Encrypt with Your Mind: Reliable and Revocable Brain Biometrics via Multidimensional Gaussian Fitted Bit Allocation.

Area of Science:

• Neuroscience
• Biometrics
• Cybersecurity

Background:

• Traditional biometrics like fingerprints and facial scans are vulnerable to cloning and theft.
• Existing brain biometrics, often relying on electroencephalogram (EEG), suffer from low signal-to-noise ratio (SNR) and unstable performance.
• The inherent randomness of brain signals makes them a promising, yet largely untapped, biometric modality.

Purpose of the Study:

• To develop a high-performance brain biometric system for secure identity authentication.
• To introduce a novel key generation method utilizing intracortical brain signals.
• To overcome the limitations of existing EEG-based brain biometrics.

Main Methods:

• Proposed a novel brain-based key generation approach: multidimensional Gaussian fitted bit allocation (MGFBA).
• Utilized intracortical brain signals, specifically local field potentials, for higher resolution and SNR.
• Extracted keys from the brain signals of ten rats.

Main Results:

• Achieved an average effective key length of 938 bits using the MGFBA method.
• Demonstrated high authentication accuracy of 88.1% with a false acceptance rate of 1.9%.
• Showcased the ability to revoke MGFBA-based keys using different motor behaviors with high entropy.

Conclusions:

• Intracortical brain signals offer a viable alternative for constructing high-performance brain biometrics.
• The MGFBA method provides a reliable and high-entropy approach for brain-based key generation.
• This research highlights the potential of brain biometrics for advanced authentication and security applications.