Maximizing information transfer rates in an SSVEP-based BCI using individualized Bayesian probability measures.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interfaces (BCIs) are crucial for translating user intentions into computer commands.
• Information Transfer Rate (ITR) is the standard metric for evaluating BCI performance.
• Existing BCI methods require optimization for speed and accuracy.

Purpose of the Study:

• To develop and validate a multi-step process for accelerating BCI intent detection and classification.
• To maximize the Information Transfer Rate (ITR) of a BCI system.
• To investigate the impact of performance feedback on BCI accuracy and ITR.

Main Methods:

• A Bayesian probability model was constructed using a ratio of Canonical Correlation Analysis (CCA) coefficients.
• A thresholding method was applied to posterior probabilities for classifying user intent.
• Probability thresholds were optimized per frequency and subject to maximize ITR.
• Two experimental sessions were conducted: one without and one with performance feedback.

Main Results:

• The proposed multi-step process achieved a maximum ITR of 39.82 bit/min.
• Performance feedback did not significantly enhance overall ITR.
• Performance feedback led to improvements in classification accuracy.
• The developed Bayesian model and thresholding method effectively classified user intent.

Conclusions:

• The novel multi-step BCI approach significantly enhances information transfer rates.
• While performance feedback boosts accuracy, its effect on ITR requires further investigation.
• This optimized BCI methodology holds potential for medical applications.