Frequency and phase mixed coding in SSVEP-based brain--computer interface.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Traditional frequency coding in steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) limits the number of selectable targets.
• This limitation is particularly restrictive for applications utilizing liquid crystal displays (LCDs) with multiple stimuli.

Purpose of the Study:

• To propose an innovative coding method for SSVEP that enhances the number of targets and improves the information transfer rate (ITR).
• To develop a BCI system capable of handling a larger number of targets than conventional methods.

Main Methods:

• A novel SSVEP coding method combining frequency and phase was developed.
• A BCI system with 15 targets was implemented using three stimulus frequencies.
• Electroencephalography (EEG) decoding was performed using Fourier coefficient projections onto reference phase directions.
• Optimization of lead position, reference phase, data segment length, and harmonic components was conducted.

Main Results:

• The proposed method enabled a BCI system with 15 targets, five times more than traditional methods.
• The average information transfer rate (ITR) exceeded 60 bits/min in a simulated online test with ten subjects.
• The decoding accuracy was enhanced through optimized parameters and the novel phase-coding approach.

Conclusions:

• The combined frequency and phase coding method significantly expands the target capacity of SSVEP-BCIs.
• This approach offers a viable solution for applications requiring a high number of stimuli, such as advanced LCD-based BCIs.
• The developed BCI system demonstrates a substantial improvement in ITR, paving the way for more complex and user-friendly brain-computer interfaces.