Does Distinctiveness in Brain Signals of Parkinson's Impact Biometric Verification Performance?

Behavioral biometrics using brainwave signals captured via Electroencephalography (EEG) have gained increasing interest as a promising method for user verification. While most prior research has focused on EEG-based authentication for healthy individuals, the potential of brainwave signals from data subjects with neurological disorders remains unexplored. In this study, we investigate the feasibility of using EEG signals from individuals diagnosed with Parkinson's disease (PD) for biometric verification. To the best of our knowledge, this is the first study to explore neurological disorder-based biometric verification. We introduced Parkinson's EEG database (ParEEG database) of 5400 EEG samples obtained from both normal with Parkinson's disease subjects. Extensive experiments were performed to evaluate the verification performance using two handcrafted and two deep learning-based approaches on both normal and Parkinson's disordered datasets. The results obtained using the deep learning-based approach demonstrated outstanding performance for human verification with normal and Parkinson's disease subjects.