U.S. Department of Energy

Pacific Northwest National Laboratory

Mitigating Cognitive Depletion in Streaming Environments

Interpretations of streaming data often require rapid, informed human decisions. Fatigue worsens decision accuracy and executive function.  Anyone who has had to perform decision making over extended periods of time will be familiar with the feeling of cognitive depletion – increased impulsiveness and errors, slower evaluation of new information, and decreased memory retention. Cognitive depletion is poorly understood; we rely on people to self-assess their levels of fatigue or restrict them to limited periods of time on or off the job. This reduces overall performance and increases stress.


Our research seeks to leverage techniques for modeling cognitive resources to create adaptive systems that extend peak performance. We will use models to understand cognitive depletion and then predict the state of the user.  We are also taking a more pragmatic approach using traditional machine learning to predict when a user should take a break. Using a combination of these two methods, we can make recommendations to users as to when they should take a break – before the effects of cognitive depletion start to degrade performance.


We have identified cognitive depletion in two distinct tasks – standardized tests and video games. Our streaming system for evaluating cognitive depletion increases performance by avoiding periods of fatigue – not simply waiting to identify the errors associated with fatigue. By identifying cognitive depletion, we will allow users to take breaks before their fatigue levels make them error prone. For users that are not able to selectively take breaks, quantifying cognitive depletion allows systems to engage increased automation or decrease the impact of user’s potentially impaired decisions.

| Pacific Northwest National Laboratory