Toward practical BCIs: a BMNABC-based feature selection and sensor optimization framework for implicit learning...

Area of Science:

• Cognitive Neuroscience
• Biomedical Engineering

Background:

• Implicit learning is crucial for cognition and training.
• Identifying implicit learning requires advanced neuroimaging techniques.
• Current methods for analyzing multimodal brain data are limited.

Purpose of the Study:

• To develop a generalizable framework for feature selection and sensor optimization using simultaneous EEG and fNIRS.
• To identify implicit learning events more effectively.
• To improve the efficiency and user-friendliness of brain-computer interfaces (BCIs).

Main Methods:

• A two-stage optimization process using a binary multi-neighbor artificial bee colony (BMNABC) algorithm.
• Prioritizing optimal features from multimodal data using a normalized weighted sum (NWS) metric.
• Recursive backward elimination for sensor reduction in BCIs.

Main Results:

• The BMNABC framework successfully identified a superior feature set, significantly improving classification accuracy compared to single modalities.
• Selected features provided neurophysiological validation, isolating key biomarkers in the prefrontal cortex.
• High performance was maintained with up to 66% fewer sensors, demonstrating a sparse yet effective sensor configuration.

Conclusions:

• The proposed framework offers a data-driven method for detecting implicit learning.
• This approach advances the design of more efficient and user-friendly BCI systems.
• Simultaneous EEG and fNIRS combined with advanced optimization techniques show great promise for cognitive neuroscience research and BCI development.