Pattern recognition and direct control home use of a multi-articulating hand prosthesis

Area of Science:

• Prosthetics and Bionics
• Rehabilitation Engineering
• Neuroprosthetics

Background:

• Multi-articulating hand prostheses are increasingly available, but effective user control remains a significant challenge.
• Transradial amputation presents unique challenges for restoring hand function and dexterity.

Purpose of the Study:

• To investigate and compare the home use of myoelectric direct control versus pattern recognition control for multi-articulating hand prostheses.
• To evaluate the usability and effectiveness of different control strategies in individuals with transradial amputation.

Main Methods:

• Four participants with transradial amputation were fitted with a multi-articulating hand prosthesis and a Coapt COMPLETE CONTROL system.
• Training was provided for both direct and pattern recognition control styles, with grip selection optimized based on clinician and user feedback.
• Home trial data on usage patterns, including trial length and time powered on, were collected and analyzed.

Main Results:

• No significant differences were observed in trial length, days powered on, or total time powered on between direct and pattern recognition control.
• Direct control utilized a higher average number of configured grips (4.8) compared to pattern recognition (3.8), though this difference was not statistically significant.
• Users spent over 80% of hand closing time in a single grip with direct control, while pattern recognition usage was distributed across multiple grips (45% in one, 25% in a second, 20% in a third).

Conclusions:

• Pattern recognition control may offer a more intuitive method for users to select and utilize the diverse grips of multi-articulating hand prostheses.
• While direct control allowed for more configured grips, pattern recognition demonstrated a more distributed and potentially versatile grip usage pattern in home use.
• Further research is warranted to optimize pattern recognition algorithms for enhanced intuitive control and user experience in upper limb prosthetics.