Muscle Synergy Patterns During Running: Coordinative Mechanisms From a Neuromechanical Perspective

TL;DR

This review examines muscle synergy analysis during running, highlighting neural control theories, decomposition methods, and how synergies are modulated by various factors, with implications for sports and rehabilitation.

Contribution

It synthesizes current knowledge on muscle synergy patterns during running, compares methods, and identifies gaps for future research in neuromechanical control.

Findings

Number and structure of synergies are stable across conditions.

Muscle weightings and primitives are highly adaptable.

Methodological variability affects synergy analysis results.

Abstract

Running is a fundamental form of human locomotion and a key task for evaluating neuromuscular control and lower-limb coordination. In recent years, muscle synergy analysis based on surface electromyography (sEMG) has become an important approach in this area. This review focuses on muscle synergies during running, outlining core neural control theories and biomechanical optimization hypotheses, summarizing commonly used decomposition methods (e.g., PCA, ICA, FA, NMF) and emerging autoencoder-based approaches. We synthesize findings on the development and evolution of running-related synergies across the lifespan, examine how running surface, speed, foot-strike pattern, fatigue, and performance level modulate synergy patterns, and describe characteristic alterations in populations with knee osteoarthritis, patellofemoral pain, and stroke. Current evidence suggests that the number and basic structure of lower-limb synergies during running are relatively stable, whereas spatial muscle weightings and motor primitives are highly plastic and sensitive to task demands, fatigue, and pathology. However, substantial methodological variability remains in EMG channel selection, preprocessing pipelines, and decomposition algorithms, and direct neurophysiological validation and translational application are still limited. Future work should prioritize standardized processing protocols, integration of multi-source neuromusculoskeletal data, nonlinear modeling, and longitudinal intervention studies to better exploit muscle synergy analysis in sports biomechanics, athletic training, and rehabilitation medicine.