John JB Ayres - US grants

Backward conditioning has always been controversial, and even in the most convincing demonstrations has been weak. Still, it is of great theoretical importance because its existence argues against the widely accepted view that neutral events must predict the occurrence of non-neutral events in order to acquire conditioned strength. Dr. Ayres will carry out experiments using rats in order to study backward excitatory fear conditioning, in which a once-neutral stimulus (CS) comes to arouse fear because it has in the past closely followed (rather than preceded) a non-neutral stimulus (US). In Dr. Ayres' work, the US is a brief electric shock, and the CS is a tone, click, or light. Dr. Ayres will manipulate such variables as CS and US duration, the length of the gap between US termination and CS onset, etc. The aim is to discover parameters that will make the backward- conditioning effect as robust as possible and thus to facilitate future experimental and theoretical analysis. The discovery of a strong effect could, moreover, implicate backward conditioning as a hitherto-neglected source of human anxiety disorders. Dr. Ayres' research is unique in focusing mainly on one-trial conditioning procedures. Human anxiety disorders (and learning by animals in the wild) often appear to be based on a single experience, yet one-trial conditioning has received relatively little experimental attention. Such procedures also have important methodological properties. Finally, the research is also unique in that the bulk of the studies are guided by the computer-simulated predictions of a formal model of conditioning. The studies thus provide strong tests of that model.

Central to research in Pavlovian conditioning over the last 30 years has been the issue of whether the temporal contiguity of events is or is not sufficient for associative learning. The current concensus is that it is not. This concensus is, however, now being challenged by a new and growing body of evidence. The new evidence suggests that contiguity may be sufficient for learning after all yet may be insufficient for the expression of that learning in performance. Three phenomena widely regarded as showing that contiguity is not sufficient for learning are blocking, overshadowing, and latent inhibition. Virtually all contemporary theories of learning view these phenomena as learning failures that occur despite favorable temporal contiguity. Using rat subjects, the proposed work would subject each phenomenon to three assays that have become standard tools for differentiating between learning and performance effects. The assays are blocking, second-order conditioning, and sensory preconditioning. The goal is to determine whether latent inhibition, blocking, and overshadowing really reflect learning deficits or whether they reflect performance deficits instead. Results supporting the performance-deficit-view would invalidate the mainstay assumption of nearly all modern theories of learning. Results supporting the learning- deficit view, would do so more strongly than any previous evidence. Results of either type would have direct relevance for Pavlovian conditioning models of human phobia. Such models have been criticized in part because the occurrence of trauma often fails to make people phobic about the situations in which the trauma occurred. Latent inhibition, blocking, and overshadowing are among the Pavlovian mechanisms that could explain these failures.

DESCRIPTION (Adapted from applicant's abstract): Humans acquire fears in the real world (Context a). They then receive treatment for those fears in a therapist's office (Context b). Finally, they return to the real world (Context a). Unfortunately, animal research shows clearly that this a-b-a progression is one that, in its final stage (the return to Context a), leads to the renewal of an eliminated fear. Because therapists often have no choice but to follow this a-b-a progression, it is important to search for fear-elimination procedures that make fear resistant to renewal. The proposed research would exploit animal models to do so, examining a wide variety of procedures whose possible success is predicted by contemporary theory. All of the procedures use standard and well understood assays of fear in laboratory animals: Experiments 1 through 7 use the conditioned suppression (CER) procedure. Experiments 8+9 use both CER and conditioned punishment procedures. While continuing the search for procedures that resist renewal, the latter experiments provide an analysis of the role of a response contingency that is thought to be a crucial feature of participant modeling -- a type of therapy thought to be successful in treatments of fear in humans.