Divergence-based framework for common spatial patterns algorithms

Wojciech Samek, Motoaki Kawanabe, Klaus Muller

Research output: Contribution to journalArticle

97 Citations (Scopus)

Abstract

Controlling a device with a brain-computer interface requires extraction of relevant and robust features from high-dimensional electroencephalographic recordings. Spatial filtering is a crucial step in this feature extraction process. This paper reviews algorithms for spatial filter computation and introduces a general framework for this task based on divergence maximization. We show that the popular common spatial patterns (CSP) algorithm can be formulated as a divergence maximization problem and computed within our framework. Our approach easily permits enforcing different invariances and utilizing information from other subjects; thus, it unifies many of the recently proposed CSP variants in a principled manner. Furthermore, it allows to design novel spatial filtering algorithms by incorporating regularization schemes into the optimization process or applying other divergences. We evaluate the proposed approach using three regularization schemes, investigate the advantages of beta divergence, and show that subject-independent feature spaces can be extracted by jointly optimizing the divergence problems of multiple users. We discuss the relations to several CSP variants and investigate the advantages and limitations of our approach with simulations. Finally, we provide experimental results on a dataset containing recordings from 80 subjects and interpret the obtained patterns from a neurophysiological perspective.

Original languageEnglish
Article number6662468
Pages (from-to)50-72
Number of pages23
JournalIEEE Reviews in Biomedical Engineering
Volume7
DOIs
Publication statusPublished - 2014 Jan 1

Fingerprint

Brain-Computer Interfaces
Brain computer interface
Invariance
Feature extraction
Equipment and Supplies
Datasets

Keywords

  • Brain-computer interfaces
  • information geometry
  • spatial filters

ASJC Scopus subject areas

  • Biomedical Engineering

Cite this

Divergence-based framework for common spatial patterns algorithms. / Samek, Wojciech; Kawanabe, Motoaki; Muller, Klaus.

In: IEEE Reviews in Biomedical Engineering, Vol. 7, 6662468, 01.01.2014, p. 50-72.

Research output: Contribution to journalArticle

Samek, Wojciech ; Kawanabe, Motoaki ; Muller, Klaus. / Divergence-based framework for common spatial patterns algorithms. In: IEEE Reviews in Biomedical Engineering. 2014 ; Vol. 7. pp. 50-72.
@article{8f676f66a89e427190376e5cd1b4b471,
title = "Divergence-based framework for common spatial patterns algorithms",
abstract = "Controlling a device with a brain-computer interface requires extraction of relevant and robust features from high-dimensional electroencephalographic recordings. Spatial filtering is a crucial step in this feature extraction process. This paper reviews algorithms for spatial filter computation and introduces a general framework for this task based on divergence maximization. We show that the popular common spatial patterns (CSP) algorithm can be formulated as a divergence maximization problem and computed within our framework. Our approach easily permits enforcing different invariances and utilizing information from other subjects; thus, it unifies many of the recently proposed CSP variants in a principled manner. Furthermore, it allows to design novel spatial filtering algorithms by incorporating regularization schemes into the optimization process or applying other divergences. We evaluate the proposed approach using three regularization schemes, investigate the advantages of beta divergence, and show that subject-independent feature spaces can be extracted by jointly optimizing the divergence problems of multiple users. We discuss the relations to several CSP variants and investigate the advantages and limitations of our approach with simulations. Finally, we provide experimental results on a dataset containing recordings from 80 subjects and interpret the obtained patterns from a neurophysiological perspective.",
keywords = "Brain-computer interfaces, information geometry, spatial filters",
author = "Wojciech Samek and Motoaki Kawanabe and Klaus Muller",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/RBME.2013.2290621",
language = "English",
volume = "7",
pages = "50--72",
journal = "IEEE Reviews in Biomedical Engineering",
issn = "1937-3333",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Divergence-based framework for common spatial patterns algorithms

AU - Samek, Wojciech

AU - Kawanabe, Motoaki

AU - Muller, Klaus

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Controlling a device with a brain-computer interface requires extraction of relevant and robust features from high-dimensional electroencephalographic recordings. Spatial filtering is a crucial step in this feature extraction process. This paper reviews algorithms for spatial filter computation and introduces a general framework for this task based on divergence maximization. We show that the popular common spatial patterns (CSP) algorithm can be formulated as a divergence maximization problem and computed within our framework. Our approach easily permits enforcing different invariances and utilizing information from other subjects; thus, it unifies many of the recently proposed CSP variants in a principled manner. Furthermore, it allows to design novel spatial filtering algorithms by incorporating regularization schemes into the optimization process or applying other divergences. We evaluate the proposed approach using three regularization schemes, investigate the advantages of beta divergence, and show that subject-independent feature spaces can be extracted by jointly optimizing the divergence problems of multiple users. We discuss the relations to several CSP variants and investigate the advantages and limitations of our approach with simulations. Finally, we provide experimental results on a dataset containing recordings from 80 subjects and interpret the obtained patterns from a neurophysiological perspective.

AB - Controlling a device with a brain-computer interface requires extraction of relevant and robust features from high-dimensional electroencephalographic recordings. Spatial filtering is a crucial step in this feature extraction process. This paper reviews algorithms for spatial filter computation and introduces a general framework for this task based on divergence maximization. We show that the popular common spatial patterns (CSP) algorithm can be formulated as a divergence maximization problem and computed within our framework. Our approach easily permits enforcing different invariances and utilizing information from other subjects; thus, it unifies many of the recently proposed CSP variants in a principled manner. Furthermore, it allows to design novel spatial filtering algorithms by incorporating regularization schemes into the optimization process or applying other divergences. We evaluate the proposed approach using three regularization schemes, investigate the advantages of beta divergence, and show that subject-independent feature spaces can be extracted by jointly optimizing the divergence problems of multiple users. We discuss the relations to several CSP variants and investigate the advantages and limitations of our approach with simulations. Finally, we provide experimental results on a dataset containing recordings from 80 subjects and interpret the obtained patterns from a neurophysiological perspective.

KW - Brain-computer interfaces

KW - information geometry

KW - spatial filters

UR - http://www.scopus.com/inward/record.url?scp=84899454044&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84899454044&partnerID=8YFLogxK

U2 - 10.1109/RBME.2013.2290621

DO - 10.1109/RBME.2013.2290621

M3 - Article

VL - 7

SP - 50

EP - 72

JO - IEEE Reviews in Biomedical Engineering

JF - IEEE Reviews in Biomedical Engineering

SN - 1937-3333

M1 - 6662468

ER -