Classification of EEG features for prediction of working memory load

Anthony Abrantes, Elizabeth Comitz, Prithima Mosaly, Lukasz Mazur

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

Abstract

The objective of this research was to compare classification methods aimed at predicting working memory (WM) load. Electroencephalogram (EEG) data was collected from physicians while performing basic WM tasks and simulated medical scenarios. Data processing was performed to remove noise from the signal used for analysis (e.g., muscle activity, eye-blinks). The data from basic WM tasks was used to develop and test the four classification models (LASSO regression, support vector machines (SVM), nearest shrunken centroids (NSC), and iterated supervised principal components (ISPC) to predict a WM state indicative of physicians’ optimal performance. The na�ve misclassification rate was 19.74 %; LASSO and SVM outperformed this threshold: 18.10 and 12.21 % respectively). Both classification models had relatively high-specificity (LASSO: 97.2 %; SVM: 99.8 %); but relatively low-sensitivity LASSO: 20.7 %; SVM: 39.6 %). Results from simulated medical scenarios suggest that physicians were approximately 83 % of the time in the WM state that is likely indicative of optimal performance.

LanguageEnglish (US)
Title of host publicationAdvances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016
PublisherSpringer Verlag
Pages115-126
Number of pages12
ISBN (Print)9783319419466
DOIs
StatePublished - Jan 1 2017
EventInternational Conference on The Human Side of Service Engineering, 2016 - Walt Disney World, United States
Duration: Jul 27 2016Jul 31 2016

Publication series

NameAdvances in Intelligent Systems and Computing
Volume494
ISSN (Print)2194-5357

Other

OtherInternational Conference on The Human Side of Service Engineering, 2016
CountryUnited States
CityWalt Disney World
Period7/27/167/31/16

Fingerprint

Electroencephalography
Support vector machines
Data storage equipment
Muscle

Keywords

  • Classification methods
  • Cognitive workload (CWL)
  • Electronic medical records (EMR)
  • Machine learning
  • Physicians
  • Working memory (WM)

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Computer Science(all)

Cite this

Abrantes, A., Comitz, E., Mosaly, P., & Mazur, L. (2017). Classification of EEG features for prediction of working memory load. In Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016 (pp. 115-126). (Advances in Intelligent Systems and Computing; Vol. 494). Springer Verlag. DOI: 10.1007/978-3-319-41947-3_12

Classification of EEG features for prediction of working memory load. / Abrantes, Anthony; Comitz, Elizabeth; Mosaly, Prithima; Mazur, Lukasz.

Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016. Springer Verlag, 2017. p. 115-126 (Advances in Intelligent Systems and Computing; Vol. 494).

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

Abrantes, A, Comitz, E, Mosaly, P & Mazur, L 2017, Classification of EEG features for prediction of working memory load. in Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016. Advances in Intelligent Systems and Computing, vol. 494, Springer Verlag, pp. 115-126, International Conference on The Human Side of Service Engineering, 2016, Walt Disney World, United States, 7/27/16. DOI: 10.1007/978-3-319-41947-3_12
Abrantes A, Comitz E, Mosaly P, Mazur L. Classification of EEG features for prediction of working memory load. In Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016. Springer Verlag. 2017. p. 115-126. (Advances in Intelligent Systems and Computing). Available from, DOI: 10.1007/978-3-319-41947-3_12
Abrantes, Anthony ; Comitz, Elizabeth ; Mosaly, Prithima ; Mazur, Lukasz. / Classification of EEG features for prediction of working memory load. Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016. Springer Verlag, 2017. pp. 115-126 (Advances in Intelligent Systems and Computing).
@inproceedings{6d057cc50f7c4b72ab52a017de26ab52,
title = "Classification of EEG features for prediction of working memory load",
abstract = "The objective of this research was to compare classification methods aimed at predicting working memory (WM) load. Electroencephalogram (EEG) data was collected from physicians while performing basic WM tasks and simulated medical scenarios. Data processing was performed to remove noise from the signal used for analysis (e.g., muscle activity, eye-blinks). The data from basic WM tasks was used to develop and test the four classification models (LASSO regression, support vector machines (SVM), nearest shrunken centroids (NSC), and iterated supervised principal components (ISPC) to predict a WM state indicative of physicians’ optimal performance. The na{\"i}¿½ve misclassification rate was 19.74 {\%}; LASSO and SVM outperformed this threshold: 18.10 and 12.21 {\%} respectively). Both classification models had relatively high-specificity (LASSO: 97.2 {\%}; SVM: 99.8 {\%}); but relatively low-sensitivity LASSO: 20.7 {\%}; SVM: 39.6 {\%}). Results from simulated medical scenarios suggest that physicians were approximately 83 {\%} of the time in the WM state that is likely indicative of optimal performance.",
keywords = "Classification methods, Cognitive workload (CWL), Electronic medical records (EMR), Machine learning, Physicians, Working memory (WM)",
author = "Anthony Abrantes and Elizabeth Comitz and Prithima Mosaly and Lukasz Mazur",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-41947-3_12",
language = "English (US)",
isbn = "9783319419466",
series = "Advances in Intelligent Systems and Computing",
publisher = "Springer Verlag",
pages = "115--126",
booktitle = "Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016",
address = "Germany",

}

TY - GEN

T1 - Classification of EEG features for prediction of working memory load

AU - Abrantes,Anthony

AU - Comitz,Elizabeth

AU - Mosaly,Prithima

AU - Mazur,Lukasz

PY - 2017/1/1

Y1 - 2017/1/1

N2 - The objective of this research was to compare classification methods aimed at predicting working memory (WM) load. Electroencephalogram (EEG) data was collected from physicians while performing basic WM tasks and simulated medical scenarios. Data processing was performed to remove noise from the signal used for analysis (e.g., muscle activity, eye-blinks). The data from basic WM tasks was used to develop and test the four classification models (LASSO regression, support vector machines (SVM), nearest shrunken centroids (NSC), and iterated supervised principal components (ISPC) to predict a WM state indicative of physicians’ optimal performance. The na�ve misclassification rate was 19.74 %; LASSO and SVM outperformed this threshold: 18.10 and 12.21 % respectively). Both classification models had relatively high-specificity (LASSO: 97.2 %; SVM: 99.8 %); but relatively low-sensitivity LASSO: 20.7 %; SVM: 39.6 %). Results from simulated medical scenarios suggest that physicians were approximately 83 % of the time in the WM state that is likely indicative of optimal performance.

AB - The objective of this research was to compare classification methods aimed at predicting working memory (WM) load. Electroencephalogram (EEG) data was collected from physicians while performing basic WM tasks and simulated medical scenarios. Data processing was performed to remove noise from the signal used for analysis (e.g., muscle activity, eye-blinks). The data from basic WM tasks was used to develop and test the four classification models (LASSO regression, support vector machines (SVM), nearest shrunken centroids (NSC), and iterated supervised principal components (ISPC) to predict a WM state indicative of physicians’ optimal performance. The na�ve misclassification rate was 19.74 %; LASSO and SVM outperformed this threshold: 18.10 and 12.21 % respectively). Both classification models had relatively high-specificity (LASSO: 97.2 %; SVM: 99.8 %); but relatively low-sensitivity LASSO: 20.7 %; SVM: 39.6 %). Results from simulated medical scenarios suggest that physicians were approximately 83 % of the time in the WM state that is likely indicative of optimal performance.

KW - Classification methods

KW - Cognitive workload (CWL)

KW - Electronic medical records (EMR)

KW - Machine learning

KW - Physicians

KW - Working memory (WM)

UR - http://www.scopus.com/inward/record.url?scp=84986331390&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84986331390&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-41947-3_12

DO - 10.1007/978-3-319-41947-3_12

M3 - Conference contribution

SN - 9783319419466

T3 - Advances in Intelligent Systems and Computing

SP - 115

EP - 126

BT - Advances in the Human Side of Service Engineering - Proceedings of the AHFE International Conference on the Human Side of Service Engineering, 2016

PB - Springer Verlag

ER -