An Authentication System using Neurological Responses to Music

As attacks against password-based authentication increase, the need for robust biometrics becomes apparent. Currently, two of the most popular biometric authentication systems are fingerprint and facial recognition. However, both of these biometrics become unusable once compromised. Also, an attacker might coerce the user to force authentication. Therefore, we propose an authentication mechanism that depends on the participant’s neurological responses to chosen pieces of music measured using electroencephalographic (EEG) signals. The current study proposes an authentication system that uses neurological responses to music for classification. Participants listened to individually selected music and music selected by other participants during an EEG reading. The change in the Alpha and Beta band frequencies across seven electrodes served as the input to a user specific K-Nearest Neighbors (KNN). The classifier attempts to determine if we can identify a user based on their EEG response to music. Our pilot data collection and analysis has shown promise of this authentication system with an accuracy rate between 76.4%-92.3%.