Pass-thoughts Authentication System based on EEG Signals Using Artificial Neural Network

Abstract

Authentication with textual password has several limitations: passwords have low entropy in practice, are often difficult to remember, are vulnerable "shoulder surfing". Biometric system does not meet requirement as well. It relies upon unchanging features that have a lifetime as long as the individual. To avoid this limitation, we start to authenticate with thinking pass thought. User performs one mental task such as thinking of a word or phrase. In this study, Electroencephalography (EEG) was used as method for monitoring and recording the electrical activity of the brain. These signals can be captured and processed to get the useful information that can be used in pas-thoughts authentication system. Suitable analysis is essential for EEG to differentiate between best and worst tasks used for authentication. This study focuses on usefulness of EEG signal to identify best tasks suitable for the pass-thoughts authentication system. Artificial neural network (ANN) is used to train the data set. Then tests are conducted on the testing data of EEG signal to identify best and worst tasks suitable for authentication. Finally, the system performance was evaluated by computing the accuracy and therefore promising results were obtained.