A novel method for movement quality analysis of lower limb joints using surface electromyography signals and k-means clustering technique

• Published in: Biomedical signal processing and control 2024-09, Vol.95, p.106455, Article 106455
• Main Authors: Sam Jeeva Raj, Edward Jero, Palaniappan, Rajinikumar
• Format: Article
• Language: English
• Subjects: Coactivation index; K-means clustering; Movement quality; Muscle synergy; Nonnegative matrix factorization method; Surface electromyogram
• ISSN: 1746-8094; 1746-8108
• DOI: 10.1016/j.bspc.2024.106455

•A new metric, synergy index, is introduced to analyze hip, knee, and ankle movements.•Using this metric, we obtained a functional interpretation of lower limb joint movements in five different walking tasks.•This study reveals distinct muscle coordination patterns in Heel and Toe walking tasks.•The k-means method is employed to extract commonalities in joint movements across different walking tasks.•We address synergy weight variations within the common muscle synergy across different walking tasks. The limb is the most common site for sports injuries. The objective of this study is to validate lower limb muscle synergy as a biomarker for movement quality. The proposed method comprises the following steps: (i) extraction of muscle synergies using nonnegative matrix factorization, (ii) estimation of hip, knee, and ankle joint movements from synergy weights, (iii) identification of common muscle synergies across five walking tasks using k-means clustering and (iv) conducting cluster-wise analysis of muscle contraction intensity during lower limb joint movements using synergy weights. In this study, we used surface Electromyography signals from the Lencioni et al. 2019 dataset recorded during normal walking, heel walking, toe walking, stair ascending and stair descending tasks involving fifty heterogeneous subjects. For the first time in the literature, we performed a functional interpretation of lower limb joint movements during these five walking tasks based on synergy weights and compared the results with the multiple muscle co-activation index (p