BioTDMS: A Multi-Sensor and Data-driven System for Real-Time Team Performance Assessment

Abstract

The Biobehavioral Team Dynamics Measurement System (BioTDMS) presents a novel, multimodal approach for assessing real-time team performance in high-stakes, dynamic environments. Addressing the limitations of post hoc, subjective evaluations, BioTDMS integrates physiological (EEG, fNIRS, ECG, respiration), behavioral (gaze, speech, input logs), and communication data to generate dynamic, objective metrics of team cognition and adaptability. Grounded in interactive team cognition theory and layered dynamics modeling, the system captures reorganization and synchrony in team states in response to task perturbations. Using data collected from eight dyadic teams (16 participants) performing a time-constrained Noncombatant Evacuation Operation (NEO) planning task, BioTDMS identifies bio-behavioral signatures predictive of team effectiveness and stress resilience. Machine learning models, particularly logistic regression and support vector machines demonstrated high predictive performance (F1 ≈ 0.96) when leveraging team-level synchrony features, outperforming models trained on individual-level data alone. These findings underscore the importance of interactional metrics in team performance assessment and offer a pathway toward adaptive, data-driven training systems for mission-critical operations.