Exploring EEG-based biometrics for user identification and authentication

As human brain activities, represented by EEG brainwave signals, are more confidential, sensitive, and hard to steal and replicate, they hold great promise to provide a far more secure biometric approach for user identification and authentication. In this study, we present an EEG-based biometric security framework. Specifically, we propose to reduce the noise level through ensemble averaging and low-pass filter, extract frequency features using wavelet packet decomposition, and perform classification based on an artificial neural network. We explicitly discuss four different scenarios to emulate different application cases in authentication. Experimental results show that: the classification rates of distinguishing one subject or a small group of individuals (e.g., authorized personnel) from others (e.g., unauthorized personnel) can reach around 90%. However, it is also shown that recognizing each individual subject from a large pool has the worst performance with a classification rate of less than 11%. The side-by-side method shows an improvement on identifying all the subjects with classification rates of around 40%. Our study lays a solid foundation for future investigation of innovative, brainwave-based biometric approaches.