A brain-controlled lower-limb exoskeleton for human gait training.

Area of Science:

• Neuroscience
• Robotics
• Rehabilitation Engineering

Background:

• Brain-computer interfaces (BCIs) offer a novel method for translating human intentions into robotic movement commands.
• Existing BCIs often rely on single electroencephalogram (EEG) signal types, limiting their effectiveness.

Purpose of the Study:

• To develop and validate an EEG-based brain-controlled lower-limb exoskeleton for gait training.
• To propose a novel BCI framework integrating two asynchronous EEG signal modalities: sensorimotor rhythms (SMRs) and movement-related cortical potentials (MRCPs).

Main Methods:

• Experiments were conducted using a customized lower-limb exoskeleton with six human subjects.
• Subjects controlled the exoskeleton using their brain signals through a framework combining SMRs and MRCPs.
• Data collection involved offline training, online visual feedback, and online robot control, followed by post hoc analysis.

Main Results:

• The SMR-based method achieved an average robot-control accuracy of 80.16% ± 5.44%.
• The MRCP-based method yielded an average performance of 68.62% ± 8.55%.
• All subjects successfully controlled the exoskeleton, demonstrating the feasibility of the proposed framework.

Conclusions:

• The integrated BCI framework is feasible for controlling a lower-limb exoskeleton.
• This approach shows promise for human-in-the-loop rehabilitation and gait training.
• The paradigm could be extended to clinical trials for paraplegic patients.