Cognitive Cyber Crimes in the Era of Artificial Intelligence

10Brainwave Authentication:Threats and Vulnerabilities

Hitesh Rawat¹* and Anand Rajavat²

¹Computer Science Department, Shri Vaishnav Vidyapeeth Vishwavidyalaya, Indore, India

²Department of Computer Science Engineering, Shri Vaishnav Institute of Information Technology, Shri Vaishnav Vidyapeeth Vishwavidyalaya, Indore, India

Abstract

Brainwave authentication systems, leveraging electroencephalogram (EEG) signals, have emerged as a secure and unique biometric modality due to the inherent complexity and uniqueness of brainwave patterns. However, these systems face critical threats and vulnerabilities including spoofing, replay attacks, and signal injection. This study investigates these security gaps and proposes an enhanced deep wavelet convolutional neural network (DW-CNN) to improve the robustness of EEG-based authentication. Using the EEG motor movement/imagery dataset (PhysioNet), which includes EEG recordings from 109 subjects performing specific mental tasks, we evaluated the proposed method against common attack vectors. The DW-CNN achieved an authentication accuracy of 96.87%, a false acceptance rate of 1.21%, and a false rejection rate of 1.92%. Novel threat indicators such as Cognitive Disruption Index) and Signal Integrity Confidence Score were introduced, measuring at 0.97% and 93.8%, respectively, to further assess system resilience. The results indicate that advanced brainwave authentication systems can be significantly fortified against adversarial threats using ...