A strategy for minimizing the effect of misclassifications during real time pattern recognition myoelectric control

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Neuroprosthetics

Background:

• Pattern recognition myoelectric control combined with targeted muscle reinnervation (TMR) offers advanced real-time control for upper limb prostheses.
• Current systems achieve high off-line accuracy but face challenges with motion misclassifications during real-world use by amputees.

Purpose of the Study:

• To investigate the efficacy of a decision-based velocity profile to improve prosthesis positioning.
• To minimize unintended movements by limiting speed during classifier decision changes.

Main Methods:

• Two TMR surgery patients controlled virtual or physical prostheses.
• A Target Achievement Control Test was used to assess performance with and without the velocity ramp.
• A novel velocity ramp strategy was implemented to manage classifier decision changes.

Main Results:

• Participants demonstrated a 34% increase in completion rate and were 13% faster overall when using the velocity ramp.
• One participant using a physical prosthesis built a 7-block tower with the ramp versus a 2-block tower without it.
• The velocity ramp significantly improved performance metrics in prosthesis control tasks.

Conclusions:

• A decision-based velocity profile integrated with pattern recognition myoelectric control can enhance prosthesis performance.
• This approach mitigates issues from motion misclassifications, leading to better user control and task efficiency.
• The findings suggest a promising strategy for improving functional outcomes in upper limb prosthesis users.