| The following dissertation explored the potential contributions of a hypothesized category learning system, the perceptual learning system, that is posited to rely on learning related changes in sensory cortical areas. This type of learning system is thought to be engaged in category learning situations that require successful learning of only one category (i.e.,' A, not A' learning). This is in contrast to other category learning situations that require individuals to learn more than one category ('A,B' learning). We explored the potential contributions of the perceptual learning system by directly comparing (A, not A) and (A,B) learning using a two variants of the prototype distortion task. In prototype distortion tasks, category members are created by initially defining a category prototype. Category members are then created by perturbing each dot in the prototype a given amount to create 'prototype distortions'. Three behavioral experiments were conducted in order to examine the role of a posited perceptual learning system in categorization. Experiment 1 tested the prediction that perceptual learning is primarily responsible for mediating learning in (A, not A) prototype distortion tasks, but not (A,B). The results obtained were consistent with this prediction. As category members increased in the amount of distortion, individuals were less likely to endorse them as category members. This rate of endorsement increased more steeply for individuals in the (A, not A) versus individuals in the (A,B) task. In addition, the results from Experiments 2 and 3 tested the predictions that (1) the perceptual learning system can successfully learn in the absence of trial-by-trial feedback and (2) the locus of perceptual learning occurs in higher-level visual cortical areas, respectively. Again, the results obtained were consistent with the predictions. In addition, the results from an fMRI study using both variants of the prototype distortion task are discussed. Taken together, these results provide an important first step towards establishing a role for perceptual learning in categorization. |
