A Software Toolkit for Predicting the Neural Signatures of Cognitive States

Trends in human-machine teaming and multi-domain operations will only increase the complexity of the tasks facing operators. This complexity comes with design risks: will operators be capable of performing the required tasks within acceptable levels of workload, stress, etc. Often, problems are not discovered until human testing is conducted, late in the process. At that point, design flaws may be expensive, or impossible, to correct. Sonalysts and Miami University are developing a software tool that will predict operator state across a range of circumstances. Advanced cognitive modeling techniques will be used to generate synthetic electroencephalogram (EEG) signals which will, in turn, drive state-recognition algorithms. The tool, known as EEG Modeling Architecture for Predicting States (EMAPS), will be used to guide research and development of advanced user interfaces (e.g., adaptive automation systems) and training systems. EMAPS will include an “authoring tool� that allows engineers to rapidly develop the models that reflect an operator performing tasks with a specified system. EMAPS will then run a Monte Carlo simulation and present the results to engineers to allow them to identify portions of the design or segments of the task that are likely to give human operators an unacceptable level of difficulty.