Psychophysiological Models of Cognitive States Can Be Operator-Agnostic

Real-time prediction of a person's trust (T), mental workload (W), and situation awareness (SA) can improve safety and performance in operational environments. We develop psychophysiological models of TWSA both with and without operator-specific demographic information available to them and we assess the impact of this constraint on the models' performance. We demonstrate functional model fit (Adjusted 𝑅 2 : T = 0.67, W = 0.65, and SA = 0.88) and predictive ability with operator-agnostic models, assessed via leaveone-participant-out cross validation (𝑄 2 : T = 0.58, W = 0.45, and SA = 0.79). Our findings help establish the viability of operatoragnostic psychophysiological models of TWSA which could be used to inform an autonomous agent or manage multi-agent teams.