RT info:eu-repo/semantics/article
T1 EEGSym: Overcoming Inter-Subject Variability in Motor Imagery Based BCIs With Deep Learning
A1 Perez-Velasco, Sergio
A1 Santamaria-Vazquez, Eduardo
A1 Martinez-Cagigal, Victor
A1 Marcos-Martinez, Diego
A1 Hornero, Roberto
K1 Brain computer interface (BCI), deep learning (DL), motor imagery, transfer learning, inter-subject
AB In this study, we present a new Deep Learning (DL) architecture for Motor Imagery (MI) based Brain Computer Interfaces (BCIs) called EEGSym. Our implementation aims to improve previous state-of-the-art performances on MI classification by overcoming inter-subject variability and reducing BCI inefficiency, which has been estimated to affect 10-50% of the population. This convolutional neural network includes the use of inception modules, residual connections and a design that introduces the symmetry of the brain throughthemid-sagittalplane into the network architecture. It is complemented with a data augmentation technique that improves the generalization of the model and with the use of transfer learning across different datasets. We compare EEGSym’s performance on inter-subject MI classification with ShallowConvNet, Deep-ConvNet, EEGNet and EEG-Inception. This comparison is performed on 5 publicly available datasets that include left or right hand motor imagery of 280 subjects. This population isthe largest that has been evaluated in similar studies to date. EEGSym significantly outperforms the baseline models reaching accuracies of 88.6±9.0 on Physionet, 83.3±9.3 on OpenBMI, 85.1±9.5 on Kaya2018, 87.4±8.0 on Meng2019 and 90.2±6.5 on Stieger2021. At the same time, it allows 95.7% of the tested population (268 out of 280 users) to reach BCI control ( 70% accuracy). Furthermore, these results areachieved using only 16 electrodes of themore than 60 available on some datasets. Our implementation of EEGSym, which includes new advances for EEG processing with DL, outperforms previous state-of-the-art approaches on intersubject MI classification.
PB IEEE
SN 1534-4320
YR 2022
FD 2022-06-27
LK https://uvadoc.uva.es/handle/10324/81636
UL https://uvadoc.uva.es/handle/10324/81636
LA eng
NO IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 30, pp. 1766-1775, 2022
NO Producción Científica
DS UVaDOC
RD 16-ene-2026