Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions

Publikation: KonferenzbeitragPosterForschungBegutachtung

Abstract

Brain-computer interfaces (BCIs) might provide
an intuitive way for severely motor impaired persons to operate
assistive devices to perform daily life activities. Recent studies
have shown that complex hand movements, such as reachand-
grasp tasks, can be decoded from the low frequency of
the electroencephalogram (EEG). In this work we investigated
whether additional features extracted from the frequencydomain
of alpha and beta bands could improve classification
performance of rest vs. palmar vs. lateral grasp. We analysed
two multi-class classification approaches, the first using features
from the low frequency time-domain, and the second in which
we combined the time-domain with frequency-domain features
from alpha and beta bands. We measured EEG of ten participants
without motor disability which performed self-paced
reach-and-grasp actions on objects of daily life. For the timedomain
classification approach, participants reached an average
peak accuracy of 65%. For the combined approach, an average
peak accuracy of 75% was reached. In both approaches and for
all subjects, performance was significantly higher than chance
level (38.1%, 3-class scenario). By computing the confusion
matrices as well as feature rankings through the Fisher score,
we show that movement vs. rest classification performance
increased considerably in the combined approach and was
the main responsible for the multi-class higher performance.
These findings could help the development of BCIs in real-life
scenarios, where decreasing false movement detections could
drastically increase the end-user acceptance and usability of
BCIs.
Originalspracheenglisch
PublikationsstatusVeröffentlicht - 23 Jul 2019
Veranstaltung41st International Engineering in Medicine and Biology Conference 2019 - Berlin, Deutschland
Dauer: 23 Jul 201927 Jul 2019
Konferenznummer: 41
https://embc.embs.org/2019/

Konferenz

Konferenz41st International Engineering in Medicine and Biology Conference 2019
KurztitelIEEE EMBC 2019
LandDeutschland
OrtBerlin
Zeitraum23/07/1927/07/19
Internetadresse

Fingerprint

Brain computer interface
Electroencephalography

Dies zitieren

Schwarz, A., Pereira, J., Lindner, L., & Müller-Putz, G. (2019). Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions. Postersitzung präsentiert bei 41st International Engineering in Medicine and Biology Conference 2019, Berlin, Deutschland.

Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions. / Schwarz, Andreas; Pereira, Joana; Lindner, Lydia; Müller-Putz, Gernot.

2019. Postersitzung präsentiert bei 41st International Engineering in Medicine and Biology Conference 2019, Berlin, Deutschland.

Publikation: KonferenzbeitragPosterForschungBegutachtung

Schwarz A, Pereira J, Lindner L, Müller-Putz G. Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions. 2019. Postersitzung präsentiert bei 41st International Engineering in Medicine and Biology Conference 2019, Berlin, Deutschland.
Schwarz, Andreas ; Pereira, Joana ; Lindner, Lydia ; Müller-Putz, Gernot. / Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions. Postersitzung präsentiert bei 41st International Engineering in Medicine and Biology Conference 2019, Berlin, Deutschland.
@conference{2ce99c4fae02454abe97e991817f61ae,
title = "Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions",
abstract = "Brain-computer interfaces (BCIs) might providean intuitive way for severely motor impaired persons to operateassistive devices to perform daily life activities. Recent studieshave shown that complex hand movements, such as reachand-grasp tasks, can be decoded from the low frequency ofthe electroencephalogram (EEG). In this work we investigatedwhether additional features extracted from the frequencydomainof alpha and beta bands could improve classificationperformance of rest vs. palmar vs. lateral grasp. We analysedtwo multi-class classification approaches, the first using featuresfrom the low frequency time-domain, and the second in whichwe combined the time-domain with frequency-domain featuresfrom alpha and beta bands. We measured EEG of ten participantswithout motor disability which performed self-pacedreach-and-grasp actions on objects of daily life. For the timedomainclassification approach, participants reached an averagepeak accuracy of 65{\%}. For the combined approach, an averagepeak accuracy of 75{\%} was reached. In both approaches and forall subjects, performance was significantly higher than chancelevel (38.1{\%}, 3-class scenario). By computing the confusionmatrices as well as feature rankings through the Fisher score,we show that movement vs. rest classification performanceincreased considerably in the combined approach and wasthe main responsible for the multi-class higher performance.These findings could help the development of BCIs in real-lifescenarios, where decreasing false movement detections coulddrastically increase the end-user acceptance and usability ofBCIs.",
author = "Andreas Schwarz and Joana Pereira and Lydia Lindner and Gernot M{\"u}ller-Putz",
note = "Poster ; 41st International Engineering in Medicine and Biology Conference 2019, IEEE EMBC 2019 ; Conference date: 23-07-2019 Through 27-07-2019",
year = "2019",
month = "7",
day = "23",
language = "English",
url = "https://embc.embs.org/2019/",

}

TY - CONF

T1 - Combining frequency and time-domain EEG features for classification of self-paced reach-and-grasp actions

AU - Schwarz, Andreas

AU - Pereira, Joana

AU - Lindner, Lydia

AU - Müller-Putz, Gernot

N1 - Poster

PY - 2019/7/23

Y1 - 2019/7/23

N2 - Brain-computer interfaces (BCIs) might providean intuitive way for severely motor impaired persons to operateassistive devices to perform daily life activities. Recent studieshave shown that complex hand movements, such as reachand-grasp tasks, can be decoded from the low frequency ofthe electroencephalogram (EEG). In this work we investigatedwhether additional features extracted from the frequencydomainof alpha and beta bands could improve classificationperformance of rest vs. palmar vs. lateral grasp. We analysedtwo multi-class classification approaches, the first using featuresfrom the low frequency time-domain, and the second in whichwe combined the time-domain with frequency-domain featuresfrom alpha and beta bands. We measured EEG of ten participantswithout motor disability which performed self-pacedreach-and-grasp actions on objects of daily life. For the timedomainclassification approach, participants reached an averagepeak accuracy of 65%. For the combined approach, an averagepeak accuracy of 75% was reached. In both approaches and forall subjects, performance was significantly higher than chancelevel (38.1%, 3-class scenario). By computing the confusionmatrices as well as feature rankings through the Fisher score,we show that movement vs. rest classification performanceincreased considerably in the combined approach and wasthe main responsible for the multi-class higher performance.These findings could help the development of BCIs in real-lifescenarios, where decreasing false movement detections coulddrastically increase the end-user acceptance and usability ofBCIs.

AB - Brain-computer interfaces (BCIs) might providean intuitive way for severely motor impaired persons to operateassistive devices to perform daily life activities. Recent studieshave shown that complex hand movements, such as reachand-grasp tasks, can be decoded from the low frequency ofthe electroencephalogram (EEG). In this work we investigatedwhether additional features extracted from the frequencydomainof alpha and beta bands could improve classificationperformance of rest vs. palmar vs. lateral grasp. We analysedtwo multi-class classification approaches, the first using featuresfrom the low frequency time-domain, and the second in whichwe combined the time-domain with frequency-domain featuresfrom alpha and beta bands. We measured EEG of ten participantswithout motor disability which performed self-pacedreach-and-grasp actions on objects of daily life. For the timedomainclassification approach, participants reached an averagepeak accuracy of 65%. For the combined approach, an averagepeak accuracy of 75% was reached. In both approaches and forall subjects, performance was significantly higher than chancelevel (38.1%, 3-class scenario). By computing the confusionmatrices as well as feature rankings through the Fisher score,we show that movement vs. rest classification performanceincreased considerably in the combined approach and wasthe main responsible for the multi-class higher performance.These findings could help the development of BCIs in real-lifescenarios, where decreasing false movement detections coulddrastically increase the end-user acceptance and usability ofBCIs.

M3 - Poster

ER -