Central-tendency estimation and nearest-estimate classification of event related potentials

Lalit Gupta, Srinivas Kota, Phani Yarlagadda, Dennis L. Molfese

Research output: Contribution to journalArticle

Abstract

Event related potentials (ERPs) are modeled as random vectors in order to determine multivariate central-tendency (C-T) estimates of ERPs such as the arithmetic mean, geometric mean, harmonic mean, median, tri-mean, trimmed-mean, and the Winsorized mean. Additionally, it is shown that the C-T estimates can be used to implement various forms of minimum-distance classifiers for individual channels and for single-channel heterogeneous, multi-channel homogeneous, and multi-channel heterogeneoushomogenous ERP classification through decision fusion. The study also focuses on answering the following related questions: (a) How do the C-T ERP estimates compare with each other? (b) How do the performances of nearest-estimate classifiers compare with each other? (c) For a given ERP channel, do the heterogeneous nearest-estimate classifiers offer complementary information for improving performance through decision fusion? (d) Do the homogeneous nearest-estimate classifiers of different channels offer complementary information for improving performance through decision fusion? (e) Can the performance be improved by fusing the decisions of all or a selected subset of the entire classifier ensemble? These questions are answered by designing estimation and classification experiments using real 6-channel ERPs. It is shown that although the operations to compute the vector C-T estimates can be quite different, the ERP estimates are similar with respect to their overall waveform shapes and peak latencies. Furthermore, the results of the classification experiments show that by fusing homogeneous nearest-estimate classifier decisions across multiple channels, the classification accuracy can be improved significantly when compared with the accuracies of individual channel classifiers.

Original languageEnglish (US)
Pages (from-to)1418-1425
Number of pages8
JournalPattern Recognition
Volume44
Issue number7
DOIs
StatePublished - Jul 1 2011

Fingerprint

Classifiers
Fusion reactions
Experiments

Keywords

  • Decision fusion
  • ERP classification
  • ERP estimation
  • Event related potentials
  • Multivariate central-tendency estimation

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

Central-tendency estimation and nearest-estimate classification of event related potentials. / Gupta, Lalit; Kota, Srinivas; Yarlagadda, Phani; Molfese, Dennis L.

In: Pattern Recognition, Vol. 44, No. 7, 01.07.2011, p. 1418-1425.

Research output: Contribution to journalArticle

Gupta, Lalit ; Kota, Srinivas ; Yarlagadda, Phani ; Molfese, Dennis L. / Central-tendency estimation and nearest-estimate classification of event related potentials. In: Pattern Recognition. 2011 ; Vol. 44, No. 7. pp. 1418-1425.
@article{e9dacb907ef348f096c3da6cc9b6d106,
title = "Central-tendency estimation and nearest-estimate classification of event related potentials",
abstract = "Event related potentials (ERPs) are modeled as random vectors in order to determine multivariate central-tendency (C-T) estimates of ERPs such as the arithmetic mean, geometric mean, harmonic mean, median, tri-mean, trimmed-mean, and the Winsorized mean. Additionally, it is shown that the C-T estimates can be used to implement various forms of minimum-distance classifiers for individual channels and for single-channel heterogeneous, multi-channel homogeneous, and multi-channel heterogeneoushomogenous ERP classification through decision fusion. The study also focuses on answering the following related questions: (a) How do the C-T ERP estimates compare with each other? (b) How do the performances of nearest-estimate classifiers compare with each other? (c) For a given ERP channel, do the heterogeneous nearest-estimate classifiers offer complementary information for improving performance through decision fusion? (d) Do the homogeneous nearest-estimate classifiers of different channels offer complementary information for improving performance through decision fusion? (e) Can the performance be improved by fusing the decisions of all or a selected subset of the entire classifier ensemble? These questions are answered by designing estimation and classification experiments using real 6-channel ERPs. It is shown that although the operations to compute the vector C-T estimates can be quite different, the ERP estimates are similar with respect to their overall waveform shapes and peak latencies. Furthermore, the results of the classification experiments show that by fusing homogeneous nearest-estimate classifier decisions across multiple channels, the classification accuracy can be improved significantly when compared with the accuracies of individual channel classifiers.",
keywords = "Decision fusion, ERP classification, ERP estimation, Event related potentials, Multivariate central-tendency estimation",
author = "Lalit Gupta and Srinivas Kota and Phani Yarlagadda and Molfese, {Dennis L.}",
year = "2011",
month = "7",
day = "1",
doi = "10.1016/j.patcog.2011.01.002",
language = "English (US)",
volume = "44",
pages = "1418--1425",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "7",

}

TY - JOUR

T1 - Central-tendency estimation and nearest-estimate classification of event related potentials

AU - Gupta, Lalit

AU - Kota, Srinivas

AU - Yarlagadda, Phani

AU - Molfese, Dennis L.

PY - 2011/7/1

Y1 - 2011/7/1

N2 - Event related potentials (ERPs) are modeled as random vectors in order to determine multivariate central-tendency (C-T) estimates of ERPs such as the arithmetic mean, geometric mean, harmonic mean, median, tri-mean, trimmed-mean, and the Winsorized mean. Additionally, it is shown that the C-T estimates can be used to implement various forms of minimum-distance classifiers for individual channels and for single-channel heterogeneous, multi-channel homogeneous, and multi-channel heterogeneoushomogenous ERP classification through decision fusion. The study also focuses on answering the following related questions: (a) How do the C-T ERP estimates compare with each other? (b) How do the performances of nearest-estimate classifiers compare with each other? (c) For a given ERP channel, do the heterogeneous nearest-estimate classifiers offer complementary information for improving performance through decision fusion? (d) Do the homogeneous nearest-estimate classifiers of different channels offer complementary information for improving performance through decision fusion? (e) Can the performance be improved by fusing the decisions of all or a selected subset of the entire classifier ensemble? These questions are answered by designing estimation and classification experiments using real 6-channel ERPs. It is shown that although the operations to compute the vector C-T estimates can be quite different, the ERP estimates are similar with respect to their overall waveform shapes and peak latencies. Furthermore, the results of the classification experiments show that by fusing homogeneous nearest-estimate classifier decisions across multiple channels, the classification accuracy can be improved significantly when compared with the accuracies of individual channel classifiers.

AB - Event related potentials (ERPs) are modeled as random vectors in order to determine multivariate central-tendency (C-T) estimates of ERPs such as the arithmetic mean, geometric mean, harmonic mean, median, tri-mean, trimmed-mean, and the Winsorized mean. Additionally, it is shown that the C-T estimates can be used to implement various forms of minimum-distance classifiers for individual channels and for single-channel heterogeneous, multi-channel homogeneous, and multi-channel heterogeneoushomogenous ERP classification through decision fusion. The study also focuses on answering the following related questions: (a) How do the C-T ERP estimates compare with each other? (b) How do the performances of nearest-estimate classifiers compare with each other? (c) For a given ERP channel, do the heterogeneous nearest-estimate classifiers offer complementary information for improving performance through decision fusion? (d) Do the homogeneous nearest-estimate classifiers of different channels offer complementary information for improving performance through decision fusion? (e) Can the performance be improved by fusing the decisions of all or a selected subset of the entire classifier ensemble? These questions are answered by designing estimation and classification experiments using real 6-channel ERPs. It is shown that although the operations to compute the vector C-T estimates can be quite different, the ERP estimates are similar with respect to their overall waveform shapes and peak latencies. Furthermore, the results of the classification experiments show that by fusing homogeneous nearest-estimate classifier decisions across multiple channels, the classification accuracy can be improved significantly when compared with the accuracies of individual channel classifiers.

KW - Decision fusion

KW - ERP classification

KW - ERP estimation

KW - Event related potentials

KW - Multivariate central-tendency estimation

UR - http://www.scopus.com/inward/record.url?scp=79952157168&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79952157168&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2011.01.002

DO - 10.1016/j.patcog.2011.01.002

M3 - Article

AN - SCOPUS:79952157168

VL - 44

SP - 1418

EP - 1425

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 7

ER -