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

In the realm of EEG decoding, enhancing the performance of artificial neuralnetworks (ANNs) carries significant potential. This study introduces a novelapproach, termed "weight freezing", that is anchored on the principles of ANNregularization and neuroscience prior knowledge. The concept of weight freezingrevolves around the idea of reducing certain neurons' influence on thedecision-making process for a specific EEG task by freezing specific weights inthe fully connected layer during the backpropagation process. This isactualized through the use of a mask matrix and a threshold to determine theproportion of weights to be frozen during backpropagation. Moreover, by settingthe masked weights to zero, weight freezing can not only realize sparseconnections in networks with a fully connected layer as the classifier but alsofunction as an efficacious regularization method for fully connected layers.Through experiments involving three distinct ANN architectures and three widelyrecognized EEG datasets, we validate the potency of weight freezing. Our methodsignificantly surpasses previous peak performances in classification accuracyacross all examined datasets. Supplementary control experiments offer insightsinto performance differences pre and post weight freezing implementation andscrutinize the influence of the threshold in the weight freezing process. Ourstudy underscores the superior efficacy of weight freezing compared totraditional fully connected networks for EEG feature classification tasks. Withits proven effectiveness, this innovative approach holds substantial promisefor contributing to future strides in EEG decoding research.

Weight Freezing: A Regularization Approach for Fully Connected Layers with an Application in EEG Classification