A Software Toolkit for Predicting the Neural Signatures of Cognitive States

The United States Air Force (USAF) has a long history of using human performance models to increase the effectiveness of training, and predict the impact of physical factors (like fatigue) and environmental factors (like time pressures and information uncertainty). Within Phase I, Sonalysts and Miami University worked to improve the quality of these models through the development of the EEG Modeling Architecture for Predicting States (EMAPS) tool. EMAPS is part of a growing collection of research that merges the mathematical and neurophysiological modeling approaches. EMAPS will enhance USAF modeling in several ways. First, as the military increasingly moves toward human-machine teaming, developers can use these models to create more effective synthetic teammates. Second, in training contexts, developers can use the models to create more realistic constructive teammates, adversaries, or even teachers. Third, this capability can be used to more reliably assess system design and predict human performance in advanced man-machine systems. Fourth, superior models can be used to drive more responsive adaptive automation systems. Beyond extending the accuracy with which models can make predictions for both individuals and groups, EMAPS will guide the efficiency of modeling efforts by enabling multi-scale modeling for the observed phenomena.