An Approach To Developing Eeg-Based Person Authentication System

The need for a new authentication method such as biometrics becomes apparent as the data breaches on password-based authentication increase. However, current biometric forms of authentication become unusable once compromised. Additional limits are realized when an attacker coerces an authorized user into a forced authentication. To resolve both issues, we propose creating an authentication mechanism that depends on the user's neurophysiological responses to chosen pieces of music (non-lyrical) measured using electroencephalographic (EEG) signals. This research paper provides a guide for creating and presenting a system that incorporates such idea for person classification and authentication. In a group study, the aim is that participant listen to individually selected music and music selected by other participants during an EEG reading. The change in the Alpha and Beta band frequencies across eight electrode EEG sensors serves as the input feature vector for a supervised machine learning algorithm that trains on the user and attacker EEG readings. Ultimately, the goal of the algorithm is to create a user-specific model to uniquely identify the respective user based on the corresponding EEG response to music and grant authentication. Our study lays a solid foundation for creating a promising EEG-based authentication system by solving the drawbacks of current biometric authentication methods.