Scribd Logo
Upload

Welcome to Scribd!

  • 11 views

    Fabio R. Llorella, Gustavo Patow, José M. Azorín

    Uploaded by

    dagush
    This document presents a new method for detecting imagined motion from EEG signals using spectral entropy analysis and support vector machines (SVMs). The method was tested on 8 subjects performing imagined hand movements. Spectral entropy was able to characterize changes in EEG signals from the motor cortex during imagined motion. SVMs then classified the spectral entropy features to detect imagined movements, achieving average success rates of 71-82% depending on the subject. The results demonstrate that spectral entropy is useful for characterizing EEG signals during imagined motion tasks and can provide features for classifying such signals using machine learning methods like SVMs.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Fabio R. Llorella, Gustavo Patow, José M. Azorín

    Uploaded by

    dagush
    11 views1 page
    This document presents a new method for detecting imagined motion from EEG signals using spectral entropy analysis and support vector machines (SVMs). The method was tested on 8 subjects performing imagined hand movements. Spectral entropy was able to characterize changes in EEG signals from the motor cortex during imagined motion. SVMs then classified the spectral entropy features to detect imagined movements, achieving average success rates of 71-82% depending on the subject. The results demonstrate that spectral entropy is useful for characterizing EEG signals during imagined motion tasks and can provide features for classifying such signals using machine learning methods like SVMs.

    Original Description:

    Poster BCI

    Original Title

    Poster Final

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document presents a new method for detecting imagined motion from EEG signals using spectral entropy analysis and support vector machines (SVMs). The method was tested on 8 subjects performing imagined hand movements. Spectral entropy was able to characterize changes in EEG signals from the motor cortex during imagined motion. SVMs then classified the spectral entropy features to detect imagined movements, achieving average success rates of 71-82% depending on the subject. The results demonstrate that spectral entropy is useful for characterizing EEG signals during imagined motion tasks and can provide features for classifying such signals using machine learning methods like SVMs.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    11 views1 page

    Fabio R. Llorella, Gustavo Patow, José M. Azorín

    Uploaded by

    dagush
    This document presents a new method for detecting imagined motion from EEG signals using spectral entropy analysis and support vector machines (SVMs). The method was tested on 8 subjects performing imagined hand movements. Spectral entropy was able to characterize changes in EEG signals from the motor cortex during imagined motion. SVMs then classified the spectral entropy features to detect imagined movements, achieving average success rates of 71-82% depending on the subject. The results demonstrate that spectral entropy is useful for characterizing EEG signals during imagined motion tasks and can provide features for classifying such signals using machine learning methods like SVMs.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 1

    Spectral Entropy and Support Vector Machines For Imagined

    Motion Detection In Brain-Computer Interfaces


    Fabio R. Llorella1, Gustavo Patow1, Jos M. Azorn2
    1Universitat

    de Girona-ViRVIG,2Universidad Miguel Hernndez de Elche-Brain-Machine Interface Systems


    Lab
    Email:[email protected]

    Introduction
    Brain-Computer Interfaces (BCIs) aim to create a channel of communication between a person and a device without any physical
    action on the environment by the user. In this work we present a new method based on spectral entropy to detect changes in the
    motor area from EEG signals and their possible application in the detection of imagined movement. The success rate obtained
    with this technique are about 76%.

    Objective

    Spectral Entropy

    Demonstrate that the spectral entropy is useful to


    characterize the EEG signals
    Minimize the number of features

    Register

    Features and Classification

    g.USB amps from g.tec


    Sample frecuency: 256 Hz

    Real EEG Signals with Spectral Entropy


    Classification with SVM

    Bandpass from 0.1 to 100 Hz


    Notch filter: 50 Hz
    2 electrodes: C3 and C4
    (motor area)

    Results and Conclusions

    Success rates (%)


    User

    #1

    #2

    #3

    #4

    #5

    #6

    #7

    #8

    Avg

    50

    63,02

    70,16

    76,47

    65,12

    76,05

    71,84

    68,48

    78,99

    71,26

    55

    77,31

    85,29

    85,30

    83,19

    87,39

    86,97

    72,26

    80,67

    82,29

    56

    83,19

    85,71

    81,93

    84,45

    75,21

    89,07

    81,09

    73,94

    81,82

    62

    65,54

    60,08

    66,38

    69,74

    80,25

    73,10

    71,84

    71,00

    69,74

    Spectral Entropy is good for characterizing EEG


    signals because their frecuencies can be put in
    order when a motor imagination occurs.
    The frequencies are attenuated in the alpha and
    beta band. Therefore the signal is more in order
    Than when these signals are not attenuated
    ( in relax ).

    This research has been carried out in the framework of the project Associate - Decoding and stimulation of motor and sensory brain activity to support long term potentiation through Hebbian and paired associative
    stimulation during rehabilitation of gait (DPI2014-58431-C4-2-R), funded by the Spanish Ministry of Economy and Competitiveness and by the European Union through the European Regional Development Fund (ERDF)
    A way to build Europe. In addition, this work was partially funded by the TIN2014-52211-C2-2-R project from Ministerio de Economa y Competitividad, Spain.

    You might also like