Toward unsupervised adaptation of LDA for brain-computer interfaces

C. Vidaurre, M. Kawanabe, P. Von Bünau, B. Blankertz, Klaus Muller

Research output: Contribution to journalArticle

168 Citations (Scopus)

Abstract

There is a step of significant difficulty experienced by brain-computer interface (BCI) users when going from the calibration recording to the feedback application. This effect has been previously studied and a supervised adaptation solution has been proposed. In this paper, we suggest a simple unsupervised adaptation method of the linear discriminant analysis (LDA) classifier that effectively solves this problem by counteracting the harmful effect of nonclass-related nonstationarities in electroencephalography (EEG) during BCI sessions performed with motor imagery tasks. For this, we first introduce three types of adaptation procedures and investigate them in an offline study with 19 datasets. Then, we select one of the proposed methods and analyze it further. The chosen classifier is offline tested in data from 80 healthy users and four high spinal cord injury patients. Finally, for the first time in BCI literature, we apply this unsupervised classifier in online experiments. Additionally, we show that its performance is significantly better than the state-of-the-art supervised approach.

Original languageEnglish
Pages (from-to)587-597
Number of pages11
JournalIEEE Transactions on Biomedical Engineering
Volume58
Issue number3 PART 1
DOIs
Publication statusPublished - 2011 Mar 1
Externally publishedYes

Fingerprint

Brain-Computer Interfaces
Brain computer interface
Discriminant Analysis
Discriminant analysis
Classifiers
Imagery (Psychotherapy)
Electroencephalography
Spinal Cord Injuries
Calibration
Feedback
Experiments

Keywords

  • Adaptive signal processing
  • brain computer interfaces
  • linear discriminant analysis
  • unsupervised learning

ASJC Scopus subject areas

  • Biomedical Engineering

Cite this

Vidaurre, C., Kawanabe, M., Von Bünau, P., Blankertz, B., & Muller, K. (2011). Toward unsupervised adaptation of LDA for brain-computer interfaces. IEEE Transactions on Biomedical Engineering, 58(3 PART 1), 587-597. https://doi.org/10.1109/TBME.2010.2093133

Toward unsupervised adaptation of LDA for brain-computer interfaces. / Vidaurre, C.; Kawanabe, M.; Von Bünau, P.; Blankertz, B.; Muller, Klaus.

In: IEEE Transactions on Biomedical Engineering, Vol. 58, No. 3 PART 1, 01.03.2011, p. 587-597.

Research output: Contribution to journalArticle

Vidaurre, C, Kawanabe, M, Von Bünau, P, Blankertz, B & Muller, K 2011, 'Toward unsupervised adaptation of LDA for brain-computer interfaces', IEEE Transactions on Biomedical Engineering, vol. 58, no. 3 PART 1, pp. 587-597. https://doi.org/10.1109/TBME.2010.2093133
Vidaurre, C. ; Kawanabe, M. ; Von Bünau, P. ; Blankertz, B. ; Muller, Klaus. / Toward unsupervised adaptation of LDA for brain-computer interfaces. In: IEEE Transactions on Biomedical Engineering. 2011 ; Vol. 58, No. 3 PART 1. pp. 587-597.
@article{d15190014507418ba569bf6308ba1091,
title = "Toward unsupervised adaptation of LDA for brain-computer interfaces",
abstract = "There is a step of significant difficulty experienced by brain-computer interface (BCI) users when going from the calibration recording to the feedback application. This effect has been previously studied and a supervised adaptation solution has been proposed. In this paper, we suggest a simple unsupervised adaptation method of the linear discriminant analysis (LDA) classifier that effectively solves this problem by counteracting the harmful effect of nonclass-related nonstationarities in electroencephalography (EEG) during BCI sessions performed with motor imagery tasks. For this, we first introduce three types of adaptation procedures and investigate them in an offline study with 19 datasets. Then, we select one of the proposed methods and analyze it further. The chosen classifier is offline tested in data from 80 healthy users and four high spinal cord injury patients. Finally, for the first time in BCI literature, we apply this unsupervised classifier in online experiments. Additionally, we show that its performance is significantly better than the state-of-the-art supervised approach.",
keywords = "Adaptive signal processing, brain computer interfaces, linear discriminant analysis, unsupervised learning",
author = "C. Vidaurre and M. Kawanabe and {Von B{\"u}nau}, P. and B. Blankertz and Klaus Muller",
year = "2011",
month = "3",
day = "1",
doi = "10.1109/TBME.2010.2093133",
language = "English",
volume = "58",
pages = "587--597",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE Computer Society",
number = "3 PART 1",

}

TY - JOUR

T1 - Toward unsupervised adaptation of LDA for brain-computer interfaces

AU - Vidaurre, C.

AU - Kawanabe, M.

AU - Von Bünau, P.

AU - Blankertz, B.

AU - Muller, Klaus

PY - 2011/3/1

Y1 - 2011/3/1

N2 - There is a step of significant difficulty experienced by brain-computer interface (BCI) users when going from the calibration recording to the feedback application. This effect has been previously studied and a supervised adaptation solution has been proposed. In this paper, we suggest a simple unsupervised adaptation method of the linear discriminant analysis (LDA) classifier that effectively solves this problem by counteracting the harmful effect of nonclass-related nonstationarities in electroencephalography (EEG) during BCI sessions performed with motor imagery tasks. For this, we first introduce three types of adaptation procedures and investigate them in an offline study with 19 datasets. Then, we select one of the proposed methods and analyze it further. The chosen classifier is offline tested in data from 80 healthy users and four high spinal cord injury patients. Finally, for the first time in BCI literature, we apply this unsupervised classifier in online experiments. Additionally, we show that its performance is significantly better than the state-of-the-art supervised approach.

AB - There is a step of significant difficulty experienced by brain-computer interface (BCI) users when going from the calibration recording to the feedback application. This effect has been previously studied and a supervised adaptation solution has been proposed. In this paper, we suggest a simple unsupervised adaptation method of the linear discriminant analysis (LDA) classifier that effectively solves this problem by counteracting the harmful effect of nonclass-related nonstationarities in electroencephalography (EEG) during BCI sessions performed with motor imagery tasks. For this, we first introduce three types of adaptation procedures and investigate them in an offline study with 19 datasets. Then, we select one of the proposed methods and analyze it further. The chosen classifier is offline tested in data from 80 healthy users and four high spinal cord injury patients. Finally, for the first time in BCI literature, we apply this unsupervised classifier in online experiments. Additionally, we show that its performance is significantly better than the state-of-the-art supervised approach.

KW - Adaptive signal processing

KW - brain computer interfaces

KW - linear discriminant analysis

KW - unsupervised learning

UR - http://www.scopus.com/inward/record.url?scp=79952167211&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79952167211&partnerID=8YFLogxK

U2 - 10.1109/TBME.2010.2093133

DO - 10.1109/TBME.2010.2093133

M3 - Article

C2 - 21095857

AN - SCOPUS:79952167211

VL - 58

SP - 587

EP - 597

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

IS - 3 PART 1

ER -