14:00   Lower extremities & Motion III Chair: Kris Cuppens Exploring Healthy Knee Kinematic Phenotypes obtained through Dynamic CT Imaging: A Cluster Analysis Study Erin Teule, Sebastiaan van de Groes, Gerjon Hannink, Nico Verdonschot, Dennis Janssen Abstract: Dynamic Computed Tomography (CT) has become an important imaging modality for analysing knee joint kinematics. However, to effectively integrate dynamic CT into clinical settings, a detailed understanding of healthy knee movement is essential given the considerable variability observed within healthy individuals. The aim of this study was to define and describe healthy kinematic phenotypes by employing clustering analysis. Bilateral dynamic CT scans were obtained from 100 healthy participants, totalling to 200 knees. Eight knees from four participants were excluded due to technical errors. To ensure clinical relevance, knees were also excluded if a too small range of motion (defined as at least 5° to 60° of knee flexion) was achieved during dynamic CT scanning, which resulted in exclusion of another 72 knees from 40 participants. From the final dataset, consisting of 120 healthy knees from 64 participants, eight tibiofemoral (TF) and patellofemoral (PF) kinematic parameters were analysed, and K-means clustering was used to identify clusters with homogenous kinematics. The median and interquartile range was calculated for each kinematic parameter per cluster. Statistical Parametric Mapping (SPM) was used to detect kinematic differences between clusters. Two distinct clusters were identified, with 53 knees in cluster 1 and 67 knees in cluster 2. Significant differences in participant characteristics between the clusters were found for gender (14.1% (0.4 – 27.8) difference) and height (5.0 cm (2.0 – 7.0) difference). Furthermore, SPM analysis revealed significant differences in six out of eight kinematic parameters, both from TF and PF kinematics. The most evident differences were seen in TF rotations, with cluster 1 demonstrating greater internal and adduction rotation of the tibia compared to cluster 2. Furthermore, knees in cluster 1 demonstrated less lateral patellar tilt and less lateral translation of the patella compared to cluster 2. These findings offer valuable insights into the natural variability in healthy knee kinematics and underscore the importance of personalized approaches in treatment planning for knee pathologies. The identified kinematic phenotypes can be used as baseline for future research exploring pathological kinematic patterns through dynamic CT imaging. Musculoskeletal Modeling Reveals Biomechanical Factors Contributing To Clubfoot Recurrence Christian Greve, Bryce Killen, Han Houdijk, Joyce Bos, Sophie Moerman, Alessio Murgia Abstract: Introduction: Recurrence of idiopathic clubfoot affects 20-40% of cases due to imbalances between muscle and ligament forces. This study used musculoskeletal modelling of the foot to explore the biomechanics of clubfoot recurrence and assess the feasibility of personalizing muscle-tendon properties of the ankle-foot complex. The aim was to understand how reductions in evertor muscle strength, increases in joint stiffness, and gradually recurring clubfoot deformities affect ankle-foot moment balances during gait. Methods: Optimal muscle fiber lengths and tendon slack lengths of a 4 degree of freedom musculoskeletal model of the ankle-foot complex were optimized in OpenSim-Moco by solving an optimal control problem with electromyographic tracking and evaluating normalized muscle-tendon lengths. Evertor muscle strength was systematically reduced by decreasing isometric muscle tension from 100% to 5% of the original value. Joint stiffness was simulated by decreasing the tibialis’ posterior optimal muscle fiber length from 100% to 5% of its original value. Gradually recurring clubfoot deformities were simulated through systematic increases in ankle plantarflexion, subtalar inversion and midfoot supination from 0 to 18 degrees. Data from a typically developing child were used to run the simulations. Results: In the absence of deforming inversion forces, maintaining 20% of typical evertor muscle strength was sufficient to follow normal ankle-foot kinematics. However, when joint stiffness was simultaneously increased by 20% this minimum strength requirement increased to 50%, indicating the model's sensitivity to changes in muscle weakness and joint stiffness. Additionally, increasing clubfoot deformities did not disadvantage evertor moment arms compared to invertor moment arms. Discussion: Only in the presence of deforming inversion forces, evertor muscle weakness is a main contributor to the risk of clubfoot recurrence. Our findings emphasize the feasibility of using advanced musculoskeletal modelling to uncover mechanical factors that contribute to clubfoot recurrence. Incorporating muscle-tendon parameter optimization and manipulations of deep ankle-foot muscles increases the model’s ecological validity, and its biomechanical and clinical relevance of the results. This proof-of-concept study establishes a foundation for future research incorporating patient-specific data to validate the clinical relevance of this approach, potentially leading to more personalized and preventive treatment strategies, improving long term outcomes for children with clubfoot pathologies. Orientation and displacement estimation from a single inertial sensor: Demonstrated on atypical walking Anne Haitjema, Frank J. Wouda, Jasper Reenalda Abstract: Inertial measurement units (IMUs) are increasingly used in ambulatory human movement analysis due to their portability and ease of use. While accurate, IMU-driven biomechanical models typically require sensors on every body segment. From a user and data handling perspective, a reduced set of IMUs is preferable. The Drift-Free Orientation and Displacement (DFOD) algorithm accurately estimates lower leg orientation and displacement during running from a single sensor by leveraging the cyclical, planar characteristics of the movement [1]. This work extends the DFOD algorithm to accommodate a wider range of movements, with preliminary results for atypical walking presented here. One male subject walked on a treadmill (2.5 km/h, 1.5 minutes) while keeping his left knee stiff to create a less cyclical and less planar movement pattern. Data were collected using optical motion capture with cluster markers (200 Hz) and two custom IMUs (1000 Hz) placed on the lower legs. The DFOD algorithm was modified by employing a second-order Taylor approximation and by using quaternions instead of rotation matrices for orientation estimation. Sixty gait cycles were evaluated. The DFOD orientation estimates in the sagittal and forward planes showed excellent agreement with the optical reference and improved performance compared to [1], with Pearson correlations above 0.93 and RMSE below 3 degrees. Position estimates in the forward direction had Pearson correlations above 0.98 and RMSE below 3.7 cm, indicating slightly lower performance than [1]. For other orientations and positions, RMSE values were low (all below 4.2 degrees and 1.6 cm), though correlations ranged from poor to good (0.20–0.82), showing comparable performance to [1]. The estimates from the modified DFOD algorithm correspond well with the optical reference for most orientations and positions. Excellent agreement was observed for movements with substantial range of motion (ROM), such as those in the sagittal plane. Poor to good agreement was observed for movements with limited ROM, such as those in the transverse plane. These preliminary results demonstrate the potential of the modified DFOD algorithm for ambulatory assessment of more complex, 3D movements. This work is partly funded by TKI HTSM and Movella Technologies BV. References 1. M. A. Zandbergen, J. Reenalda, R. P. van Middelaar, R. I. Ferla, J. H. Buurke, and P. H. Veltink, “Drift-free 3D orientation and displacement estimation for quasi-cyclical movements using one inertial measurement unit: Application to running,” Sensors, vol. 22, pp. 956, 2022, doi: 10.3390/s22030956. Evaluation of a fatiguing run using Statistical Parametric Mapping over time Robbert van Middelaar, Jasper Reenalda Abstract: Fatigue during running is often studied because of its potential relationship with injuries. Studies examining fatigue often compare the start and end of the run, occasionally including a few intermediate phases. The parameters of interest are mostly joint angles or peak accelerations, at maxima/minima or at key moments such as initial contact or midstance. However, much of the data remains unconsidered. Statistical parametric mapping (SPM) is a statistical method to approach the entire stride cycle, which can be used repeatedly over time to detect changes within the stride cycles throughout a fatiguing run. It remains unclear how SPM can enhance our understanding of kinematic changes, such as knee and ankle angles, during a fatiguing run and how these changes are reflected in the stride cycle as fatigue progresses. 8 participants (4M/4F) ran on an instrumented treadmill on their best average 5K speed until complete exhaustion. Sagittal knee and ankle angles were evaluated during the fatiguing run, measured by three inertial measurement units (Movella Xsens MVN Link) at the thigh, tibia, and foot. 60 strides were taken at the start as a reference, versus a moving window of 10 strides used in a recurrent paired two-sample t-test SPM. An alpha-correction was applied, which changed over time. The result indicates a subject-specific magnitude of significant change in joint angle within the stride cycle over the entire fatiguing run. Data from a representative participant showed an increase in knee angle around midstance and midswing, while the ankle angle increased at toe-off and decreased throughout the swing phase in the later stages of the run. However, these significant changes still alternate with non-significant changes towards the end of the fatiguing run. This indicates the degree of variability in running kinematics due to fatigue, emphasizing the importance to investigate the entire run rather than focusing solely on the start and end. SPM is a versatile tool to analyse changes in time-series data, but is also effective for multiple testing when evaluating changes over time. Knee joint distraction in tibiofemoral osteoarthritis: Evaluating the impact of bone pins and springs on the mechanical performance of the KneeReviver device Famke Janssen, Thom Bitter, Nico Verdonschot, Dennis Janssen Abstract: Introduction: Tibiofemoral osteoarthritis significantly affects quality of life, particularly in active patients aged 45 to 65. Joint-preserving treatments like knee joint distraction are preferred over total knee arthroplasty due to high costs and lower success rates of revisions of total knee arthroplasties. A previous experimental study indicated that while the KneeReviver (KR) device effectively relieves pressure by creating a gap between the femur and tibia, the joint gap was not fully maintained at physiological loads, highlighting the need for more insight into which device specifications affect joint gap narrowing. This study investigates the contribution of bone pins and internal springs of the KR device to its mechanical performance within one patient. Methods: Finite Element models were developed to assess three device specifications: bone pin length, bone pin diameter, and active or blocked internal device springs. The finite element models were first validated against experimental data. The impact of the device specifics on contact parameters, including peak pressure, mean pressure, contact area, and total load on cartilage, as well as joint gap dynamics during opening and closing, was analyzed. Results: It was found that smaller bone pin diameters resulted in higher contact parameters, characterized by increased contact pressures and larger contact areas. Models with active springs demonstrated higher contact parameters compared to those with blocked springs, while longer bone pins were also associated with increased contact parameters. Furthermore, smaller diameters, active springs, and longer pins led to smaller joint gap creation during distraction. However, these specifications also resulted in more joint gap narrowing at lower loads during axial loading, suggesting a decrease in the total stiffness of the device. The findings underscore the critical roles of bone pins and internal springs in optimizing the KR device's mechanical function. Understanding how these parameters contribute to the mechanical working is vital for enhancing surgical placement, exploring patient-specific device options, and device design, improving distraction treatment outcomes. Future research should define acceptable load limits to determine whether the device must bear the full load or if some load through the tibiofemoral joint is acceptable. Insight in this is essential for refining the device's applications. Serious gaming training using augmented feedback for knee motor skills: randomized controlled trail Yiling Zhang, Hans Timmerman, Ming Cao, Elisabeth Wilhelm Abstract: The integration of serious games into rehabilitation has been increasingly employed to enhance motor skills and physical functions, especially for lower limb recovery[1][2]. The interactive exergames offer a dynamic platform that motivates patients and improves their motor learning through repetition and engagement[3]. Previous research has demonstrated that augmented feedback (AF), which combines both auditory and visual feedback, enhances skill acquisition by reinforcing correct movement patterns and correcting errors during practice[4][5]. This randomized controlled trail investigates the effectiveness of a custom-made exergame platform which is designed to enhance knee joint motor skills through AF. 14 participants (7 in the experimental group and 7 in the control group) engaged in a one-hour training program. In the experimental group, there were 5 females and 2 males, who were all right-leg dominant with the average age being 25.29 years, and a mean body mass index (BMI) of 22.36 kg/m2 ( standard deviation (SD): 3.88 kg/m2). The control group consisted of 1 female and 6 males; one of them was left-leg dominant, and the rest were all right-leg dominant. Control group participants had an average age of 28 years, and a mean BMI of 21.66 kg/m2 (SD: 2.63 kg/m2). The experimental group utilized the exergame platform with real-time AF during knee exercises, squatting and stretching. The control group performed the same exercises without AF. Both groups of participants underwent one round of pre-test (10 squatting test and 10 stretching test) 4 rounds of squat game training, 4 rounds of stretch game training and 1 round of post-test which is as same as the pre-test. The training sequence of these two modes of game were random for each participant, and the test sequence is consistent with the training sequence. The performance of all above movements was assessed by the difference between the actual knee joint angle of each test movement and the theoretically required knee joint angle of the task knee, here abbreviated as DAT. The actual knee joint angle was measured by two IMU sensors, which were attached on thigh and shank separately. The absolute average of DAT differences between pre- and post-test in experiment group squatting test, control group squatting test, experiment group stretching test, control group stretching test were 14.9557 ° (SD: 14.6431 °), 10.0612° (SD: 6.6061°), 8.4275° (SD: 8.7519°) and 6.7953° (SD: 4.8418°), respectively. A one side Mann-Whitney-U test was utilized to analyze whether AF knee joint motor training is more effective than the training without AF. The result revealed that the experiment group improved training outcomes compared to the control group, in squatting ( U = 89.0100, p = 0.0229) and in stretching ( U=71.0142, p = 0.0456). Participants receiving AF demonstrated greater improvements in the squatting and stretching tasks compared to the control group. The findings indicate that the integration of AF in exergame platforms has a positive effect on knee joint motor training. References: 1. Warburton, D. E., Bredin, S. S., Horita, L. T., Zbogar, D., Scott, J. M., Esch, B. T., & Rhodes, R. E. (2007). The health benefits of interactive video game exercise. Applied Physiology, Nutrition, and Metabolism, 32(4), 655-663. 2. Lange, B., Koenig, S., Chang, C. Y., McConnell, E., Suma, E., Bolas, M., & Rizzo, A. (2012). Designing informed game-based rehabilitation for motor therapy in a virtual world. International Journal of Disability and Human Development, 11(1), 95-101. 3. Burke, J. W., McNeill, M. D., Charles, D. K., Morrow, P. J., Crosbie, J. H., & McDonough, S. M. (2009). Serious games for upper limb rehabilitation following stroke. Proceedings of the 2009 Conference in Games and Virtual Worlds for Serious Applications, 103-110. 4. Magill, R. A. (2001). Augmented feedback in motor skill acquisition. In: J. L. Davis (Ed.), Motor Learning and Control: Concepts and Applications (pp. 191-220). McGraw-Hill. 5. Lohse, K. R., Shirzad, N., Verster, A., Hodges, N. J., & Van der Loos, H. F. M. (2013). Video games and rehabilitation: Using design principles to enhance engagement in physical therapy. Journal of Neurologic Physical Therapy, 37(4), 166-175.

end %-->