Patient-Specific Movement Regime: Investigating the Potential of Upper-Extremity Motions for Intelligent Myoelectric Prosthetic Control

Amputation can have varying impacts on amputees depending on the extent of residual muscles left. Commercial prosthetics prioritize functional movements, neglecting patient-specific gestures related to the amputation type. Therefore, examining user-specific non-functional movements is pivotal. Rather than training on innumerable motions, prioritizing efficiently identifiable movements enhances prosthetic user-specificity. We examined 16 distinct wrist and finger motions using five classifiers (Wide Neural Networks, Ensemble Bagged Trees, Cubic SVM, Fine KNN, Logistic Regression Kernel). One-Way Anova testing (
$\text{p} < 0.05$
) revealed significant classifier differences, with Wilcoxon signed-rank tests (
$\text{p} < 0.05$
) distinguishing finger and wrist gestures. Ensemble Bagged Trees excelled (
$\text{p} < 0.05$
) for finger (98.2% ± 0.81) and wrist (99% ± 0.73) movements; Fine KNN and Cubic SVM scored up to 96.3% ± 0.81 and 95.9% ± 1.43, respectively. Notably, Abduction of thumb and flexion of others consistently showed high accuracies for finger gestures. Wrist Radial Deviation (Clockwise) and Wrist Extension emerged as top wrist movements. Our study addresses the oversight of non-functional movements in myoelectric control. Our innovative approach focuses on selecting optimal finger and wrist movements for accurate gesture recognition, a departure from prior research neglecting these critical movements. By pinpointing these movements’ significance and performance, we provide a solution to enhance personalized prosthetic development. Our findings mark a pioneering step in leveraging accurate finger and wrist gestures to craft more effective and personalized myoelectric control systems. These insights bridge the gap in existing research and pave the way for improved prosthetic device design.