A BCI based visual-haptic neurofeedback training improves cortical activations and classification performance during...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Motor imagery (MI) performance in brain-computer interfaces (BCIs) can be limited by sensorimotor cortical activation.
• Traditional neurofeedback training (NFT) methods may not fully optimize BCI capabilities.

Purpose of the Study:

• To develop and evaluate a novel BCI-based visual-haptic neurofeedback training (NFT) system.
• To enhance sensorimotor cortical activations and improve classification performance during motor imagery (MI).
• To investigate the correlations between neurofeedback patterns and brain network dynamics.

Main Methods:

• A 64-channel electroencephalographic (EEG) system was used to record data from 19 healthy subjects.
• NFT incorporated synchronous visual scene and proprioceptive electrical stimulation feedback.
• Feedback was driven by real-time lateralized relative event-related desynchronization (lrERD) during MI tasks.

Main Results:

• Significant enhancement in multi-band absolute ERD powers and lrERD patterns post-NFT.
• Mean classification accuracy improved by approximately 9% to 85%.
• Significant correlations observed between lrERD patterns and classification accuracies; increased functional connectivity in sensorimotor networks.

Conclusions:

• The proposed visual-haptic NFT is feasible and effective for improving sensorimotor cortical activation and BCI performance during MI.
• This method offers a promising optimization for conventional NFT and evaluation of motor training effectiveness.