9702291 Van Zandt An important area of research in cognitive psychology is how information, whether directly impinging on the senses or retrieved from memory, is transformed during cognition to allow performance of simple tasks. Understanding this process of transformation provides the basis for understanding human behavior more generally. This research will emphasize a theoretical, model-driven approach to the study of behavior, in which a number of models of human information processing will be developed and tested. In particular, this research will use neural network models and certain statistical models to explain the interactions between behavioral variables (such as reaction time, response accuracy and confidence) in memory retrieval tasks. In addition, the research will involve the construction of biologically plausible models of simple decision-making in an interdisciplinary, collaborative effort with Dr. Ernst Niebur of the Zanvyl Krieger Mind/Brain Institute. These models may represent a step toward the development of biologically realistic models of human cognition. As psychological science becomes more specialized, students require more mathematical and statistical training. The research and educational activities in this project will teach graduate and undergraduate psychology students about modeling human behavior. Changes to the undergraduate psychology curriculum at the Johns Hopkins University will incorporate laboratory and research experience in introductory level experimental psychology courses. Redesigned statistics and experimental design courses will emphasize computer methods and psychological applications. As a result, Hopkins undergraduates will have a deeper appreciation for scientific and quantitative methods in experimental psychology. Furthermore, students with interests in cognition and neuroscience will have opportunities to do collaborative research in these areas by assisting directly with the research in this proj ect. This research, assisted by undergraduates and graduate students, will progress toward an understanding of the differences and similarities between network models and more traditional statistical models of cognition. Finally, the application of these kinds of dynamic models of cognition to problems in memory retrieval will increase our understanding of the ways that humans use information in making decisions about the world and ultimately enable the development of automated methods to help people make such decisions in more effective ways. ***
