Self-adaptive learning of anticipation skills in virtual reality.

The main objective of this study was to determine the extent to which a virtual reality-based relay simulator can optimize the anticipation skills of novice relay runners. The simulator integrated a head-mounted display (HMD) to immerse participants in a life-sized virtual stadium, requiring them to physically trigger their start based on the approaching virtual incoming runner. A learning protocol was used to compare the respective effectiveness of a self-adaptive method, which progressively adjusts the task difficulty level to the learner's performance, and a method imposing the same progressive reduction in task difficulty on all participants. The results demonstrate a significant improvement in the anticipatory behavior of novice relay runners (e.g., reduction of absolute error from ∼131 ms to ∼58 ms). While both methods yield equivalent learning improvements, the self-adaptive approach provides a foundational benefit, as its prior deployment is essential for the subsequent implementation of the prescriptive method. Moreover, it not only allows learners to be confronted with the most appropriate level of difficulty throughout the learning process, but also adapts as closely as possible to each learner's needs. The precise impact of the learning methods on the processes underlying athletes' anticipation behavior is discussed, and the importance of testing the hypothesis of learning transfer to real-world situations is emphasized.