EEG decoding of semantic category reveals distributed representations for single concepts

B. Murphy, M. Poesio, F. Bovolo, L. Bruzzone, M. Dalponte, H. Lakany

    Research output: Contribution to journalArticle

    27 Citations (Scopus)

    Abstract

    Here we present a collection of advanced data-mining techniques that allows the semantics of individual concepts to be decoded from single trials of EEG data. Neural activity was recorded while participants silently named images of mammals and tools, and category could be detected in single trials with an accuracy well above chance, both when considering data from single participants, and when group-training across participants. The pattern of classifications made by the algorithm confirmed that the neural patterns identified are due to conceptual category, and not any of a series of processing-related confounds. The time intervals, frequency bands and scalp locations that proved most informative for prediction permit physiological interpretation: the widespread activation shortly after appearance of the stimulus (from 100ms) is consistent both with accounts of multi-pass processing, and distributed representations of categories. These methods provide a new alternative to fMRI for fine-grained investigations of the conceptual lexicon.
    LanguageEnglish
    Pages12-22
    Number of pages11
    JournalBrain and Language
    Volume117
    Issue number1
    DOIs
    Publication statusPublished - Apr 2011

    Fingerprint

    Data Mining
    Scalp
    Semantics
    Electroencephalography
    Mammals
    semantics
    Magnetic Resonance Imaging
    activation
    stimulus
    interpretation
    Distributed Representation
    Semantic Category
    Electroencephalogram
    Decoding
    Group

    Keywords

    • semantics
    • single concepts
    • neuroimaging
    • brain damage
    • data mining
    • machine learning

    Cite this

    Murphy, B., Poesio, M., Bovolo, F., Bruzzone, L., Dalponte, M., & Lakany, H. (2011). EEG decoding of semantic category reveals distributed representations for single concepts. Brain and Language, 117(1), 12-22. https://doi.org/10.1016/j.bandl.2010.09.013
    Murphy, B. ; Poesio, M. ; Bovolo, F. ; Bruzzone, L. ; Dalponte, M. ; Lakany, H. / EEG decoding of semantic category reveals distributed representations for single concepts. In: Brain and Language. 2011 ; Vol. 117, No. 1. pp. 12-22.
    @article{b0c22f9804a849a3adf3c5723ff4f301,
    title = "EEG decoding of semantic category reveals distributed representations for single concepts",
    abstract = "Here we present a collection of advanced data-mining techniques that allows the semantics of individual concepts to be decoded from single trials of EEG data. Neural activity was recorded while participants silently named images of mammals and tools, and category could be detected in single trials with an accuracy well above chance, both when considering data from single participants, and when group-training across participants. The pattern of classifications made by the algorithm confirmed that the neural patterns identified are due to conceptual category, and not any of a series of processing-related confounds. The time intervals, frequency bands and scalp locations that proved most informative for prediction permit physiological interpretation: the widespread activation shortly after appearance of the stimulus (from 100ms) is consistent both with accounts of multi-pass processing, and distributed representations of categories. These methods provide a new alternative to fMRI for fine-grained investigations of the conceptual lexicon.",
    keywords = "semantics , single concepts , neuroimaging, brain damage, data mining, machine learning",
    author = "B. Murphy and M. Poesio and F. Bovolo and L. Bruzzone and M. Dalponte and H. Lakany",
    year = "2011",
    month = "4",
    doi = "10.1016/j.bandl.2010.09.013",
    language = "English",
    volume = "117",
    pages = "12--22",
    journal = "Brain and Language",
    issn = "0093-934X",
    number = "1",

    }

    Murphy, B, Poesio, M, Bovolo, F, Bruzzone, L, Dalponte, M & Lakany, H 2011, 'EEG decoding of semantic category reveals distributed representations for single concepts' Brain and Language, vol. 117, no. 1, pp. 12-22. https://doi.org/10.1016/j.bandl.2010.09.013

    EEG decoding of semantic category reveals distributed representations for single concepts. / Murphy, B.; Poesio, M.; Bovolo, F.; Bruzzone, L.; Dalponte, M.; Lakany, H.

    In: Brain and Language, Vol. 117, No. 1, 04.2011, p. 12-22.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - EEG decoding of semantic category reveals distributed representations for single concepts

    AU - Murphy, B.

    AU - Poesio, M.

    AU - Bovolo, F.

    AU - Bruzzone, L.

    AU - Dalponte, M.

    AU - Lakany, H.

    PY - 2011/4

    Y1 - 2011/4

    N2 - Here we present a collection of advanced data-mining techniques that allows the semantics of individual concepts to be decoded from single trials of EEG data. Neural activity was recorded while participants silently named images of mammals and tools, and category could be detected in single trials with an accuracy well above chance, both when considering data from single participants, and when group-training across participants. The pattern of classifications made by the algorithm confirmed that the neural patterns identified are due to conceptual category, and not any of a series of processing-related confounds. The time intervals, frequency bands and scalp locations that proved most informative for prediction permit physiological interpretation: the widespread activation shortly after appearance of the stimulus (from 100ms) is consistent both with accounts of multi-pass processing, and distributed representations of categories. These methods provide a new alternative to fMRI for fine-grained investigations of the conceptual lexicon.

    AB - Here we present a collection of advanced data-mining techniques that allows the semantics of individual concepts to be decoded from single trials of EEG data. Neural activity was recorded while participants silently named images of mammals and tools, and category could be detected in single trials with an accuracy well above chance, both when considering data from single participants, and when group-training across participants. The pattern of classifications made by the algorithm confirmed that the neural patterns identified are due to conceptual category, and not any of a series of processing-related confounds. The time intervals, frequency bands and scalp locations that proved most informative for prediction permit physiological interpretation: the widespread activation shortly after appearance of the stimulus (from 100ms) is consistent both with accounts of multi-pass processing, and distributed representations of categories. These methods provide a new alternative to fMRI for fine-grained investigations of the conceptual lexicon.

    KW - semantics

    KW - single concepts

    KW - neuroimaging

    KW - brain damage

    KW - data mining

    KW - machine learning

    U2 - 10.1016/j.bandl.2010.09.013

    DO - 10.1016/j.bandl.2010.09.013

    M3 - Article

    VL - 117

    SP - 12

    EP - 22

    JO - Brain and Language

    T2 - Brain and Language

    JF - Brain and Language

    SN - 0093-934X

    IS - 1

    ER -