I. Introduction
Brain-computer interfaces (BCI) can be used to control devices by interpreting brain signals [1]. Electroencephalography (EEG), magnetoencephalography (MEG), and functional magnetic resonance imaging (fMRI) have been employed to measure brain activity [2]–[4]. These methods have been used to understand the different brain processes and changes in neural oscillations related to cognitive or motor functions. The use of EEG has been increasing in the last years due to its portability and low cost compared to the other above mentioned methods [5]. One of the biggest applications of EEG signals is BCI, which uses multiple paradigms such as evoked potentials [6], event-related potentials (ERP) [7], [8], motorimagery(MI) [L] and more [9].