The LDA beamformer: Optimal estimation of ERP source time series using linear discriminant analysis

Matthias S. Treder, Anne K. Porbadnigk, Forooz Shahbazi Avarvand, Klaus Muller, Benjamin Blankertz

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

We introduce a novel beamforming approach for estimating event-related potential (ERP) source time series based on regularized linear discriminant analysis (LDA). The optimization problems in LDA and linearly-constrained minimum-variance (LCMV) beamformers are formally equivalent. The approaches differ in that, in LCMV beamformers, the spatial patterns are derived from a source model, whereas in an LDA beamformer the spatial patterns are derived directly from the data (i.e., the ERP peak). Using a formal proof and MEG simulations, we show that the LDA beamformer is robust to correlated sources and offers a higher signal-to-noise ratio than the LCMV beamformer and PCA. As an application, we use EEG data from an oddball experiment to show how the LDA beamformer can be harnessed to detect single-trial ERP latencies and estimate connectivity between ERP sources. Concluding, the LDA beamformer optimally reconstructs ERP sources by maximizing the ERP signal-to-noise ratio. Hence, it is a highly suited tool for analyzing ERP source time series, particularly in EEG/MEG studies wherein a source model is not available.

Original languageEnglish
Pages (from-to)279-291
Number of pages13
JournalNeuroImage
Volume129
DOIs
Publication statusPublished - 2016 Apr 1

Fingerprint

Discriminant Analysis
Evoked Potentials
Signal-To-Noise Ratio
Electroencephalography
Passive Cutaneous Anaphylaxis

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Treder, M. S., Porbadnigk, A. K., Shahbazi Avarvand, F., Muller, K., & Blankertz, B. (2016). The LDA beamformer: Optimal estimation of ERP source time series using linear discriminant analysis. NeuroImage, 129, 279-291. https://doi.org/10.1016/j.neuroimage.2016.01.019

The LDA beamformer : Optimal estimation of ERP source time series using linear discriminant analysis. / Treder, Matthias S.; Porbadnigk, Anne K.; Shahbazi Avarvand, Forooz; Muller, Klaus; Blankertz, Benjamin.

In: NeuroImage, Vol. 129, 01.04.2016, p. 279-291.

Research output: Contribution to journalArticle

Treder, MS, Porbadnigk, AK, Shahbazi Avarvand, F, Muller, K & Blankertz, B 2016, 'The LDA beamformer: Optimal estimation of ERP source time series using linear discriminant analysis', NeuroImage, vol. 129, pp. 279-291. https://doi.org/10.1016/j.neuroimage.2016.01.019
Treder, Matthias S. ; Porbadnigk, Anne K. ; Shahbazi Avarvand, Forooz ; Muller, Klaus ; Blankertz, Benjamin. / The LDA beamformer : Optimal estimation of ERP source time series using linear discriminant analysis. In: NeuroImage. 2016 ; Vol. 129. pp. 279-291.
@article{b4f05c15fef2447ca91afb12ea86c169,
title = "The LDA beamformer: Optimal estimation of ERP source time series using linear discriminant analysis",
abstract = "We introduce a novel beamforming approach for estimating event-related potential (ERP) source time series based on regularized linear discriminant analysis (LDA). The optimization problems in LDA and linearly-constrained minimum-variance (LCMV) beamformers are formally equivalent. The approaches differ in that, in LCMV beamformers, the spatial patterns are derived from a source model, whereas in an LDA beamformer the spatial patterns are derived directly from the data (i.e., the ERP peak). Using a formal proof and MEG simulations, we show that the LDA beamformer is robust to correlated sources and offers a higher signal-to-noise ratio than the LCMV beamformer and PCA. As an application, we use EEG data from an oddball experiment to show how the LDA beamformer can be harnessed to detect single-trial ERP latencies and estimate connectivity between ERP sources. Concluding, the LDA beamformer optimally reconstructs ERP sources by maximizing the ERP signal-to-noise ratio. Hence, it is a highly suited tool for analyzing ERP source time series, particularly in EEG/MEG studies wherein a source model is not available.",
author = "Treder, {Matthias S.} and Porbadnigk, {Anne K.} and {Shahbazi Avarvand}, Forooz and Klaus Muller and Benjamin Blankertz",
year = "2016",
month = "4",
day = "1",
doi = "10.1016/j.neuroimage.2016.01.019",
language = "English",
volume = "129",
pages = "279--291",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - The LDA beamformer

T2 - Optimal estimation of ERP source time series using linear discriminant analysis

AU - Treder, Matthias S.

AU - Porbadnigk, Anne K.

AU - Shahbazi Avarvand, Forooz

AU - Muller, Klaus

AU - Blankertz, Benjamin

PY - 2016/4/1

Y1 - 2016/4/1

N2 - We introduce a novel beamforming approach for estimating event-related potential (ERP) source time series based on regularized linear discriminant analysis (LDA). The optimization problems in LDA and linearly-constrained minimum-variance (LCMV) beamformers are formally equivalent. The approaches differ in that, in LCMV beamformers, the spatial patterns are derived from a source model, whereas in an LDA beamformer the spatial patterns are derived directly from the data (i.e., the ERP peak). Using a formal proof and MEG simulations, we show that the LDA beamformer is robust to correlated sources and offers a higher signal-to-noise ratio than the LCMV beamformer and PCA. As an application, we use EEG data from an oddball experiment to show how the LDA beamformer can be harnessed to detect single-trial ERP latencies and estimate connectivity between ERP sources. Concluding, the LDA beamformer optimally reconstructs ERP sources by maximizing the ERP signal-to-noise ratio. Hence, it is a highly suited tool for analyzing ERP source time series, particularly in EEG/MEG studies wherein a source model is not available.

AB - We introduce a novel beamforming approach for estimating event-related potential (ERP) source time series based on regularized linear discriminant analysis (LDA). The optimization problems in LDA and linearly-constrained minimum-variance (LCMV) beamformers are formally equivalent. The approaches differ in that, in LCMV beamformers, the spatial patterns are derived from a source model, whereas in an LDA beamformer the spatial patterns are derived directly from the data (i.e., the ERP peak). Using a formal proof and MEG simulations, we show that the LDA beamformer is robust to correlated sources and offers a higher signal-to-noise ratio than the LCMV beamformer and PCA. As an application, we use EEG data from an oddball experiment to show how the LDA beamformer can be harnessed to detect single-trial ERP latencies and estimate connectivity between ERP sources. Concluding, the LDA beamformer optimally reconstructs ERP sources by maximizing the ERP signal-to-noise ratio. Hence, it is a highly suited tool for analyzing ERP source time series, particularly in EEG/MEG studies wherein a source model is not available.

UR - http://www.scopus.com/inward/record.url?scp=84956884007&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84956884007&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2016.01.019

DO - 10.1016/j.neuroimage.2016.01.019

M3 - Article

C2 - 26804780

AN - SCOPUS:84956884007

VL - 129

SP - 279

EP - 291

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -