Visual Summary-Statistic Processing in Infancy

North Dakota State University, Fargo ND

Investigators

Linked publications, trials & patents

Abstract

DESCRIPTION (provided by applicant): Part of the NICHD's mission is to support basic research in human development. The Developmental Cognitive Psychology, Behavioral Neuroscience, and Psychobiology program supports research to identify links between the developing brain and the environment. The proposed research has been designed to help us understand how the visual environment shapes infants' use of summary statistics to describe the things they see. A growing body of research suggests that adults summarize a great deal of visual information by describing the world with texture-like measurements - instead of measuring exactly what we saw and where we saw it, we often pool information together over large areas by measuring how simple features co-occur and forgetting about exactly where they came from. For some tasks, these kinds of descriptions are useful. In other cases, like recognizing a single object in clutter, they don't work well at all. Understanding the limits of these summary statistic descriptions of our visual world has given us insights into why adults sometimes find visual search difficult (Rosenholtz et al., 2012), why it's hard to recognize single objects in clutter (Balas et al., 2009), and may help us understand conditions like amblyopia or macular degeneration, where we think these kinds of summaries may be nearly all the visual system has to work with. In these studies, we will work with infants and adults to understand how summary-statistic descriptions are shaped by the visual environment. We will use computer graphics techniques to create artificial textures that are matched to natural textures using a model of the early visual system. Our goal is to determine how different artificial textures need to be from natural textures for infants to tell them apart. By using a computer graphics model, we can carefully control what information is available to tell the images apart, which will help us understand what measurements contribute to infants' summary-statistic descriptions as they get older. Further, we plan to measure how well infants can tell real textures apart from artificial ones by measuring how their brain responds to those images. We will use EEG to measure how the brain responds to different kinds of natural and artificial textures. Different neural response reflect different kinds of processing in the brain, and we plan to use the location and the timing of the differences we see in infants' brains to help us understand how summary statistics are applied. These studies support the NICHD's mission and will help us understand the perceptual difficulties that some children and adults have due to visual impairments that force summary statistics to be used all the time.