Brain-Computer Interface Based Security Model Using Steady State Evoked Potential Modality

The internet has evolved into an indispensable element in our daily life, making it possible for us to make online purchases, handle financial transactions, and book online tickets. However, it also poses hazards such as financial threats, data privacy, and personal security. Organizations are also exposed to malicious software, hacking, and fraudulent activity. As security issues and forgeries become more prevalent, the necessity for strong user authentication approaches emerges. As a result, a trustworthy system is required to safeguard personal and organizational confidential data. This work proposes a visual brain-computer interface (BCI) based 2 × 2 spellers that make use of steady-state evoked potential (SSVEP) reactions to boost the authentication system’s effectiveness. The experiment employed two paradigm protocols: a colored flickering screen and flickering images against a black background. To prevent mental tiredness, subjects were shown four images at four target areas, with adequate spacing between flickering spots. System performance was tested using two signal classification algorithms (Canonical Correlation Analysis (CCA) and Minimum Energy Combination (MEC)). The accuracy of the two systems is 94.02 ± 1.47% and 95.65 ± 0.99%, respectively. To validate the results, a statistical analysis was performed on the parameters such as sensitivity, specificity, and F1-score, for both algorithms.