Association for Behavior Analysis International

Basic research on the topic of conditioned reinforcement has waxed and waned, however since its initial conception an abundance of data has accumulated. Taken together, the results from innumerable studies in the area of conditioned reinforcement indicate several important questions remain regarding the establishment and maintenance of conditioned reinforcers. Additionally, there is a paucity of research aiming to reconcile existing discrepancies in effects across diverse measures of conditioned reinforcement. This symposium will highlight recent work with regard to conditioned reinforcement, including experiments investigating of the role of temporal variables in the establishment of conditioned reinforcers, assessing conditioned reinforcement within a delay-of-reinforcement context, examining the conditioned reinforcing function of multi-stimulus sequences, and exploring the effects of conditioned reinforcement on choice. The discussant will integrate the work presented with prior findings related to conditioned reinforcement. A panel on Contemporary Issues in Conditioned Reinforcement: Concept and Theory will immediately follow this symposium.

Neutral stimuli are thought to acquire the capacity to function as reinforcers for instrumental behavior via Pavlovian conditioning. Experiments in Pavlovian conditioning suggest that animals encode and remember the relative temporal proximities of conditioned stimuli (CSs) and unconditioned stimuli (USs) and integrate these relations across situations to predict significant events. This experiment examined whether temporal encoding of durations of events facilitates the acquisition of a new response with conditioned reinforcement. Two groups of rats received appetitive conditioning in which a 10-s CS predicted response-independent food deliveries. One group received food at the offset of the CS (Delay), and the other 10-s after CS offset (Trace). Both groups then experienced pairings of the 10-s training CS and a novel 10-s CS in backward order (CS1-CS2 pairing). Finally, we assessed the ability of CS2 to function as a conditioned reinforcer for a new response (lever-pressing). Results show that a backward-paired CS functioned to better support the acquisition of lever-pressing in the trace-conditioned group compared to a random control group and the delay-conditioned group. The results suggest that a CS became a better predictor of reinforcement through temporal integration of events.

Despite the plethora of studies conducted on conditioned reinforcement, the boundary conditions of conditioned reinforcement remain unclear. The present experiments, conducted with rats, investigated conditioned reinforcement within a delay-of-reinforcement context, as delays to reinforcement are ubiquitous and their effects on behavior are well understood. Experiments 1 and 2 offer a new procedure for examining conditioned reinforcement involving the observing of delay-of-reinforcement signals during non-resetting and resetting delays, respectively. In Experiment 3, preference for the delay signals observed during the first two experiments will be assessed using a well-established concurrent-chains procedure. The results of Experiments 1 and 2 indicate some of the conditions under which delay signals function as conditioned reinforcers, such as experience with signaled delays prior to experience with unsignaled delays, and reveal various mechanisms underlying conditioned reinforcement effects during reinforcement delays. The results from Experiment 3 should provide an assessment of both the validity of the novel observing response procedure and of preference for delay signals. Taken in sum, findings from the experiments presented contribute to the extant theoretical and methodological literature that informs the contemporary study of conditioned reinforcement.

Stimuli can enhance the ability of a reinforcer to control behavior at a temporal distance. This effect is traditionally thought to be a function of the stimulus becoming a conditioned reinforcer. However, exactly how reinforcement controls responding when a sequence of stimuli is presented is not well understood, as stimuli correlated with reinforcement could serve discriminative or conditioned reinforcing functions, or both. This study is a systematic replication of Cronin (1980), a study often cited as clear support for conditioned reinforcement. Pigeons chose between two stimuli. Responses to one always resulted in food after 60-s whereas responses to the other never resulted in food. In some conditions, stimuli presented during the first and last 10-s of the 60-s delay were consistent. In other conditions, stimuli presented during the first and last 10-s of the delay were inconsistent. Pigeons easily learned the task when the stimuli were consistent (i.e., they would increasingly choose the option that lead to food following the delay). Preliminary data from the inconsistent condition suggest that pigeons began to choose the option that did not lead to food more often. Taken together, the study provides additional support for the utility of the theoretical construct of conditioned reinforcement.

In human gambling, the investment is generally greater than the gain (suboptimal choice). A similar phenomenon can be found in pigeons when they show a preference for a conditioned stimulus associated with a low-probability, high-reward outcome (the jackpot) over a conditioned stimulus associated with a guaranteed low-reward outcome but one that has a higher overall value (not gambling). Research indicates that conditioned reinforcers play an important role in this suboptimal choice by pigeons. For example, it is the value of the conditioned reinforcers rather than the value of the choice alternatives that determine the pigeons choice (i.e., the results of winning rather than the probability of winning). Furthermore, the suboptimal choice is amplified by the finding that conditioned inhibition associated with the stimulus that predicts nonreinforcement declines with training (i.e., the high probability of losing fails to inhibit choice of the suboptimal alternative as one might think it should). Thus, just as with human gamblers, it is the high positive value of the rarely occurring conditioned reinforcer, together with the low negative value of the conditioned inhibitor that is responsible for pigeons (and likely humans) suboptimal choice.