Design of a brain-machine interface for reducing false activations of a lower-limb exoskeleton based on error related...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Brain-Machine Interfaces (BMIs) offer intuitive exoskeleton control via motor imagery.
• Clinical application is limited by accuracy issues and false activations.
• Integrating Error Related Potentials (ErrP) detection can improve BMI accuracy and usability.

Purpose of the Study:

• Characterize ErrP at the start of gait for lower limb exoskeleton control.
• Reduce false starts in BMI systems.
• Determine the optimal feedback type (Tactile, Visual, Visuo-Tactile) for ErrP detection.

Main Methods:

• Designed an experimental protocol to evoke ErrP using Tactile, Visual, and Visuo-Tactile stimuli.
• Utilized iterative selection to characterize ErrP in time/frequency domains.
• Configured an ensemble classification system with a generic classifier for 6 subjects.

Main Results:

• The ensembled system corrected an average of 72.60% ± 10.23% of false starts.
• Tactile feedback was the most effective stimulus for evoking and detecting ErrP.
• Demonstrated significant improvement in system accuracy and reduced false starts.

Conclusions:

• Integrating ErrP detection with BMI-MI using Tactile feedback shows promise for reducing false starts.
• Enhanced system security and usability.
• Future research will focus on ErrP detection during gait for stopping maneuvers.