Revisiting Privacy Preservation in Brain-Computer Interfaces: Conceptual Boundaries, Risk Pathways, and a Protection-Strength Grading Framework
Brain–computer interfaces (BCIs) face multidimensional privacy risks in real-world applications, including neural data leakage, inference of sensitive information, and tight coupling between models and user privacy. This work proposes the first privacy-protection strength grading framework specifically designed for BCIs, establishing a systematic taxonomy across three dimensions: protected entities, lifecycle stages, and protection strength. Through conceptual modeling, risk pathway analysis, and joint evaluation, existing approaches are categorized into four protection levels. The framework emphasizes decoupling task-irrelevant sensitive information to achieve an optimal balance between privacy preservation and functional utility. By providing a structured foundation for understanding and addressing neural privacy threats, this study offers both theoretical grounding and technical guidance for tackling emerging challenges in neuroethics and privacy-aware BCI design.