Improving BCI performance by task-related trial pruning

Claudia Sannelli, Mikio Braun, Klaus Muller

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

Noise in electroencephalography data (EEG) is an ubiquitous problem that limits the performance of brain computer interfaces (BCI). While typical EEG artifacts are usually removed by trial rejection or by filtering, noise induced in the data by the subject's failure to produce the required mental state is very harmful. Such "noise" effects are rather common, especially for naive subjects in their training phase and, thus, standard artifact removal methods would inevitably fail. In this paper, we present a novel method which aims to detect such defected trials taking into account the intended task by use of Relevant Dimensionality Estimation (RDE), a new machine learning method for denoising in feature space. In this manner, our method effectively "cleans" the training data and thus allows better BCI classification. Preliminary results conducted on a data set of 43 naive subjects show a significant improvement for 74% of the subjects.

Original languageEnglish
Pages (from-to)1295-1304
Number of pages10
JournalNeural Networks
Volume22
Issue number9
DOIs
Publication statusPublished - 2009 Nov 1
Externally publishedYes

Fingerprint

Brain-Computer Interfaces
Brain computer interface
Task Performance and Analysis
Electroencephalography
Noise
Artifacts
Learning systems

Keywords

  • Artifact Correction/Rejection
  • Brain-Computer Interface
  • Common Spatial Patterns
  • Denoising
  • Kernel Principal Component Analysis

ASJC Scopus subject areas

  • Artificial Intelligence
  • Cognitive Neuroscience

Cite this

Improving BCI performance by task-related trial pruning. / Sannelli, Claudia; Braun, Mikio; Muller, Klaus.

In: Neural Networks, Vol. 22, No. 9, 01.11.2009, p. 1295-1304.

Research output: Contribution to journalArticle

Sannelli, C, Braun, M & Muller, K 2009, 'Improving BCI performance by task-related trial pruning', Neural Networks, vol. 22, no. 9, pp. 1295-1304. https://doi.org/10.1016/j.neunet.2009.08.006
Sannelli, Claudia ; Braun, Mikio ; Muller, Klaus. / Improving BCI performance by task-related trial pruning. In: Neural Networks. 2009 ; Vol. 22, No. 9. pp. 1295-1304.
@article{17c6ff74de1e40dc9624de5a91bb3265,
title = "Improving BCI performance by task-related trial pruning",
abstract = "Noise in electroencephalography data (EEG) is an ubiquitous problem that limits the performance of brain computer interfaces (BCI). While typical EEG artifacts are usually removed by trial rejection or by filtering, noise induced in the data by the subject's failure to produce the required mental state is very harmful. Such {"}noise{"} effects are rather common, especially for naive subjects in their training phase and, thus, standard artifact removal methods would inevitably fail. In this paper, we present a novel method which aims to detect such defected trials taking into account the intended task by use of Relevant Dimensionality Estimation (RDE), a new machine learning method for denoising in feature space. In this manner, our method effectively {"}cleans{"} the training data and thus allows better BCI classification. Preliminary results conducted on a data set of 43 naive subjects show a significant improvement for 74{\%} of the subjects.",
keywords = "Artifact Correction/Rejection, Brain-Computer Interface, Common Spatial Patterns, Denoising, Kernel Principal Component Analysis",
author = "Claudia Sannelli and Mikio Braun and Klaus Muller",
year = "2009",
month = "11",
day = "1",
doi = "10.1016/j.neunet.2009.08.006",
language = "English",
volume = "22",
pages = "1295--1304",
journal = "Neural Networks",
issn = "0893-6080",
publisher = "Elsevier Limited",
number = "9",

}

TY - JOUR

T1 - Improving BCI performance by task-related trial pruning

AU - Sannelli, Claudia

AU - Braun, Mikio

AU - Muller, Klaus

PY - 2009/11/1

Y1 - 2009/11/1

N2 - Noise in electroencephalography data (EEG) is an ubiquitous problem that limits the performance of brain computer interfaces (BCI). While typical EEG artifacts are usually removed by trial rejection or by filtering, noise induced in the data by the subject's failure to produce the required mental state is very harmful. Such "noise" effects are rather common, especially for naive subjects in their training phase and, thus, standard artifact removal methods would inevitably fail. In this paper, we present a novel method which aims to detect such defected trials taking into account the intended task by use of Relevant Dimensionality Estimation (RDE), a new machine learning method for denoising in feature space. In this manner, our method effectively "cleans" the training data and thus allows better BCI classification. Preliminary results conducted on a data set of 43 naive subjects show a significant improvement for 74% of the subjects.

AB - Noise in electroencephalography data (EEG) is an ubiquitous problem that limits the performance of brain computer interfaces (BCI). While typical EEG artifacts are usually removed by trial rejection or by filtering, noise induced in the data by the subject's failure to produce the required mental state is very harmful. Such "noise" effects are rather common, especially for naive subjects in their training phase and, thus, standard artifact removal methods would inevitably fail. In this paper, we present a novel method which aims to detect such defected trials taking into account the intended task by use of Relevant Dimensionality Estimation (RDE), a new machine learning method for denoising in feature space. In this manner, our method effectively "cleans" the training data and thus allows better BCI classification. Preliminary results conducted on a data set of 43 naive subjects show a significant improvement for 74% of the subjects.

KW - Artifact Correction/Rejection

KW - Brain-Computer Interface

KW - Common Spatial Patterns

KW - Denoising

KW - Kernel Principal Component Analysis

UR - http://www.scopus.com/inward/record.url?scp=70350223661&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350223661&partnerID=8YFLogxK

U2 - 10.1016/j.neunet.2009.08.006

DO - 10.1016/j.neunet.2009.08.006

M3 - Article

VL - 22

SP - 1295

EP - 1304

JO - Neural Networks

JF - Neural Networks

SN - 0893-6080

IS - 9

ER -