Inhibitory stimulus control following errorless discrimination learning

Three generalization procedures were used to investigate inhibitory stimulus control following discrimination learning with few errors. Three groups of pigeons acquired a discrimination between a green stimulus (the positive stimulus) and a vertical or horizontal line (the negative stimulus) through differential autoshaping followed by multiple schedule presentation of the two stimuli with gradually increasing stimulus durations. Genereralization testing was along a line-tilt continuum. For one group, the test involved a resistance-to-reinforcement procedure in which responses to all line tilts were reinforced on a variable-interval schedule. For a second group, also tested with the resistance-to-reinforcement procedure, the lines were superimposed on the green field that formerly served as the positive stimulus. A third group was tested in extinction with the combined stimuli. Control groups had no discrimination training but responding to green was nondifferentially reinforced. The control subjects responded more to all line tilts during testing than did the comparable experimental subjects, indicating that the negative stimulus had become an inhibitory stimulus. Both resistance-to-reinforcement groups revealed inhibitory gradients around the negative stimulus, but the gradient for the extinction group was relatively flat. These data are consistent with others that modify Terrace's early conclusion concerning the failure of inhibition to develop during errorless training.     

The interaction of temporal generalization gradients predicts the context effect

In a temporal double bisection task, animals learn two discriminations. In the presence of Red and Green keys, responses to Red are reinforced after 1-s samples and responses to Green are reinforced after 4-s samples; in the presence of Blue and Yellow keys, responses to Blue are reinforced after 4-s samples and responses to Yellow are reinforced after 16-s samples. Subsequently, given a choice between Green and Blue, the probability of choosing Green increases with the sample duration-the context effect. In the present study we asked whether this effect could be predicted from the stimulus generalization gradients induced by the two basic discriminations. Six pigeons learned to peck Green following 4-s samples (S(+)) but not following 1-s samples (S(-)) and to peck Red following 4-s samples (S(+)) but not following 16-s samples (S(-)). Temporal generalization gradients for Green and Red were then obtained. Finally, the pigeons were given a choice between Green and Red following sample durations ranging from 1 to 16 s. Results showed that a) the two generalization gradients had the minimum at the S(-) duration, an intermediate value between the S(-) and the S(+) durations, and the maximum at the S(+) as well as more extreme durations; b) on choice trials, preference for Green over Red increased with sample duration, the context effect; and c) the two generalization gradients predicted the average context effect well. The Learning-to-Time model accounts for the major trends in the data.     

Delayed temporal discrimination in pigeons: A comparison of two procedures

A within-subjects comparison was made of pigeons' performance on two temporal discrimination procedures that were signaled by differently colored keylight samples. During stimulus trials, a peck on the key displaying a slanted line was reinforced following short keylight samples, and a peck on the key displaying a horizontal line was reinforced following long keylight samples, regardless of the location of the stimuli on those two choice keys. During position trials, a peck on the left key was reinforced following short keylight samples and a peck on the right key was reinforced following long keylight samples, regardless of which line stimulus appeared on the correct key. Thus, on stimulus trials, the correct choice key could not be discriminated prior to the presentation of the test stimuli, whereas on position trials, the correct choice key could be discriminated during the presentation of the sample stimulus. During Phase 1, with a 0-s delay between sample and choice stimuli, discrimination learning was faster on position trials than on stimulus trials for all 4 birds. During Phase 2, 0-, 0.5-, and 1.0-s delays produced differential loss of stimulus control under the two tasks for 2 birds. Response patterns during the delay intervals provided some evidence for differential mediation of the two delayed discriminations. These between-task differences suggest that the same processes may not mediate performance in each.     

Category discrimination by pigeons using five polymorphous features

Eight pigeons were trained to discriminate between sets of color photographs of natural scenes. The scenes differed along five two-valued dimensions (site, weather, camera distance, camera orientation, and camera height), and all combinations of the feature values were used. One value of each dimension was designated as positive, and slides containing three or more positive feature values were members of the positive stimulus set. Thus, each feature had an equal, low, correlation with reinforcement, and all features had zero correlations with each other. Seven of the 8 pigeons learned this discrimination, and their responding came under the control of all five features. Within the positive and negative stimulus sets, response rates were higher to stimuli that contained more positive feature values. Once discrimination had been achieved, reversal training was given using a subset of the slides. In this subset, only a single feature was correlated with reinforcement. All pigeons learned this reversal successfully and generalized it to additional photographs with the same feature content. After reversal, the original reinforcement contingencies were reinstated, and training was continued using all the slides except those that had been used in reversal. Reversal generalized to these slides to some extent. Analysis of the response rates to individual slides showed that, compared with prereversal training, only the feature that had been subjected to reversal contingencies showed a reversed correlation with response rate. The remaining features showed the same correlation with response rate as they had before reversal training. Thus, reversal on some members of a category following category discrimination training led to generalization to stimuli within the category that were not involved in the reversal, but not to features that were not reversed. It is therefore inappropriate to refer to the pigeons as learning a concept.     

DOES CONDITIONAL DISCRIMINATION LEARNING BY PIGEONS NECESSARILY INVOLVE HIERARCHICAL RELATIONSHIPS?

Four experiments demonstrate that when putative conditional and discriminative cues are presented simultaneously in the single reversal procedure, it is not possible to ascribe a uniquely conditional or uniquely discriminative function to either of the cues. In Experiment 1, pigeons were trained to respond to a blue key and not to a red key while the houselight was on; then in a different session they learned the reversal of this discrimination with the houselight off (single reversal). Separate groups were tested for color generalization with houselight conditions alternating in blocks of trials or for houselight intensity generalization with blue and red key colors alternating in blocks of trials. Both test procedures revealed a conditional relationship between houselight and key color conditions. Experiment 2 produced the same result following training in which the key colors were held constant across training sessions while the houselight and no houselight conditions varied within sessions. In Experiment 3, separate groups were trained with the two procedures but were tested with randomly ordered combinations of key colors and houselight intensities. The two groups yielded indistinguishable bidimensional generalization gradients with peaks at both previously reinforced stimulus combinations. In Experiment 4 the subjects were switched from one of these training procedures to the other with no decrement in their discriminative performance. We conclude that for successive discriminations between conditional- and discriminative-stimulus combinations, the notion of a hierarchical relation between conditional and discriminative stimuli must be extended to include a symmetrical relationship or the notion should be abandoned altogether.     

Inhibitory control and errorless discrimination learning

Pigeons learned to discriminate between a positive stimulus (white key) and a negative stimulus (red or green key, depending on the subject) via Terrace's fading procedure. Generalization tests, conducted with intermittent reinforcement for key pecking at various wavelengths, yielded minima at the value of the negative stimulus in most “errorless” birds. Terrace's contrary finding of flat gradients in errorless subjects probably resulted from a floor-effect (i.e., virtually zero responding) produced by his extinction-test procedure. The present and other findings do not support Terrace's conclusions that the negative stimulus of an errorless discrimination is behaviorally neutral; inhibition apparently develops to the nonreinforced stimulus even during errorless discrimination learning. A negative correlation between stimulus and reinforcer seems the crucial factor in producing an inhibitory stimulus.     

The effect of sample duration and cue on a double temporal discrimination

To test the assumptions of two models of timing, Scalar Expectancy Theory (SET) and Learning to Time (LeT), nine pigeons were exposed to two temporal discriminations, each signaled by a different cue. On half of the trials, pigeons learned to choose a red key after a 1.5-s horizontal bar and a green key after a 6-s horizontal bar; on the other half of the trials, they learned to choose a blue key after a 6-s vertical bar and a yellow key after a 24-s vertical bar. During subsequent test trials, they were exposed to the horizontal or vertical bar, for durations ranging from 1.5 to 24 s, and given a choice between novel key combinations: red vs. yellow, or green vs. blue. Results showed a strong effect of sample duration—as the test signal duration increased, preference for green over blue increased and preference for red over yellow decreased. The effect of sample cue was obtained only on the green-blue test trials. These effects are discussed in light of SET and LeT.     

Generalization and response mediation of a conditional discrimination

Fifteen pigeons were given conditional discrimination training in which a colored sample stimulus determined which of two line comparison stimuli (vertical and horizontal) was correct. As part of the conditional discrimination procedure, birds were required to make an "observing response" to the sample stimulus presented on a wide key. The location on this key of the required observing response for the two sample stimuli differed by 0, 3, or 6 in. (0, 7.6, or 15.2 cm) for three groups of birds. Accuracy of conditional discrimination performance was directly related to the amount of separation. In subsequent generalization tests with novel sample stimuli, both observing-response location and comparison responding changed within the same region of the wavelength continuum from that appropriate for one of the training samples to that appropriate for the other. A maintained generalization test (continued reinforcement for training stimuli) revealed this relation more strongly. A test in which observing-response location was the only sample stimulus of a conditional discrimination revealed stimulus control by this observing response, supporting a response mediation interpretation of the data.     

Context specificity of operant discriminative performance in pigeons: II. Necessary and sufficient conditions

Six experiments were performed to explore the necessary and sufficient conditions for producing context specificity of discriminative operant performance in pigeons. In Experiment 1, pigeons learned a successive discrimination (red S+/blue S−) in two chambers that had a particular odor present and between which they were frequently switched. The birds subsequently learned the reversal (blue S+/ red S−) in one of these chambers with a different odor present. When switched to the alternative chamber, although the odor and the reinforcement contingency were still appropriate to the reversal, performance appropriate to the original discrimination recurred in subjects for which the houselights were on during training and testing but not for those for which the houselights were off. This indicated the importance of visual contextual cues in producing context specificity. Experiment 2 showed that the frequent switching between boxes in initial training was of no consequence, presumably because the apparatus cues were highly salient to the subjects. Experiment 3 showed significantly less context specificity when odor cues were omitted. Experiment 4 showed that simply using a different reinforced stimulus in each phase of training was ineffective in producing context specificity. Experiment 5 showed that the generalization test procedure used in Experiment 4 was sensitive to context specificity when discrimination-reversal training was used with different odors in the two training phases. Experiment 6 replicated the results of Experiment 4, but then showed that when different odors accompanied the two training phases, context specificity was obtained with the single-stimulus paradigm. Thus in both single-stimulus and discrimination-reversal paradigms, redundant odor cues potentiated learning about apparatus cues.     

Effects of variable-interval value and amount of training on stimulus generalization.

In Experiment 1 pigeons pecked a key that was illuminated with a 501-nm light and obtained food by doing so according to a variable-interval (VI) schedule of reinforcement, the mean value of which differed across groups: either 30 s, 120 s, or 240 s. The pigeons in all three groups were trained for 10 50-min sessions. Generalization testing was conducted in extinction with different wavelengths of light. Absolute and relative generalization gradients were similar in shape for the three groups. Experiment 2 was a systematic replication of Experiment 1 using line orientation as the stimulus dimension and a mean VI value of either 30 s or 240 s. Again, gradients of generalization were similar for the two groups. In Experiment 3 pigeons pecked a key that was illuminated with a 501-nm light and obtained food reinforcers according to either a VI 30-s or a 240-s schedule. Training continued until response rates stabilized (> 30 sessions). For subjects trained with the 30-s schedule, generalization gradients were virtually identical regardless of whether training was for 10 sessions (Experiment 1) or until response rates stabilized. For subjects trained with the VI 240-s schedule, absolute generalization gradients for subjects trained to stability were displaced upward relative to gradients for subjects trained for only 10 sessions (Experiment 1), and relative generalization gradients were slightly flatter. These results indicate that the shape of a generalization gradient does not necessarily depend on the rate of reinforcement during 10-session single-stimulus training but that the effects of prolonged training on stimulus generalization may be schedule dependent.     

The structure of pigeon multiple-class same-different learning

Three experiments examined the structure of the decision framework used by pigeons in learning a multiple-class same-different task. Using a same-different choice task requiring the discrimination of odd-item different displays (one or more of the display's component elements differed) from same displays (all display components identical), pigeons were concurrently trained with sets of four discriminable display types. In each experiment, the consistent group was tested such that the same and different displays of four display types were consistently mapped onto their choice alternatives. The inconsistent group received a conflicting mapping of the same and different displays and the choice alternatives that differed across the four display types but were consistent within a display type. Experiment 1 tested experienced pigeons, and Experiment 2 tested naive pigeons. In both experiments, the consistent group learned their discrimination faster and to a higher level of choice accuracy than did the inconsistent group, which performed poorly in general. Only in the consistent group was the discrimination transferred to novel stimuli, indicative of concept formation in that group. A third experiment documented that the different display classes were discriminable from one another. These results suggest that pigeons attempt to generate a single discriminative rule when learning this type of task, and that this general rule can cover a large variety of stimulus elements and organizations, consistent with previous evidence suggesting that pigeons may be capable of learning relatively unbounded relational same-different concepts.     

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.     

Functional lateralization, interhemispheric transfer and position bias in serial reversal learning in pigeons (Columba livia)

In the present study we investigated lateralization of color reversal learning in pigeons. After monocular acquisition of a simple color discrimination with either the left or right eye, birds were tested in a serial reversal procedure. While there was only a slight and non-significant difference in choice accuracy during original color discrimination, a stable superiority of birds using the right eye emerged in serial reversals. Both groups showed a characteristic ‘learning-to-learn’ effect, but right-eyed subjects improved faster and reached a lower asymptotic error rate. Subsequent testing for interocular transfer demonstrated a difference between pre- and post-shift choice accuracy in pigeons switching from right to left eye but not vice versa. This can be accounted for by differences in maximum performance using either the left or right eye along with an equally efficient but incomplete interocular transfer in both directions. Detailed analysis of the birds’ response patterns during serial reversals revealed a preference for the right of two response keys in both groups. This bias was most pronounced at the beginning of a session. It decreased within sessions, but became more pronounced in late reversals, thus indicating a successful strategy for mastering the serial reversal task. Interocular transfer of response patterns revealed an unexpected asymmetry. Birds switching from right to left eye continued to prefer the right side, whereas pigeons shifting from left to right eye were now biased towards the left side. The results suggest that lateralized performance during reversal learning in pigeons rests on a complex interplay of learning about individual stimuli, stimulus dimensions, and lateralized response strategies. Received: 4 June 1999 / Accepted after revision: 18 August 1999     

Context effects in a temporal discrimination task" further tests of the Scalar Expectancy Theory and Learning-to-Time models

Pigeons were trained on two temporal bisection tasks, which alternated every two sessions. In the first task, they learned to choose a red key after a 1-s signal and a green key after a 4-s signal; in the second task, they learned to choose a blue key after a 4-s signal and a yellow key after a 16-s signal. Then the pigeons were exposed to a series of test trials in order to contrast two timing models, Learning-to-Time (LeT) and Scalar Expectancy Theory (SET). The models made substantially different predictions particularly for the test trials in which the sample duration ranged from 1 s to 16 s and the choice keys were Green and Blue, the keys associated with the same 4-s samples: LeT predicted that preference for Green should increase with sample duration, a context effect, but SET predicted that preference for Green should not vary with sample duration. The results were consistent with LeT. The present study adds to the literature the finding that the context effect occurs even when the two basic discriminations are never combined in the same session.     

Effects of stimulus similarity in discrimination training upon wavelength generalization in pigeons.

Thirty pigeons were given variable interval training to peck a 555-nm. light and then were tested for wavelength generalization. The subjects were later assigned to 1 of 3 groups, matched for both relative generalization slope and response rate. One group then received successive discrimination training between the 555-nm. stimulus (S+) and a vertical white line on a 555-nm. background (S minus); another group experienced the same S+ but a vertical white line on a black background as S minus. A third group received a comparable amount of single stimulus training with the 555-nm. value. On a second wavelength generalization test, the first group yielded greater sharpening of generalization than the second group, whereas the third group showed no change from Test 1. These results indicate that the sharpening of generalization gradients by discrimination training is directly related to the similarity of the discrimination training stimuli.     

Contrast and stimulus generalization following prolonged discrimination training

Different groups of pigeons received discrimination training in which the reinforcement-associated and extinction-associated stimuli were respectively either (a) a line tilt vs a blank key, (b) a blank key vs a line tilt, or (c) two different line tilts. The high response rates that developed to the positive stimulus in all groups during discrimination learning were maintained over 64 sessions of training. After these sessions, all subjects were tested for stimulus generalization along the line-tilt dimension. Gradients of relative (per cent) generalization around the stimulus associated with reinforcement (so-called excitatory gradients) and around the stimulus associated with extinction (so-called inhibitory gradients) were as steep as they typically are after much briefer training periods. These results do not support several of Terrace's predictions on the basis of the hypothesis that emotional responses develop to the stimulus associated with extinction during discrimination learning with errors, but eventually dissipate after extended training.     

Pigeons can discriminate “good” and “bad” paintings by children

Humans have the unique ability to create art, but non-human animals may be able to discriminate “good” art from “bad” art. In this study, I investigated whether pigeons could be trained to discriminate between paintings that had been judged by humans as either “bad” or “good”. To do this, adult human observers first classified several children’s paintings as either “good” (beautiful) or “bad” (ugly). Using operant conditioning procedures, pigeons were then reinforced for pecking at “good” paintings. After the pigeons learned the discrimination task, they were presented with novel pictures of both “good” and “bad” children’s paintings to test whether they had successfully learned to discriminate between these two stimulus categories. The results showed that pigeons could discriminate novel “good” and “bad” paintings. Then, to determine which cues the subjects used for the discrimination, I conducted tests of the stimuli when the paintings were of reduced size or grayscale. In addition, I tested their ability to discriminate when the painting stimuli were mosaic and partial occluded. The pigeons maintained discrimination performance when the paintings were reduced in size. However, discrimination performance decreased when stimuli were presented as grayscale images or when a mosaic effect was applied to the original stimuli in order to disrupt spatial frequency. Thus, the pigeons used both color and pattern cues for their discrimination. The partial occlusion did not disrupt the discriminative behavior suggesting that the pigeons did not attend to particular parts, namely upper, lower, left or right half, of the paintings. These results suggest that the pigeons are capable of learning the concept of a stimulus class that humans name “good” pictures. The second experiment showed that pigeons learned to discriminate watercolor paintings from pastel paintings. The subjects showed generalization to novel paintings. Then, as the first experiment, size reduction test, grayscale test, mosaic processing test and partial occlusion test were carried out. The results suggest that the pigeons used both color and pattern cues for the discrimination and show that non-human animals, such as pigeons, can be trained to discriminate abstract visual stimuli, such as pictures and may also have the ability to learn the concept of “beauty” as defined by humans.     

Effects of error and errorless discrimination acquisition on reversal learning

The effectiveness of trial-and-error, graded-choice, and verbal-instruction procedures on the acquisition and maintenance of a two-choice simultaneous color discrimination in an intradimensional double-reversal learning situation was studied using 18 first-grade children. After acquiring a red-green discrimination during one 70-trial session, the discriminative roles of the stimuli were reversed for 30 trials, followed by a second reversal for 30 trials. Children in the graded-choice and verbal-instruction groups acquired and maintained the discriminations with fewer errors than children who learned by trial and error. The importance of the results in terms of two-stage discrimination learning theories is pointed out and similarities between errorless learning and overtraining are discussed.     

Generalization gradients of inhibition after different amounts of training

Five groups of pigeons received seven sessions of variable-interval reinforcement for pecking a blank white key, followed by either 1, 2, 4, 8, or 16 sessions of training on a successive discrimination in which the positive stimulus was the blank white key and the negative stimulus was a black vertical line on the white key. After training, a generalization test was administered along the line-tilt continuum. Relative gradients of inhibition became steeper with increased amounts of training, and reliably nonhorizontal absolute gradients were obtained only from groups of subjects with at least four days of training. Therefore, inhibitory stimulus control improves with added training. Several problems with the concept of “inhibition” are examined and some implications of the results for theoretical analyses of operant discrimination learning are discussed.     

Choice biases in delayed matching-to-sample duration with pigeons: Manipulations of ITI and delayillumination

Two experiments investigated the effects of similarity between intertrial interval (ITI) and delay illumation on the choose-short effect. Different groups of pigeons learned to match 'short' (2 s) and 'long' (6 or 8 s) food samples to green and red test stimuli in a matchingto-sample procedure with a 5-s training delay.Subsequent 10- and 20-s delay tests revealed choose-short effects if the ITI and delay were both illuminated (i.e., group ON-ON), if the ITI and delay were both dark (i.e., group OFF-OFF), and if the ITI was illuminated and the delay was dark (i.e., group ON-OFF). In addition, either a choose-short effect or a chooselong effect was observed if the ITI was dark and the delay was illuminated (i.e., group OFFON). Results are incompatible with the confusion/instructionalfailure viewof the choose-short effect.