Back to Wayne Gray's home page / To Research Interests Index / To Downloadable Pubs / The CogWorks Lab / Cognitive Science Department
What is cognitive workload? As in many constructs used by the applied community, the notion of cognitive workload is ill-defined, typically being left to the reader's intuitions to fill in the gap. The mysterian position is implicit -- cognitive workload is often portrayed as something unto itself; that is, it is often implicitly portrayed as something that cannot be reduced to a combination of more fundamental processes such as working memory load, attention, and so on.
In contrast to this mysterian position, in our work we have adopted the perspective that cognitive workload is little more than busyness at a fine-grained level. If we analyze situations of high cognitive workload and compare our analysis to situations of low cognitive workload, what we will see is nothing more and nothing less than differences in low-level cognitive, perceptual, and motor operations.
The Argus Project is attempting to peel back the layers of mystery surrounding the concept of cognitive workload. Argus Prime is a complex but tractable simulated task environment designed to study cognitive workload in a task involving cognitive as well as perceptual-motor performance. In particular, our goal is to understand how subtle aspects of an interface may lead to large increases or decreases in cognitive and perceptual workload. We hypothesize that subtle differences in interface design affect cognitive workload by the set of strategies that they make available to the users. Hence, our current efforts focus on describing and modeling the set of strategies available for use with different interfaces.
If you were to interrupt someone to ask what they are doing, you would typically get an answer such as, "I am reading a book," "playing a video game," or "surfing the web." These sorts of activities take place at the bounded rationality level of activities that take place over the time span of one minute to days.
|
Scale (sec)
|
Time Units
|
System
|
Analysis
|
World (theory)
|
| 100000 | days | Task | Task Analysis | BoundedRationality |
| 10000 | hours | |||
| 1000 | 10 min | |||
| 100 | min | Subtask | Unit Task Analysis | |
| 10 | 10 sec | Unit task | Cognitive Task Analysis | Cognitive Band |
| 1 | 1 sec | Activity | Embodied Cognitiion | |
| 0.1 | 100 ms | Embodiment Level | Microstrategies | |
| 0.01 | 10 ms | Elements | Architectural | BiologicalBand |
| 0.001 | 1 ms | Parameters |
Our analysis of systems begins in the cognitive band (area in blue in the table -- between 100 msec and 10 sec in duration). We would like to determine whether cognitive workload in interactive behavior is manifested at a given level of analysis or whether each level of analysis presents its own challenges to cognitive workload. If the answer is the former, then what is done at that level is the key to cognitive workload. For example, if the critical level is the unit task level, then changing what activity operators are required for a given unit task would increment or decrement cognitive workload. However, if the answer is the latter, then attention must be paid to changing how each level is implemented. The distinction between what versus how is the difference between being able to direct researcher and developer attention to one level of the cognitive band versus having to be attentive to all levels.
At this time we are writing the first major journal article that deals with Argus Prime and cognitive workload. At present the articles that I list below deal with selected aspects of this issue. They are also very relevant to the issues of microstrategies and interactive behavior.
Gray, W. D., & Fu, W.-t. (2001). Ignoring perfect knowledge in-the-world for imperfect knowledge in-the-head: Implications of rational analysis for interface design. CHI Letters, 3(1), 112-119. Also in ACM CHI'01 Conference on Human Factors in Computing Systems
Gray, W. D., & Boehm-Davis, D. A. (2000). Milliseconds Matter: An introduction to microstrategies and to their use in describing and predicting interactive behavior. Journal of Experiment Psychology: Applied, 6(4), 322-335.
Gray, W. D. (2000). The nature and processing of errors in interactive behavior. Cognitive Science, 24(2), 205-248.
Gray, W. D., Schoelles, M. J., & Fu, W.-t. (2000). Modeling a continuous dynamic task. In N. Taatgen & J. Aasman (Eds.), Proceedings of the Third International Conference on Cognitive Modeling (pp. 158-168). Veenendal, The Netherlands: Universal Press.
Schoelles, M. J., & Gray, W. D. (2001). Argus: A suite of tools for research in complex cognition. Behavior Research Methods, Instruments, & Computers, 33(2), 130-140..
Schoelles, M. J., & Gray, W. D. (2001). Decomposing interactive behavior, In J. D. Moore & K. Stenning (Eds.), Twenty-Third Annual Conference of the Cognitive Science Society (pp. 898-903). Mahwah, NJ: Erlbaum.
Schoelles, M. J., & Gray, W. D. (2000). Argus Prime: Modeling emergent microstrategies in a complex simulated task environment. In N. Taatgen & J. Aasman (Eds.), Proceedings of the Third International Conference on Cognitive Modeling (pp. 260-270). Veenendal, NL: Universal Press.
Schoelles, M. J., & Gray, W. D. (1998). Argus, A System for Varying Cognitive Workload. Paper presented at the Human Factors and Ergonomics Society 42th Annual Meeting, Chicago.
Back to Wayne Gray's home page / To Research Interests Index / To Downloadable Pubs / The CogWorks Lab / Cognitive Science Department
Last changed: 2002-11-16 wdg