A time–frequency convolutional neural network for the offline classification of steady-state visual evoked potential responses

Hubert Cecotti

    Research output: Contribution to journalArticle

    31 Citations (Scopus)

    Abstract

    A new convolutional neural network architecture is presented. It includes the fast Fourier transform between two hidden layers to switch the signal analysis from the time domain to the frequency domain inside the network. This technique allows the signal classification without any special pre-processing and uses knowledge from the problem in the network topology. The first step allows the creation of different spatial and time filters. The second step is dedicated to the signal transformation in the frequency domain. The last step is the classification. The system is tested offline on the classification of EEG signals that contain steady-state visual evoked potential (SSVEP) responses. The mean recognition rate of the classification of five different types of SSVEP response is 95.61% on a time segment length of 1 s. The proposed strategy outperforms other classical neural network architecures.
    LanguageEnglish
    Pages1145-1153
    JournalPattern Recognition Letters
    Volume32
    DOIs
    Publication statusPublished - 11 Mar 2011

    Fingerprint

    Bioelectric potentials
    Neural networks
    Signal analysis
    Electroencephalography
    Network architecture
    Fast Fourier transforms
    Switches
    Topology
    Processing

    Cite this

    @article{b956f40036c44a2a84986ec8ef145be1,
    title = "A time–frequency convolutional neural network for the offline classification of steady-state visual evoked potential responses",
    abstract = "A new convolutional neural network architecture is presented. It includes the fast Fourier transform between two hidden layers to switch the signal analysis from the time domain to the frequency domain inside the network. This technique allows the signal classification without any special pre-processing and uses knowledge from the problem in the network topology. The first step allows the creation of different spatial and time filters. The second step is dedicated to the signal transformation in the frequency domain. The last step is the classification. The system is tested offline on the classification of EEG signals that contain steady-state visual evoked potential (SSVEP) responses. The mean recognition rate of the classification of five different types of SSVEP response is 95.61{\%} on a time segment length of 1 s. The proposed strategy outperforms other classical neural network architecures.",
    author = "Hubert Cecotti",
    year = "2011",
    month = "3",
    day = "11",
    doi = "10.1016/j.patrec.2011.02.022",
    language = "English",
    volume = "32",
    pages = "1145--1153",
    journal = "Pattern Recognition Letters",
    issn = "0167-8655",
    publisher = "Elsevier",

    }

    A time–frequency convolutional neural network for the offline classification of steady-state visual evoked potential responses. / Cecotti, Hubert.

    In: Pattern Recognition Letters, Vol. 32, 11.03.2011, p. 1145-1153.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - A time–frequency convolutional neural network for the offline classification of steady-state visual evoked potential responses

    AU - Cecotti, Hubert

    PY - 2011/3/11

    Y1 - 2011/3/11

    N2 - A new convolutional neural network architecture is presented. It includes the fast Fourier transform between two hidden layers to switch the signal analysis from the time domain to the frequency domain inside the network. This technique allows the signal classification without any special pre-processing and uses knowledge from the problem in the network topology. The first step allows the creation of different spatial and time filters. The second step is dedicated to the signal transformation in the frequency domain. The last step is the classification. The system is tested offline on the classification of EEG signals that contain steady-state visual evoked potential (SSVEP) responses. The mean recognition rate of the classification of five different types of SSVEP response is 95.61% on a time segment length of 1 s. The proposed strategy outperforms other classical neural network architecures.

    AB - A new convolutional neural network architecture is presented. It includes the fast Fourier transform between two hidden layers to switch the signal analysis from the time domain to the frequency domain inside the network. This technique allows the signal classification without any special pre-processing and uses knowledge from the problem in the network topology. The first step allows the creation of different spatial and time filters. The second step is dedicated to the signal transformation in the frequency domain. The last step is the classification. The system is tested offline on the classification of EEG signals that contain steady-state visual evoked potential (SSVEP) responses. The mean recognition rate of the classification of five different types of SSVEP response is 95.61% on a time segment length of 1 s. The proposed strategy outperforms other classical neural network architecures.

    U2 - 10.1016/j.patrec.2011.02.022

    DO - 10.1016/j.patrec.2011.02.022

    M3 - Article

    VL - 32

    SP - 1145

    EP - 1153

    JO - Pattern Recognition Letters

    T2 - Pattern Recognition Letters

    JF - Pattern Recognition Letters

    SN - 0167-8655

    ER -