COLLABORATIVE RESEARCH: Mechanisms of Perceptual Learning

点击翻译按钮获取中文摘要

Collaborative Research: Mechanisms of Perceptual LearningAdult humans can show large improvements in the performance of even the simplest perceptual tasks as the result of training and/or practice. These improvements have been observed in tasks in virtually every sensory modality. This project investigates the brain mechanisms of perceptual learning and the circumstances under which perceptual learning occurs. The findings will help us understand the adaptive nature of human behavior, and may generate new computational and training principles for performance optimization in particular task environments. Our research applies a powerful method for identifying and characterizing the mechanism of perceptual learning in visual tasks. The method adds systematically increasing amounts of external noise - random visual noise (similar to random TV noise) to the visual stimulus and observes the effect on a perceptual task as perceptual learning takes place under different training protocols. Performance in clear and in noisy visual task environments can be modeled quantitatively to identify three mechanisms of perceptual learning. Each mechanism has a "signature". (E.g., modification of the observer's perceptual template through training only affects performance at high levels of external visual noise, where the external noise is large enough to be the limiting factor. Experts can eliminate external noise more efficiently.) We investigate the mechanisms of perceptual learning in a wide variety of perceptual tasks including discrimination and identification or classification of both simple and complex visual patterns. Combining our methods with transfer manipulations that test for critical properties of learning will validate the measurements and provide information about the level of learning. The proposed work will improve our empirical and theoretical understanding of the nature of perceptual learning. Characterization of perceptual learning mechanisms is necessary to a full understanding of the adaptive nature of the adult human brain. Our theory and methods provide a basis for developing adaptive models of the human brain and may contribute to the development of efficient training procedures in applied settings.