Deep learning played a vital role in the seizure prediction challenge. Nevertheless, most studies used generic architectures that fail to consider the distinct characteristics of multivariate time-series Electroencephalography ...