Discrimination Training

Description of the Strategy

Discrimination training involves the use of selective reinforcement and extinction to generate differential responding to two or more stimuli. Training automobile drivers to stop when the light is red and to go when the light is green is an example. The act of responding differentially to stimuli (e.g., different colored lights) is called stimulus discrimination. It is the goal of discrimination training to elicit stimulus discrimination.

The process of discrimination training within laboratory settings involves repeated trials in which two or more stimuli are presented (concurrently or sequentially) and the subject is provided an opportunity to obtain reinforcement, given the presentation of one but not the other. As a result of such training, responding will come to occur in the presence of the stimulus that signals the availability of reinforcement. That stimulus is often called the S+ or SD (pronounced “ess dee”). The stimulus that signals the lack of reinforcement is often called the S− or SD (pronounced “ess delta”). For example, during discrimination training, a 400-Hz tone might be consistently followed by the presentation of reinforcement, while a 600-Hz tone is not. In this case, the 400-Hz tone is the S+ or SD and the 600-Hz tone is the S− or SD.

Stimulus discrimination training often involves a graduated approach to reducing the stimuli that give rise to a certain response. To accomplish this, successively similar stimuli are presented, and responding to one type is reinforced, while responding to the other is not. In the above example, on subsequent trials reinforcement will continue to be given to responses that follow a 400-Hz tone. However, presentations that do [Page 811]not signal the availability of reinforcement may include not only 600-Hz tones but also 500-Hz tones and 450-Hz tones. In this example, training leads to the discrimination of stimuli that are increasingly similar and more difficult to discriminate.

Early experimental studies of discrimination training revealed a phenomenon referred to as experimental neurosis, in which experimental participants illustrated unusual behavior arising out of exposure to increasingly difficult discriminations. In a classic study, a dog was trained to salivate to a circle but not to ellipses. Over the course of repeated trials, the shape of the ellipses presented to the dog became more and more similar to that of a circle. As the discrimination became more difficult to make due to the increased similarity between the shapes, the dog began to display heightened levels of restlessness, agitation, and unprovoked aggressiveness—behaviors resembling anxiety or neurosis observed in humans. Pavlov explained this unusual behavior as arising from a conflict between cortical excitation and inhibition.

It was subsequently argued that the conditions giving rise to experimental neuroses in animals may resemble the nonlaboratory conditions that generate neuroses in humans. In particular, the experimental conditions were seen as resembling human environments in which it is difficult to identify conditions leading to reinforcement and those leading to no reinforcement or to punishment. Such unpredictability—arising from either randomness or stimulus discrimination failure—is thought to be an environmental condition leading to a subjective experience of anxiety and also overtly anxious and, perhaps, aggressive behavior.

...