Understanding whole-body inter-personal dynamics between two players using neural granger causality as the explainable artificial intelligence.

Understanding the dynamics of complex, whole-body interpersonal coordination behavior in humans is an important subject in behavioral science. However, due to the challenges of analyzing complex causal relationships among multiple body components with conventional techniques, this area remains underexplored. To address this issue, this study proposes a new analytical framework that attempts to understand the underlying causal structures behind each joint movement of individual players using neural Granger causality (NGC) as the explainable artificial intelligence (XAI). In the NGC analysis, causal relationships were defined as the size of the weight parameters of the first layer of a machine-learning model trained to predict the future state of a specific time-series variable. To verify this approach practically, we conducted an experiment with 16 pairs of expert baseball pitchers and batters, and input datasets with 27 joint resultant velocity (13 pitchers' and 14 batters' joints) were generated and used for model training. The results revealed that significant causal relations exist among intra- and inter-individual body components, such as "the batter's hands have a causal effect from pitcher's throwing arm." Although the causality from the batters to the pitcher's body is significantly lower than that from the pitchers to the batter's body, it exhibits a significant correlation with the performance outcomes of batters (R = 0.69). These results suggest the effectiveness of the NGC analysis for understanding whole-body inter-personal coordination dynamics and, more broadly, the XAI technique as a new approach for analyzing complex human behavior from a perspective different from conventional techniques.