Analysis of the Capability of Deep Learning Algorithms for EEG-based Brain-Computer Interface Implementation

Machine learning models have received significant attention for their exceptional performance in classifying electroencephalography (EEG) data. They have proven to be highly effective in extracting intricate patterns and features from the raw signal data, thereby contributing to their success in EEG classification tasks. In this study, we explore the possibilities of utilizing contemporary machine learning algorithms in decoding brain activity signals for a quick and efficient feature extraction in a potential BCI application. Specifically, the EEG data is associated with movement imagination as well as the state of relaxation. A total of 4 models based on neural networks, with distinct structures, were implemented and evaluated on a proprietary subject-specific dataset: EEGNet, EEG Inception, Spatial-Temporal Tiny Transformer (S3T), DeepConvNet. The experiments resulted in promising prediction accuracy. However, the performance of classifiers was not evaluated for new subjects or different hardware.