Passthoughts authentication with low cost EarEEG.

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Neuroscience

Background:

• Wearable computing trends toward smaller, integrated devices.
• Existing authentication methods face challenges with seamless integration and user convenience.
• Electroencephalography (EEG) offers a potential biometric modality.

Purpose of the Study:

• To explore the feasibility of ear-based EEG for user authentication.
• To evaluate the performance of an earEEG authentication system using consumer-grade hardware.
• To assess the potential of mental gestures as an invisible authentication factor.

Main Methods:

• Collected EEG data from 12 participants using a low-cost, single-channel device.
• Participants performed 5 distinct mental tasks.
• Evaluated authentication accuracy using within-participant and between-participant analyses.

Main Results:

• Achieved a 44% reduction in Half Total Error Rate (HTER) compared to random classification.
• Attained 72% authentication accuracy in within-participant analyses.
• Demonstrated 60% reduction in HTER and 80% accuracy in between-participant analyses.

Conclusions:

• Ear-based EEG (earEEG) shows promise as a convenient and secure authentication method.
• The system is integrable into common wearable devices like earbuds.
• Mental gestures used for authentication are imperceptible to observers, enhancing security.