Non-invasive brain-computer interfaces (BCIs) using EEG offers a new communication channel for humans with limited communication abilities. One type of such an EEG-based BCI can build upon decoding musical information to enable people to generate musical activity only by brain signals (Fig. 10). Music BCIs may rely on mental imagination of musical activity due to physical disability. Hence, it is essential to find EEG signatures of actively imagining music which are different from passively listening to music. This project addresses this by attempting to decode EEG into either imagining or listening to music. It also includes a control condition where participants simply watching musical instruments (e.g., piano keyboard). Using MLP, it can discriminate music imagining condition from listening or control EEG conditions with accuracy above 80%.