An asynchronous artifact-enhanced electroencephalogram based control paradigm assisted by slight facial expression.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Existing electroencephalogram (EEG)-based control systems often lack real-time capability, asynchronous logic, and robustness.
• Artifacts in EEG signals can hinder accurate and reliable control.
• Slight facial expressions (sFE) offer a potential non-invasive modality for augmenting EEG control.

Purpose of the Study:

• To develop and evaluate an asynchronous artifact-enhanced EEG-based control paradigm assisted by slight facial expressions (sFE-paradigm).
• To improve the real-time performance, asynchronous functionality, and robustness of EEG control.
• To assess the sFE-paradigm's effectiveness in real-world manipulation tasks.

Main Methods:

• Developed a core algorithm involving exclusion of non-sFE-EEGs, interface 'ON' detection, sFE-EEGs real-time decoding, and validity judgment.
• Utilized component analysis for dominant component estimation and signal processing.
• Conducted brain connectivity analysis to understand dynamic directional interactions.

Main Results:

• Offline assessment achieved 96.46% accuracy for interface 'ON' detection and 92.68% for sFE-EEG decoding, with a <200 ms output timespan.
• Online evaluations demonstrated stability and agility in object-moving (60.03% IoU) and water-pouring (202.5 ml) tasks.
• Performance showed no significant difference compared to commercial control methods (FlexPendant, Joystick).

Conclusions:

• The sFE-paradigm effectively addresses limitations in real-time capability, asynchronous logic, and robustness for EEG-based control.
• This novel approach enables reliable non-invasive EEG control for real-world applications.
• The sFE-paradigm offers a promising solution for enhanced human-computer interaction.