A 3-channel Active Electrode EEG Device for the Classification of MotorImagery Brain waves for Brain Computer Interface
Date
2017-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Pulchowk Campus
Abstract
Thisthesisworkpresentsacosteffectivemethodtorecordbrainwavesignalsusing
threechannelactiveelectrodeEEGdeviceandclassifybrainwavesrelatedtomotor
imagery(MI)leftandrighthandmovement,basedonelectroencephalography(EEG)
measuredfromthecentrallobe,thatcouldbeusedfortheBrainComputerInterface
(BCI)systems.ThegoalofthisthesisistouseIndependentComponentAnalysis
(ICA)fortheremovalofEEGartifacts,andthenextractthebrainwavesfeaturesforMI
lefthandandMIrighthandmovementusingWaveletDecomposition(WD).The‘Mor-
let’motherwaveletisusedforwaveletdecompositionasitshowsbetterperformance
foranalysisofnon-stationarybiomedicalsignalslikeEEG.Thebrainwavefeatureslike
MaximumPoweramongalldecompositionlevel(MMP),Frequencycorrespondingto
MMP(MAF),andMaximumAmplitudeofthesignalwithMAF(MMA)ischosen
astheclassificationfeaturesfortheclassificationofMIbrainwaves.Theclassifica-tionofMIbrainwavesignalsisdoneusingLinearDiscriminantAnalysis(LDA)which
showedtheaccuracyof81.6%.Thus,thedesignedthreechannelactiveelectrodeEEG
deviceusedshowedgoodperformanceforrecordingEEGsignals.Furthermore,signal
preprocessingalgorithmICA,featureextractionmethodWaveletDecomposition,and
classificationmethodLDAshowedgoodperformancefortheclassificationofMIleft
handandMIrighthandactivities.
Description
This thesis work presents a cost effective method to record brain wave signals using
three channel active electrode EEG device and classify brain waves related to motor
imagery(MI) left and right hand movement, based on electroencephalography (EEG)
measured from the central lobe ,that could be used for the Brain Computer Interface
(BCI) systems.
Keywords
LDA, Morlet, Wavelet Decomposition, ICA