Feature Selection Combining Filter and Wrapper Methods for Motor-Imagery Based Brain-Computer Interfaces.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Machine Learning

Background:

• Brain-computer interfaces (BCIs) offer a way for individuals with movement disorders to control external devices.
• Common Spatial Pattern (CSP) is a key algorithm for extracting features in motor imagery (MI) tasks.
• Traditional CSP methods face challenges due to noise and non-stationarity in electroencephalography (EEG) data, limiting feature optimality.

Purpose of the Study:

• To develop an improved CSP feature selection framework for more effective MI-based BCIs.
• To enhance the accuracy and reliability of decoding MI tasks.
• To optimize feature extraction for BCIs in patients with movement disorders.

Main Methods:

• A novel CSP feature selection framework was designed, integrating filter and wrapper methods.
• Feature importance was evaluated using infinite latent feature selection and Wasserstein distance.
• An improved binary gravitational search algorithm (IBGSA) was developed and applied to a refined CSP feature subspace.

Main Results:

• The proposed method effectively reduced CSP features by half, creating a new feature subspace.
• Experiments on three public BCI datasets demonstrated comparable classification accuracies to existing studies.
• The presented model showed superior performance compared to other methods on the same datasets.

Conclusions:

• The novel CSP feature selection framework significantly improves MI decoding accuracy.
• The developed IBGSA algorithm effectively identifies optimal feature sets within the refined subspace.
• This approach holds promise for advancing BCIs for individuals with movement disorders.