The natural environment poses difficult categorization problems for nonhuman animals by placing a premium on nearly optimal performance, while providing stimuli that are ambiguous and only probabilistically related to outcomes such as food. The proposed research is part of a long-term effort to develop method and theory for the study of how animals learn to perform optimally when they confront visual environments with complex perceptual meaning and ambiguous historical implications. The five proposed experiments accordingly will develop an animal model of a well-known procedure developed by Ashby and his colleagues for the study of how humans optimally categorize two-dimensional exemplars of ill-defined concepts. The experiments will identify the kinds of decision rules animals can learn and these rules will in turn discriminate among competing theories for the nature of ill-defined concepts. Specifically, it will be determined how well birds' decision rules approximate varying types of optimal rules, and the rules the birds learn will diagnose among prototype, exemplar-based, and decision-rule based theories of ill-defined concepts. Furthermore, a computer-simulation theory will provide interpretations of the processes involved when animals learn these rules. The long-term health related implications derive from the animal model the proposed research will produce. This model will open up the possibility of future research into the biological basis of optimal categorizations of complex stimuli. It will also provide an animal model with which to explore suggestions from current neuropsychological research with human patients, regarding the biological basis of cognitive deficits seen in Parkinson's patients.
| Herbranson, W T; Fremouw, T; Shimp, C P (1999) The randomization procedure in the study of categorization of multidimensional stimuli by pigeons. J Exp Psychol Anim Behav Process 25:113-34 |
