Computational Models of Individual Differences in Working Memory Capacity

摘要

This project involved several inter-related goals. The first was to study how working memory affects performance of cognitive tasks. We then demonstrated that a formally specified, computer implemented model can account for performance at a fine level of detail including performance differences among individuals. Finally, we demonstrated that differences in working memory capacity are stable and predictive of performance across several qualitatively different tasks. Frank Lee and Glenn Gunzelmann both did research on the spatial information processing requirements of complex systems. Frank did work on how people learn to attend to appropriate regions of these visual displays. Working with an air-traffic control simulation, he showed that people came to focus their attention on those regions that were most informative. Glenn Gunzelmann looked at how people relate camera information to map information in the unmanned flight simulation. He showed that participants developed one of two strategies for relating the two sources of information. Both Frank and Glenn developed successful ACT-R models of their respective tasks which were used to direct subsequent research. Michaela Spehn conducted research concerned with how prior knowledge, specifically familiarity with a concept and number of prior associations affects learning of new information. She was also involved in modeling her behavioral data.