Surface electromyography evaluation for decoding hand motor intent in children with congenital upper limb deficiency

Area of Science:

• Biomedical Engineering
• Neuroscience
• Rehabilitation Technology

Background:

• Children with congenital upper limb absence possess measurable muscle control via surface electromyography (sEMG).
• Conventional sEMG classification methods are optimized for adults with acquired amputations, limiting their direct application to children with congenital limb differences.
• Adapting sEMG classification algorithms is essential for enabling dexterous prosthesis use in pediatric populations.

Purpose of the Study:

• To investigate the efficacy of sEMG classification techniques for decoding motor intentions in children with congenital upper limb absence.
• To develop and validate a generalized feature set for tuning sEMG classification algorithms in this specific cohort.
• To assess the potential of these tuned algorithms for controlling dexterous prostheses.

Main Methods:

• Collected sEMG data from 9 children with unilateral congenital below-elbow deficiency during 11 distinct hand movements.
• Employed five classification algorithms, utilizing time, frequency, and time-frequency domain features.
• Derived and validated a Congenital Feature Set (CFS) from participant-specific tuned algorithms, assessing its generalizability.

Main Results:

• Achieved an average offline classification accuracy of 73.8% ± 13.8% for 11 hand movements using the CFS across participants.
• Accuracy increased to 96.5% ± 6.6% when focusing on a reduced set of five movements.
• Demonstrated the generalizability of the CFS across the cohort, indicating robust motor intent decoding.

Conclusions:

• Individuals with congenital upper limb absence exhibit sufficient muscle control for effective sEMG-based prosthetic operation.
• The developed CFS shows promise for tuning sEMG classification algorithms, facilitating the translation of dexterous prostheses for children.
• This research supports the potential for advanced myoelectric prosthetic control in pediatric populations with limb differences.