Main Subcommittee Category
Human Performance Analysis and EngineeringHow would you label your submission?
IndustryAbstract
The computerized multi-attribute task battery (MATB) was introduced for cognitive workload research by NASA in the 1990s as a low-fidelity flight-simulation platform in which users attend to multiple simultaneous tasks designed to engage multiple cognitive faculties at once (e.g., auditory/linguistic processing, visual processing, logical reasoning, visuo-motor coordination) while emulating representative piloting tasks. The MATB has since been widely utilized in research investigating human performance, cognitive workload, trust in automation, strategic behavior, and related topics. Although it has been updated and re-implemented for a broader range of applications and greater researcher configurability by NASA, the U.S. Air Force, and others (e.g., MATB-II, AF-MATB, and OpenMATB, respectively), it still lacks key elements for maintaining participant engagement such as adequate real-time performance feedback and individualized adjustment of task difficulty.
This paper describes novel MATB enhancements designed for application in cognitive workload research and related domains. Specifically, the MATB was extended with real-time performance feedback, including auditory and haptic feedback and noxious electrical stimuli (i.e., shocks, individually calibrated to be unpleasant but not painful), designed to increase participant engagement, self-awareness, and motivation. Feedback of correct/incorrect responses and error punishment were expected to reduce the attenuation of physiological reactions to cognitive workload observed in simulated work environments relative to real ones. Additional MATB enhancements include periodic, instantaneous task-load self-assessment, an adaptive approach to manipulating task demands designed to improve experimental control for potentially confounding effects of learning and/or fatigue, and a novel task automation mode designed to bring “automation surprise” and/or “automation frustration” under experimental control. These modifications were tested in the context of a DARPA-supported human study investigating physiological responses to cognitive workload in human-automation teams (e.g., aircrew-automation teams). The details of the enhancements and their evaluation are presented, along with recommendations for researchers applying these or similar mechanisms in their experimental designs.