Analysis of Prefrontal Single-Channel EEG Data for Portable Auditory ERP-Based Brain-Computer Interfaces

Area of Science:

• Neuroscience and Biomedical Engineering
• Brain-Computer Interfaces (BCI)
• Assistive Technology

Background:

• Electroencephalogram (EEG)-based brain-computer interfaces (BCIs) offer non-invasive control by translating brain activity.
• Auditory oddball paradigms enable event-related potential (ERP)-based BCIs without ocular activity, but traditional systems are costly and time-consuming.
• Amyotrophic lateral sclerosis (ALS) patients require accessible communication tools due to severe generalized myopathy.

Purpose of the Study:

• To develop a portable, cost-effective, single-channel EEG auditory oddball BCI system.
• To evaluate the performance of this BCI using natural sound stimuli.
• To assess the feasibility of using prefrontal single-channel EEG for practical communication aids.

Main Methods:

• Analyzed prefrontal single-channel EEG data from a consumer-grade device using a natural sound-based auditory oddball paradigm.
• Collected data from nine healthy subjects and one ALS patient.
• Quantified BCI performance using offline cross-validation and two online conditions.

Main Results:

• Offline analysis showed high detection accuracy and information transfer rates (ITR) across multiple command sets (e.g., 85.7% accuracy, 0.63 bits/min for two commands).
• First online analysis demonstrated strong performance for new data (80.0% accuracy, 1.16 bits/min for three commands).
• Second online analysis showed reduced performance on a subsequent day (62.5% accuracy, 0.43 bits/min for three commands), indicating potential for improvement.

Conclusions:

• Prefrontal single-channel EEG data is viable for developing user-friendly, portable auditory ERP-based BCIs.
• The developed paradigm shows promise for enabling communication in individuals with severe motor impairments.
• Further research is needed to optimize performance and ensure long-term reliability for daily use.