Event Models in Cognition and Perception: From Text to Real-Time

Washington University, Saint Louis MO

Investigators

Abstract

In talking about the past or about a movie, we often describe things in terms of discrete events, breaking continuous activity into meaningful parts. A child asked about a day at school might first describe reading, then recess, and maybe lunch. If asked further about recess, the child might divide it into smaller parts like soccer or tag or talking to friends. Psychological studies show that events and their sub-events play discernable roles in how we understand stories, for example, or remember them later. However, little is known about how this way of structuring of experience guides ongoing perception and comprehension. In research supported by the National Science Foundation, Dr. Jeffrey M. Zacks studies how events and sub-events drive ongoing perceptions and later memories. His work incorporates perceptual judgments, tests of memory, reading, and he constructs computer models to make his hypotheses explicit. Dr. Zacks's research uses novel methods for the real-time study of perception of dynamic events. The broader impacts of this work include training graduate and undergraduate students in these methods and the tools used to implement them. These results also contribute to the ongoing development of tools for procedural learning and for navigating video.