Enhancing performances of SSVEP-based brain-computer interfaces via exploiting inter-subject information

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Steady-state visual evoked potential (SSVEP) based brain-computer interfaces (BCIs) are crucial for human-computer interaction.
• Target detection in SSVEP BCIs often requires extensive subject-specific training data.
• Existing methods may struggle with limited data or inter-subject variability.

Purpose of the Study:

• To introduce a novel training-free framework for SSVEP target detection.
• To enhance SSVEP detection by transferring templates between subjects.
• To improve the efficiency and accuracy of SSVEP-based BCIs.

Main Methods:

• Developed a framework utilizing joint frequency-phase coding for SSVEP detection.
• Introduced transfer template-based canonical correlation analysis (tt-CCA) for single and multi-channel data.
• Proposed an online adaptation method (ott-CCA) for updating electroencephalogram (EEG) templates.

Main Results:

• The tt-CCA method demonstrated an 18.78% increase in accuracy compared to standard CCA with a 1.5s data length in simulated experiments.
• The ott-CCA method achieved an additional 2.99% accuracy increase in simulated tests.
• The framework proved effective in enhancing SSVEP detection efficiency.

Conclusions:

• The proposed framework significantly improves the usability of SSVEP BCIs.
• Joint frequency-phase coding combined with inter-subject information transfer offers a promising approach for EEG-based BCIs.
• This study highlights the value of leveraging inter-subject data for enhanced BCI performance.