Semi-simulation Experiments for Quantifying the Performance of SSVEP-based BCI after Reducing Artifacts from...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Muscular artifacts frequently contaminate electroencephalograms (EEGs), impairing brain-computer interface (BCI) functionality.
• Existing artifact reduction methods primarily focus on minimizing reconstruction errors, with limited evaluation of their impact on BCI performance.

Purpose of the Study:

• To evaluate a state-of-the-art muscular artifact reduction technique for steady-state visual evoked potentials (SSVEPs)-based BCIs.
• To assess the relationship between artifact reduction effectiveness and BCI classification accuracy.

Main Methods:

• A semi-simulation approach was employed using a benchmark dataset of SSVEPs contaminated with muscular artifacts from the trapezius muscle.
• The performance of the artifact reduction technique was evaluated in conjunction with a task-related component analysis (TRCA) classification algorithm.

Main Results:

• Combining the artifact reduction method with the TRCA classifier led to enhanced classification accuracy.
• The artifact reduction settings that minimized reconstruction errors did not consistently maximize BCI classification performance.

Conclusions:

• Artifact reduction methods for BCIs should be optimized to maximize BCI performance rather than solely focusing on minimizing waveform reconstruction errors.
• Tailoring artifact removal strategies to specific BCI applications is essential for achieving optimal performance.