Rule-based and information-integration categorization during an incidental learning task

Abstract

Research in visual category learning supports the existence of dual category-learning systems with distinct neurobiological substrates. Rule-based category learning is engaged by category input distributions that vary orthogonally in perceptual space whereas information integration learning is engaged when categories are defined across multiple input dimensions. Investigators have recently begun to explore these dual systems in audition, including phonetic category learning. However, investigations of dual category learning systems have tended to use category training tasks in which participants actively search for category-relevant dimensions, make overt category labeling decisions, and receive explicit feedback about the correctness of their responses. Recent neuroimaging research highlights the fact that overt category training tasks like this engage learning systems that differ from those engaged by more incidental category learning. Since most natural category learning occurs under more incidental conditions, it is important to understand how training tasks interact with category learning systems. The aim of the present research is to investigate the interaction of input distribution and task. We examine nonspeech auditory category learning across the same perceptual input space for rule-based and information-integration categories in incidental and overt training tasks to characterize how training task interacts with input category distributions.