Comparison of manual and semi-automated methods for measuring fascicle contraction velocity following neuromuscular electrical stimulation from ultrafast ultrasound

Abstract

Introduction & Purpose

Ultrasound imaging is widely used to assess muscle and tendon properties and function, including muscle cross-sectional area, pennation angle, fascicle length, and musculotendinous behavior in vivo (Van Hooren et al., 2020). Recent technological advances in ultrafast ultrasound imaging (UUI) allow recordings at up to 5000 frames per second, thereby enabling the investigation of muscle dynamics even during rapid contractions (Deffieux et al., 2006).

However, accurately and reliably quantifying metrics such as fascicle contraction velocity from these recordings remains challenging. One approach involves manually measuring fascicle length to calculate shortening velocity. As this process is highly time-consuming, semi-automated tracking methods have been developed to estimate fascicle length frame by frame. One commonly used tool is UltraTrack (UT), which has been shown to provide reliable and accurate measurements (Farris & Lichtwark, 2016; Gillett et al., 2013). Nevertheless, UT has not yet been evaluated during rapid contractions recorded with UUI. Furthermore, findings by Cronin et al. (2011) suggest that the accuracy of this algorithm may decrease as movement velocity increases.

Therefore, this study aimed to investigate the reliability of UT and a custom semi-automated tracking method for assessing fascicle contraction velocity from UUI recordings of fast muscle contractions induced by supramaximal neuromuscular electrical stimulation (NMES), using manual analysis (MA) as the reference method.

Methods

Fifteen healthy physically active male participants completed two identical testing sessions on consecutive days using a test-retest design. Muscle contractions of the gastrocnemius muscle were induced via supramaximal NMES (STMISOLA, Biopac®, rectangular pulse, 1 ms) and recorded at 1000 frames/s using UUI (Aixplorer, Supersonic Imaging®). Several fascicle contraction velocity parameters were assessed using three analysis approaches: MA, a custom-developed block tracking method (BT), and UT. Test-retest reliability was determined using intraclass correlation coefficients (ICC 3, k), and agreement between methods was evaluated by comparing extracted contraction velocity parameters across sessions and analysis techniques.

Results

Manual analysis demonstrated good to excellent test-retest reliability (ICC = 089–0.92 for the analyzed velocity parameters, with average contraction velocity in the first 40 ms showing the highest reliability values). UT showed higher reliability than BT (ICC = 0.62–0.88 vs. ICC = 0.39–0.79). Between methods, only BT and UT showed significant correlation (r = 0.72; p < 0.05 for 40 ms mean velocity).

Discussion

The findings suggest that MA remains the most reliable approach for assessing contraction velocity from UUI recordings. Among the semi-automated methods, UT appears to be the more reliable. These findings confirm previous results reported by Gillett et al. (2013) and extend their applicability to UUI recordings. However, the low interchangeability between methods limits the significance of these findings.

Conclusion

This study demonstrates that UT represents a reliable semi-automated alternative for assessing fascicle contraction velocity from UUI. However, the low agreement between methods limits the interpretability of the findings. Future studies should include a broader comparison of currently available tracking approaches like UltraTimTrack (Zee et al., 2025) or the hybrid tracking method by Verheul & Yeo (2023) to identify the most accurate and reliable method.

References

Cronin, N. J., Carty, C. P., Barrett, R. S., & Lichtwark, G. (2011). Automatic tracking of medial gastrocnemius fascicle length during human locomotion. Journal of Applied Physiology, 111(5), 1491–1496. https://doi.org/10.1152/japplphysiol.00530.2011

Deffieux, T., Gennisson, J.-L., Tanter, M., Fink, M., & Nordez, A. (2006). Ultrafast imaging of in vivo muscle contraction using ultrasound. Applied Physics Letters, 89(18), 184107. https://doi.org/10.1063/1.2378616

Farris, D. J., & Lichtwark, G. A. (2016). UltraTrack: Software for semi-automated tracking of muscle fascicles in sequences of B-mode ultrasound images. Computer Methods and Programs in Biomedicine, 128, 111–118. https://doi.org/10.1016/j.cmpb.2016.02.016

Gillett, J. G., Barrett, R. S., & Lichtwark, G. A. (2013). Reliability and accuracy of an automated tracking algorithm to measure controlled passive and active muscle fascicle length changes from ultrasound. Computer Methods in Biomechanics and Biomedical Engineering, 16(6), 678–687. https://doi.org/10.1080/10255842.2011.633516

Van Hooren, B., Teratsias, P., & Hodson-Tole, E. F. (2020). Ultrasound imaging to assess skeletal muscle architecture during movements: a systematic review of methods, reliability, and challenges. Journal of Applied Physiology, 128(4), 978–999. https://doi.org/10.1152/japplphysiol.00835.2019

Verheul, J., & Yeo, S.-H. (2023). A Hybrid Method for Ultrasound-Based Tracking of Skeletal Muscle Architecture. IEEE Transactions on Biomedical Engineering, 70(4), 1114–1124. https://doi.org/10.1109/TBME.2022.3210724

Zee, T., Tecchio, P., Hahn, D., & Raiteri, B. (2025). UltraTimTrack: a Kalman-filter-based algorithm to track muscle fascicles in ultrasound image sequences. PeerJ Computer Science, 11, e2636. https://doi.org/10.7717/peerj-cs.2636