I. Introduction
Brain-computer interface (BCI) systems can assist disabled people who cannot use their bodies to communicate with the external environment [1]–[3]. To date, electroencephalography (EEG) has been mainly applied to the development of BCI systems, compared to other brain imaging modalities, such as functional magnetic resonance imaging (fMRI) and magnetoencephalography (MEG), due to relatively low cost and easy set up. BCI systems based on conventional EEGs require the attachment of recording electrodes on the scalp with conductive gels for accurate signal measurements, which limits the application value of BCI systems in terms of practical use. In recent, to overcome the limitation of conventional EEG-based BCI systems, EEGs measured around the ears, called Ear-EEG, has been proposed that can allow unobtrusive EEG measurement [4]–[6].