Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network

Christian Oliva; Vinicio Changoluisa; Francisco B. Rodríguez; Luis F. Lago-Fernández

doi:10.1007/978-3-031-34111-3_9

Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network

Christian Oliva, Vinicio Changoluisa, Francisco B. Rodríguez, Luis F. Lago-Fernández

Electronics and Telematics Research Group (GIETEC)

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

1 Scopus citations

Abstract

We introduce a novel approach for detecting the sample-level temporal structure of P300 event-related potentials. It consists of extracting the most informative neurons from a Recurrent Neural Network and building a post-validation neural ensemble (PVNE). The weights connecting the recurrent and the output layers are used to rank the recurrent neurons according to their relevance when generating the network’s output. A set of neurons is selected according to their positions in this ranking, and their individual predictions are then combined to obtain the final model’s output. This procedure discards neurons whose role could be more related to maintaining the network’s hidden state than to detecting the P300 events, with an overall performance increase. The use of L1 regularization notably emphasizes this effect. We compare the performance of this approach with both Elman and LSTM RNNs and show that the PVNE is able to detect the sample-level temporal structure of P300 event-related potentials, outperforming the standard models. Sample-level prediction also allows for real-time monitoring of the EEG signal generation related to ERPs.

Original language	English
Title of host publication	Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings
Editors	Ilias Maglogiannis, Lazaros Iliadis, John MacIntyre, Manuel Dominguez
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	90-101
Number of pages	12
ISBN (Print)	9783031341106
DOIs	https://doi.org/10.1007/978-3-031-34111-3_9
State	Published - 2023
Event	19th IFIP WG 12.5 International Conference on Artificial Intelligence Applications and Innovations, AIAI 2023 - León, Spain Duration: 14 Jun 2023 → 17 Jun 2023

Publication series

Name	IFIP Advances in Information and Communication Technology
Volume	675 IFIP
ISSN (Print)	1868-4238
ISSN (Electronic)	1868-422X

Conference

Conference	19th IFIP WG 12.5 International Conference on Artificial Intelligence Applications and Innovations, AIAI 2023
Country/Territory	Spain
City	León
Period	14/06/23 → 17/06/23

Bibliographical note

Funding Information:
This work has been partially funded by Spanish project PID2020-114867RB-I00, (MCIN/AEI and ERDF-“A way of making Europe”), Uni-versidad Politécnica Salesiana 034-02-2022-03-31 and by Predoctoral Research Grants 2015-AR2Q9086 of the Government of Ecuador through SENESCYT.

Publisher Copyright:
© 2023, IFIP International Federation for Information Processing.

Keywords

Bayesian Linear Discriminant Analysis
Brain-machine interface
Elman RNN
Inter- and intra-subject variability
LSTM
Neurons interpretability
Neurons relevance
Output Neurons Ensemble

Access to Document

10.1007/978-3-031-34111-3_9

Cite this

Oliva, C., Changoluisa, V., Rodríguez, F. B., & Lago-Fernández, L. F. (2023). Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network. In I. Maglogiannis, L. Iliadis, J. MacIntyre, & M. Dominguez (Eds.), Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings (pp. 90-101). (IFIP Advances in Information and Communication Technology; Vol. 675 IFIP). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-34111-3_9

Oliva, Christian ; Changoluisa, Vinicio ; Rodríguez, Francisco B. et al. / Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network. Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings. editor / Ilias Maglogiannis ; Lazaros Iliadis ; John MacIntyre ; Manuel Dominguez. Springer Science and Business Media Deutschland GmbH, 2023. pp. 90-101 (IFIP Advances in Information and Communication Technology).

@inproceedings{87a3b8a2becd43409ebf85fbaa17dea1,

title = "Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network",

abstract = "We introduce a novel approach for detecting the sample-level temporal structure of P300 event-related potentials. It consists of extracting the most informative neurons from a Recurrent Neural Network and building a post-validation neural ensemble (PVNE). The weights connecting the recurrent and the output layers are used to rank the recurrent neurons according to their relevance when generating the network{\textquoteright}s output. A set of neurons is selected according to their positions in this ranking, and their individual predictions are then combined to obtain the final model{\textquoteright}s output. This procedure discards neurons whose role could be more related to maintaining the network{\textquoteright}s hidden state than to detecting the P300 events, with an overall performance increase. The use of L1 regularization notably emphasizes this effect. We compare the performance of this approach with both Elman and LSTM RNNs and show that the PVNE is able to detect the sample-level temporal structure of P300 event-related potentials, outperforming the standard models. Sample-level prediction also allows for real-time monitoring of the EEG signal generation related to ERPs.",

keywords = "Bayesian Linear Discriminant Analysis, Brain-machine interface, Elman RNN, Inter- and intra-subject variability, LSTM, Neurons interpretability, Neurons relevance, Output Neurons Ensemble",

author = "Christian Oliva and Vinicio Changoluisa and Rodr{\'i}guez, {Francisco B.} and Lago-Fern{\'a}ndez, {Luis F.}",

note = "Publisher Copyright: {\textcopyright} 2023, IFIP International Federation for Information Processing.; 19th IFIP WG 12.5 International Conference on Artificial Intelligence Applications and Innovations, AIAI 2023 ; Conference date: 14-06-2023 Through 17-06-2023",

year = "2023",

doi = "10.1007/978-3-031-34111-3_9",

language = "English",

isbn = "9783031341106",

series = "IFIP Advances in Information and Communication Technology",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "90--101",

editor = "Ilias Maglogiannis and Lazaros Iliadis and John MacIntyre and Manuel Dominguez",

booktitle = "Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings",

address = "Germany",

}

Oliva, C, Changoluisa, V, Rodríguez, FB & Lago-Fernández, LF 2023, Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network. in I Maglogiannis, L Iliadis, J MacIntyre & M Dominguez (eds), Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings. IFIP Advances in Information and Communication Technology, vol. 675 IFIP, Springer Science and Business Media Deutschland GmbH, pp. 90-101, 19th IFIP WG 12.5 International Conference on Artificial Intelligence Applications and Innovations, AIAI 2023, León, Spain, 14/06/23. https://doi.org/10.1007/978-3-031-34111-3_9

Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network. / Oliva, Christian; Changoluisa, Vinicio; Rodríguez, Francisco B. et al.
Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings. ed. / Ilias Maglogiannis; Lazaros Iliadis; John MacIntyre; Manuel Dominguez. Springer Science and Business Media Deutschland GmbH, 2023. p. 90-101 (IFIP Advances in Information and Communication Technology; Vol. 675 IFIP).

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

TY - GEN

T1 - Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network

AU - Oliva, Christian

AU - Changoluisa, Vinicio

AU - Rodríguez, Francisco B.

AU - Lago-Fernández, Luis F.

PY - 2023

Y1 - 2023

N2 - We introduce a novel approach for detecting the sample-level temporal structure of P300 event-related potentials. It consists of extracting the most informative neurons from a Recurrent Neural Network and building a post-validation neural ensemble (PVNE). The weights connecting the recurrent and the output layers are used to rank the recurrent neurons according to their relevance when generating the network’s output. A set of neurons is selected according to their positions in this ranking, and their individual predictions are then combined to obtain the final model’s output. This procedure discards neurons whose role could be more related to maintaining the network’s hidden state than to detecting the P300 events, with an overall performance increase. The use of L1 regularization notably emphasizes this effect. We compare the performance of this approach with both Elman and LSTM RNNs and show that the PVNE is able to detect the sample-level temporal structure of P300 event-related potentials, outperforming the standard models. Sample-level prediction also allows for real-time monitoring of the EEG signal generation related to ERPs.

AB - We introduce a novel approach for detecting the sample-level temporal structure of P300 event-related potentials. It consists of extracting the most informative neurons from a Recurrent Neural Network and building a post-validation neural ensemble (PVNE). The weights connecting the recurrent and the output layers are used to rank the recurrent neurons according to their relevance when generating the network’s output. A set of neurons is selected according to their positions in this ranking, and their individual predictions are then combined to obtain the final model’s output. This procedure discards neurons whose role could be more related to maintaining the network’s hidden state than to detecting the P300 events, with an overall performance increase. The use of L1 regularization notably emphasizes this effect. We compare the performance of this approach with both Elman and LSTM RNNs and show that the PVNE is able to detect the sample-level temporal structure of P300 event-related potentials, outperforming the standard models. Sample-level prediction also allows for real-time monitoring of the EEG signal generation related to ERPs.

KW - Bayesian Linear Discriminant Analysis

KW - Brain-machine interface

KW - Elman RNN

KW - Inter- and intra-subject variability

KW - LSTM

KW - Neurons interpretability

KW - Neurons relevance

KW - Output Neurons Ensemble

UR - http://www.scopus.com/inward/record.url?scp=85163393300&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-34111-3_9

DO - 10.1007/978-3-031-34111-3_9

M3 - Conference contribution

AN - SCOPUS:85163393300

SN - 9783031341106

T3 - IFIP Advances in Information and Communication Technology

SP - 90

EP - 101

BT - Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings

A2 - Maglogiannis, Ilias

A2 - Iliadis, Lazaros

A2 - MacIntyre, John

A2 - Dominguez, Manuel

PB - Springer Science and Business Media Deutschland GmbH

T2 - 19th IFIP WG 12.5 International Conference on Artificial Intelligence Applications and Innovations, AIAI 2023

Y2 - 14 June 2023 through 17 June 2023

ER -

Oliva C, Changoluisa V, Rodríguez FB, Lago-Fernández LF. Detecting P300-ERPs Building a Post-validation Neural Ensemble with Informative Neurons from a Recurrent Neural Network. In Maglogiannis I, Iliadis L, MacIntyre J, Dominguez M, editors, Artificial Intelligence Applications and Innovations - 19th IFIP WG 12.5 International Conference, AIAI 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 90-101. (IFIP Advances in Information and Communication Technology). doi: 10.1007/978-3-031-34111-3_9