Estimation of the most effective rhythm for human identification using EEG signal

Human identification using a special biological feature has become a promising field for the purpose of security system. Electroencephalogram (EEG) signals are the signature of human mind and can be used confidently as a strong biometric identifier. As EEG signals are consist of five different frequency bands, this paper represents a general methodology to determine the most effective rhythm for human identification. Using the different features from different rhythms in time and frequency domain, four neural networks are developed for the classification approach. Comparison of the designed neural networks shows that beta rhythm gives the best performance with a very low mean square error whereas delta rhythm gives the worst performance with comparative higher mean square error for identifying a person. It is concluded that beta rhythm is the most effective frequency band for human identification using EEG in resting and problem solving condition.