Research on motor decision making in collaboration with machine learning and computational modeling: a scoping review

This study is a scoping review following the PRISMA-ScR specification, aiming to systematically sort out the application of machine learning and computational modeling in sports decision analysis and training. In the face of the fragmentation between these fields, this study conducted a comparative analysis of the literature included in the two-path search strategy. It is found that machine learning paths limit the mechanistic interpretation due to the "black box" property. The ecological validity and intervention translation of computational modeling pathways were insufficient. At the same time, based on the analysis, this study innovatively proposes a new cognitive computing framework, which aims to integrate the dual advantages of data and theory at the theoretical level, and provide a clear integration blueprint and development agenda for the future development of intelligent decision assistance systems with high performance, high interpretation and compliance with cognitive laws.