User authentication using EEG signal

Abstract

A security system can be defined as a procedure to give protection for different types of data. A security system must perform a sequential process to achieve good protection. User authentication is a method to verify a user’s identity to allow them to access and use the system. A biometric system is one of the best security systems available in this technologydriven era. In today’s world, Biometric authentication is widely used to protect personal data against unauthorized access. Electroencephalography (EEG)-based authentication is considered one of the most recent developments. The subject-specific nature of brain responses in EEG, coupled with the uniqueness of everyone’s brain wave pattern, poses a significant challenge in artificially recreating it. This paper introduces a user authentication system that utilizes EEG signals. The system selects different groups of electrode channels (8, 16, 32, or 64) from the EEG Motor Movement/Imagery dataset and extracts features from each channel. To condense the feature matrix, principal component analysis (PCA) is employed. The Support Vector Machine (SVM) serves as the classifier for subject authentication. The results demonstrate an accuracy of 92.20%, 92.70%, 95.13%, and 95.41% for the 8, 16, 32, and 64 channels, respectively.