Perceptual and Cognitive Processes in Category Learning

点击翻译按钮获取中文摘要

The research funded by this award will investigate how people learn newcategories of perceptual objects and events. Categorization is a cognitiveprocess of fundamental importance that all people perform countless timesevery day. It allows people to read, to select foods that are safe to eat,to use objects and tools in the environment, and to recognize friends andkin. Unfortunately, it is also a fragile process, in the sense that anumber of normal and pathological conditions have been shown to causecategory learning deficits. These include normal aging, major depression,Parkinson's disease, Alzheimer's disease, Huntington's disease, strokes,and schizophrenia. Previous theories of category learning were testedprimarily on categorization data from healthy university students. Tounderstand category learning in these neglected populations requires anunderstanding of the brain structures and pathways that mediate thisimportant skill. Toward this end, this research will test and extend thefirst neuropsychological theory of category learning, called COVIS(COmpetition between Verbal and Implicit Systems; Ashby, Alfonso-Reese,Turken, & Waldron, 1998, Psychological Review, 105, 442-481), which assumesthat category learning is a competition between separate explicit andimplicit categorization systems. The explicit system is under consciouscontrol and engages in a systematic process of hypothesis testing orrule-based reasoning. The implicit system, which is largely not accessibleto conscious awareness, engages in a form of procedural learning. Ashby etal. (1998) developed a formal version of COVIS that makes precisequantitative predictions in a wide variety of categorization tasks, andalso a neuropsychological version that describes the underlying neuralmechanisms and pathways that are assumed to mediate category learning. Thecurrent research includes many new experiments with healthy young adultsthat were motivated by the COVIS theory. Some experiments will test theCOVIS assumption that normal category learning is a competition betweenfunctionally separate explicit and implicit systems. Many of theexperiments will isolate the explicit system and investigate itsproperties, including the rules that govern how it learns, both with andwithout trial-by-trial feedback about response accuracy. Finally, manyanalogous experiments will attempt to achieve the same goals with theimplicit system (i.e., isolate it and investigate its properties). Thecompletion of these studies will not only provide a rigorous test of COVIS,but regardless of their outcome, it will also provide valuable new insightsinto the perceptual and cognitive processes and the underlying neuralmechanisms and pathways that mediate one of the most basic of all humanacts - grouping objects and events in the environment by category.