vol. 11309).

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

TY - GEN

T1 - Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness

AU - Patlatzoglou, K.

AU - Chennu, S.

AU - Boly, M.

AU - Noirhomme, Q.

AU - Bonhomme, V.

AU - Brichant, J.F.

AU - Gosseries, O.

AU - Laureys, S.

AU - Wang, S.

AU - Yamamoto, V.

AU - Su, J.

AU - Yang, Y.

AU - Jones, E.

AU - Iasemidis, L.

AU - Mitchell, T.

PY - 2018

Y1 - 2018

N2 - Despite the common use of anesthetics to modulate consciousness in the clinic, brain-based monitoring of consciousness is uncommon. We combined electroencephalographic measurement of brain activity with deep neural networks to automatically discriminate anesthetic states induced by propofol. Our results with leave-one-participant-out-cross-validation show that convolutional neural networks significantly outperform multilayer perceptrons in discrimination accuracy when working with raw time series. Perceptrons achieved comparable accuracy when provided with power spectral densities. These findings highlight the potential of deep convolutional networks for completely automatic extraction of useful spatio-temporo-spectral features from human EEG.

AB - Despite the common use of anesthetics to modulate consciousness in the clinic, brain-based monitoring of consciousness is uncommon. We combined electroencephalographic measurement of brain activity with deep neural networks to automatically discriminate anesthetic states induced by propofol. Our results with leave-one-participant-out-cross-validation show that convolutional neural networks significantly outperform multilayer perceptrons in discrimination accuracy when working with raw time series. Perceptrons achieved comparable accuracy when provided with power spectral densities. These findings highlight the potential of deep convolutional networks for completely automatic extraction of useful spatio-temporo-spectral features from human EEG.

KW - Consciousness

KW - Anesthesia

KW - EEG

KW - Deep learning

U2 - 10.1007/978-3-030-05587-5_21

DO - 10.1007/978-3-030-05587-5_21

M3 - Conference article in proceeding

SN - 9783030055868

T3 - Lecture Notes in Artificial Intelligence

SP - 216

EP - 225

BT - Brain Informatics

A2 - Wang, Shouyi

A2 - Yamamoto, Vicky

A2 - Su, Jianzhong

A2 - Yang, Yang

A2 - Jones, Erick

A2 - Iasemidis, Leon

A2 - Mitchell, Tom

PB - Springer

T2 - International Conference on Brain Informatics (BI)

Y2 - 7 December 2018 through 9 December 2018

ER -

Patlatzoglou K, Chennu S, Boly M, Noirhomme Q, Bonhomme V, Brichant JF et al. Deep Neural Networks for Automatic Classification of Anesthetic-Induced Unconsciousness. In Wang S, Yamamoto V, Su J, Yang Y, Jones E, Iasemidis L, Mitchell T, editors, Brain Informatics: International Conference, BI 2018, Arlington, TX, USA, December 7–9, 2018, Proceedings. Springer. 2018. p. 216-225. (Lecture Notes in Artificial Intelligence, Vol. 11309). doi: 10.1007/978-3-030-05587-5_21