Brian M. Kruger

Free choice of signaled vs unsignaled scrambled electric shock with rats

After forced exposure to both sides of a shuttlebox, 15 rats were given a choice between signaled scrambled electric shock in one half and unsignaled scrambled shock in the other. The interstimulus intervals between tone and shock were 2, 4, 8, 16, or 32 sec. Signal offset contingencies were: off 1 sec before shock, off at shock onset, or off at shock offset. The main finding of no preference between signaled and nonsignaled shock supports the view that signaled shock will be preferred only when it is modifiable.

Some effects of type of auditory CS on self-punitive running in rats

Self-punitive behavior was demonstrated in prepunishment speeds during extinction following shock escape training in a straight runway under buzzer-CS and tone-CS conditions. Relative to the tone, the buzzer enhanced punished running (shock in the last half of the runway) and nonpunished running. These results extend the findings of Myers (1962), who reported better avoidance conditioning with a buzzer than with a tone. Self-punitive behavior, defined as the difference between punishment conditions with faster running under punishment conditions, was not significantly enhanced by the buzzer. These results do not support the suggestion (Delude, 1973) that self-punitive behavior may be uniquely affected by the use of a buzzer CS.

Magnitude of water reward in the runway: A parametric investigation

Forty rats were given 42 acquisition and 15 extinction trials in a runway at 1 trial/day. The magnitude of water reward during acquisition was varied across five levels for independent groups (.1, 1.1, 2.1, 3.1, and 4.1 ml). Acquisition response speeds were an inverted-U-shaped function of the magnitude of reward. The results of the extinction phase showed that the rats that received the smallest reward during training showed the greatest persistence during extinction. The extinction performance of the rats that received the larger rewards during training did not differ. The results of both the acquisition and extinction phases were seen as generally consistent with the literature on the magnitude of food reward.

Runway performance as a function of the schedule and magnitude of water reward

Sixty rats were trained in a runway for 42 days of acquisition and 15 days of extinction at the rate of one trial a day. Half of the animals received a 50% partial reinforcement (PR) schedule and half received a 100% reinforcement schedule (CRF). For one-third of the rats within each schedule the magnitude of water reward was.1,.4, or.8 ml. Persistence during extinction was an increasing function of reward magnitude for the PR animals, but a decreasing function of reward magnitude for the CRF animals. A partial reinforcement extinction effect was observed with the medium and large reward but not with the smallest reward. These data are consistent with the literature dealing with food reward.

The effect of reversal shifts and scrambled shock on preference for signaled shock established with unscrambled shock

Four rats were given free choice of sides in a shuttlebox. Each rat was enclosed in each side for 30 min prior to each daily 1-h choice session. Shocks of 2 sec duration were administered on both sides at irregular intervals. Shocks were preceded by a tone signal on one side. All rats showed a reliable preference for the signaled shock side in the first phase. Reversal of the visual stimuli differentiating the signal side from the nonsignal side eliminated the preference for all rats. Reinstatement of the original relations resulted in the reestablishment of the preference for two rats. In the next phase, changing to scrambled shock from the previously used unscrambled shock eliminated the preference for these two rats. The preference was not reestablished when unscrambled shock was again used in the final phase. The results are interpreted as supporting preparatory response and conditioned aversiveness interpretations of some of the phenomena in signaled shock studies.

Some effects of methylphenidate on self-punitive running in rats

Self-punitive behavior was demonstrated in prepunishment speeds during extinction following shock-escape training in a straight runway under no-dose, low-dose (1-mg/kg), and high-dose (10-mg/kg) methylphenidate conditions. Increased dosage enhanced punished running (shock in the last half of the runway) and nonpunished running. Self-punitive behavior, defined as the difference between punishment conditions with faster running or increased persistence under punishment conditions, was not significantly affected by the drug variable. The results were interpreted as compatible with conditioned fear interpretations of self-punitive behavior but nonsupportive of cognitive interpretations.

Some effects of a buzzer CS and a novel buzzer on self-punitive running in rats

Self-punitive behavior was demonstrated in prepunishment speeds during extinction following shock escape training in a straight runway under buzzer-CS and under no-CS conditions. The buzzer enhanced punished running (shock in the last half of the runway) and nonpunished running. The buzzer did not significantly enhance self-punitive behavior, defined as the difference between punishment conditions with faster running for punished rats than for nonpunished rats. Introduction of a novel buzzer at the beginning of extinction enhanced both punished and nonpunished running relative to no-CS conditions but had no effect on selfpunitive behavior. Discontinuance of the buzzer CS at the end of acquisition eliminated selfpunitive behavior during extinction. The results were interpreted as supporting explanations of self-punitive behavior in terms of fear conditioning and conflict resolution.

Rats can learn a probability discrimination based on previous trial outcomes in partial reward schedules

In each of two experiments, rats were trained in a runway with either an alternating or an irregular schedule of partial reward. One-third of the animals within each schedule were trained with a small, medium, or large reward. In Experiment 1, the irregular schedule was constructed in such a way that the probability of an R trial following an N trial was.75 and the probability of an N trial following an R trial was also.75. Of course, in an alternating schedule the probabilities are 1.0. The rats learned to run rapidly following N trials and slowly following R trials when trained under either schedule. This effect, however, was greater in the alternating condition and was facilitated by larger rewards. In Experiment 2, the irregular schedule was constructed so that the occurrence of R or N trials was not a reliable predictor of the following trial. Under these conditions, differential running speeds did not develop in the irregular schedule, but the rats in the alternating schedule performed as in Experiment 1. The data are discussed within the context of probability learning and are related to methodological issues of importance to sequential trial experiments.

Barpress variability as a function of two methods of body-weight control

Two methods of food deprivation used to control hunger motivation were compared on their ability to control body weight and to control response variability in an appetitive conditioning situation. Percent-weight rats received individually adjusted daily rations to maintain them at 85% of free-feeding body weight. Fixed-food rats each received the average daily ration given the percent-weight subjects. All were given 34 days of training on an FR 10 schedule followed by 15 days on a VI 30-sec schedule. There were no significant differences in mean body weights or mean barpressing rates. Variability in body weight was significantly greater for fixed-food rats, but there was no significant difference in barpress variability. Since experimental error was not reduced by the more costly percent-weight method, it is suggested that the fixed-food method be used for experimental conditions similar to those used in the present study.

SELECTIVE LEARNING UNDER TWO LEVELS OF DRIVE: CONTIGUITY VS REINFORCEMENT REVISITED

Theoretical analyses, using Spences (1958) theory, of two idealized discrimination situations show that the contiguity and reinforcement assumptions of habit strength development need not result in differential predictions for percent-choice performance of two drive-level groups when the numbers of occurrences of responses are equated. For actual situations approximating 100% generalization (all common cues), both assumptions effectively generate the prediction of no difference in performance. Review of the studies cited by Spence, Goodrich, and Ross (1959) as supporting the contiguity view suggests that situations of this type were employed.