Authentication using c-VEP evoked in a mild-burdened cognitive task

Area of Science:

• Biometric authentication
• Neuroscience
• Brain-computer interfaces

Background:

• Biomarker-based authentication is a growing research area.
• Code-modulated visual evoked potentials (c-VEP) are increasingly important for brain-computer interfaces.
• Individual discriminative characteristics within c-VEP signals are being explored for authentication.

Purpose of the Study:

• To investigate the feasibility of authentication using c-VEP signals.
• To develop and validate a deep learning method for c-VEP-based authentication.
• To assess the effectiveness of a mild-burdened cognitive task (MBCT) in evoking attention-modulated c-VEP signals for authentication.

Main Methods:

• A mild-burdened cognitive task (MBCT) was designed to ensure participant attention to visual stimuli.
• Deep learning models were trained using c-VEP data from a first session.
• The trained models were verified using c-VEP data from a second session with 17 participants.

Main Results:

• The deep learning models achieved an average accuracy of 0.92.
• The models demonstrated an average F1 score of 0.89.
• The study confirmed that c-VEP signals possess unique individual characteristics.

Conclusions:

• Authentication based on c-VEP signals within an MBCT is feasible.
• Deep learning effectively extracts discriminative features from c-VEP for authentication.
• The findings support the development of practical authentication systems utilizing c-VEP.