Umeå universitets logga

INTRODUCTION: Motor units (MUs) exhibit a continuum of phenotypes, often described along a spectrum from slower to faster units, that is reflected in differences of their firing behavior and mechanical twitch dynamics. In intact humans in vivo, understanding how these distributed MU properties relate to contractile speed and rate of force development remains an open challenge. Recent work has proposed a statistical approach to characterize MU twitch responses from firing properties [1], and large-cohort analyses have shown that MU conduction velocity systematically varies with recruitment threshold, consistent with differences in underlying fiber characteristics [2]. Despite these advances, no study has directly linked MU firing properties and conduction velocity to their actual mechanical twitch characteristics. Combining high-density surface EMG (HDsEMG) with ultrafast ultrasound (UUS) enables non-invasive, MU-specific observation of electrical activity and local tissue motion [3] creating the opportunity to resolve MU populations along a slow–fast phenotypic spectrum in vivo. In this study, we extract MU firing behavior from HDsEMG and estimate contractile features, and we correlate these with MU-specific twitch profiles derived from synchronized UUS recordings to probe how electrophysiological properties map onto mechanical twitch dynamics.

METHODS: The experimental protocol was described in a previous study [4]. Briefly, HDsEMG signals and UUS data were synchronously recorded from the tibialis anterior muscle of 10 healthy participants. HDsEMG signals were decomposed into MU spike trains across four trapezoidal contraction at 2, 5, 10, 20% MVC. Tissue velocity maps were computed through detecting the phase shifts of beamformed UUS signals along the depth direction from frame to frame. For each MU, spike-triggered averaging of velocity maps was performed within a 100-ms interval to extract MU-specific velocity twitch profiles [4]. For each MU, a proxy for contraction time was derived from the velocity twitch and correlated with a compound feature extracted using an eigenvector approach derived from firing properties, designed to resemble twitch characteristics [1]. This feature was obtained by projecting normalized firing properties (discharge rate and recruitment threshold) onto their first principal component (PC1 score).

RESULTS: MU-specific contraction time were extracted from UUS-derived velocity profiles (positive to negative peak duration 21.21±9.24 ms). A progressive reduction in the contraction time was observed for MUs recruited at 2%, 5%, and 10% MVC, with a small but statistically significant difference between 2% and 10% MVC (Kruskal-Wallis, p<0.05). A moderate linear relationship (R²=0.11) was observed between the contraction time and the PC1 score, which was statistically significant (p<0.005).

CONCLUSION: The eigenvector-based method appears to classify MUs from slower to faster units in a manner consistent with longer to shorter tissue velocity contraction times as derived from UUS, even within a limited range of contraction levels. This study demonstrates the feasibility of using HDsEMG to characterize MU-specific contractile properties and, in combination with UUS, to provide new insights into how MU phenotypes are reflected in force generation.

REFERENCES

[1] A. Gogeascoechea, et al., 10.1109/TNSRE.2023.3319959.

[2] A. Del Vecchio, et al., 10.1111/apha.12930.

[3] M. Carbonaro, et al., 10.1038/s41598-022-12999-4.

[4] E. Lubel et al., 10.1109/TNSRE.2023.3315146.