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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We present a simple deep learning-based framework commonly used in computervision and demonstrate its effectiveness for cross-dataset transfer learning inmental imagery decoding tasks that are common in the field of Brain-ComputerInterfaces (BCI). We investigate, on a large selection of 12 motor-imagerydatasets, which ones are well suited for transfer, both as donors and asreceivers. Challenges. Deep learning models typically require long trainingtimes and are data-hungry, which impedes their use for BCI systems that have tominimize the recording time for (training) examples and are subject toconstraints induced by experiments involving human subjects. A solution to bothissues is transfer learning, but it comes with its own challenge, i.e.,substantial data distribution shifts between datasets, subjects and evenbetween subsequent sessions of the same subject. Approach. For every pair ofpre-training (donor) and test (receiver) dataset, we first train a model on thedonor before training merely an additional new linear classification layerbased on a few receiver trials. Performance of this transfer approach is thentested on other trials of the receiver dataset. Significance. First, we lowerthe threshold to use transfer learning between motor imagery datasets: theoverall framework is extremely simple and nevertheless obtains decentclassification scores. Second, we demonstrate that deep learning models are agood option for motor imagery cross-dataset transfer both for the reasonsoutlined in the first point and because the framework presented is viable inonline scenarios. Finally, analysing which datasets are best suited fortransfer learning can be used as a reference for future researchers todetermine which to use for pre-training or benchmarking.

Transfer Learning between Motor Imagery Datasets using Deep Learning -- Validation of Framework and Comparison of Datasets