Extending motor imagery by speech imagery for brain-computer interface

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Electroencephalogram (EEG)-based brain-computer interfaces (BCIs) translate brain activity into control signals.
• Current BCIs primarily rely on motor imagery, which has limited operational dimensions.
• Expanding BCI capabilities is crucial for broader applications.

Purpose of the Study:

• To extend the operational dimensions of EEG-based BCIs.
• To investigate the integration of speech imagery with motor imagery for enhanced BCI control.
• To evaluate the effectiveness of this combined approach in Chinese native speakers.

Main Methods:

• Developed a novel BCI system incorporating both motor and speech imagery.
• Collected EEG data from subjects performing speech imagery and combined motor-speech imagery tasks.
• Extracted feature vectors using common spatial patterns (CSP) and cross-correlation functions.
• Classified feature vectors using a support vector machine (SVM).

Main Results:

• The system demonstrated distinguishing accuracies between 79.33% and 88.26% for two intentions.
• Successful integration of speech imagery significantly enhanced BCI performance.
• The approach was validated with native Chinese speakers.

Conclusions:

• Combining motor imagery and speech imagery effectively extends the capabilities of EEG-based BCIs.
• This hybrid approach offers a promising direction for developing more versatile and powerful brain-computer interfaces.
• The findings have implications for improving human-computer interaction and assistive technologies.