Pilot study: Insights into the validity of Opencap to assess knee kinematics during the back squat

Abstract

Introduction

Advancements in technology have revolutionized the field of sports science, enabling researchers to gain deeper insights into the biomechanics of athletes' movements (Baca et al., 2022). Among these technological advancements, OpenCap has emerged as a powerful tool for biomechanical analysis (Uhlrich et al., 2023).  Integrating OpenCap with other innovative technologies, such as motion capture systems and data analytics, further enhances the understanding of human movement. However, the effective implementation of new technologies requires rigorous scientific validation and a nuanced approach to their comparison with already established techniques. The primary goal of this research is to establish basic characteristics for comparing biomechanical analysis results obtained from OpenCap and VICON.

Methods

A total of 3 healthy 23-29-year-old women with a minimum of one year of experience in strength training were recruited for the study. For OpenCap (opensimModel: LaiArnoldModified2017_poly_withArms_weldHand, posemodel: openpose, augmentermodel: v0.2), the system paired cameras of two iOS devices with a web app running on a standard laptop recording videos at 60Hz. A VICON system consisting of 10 cameras (200 Hz, Oxford Metrics Group, Oxford UK) was used for 3D motion analysis. The Plug-In Gait marker set (from the VICON system), comprising 42 markers with a diameter of 16mm, was employed along with two manual markers for tracking the barbell position. Ground reaction forces were measured using two force plates (1000 Hz, Kistler AG, Winterthur, CH). The collected data served as reference data for scaling the models and running OpenSim simulations. Barbell loads were adjusted based on body weight, with loads of 25% or 50% for the free-weight back squat. Each exercise consisted of three cycles of five repetitions, enabling subsequent calculation of average values for further analysis and evaluation.

Results

The analysis incorporated 22 valid squats. Statistical mapping revealed significant discrepancies in knee angles between Opencap and the OpenSim model throughout most of the back squat cycle, except from 10 to 35%. The average difference was 16.9 degrees (SD = 18.3 degrees). An RMSE of 24.9 and an ICC 3.1 of 0.503 (p = 0.006076) suggest moderate agreement between the models.

Discussion/Conclusion

Ongoing technological advancements are pushing the frontiers of sport biomechanics through groundbreaking innovations. To achieve a comprehensive assessment and comparison of new technologies, it is vital to perform correlation analyses on identical parameters produced by diverse models (Jing et al., 2023). The ICC results indicated a moderate agreement between the models, with the OpenCap model significantly underestimating knee angles. This lesser agreement, as compared to existing literature (Lima et al., 2023), may be attributed to the potential misinterpretation of the barbell by the model. Pending updates from the forthcoming augmenter model v0.3, we advise cautious application of Opencap technology in practical scenarios. However, we believe that with further technological advancement and continued refinement, it has the potential to greatly benefit the field of sports biomechanics.

References

Baca, A., Dabnichki, P., Hu, C. W., Kornfeind, P., & Exel, J. (2022). Ubiquitous computing in sports and physical activity—Recent trends and developments. Sensors, 22(21), Article 8370. https://www.mdpi.com/1424-8220/22/21/8370

Jing, Z., Han, J., & Zhang, J. (2023). Comparison of biomechanical analysis results using different musculoskeletal models for children with cerebral palsy. Frontiers in Bioengineering and Biotechnology, 11, Article 1217918. https://doi.org/10.3389/fbioe.2023.1217918

Lima, Y., Collings, T., Hall, M., Bourne, M., Diamond, L. (2023). Assessing lower-limb kinematics via OpenCap during dynamic tasks relevant to anterior cruciate ligament injury: A validity study. Journal of Science and Medicine in Sport, 26(Suppl2), S105. https://doi.org/10.1016/j.jsams.2023.08.123

Uhlrich, S. D., Falisse, A., Kidziński, Ł., Muccini, J., Ko, M., Chaudhari, A. S., Hicks, J. L., & Delp, S. L. (2023). OpenCap: Human movement dynamics from smartphone videos. PLoS Computational Biology, 19(10), Article e1011462. https://doi.org/10.1371/journal.pcbi.1011462

Acknowledgement

This study was funded by the Swiss National Foundation (192289).