E. J. Capaldi

Stimulus control of responding in the early trials of differential conditioning

Abstract In Experiment 1 two groups of rats were given 12 differential conditioning trials, seven to the rewarded alley (S+) and five to the nonrewarded alley (S−), prior to being extinguished in both alleys. Group S−S+ received S+ trials, following S− trials in acquisition, while Group S+S− did not receive S+ trials following S− trials in acquisition. In extinction S+ and S− trials were presented according to a quasi-random sequence for both groups. Running on the last 3 trials of acquisition was found to be faster following S+ than following S− trials. Group S−S+ showed greater resistance to extinction and less discriminative responding in extinction than Group S+S−. These results suggest that responding in differential conditioning is controlled not merely by S+ and S− but by the memories of reward (SR) and of nonreward (SN) as well. When the joint effects of both classes of cues were considered, e.g., SR+S+, responding in the early trials of differential conditioning was shown to be highly orderly. Experiment 2 was highly similar to Experiment 1 except that Groups S−S+ and S+S− were equated along dimensions not equated in Experiment 1. The results obtained in Experiment 2 were highly similar to those obtained in Experiment 1.

A role of stimulus compounds in eliciting responses: Relatively spaced extinction following massed acquisition

One of two schedules of rewarded (R) and nonrewarded (N) trials (RNR vs. RRN) was combined factorially with intertrial interval (ITI) in acquisition (1 vs. 45 min), with extinction occurring at a 45-min ITI. The RNR schedule produced greater resistance to extinction than the RRN schedule regardless of acquisition ITI. The shift in ITI from acquisition to extinction reduced resistance to extinction slightly in the RNR group but not in the RRN group. These findings suggest that there are cues common to 1-min and 45-min ITIs. It was suggested that these ITI-associated cues enter into compound with memories to control instrumental responding.

Aversive Learning Situations: Apparatus and Procedures

Publisher Summary This chapter discusses the apparatus and procedures of aversive learning situations. The ideal punishing or aversive stimulus has the following characteristics. It should have precise physical specification, the contact the stimulus makes with the subject should be constant, the subject should not be able to escape or minimize stimulation by means of some unauthorized or unwanted behavior, there should be few skeletal reactions to the stimulus, and the experimenter should have the option of varying the aversive stimulus over a wide range of values. The aversive stimulus that best meets these criteria is, perhaps, shock. Footshock has been the most extensively used aversive stimulus. Other aversive stimuli that have been used are air blasts with cats, bar slap with rats, noise with humans, extremes of water temperature, and even objects that elicit fear reactions in animals, for example, monkeys fear moving toy bears. A number of parameters of aversive stimuli have been manipulated. Among such variations are the intensity of the aversive stimulus, the duration of the aversive stimulus, the frequency or percentage of the aversive stimulus, the schedule such as fixed ratio according to which the aversive stimulus is delivered, the immediacy or delay of the aversive stimulus, and whether the aversive stimulus is introduced at full strength or increased gradually over a series of trials.

An instrumental partial reinforcement effect in the absence of any overt instrumental acquisition training

Rats were placed directly over a food cup or received ordinary runway trials under either partial or consistent food reward, no other acquisition training having been given. Also included was a group that was nonrewarded on all running trials. A gradient occurred in extinction such that type of acquisition training (placed vs run) exercised less influence over performance and schedule of reward (partial vs continuous) exercised more influence over performance both as extinction training progressed and as the goal section was approached. The results suggest that rats make small terminal instrumental reactions on placement trials, reactions which, if strongly conditioned, can support a partial reinforcement effect in the full instrumental reaction, particularly in the goal area.

Amobarbital vs saline extinction following different magnitudes of consistent reinforcement

Rats were extinguished in a runway following acquisition under large consistent reward (20 45-mg pellets) vs small consistent reward (2 pellets). Large reward reduced resistance to extinction under amobarbital and under saline; in the goal section, the reducing effects of large reward were greater under amobarbital than under saline. The reducing effect of large reward on resistance to extinction thus appears to be independent of the absolute level of frustration occurring in extinction.

A discrepancy between anticipated reward and obtained reward with no increase in resistance to extinction

One of two varied magnitude-of-reward groups, trained in a runway under conditions designed to preclude anticipation of the smaller reward magnitude, nevertheless failed to show increased resistance to extinction, a result consistent with the sequential hypothesis. Since this group must have experienced frustration or dissonance in acquisition, its failure to show increased resistance to extinction is inconsistent with those hypotheses. This result suggests either that frustration or dissonance is irrelevant in determining extinction performance or that both hypotheses have failed to specify and identify the conditions that are both necessary and sufficient for increasing resistance to extinction.