Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface?

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

Abstract

Convolutional Neural Network (CNN) has outperformed many traditional linear and polynomial classifiers in various domains. CNN and other deep learning methods have gained attention in the Brain-Computer Interface (BCI) domain also. Here, we investigate a CNN-based model with two different optimizers for reducing error of classification of brain states using EEG motor imagery data. Two different optimizers namely stochastic gradient descent (SGD) and adaptive momentum (Adam) are investigated for increasing classification accuracy using the well-known BCI competition IV 2b dataset. The study was conducted to investigate the feasibility of a single deep learning model for all subjects without compromising on information decoding rate for any of the BCI participants. Using two different models, mean cross-validation accuracy of 80.32% (±2.2) was achieved across participants, which is significantly higher (p<0.05) compared to a state-of-the-art deep learning approach.
LanguageEnglish
Title of host publicationProc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy
Number of pages5
VolumeIEEE
Publication statusAccepted/In press - 23 Jun 2019
Event2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) - Bari, Italy
Duration: 6 Oct 20199 Oct 2019

Conference

Conference2019 IEEE International Conference on Systems, Man and Cybernetics (SMC)
CountryItaly
CityBari
Period6/10/199/10/19

Fingerprint

Brain computer interface
Electroencephalography
Neural networks
Decoding
Brain
Momentum
Classifiers
Polynomials
Deep learning

Keywords

  • Convolutional Neural Network (CNN)
  • BCI
  • stochastic gradient descent (SGD)
  • adaptive momentum (Adam)
  • Deep Learning

Cite this

Roy, S., McCreadie, K., & Prasad, G. (Accepted/In press). Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface? In Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy (Vol. IEEE)
Roy, Sujit ; McCreadie, Karl ; Prasad, Girijesh. / Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface?. Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy. Vol. IEEE 2019.
@inproceedings{2e61b1acfcfb4710a22a683bd4bf0d70,
title = "Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface?",
abstract = "Convolutional Neural Network (CNN) has outperformed many traditional linear and polynomial classifiers in various domains. CNN and other deep learning methods have gained attention in the Brain-Computer Interface (BCI) domain also. Here, we investigate a CNN-based model with two different optimizers for reducing error of classification of brain states using EEG motor imagery data. Two different optimizers namely stochastic gradient descent (SGD) and adaptive momentum (Adam) are investigated for increasing classification accuracy using the well-known BCI competition IV 2b dataset. The study was conducted to investigate the feasibility of a single deep learning model for all subjects without compromising on information decoding rate for any of the BCI participants. Using two different models, mean cross-validation accuracy of 80.32{\%} (±2.2) was achieved across participants, which is significantly higher (p<0.05) compared to a state-of-the-art deep learning approach.",
keywords = "Convolutional Neural Network (CNN), BCI, stochastic gradient descent (SGD), adaptive momentum (Adam), Deep Learning",
author = "Sujit Roy and Karl McCreadie and Girijesh Prasad",
year = "2019",
month = "6",
day = "23",
language = "English",
volume = "IEEE",
booktitle = "Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy",

}

Roy, S, McCreadie, K & Prasad, G 2019, Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface? in Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy. vol. IEEE, 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC), Bari, Italy, 6/10/19.

Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface? / Roy, Sujit; McCreadie, Karl; Prasad, Girijesh.

Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy. Vol. IEEE 2019.

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

TY - GEN

T1 - Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface?

AU - Roy, Sujit

AU - McCreadie, Karl

AU - Prasad, Girijesh

PY - 2019/6/23

Y1 - 2019/6/23

N2 - Convolutional Neural Network (CNN) has outperformed many traditional linear and polynomial classifiers in various domains. CNN and other deep learning methods have gained attention in the Brain-Computer Interface (BCI) domain also. Here, we investigate a CNN-based model with two different optimizers for reducing error of classification of brain states using EEG motor imagery data. Two different optimizers namely stochastic gradient descent (SGD) and adaptive momentum (Adam) are investigated for increasing classification accuracy using the well-known BCI competition IV 2b dataset. The study was conducted to investigate the feasibility of a single deep learning model for all subjects without compromising on information decoding rate for any of the BCI participants. Using two different models, mean cross-validation accuracy of 80.32% (±2.2) was achieved across participants, which is significantly higher (p<0.05) compared to a state-of-the-art deep learning approach.

AB - Convolutional Neural Network (CNN) has outperformed many traditional linear and polynomial classifiers in various domains. CNN and other deep learning methods have gained attention in the Brain-Computer Interface (BCI) domain also. Here, we investigate a CNN-based model with two different optimizers for reducing error of classification of brain states using EEG motor imagery data. Two different optimizers namely stochastic gradient descent (SGD) and adaptive momentum (Adam) are investigated for increasing classification accuracy using the well-known BCI competition IV 2b dataset. The study was conducted to investigate the feasibility of a single deep learning model for all subjects without compromising on information decoding rate for any of the BCI participants. Using two different models, mean cross-validation accuracy of 80.32% (±2.2) was achieved across participants, which is significantly higher (p<0.05) compared to a state-of-the-art deep learning approach.

KW - Convolutional Neural Network (CNN)

KW - BCI

KW - stochastic gradient descent (SGD)

KW - adaptive momentum (Adam)

KW - Deep Learning

UR - https://conf.papercept.net/conferences/conferences/SMC19/program/SMC19_ContentListWeb_2.html#moct8

M3 - Conference contribution

VL - IEEE

BT - Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy

ER -

Roy S, McCreadie K, Prasad G. Can a Single Model Deep Learning Approach Enhance Classification Accuracy of an EEG-based Brain-Computer Interface? In Proc. 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) October 6-9, 2019, Bari, Italy. Vol. IEEE. 2019