Human Intracranial EEG Biometric Identification

Abstract

This paper investigates the feasibility of using intracranial electroencephalographic (iEEG) signals for human biometric identification-a novel research area not previously explored. The study aims to address the limitations of traditional EEG biometric methods and to explore the potential of intracranial devices to mitigate these challenges. We developed a comprehensive pipeline that processes raw iEEG signals and assessed its effectiveness through same-session and cross-session testing. Utilising a repurposed iEEG dataset, our study achieved a same-session identification accuracy of 95.84% on 78 subjects. Our cross-session analysis further affirmed the resilience of iEEG signals against template ageing, a significant issue in traditional EEG biometrics, and demonstrated incremental learning capabilities. This proof-of-concept work demonstrates the potential of iEEG signals to be effectively used in human biometric identification.Clinical relevance- Establishes how intracranial electroencephalography (iEEG) can be leveraged for human biometrics.