THEORY OF ANIMAL TIMING: LEARNING AND REMEMBERING AGGRE

在过去的几年里，一个正式的动物计时理论，标量

Over the past several years a formal theory of animal timing, Scalar Expectancy Theory (SET), has been developed and applied to a wide range of temporal control findings in animal learning. Theory and experiment to date have been focused on understanding anticipation, or expectancy, of reinforcement at fixed, well learned delays. This has meant that applications of the theory have been mainly focused on animal psychophysics preparations and instrumental and classical conditioning situations involving standardized, fixed interstimulus intervals. While considerable success has been found here, a frequent technique for studying choice behavior of animals under temporally based schedules of reward utillizes variable delays, in which the choice alternatives are stochastic, although generally constant in the mean. Variability introduced experimentally has revealed some important regularities in the learning laboratory. The most notable of these perhaps the "Matching Law", according to which choice proportions approximately match corresponding reinforcer proportions. Matching, and corollary regularities in choice for variable, time-based alternatives, are perforce based on some appreciation of aggregations of these delays. This proposal addresses the problem of how such aggregates are learned, remembered, and discriminated in a variety of settings. In particular, the theory is adapted to describe constant probability (Poisson process) variable interval schedules of reward, and addresses the Matching Law and regulated findings. A second focus is addressed to understanding the memory structure for these aggregates when they constitute discriminative stimuli rather than delays to reinforcement. Subjects will be tested in an animal psychophysics setting for their ability to discriminate variable ensembles from fixed and variable alternatives. The psychophysical task requires a very different response strategy, but it is based on the same memory for time. Hence, the research should reveal structural properties of memories.