Adaptive digital twin framework for real-time cognitive and emotional state decoding

This project develops a physics-based Adaptive Digital Twin (ADT) framework for decoding human cognitive and emotional states from EEG and bio-signals in real time. Unlike conventional AI approaches that depend heavily on large training datasets and black-box models, this effort combines systems theory, operational modal analysis (OMA), adaptive estimation, and dynamical systems modeling to identify evolving cognitive “modes” directly from measured data. The methodology treats the brain as a high-dimensional dynamical system whose dominant modal structures evolve over time. EEG and physiological measurements are used to continuously update a reduced-order state-space model and extract interpretable features such as modal frequency, damping, energy participation, complexity, and traveling-wave behavior. These dynamical signatures are being investigated for correlation with stress, fatigue, attention, emotional response, cognitive workload, and human performance. Current work includes analysis of DEAP emotion datasets and experimental EEG recordings, with validation through reconstruction accuracy, modal consistency, and cognitive-state correlation studies. Preliminary results suggest that dominant dynamical modes may provide a compact and explainable representation of cognitive and emotional behavior suitable for real-time human-machine interaction. Potential applications include adaptive autonomy, driver/operator monitoring, emotionally aware AI systems, aerospace human factors, neuroergonomics, rehabilitation, and cognitive health assessment. The project is seeking collaborators in neuroscience, AI, robotics, autonomous systems, transportation, defense, and human-performance monitoring for experimental validation and real-time deployment studies.