Hierarchical feature fusion framework for frequency recognition in SSVEP-based BCIs

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Science

Background:

• Steady-state visual evoked potential (SSVEP) based brain-computer interfaces (BCIs) require effective frequency recognition algorithms.
• Current methods may have limitations in accurately identifying distinct brain responses to visual stimuli.

Purpose of the Study:

• To propose a novel hierarchical feature fusion framework for enhancing frequency recognition in SSVEP-BCIs.
• To introduce spatial dimension (SD) and frequency dimension (FD) fusion techniques within this framework.
• To evaluate the framework's performance using the correlated component analysis (CORRCA) method.

Main Methods:

• Developed a hierarchical feature fusion framework incorporating spatial dimension (SD) and frequency dimension (FD) fusion.
• Utilized a weighted strategy with a nonlinear function for both SD and FD fusions.
• Applied an extended correlated component analysis (CORRCA) method to assess the framework's efficacy on a benchmark dataset.

Main Results:

• The proposed hierarchical feature fusion framework, when integrated with an extended CORRCA method, demonstrated significant performance improvements.
• Experimental results from thirty-five subjects confirmed the superiority of the enhanced framework over the original CORCCA method.
• The framework effectively leverages feature fusion to enhance frequency recognition accuracy.

Conclusions:

• The developed hierarchical feature fusion framework offers a promising approach to significantly improve frequency recognition in SSVEP-BCIs.
• This framework has the potential to enhance the overall performance and reliability of SSVEP-based brain-computer interfaces.
• Further research can explore optimizing the fusion strategies and nonlinear functions for broader BCI applications.