Optimized SSVEP-Based Framework for Improving Accuracy and Reliability in Brain–Computer Interfaces

Abstract

Steady-State Visual Evoked Potentials (SSVEPs) are the foundation of the non-invasive Brain–Computer Interfaces (BCIs), offering a stable method of neural communication.  Reliability and precision are, however, areas of greatest concern regarding practical application. The conventional methods of Canonical Correlation Analysis (CCA), Task-Related Component Analysis (TRCA), and Filter Bank Canonical Correlation Analysis (FBCCA) have been widely used but each of them has limitations—CCA is troubled by its sensitivity to noise, TRCA has greater need for computational demand, while FBCCA requires intensive calibration. To break these limitations, the present study proposes an Optimized SSVEP-Based Framework (OSF) that integrates an Adaptive Feature Optimization Method (AFOM) for pursuing higher classification efficiency and robustness. The framework utilizes optimized signal separation alongside adaptive filtering for enhancing detection efficiency amidst fluctuating situations. Comparative analysis with conventional approaches reveals that OSF achieves stunning improvements in correct classification percentage increased by 18%, information transmission rate increased by 21%, decision reliability enhanced by 16%, and latency reduced by 12%. These results also prove that the devised OSF not only strengthens the stability of SSVEP-based BCIs but also delivers greater consistency and adaptability that hold promise for effective utilization in assistive communications as well as neurorehabilitative applications.