close
Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Aug 30;6(1):891.
doi: 10.1038/s42003-023-05204-3.

Network of muscle fibers activation facilitates inter-muscular coordination, adapts to fatigue and reflects muscle function

Sergi Garcia-Retortillo  1   2   3 , Carlos Romero-Gómez  3 , Plamen Ch Ivanov  4   5   6
Affiliations

Network of muscle fibers activation facilitates inter-muscular coordination, adapts to fatigue and reflects muscle function

Sergi Garcia-Retortillo et al. Commun Biol. .

Abstract

Fundamental movement patterns require continuous skeletal muscle coordination, where muscle fibers with different timing of activation synchronize their dynamics across muscles with distinct functions. It is unknown how muscle fibers integrate as a network to generate and fine tune movements. We investigate how distinct muscle fiber types synchronize across arm and chest muscles, and respond to fatigue during maximal push-up exercise. We uncover that a complex inter-muscular network of muscle fiber cross-frequency interactions underlies push-up movements. The network exhibits hierarchical organization (sub-networks/modules) with specific links strength stratification profile, reflecting distinct functions of muscles involved in push-up movements. We find network reorganization with fatigue where network modules follow distinct phase-space trajectories reflecting their functional role and adaptation to fatigue. Consistent with earlier observations for squat movements under same protocol, our findings point to general principles of inter-muscular coordination for fundamental movements, and open a new area of research, Network Physiology of Exercise.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Experimental set up, exercise protocol, raw EMG data and spectral power dynamics of different EMG frequency bands.
a Red dots mark EMG electrodes placement of selected muscles for simultaneous recording: left and right pectoralis major (ChestL and ChestR); left and right triceps brachii (ArmL and ArmR). b Schematic representation of push-ups. This fundamental movement requires continuous coordination of chest and arm muscles activity. c Dynamics in EMG signals recorded from each muscle represent the evolution of myoelectrical activity during push-up exercise. The exercise protocol includes three consecutive push-up bouts (Exercise 1, 2 and 3), each performed until exhaustion and separated by 5-min rest periods. EMG amplitude gradually increases with accumulation of fatigue during each exercise bout and higher EMG amplitude at Beginning of Exercise 2 and 3 indicates residual fatigue (“Study design and test protocol” section, Methods). d Time series representing spectral power dynamics of EMG frequency bands (corresponding to activation of different muscle fiber types), show synchronous spectral power modulation for different pairs of frequency bands in both muscle pairs ChestL-ChestR and ArmL-ArmR during Exercise 1 for two representative subjects. Blue lines show spectral power of frequency bands Fi from ChestL and ArmL muscles (left vertical axis in the panels), and red lines show spectral power of the bands from ChestR and ArmR muscles (right vertical axis in the panels). During the Beginning and Middle segment of Exercise 1 there is an almost perfect alignment of bursts in the spectral power of muscle fibers from different muscles with practically zero time delay — a behavior observed for frequency bands from both muscle pairs ChestL-ChestR and ArmL-ArmR (thus, our choice to use cross-correlation of the spectral power time series with zero time lag as a measure of coupling, Methods, “Cross-correlations between time series of EMG spectral power in different frequency bands”). Reduction in the degree of synchronous muscle fiber activation during the End segment of Exercise 1 is observed for both muscle pairs and is associated with accumulation of fatigue. This leads to reduction of the cross-correlation measure C(Fi,Fj) that results from two separate effects: (i) time shift of peaks in spectral power related to rhythmicity of the push-ups and (ii) loss of oscillatory activation pattern in the activity of different muscle fibers (frequency bands) and transition to a more stochastic/noisy behavior. Images used in Fig. 1a, b were extracted from the Shutterstock database, and appropriate permissions have been obtained for their usage.
Fig. 2
Fig. 2. Cross-correlation matrices representing networks of inter-muscular interactions among EMG rhythms during push-up movement.
a Intermuscular cross-correlation matrices where matrix elements represent the group-average coupling (degree of synchronization, Fig. 1d) between rhythms of myoelectrical activation for all pairs of muscles (ChestL-ChestR, ChestL-ArmL, ChestL-ArmR, ChestR-ArmL, ChestR-ArmR, ArmL-ArmR). Each matrix represents a subnetwork of intermuscular interactions among myoelectrical rhythms (10 frequency bands F1, F2, …, F10 with equal width of 19.5 Hz in the intervals [5-45 Hz] and [55-215 Hz]; for each pair of muscles. The Fi frequency bands correspond to the range of activity of different types of muscle fibers in each Chest and Arm muscle, i.e., F1 = [5-24.5 Hz], F2 = [25-44.5 Hz], F3 = [55-74.5 Hz], F4 = [75-94.5 Hz], F5 = [95-114.5 Hz], F6 = [115-134.5 Hz], F7 = [135-154.5 Hz], F8 = [155-174.5 Hz], F9 = [175-194.5 Hz] and F10 = [195-214.5 Hz]. Matrix rows show myoelectrical rhythms derived from the first muscle and matrix columns correspond to the rhythms in the second muscle for each pair; color code in vertical bars indicates coupling strength (based on cross-correlation C(Fi,Fj) of the spectral power amplitude of Fi and Fj, Fig. 1d). Heterogeneity of the matrices indicates complex organization of network links (coupling strength). b The average coupling strength for each intermuscular subnetwork exhibits a pronounced stratification pattern during Exercise 1, 2 and 3—bars represent the average value of all matrix elements for each matrix in (a), i.e., average links strength for each subnetwork. Error bars on top of each bar indicate the standard error; horizontal black dotted line marks the threshold of physiological significance for network interactions of myoelectrical rhythms across muscles based on a surrogate test where rhythms from different subjects are randomly coupled (“Fourier phase randomization surrogate test and significance threshold for links strength in networks of intermuscular interactions”, Methods).
Fig. 3
Fig. 3. Interaction networks of myoelectrical rhythms across muscles and their evolution with accumulation of fatigue during exercise.
a Dynamic networks of intermuscular interactions between major muscles involved in push-up movement show network reorganization with accumulation of fatigue during Exercise 1. Group-averaged network maps derived from the cross-correlation matrices for Exercise 1 (Fig. 2a), where network links are the cross-correlation matrix elements and represent the coupling strength (degree of synchronization) between myoelectrical rhythms (frequency bands Fi) corresponding to the degree of synchronous activity of muscle fibers from different muscles. Networks are derived from data corresponding the Beginning (first third) and End (last third) segments of Exercise 1. Each muscle is shown as semicircle with color nodes for the different frequency bands Fi (“Spectral decomposition”, Methods) within the muscle. Line color and width mark the links strength (Section “Intermuscular interaction networks”, Methods). Intermuscular interactions among muscle fibers form a multiplex network with pronounced heterogeneity characterized by distinct topology and hierarchical organization of links strength for subnetworks representing pairs of same- and different-type muscles. b Same-type muscle ChestL-ChestR and ArmL-ArmR subnetworks play main role in the push-up movement, exhibit hierarchical structure of links strength, and reorganize with accumulation of fatigue from Beginning to End segments of Exercise 1. Diagrams show distinct modules in these subnetworks for low (F1, F3), intermediate (F5) and high (F8, F10) frequency bands. These modules are obtained from the network maps in (a) using the same color code. c Stratification profiles of links strength for the ChestL-ChestR and ArmL-ArmR subnetworks reorganize with accumulation of fatigue from Beginning to End of Exercise 1. Six bars in each profile correspond to the group-average links strength of all muscle fiber interaction modules within each subnetwork. Error bars indicate standard error. Red stars mark statistically significant differences in links strength for all modules comparing Beginning vs End (Wilcoxon test p values < 0.05).
Fig. 4
Fig. 4. Interactions networks of myoelectrical rhythms across muscles and their evolution with fatigue.
a Dynamic networks of intermuscular interactions between major muscles involved in push-up movement show network reorganization with progression of fatigue for repeated bouts of exercise. Group-averaged network maps derived from the cross-correlation matrices for Exercise 1, 2 and 3 shown in Fig. 2a, where network links are the cross-correlation matrix elements and represent the coupling strength (degree of synchronization) between myoelectrical rhythms (frequency bands) corresponding to the degree of synchronous activity of muscle fibers from different muscles. Intermuscular interactions form a multiplex network with pronounced heterogeneity characterized by distinct topology and hierarchical organization of links strength for subnetworks representing pairs of same- and different-type muscles. Each muscle is shown as semicircle with color nodes for the different frequency bands (“Cross-correlations between time series of EMG spectral power in different frequency bands”, Methods) within the muscle. Line color and width mark the links strength (“Intermuscular interaction networks”, Methods). b Same-type muscle ChestL-ChestR and ArmL-ArmR subnetworks play main role in the push-up movement, exhibit hierarchical structure of links strength, and reorganize with fatigue during Exercise 2 and 3. Diagrams show modules in these subnetworks for low (F1, F3), intermediate (F5) and high (F8, F10) frequency bands. These modules are obtained from the network maps in (a) using the same color code. c Stratification profiles of links strength for the ChestL-ChestR and ArmL-ArmR subnetworks. Stratification profiles reorganize in response to fatigue during consecutive exercise bouts. Six bars in each profile for the ChestL-ChestR and ArmL-ArmR subnetwork correspond to the group-average links strength of all muscle fiber interaction modules within each subnetwork. Error bars indicate standard error. Red stars mark statistically significant differences in links strength for all modules comparing Exercise 3 vs Exercise 1 (Wilcoxon test p values with multiple tests Bonferroni correction < 0.025).
Fig. 5
Fig. 5. Subnetworks of synchronous activity among myoelectrical rhythms for pairs of different muscle types and their reorganization with fatigue.
a Muscle fiber interactions for the different-type muscle ChestR-ArmR and ChestR-ArmL subnetworks. These subnetworks play supportive/secondary role in push-up movement and exhibit specific hierarchical organization of links strength. Shown are the subnetwork modules for low (F1, F3), intermediate (F5) and high (F8, F10) frequency bands extracted from the network maps in Fig. 4a using the same color code. The other two subnetworks ChestL-ArmL and ChestL-ArmR of different muscle types exhibit similar network structure (modules) and reorganization with fatigue for repeated exercise bouts (matrices in Fig. 2a). b Links strength stratification profiles for the ChestR-ArmR and ChestR-ArmL subnetworks and profile reorganization for repeated exercise bouts with fatigue. Six bars in each profile correspond to the group-average links strength of all muscle fiber interaction modules within each subnetwork. Error bars indicate standard error. Red stars mark statistically significant differences in links strength for all modules comparing Exercise 3 vs. Exercise 1 (Wilcoxon test p values with multiple tests Bonferroni correction < 0.025).
Fig. 6
Fig. 6. Data statistics for average links strength in subnetwork modules and effects of accumulation and residual fatigue.
Boxplots display the median links strength and interquartile range for the average links strength of network modules within subnetworks of different muscle pairs obtained separately for each subject in the database. a Boxplots representing subject statistics for the average links strength of the six modules (shown in different box color) in the same-type muscle subnetworks ChestL-ChestR and ArmL-ArmR, comparing the Beginning and End segments of Exercise 1. Boxplots correspond to the links strength stratification profiles in Fig. 3c, and represent dispersity of data from individual subjects with the effect of accumulated fatigue from Beginning to End of Exercise 1. b Boxplots representing subject statistics for the average links strength of the six modules (shown in different box color) for the subnetworks of all muscle pairs, comparing Exercise 1 with Exercise 2 and 3. Boxplots correspond to the links strength stratification profiles in Figs. 4c and 5b, and show data dispersity from individual subjects with the effect of residual fatigue for Exercise 2 and 3. Statistics support the presence of distinct links strength profiles that uniquely quantify each muscle pair subnetwork.
Fig. 7
Fig. 7. Profiles of links strength and network modules within same-type muscle subnetworks and their modulation in response to accumulation of fatigue during exercise.
a Network interactions of frequency bands Fi (representing the activity of different muscle fiber types) within the ChestL-ChestR subnetwork. Subnetwork structure (left panels) and links strength profiles (right panels) of the ten basic network modules within the ChestL-ChestR subnetwork (each module represents the interaction of a given frequency band Fi from one muscle ChestL with all frequency bands from the other muscle ChestR) for the Beginning and End segments (top and bottom panels) of Exercise 1. Network structure and profiles of links strength dramatically reorganize with accumulation of fatigue during Exercise 1. Dynamic networks are derived from the group-averaged cross-correlation matrices for Exercise 1 in Fig. 2a, where network links correspond to the matrix elements and show the coupling strength (degree of synchronous activity; Fig. 1d) between distinct muscle fiber types from the ChestL and ChestR muscles (line color and width indicates network links strength; “Intermuscular interaction networks”, Methods). The ChestL-ChestR subnetwork topology is defined by the structure of the basic subnetwork modules — ten modules, with ten links each, form a complex hierarchical organization of links strength in the subnetwork which changes in response to accumulation of fatigue. Left panels: frequency bands Fi of the ChestL muscle are marked by circles and bands Fi of the ChestR muscle are marked by black squares. Same notation for Fi is used on the horizontal axis of the right panels with bar charts showing the link strength interaction profiles of the ten modules, where color bars within each profile correspond to the color of the network node (left panels) associated with a given frequency band in the Chest muscle. Red stars indicate statistically significant reduction in links strength for all modules comparing Beginning vs End of Exercise 1. Modulation of the profiles is marked by dash lines in color. b Same-type muscle ArmL-ArmR subnetwork (left panels) and links strength profiles for the ten subnetwork modules (right panels) with same notations as in (a). Modules in the ArmL-ArmR subnetwork exhibit similar links strength profiles as modules in the ChestL-ChestR subnetwork, however, with (i) weaker links and significantly higher stratification of links strength within each module for the Beginning segment of Exercise 1 (top right panel), (ii) more pronounced decline in links strength with fatigue accumulation for the End segment (bottom right panel), and (iii) preserved degree of links strength stratification within profiles comparing Beginning vs End (in contrast to the ChestL-ChestR subnetwork which shows less decline in link strength but a dramatic increase in profile stratification with accumulation of fatigue).
Fig. 8
Fig. 8. Profiles of links strength and network modules within same-type muscle subnetworks.
a Network interactions of frequency bands Fi (representing the activity of different types of muscle fibers) within the ChestL-ChestR subnetwork. Links strength in the ten basic network modules within the subnetwork (each module represents the interaction of a given frequency band Fi from one muscle ChestL with all frequency bands from the other muscle ChestR) dramatically reorganize with fatigue for consecutive exercise bouts (left panels with networks). Dynamic networks are derived from the group-averaged cross-correlation matrices for Exercise 1, 2 and 3 in Fig. 2a, where network links correspond to the matrix elements and show the coupling strength (degree of synchronous activity) between distinct muscle fiber types from the ChestL and ChestR muscles (Line color and width indicates network links strength; “Intermuscular interaction networks”, Methods). The ChestL-ChestR subnetwork topology is defined by the structure of the basic subnetwork modules —ten modules, with ten links each, form a complex hierarchical organization of links strength in the subnetwork which changes in response to fatigue. Left panels: frequency bands Fi of the ChestL muscle are marked by circles and bands Fi of the ChestR muscle are marked by black squares. Same notation for Fi is used on the horizontal axis of the right panels with bar charts showing the link strength interaction profiles of the ten modules, where color bars within each profile correspond to the color of the network node (left panels) associated with a given frequency band in the Chest muscle. Red stars indicate statistically significant differences in links strength for all modules comparing Exercise 1 vs Exercise 3 (Wilcoxon test p values with multiple tests Bonferroni correction < 0.025). Modulation of the profiles is marked by dash lines in color. b Same-type muscle ArmL-ArmR subnetwork (left panels) and links strength profiles for the ten subnetwork modules (right panels) with same notations as in (a). Modules in the ArmL-ArmR subnetwork exhibit similar links strength profiles as modules in the ChestL-ChestR subnetwork, however, with significantly higher stratification of links strength within each module, less pronounced decline in links strength with fatigue, and preserved degree of links strength stratification within profiles for all exercise bouts.
Fig. 9
Fig. 9. Profiles of links strength and network modules within different-type muscle subnetworks.
a Interaction networks for myoelectrical rhythms within the ChestR-ArmR subnetwork where links strength represents the degree of synchronous activation in the spectral power of distinct muscle fibers working in different frequency bands Fi. During Exercise 1, the ChestR-ArmR subnetwork topology is characterized by a hierarchical organization of ten basic modules, where each module represents the interaction of a given frequency band Fi from ChestR with all frequency bands from ArmR in the subnetwork. Right panels: frequency bands Fi for the ChestR muscle are marked by circles on the horizontal axis of each panel, and for the ArmR muscle are marked by black squares within each subnetwork module. Bars color in each module profile corresponds to the color of the node (left panels) associated with a given frequency band in the muscle. Red stars indicate statistically significant differences in links strength for all modules comparing Exercise 3 vs Exercise 1 (Wilcoxon test p values with multiple tests Bonferroni correction < 0.025). Color dash lines mark changes in the profile for the ten modules. b Same network representation and links strength profiles of subnetwork modules as in (a) are shown for the different-type muscle ChestR-ArmL subnetwork. Modules in the ChestR-ArmL subnetwork exhibit (i) similar shape for their links strength profiles as the modules in ChestR-ArmR subnetwork; (ii) similar modulation of the profiles shape with accumulation of fatigue during Exercise 2 and 3.
Fig. 10
Fig. 10. Degree of intermuscular coupling for the subnetworks involved in push-up movement, and change in the strength of network interactions with fatigue.
During Exercise 1, the average links strength for the same-type muscle subnetworks ChestL-ChestR and ArmL-ArmR exhibit intermediate to high degree of coupling. Note that ArmL-ArmR interactions are weaker than ChestL-ChestR. In contrast, the different-type muscle subnetworks ChestR-ArmR, ChestR-ArmL, ChestL-ArmL and ChestR-ArmR present low degree of coupling, indicating less synchronous activation of distinct types muscle fibers across these muscles. Further, this reflects distinct link strength profiles of subnetwork interactions that play role in coordinating muscle fibers activity between different-type muscles (Fig. 9) compared to the same-type muscle subnetworks (Fig. 8). With accumulation of fatigue during Exercise 2 and 3, interactions in the ChestL-ChestR and ArmL-ArmR subnetworks become remarkably weaker as these subnetworks account for the major dynamic effort in generating push-up movement, and are quickly affected by fatigue. Similarly, coupling strength for the different-type muscle Chest-Arm subnetworks is reduced with fatigue, however, the reduction is less pronounced compared to same-type muscle ChestL-ChestR and ArmL-ArmR subnetworks. Notably, Chest-Arm interactions on the right side of the body (ChestR-ArmR, ChestR-ArmL) are stronger than interactions in left side (ChestL-ArmL, ChestR-ArmR) since all participants in this study were right-handed.
See this image and copyright information in PMC

References

    1. Prilutsky B. l. Coordination of Two- and One-Joint Muscles: Functional Consequences and implications for Motor Control. Mot. Control. 2000;4:1–44. - PubMed
    1. Kristiansen M, Samani A, Madeleine P, Hansen EA. Effects of 5 Weeks of Bench Press Training on Muscle Synergies: A Randomized Controlled Study. J. Strength Cond. Res. 2016;30:1948–1959. - PubMed
    1. Comaduran Marquez D, von Tscharner V, Murari K, Nigg BM. Development of a multichannel current-EMG system for coherence modulation with visual biofeedback. PLOS ONE. 2018;13:e0206871. - PMC - PubMed
    1. Casolo, A. et al. Non-invasive muscle biopsy: estimation of muscle fibre size from a neuromuscular interface. bioRxiv 2022.10.21.513157 10.1101/2022.10.21.513157 (2022).
    1. Farina D. Counterpoint: Spectral properties of the surface emg do not provide information about motor unit recruitment and muscle fiber type. J. Appl Physiol. 2008;105:1673–1674. - PubMed

Publication types

LinkOut - more resources