English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

EEG technology finds applications in several domains. Currently, most EEGsystems require subjects to wear several electrodes on the scalp to beeffective. However, several channels might include noisy information, redundantsignals, induce longer preparation times and increase computational times ofany automated system for EEG decoding. One way to reduce the signal-to-noiseratio and improve classification accuracy is to combine channel selection withfeature extraction, but EEG signals are known to present high inter-subjectvariability. In this work we introduce a novel algorithm forsubject-independent channel selection of EEG recordings. Consideringmulti-channel trial recordings as statistical units and the EEG decoding taskas the class of reference, the algorithm (i) exploits channel-specific1D-Convolutional Neural Networks (1D-CNNs) as feature extractors in asupervised fashion to maximize class separability; (ii) it reduces a highdimensional multi-channel trial representation into a unique trial vector byconcatenating the channels' embeddings and (iii) recovers the complexinter-channel relationships during channel selection, by exploiting an ensembleof AutoEncoders (AE) to identify from these vectors the most relevant channelsto perform classification. After training, the algorithm can be exploited bytransferring only the parametrized subgroup of selected channel-specific1D-CNNs to new signals from new subjects and obtain low-dimensional and highlyinformative trial vectors to be fed to any classifier.

Learning Signal Representations for EEG Cross-Subject Channel Selection and Trial Classification