LASSO Homotopy-Based Sparse Representation Classification for fNIRS-BCI.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Brain-computer interface (BCI) systems enhance communication and control for individuals with motor impairments.
• Functional near-infrared spectroscopy (fNIRS) offers a non-invasive method for monitoring brain activity.
• Effective channel selection is crucial for improving the accuracy of fNIRS-based BCIs.

Purpose of the Study:

• To enhance the classification accuracy of fNIRS-BCI systems for motor rehabilitation.
• To investigate the efficacy of Least Absolute Shrinkage and Selection Operator (LASSO) homotopy-based sparse representation for channel selection.
• To decode walking activity and resting states using fNIRS signals.

Main Methods:

• Collected fNIRS signals from nine subjects' left motor cortex during walking and resting states.
• Applied filtering to remove motion artifacts and physiological noise.
• Utilized LASSO homotopy-based sparse representation for significant channel selection and classification.

Main Results:

• The LASSO homotopy-based sparse representation classification method successfully discriminated between walking and resting states.
• Achieved an average classification accuracy of 91.32% (p < 0.016).
• Outperformed classification based on traditional statistical spatial features.

Conclusions:

• LASSO homotopy-based sparse representation is effective for channel selection in fNIRS-BCI systems.
• The proposed methodology significantly improves classification accuracy for motor state decoding.
• This approach holds potential for controlling assistive devices and developing active rehabilitation techniques.