Background

Knee injuries are common during sports that require fast change-of-direction (COD) movements such as sidestepping and pivoting during soccer, basketball, handball, and related sports. COD movements expose the knee joint to large external forces, particularly if players show a poor COD technique such as lateral trunk lean towards the plant foot or a strong knee valgus of the cutting leg. Larger external forces and moments that act on the knee joint are expected to result in larger strain of the anterior cruciate ligament (ACL) and thus a higher risk of ACL rupture. Consequently, during sports like soccer and basketball, many non-contact ACL injuries occur during COD tasks. While neuromuscular training (NMT) programs have been developed to effectively reduce the risk of sports injury including ACL tears, ACL injury rates have not declined in the last years. One of the reasons for this paradox may be that many NMT programs such as the FIFA11+ program, which were developed to protect from injury do not actually improve COD movement strategies. It may be assumed that FIFA11+ does reduce the overall risk of sports injury through general improvements in strength and balance as well as safer jump landing technique but not through safer COD technique. Further, if training interventions were successful in reducing 'high-risk' movement patterns and in developing knee-stabilizing muscle synergies during COD movements, it remains unclear whether the improved movement strategy, e.g. the reduction in external knee valgus moments, actually corresponds to reduced ACL strain. In consequence, there is the need for a comprehensive investigation to determine whether a NMT program focused on improving COD technique will improve COD movement and muscle activation strategies and whether these improvements are correlated with estimated ACL strain.

A second reason for the paradox may be that current experimental protocols to investigate COD movement strategies in the laboratory are not a good indicator for actual player behavior on the field thus masking potential benefits of NMT on lateral movements. Therefore, the sports injury prevention community should aim to move the assessment of COD movement strategy onto the playing field and into a more realistic playing environment while characterizing the kinematics and kinetics of sidestepping based on wearable sensors. In consequence, novel analytical frameworks based on wearables need to be developed, which can capture full-body kinematics and the underlying forces during COD movements on the playing field. In the long run, such systems could facilitate real-time feedback with respect to COD technique on the playing field and thus enhance motor learning of the players as well as characterize real-world player agility.

Research objectives & hypotheses

Objective 1: To determine the effect of an 8-week NMT and COD technique modification intervention (multidirection training, MD) on 1) COD movement strategies as characterized by the lateral trunk angle and knee valgus moment and 2) estimated ACL strain during 45- and 135-degree COD movements in comparison to an 8-week NMT and linear sprint training intervention (linear sprint training, LS) in sports science students.

Hypothesis 1: There will be a larger reduction in lateral trunk angle and knee valgus moment and an associated reduction in ACL strain in the MD group compared to the LS group following the 8-week intervention, which will be retained four weeks later.

Objective 2: To determine the effect of an 8-week NMT and COD technique modification intervention on leg muscle synergies as characterized by the number of muscles and the structure of the synergy vector for each identified muscle synergy in comparison to an 8-week NMT and linear sprint training intervention in sports science students.

Hypothesis 2: One or multiple muscle synergy vectors will show an increased contribution of hip abductor muscle activity in the MD group following training and there will be a lower number of activated muscles per identified synergy, i.e. a more selective muscle activation in comparison to the LS group. These improvements will be retained four weeks later.

Objective 3: To determine the validity of an analysis framework to estimate COD movement strategy (lateral trunk angle, foot progression angle, knee valgus moment) and ACL strain based solely on inertial motion capture data in comparison to the gold-standard of 3D optimal motion capture.

Study design and participants

This is a quasi-randomized trial to investigate the effects of effects of two neuromuscular training interventions (multidirectional vs. linear sprint training) on COD movement strategies in sports science students and in parallel to develop a methodological framework for field-based COD movement assessments. This trial will be registered with ClinicalTrials.gov after ethical approval has been granted. An a-priori sample size estimation was conducted based on a previous non-randomized controlled trial reporting a strong, beneficial effect of a 6-week COD technique modification on the Cutting Movement Assessment Score (CMAS), a qualitative screening tool to quantify biomechanical and neuromuscular deficits during COD tasks [significant group x time interaction, Cohen's f2 = 0.63 (right leg) - 1.00 (left leg)]. With an a-priori significance level of alpha = 0.05 and a desired power of 0.8, the minimum required sample size was estimated in G*Power v3.1.9.7 as only N = 8 (n = 4 in each training group). One concern with this previous study is, however, that CMAS raters were not blinded to group assignment leading to a potential systematic bias and overestimated effect size. A second concern is that the CMAS, which combines multiple biomechanical and neuromuscular deficits (e.g. knee valgus, lateral trunk flexion, hip internal rotation) into one aggregate score may show a larger effect compared to the analysis of single biomechanical variables (e.g. knee abduction moment) like in the current study. Based on these considerations the required sample size of N = 40 sports science students (n = 20 for the MD and LS) was selected. With an expected drop-out rate of 25%, this sample of at least N = 30 participants will enable to detect moderate to strong interaction effects (Cohen's f2 > 0.3) with a power of at least 0.8.

Experimental protocol

All participants will take part in a first familiarization session (WN 40) to practice the execution of the COD task and receive information about the content and goals of the training intervention. Baseline testing (WN 40-41), follow-up testing (WN 49-50), and retention testing (WN2-4, 2022) will be conducted in the laboratory Pulverturm at the ISW Innsbruck and will follow the same testing protocol. Participants will be equipped with 47 retroreflective markers for 3D motion analysis (extended full-body PlugIn Gait model, 10 cameras, 200 Hz) and 14 sensors for combined measurements of surface electromyography (sEMG, 2000 Hz) and inertial segment data (200 Hz) (Noraxon Ultium, Noraxon Inc., Scottsdale, AZ, USA) of the right leg. Afterwards, the participants will be asked to warm-up on a stationary bike for five minutes followed by the initial running exercises of the FIFA11+ program. Maximum voluntary contractions (MVCs) will be obtained for ankle plantarflexion/dorsiflexion, knee flexion/extension, hip extension/flexion, and hip abduction according to standardized manual tests. IMU position and orientation with respect to the participants' anatomy will be determined based on images and calibration movements taken before the start of the measurements. After four successful practice trials, participants will be asked to complete 10 COD movements with a 45-degree cutting angle (COD45) and 10 COD movements with a 135-degree cutting angle (COD135) always planting and cutting with the right foot on a ground-embedded force plate (1000 Hz, Kistler, Winterthur, Switzerland). The approach and exit speed for each COD movement will be controlled to approximately 4.5 m/s (COD45) and 4 m/s (COD135) as measured by optical timing gates placed within the approach and exit zones (Witty, Microgate, Bolzano, Italy). Rest periods of one minute in between trials and five minutes in between blocks (COD45 vs. COD135) will avoid accumulation of neuromuscular fatigue.

Data analysis

Using the 3D marker trajectories and force data as inputs (optimal motion capture, OMC, model), full-body kinematics, 3D joint moments, muscle forces, and ACL strain will be estimated using inverse kinematics and EMG-informed optimization procedures in OpenSim (v.4.2) combined with FE simulations.

Using only the inertial sensor data as inputs (inertial motion capture, IMC, model), full-body kinematics, 3D joint moments, muscle forces, and ACL strain will be estimated using an IMU-tracking and optimal control algorithm combined with FE simulations. This latter approach is novel and is currently being developed.

The muscle activation pattern of the right leg will be further analyzed via a muscle synergy analysis [one muscle synergy model for each testing time point (baseline, follow-up 1, follow-up 2] and quantified according to its muscle synergy vectors and synergy activation profiles. Specifically, the number of muscles per synergy vector and the shape of the muscle synergy vectors will serve as additional dependent variables.

Statistics

Objective 1 & 2: Effects of the NMT and COD technique modification intervention on the kinematic, kinetic (OMC model), and EMG-related outcome variables will be tested according to a repeated measures ANOVA with the between-subject factor 'group' (MD vs. LS) and the within-subject factor 'time point' (baseline, follow-up, retention). The null hypothesis H1 will be rejected if there is a significant interaction effect 'group x time point' at an a-priori significance level of alpha = 0.05. Interaction effects, i.e. effect modification, due to the participants' sex or potential other covariates (e.g. baseline COD performance) will be explored in additional three-way repeated measures ANOVAs. Post-hoc testing will investigate changes in dependent variables between time points for each group.

Objective 3: The validity of the IMC model will be determined based on Bland & Altman limits of agreement to estimate the systematic bias and random error of the IMC model-based dependent variables in comparison to the OMC model-based variables. The IMC model will be called valid if the limits of agreement are smaller than the observed changes in the corresponding variables between baseline and follow-up testing in the intervention group.

Ethical considerations / risk-benefit analysis / insurance information

All participants will be sports science students and thus frequently encounter the injury risks associated with multi-directional sports. Therefore, the risk of injury for participants in this study will be no higher than in their everyday life. In the rare case that a participant sustains an injury during this study, he/she will be covered by accident insurance provided to all University of Innsbruck students. Furthermore, all participants will benefit from participating in the current study given that the FIFA11+ NMT program, which is part of the training in both training groups, has been shown to effectively reduce the risk of sports injury. Overall, the benefits of this study outweigh the risks.

Inclusion Criteria:

• Enrolled as a student at the Department of Sport Science in Innsbruck

Exclusion Criteria:

• Clinical diagnosis of a lower extremity muscle or joint injury within the last six months that led to a disruption of sport participant by at least two weeks
• Clinical diagnosis of a balance disorder

The raw data (motion capture, EMG, and anonymous participant information) will be shared via an open access data repository along with the publications that present the study results.

Data will be shared via an open access data repository along with the publications that present the study results.

Open access

While the study team will not point participants to the specific differences between the two training programs (MD vs. LS), an exchange between participants about the training elements in each group cannot be avoided given that the participants are all students at the same faculty. Given that the study team will supervise the training sessions, collect the data, and assess the outcomes, blinding of investigators will not be possible but the analysis will be designed to minimize sources of investigator bias.