Feature selection on movement imagery discrimination and attention detection

N. S. Dias, M. Kamrunnahar, P. M. Mendes, S. J. Schiff, J. H. Correia

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Noninvasive brain-computer interfaces (BCI) translate subject's electroencephalogram (EEG) features into device commands. Large feature sets should be downselected for efficient feature translation. This work proposes two different feature down-selection algorithms for BCI: (a) a sequential forward selection; and (b) an acrossgroup variance. Power rar ratios (PRs) were extracted from the EEG data for movement imagery discrimination. Event-related potentials (ERPs) were employed in the discrimination of cue-evoked responses. While center-out arrows, commonly used in calibration sessions, cued the subjects in the first experiment (for both PR and ERP analyses), less stimulating arrows that were centered in the visual field were employed in the second experiment (for ERP analysis). The proposed algorithms outperformed other three popular feature selection algorithms in movement imagery discrimination. In the first experiment, both algorithms achieved classification errors as low as 12.5% reducing the feature set dimensionality by more than 90%. The classification accuracy of ERPs dropped in the second experiment since centered cues reduced the amplitude of cue-evoked ERPs. The two proposed algorithms effectively reduced feature dimensionality while increasing movement imagery discrimination and detected cue-evoked ERPs that reflect subject attention.

Original languageEnglish (US)
Pages (from-to)331-341
Number of pages11
JournalMedical and Biological Engineering and Computing
Volume48
Issue number4
DOIs
StatePublished - Apr 1 2010

Fingerprint

Feature extraction
Brain computer interface
Electroencephalography
Experiments
Bioelectric potentials
Calibration

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering
  • Computer Science Applications

Cite this

Dias, N. S. ; Kamrunnahar, M. ; Mendes, P. M. ; Schiff, S. J. ; Correia, J. H. / Feature selection on movement imagery discrimination and attention detection. In: Medical and Biological Engineering and Computing. 2010 ; Vol. 48, No. 4. pp. 331-341.
@article{9dd05222d24a4cb28b52fb4e30e5a130,
title = "Feature selection on movement imagery discrimination and attention detection",
abstract = "Noninvasive brain-computer interfaces (BCI) translate subject's electroencephalogram (EEG) features into device commands. Large feature sets should be downselected for efficient feature translation. This work proposes two different feature down-selection algorithms for BCI: (a) a sequential forward selection; and (b) an acrossgroup variance. Power rar ratios (PRs) were extracted from the EEG data for movement imagery discrimination. Event-related potentials (ERPs) were employed in the discrimination of cue-evoked responses. While center-out arrows, commonly used in calibration sessions, cued the subjects in the first experiment (for both PR and ERP analyses), less stimulating arrows that were centered in the visual field were employed in the second experiment (for ERP analysis). The proposed algorithms outperformed other three popular feature selection algorithms in movement imagery discrimination. In the first experiment, both algorithms achieved classification errors as low as 12.5{\%} reducing the feature set dimensionality by more than 90{\%}. The classification accuracy of ERPs dropped in the second experiment since centered cues reduced the amplitude of cue-evoked ERPs. The two proposed algorithms effectively reduced feature dimensionality while increasing movement imagery discrimination and detected cue-evoked ERPs that reflect subject attention.",
author = "Dias, {N. S.} and M. Kamrunnahar and Mendes, {P. M.} and Schiff, {S. J.} and Correia, {J. H.}",
year = "2010",
month = "4",
day = "1",
doi = "10.1007/s11517-010-0578-1",
language = "English (US)",
volume = "48",
pages = "331--341",
journal = "Medical and Biological Engineering and Computing",
issn = "0140-0118",
publisher = "Springer Verlag",
number = "4",

}

Feature selection on movement imagery discrimination and attention detection. / Dias, N. S.; Kamrunnahar, M.; Mendes, P. M.; Schiff, S. J.; Correia, J. H.

In: Medical and Biological Engineering and Computing, Vol. 48, No. 4, 01.04.2010, p. 331-341.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Feature selection on movement imagery discrimination and attention detection

AU - Dias, N. S.

AU - Kamrunnahar, M.

AU - Mendes, P. M.

AU - Schiff, S. J.

AU - Correia, J. H.

PY - 2010/4/1

Y1 - 2010/4/1

N2 - Noninvasive brain-computer interfaces (BCI) translate subject's electroencephalogram (EEG) features into device commands. Large feature sets should be downselected for efficient feature translation. This work proposes two different feature down-selection algorithms for BCI: (a) a sequential forward selection; and (b) an acrossgroup variance. Power rar ratios (PRs) were extracted from the EEG data for movement imagery discrimination. Event-related potentials (ERPs) were employed in the discrimination of cue-evoked responses. While center-out arrows, commonly used in calibration sessions, cued the subjects in the first experiment (for both PR and ERP analyses), less stimulating arrows that were centered in the visual field were employed in the second experiment (for ERP analysis). The proposed algorithms outperformed other three popular feature selection algorithms in movement imagery discrimination. In the first experiment, both algorithms achieved classification errors as low as 12.5% reducing the feature set dimensionality by more than 90%. The classification accuracy of ERPs dropped in the second experiment since centered cues reduced the amplitude of cue-evoked ERPs. The two proposed algorithms effectively reduced feature dimensionality while increasing movement imagery discrimination and detected cue-evoked ERPs that reflect subject attention.

AB - Noninvasive brain-computer interfaces (BCI) translate subject's electroencephalogram (EEG) features into device commands. Large feature sets should be downselected for efficient feature translation. This work proposes two different feature down-selection algorithms for BCI: (a) a sequential forward selection; and (b) an acrossgroup variance. Power rar ratios (PRs) were extracted from the EEG data for movement imagery discrimination. Event-related potentials (ERPs) were employed in the discrimination of cue-evoked responses. While center-out arrows, commonly used in calibration sessions, cued the subjects in the first experiment (for both PR and ERP analyses), less stimulating arrows that were centered in the visual field were employed in the second experiment (for ERP analysis). The proposed algorithms outperformed other three popular feature selection algorithms in movement imagery discrimination. In the first experiment, both algorithms achieved classification errors as low as 12.5% reducing the feature set dimensionality by more than 90%. The classification accuracy of ERPs dropped in the second experiment since centered cues reduced the amplitude of cue-evoked ERPs. The two proposed algorithms effectively reduced feature dimensionality while increasing movement imagery discrimination and detected cue-evoked ERPs that reflect subject attention.

UR - http://www.scopus.com/inward/record.url?scp=77954458083&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77954458083&partnerID=8YFLogxK

U2 - 10.1007/s11517-010-0578-1

DO - 10.1007/s11517-010-0578-1

M3 - Article

C2 - 20112135

AN - SCOPUS:77954458083

VL - 48

SP - 331

EP - 341

JO - Medical and Biological Engineering and Computing

JF - Medical and Biological Engineering and Computing

SN - 0140-0118

IS - 4

ER -