Multifactor Authentication System Using Simplified EEG Brain–Computer Interface

Abstract

This article proposes a scheme for multifactor authentication based on electroencephalography (EEG) signal analysis. A solution for EEG signal acquisition and recording of acquisition results has been implemented, a machine learning model has been developed and trained, then a classifier that determines the user's login procedure familiarity has been built, and a solution to carry out the described experiments has been implemented, in the form of a mobile application. Besides, a multifactor authentication system, based on the EEG signal combined with user image verification, using the brain–computer interface system with a single EEG electrode based on the NeuroSky MindWave device was proposed. Based on the defined scenarios, experiments were conducted, followed by a survey on the research group, and the obtained results were analyzed. In the case of an experiment related to login simulation, a high classification accuracy rate was obtained, both for the classifier itself (83.33%) and the proposed user authentication system (77.78%). Analyzing the results of the EEG signal recording used in the classification, it seems that the proposed solution is promising not only due to the high accuracy and a low false rejections rate but also through confirmed associations in the analysis of brain wave signal, corresponding to the results of research in the literature. A proposed multi-factor authentication system based on image selection and EEG analysis can be implemented in many areas as a modern solution in securing IT systems.