Online asynchronous decoding of error-related potentials during the continuous control of a robot

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Error-related potentials (ErrPs) are neural signals indicating error processing.
• Detecting ErrPs offers a pathway to enhance brain-computer interface (BCI) performance.
• Asynchronous ErrP detection in BCIs remains an emerging research area.

Purpose of the Study:

• To investigate the feasibility of asynchronously decoding ErrPs in a real-time BCI setting.
• To evaluate the performance of online ErrP detection during robotic arm control.

Main Methods:

• Electroencephalography (EEG) was recorded from 15 participants controlling a robotic arm.
• A subset of trials (30%) featured unexpected halts to elicit ErrPs.
• An asynchronous classification approach was employed to detect ErrPs online.

Main Results:

• The system achieved an average true positive rate (TPR) of 70% for ErrP detection.
• An average true negative rate (TNR) of 86.8% was recorded, indicating reliable non-detection.
• Successful ErrP detection allowed participants to regain control of the robotic arm.

Conclusions:

• Online asynchronous decoding of ErrPs is feasible and reliable.
• This finding supports the integration of ErrP detection for advanced BCI applications.
• The study validates a promising method for real-time error monitoring in BCI systems.