Development of neuroauthenticator: Feasibility of an EEG-based authentication

摘要

We have been developing a cognitive type of brain- machine interface (BMI) system "Neurocommunicator," an EEG-based communication aid for people with severe motor disabilities. The system analyzes an event-related potential (ERP) to the sequentially flashed pictograms to indicate a desired message, and predicts the user's choice (target) in the brain. In this study, we focused on the individual differences of the waveform of ERP to the target, and we compared the pattern recognition model for the detection of the ERP among many subjects. The results showed that the user's own model was bestfitted to the data, indicating feasibility of an EEG-based authentication system "Neuroauthenticator".