Partial reinforcement effects on learning and extinction of place preferences in the water maze

In two experiments, two groups of rats were trained in a navigation task according to either a continuous or a partial schedule of reinforcement. In Experiment 1, animals that were given continuous reinforcement extinguished the spatial response of approaching the goal location more readily than animals given partial reinforcement—a partial reinforcement extinction effect. In Experiment 2, after partially or continuously reinforced training, animals were trained in a new task that made use of the same reinforcer according to a continuous reinforcement schedule. Animals initially given partial reinforcement performed better in the novel task than did rats initially given continuous reinforcement. These results replicate, in the spatial domain, well-known partial reinforcement phenomena typically observed in the context of Pavlovian and instrumental conditioning, suggesting that similar principles govern spatial and associative learning. The results reported support the notion that salience modulation processes play a key role in determining partial reinforcement effects.     

Latent discrimination learning under a regular schedule of partial reinforcement

In this investigation, which employed rats in a runway, discriminative responding consisted of faster running on the reinforced than on the nonreinforced trials of either the 4NR or R4N schedule, both schedules containing fixed, repeated sequences of nonreinforced and reinforced trials. Under the 4NR schedule, four nonreinforced trials preceded a reinforced trial each day, and under the R4N schedule, a reinforced trial was followed by four nonreinforced trials each day. The major finding obtained was that under the 4NR schedule, discriminative responding was improved very substantially by a shift to extinction. Rats maintained on the 4NR schedule did not show improved discriminative responding, nor did discriminative responding improve in extinction following training under either the R4N schedule or a schedule of consistent reinforcement. Latent discrimination learning was defined as discriminative responding which fails to reflect adequately the amount of discrimination learning accomplished. The present findings demonstrate latent discrimination learning for regular schedules of partial reinforcement, something already demonstrated for brightness differential conditioning and possibly DRL schedules, as well.     

Contrast effects in simultaneous discrimination learning

Pigeons were trained on a form discrimination that alternated across trials with an already-learned color discrimination, which was associated with different probabilities of reinforcement for different groups. The rate of acquisition of the form discrimination varied inversely with the rate of reinforcement for the color discrimination, thus demonstrating a “contrast effect” in simultaneous discrimination learning.     

Feature-positive and feature-negative discrimination learning maintained by negative reinforcement in rhesus monkeys

Rhesus monkeys were trained to press panels to escape or avoid shock. Escape and avoidance training were combined factorially with feature-positive (FP) and feature-negative (FN) discriminations. In the FP situation, S+ trials were characterized by one red and two green panel displays and S? consisted of an all green panel array. In the FN arrangement, S+ and S? arrays were reversed. When compared properly, FP discriminations were learned more readily than FN discriminations in both escape and avoidance training conditions. Positive tracking was associated with FP discriminations, and feature avoidance characterized the FN tasks. These data suggest that in primates, feature responses (i.e., tracking behaviors) are more a function of the operant contingency than of the Pavlovian correlation.     

A behavioral field approach to instrumental learning in the rat: I. Partial reinforcement effects and sex differences

The behavioral field approach employs naturalistic observation and simultaneous, multiple recordings of ecologically relevant aspects of behavior-environment interactions. Applying this approach to runway learning in the rat, the inferior acquisition speed of partial reinforcement (PRF) subjects as compared to continuous reinforcement (CRF) subjects was found to result from greater response variability, more sniffing and goal-avoidance behavior, and slower dropping out of collateral behaviors (e.g., drinking and sand-digging). Extinction first produced an increase in exploratory behavior, then displacement activities (e.g., grooming and biting) and goal-avoidance. CRF subjects showed greater response persistence as measured by number of extinction trials to disrupt an established, favored path. PRF subjects showed greater goal persistence as measured by trials to retrace from goalbox. In extinction, while CRF subjects were more inclined to engage in drinking and sand-digging in the startbox, PRF subjects exhibited more biting behavior in the goalbox. The only sex-related differences were superior speeds by female CRF subjects, inferior goal speeds by female PRF subjects during acquisition, and superior goalbox escape learning by females in extinction.     

Behavioral and associative effects of differential outcomes in discrimination learning

The role of the reinforcer in instrumental discriminations has often been viewed as that of facilitating associative learning between a reinforced response and the discriminative stimulus that occasions it. The differential-outcome paradigm introduced by Trapold (1970), however, has provided compelling evidence that reinforcers are also part of what is learned in discrimination tasks. Specifically, when the availability of different reinforcing outcomes is signaled by different discriminative stimuli, the conditioned anticipation of those outcomes can provide another source of stimulus control over responding. This article reviews how such control develops and how it can be revealed, its impact on behavior, and different possible mechanisms that could mediate the behavioral effects. The main conclusion is that differential-outcome effects are almost entirely explicable in terms of the cue properties of outcome expectancies—namely, that conditioned expectancies acquire discriminative control just like any other discriminative or conditional stimulus in instrumental learning.     

Facilitating stimulus effects of reward and punishment in discrimination learning

To demonstrate a facilitating stimulus effect, as opposed to an incentive effect, of food reward, rats were trained on an easy, light-dark discrimination with different amounts of reward for correct and incorrect responses (1-0, 2-0, 3-1, and 5-1 pellets, respectively), and with shock or no shock administered in the correct goalbox. Both errors and trials to criterion were fewer with a large reward differential (LRD: 2-0 and 5-1), as compared with a small reward differential (SRD: 1-0 and 3-1), but were not affected by the “base” reinforcement condition of either 1 or 0 pellets for the incorrect response. In addition, choice and arm speeds during early training were positively related to the combined, or average, number of pellets contingent upon both correct and incorrect responses, indicating a generalization of reward expectancies. Although shock uniformly suppressed arm speeds under all reward conditions, it facilitated discrimination learning in the SRD conditions. That such facilitation occurred only when the conditions of reward for correct and incorrect responses were relatively similar indicates that not only shock, but also food can function as a distinctive cue: As a stimulus selectively applied to one response, it can decrease the similarity of the alternatives, and, in this manner, it can faciltate performance.     

The effects of intertrial and feature-target intervals on operant serial feature-positive discrimination learning

The effects of intertrial interval (ITI) and feature-target interval (FTI) on the nature of learning in discrete-trial operant serial feature-positive (feature → target+ / target?) discrimination training were examined in two experiments with rats. Discrimination performance was acquired more rapidly with longer ITIs and shorter FTIs. In contrast, transfer to a separately trained target was greater with shorter ITIs regardless of FTI. Persistence of discrimination performance after feature extinction was greater with longer ITIs. Only the last of these performance measures showed evidence for invariance with constant ITI/FTI ratios. The results are discussed in the context of theories of occasion setting.     

Serial discrimination reversal learning in pigeons as a function of intertrial interval and delay of reinforcement

Pigeons learned a series of reversals of a simultaneous red-green visual discrimination. Delay of reinforcement (0 vs. 2 sec) and intertrial interval (ITI; 4 vs. 40 sec) were varied across blocks of reversals. Learning was faster with 0-sec than with 2-sec delays for both ITI values and faster with 4-sec ITIs than with 40-sec ITIs for both delays. Improvement in learning across successive reversals was evident throughout the experiment, furthermore, even after more than 120 reversals. The potent effects of small differences in reinforcement delay provide evidence for associative accounts and appear to be incompatible with accounts of choice that attempt to encompass the effects of temporal parameters in terms of animals’ timing of temporal intervals.     

Partial reinforcement in autoshaping with pigeons

The acquisition, maintenance, and extinction of autoshaped responding in pigeons were studied under partial and continuous reinforcement. Five values of probability of reinforcement, ranging from .1 to 1.0, were combined factorially with five values of intertrial interval ranging from 15 to 250 sec for different groups. The number of trials required before autoshaped responding emerged varied inversely with the duration of the intertriai interval and probability of reinforcment, but partial reinforcement did not increase the number of reinforcers before acquisition. During maintained training, partial reinforcement increased the overall rate of responding. A temporal gradient of accelerated responding over the trial duration emerged during maintenance training for partial reinforcement groups, and was evident for all groups in extinction. Partial reinforcement groups responded more than continuous reinforcement groups over an equivalent number of trials in extinction. However, this partial-reinforcment extinction effect disappeared when examined in terms of the omission of “expected” reinforcers.     

Partial reinforcement and resistance to extinction in honeybees

Previous experiments have shown the partial reinforcement effect in honeybees under conditions which permit an interpretation in terms of sensory carryover. In the five experiments reported here, the effect was sought under conditions which would require an interpretation in terms of associative reinstatement. Since it is not feasible to train honeybees in widely spaced trials, several different interpolated-trials procedures were employed which had in common the feature that nonrewarded response to a stimulus never was followed by rewarded response to the same stimulus. Implications of the negative results for the interpretation of the overlearning-extinction effect and successive negative contrast in honeybees are considered.     

Partial reinforcement effect: The expectancy of reward on nonreward trials

In order to determine the importance of the development of expectancy of reward prior to partial reward trials; rats were given 20 continuously reinforced trials prior to 20 partially reinforced trials (CRF-PRF) and compared to Ss given only 20 partially reinforced trials (PRF). Control groups received 20 or 40 continuously reinforced trials (CRF-20, CRF-40) to determine the effect of differing numbers of acquisition trials. Results showed that terminal acquisition differences were minimal in the run segment of the alley and that Group CRF-PRF was more resistant to extinction than Group PRF, and both were more resistant to extinction than the CRF-20 and CRF-40 groups, which did not differ from each other. These results were interpreted as supporting the notion that the expectancy of reward on nonreward trials during partial reinforcement acquisition is a determiner of the magnitude of the partial reinforcement extinction effect.     

Reward pretraining and discrimination learning set

Pigtailed monkeys were trained on discrimination problems involving objects that had been previously paired with reward, previously paired with nonreward, or objects not associated with any pretraining. In two separate studies, no differences on these various problem formats were observed, but performance improved considerably with practice. These results are contrary to the theories for learning set formation of Spence and Medin.     

Simultaneous discrimination learning: Stimulus interactions

Although considerable attention has been given to interactions between events serving as the positive (S+) and negative (S−) stimuli in successive discriminations, there has been little study of similar interactions in simultaneous discriminations. We propose that in a simultaneous discrimination, given what is typically relatively little experience with the consequences of responding to the S−, some of the value of the S+ transfers to the S− with which it was paired. Furthermore, the mechanisms responsible for this transfer of value appear to be the higher order Pavlovian association between the S− and the S+, as well as within-event associations between them. Although in typical simultaneous discriminations, negative value does not appear to transfer from the S− to the S+, when adequate experience is provided with the S−, contrast typically develops, reducing the value of the S− (negative contrast) and enhancing the value of the S+ (positive contrast). This model of stimulus interactions has implications not only for simple simultaneous discrimination learning, but also for research using combinations of discriminations (e.g., transitive inference in animals and humans). This model may also have implications for a number of human social psychological phenomena.