Quantitative modeling of category learning in amnesic patients

Category rule learning was examined in two amnesic patients using the perceptual categorization task (e.g., Ashby & Gott, 1988; Filoteo & Maddox, 1999). Traditional accuracy-based analyses as well as quantitative model-based analyses were performed. Unlike accuracy-based analyses, the model-based approach allowed us to examine both categorization rule learning and variability in the trial-by-trial application of the participant's categorization rule. The results indicated that the amnesic patients were as accurate as the controls in learning a complex, nonlinear rule over a large number of trials. The model-based analysis indicated that, in general, the amnesic patients learned the categorization rule as well as controls and applied their rule as consistently as controls. Categorization performance on a second day of testing revealed that amnesic patients can retain the categorization rule over a 24-h period. These results suggest that the brain regions damaged in amnesia are not involved in category learning or memory for the category structures.