Dataset Evaluation Method and Application for Performance Testing of SSVEP-BCI Decoding Algorithm.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Science

Background:

• Steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI) systems are widely researched.
• Existing SSVEP datasets have varying sample distributions and collection equipment, lacking a unified evaluation method.
• Current algorithm testing on limited datasets may not reflect real-world performance.

Purpose of the Study:

• To propose a unified SSVEP dataset evaluation method.
• To establish a comprehensive SSVEP algorithm evaluation dataset system.
• To analyze the performance of existing SSVEP decoding algorithms across diverse datasets.

Main Methods:

• Developed a novel SSVEP dataset evaluation methodology.
• Integrated six diverse SSVEP datasets using frequency and phase modulation paradigms.
• Tested four existing SSVEP decoding algorithms on the established dataset system.

Main Results:

• Algorithm performance varied significantly across different datasets.
• Substantial inter-subject performance differences were observed.
• The proposed evaluation system effectively highlights performance variations.

Conclusions:

• The developed SSVEP dataset evaluation system provides a robust platform for testing BCI algorithms.
• This system enables comprehensive algorithm verification across subjects, environments, and equipment.
• The findings offer valuable insights for advancing SSVEP decoding algorithms and BCI research.