Rapid Nasal Breathing as a Biometric Trigger: High-Accuracy Electroencephalogram-Based Authentication for Clinical Applications.

BACKGROUND: Traditional biometric systems are vulnerable to forgery, highlighting the need for secure alternatives. Electroencephalography (EEG) offers inherent advantages in liveness detection and antispoofing but typically requires external stimuli. We propose a novel paradigm leveraging intrinsic respiratory-evoked EEG signals for identity authentication, with potential applications in clinical settings where unobtrusive monitoring is critical. METHODS: We developed a 64-channel EEG acquisition system with synchronized respiratory event monitoring. Thirteen healthy volunteers performed four breathing patterns: oral, nasal, slow nasal, and rapid nasal breathing. A hybrid deep learning model was designed to optimize spatial-temporal feature extraction from EEG signals. RESULTS: The model achieved 98.3% accuracy in identity recognition using rapid nasal breathing-evoked EEG, outperforming traditional biometric methods. Nasal breathing patterns consistently yielded higher accuracy than oral breathing, with rapid nasal breathing showing the strongest discriminative power. CONCLUSIONS: Respiratory-evoked EEG signals provide a viable, noninvasive biometric identifier. The high accuracy of rapid nasal breathing opens avenues for clinical integration, such as continuous patient authentication in respiratory monitoring devices or secure access to electronic health records.