Noninvasive neuroimaging enhances continuous neural tracking for robotic device control.

Area of Science:

• Neuroscience
• Robotics
• Biomedical Engineering

Background:

• Intracortical brain-computer interfaces (BCIs) offer high-dimensional robotic control but require invasive surgery and specialized expertise.
• The limited accessibility of current BCIs hinders their widespread clinical and home use.
• A noninvasive BCI with high-quality control is needed to broaden BCI integration.

Purpose of the Study:

• To develop and validate a noninvasive framework using electroencephalography (EEG) for controlling robotic devices.
• To enhance user engagement and the spatial resolution of neural data for improved BCI performance.
• To demonstrate the practical application of this noninvasive BCI framework in controlling a robotic arm.

Main Methods:

• Utilized electroencephalography (EEG) for noninvasive neural signal acquisition.
• Implemented EEG source imaging to improve the spatial resolution of neural data.
• Developed a continuous pursuit task and training paradigm to enhance user engagement.
• Integrated online noninvasive neuroimaging for real-time BCI control enhancement.

Main Results:

• The framework significantly enhanced BCI learning, improving performance by nearly 60% on center-out tasks and over 500% on continuous pursuit tasks.
• Online noninvasive neuroimaging further improved BCI control by almost 10%.
• Successfully transitioned control from a virtual cursor to a physical robotic arm in real-time.

Conclusions:

• The developed noninvasive EEG-based BCI framework offers a practical and effective solution for robotic device control.
• This advancement significantly improves BCI learning, neural decoding quality, and practical utility.
• The framework holds major implications for the future development and implementation of noninvasive neurorobotics.