Enhancing Motor Imagery EEG Classification Using a Modified Bidirectional LSTM Network
International Journal of Scientific Research in Science, Engineering and Technology · 05.03.2026 · 13 (2) · 01-17
DOI: 10.32628/ijsrset261306 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Scientific Research in Science, Engineering and Technology, online erschienen am 05.03.2026. Verfasst von Iman Hussein AL-Qinani und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Motor imagery (MI) Electroencephalography (EEG) classification is a key step for supporting brain computer interface (BCI) applications, yet high accuracy remains challenging due to inter subject variability and signal noise. This paper uses the BNCI Horizon 2020 dataset under an all subjects together training setting to build general model that learns from all participants. Signal representation is improved through preprocessing scenarios including Normalization and principal component analysis (PCA), followed by two Bidirectional Long-Short-Term Memory (BiLSTM) architectures, a baseline design and a modified design with higher capacity and added batch normalization and dropout to stabilize training and reduce overfitting. Results show that the modified BiLSTM achieves strong multiclass performance, reaching 92.35% training accuracy and 92.23% testing accuracy for the 8-class task, and approximately 99.98% accuracy for both training and testing in the 7-class task. These findings highlight that combining appropriate preprocessing with an improved BiLSTM design can increase the reliability of motor imagery EEG classification for practical BCI use.
Zitieren
APA
AL-Qinani, I. H., & Alazzawi, Y. (2026). Enhancing Motor Imagery EEG Classification Using a Modified Bidirectional LSTM Network. International Journal of Scientific Research in Science, Engineering and Technology, 13(2), 01-17. https://doi.org/10.32628/ijsrset261306
Vancouver
AL-Qinani IH, Alazzawi Y. Enhancing Motor Imagery EEG Classification Using a Modified Bidirectional LSTM Network. International Journal of Scientific Research in Science, Engineering and Technology. 2026;13(2):01-17. doi:10.32628/ijsrset261306
BibTeX
@article{ALQinani2026,
author = {AL-Qinani, Iman Hussein and Alazzawi, Yarub},
title = {Enhancing Motor Imagery EEG Classification Using a Modified Bidirectional LSTM Network},
journal = {International Journal of Scientific Research in Science, Engineering and Technology},
year = {2026},
volume = {13},
number = {2},
pages = {01--17},
doi = {10.32628/ijsrset261306}
}
EUVIMED