EEG-Based Lower-Limb Movement Onset Decoding: Continuous Classification and Asynchronous Detection.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Brain-machine interfaces (BMIs) leverage electroencephalography (EEG) to decode user intentions for controlling robotic devices, crucial for promoting neural plasticity and motor recovery.
• Active neural participation in rehabilitation enhances motor recovery opportunities.

Purpose of the Study:

• To present a novel framework for decoding lower-limb movement-related cortical potentials using continuous classification and asynchronous detection.
• To evaluate the performance of this framework in a customized gait trainer setting.

Main Methods:

• Experiments were conducted with 10 healthy subjects performing self-initiated ankle plantar flexion in a gait trainer.
• Movement onset was decoded from EEG signals using continuous classification and asynchronous detection.
• Features were analyzed, and the impact of limb side (left vs. right) on neural signatures and classification performance was evaluated.
• The proposed decoding framework was compared with existing online detection methods.

Main Results:

• No significant differences in neural signatures or classification performance were found between the left and right legs.
• The proposed framework achieved a higher true positive rate and lower false positive rate compared to existing methods.
• Comparable latencies were observed with current online detection techniques.

Conclusions:

• The study demonstrates the feasibility of the proposed framework for developing a closed-loop gait trainer.
• This technology holds significant potential for gait training and neurorehabilitation applications, particularly in clinical trials.