Enhanced low-latency detection of motor intention from EEG for closed-loop brain-computer interface applications

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interfaces (BCIs) utilize electroencephalogram (EEG) signals to control external devices.
• Movement-related cortical potentials (MRCPs) are increasingly used for detecting voluntary motor intentions in BCIs.
• Enhancing the real-time detection efficacy of MRCPs is crucial for improving closed-loop BCI performance.

Purpose of the Study:

• To propose and evaluate a novel manifold method, Locality Preserving Projection coupled with Linear Discriminant Analysis (LPP-LDA), for real-time MRCP detection from scalp EEG.
• To compare the performance of the LPP-LDA method against traditional approaches like the matched filter method.

Main Methods:

• The study employed the LPP-LDA algorithm for real-time detection of MRCPs from EEG data.
• An online experiment was conducted with nine healthy subjects to assess the system's performance.
• Performance metrics included true positive rate, false positive rate, and detection latency.

Main Results:

• The LPP-LDA method demonstrated a statistically significant higher true positive rate (79% vs. 68%) and fewer false positives (1.4/min vs. 2.3/min) compared to the matched filter approach.
• The proposed system achieved significantly shorter detection latency (315 ms vs. 460 ms).
• These improvements are critical for inducing neuroplasticity in closed-loop BCIs.

Conclusions:

• The LPP-LDA system provides a highly accurate and low-latency solution for real-time MRCP detection.
• The system is easily implemented online and functions as a versatile brain switch.
• It holds potential as a core component for BCI systems aimed at neuromodulation and motor function rehabilitation.