Decision-Making Selector (DMS) for Integrating CCA-Based Methods to Improve Performance of SSVEP-Based BCIs.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interfaces (BCIs) are crucial for assistive technologies.
• Fusion-based approaches show promise for enhancing BCI performance.
• Decoding steady state visual evoked potentials (SSVEPs) is a key BCI application.

Purpose of the Study:

• To propose a novel decision-making selector (DMS) for integrating classification decisions from different canonical correlation analysis (CCA) methods.
• To improve the performance of SSVEP-based BCIs through a fusion strategy.

Main Methods:

• Developed a DMS to select the most accurate classification decision between two methods (M1 and M2) by analyzing false trial separation.
• Extracted feature vectors using correlation coefficients to measure decision uncertainty.
• Evaluated the DMS by integrating pairs of seven CCA-based algorithms (CCA, ITCCA, MsetCCA, L1-MCCA, FBCCA, ECCA, TRCA).

Main Results:

• The proposed DMS method demonstrated enhanced performance in integrating algorithms with similar accuracies.
• Experimental results were validated on a 40-target dataset from thirty-five subjects.
• Integration of algorithms using DMS led to improved classification accuracy.

Conclusions:

• The DMS is an effective method for improving SSVEP-based BCI performance.
• Decision fusion using the proposed DMS enhances the reliability and accuracy of BCIs.
• This approach offers a robust strategy for combining multiple decoding algorithms in BCI research.