Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern

T. Shiratori, H. Tsubakida, Atsushi Ishiyama, Y. Ono

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    4 Citations (Scopus)

    Abstract

    Our purpose is to develop the 3-class Brain Machine Interface (BMI) incorporating the classification of resting state using Electroencephalography (EEG). Conventionally the most of BMI systems only accept EEG data when a subject performs some kind of task, such as motor imagery and gaze at visual stimuli. However, performing task causes fatigue of the subject. It is therefore important to develop classification algorithm for BMI system that utilizes rest state-EEG as one of the classes. The 3 classes we defined in this experiment were: (1) motor imagery of moving right hand; (2) motor imagery of moving left hand; and (3) rest state. And, we also measured EEG in an actual moving task (finger tapping) to ascertain validity of algorithm. We extracted feature vector using Finite Impulse Response (FIR) digital filter Filter Bank and multi-class Common Spatial Filter (mCSP) from EEG data, selected the feature by Mutual Information (MI), and made three 3-class classifiers using Random Forest (RF). The mean classification rate was 56.7±4.43% at motor imagery task and 88.7±4.54% at actual finger tapping task. And we compared the time required to extract features and compute classifiers with those of other methods. Our method is effective to some extent. (1) parameter selection time was better than choosing single band-pass filter that best discriminate classes among possible options of frequency bands; and (2) accuracy rate was better than our previous method using majority vote.

    Original languageEnglish
    Title of host publication3rd International Winter Conference on Brain-Computer Interface, BCI 2015
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    ISBN (Print)9781479974948
    DOIs
    Publication statusPublished - 2015 Mar 30
    Event2015 3rd International Winter Conference on Brain-Computer Interface, BCI 2015 - Gangwon-Do, Korea, Republic of
    Duration: 2015 Jan 122015 Jan 14

    Other

    Other2015 3rd International Winter Conference on Brain-Computer Interface, BCI 2015
    CountryKorea, Republic of
    CityGangwon-Do
    Period15/1/1215/1/14

    Fingerprint

    Imagery (Psychotherapy)
    Filter banks
    Electroencephalography
    Brain-Computer Interfaces
    Brain
    Fingers
    Classifiers
    Hand
    FIR filters
    Digital filters
    Bandpass filters
    Frequency bands
    Fatigue
    Fatigue of materials
    Experiments

    Keywords

    • EEG classification
    • filter bank common spatial pattern
    • motor imagery
    • multi-class common spatial pattern
    • mutual information
    • random forest
    • reststate

    ASJC Scopus subject areas

    • Human-Computer Interaction
    • Cognitive Neuroscience
    • Sensory Systems

    Cite this

    Shiratori, T., Tsubakida, H., Ishiyama, A., & Ono, Y. (2015). Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern. In 3rd International Winter Conference on Brain-Computer Interface, BCI 2015 [7073053] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IWW-BCI.2015.7073053

    Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern. / Shiratori, T.; Tsubakida, H.; Ishiyama, Atsushi; Ono, Y.

    3rd International Winter Conference on Brain-Computer Interface, BCI 2015. Institute of Electrical and Electronics Engineers Inc., 2015. 7073053.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Shiratori, T, Tsubakida, H, Ishiyama, A & Ono, Y 2015, Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern. in 3rd International Winter Conference on Brain-Computer Interface, BCI 2015., 7073053, Institute of Electrical and Electronics Engineers Inc., 2015 3rd International Winter Conference on Brain-Computer Interface, BCI 2015, Gangwon-Do, Korea, Republic of, 15/1/12. https://doi.org/10.1109/IWW-BCI.2015.7073053
    Shiratori T, Tsubakida H, Ishiyama A, Ono Y. Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern. In 3rd International Winter Conference on Brain-Computer Interface, BCI 2015. Institute of Electrical and Electronics Engineers Inc. 2015. 7073053 https://doi.org/10.1109/IWW-BCI.2015.7073053
    Shiratori, T. ; Tsubakida, H. ; Ishiyama, Atsushi ; Ono, Y. / Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern. 3rd International Winter Conference on Brain-Computer Interface, BCI 2015. Institute of Electrical and Electronics Engineers Inc., 2015.
    @inproceedings{5d929c2b122248c3af4219769bf97984,
    title = "Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern",
    abstract = "Our purpose is to develop the 3-class Brain Machine Interface (BMI) incorporating the classification of resting state using Electroencephalography (EEG). Conventionally the most of BMI systems only accept EEG data when a subject performs some kind of task, such as motor imagery and gaze at visual stimuli. However, performing task causes fatigue of the subject. It is therefore important to develop classification algorithm for BMI system that utilizes rest state-EEG as one of the classes. The 3 classes we defined in this experiment were: (1) motor imagery of moving right hand; (2) motor imagery of moving left hand; and (3) rest state. And, we also measured EEG in an actual moving task (finger tapping) to ascertain validity of algorithm. We extracted feature vector using Finite Impulse Response (FIR) digital filter Filter Bank and multi-class Common Spatial Filter (mCSP) from EEG data, selected the feature by Mutual Information (MI), and made three 3-class classifiers using Random Forest (RF). The mean classification rate was 56.7±4.43{\%} at motor imagery task and 88.7±4.54{\%} at actual finger tapping task. And we compared the time required to extract features and compute classifiers with those of other methods. Our method is effective to some extent. (1) parameter selection time was better than choosing single band-pass filter that best discriminate classes among possible options of frequency bands; and (2) accuracy rate was better than our previous method using majority vote.",
    keywords = "EEG classification, filter bank common spatial pattern, motor imagery, multi-class common spatial pattern, mutual information, random forest, reststate",
    author = "T. Shiratori and H. Tsubakida and Atsushi Ishiyama and Y. Ono",
    year = "2015",
    month = "3",
    day = "30",
    doi = "10.1109/IWW-BCI.2015.7073053",
    language = "English",
    isbn = "9781479974948",
    booktitle = "3rd International Winter Conference on Brain-Computer Interface, BCI 2015",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",

    }

    TY - GEN

    T1 - Three-class classification of motor imagery EEG data including 'rest state' using filter-bank multi-class Common Spatial pattern

    AU - Shiratori, T.

    AU - Tsubakida, H.

    AU - Ishiyama, Atsushi

    AU - Ono, Y.

    PY - 2015/3/30

    Y1 - 2015/3/30

    N2 - Our purpose is to develop the 3-class Brain Machine Interface (BMI) incorporating the classification of resting state using Electroencephalography (EEG). Conventionally the most of BMI systems only accept EEG data when a subject performs some kind of task, such as motor imagery and gaze at visual stimuli. However, performing task causes fatigue of the subject. It is therefore important to develop classification algorithm for BMI system that utilizes rest state-EEG as one of the classes. The 3 classes we defined in this experiment were: (1) motor imagery of moving right hand; (2) motor imagery of moving left hand; and (3) rest state. And, we also measured EEG in an actual moving task (finger tapping) to ascertain validity of algorithm. We extracted feature vector using Finite Impulse Response (FIR) digital filter Filter Bank and multi-class Common Spatial Filter (mCSP) from EEG data, selected the feature by Mutual Information (MI), and made three 3-class classifiers using Random Forest (RF). The mean classification rate was 56.7±4.43% at motor imagery task and 88.7±4.54% at actual finger tapping task. And we compared the time required to extract features and compute classifiers with those of other methods. Our method is effective to some extent. (1) parameter selection time was better than choosing single band-pass filter that best discriminate classes among possible options of frequency bands; and (2) accuracy rate was better than our previous method using majority vote.

    AB - Our purpose is to develop the 3-class Brain Machine Interface (BMI) incorporating the classification of resting state using Electroencephalography (EEG). Conventionally the most of BMI systems only accept EEG data when a subject performs some kind of task, such as motor imagery and gaze at visual stimuli. However, performing task causes fatigue of the subject. It is therefore important to develop classification algorithm for BMI system that utilizes rest state-EEG as one of the classes. The 3 classes we defined in this experiment were: (1) motor imagery of moving right hand; (2) motor imagery of moving left hand; and (3) rest state. And, we also measured EEG in an actual moving task (finger tapping) to ascertain validity of algorithm. We extracted feature vector using Finite Impulse Response (FIR) digital filter Filter Bank and multi-class Common Spatial Filter (mCSP) from EEG data, selected the feature by Mutual Information (MI), and made three 3-class classifiers using Random Forest (RF). The mean classification rate was 56.7±4.43% at motor imagery task and 88.7±4.54% at actual finger tapping task. And we compared the time required to extract features and compute classifiers with those of other methods. Our method is effective to some extent. (1) parameter selection time was better than choosing single band-pass filter that best discriminate classes among possible options of frequency bands; and (2) accuracy rate was better than our previous method using majority vote.

    KW - EEG classification

    KW - filter bank common spatial pattern

    KW - motor imagery

    KW - multi-class common spatial pattern

    KW - mutual information

    KW - random forest

    KW - reststate

    UR - http://www.scopus.com/inward/record.url?scp=84927949074&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84927949074&partnerID=8YFLogxK

    U2 - 10.1109/IWW-BCI.2015.7073053

    DO - 10.1109/IWW-BCI.2015.7073053

    M3 - Conference contribution

    AN - SCOPUS:84927949074

    SN - 9781479974948

    BT - 3rd International Winter Conference on Brain-Computer Interface, BCI 2015

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -