Exploratory Analysis Between Entropy-Based Team Dynamics and Team Performance in Immersive Virtual Reality Training

This exploratory study examined associations between entropy-based measures of team dynamics and team performance in an immersive virtual reality (VR) team training environment. Six teams ( n  = 6) completed a VR-based Tactical Combat Casualty Care (TC3) training session within the Team Dynamics Measurement Framework (TDMF) architecture, consisting of two scenarios with pre-, break, and post-measurements. Behavioral data were continuously captured across four interaction layers: communication, gaze, patient, and system. Entropy time series were computed using a moving window approach to continuously measure team adaptation, where low entropy corresponds to more uniform, stable team dynamics and high entropy means that team is momentarily unstable and adapting. Mean and range entropy metrics (measures of amount of adaptation and adaptive capacity) were correlated with team dynamics ratings, self-efficacy, satisfaction, and system usability scores. Correlations were computed in two analytic variants: (a) correlations using raw pre-, break-, and post-scenario outcome scores, and (b) correlations using delta-based change scores to examine performance shifts across scenario segments. Results indicated moderate to large associations between entropy-based measures and performance outcomes, with range entropy demonstrating stronger and more consistent associations than mean entropy across layers, with the strongest observed effect between early self-efficacy increases during Scenario 1 and range patient interaction entropy ( r (4) = 0.93; p  = .007), providing preliminary proof-of-concept support for entropy-based metrics as dynamic indicators of adaptive team functioning within immersive VR training systems.