Probability-dependent Changes in Movement Preparation During Change-of-direction Movements: Sex Differences and the Influence of Risk-taking

Introduction & Purpose Anterior cruciate ligament (ACL) injuries often occur during change-of-direction (COD) movements without opponent contact (Krosshaug et al., 2007), with some injuries following neurocognitive errors that elicit injury-prone movements (Gokeler et al., 2024). Especially female populations are known to have higher incidences of ACL injuries (Bram et al., 2021) and COD movement strategies differ between the sexes (Donelon et al., 2024). Previous biomechanical studies typically investigated this injury scenario based on “unplanned” COD movements where players react to directional cues with equal left-right probabilities (50:50). However, this may not reflect the motor control strategy during real play. Athletes likely estimate directional probabilities from opponent cues and game flow, anticipate the most probable action and pre-plan movements accordingly, e.g., to gain a performance advantage (Gredin et al., 2023). If the plan does not match the required action, erroneous movement or imprecise control might occur and in turn elevate ACL injury risk. For example, pelvis rotation towards the new movement direction is a modulating factor to mitigate injury risk (while maintaining performance), by taking rotational demands off of the lower extremities (David et al., 2018). This strategy, however, requires a certain amount of preparation for the new movement direction. On the other hand, pelvis rotation towards the wrong direction can increase injury risk. Risk-taking attitudes may shape this anticipation and pre-planning process. Males, especially younger males, show a higher risk-taking propensity than females in computer-based tests like the Balloon Analogue Risk Test (BART; Lejuez et al., 2002). It is yet unknown whether this tendency carries over to motor planning and sports-specific, decisions and whether risk-taking propensities (measured by a general behavioral proxy like BART) can explain sex differences in COD movement strategies and ACL injury risk. Here, we investigated whether unequal directional probabilities (80% to the right) in reactive COD movements lead to biased movement preparation toward the more likely direction, measurable as greater pelvis rotation at initial contact of the sidestepping leg and improved performance as compared to a 50:50 condition. Further, we investigated whether the extent of biased movement preparation is modulated by sex. Finally, we examined whether individual risk-taking propensity, as measured by BART, and its interaction with sex is associated with preparatory pelvis rotation under unequal probabilities. We hypothesized that players would show greater pelvis rotation towards the right and faster completion times during CODs when the directional probability was 80% compared to the standard 50% (H1), and that sex-related differences in risk-taking propensity can explain preparatory pelvis rotation (H2). Methods Fifty-two soccer players participated: 36 male players of an U18 youth team (age: 16.5 ± 0.5 years) and an U23 adult team (age: 18.5 ± 2.2 years) and 16 players of a female youth / adult team (age: 17.1 ± 2.7 years). The sample was a convenience sample and unbalanced across sexes (36 male, 16 female), limiting power to detect interaction effects. All gave written informed consent. The protocol was approved by the University of Innsbruck Ethics Review Board Sports Science (ID 76/2025). Players performed 16 unplanned 90° CODs on artificial turf. Eight IMUs (Ultium Motion by Noraxon, Scottsdale, Arizona, USA) across lower limbs, pelvis, and trunk were used to record COD kinematics: this analysis focused on pelvis rotation (yaw) at the initial contact of the sidestepping leg. Pelvis yaw was derived after sensor-to-segment calibration (static neutral stance) from a pelvis sensor fixated above the sacrum with adhesive and an elastic-belt to minimize skin motion. It was defined positive toward the right cutting direction, therefore values reported for rightward CODs reflect rotation toward the intended cutting direction. IC was detected from peak resultant acceleration of the foot sensor. Earlier in-lab validation against optical motion capture for a similar cutting task indicated RMSEs for pelvis yaw of ~8° and excellent Waveform correlations (r = 1.00) (unpublished), which contextualizes the magnitude of the observed differences. After a short run up, a verbal cue indicated the cutting direction 3 m before a dedicated COD area of 1 m diameter. In six trials, COD directions were distributed equally (50:50 condition) and in another set of ten trials, the cue had an 80% bias to the right side (80:20 condition). The order of the 50:50 and 80:20 conditions was randomized across participants. Within each condition, left-right cues were presented in a pseudo-random sequence generated individually for each participant: the 50:50 block contained three left and three right trials, the 80:20 block contained exactly 8 right and 2 left cues. While participants were unaware of the required number of trials per condition, cue predictability may have increased toward the end of the blocks. Participants were informed about the probabilities of the trials before each condition. For this analysis, we solely considered CODs towards the right, i.e., 3 trials for 50:50 and 8 trials for 80:20 to ensure comparability across conditions with the same biomechanical direction and to directly test bias toward the more likely side in the 80:20 block. Performance was assessed based on COD completion time and recorded from run-up start to a target 5 m beyond the COD zone using BlazePods (BlazePod Ltd., Miami, Florida, USA). Afterward, players completed the computer-based BART (Lejuez et al., 2002) on-site. In this test, participants pumped up a number of virtual balloons and received virtual money for each pump. If a balloon pops, the money is lost. The outcome was the adjusted average pumps (AAP), which averages the pumps on balloons that did not explode and is considered a measure of risk-taking propensity (Lejuez et al., 2002). Statistical analyses included linear mixed models (random intercept for ‘participant’ and fixed factors ‘condition’ with two levels (‘5050’, ‘8020’) and ‘sex’ with two levels (‘f’, ‘m’)) to investigate the effects of the probability condition and the sex on the pelvis rotation as well as on performance. A t-test was used to investigate sex-differences in risk-taking propensity. A linear regression (predictors: AAP*sex) tested, whether risk-taking behavior and sex can explain the pelvis rotation in the 80% condition. Approximate normal distribution of AAP and regression model residuals were tested and confirmed based on the Shapiro-Wilk Test. The significance level for all statistical analyses was set to alpha = 0.05. Results We found a trend but no statistical significance for an interaction effect of sex and probability condition (50:50 versus 80:20) for pelvis rotation (F(1,519.4) = 2.95, p = 0.086) (Figure 1a). The trend shows increased pelvic rotation in 80:20 conditions for male players and decreased pelvic rotation for female players. Main effects were not significant (sex: F(1,53.1) = 0.70, p = 0.406, condition: F(1,519.4) = 0.09,  p = 0.768). For performance, a significant main effect of sex (F(1,54.6) = 45.8, p < 0.001) revealed longer completion times for female players compared to male players, but no effect of probability condition (F(1,513.4) = 0.01, p = 0.909) nor an interaction effect with sex (F(1,513.4) = 0.01, p = 0.931). A t-test revealed significant differences (t(50) = 3.46, p = 0.001) in AAP between females and males with increased AAP in male players (M = 11.4, SD = 1.8) compared to female players (M = 9.59, SD = 1.5) (Figure 1b). In an overall non-significant regression model (F(3,48) = 1.87, p = 0.147, R² = 0.11), AAP in isolation was not a significant predictor of pelvis rotation during the 80:20 condition (p = 0.586) but its interaction with sex approached the significance level (p = 0.076). Discussion This study investigated the role of unequal directional probabilities and risk-taking propensity for movement strategies in COD movements. Our findings do not support our H1: there were no statistically significant adaptations in pelvis rotation nor COD performance when the movement direction was more probable (80%), relative to a 50:50 condition. H2 was also not supported: although males exhibited higher risk-taking propensity than females, risk-taking propensity in isolation did not explain pelvic pre-rotation under unequal probabilities. Males showed higher risk-taking propensity than females based on BART AAP scores. This observation did not translate into clear kinematic adaptations between the sexes. Trends in the interaction pattern suggested that males may have increased pelvis rotation toward the more likely direction, whereas females may not have adapted similarly, possibly adopting a more conservative strategy. However, these trends were not statistically significant and should be interpreted cautiously. Performance advantages relating to earlier rotation trends in males could not be observed. Given the absence of significant interaction effects and uncertainty around whether BART reflects sport-specific risk-taking, our findings, based on pelvis rotation and AAP, do not allow firm inferences about the role of risk-taking in sex differences in COD strategies and in explaining female’s elevated ACL injury risk. More analyses including whole-body kinematics over more than one time point could give further insight into the COD movements elicited in these decision-making tasks including unequal probabilities. Conclusion Unequal directional probabilities did not yield significant group-level changes in pelvic pre-rotation or COD performance relative to 50:50 trials. A trend indicates that males may increase rotation toward the likely direction, while females do not. Although males demonstrated higher risk-taking propensity, it did not explain preparatory pelvis rotation under unequal probabilities. Within the constraints of this protocol and measures, we did not find statistical evidence that (BART-assessed) risk-taking propensity explains sex differences in ACL-injury related kinematics during COD tasks. However, the observed trends raise more questions regarding motor planning in scenarios with unequal probabilities, e.g., to understand how neurocognitive errors occur and lead to injuries. Future whole-body biomechanical analyses might help to better understand risk-taking strategies in COD tasks and their effect on ACL injury risk. References Bram, J. T., Magee, L. C., Mehta, N. N., Patel, N. M., & Ganley, T. J. (2021). Anterior cruciate ligament injury incidence in adolescent athletes: a systematic review and meta-analysis. The American journal of sports medicine, 49(7), 1962-1972. https://doi.org/10.1177/0363546520959619 David, S., Mundt, M., Komnik, I., & Potthast, W. (2018). Understanding cutting maneuvers–The mechanical consequence of preparatory strategies and foot strike pattern. Human movement science, 62, 202-210. https://doi.org/10.1016/j.humov.2018.10.005 Donelon, T. A., Edwards, J., Brown, M., Jones, P. A., O’Driscoll, J., & Dos’ Santos, T. (2024). Differences in biomechanical determinants of ACL injury risk in change of direction tasks between males and females: A systematic review and meta-analysis. Sports Medicine-Open, 10(1), 29. https://doi.org/10.1186/s40798-024-00701-z Gokeler, A., Tosarelli, F., Buckthorpe, M., & Della Villa, F. (2024). Neurocognitive errors and noncontact anterior cruciate ligament injuries in professional male soccer players. Journal of Athletic Training, 59(3), 262-269. https://doi.org/10.4085/1062-6050-0209.22 Gredin, N. V., Bishop, D. T., Williams, A. M., & Broadbent, D. P. (2023). The use of contextual priors and kinematic information during anticipation in sport: toward a Bayesian integration framework. International Review of Sport and Exercise Psychology, 16(1), 286-310. https://doi.org/10.1080/1750984X.2020.1855667 Krosshaug, T., Nakamae, A., Boden, B. P., Engebretsen, L., Smith, G., Slauterbeck, J. R., ... & Bahr, R. (2007). Mechanisms of anterior cruciate ligament injury in basketball: video analysis of 39 cases. The American journal of sports medicine, 35(3), 359-367. https://doi.org/10.1177/0363546506293899   Lejuez, C. W., Read, J. P., Kahler, C. W., Richards, J. B., Ramsey, S. E., Stuart, G. L., ... & Brown, R. A. (2002). Evaluation of a behavioral measure of risk taking: the Balloon Analogue Risk Task (BART). Journal of Experimental Psychology: Applied, 8(2), 75. https://doi.org/10.1037//1076-898X.8.2.75