|
| SQAB 2005 Tutorial: Temporal Regulation of Choice |
| Saturday, May 28, 2005 |
| 4:00 PM–4:50 PM |
| International South (2nd floor) |
| Domain: Basic Research |
| Chair: Robert W. Allan (Lafayette College) |
| Presenting Authors: : DANIEL T. CERUTTI (Duke University) |
| Abstract: In the simplest example of temporal control, animals responding on a fixed interval (FI) schedule show a characteristic pause-and-respond pattern—the FI “scallop.” Parametric research on FI schedules reveals a lawful relationship between post-reinforcement-pausing and the inter-food interval (IFI), pausing is a fixed proportion of IFI. If temporal control of any sort is common to all schedules of reinforcement, it should play a role in conditioned reinforcement (chained schedules), choice (concurrent schedules), and choice for different magnitudes of reinforcement (self-control). Empirical studies confirmed informal predictions of differential pausing in concurrent and concurrent-chain schedules, with greater pausing shown on schedules with longer average IFIs. Thus, rates of responding on concurrent schedules may derive from differential pausing. The most parsimonious model to fit choice data is one that assumes a single representation of time from the last reinforcer, ignoring the response that produced it, with pauses on each response proportional to the corresponding last IFI, constant running rates after pauses. Subsequent findings confirm that pigeons responding on concurrent schedules is timed from the last reinforcer, but further suggest that behavior on concurrent schedules can be understood better as a temporal-stochastic process in which (a) the overall reinforcement delay determines the overall rate of behavior and (b) the individual schedules’ reinforcement delays determine the allocation of responses to each alternative. Taken together, findings suggest that animals responding on simple schedules learn when to respond, and animals responding on choice tasks learn when to respond where. |
| |
| DANIEL T. CERUTTI (Duke University) |
| Dan Cerutti is a Research Associate Professor of Psychology at Duke University. He has done experimental and theoretical work on operant conditioning, rule-governed behavior, and discrimination. His previous work includes demonstrations of recombinant generalization in the pigeon (with A.C. Catania), studies on the interaction between instructions and schedule performance in humans, experiments on the spatial-attentional biases in children (with C. Arciniega), and studies on choice in pigeons. In addition to teaching classes in learning and research methodology at Duke, Dan maintains two labs, one for pigeons and another for fish. His pigeon experiments, done in collaboration with John Staddon and Jeremie Josefowiez, are concerned with understanding temporal variables in choice. Those experiments appear to show that what pigeons learn about reinforcers in a choice task is a simple rule of /when/ and /where/. That is, when the pigeons are faced with two choices, each presenting a different rate of reinforcement, they begin responding on the choice with the shortest delay to reinforcement, and only later respond on the choice with the longer delay to reinforcement. Dan has spent the last few years developing technology to study operant behavior in fish to study the generality of timing processes, their ecological basis, and perhaps shed light on their evolutionary origins. |
|
| |