Attention-based EEG classification

In recent years, deep learning models become one of the more used techniques to perform EEG classification. In this work, we aim to explore the performance of different architectures, including the new idea of attention model, on different datasets related to different EEG conditions. The first dataset contains data from normal subjects and data from patients, the second clean signals and artifactual ones, while the third collects signals with focal and global seizures. The used architectures allow us to perform binary classification and discriminate among the two defined classes for each dataset.