A High-Speed Visual BCI Based on Hybrid Frequency-Phase-Space Encoding and High-Density EEG Decoding

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interface (BCI) technology enables direct brain-device communication.
• Current visual BCIs have limited information transfer rates (ITRs) due to underexploited spatial information and low recording resolution.
• High spatiotemporal resolution is crucial for capturing brain signal dynamics.

Purpose of the Study:

• To develop high-speed BCI systems by enhancing spatial information utilization.
• To propose a hybrid frequency-phase-space encoding method integrated with high-density electroencephalogram (EEG) recordings.
• To investigate the impact of electrode configuration and density on BCI performance.

Main Methods:

• Recorded EEG data using a 256-channel cap.
• Compared four parieto-occipital electrode configurations (66/256, 32/128, 21/64, 9/64).
• Implemented classical frequency-phase encoding and a novel hybrid frequency-phase-space encoding method in BCI paradigms.

Main Results:

• The hybrid method with higher electrode density configurations (66/256, 32/128, 21/64) significantly increased theoretical ITRs compared to the traditional 9/64 setup (up to 195.56%).
• An online BCI system achieved a high average actual ITR of (472.72 ± 15.06) bits per minute.
• Demonstrated that spatiotemporal encoding strategy and electrode density jointly determine achievable ITRs.

Conclusions:

• The proposed hybrid frequency-phase-space encoding method substantially improves BCI speed and efficiency.
• High-density EEG recordings and optimized electrode configurations are critical for high-speed visual BCIs.
• Findings provide quantitative guidelines for designing advanced high-speed visual BCI systems.