Main content
Cognitive Workload Measurement and Modeling under Divided Attention
- Spencer C. Castro
- Andrew Heathcote
- David Strayer
- Dora Matzke
Date created: | Last Updated:
: DOI | ARK
Creating DOI. Please wait...
Category: Project
Description: Overloading a person’s limited capacity for attention results in decreased performance during goal-oriented tasks. Currently, cognitive researchers have demonstrated the relationship between increasing task difficulty, activation in neurological correlates of executive control, and behavioral outcomes. However, questions remain as to the source of performance limitations. The parameters of sequential sampling models of simple choices may provide scientists with tools to explore the nature of this limitation. In these models, the overload can be thought of as either a maximization of the information-processing rate, or a mechanism for improving certainty classified as response caution (i.e., a threshold of evidence). Both parameters could result in slowed reaction times (RTs) and increased errors, but the characteristics of RT and error distributions may favor one parameter over the other. In order to determine how the parameters of these models vary with cognitive workload, we modeled a distracted driving task. The results demonstrate that a linear model of evidence accumulation accurately predicts response time distributions to an ISO standard and a modified choice “Detection Response Task” (DRT) under cognitive workload. Additionally, the most parsimonious model favors an interaction of a dominant threshold effect and an effect of the drift rate parameter. The application of sequential sampling models to future studies of load may reshape our understanding of the traditional limited capacity information processing framework.