“EMG Analysis of Redundant and Synergistic Muscle Coordination and Cross-Frequency Coupling During Grip Force Control”

Abstract

Understanding how the central nervous system coordinates synergistic and redundant muscle activations to achieve stable and adaptable movement remains a central question in motor neuroscience. This study investigates the structure of redundant and synergistic muscle coordination during grip force control and explores the modulation of high-frequency electromyographic activity by low-frequency oscillations, with cross-frequency coupling. Using EMG recordings from upper-limb muscles during object-lifting tasks, we applied an information-theoretic framework to decompose muscle interactions into redundant and synergistic components based on mutual information metrics. These metrics quantified how pairs of muscles shared or complemented task-relevant information about grip force. Non-negative matrix tri-factorization was used to extract low-dimensional spatial and temporal synergy modules. The analysis revealed that muscle redundancy was negligible, while spatial synergy typically featured one primary muscle exhibiting strong coordination with multiple others. Although the dominant synergistic muscle varied across participants, the deltoid consistently showed high synergy. Temporal synergy was more diverse and most pronounced during the early phase of grip control. Phase-amplitude coupling analysis was conducted to investigate whether low-frequency EMG phase modulated high-frequency amplitude, reflecting potential shared neural drives. Initial PAC estimation using Hilbert-transform based filtering without surrogates indicated modulation centered around a 2 Hz phase frequency, possibly influenced by trial alignment. A subsequent analysis using Morlet wavelet convolution with surrogate normalization revealed more individualized PAC patterns, where the most modulating phase frequency varied across participants. For each participant, two EMG channels were selected based on synergy strength, and a composite channel was formed by averaging their activity to represent a synergistic unit. Correlations between grip force and the three PAC measures were examined, along with synergy-grip force correlations. The deltoid-dominated channel exhibited the highest, though not statistically significant, correlation with grip force. These findings suggest that low-frequency oscillations modulate high-frequency EMG activity in a muscle and participant-specific manner and that both synergy and PAC may encode shared neural drives related to grip force control. Limitations include the restricted participant subset used for PAC, those with high synergy values, selective channel analysis, and the use of a simple averaging method to generate composite synergy signals. Future work should refine composite signal computation and directly relate PAC and synergy to enhance predictive modeling of physiological outcomes.