Attention-Driven EEG Signal Analysis for Robust Brain-Computer Interface Applications

Abstract

Brain-Computer Interfaces (BCIs) hold great potential for enabling nonmuscular communication by translating neural activity into actionablecommands. However, the inherent variability and noise withinelectroencephalography (EEG) signals pose significant challenges toaccuracy and reliability. This study presents a comprehensive investigationinto the role of attention mechanisms—both traditional and Transformerbased—in enhancing EEG signal classification for BCI applications.Through empirical evaluation across age-based user groups andvariousclassifiers, our results demonstrate that Support Vector Machines and