Optimizing risk in action: Differences between fixated aim points and movement outcomes in throwing

Introduction Movement outcomes are inherently subject to variance. Handling this variance is crucial for successful sensorimotor behaviour – whether in everyday life or in sports –, particularly in high-risk situations. Research using finger-pointing tasks has shown that humans take into account their own motor variance and costs of potential outcomes in movement planning (Trommershäuser et al., 2008). However, the question remains whether this mechanism extends to more complex tasks (Beck et al., 2023). Here, we investigate sensorimotor behaviour under risk in throwing, across three experiments with 20 participants each. Methods Participants’ task was to throw balls on a target circle in a virtual reality (VR) setup, gaining 100 points for each hit. The target was partially overlapped by a penalty circle. We manipulated the consequences of hitting the penalty circle (0 points vs -500 points vs -2000 points) and the distance between both circles (30 cm vs 45 cm vs 60 cm). This task challenged participants to find strategies that optimally trade-off potential penalties and rewards. To capture participants’ strategies, we measured the location of their final gaze fixation before movement – as an indicator of their planned aiming point – and the ball’s impact location. Models of statistical decision theory (Trommershäuser et al., 2008) predict that the optimal aim point horizontally shifts away from the centre of the target circle as soon as the penalty is non-zero. In other words, participants should incorporate a “safety margin”. This horizontal shift should be larger with (1) higher penalties, (2) smaller distances between the target and penalty circle and (3) with higher motor variance. Results In the no-penalty condition, the final fixation and the ball’s impact location were both centred on the target. In the penalty condition, both their final fixations and the ball’s impact location shifted away from the penalty circle, with larger shifts for higher penalties and smaller distances. Intriguingly, the shifts in the ball’s actual impact location were not only significantly larger (“more conservative”) but also closer to the statistically optimal location compared to the initially fixated aim points. Analysis of movement trajectories shows that, in penalty conditions, the shifts away from the penalty zone increased progressively until the final phases of the movement. Conclusion Our findings show that principles of statistical decision theory generalize to more complex tasks. Extending Trommershäuser et al. (2008)., our results suggest that risk evaluation is not completed in a planning phase before movement execution but is optimized during ongoing movements. ReferencesBeck, D., Hossner, E.-J., & Zahno, S. (2023). Mechanisms for handling uncertainty in sensorimotor control in sports: A scoping review. International Review of Sport and Exercise Psychology, 1–35. https://doi.org/10.1080/1750984X.2023.2280899 Trommershäuser, J., Maloney, L. T., & Landy, M. S. (2008). Decision making, movement planning and statistical decision theory. Trends in Cognitive Sciences, 12(8), 291–297. https://doi.org/10.1016/j.tics.2008.04.010