Feature Extraction and Recognition Methods for Brain-Computer Interfaces（Jinjing）
Brain-Computer Interfaces (BCIs) have reshaped the paradigm of human-computer interaction and represent a frontier field within national science and technology strategies.By investigating the inherent mechanisms of neural signal distortion, this work innovatively designs optimal encoding combination methods, breaking through the limitations of global information dependency and the lack of neurophysiological constraints. Furthermore, by integrating the dynamic evolution patterns of neural responses, the research enables precise task-target localization, effectively overcoming the challenge of feature misalignment. By analyzing the commonality patterns of signal populations, the project resolves the long-standing reliance of global data modeling on empirical data.The achievements of this work have been implemented in clinical applications across 30 domestic hospitals and more than 10 countries, serving over 1,000 patients. The resulting products were nominated for the Austrian National Innovation Award, recognized as a Most Investable Technological Achievement, and have successfully achieved large-scale industrialization.
