Stimulus generalization and the peak shift with movement stimuli

Stimulus generalization is suggested as an alternative method for examination of the "novelty" problem in motor learning. These experiments demonstrated that stimulus generalization occurs using simple movements as stimuli. The phenomenon of the "peak shift" in post-discrimination generalization gradients was also examined. The first experiment demonstrated that a peak shift occurred using linear movements as stimuli and that the magnitude of the peak shift increased as the difference between the training stimuli decreased. The second experiment showed similar results when the stimuli consisted of a range of movements rather than single movement length. The final experiment provided evidence that perception of movement length is influenced by the magnitude of an immediately preceding movement. The relevance of these studies to current motor-learning theory is discussed.     

Spatial generalization and peak shift in humans

Using a computer betting game, five experiments tested university students on spatial generalization and peak shift. On each trial, one location was marked and the subject was invited to bet 0-4 points. At the winning location (S+), bets won four times the points betted. At nearby losing locations (S−s), points betted were lost. Generalization gradients were exponential in shape, supporting Shepard's (1987), law (Experiment 1). With peak shift manipulations, three kinds of peak shift or area shift were found. (1) Subjects betted more on the S+ side than on the S− side (Experiments 2-4). (2) When asked if a location was the winning location, subjects responded “yes” more often to locations on the S+ side than to locations on the S− side (Experiments 3-5). (3) When asked to point to the winning location on the screen, subjects' errors indicated peak shift (Experiment 5).     

Two determinants of the peak shift in human voluntary stimulus generalization

Two hundred and forty college students were divided into two groups, with training stimuli (from the brightness dimension) selected to produce small and large adaptation level shifts between discrimination training (to respond “same” to S+ and “different” to S—) and gen-eralization testing. These were further divided into three groups with discriminations expected to yield positive (toward brighter values), negative (toward dimmer values), or zero post-discrimination peak shifts. Half the subjects received brief discrimination training while half received extended training. A further group of 60 subjects were given exposure to the stimuli comparable to that of the extended training subjects, but were asked to rate the stimuli instead of being given discrimination training. The results indicated that two independent, additive sources of shift were active. One source, occurring in all groups, was interpreted as being due to the change in adaptation level from training to test. The other source of shift, occurring only in the groups with extended discrimination training, was interpreted as due to the establishment of asymmetrical decision criteria; the more traditional interpretation in terms of the interaction between excitatory and inhibitory gradients of response strength was also considered.     

Stimulus generalization, discrimination learning, and peak shift in horses.

Using horses, we investigated three aspects of the stimulus control of lever-pressing behavior: stimulus generalization, discrimination learning, and peak shift. Nine solid black circles, ranging in size from 0.5 in. to 4.5 in. (1.3 cm to 11.4 cm) served as stimuli. Each horse was shaped, using successive approximations, to press a rat lever with its lip in the presence of a positive stimulus, the 2.5-in. (6.4-cm) circle. Shaping proceeded quickly and was comparable to that of other laboratory organisms. After responding was maintained on a variable-interval 30-s schedule, stimulus generalization gradients were collected from 2 horses prior to discrimination training. During discrimination training, grain followed lever presses in the presence of a positive stimulus (a 2.5-in circle) and never followed lever presses in the presence of a negative stimulus (a 1.5-in. [3.8-cm] circle). Three horses met a criterion of zero responses to the negative stimulus in fewer than 15 sessions. Horses given stimulus generalization testing prior to discrimination training produced symmetrical gradients; horses given discrimination training prior to generalization testing produced asymmetrical gradients. The peak of these gradients shifted away from the negative stimulus. These results are consistent with discrimination, stimulus generalization, and peak-shift phenomena observed in other organisms.     

The peak shift in stimulus generalization: equivalent effects of errors and noncontingent shock

Terrace suggested that the peak shift in stimulus generalization occurs because the training stimulus not correlated with reinforcement has become aversive. This hypothesis is plausible in the light of instances where the peak shift is obtained compared with those where it fails to appear. The present experiment attempted to test implications of this hypothesis. Two groups of pigeons learned the same two-stimulus discrimination between colors by different training methods in a free-operant situation. When the discrimination was trained with many errors, a large peak shift was obtained in a subsequent generalization test of wavelength; after discrimination training with few errors, a negligible shift was observed. Half of each group then received noncontingent aversive shock during presentations of the stimulus not correlated with reinforcement in continued discrimination training. After this treatment, the errorless-shock subgroup showed a large peak shift and the error-shock subgroup tended to show a larger shift than before. Nonshocked control groups showed little change in the peak shift. It was concluded that pairing aversive shock with a stimulus not correlated with reinforcement is sufficient to produce or enhance a peak shift. In their effect on the peak shift, aversive shock and large amounts of nonreinforced responding appear to be equivalent.     

Relative and absolute density of reinforcement as factors influencing the peak shift

Four groups of six pigeons each were given nondifferential training on multiple variable-interval variable-interval reinforcement schedules and then were switched to differential training involving a multiple schedule in which reinforcement density was reduced in one of the two components. The multiple schedules used in the four groups had mean interreinforcement intervals of 1 min and 1 min in the two components changed to 1 min, 5 min; 2.5 min, 2.5 min changed to 2.5 min, 5 min; 12 sec, 12 sec changed to 12 sec, 24 sec; and 12 sec, 12 sec changed to 12 sec, 60 sec. In subsequently administered wavelength generalization tests, some peak shifts were observed in each condition and occurred occasionally in the absence of behavioral contrast or rate reduction in the less-reinforced component. The best predictor of peak shift was a high proportion of total responses emitted during the more-reinforced component at the end of differential training.     

Reinforcement duration and the peak shift in post-discrimination gradients

Pigeons were trained to key-peck for food, first with single-stimulus training and then with successive discrimination (multiple schedule) training. In the multiple schedule, two different wavelengths were each correlated with equally frequent variable-interval reinforcement but different durations (6 sec vs. 2 sec) of access to grain. For some birds, the different durations of feeding cycle were cued by different intensities of the food hopper light. For some of these “cued” birds, single-stimulus training had been carried out with 6-sec feedings and when multiple-schedule training was introduced, the novel stimulus was correlated with 2-sec feedings. For the others, 2-sec feedings were originally used, and the novel stimulus was then present during the 6-sec reinforcement duration. The cueing procedure enhanced discrimination performance, and was necessary for the consistent production of a peak shift. In addition, the condition in which original training had been carried out with 6-sec feedings, and thus reinforcement duration was reduced in the presence of the novel stimulus, led to the best performance.     

Stimulus generalization: the ordering and spacing of test stimuli

Twenty-four pigeons learned a successive discrimination between 500 mmu (S+) and 574 mmu (S-). When tested in extinction, some birds received stimuli around S+, with no S- presentations. These birds showed a positive peak shift, with maximum responding not at 550 mmu, but displaced to 538 mmu and 544 mmu. Other birds were tested with stimuli around S-, with no S+ presentations. These birds showed a negative shift, with least responding not at 574 mmu, but at 586 mmu. Though the first group was tested around S+ and the second around S-, total responding between groups did not differ. When retested on the other half of the continuum, however, birds that had gone from the S+ half to the S- half responded fewer times than those that had gone from the S- half to the S+ half. In a second experiment, reducing stimulus spacing from 6 mmu to 2 mmu produced flatter gradients and decreased the amount of positive shift. In a third experiment, birds were tested across the whole continuum with stimuli presented in serial order. A sequence from 538 mmu to 586 mmu produced no responding after the first part of the session; a sequence from 586 mmu to 538 mmu produced responding throughout the session.     

An explanation for the presentation-order effect in the method of constant stimuli

The point of subjective equality obtained by the method of constant stimuli depends to a great extent on whether the standard (S) or the variable (V) stimulus occurs first. This presentation-order effect was studied using lines as stimuli. Successive S, V pairs were presented, with inter-stimulus and interpair intervals equal. Observers, who were not told which was S or V, reported whether a given line was longer or shorter than the immediately preceding line. Although the observers' subjective experience was of a train of lines that was not organized into pairs, the presentation-order effect still occurred. This implies that the effect does not depend on the order of presentation of the stimuli in an experienced pair. It was also shown that the observers could categorize line lengths, since they could identify stochastically the most frequent stimulus (S). We propose that the presentation-order effect depends on a decision process based on response probabilities inferred from length categories.