Beyond detection rate: understanding the vigilance decrement using signal detection theory

The vigilance decrement has been classically characterized as the decline in performance across time as individuals continuously attend to a task. Errors during these periods of degraded performance are often collectively characterized as failures of attention. Methodologically, the classic characterization of the vigilance decrement relies upon declines in detection rate, a binary measure that is unable to characterize performance beyond a single dimension. Theoretically, using a single construct, such as attention, to describe impaired performance obscures what is likely a range of behaviors. This is a critical issue for the study of vigilance because detection rate can be impacted both by changes in sensitivity and decision criterion. Commonly used tasks do not allow for the reliable computation of these metrics because they elicit a low number of false alarms or because they introduce confounding response demands. To address these shortcomings, we propose the use of a paradigm amenable to the application of the signal detection framework, which permits the reliable and isolated investigation of the vigilance decrement across multiple measures.