Direct information transfer rate optimisation for SSVEP-based BCI.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Science

Background:

• Brain-computer interfaces (BCIs) enable communication and control through neural signals.
• Steady-state visual evoked potentials (SSVEPs) are commonly used in BCIs due to their robustness.
• Accurate classification of SSVEP signals is crucial for effective BCI performance.

Purpose of the Study:

• To propose a novel classification method for SSVEP-based BCIs.
• To optimize target classification by maximizing the information transfer rate (ITR).
• To develop a method that avoids common pitfalls in ITR calculation.

Main Methods:

• Feature extraction from traditional SSVEP BCI methods.
• Optimization of discrimination thresholds for each feature.
• Derivation of a generalized formula for ITR calculation, avoiding standard assumptions.

Main Results:

• The proposed method achieved a significant improvement in target classification performance.
• Achieved an information transfer rate (ITR) of 62 bits/min, doubling previous results on the same dataset.
• Demonstrated superior performance compared to existing methods on the benchmark dataset.

Conclusions:

• The developed method automatically calculates optimal discrimination thresholds, eliminating manual tuning and grid searches.
• This approach enhances the reliability of SSVEP-based BCIs by reducing false classifications.
• The method holds promise for practical, real-world BCI applications requiring high accuracy and efficiency.