Machine learning application in soccer: a systematic review.

Due to the chaotic nature of soccer, the predictive statistical models have become in a current challenge to decision-making based on scientific evidence. The aim of the present study was to systematically identify original studies that applied machine learning (ML) to soccer data, highlighting current possibilities in ML and future applications. A systematic review of PubMed, SPORTDiscus, and FECYT (Web of Sciences, CCC, DIIDW, KJD, MEDLINE, RSCI, and SCIELO) was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. From the 145 studies initially identified, 32 were fully reviewed, and their outcome measures were extracted and analyzed. In summary, all articles were clustered into three groups: injury (n = 7); performance (n = 21), which was classified in match/league outcomes forecasting, physical/physiological forecasting, and technical/tactical forecasting; and the last group was about talent forecasting (n = 5). The development of technology, and subsequently the large amount of data available, has become ML in an important strategy to help team staff members in decision-making predicting dose-response relationship reducing the chaotic nature of this team sport. However, since ML models depend upon the amount of dataset, further studies should analyze the amount of data input needed make to a relevant predictive attempt which makes accurate predicting available.

Ubiquitous Computing in Sports and Physical Activity-Recent Trends and Developments.

The use of small, interconnected and intelligent tools within the broad framework of pervasive computing for analysis and assessments in sport and physical activity is not a trend in itself but defines a way for information to be handled, processed and utilised: everywhere, at any time. The demand for objective data to support decision making prompted the adoption of wearables that evolve to fulfil the aims of assessing athletes and practitioners as closely as possible with their performance environments. In the present paper, we mention and discuss the advancements in ubiquitous computing in sports and physical activity in the past 5 years. Thus, recent developments in wearable sensors, cloud computing and artificial intelligence tools have been the pillars for a major change in the ways sport-related analyses are performed. The focus of our analysis is wearable technology, computer vision solutions for markerless tracking and their major contribution to the process of acquiring more representative data from uninhibited actions in realistic ecological conditions. We selected relevant literature on the applications of such approaches in various areas of sports and physical activity while outlining some limitations of the present-day data acquisition and data processing practices and the resulting sensors' functionalities, as well as the limitations to the data-driven informed decision making in the current technological and scientific framework. Finally, we hypothesise that a continuous merger of measurement, processing and analysis will lead to the development of more reliable models utilising the advantages of open computing and unrestricted data access and allow for the development of personalised-medicine-type approaches to sport training and performance.

The Effects of Visual Feedback on Performance in Heart Rate- and Power-Based-Tasks during a Constant Load Cycling Test.

Performance feedback can be essential for cyclists to help with pacing their efforts during competitions and also during standardized performance tests. However, the choice of feedback options on modern bike computers is limited. Moreover, little research on the effectiveness of the currently used feedback methods is available. In this study, two novel feedback variants using a bar or a tacho to visualize targets and deviation from targets were compared to a classic design using only numbers. Participants (6 female and 25 male trained to well-trained athletes) completed a protocol consisting of three heart rate-based tasks and one power-based task. The displays were compared with respect to their ability to guide athletes during their trials. Results showed lower root mean square error (RMSE) of the novel variants, but no significant effect of feedback variant on RMSE was found for both tasks (p > 0.05). However, when comparing the feedback variants on a person to person basis, significant differences were found for all investigated scenarios (p < 0.001). This leads to the conclusion that novel feedback variants can improve athletes' ability to follow heart rate-based and power-based protocols, but even better results might be achieved by individualizing the feedback.

Concurrent Validity of Field-Based Diagnostic Technology Monitoring Movement Velocity in Powerlifting Exercises.

ABSTRACT: Mitter, B, Hölbling, D, Bauer, P, Stöckl, M, Baca, A, and Tschan, H. Concurrent validity of field-based diagnostic technology monitoring movement velocity in powerlifting exercises. J Strength Cond Res 35(8): 2170-2178, 2021-The study was designed to investigate the validity of different technologies used to determine movement velocity in resistance training. Twenty-four experienced powerlifters (18 male and 6 female; age, 25.1 ± 5.1 years) completed a progressive loading test in the squat, bench press, and conventional deadlift until reaching their 1 repetition maximum. Peak and mean velocity were simultaneously recorded with 4 field-based systems: GymAware (GA), FitroDyne (FD), PUSH (PU), and Beast Sensor (BS). 3D motion capturing was used to calculate specific gold standard trajectory references for each device. GA provided the most accurate output across exercises (r = 0.99-1, ES = -0.05 to 0.1). FD showed similar results for peak velocity (r = 1, standardized mean bias [ES] = -0.1 to -0.02) but considerably less validity for mean velocity (r = 0.92-0.95, ES = -0.57 to -0.29). Reasonably valid to highly valid output was provided by PU in all exercises (r = 0.91-0.97, ES = -0.5 to 0.28) and by BS in the bench press and for mean velocity in the squat (r = 0.87-0.96, ES = -0.5 to -0.06). However, BS did not reach the thresholds for reasonable validity in the deadlift and for peak velocity in the squat, mostly due to high standardized mean bias (ES = -0.78 to -0.63). In conclusion, different technologies should not be used interchangeably. Practitioners who require negligible measurement error in their assessment of movement velocity are advised to use linear position transducers over inertial sensors.

Medium-term effects of a two-desk sit/stand workstation on cognitive performance and workload for healthy people performing sedentary work: a secondary analysis of a randomised controlled trial.

Implementing sit/stand workstations in sedentary work environments is a common way to reduce sedentary time, but their medium-term effect on cognitive performance is unclear. To address this circumstance, eighteen office workers participated in a two-arm, randomised controlled cross-over trial (ClinicalTrials.gov Identifier: NCT02825303), either working at a traditional (sit) or an interventional (sit/stand) workplace for 23 weeks. Cognitive performance (working speed, reaction time, concentration performance, accuracy), workload and relevant covariates (salivary cortisol level, heart rate, physical activity, sitting time) were measured pre- and post-intervention under laboratory conditions. MANOVA and RMANOVA results did not show differences in performance parameters and workload, respectively, between sit/stand and traditional workplace users. Differences in text editing accuracy and cortisol levels for sit/stand workstation users indicate potential connectivity to cognitive parameters which should be further examined with large-scale studies. Practitioner summary: Medium-term effects of working at sit/stand workstations on cognitive performance and workload are unexplored. This randomised controlled trial suggests that cognitive performance and workload are unaffected for sit/stand workstation users after 23 weeks of use. However, accuracy appeared to improve and physiological stress appeared to be altered. Abbreviations: BMI: body mass index; IPAQ: International physical activity questionnaire; MET: metabolic equivalent of task; MANOVA: multivariate ANOVA; NASA TLX: NASA task load index; RMANOVA: repeated measures ANOVA.

The Contraction Modalities in a Stretch-Shortening Cycle in Animals and Single Joint Movements in Humans: A Systematic Review.

A systematic literature search was conducted to review the force-enhancing mechanisms caused by a stretch-shortening cycle (SSC). The review aims to yield an overview of the contraction modalities influencing the SSC performance in animals and single joint movements in humans. The search was executed in common with the PRISMA statement. CINAHL, MEDLINE (via ProQuest), PubMed, ScienceDirect, Scopus and Web of Science databases were used for the systematic search from its inception until February 2019. A quality assessment was conducted with a modified Downs and Black checklist. Twenty-five studies were included. SSC effects, leading to increased force/work during a SSC and a reduced force depression (FD) compared to a pure shortening contraction, are existent on different levels of the muscle, from single fiber experiments to the level of in vivo muscle-tendon complex. Muscle performance is dependent on shortening velocity, shortening distance, stretch distance, the time (transition phase) between stretch and shortening and the active prephase duration. Concerning stretch velocity we found conflicting results. The findings from this systematic review indicate that the mechanisms in the early phase of shortening are associated with pre-activation effects, elastic recoil and stretch reflex. Furthermore, we speculate that residual force enhancement (RFE) is mainly responsible for an increased steady-state force compared to a pure shortening contraction.

Effect of alternating postures on cognitive performance for healthy people performing sedentary work.

Prolonged sitting is a risk factor for several diseases and the prevalence of worksite-based interventions such as sit-to-stand workstations is increasing. Although their impact on sedentary behaviour has been regularly investigated, the effect of working in alternating body postures on cognitive performance is unclear. To address this uncertainty, 45 students participated in a two-arm, randomised controlled cross-over trial under laboratory conditions. Subjects executed validated cognitive tests (working speed, reaction time, concentration performance) either in sitting or alternating working postures on two separate days (ClinicalTrials.gov Identifier: NCT02863731). MANOVA results showed no significant difference in cognitive performance between trials executed in alternating, standing or sitting postures. Perceived workload did not differ between sitting and alternating days. Repeated measures ANOVA revealed significant learning effects regarding concentration performance and working speed for both days. These results suggest that working posture did not affect cognitive performance in the short term. Practitioner Summary: Prior reports indicated health-related benefits based on alternated (sit/stand) body postures. Nevertheless, their effect on cognitive performance is unknown. This randomised controlled trial showed that working in alternating body postures did not influence reaction time, concentration performance, working speed or workload perception in the short term.

Kinematic analysis of the double side kick in pointfighting, kickboxing.

The double side kick is a frequently used technique in pointfighting on which most offensive kicking techniques are based on. The aim of the study was to identify parameters, e.g., trunk and leg positions, durations and velocities that affect successful technique execution. 44 fighters, male and female, including European and World champions, participated in the research. The participants performed double side kicks to a punching bag with the target at chest level. 10 international experts scored (1-10 points; interrater correlation ICC (3,1): 0.952) videos of the best individual trial for quality of technique execution. Based on the average of the scores participants were categorised into 2 equally sized groups. For kinematic data acquisition, a Vicon 3D-motion capturing system was used. The normalised knee height of the kick leg (KHK), normalised absolute distance to the frontal shoulder (DKS) at the end of both chambering phases, the horizontal velocity during both chambering phases (KEV), the durations of all 6 functional phases and total duration, were analysed. KHK1, KHK2, DKS1 and DKS2, total duration as well as 4 out of 6 phase durations and KEV1 showed significant differences between groups in two-sided t-test or Mann-Whitney U-test for not normally distributed variables.

Effect of a novel two-desk sit-to-stand workplace (ACTIVE OFFICE) on sitting time, performance and physiological parameters: protocol for a randomized control trial.

BACKGROUND: Prolonged sitting is ubiquitous in modern society and linked to several diseases. Height-adjustable desks are being used to decrease worksite based sitting time (ST). Single-desk sit-to-stand workplaces exhibit small ST reduction potential and short-term loss in performance. The aim of this paper is to report the study design and methodology of an ACTIVE OFFICE trial. DESIGN: The study was a 1-year three-arm, randomized controlled trial in 18 healthy Austrian office workers. Allocation was done via a regional health insurance, with data collection during Jan 2014 - March 2015. Participants were allocated to either an intervention or control group. Intervention group subjects were provided with traditional or two-desk sit-to-stand workstations in either the first or the second half of the study, while control subjects did not experience any changes during the whole study duration. Sitting time and physical activity (IPAQ-long), cognitive performance (text editing task, Stroop-test, d2R test of attention), workload perception (NASA-TLX) and physiological parameters (salivary cortisol, heartrate variability and body weight) were measured pre- and post-intervention (23 weeks after baseline) for intervention and control periods. Postural changes and sitting/standing time (software logger) were recorded at the workplace for the whole intervention period. DISCUSSION: This study evaluates the effects of a novel two-desk sit-to-stand workplace on sitting time, physical parameters and work performance of healthy office based workers. If the intervention proves effective, it has a great potential to be implemented in regular workplaces to reduce diseases related to prolonged sitting. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02825303 , July 2016 (retrospectively registered).

A Kinematic Analysis of the Jumping Front-Leg Axe-Kick in Taekwondo.

The jumping front-leg axe-kick is a valid attacking and counterattacking technique in Taekwondo competition (Streif, 1993). Yet, the existing literature on this technique is sparse (Kloiber et al., 2009). Therefore, the goal of this study was to determine parameters contributing significantly to maximum linear speed of the foot at impact. Parameters are timing of segment and joint angular velocity characteristics and segment lengths of the kicking leg. Moreover, we were interested in the prevalence of proximal-to-distal-sequencing. Three-dimensional kinematics of the kicks of 22 male Taekwondo-athletes (age: 23.3 ± 5.3 years) were recorded via a motion capturing system (Vicon Motion Systems Limited, Oxford, UK). The participants performed maximum effort kicks onto a rack-held kicking pad. Only the kick with the highest impact velocity was analysed, as it was assumed to represent the individual's best performance. Significant Pearson correlations to impact velocity were found for pelvis tilt angular displacement (r = 0.468, p < 0.05) and for hip extension angular velocity (r = -0.446, p < 0.05) and for the timing of the minima of pelvis tilt velocity (r = -0.426, p < 0.05) and knee flexion velocity (r = -0.480, p < 0.05). Backward step linear regression analysis suggests a model consisting of three predictor variables: pelvis tilt angular displacement, hip flexion velocity at target contact and timing of pelvic tilt angular velocity minimum (adjusted R(2) = 0.524). Results of Chi-Squared tests show that neither for the leg-raising period (χ(2) = 2.909) of the technique, nor for the leg-lowering period a pattern of proximal-to-distal sequencing is prevalent (χ(2) = 0.727). From the results we conclude that the jumping front-leg axe-kick does not follow a proximal-to-distal pattern. Raising the leg early in the technique and apprehending the upper body to be leant back during the leg-lowering period seems to be beneficial for high impact velocity. Furthermore, striking by extending the hip rather than by flexing the knee could raise impact velocity. Key pointsAngular velocity characteristics of the pelvis segment and the kicking leg's hip and knee joint show no proximal-to-distal sequencing, neither for the leg-raising or leg-lowering period in a jumping front-leg axe-kick.Anthropometric parameters of taekwondo athlete's do not influence their impact velocities.In order to raise the impact velocity in the jumping front-leg axe-kick an athlete should avoid tilting back with the torso. Instead, an upright position should be maintained.In the leg-lowering period, we suggest hitting the target by using hip extension with a rather straight knee, instead of flexing the knee.

Event detection in football: Improving the reliability of match analysis.

With recent technological advancements, quantitative analysis has become an increasingly important area within professional sports. However, the manual process of collecting data on relevant match events like passes, goals and tacklings comes with considerable costs and limited consistency across providers, affecting both research and practice. In football, while automatic detection of events from positional data of the players and the ball could alleviate these issues, it is not entirely clear what accuracy current state-of-the-art methods realistically achieve because there is a lack of high-quality validations on realistic and diverse data sets. This paper adds context to existing research by validating a two-step rule-based pass and shot detection algorithm on four different data sets using a comprehensive validation routine that accounts for the temporal, hierarchical and imbalanced nature of the task. Our evaluation shows that pass and shot detection performance is highly dependent on the specifics of the data set. In accordance with previous studies, we achieve F-scores of up to 0.92 for passes, but only when there is an inherent dependency between event and positional data. We find a significantly lower accuracy with F-scores of 0.71 for passes and 0.65 for shots if event and positional data are independent. This result, together with a critical evaluation of existing methodologies, suggests that the accuracy of current football event detection algorithms operating on positional data is currently overestimated. Further analysis reveals that the temporal extraction of passes and shots from positional data poses the main challenge for rule-based approaches. Our results further indicate that the classification of plays into shots and passes is a relatively straightforward task, achieving F-scores between 0.83 to 0.91 ro rule-based classifiers and up to 0.95 for machine learning classifiers. We show that there exist simple classifiers that accurately differentiate shots from passes in different data sets using a low number of human-understandable rules. Operating on basic spatial features, our classifiers provide a simple, objective event definition that can be used as a foundation for more reliable event-based match analysis.

Muscle synergies are shared across fundamental subtasks in complex movements of skateboarding.

A common theory of motor control posits that movement is controlled by muscle synergies. However, the behavior of these synergies during highly complex movements remains largely unexplored. Skateboarding is a hardly researched sport that requires rapid motor control to perform tricks. The objectives of this study were to investigate three key areas: (i) whether motor complexity differs between skateboard tricks, (ii) the inter-participant variability in synergies, and (iii) whether synergies are shared between different tricks. Electromyography data from eight muscles per leg were collected from seven experienced skateboarders performing three different tricks (Ollie, Kickflip, 360°-flip). Synergies were extracted using non-negative matrix factorization. The number of synergies (NoS) was determined using two criteria based on the total variance accounted for (tVAF > 90% and adding an additional synergy does not increase tVAF > 1%). In summary: (i) NoS and tVAF did not significantly differ between tricks, indicating similar motor complexity. (ii) High inter-participant variability exists across participants, potentially caused by the low number of constraints given to perform the tricks. (iii) Shared synergies were observed in every comparison of two tricks. Furthermore, each participant exhibited at least one synergy vector, which corresponds to the fundamental 'jumping' task, that was shared through all three tricks.

Reduction in systemic muscle stress markers after exercise-induced muscle damage following concurrent training and supplementation with specific collagen peptides - a randomized controlled trial.

Introduction: Collagen peptide supplementation in conjunction with exercise has been shown to improve structural and functional adaptations of both muscles and the extracellular matrix. This study aimed to explore whether specific collagen peptide (SCP) supplementation combined with a concurrent training intervention can improve muscular stress after exercise-induced muscle damage, verified by reliable blood markers. Methods: 55 sedentary to moderately active males participating in a concurrent training (CT) intervention (3x/week) for 12 weeks were administered either 15 g of SCP or placebo (PLA) daily. Before (T1) and after the intervention (T2), 150 muscle-damaging drop jumps were performed. Blood samples were collected to measure creatine kinase (CK), lactate dehydrogenase (LDH), myoglobin (MYO) and high-sensitivity C-reactive protein (hsCRP) before, after, and at 2 h, 24 h and 48 h post exercise. Results: A combination of concurrent training and SCP administration showed statistically significant interaction effects, implying a lower increase in the area under the curve (AUC) of MYO (p = 0.004, ηp2 = 0.184), CK (p = 0.01, ηp2 = 0.145) and LDH (p = 0.016, ηp2 = 0.133) in the SCP group. On closer examination, the absolute mean differences (ΔAUCs) showed statistical significance in MYO (p = 0.017, d = 0.771), CK (p = 0.039, d = 0.633) and LDH (p = 0.016, d = 0.764) by SCP supplementation. Conclusion: In conclusion, 12 weeks of 15 g SCP supplementation combined with CT intervention reduced acute markers of exercise-induced muscle damage and improved post-exercise regenerative capacity, as evidenced by the altered post-exercise time course. The current findings indicate that SCP supplementation had a positive effect on the early phase of muscular recovery by either improving the structural integrity of the muscle and extracellular matrix during the training period or by accelerating membrane and cytoskeletal protein repair. Clinical trial registration: https://www.clinicaltrials.gov/study/NCT05220371?cond=NCT05220371&rank=1, NCT05220371.

Increased trial-to-trial similarity and reduced temporal overlap of muscle synergy activation coefficients manifest during learning and with increasing movement proficiency.

Muscle synergy analyses are used to enhance our understanding of motor control. Spatially fixed synergy weights coordinate multiple co-active muscles through activation commands, known as activation coefficients. To gain a more comprehensive understanding of motor learning, it is essential to understand how activation coefficients vary during a learning task and at different levels of movement proficiency. Participants walked on a line, a beam, and learned to walk on a tightrope-tasks that represent different levels of proficiency. Muscle synergies were extracted from electromyography signals across all conditions and the number of synergies was determined by the knee-point of the total variance accounted for (tVAF) curve. The results indicated that the tVAF of one synergy decreased with task proficiency, with the tightrope task resulting in the highest tVAF compared to the line and beam tasks. Furthermore, with increasing proficiency and after a learning process, trial-to-trial similarity increased and temporal overlap of synergy activation coefficients decreased. Consequently, we propose that precise adjustment and refinement of synergy activation coefficients play a pivotal role in motor learning.

Increased knee flexion in participants with cerebral palsy results in altered stresses at the distal femoral growth plate compared to a typically developing cohort.

INTRODUCTION: Femoral deformities are highly prevalent in children with cerebral palsy (CP) and can have a severe impact on patients' gait abilities. While the mechanical stress regime within the distal femoral growth plate remains underexplored, understanding it is crucial given bone's adaptive response to mechanical stimuli. We quantified stresses at the distal femoral growth plate to deepen our understanding of the relationship between healthy and pathological gait patterns, internal loading, and femoral growth patterns. METHODS: This study included three-dimensional motion capture data and magnetic resonance images of 13 typically developing children and twelve participants with cerebral palsy. Employing a multi-scale mechanobiological approach, integrating musculoskeletal simulations and subject-specific finite element analysis, we investigated the orientation of the distal femoral growth plate and the stresses within it. Limbs of participants with CP were grouped depending on their knee flexion kinematics during stance phase as this potentially changes the stresses induced by knee and patellofemoral joint contact forces. RESULTS: Despite similar growth plate orientation across groups, significant differences were observed in the shape and distribution of growth values. Higher growth rates were noted in the anterior compartment in CP limbs with high knee flexion while CP limbs with normal knee flexion showed high similarity to the group of healthy participants. DISCUSSION: Results indicate that the knee flexion angle during the stance phase is of high relevance for typical bone growth at the distal femur. The evaluated growth rates reveal plausible results, as long-term promoted growth in the anterior compartment leads to anterior bending of the femur which was confirmed for the group with high knee flexion through analyses of the femoral geometry. The framework for these multi-scale simulations has been made accessible on GitHub, empowering peers to conduct similar mechanobiological studies. Advancing our understanding of femoral bone development could ultimately support clinical decision-making.

Artificial intelligence in sports on the example of weight training.

The overall goal of the present study was to illustrate the potential of artificial intelligence (AI) techniques in sports on the example of weight training. The research focused in particular on the implementation of pattern recognition methods for the evaluation of performed exercises on training machines. The data acquisition was carried out using way and cable force sensors attached to various weight machines, thereby enabling the measurement of essential displacement and force determinants during training. On the basis of the gathered data, it was consequently possible to deduce other significant characteristics like time periods or movement velocities. These parameters were applied for the development of intelligent methods adapted from conventional machine learning concepts, allowing an automatic assessment of the exercise technique and providing individuals with appropriate feedback. In practice, the implementation of such techniques could be crucial for the investigation of the quality of the execution, the assistance of athletes but also coaches, the training optimization and for prevention purposes. For the current study, the data was based on measurements from 15 rather inexperienced participants, performing 3-5 sets of 10-12 repetitions on a leg press machine. The initially preprocessed data was used for the extraction of significant features, on which supervised modeling methods were applied. Professional trainers were involved in the assessment and classification processes by analyzing the video recorded executions. The so far obtained modeling results showed good performance and prediction outcomes, indicating the feasibility and potency of AI techniques in assessing performances on weight training equipment automatically and providing sportsmen with prompt advice. Key pointsArtificial intelligence is a promising field for sport-related analysis.Implementations integrating pattern recognition techniques enable the automatic evaluation of data measurements.Artificial neural networks applied for the analysis of weight training data show good performance and high classification rates.

A kinematic model for assessment of hip joint range-of-motion in fast sport movements using spreading angles.

The work investigates the relationship between the leg spread and hip joint flexibility during kicking action of highly skilled martial artists. Vector-based spreading angles (VSAs), as a simplified way to assess kick execution patterns and movement ranges, are proposed. The first VSA measures the angle between the two femur vectors (i.e., upper leg spread), whereas the second describes the full-leg spread. The proposed measures were applied to 3D motion analysis performed on 33 participants at various skill levels who executed double side kicks and performed two static flexibility tests (hip joint flexion and abduction). Statistical analysis showed that the proposed parameters differentiate between skill levels. Both VSAs and VSA rate of change showed high Pearson correlation to the quality of execution (the average of 10 subjective kickboxing experts' evaluation scores) at critical instants of the kick execution , i.e., during the first leg elevation and kick (angles: ρ > 0.8; p < 0.001). In addition, they significantly correlated to maximum static hip abduction at any instant (ρ = 0.37-0.69; p < 0.05). The work established that both static and dynamic joint flexibility contribute to high-performance levels and could be used for initial selection and assessment of training effectiveness.

Modeling the formation of defensive gaps in basketball: Cutting on a teammate's drive.

Basketball is a game of simultaneous actions, and inter-player coordination is key for offensive success. One of the most challenging aspects in this regard is basket cutting on a teammate's drive. The ability to make these cuts is considered to be an artistic skill, mastered by only a handful of players. This skill is also hard to assess, as there is no method to measure the players' capability with respect to this quality-especially not automatically. Using SportVU data from the NBA, we created a mathematical model that identifies the openings in the defense which allow to perform a cut. Our model succeeds to generalize, as it detects these openings on average 139ms earlier than the actual cuts start and has an overall (balanced) accuracy of 0.818 on the test set. Having a tree-based gradient boosting classifier, we received a clear hierarchy of feature importance and were able to inspect the interactions between these attributes during action. This way, the model gives insights about the kind of defensive movements needed for a player to allow enough space to cut while in practical usage the analysis of the output can also help the coaching staff in designing play options and assessing player abilities. By paying more attention to the possible off ball movements during drives, offensive plays can become more versatile-benefiting the participants and the spectators alike.

The gait pattern and not the femoral morphology is the main contributor to asymmetric hip joint loading.

Gait asymmetry and skeletal deformities are common in many children with cerebral palsy (CP). Changes of the hip joint loading, i.e. hip joint contact force (HJCF), can lead to pathological femoral growth. A child's gait pattern and femoral morphology affect HJCFs. The twofold aim of this study was to (1) evaluate if the asymmetry in HJCFs is higher in children with CP compared to typically developing (TD) children and (2) identify if the bony morphology or the subject-specific gait pattern is the main contributor to asymmetric HJCFs. Magnetic resonance images (MRI) and three-dimensional gait analysis data of twelve children with CP and fifteen TD children were used to create subject-specific musculoskeletal models and calculate HJCF using OpenSim. Root-mean-square-differences between left and right HJCF magnitude and orientation were computed and compared between participant groups (CP versus TD). Additionally, the influence on HJCF asymmetries solely due to the femoral morphology and solely due to the gait pattern was quantified. Our findings demonstrate that the gait pattern is the main contributor to asymmetric HJCFs in CP and TD children. Children with CP have higher HJCF asymmetries which is probably the result of larger asymmetries in their gait pattern compared to TD children. The gained insights from our study highlight that clinical interventions should focus on normalizing the gait pattern and therefore the hip joint loading to avoid the development of femoral deformities.

Intra- and inter-subject variability of femoral growth plate stresses in typically developing children and children with cerebral palsy.

Little is known about the influence of mechanical loading on growth plate stresses and femoral growth. A multi-scale workflow based on musculoskeletal simulations and mechanobiological finite element (FE) analysis can be used to estimate growth plate loading and femoral growth trends. Personalizing the model in this workflow is time-consuming and therefore previous studies included small sample sizes (N < 4) or generic finite element models. The aim of this study was to develop a semi-automated toolbox to perform this workflow and to quantify intra-subject variability in growth plate stresses in 13 typically developing (TD) children and 12 children with cerebral palsy (CP). Additionally, we investigated the influence of the musculoskeletal model and the chosen material properties on the simulation results. Intra-subject variability in growth plate stresses was higher in cerebral palsy than in typically developing children. The highest osteogenic index (OI) was observed in the posterior region in 62% of the TD femurs while in children with CP the lateral region was the most common (50%). A representative reference osteogenic index distribution heatmap generated from data of 26 TD children's femurs showed a ring shape with low values in the center region and high values at the border of the growth plate. Our simulation results can be used as reference values for further investigations. Furthermore, the code of the developed GP-Tool ("Growth Prediction-Tool") is freely available on GitHub (https://github.com/WilliKoller/GP-Tool) to enable peers to conduct mechanobiological growth studies with larger sample sizes to improve our understanding of femoral growth and to support clinical decision making in the near future.