Feature- and rule-based generalization in human associative learning.

Two experiments examined the contributions of feature- and rule-based knowledge in a human associative learning task. Participants were presented with concurrent negative (A?→?O, B?→?O, AB?→?no O) and positive (C?→?no O, D?→?no O, CD?→?O) patterning problems in which certain combinations of foods were associated with an allergy outcome (O). In the test stage, some participants showed normal feature-based generalization to novel trial types, whereas other participants transferred the patterning rule (i.e., a compound and its elements signal opposite outcomes). Mastery of the discrimination presented in the training phase was strongly linked to rule-based generalization. The results suggest that models of human associative learning need to incorporate mechanisms for rule-based as well as for feature-based generalization. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Solving XOR.

[Correction Notice: An erratum for this article was reported in Vol 35(1) of Journal of Experimental Psychology: Animal Behavior Processes (see record 2009-00257-012). Figure 2 in the article was printed incorrectly due to an editing error. Below is the correct version of Figure 2.] Three experiments examined the way in which exclusive-or (XOR) problems are solved by rats. All rats first received food-rewarded positive and negative patterning problems with two stimulus sets: either A+, B+, AB- and C-, D-, CD+, or A-, B-, AB + and C+, D +, and CD-. Subsequently, rats received revaluation trials in which A was paired with shock and C was not, prior to generalization test trials with B, D, AB, and CD (Experiments 1 & 2); or received A→shock trials prior to tests with B and CD (Experiment 3). There was greater generalized fear to B than to either D (Experiments 1 & 2) or AB (Experiment 2) and CD (Experiments 2 & 3). These results are inconsistent with configural, connectionist models, but are consistent with an alternative connectionist model that can represent the logical structure of XOR problems. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Feature-positive discrimination learning.

In Experiments 1, 2, and 3, pigeons were trained with an ABC+ BCo discrimination, in which three stimuli, A, B, and C, were presented together and paired with food, and the compound BC was followed by nothing; they were also trained with a DEF+ Eo Fo discrimination in which stimuli E and F were presented separately and followed by nothing, whereas the compound DEF was paired with food. On completion of discrimination training, test trials with the feature A consistently revealed a higher rate of responding than with D. In Experiment 4, reinforced presentations of D were intermixed with the DEF+ Eo Fo discrimination. Test trials revealed that E enhanced responding when it was paired with F, but it had the opposite effect when paired with D. The results are seen as being more consistent with a configural than an elemental model of conditioning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Configural associations in discrimination learning.

In 4 experiments, Sprague-Dawley rats and homing pigeons received training with an A+ AB0 BC+ discrimination, in which food (+) accompanied trials with A and BC. Food was not presented (0) on trials with the compound AB. Subsequent test trials revealed that responding during C by itself, or the compound ABC, was slower than during either A or BC. Responding during the ABC compound was also found to be slower after training with the A+ AB0 BC+ than an A0 AB+ BC+ discrimination. We argue that these findings demonstrate the importance of configural associations in discrimination learning. Two accounts for the way in which these associations exert their influence are considered. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Acquired relational equivalence: Implications for the nature of associative structures.

In 2 experiments, rats received discrimination training in which separate presentations of A and B signaled a common pair of relationships or associations (X?→?food and Y?→?no food), whereas presentations of C and D signaled a different pair of relationships (X?→?no food and Y?→?food). To assess the nature of the associative structures acquired during this training, rats then received 2 types of revaluation procedure: In Experiment 1, A was paired with shock and C was not. In Experiment 2, the relationships that A and B had previously signaled (X?→?food and Y?→?no food) were paired with shock, whereas those that C and D had signaled (Y?→?food and X?→?no food) were not. After both types of revaluation treatment, rats showed greater generalized conditioned suppression in the presence of B than D. These results indicate that A, B, C, and D come to evoke memories of the relationships or associations that they have signaled. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Trial spacing and trial distribution effects in Pavlovian conditioning: Contributions of a comparator mechanism.

Two experiments, involving a total of 144 rats, investigated a potential basis for trial spacing and trial distribution effects. In Exp 1, a CS (e.g., CS A) was trained with either massed (e.g., A?→?A?→?A) or spaced (e.g., A A A) trials. When trials were massed, brief exposure to the training context (a condition typical of massed training) impaired responding, whereas more extensive exposure to the context during or after training reduced this apparent massed trials deficit. In Exp 2, different CSs were trained in either a massed (e.g., A?→?A?→?A?→?B?→?B?→?B?→?C?→?C?→?C) or a distributed (e.g., A?→?B?→?C?→?A?→?B?→?C, etc.) manner. Trials massed in this sense resulted in impaired responding to the CS, and this impairment was attenuated by posttraining extinction of the context cues. Thus, trial distribution and apparent trial spacing effects were, at least in part, reversible deficits in performance rather than failures of learning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Intertrial interval as a contextual stimulus: Further analysis of a novel asymmetry in temporal discrimination learning.

Four experiments investigated discrimination learning when the duration of the intertrial interval (ITI) signaled whether or not the next conditional stimulus (CS) would be paired with food pellets. Rats received presentations of a 10-s CS separated half the time by long ITIs and half the time by short ITIs. When the long ITI signaled that the CS would be reinforced and the short interval signaled that it would not be (Long+/Short?), rats learned the discrimination readily. However, when the short ITI signaled that the CS would be reinforced and the long interval signaled that it would not (Short+/Long?), discrimination learning was much slower. Experiment 1 compared Long+/Short? and Short+/Long? discrimination learning with 16-min/4-min or 4-min/1-min ITI combinations. Experiment 2 found no evidence that Short+/Long? learning is inferior because the temporal cue corresponding to the short interval is ambiguous. Experiment 3 found no evidence that Short+/Long? learning is poor because the end of a long ITI signals a substantial reduction in delay to the next reinforcer. Long+/Short? learning may be faster than Short+/Long?because elapsing time involves exposure to a sequence of hypothetical stimulus elements (e.g., A then B), and feature-positive discriminations (AB+/A?) are learned quicker than feature-negative discriminations (A+/AB?). Consistent with this view, Experiment 4 found a robust feature-positive effect when sequentially presented CSs played the role of elements A and B. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effects of discrimination training on responding during a compound conditioned stimulus.

Four experiments examined responding in the presence of a triple-element compound ABC after discrimination training in which 2 compounds, AB and BC, signaled the delivery of food and 1 element alone, B, signaled the absence of food. In Exps 1 and 2, using rats, responding during ABC was more vigorous than in a control group for which A and C but not B had been individually paired with food. This finding was replicated in Exp 3, which used pigeons, and in Exp 4, where all 3 stimuli of the control condition were individually paired with food. The results are more consistent with a configural than with an elemental theory of learning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Inhibition and facilitation by B over C after A+, AB–, and ABC+ training with multimodality stimulus combinations.

Previous research with keylight conditioned stimuli has revealed that pigeons failed to show inhibition by Stimulus B over Stimulus C in BC versus C testing after A+, AB–, ABC+ training where Stimulus A and Stimulus Compound ABC had signaled food, and Stimulus Compound AB had signaled no food. Indeed, B slightly facilitated responding to C on the BC trials. The present research addressed the same issue with multimodality stimulus arrangements in autoshaping with pigeons, conditioned suppression with rats, and instrumental discrimination learning with rats. Stimulus B facilitated responding to C if A and B were of the same modality and C was of a different modality. However, B inhibited responding to C if A and C were of the same modality and B was of a different modality, or if B and C were of the same modality and A was of a different modality. These results are correctly predictable by Pearce's configural model with a minor modification. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

"Solving XOR": Correction to Grand and Honey (2008).

Reports an error in "Solving XOR" by C. Grand and R. C. Honey (Journal of Experimental Psychology: Animal Behavior Processes, 2008[Oct], Vol 34[4], 486-493). Figure 2 in the article was printed incorrectly due to an editing error. The correct version of Figure 2 is provided in the erratum. (The following abstract of the original article appeared in record 2008-14849-005.) Three experiments examined the way in which exclusive-or (XOR) problems are solved by rats. All rats first received food-rewarded positive and negative patterning problems with two stimulus sets: either A+, B+, AB- and C-, D-, CD+, or A-, B-, AB + and C+, D +, and CD-. Subsequently, rats received revaluation trials in which A was paired with shock and C was not, prior to generalization test trials with B, D, AB, and CD (Experiments 1 & 2); or received A→shock trials prior to tests with B and CD (Experiment 3). There was greater generalized fear to B than to either D (Experiments 1 & 2) or AB (Experiment 2) and CD (Experiments 2 & 3). These results are inconsistent with configural, connectionist models, but are consistent with an alternative connectionist model that can represent the logical structure of XOR problems. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning about associations: Evidence for a hierarchical account of occasion setting.

In 2 experiments rats were trained on a switching discrimination, with 4 occasion setters, A, B, C, and D and 2 target stimuli, x and y. When signaled either by A or by B, x was reinforced with food and y was not, whereas when signaled either by C or by D these reinforcement relations were reversed (i.e., A: → x+, A: y → ?, B: x → +, B: y → ?, C: x → ?, C: y → +, D: x → ?, D: y → +). In a subsequent Stage A was paired with shock, and then the degree to which food–reinforced (Experiment 1a) and nonreinforced (Experiment 1b) presentations of x and y were capable of eliciting fear was assessed. Those conditioned stimulus (CS)/unconditioned stimulus (US) relations that had been operative in the presence of the fear-eliciting occasion setter A (i.e., x → +, y → ?) elicited more fear than the alternative CS/US combinations (i.e., x → ?, y → +). The implications of these findings are discussed with reference to theories of occasion setting and of configural learning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Acquired relational equivalence between contexts and features.

In 2 experiments, rats received a biconditional discrimination wherein separate presentations of A and B signaled 1 pair of associations (X?→?food and Y?→?no food) and presentations of C and D signaled a different pair of associations (X?→?no food and Y?→?food). In Exp 1, A, B, C, and D were diffuse contextual stimuli in which the associations were embedded. In Exp 2, A and C were contextual stimuli whereas B and D were features that immediately preceded the associations. To assess the associative structures acquired during training, all rats then received a revaluation procedure in which A was paired with shock and C was not. In both experiments, greater generalized suppression of behavior was observed in the presence of B than in the presence of D. These results indicate that contextual stimuli share with features the capacity to evoke the associations that they have signaled. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Positive transfer from sequence production to sequence discrimination in a nonverbal organism.

Examined the relation between a pigeon's ability to produce and to recognize a particular arbitrary sequence of arbitrary elements, in 3 experiments involving a total of 100 naive male White Carneaux pigeons. In Exp I, 16 Ss were trained to discriminate sequences composed of 3 successively presented colors (A, B, and C). A yes response was rewarded for identifying A?→?B?→?C sequences; a no response for identifying non-A?→?B?→?C sequences. Exp II showed that the acquisition of this discrimination was facilitated by prior training on the production of the sequence A?→?B?→?C. Exp III showed that the facilitation of the discrimination of A?→?B?→?C sequences observed in Exp II could be attributed specifically to knowledge of the order of the elements A, B, and C that was acquired during the production of the sequence A?→?B?→?C. The control conditions of Exp III were (a) a pseudo-production task in which the Ss pecked the colors A, B, and C as presented successively; (b) a nonsequential task that established discriminative performance with respect to A, B, and C; (c) pseudo-production training that was preceded by discrimination training; and (d) training to produce either the sequence C?→?B?→?A or B?→?A?→?C. (46 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Configural analysis of summation in autoshaping.

In 4 experiments, pigeons received autoshaping with various combinations of three stimuli, A, B, and C, before test trials in which responding during all three stimuli, ABC, was compared with that during a three-element control compound, DEF, which had been consistently paired with food. Pairing A, B, and C individually with food resulted in similar rates of responding during ABC and DEF (Experiments 1 and 2). Responding was faster, however, during ABC than during DEF after training in which food was signaled by the pairs of stimuli (AB, AC, and BC; Experiment 1). Responding was also faster during ABC than during DEF after training involving reinforced (+) and nonreinforced (°) trials of the form ABC+ A° BC°, followed by A+ BC+ (Experiment 2), or AB+ BC+ B° (Experiments 3 and 4). The results are consistent with those of a configural analysis of summation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Supernormal conditioning.

In Experiment 1, rats received an A+AX° discrimination in which food was presented after Stimulus A by itself but not after a simultaneous compound of A with Stimulus X. AX was then paired with food in a 2nd stage, followed by test trials with A alone. Responding on the test trials with A was more vigorous than during a control stimulus that had been consistently paired with food. The remaining experiments were of similar design to Experiment 1, except that the 2nd stage also contained conditioning trials with X. The results from the test trials were essentially the same as for Experiment 1. The high level of responding during the test trials with Stimulus A is regarded as evidence of supernormal conditioning. Overall, the results are more consistent with a configural than an elemental theory of conditioning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The influence of an irrelevant stimulus on two discriminations.

In 3 experiments, the effect of adding an irrelevant stimulus to a discrimination was examined. In Exp 1, a group of pigeons received autoshaping with an A+Bo discrimination in which 1 stimulus signaled food, A+, and a simultaneous compound of A with another stimulus, B, signaled the absence of food, ABo. A 2nd group received similiar training, except that a 3rd stimulus, C, was present in both types of trials, AC+BCo. The A+Bo discrimination was acquired more readily than the AC+BCo discrimination. Exps 2 and 3 used a negative-patterning design, A+Bo+. In both experiments, this problem was mastered more readily than when an irrelevant stimulus was used to create an AC+BCoC+ discrimination. The results fail to confirm predictions derived from elemental theories of conditioning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Prototype effects in categorization by pigeons.

In 4 experiments involving a total of 78 pigeons, Ss received a categorization task involving 6 simultaneous compounds in which the elements A, B, and C were more frequently paired with food than were the elements D, E, and F. Food was delivered after compounds ABF, AEC, and DBC but not after DEC, DBF, and AEF. Subsequent testing revealed a higher rate of responding during ABC than during any of the compounds that had signaled food and a lower rate of responding during DEF than during any of the compounds that had not signaled food. Exps 2, 3, and 4 further demonstrated that the rate of responding during test trials with ABC was faster than during a compound composed of 3 elements that had individually been paired with food. Results are more consistent with a configural than an elemental analysis of discrimination and categorization. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Inhibition of fear-potentiated startle can be detected after the offset of a feature trained in a serial feature-negative discrimination.

Using the fear-potentiated startle paradigm in rats, 4 experiments examined whether the inhibitory effect of a feature is evident after its offset following serial feature-negative discrimination training (A+ and X?→?A–). When startle probes were presented shortly after the offset of X on X?→?A test trials, the inhibitory properties of X were observed immediately after its offset. Furthermore, trace reinforcement of X (X?→?+), but not delay reinforcement (X+), disrupted the ability of X to inhibit fear-potentiated startle on X?→?A trials. Trace conditioning to X was also retarded after A+ and X?→?A– training. These results suggest that the inhibitory properties of the serially trained feature are present after its offset and raise the possibility that either temporal information regarding nonreinforcement or poststimulus attributes of X acquire inhibitory properties. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Encoding specific associative memory: Evidence from behavioral and neural manipulations.

Within-subjects procedures with rats assessed the associative structures acquired during conditioning trials in which the interval between the stimuli and food was either short or long (i.e., A–10 s→food and B–40 s→food). In Experiments 1 and 2, after these conditioning trials, A and B served as second-order reinforcers for 2 further stimuli (i.e., X→A and Y→B); whereas Experiment 3 used a sensory preconditioning procedure in which X→A and Y→B trials occurred before the conditioning trials, and rats were finally tested with X and Y. In each experiment, Y elicited greater responding at test than did X. This finding supports the contention that the long-lived trace of B (associated with food on B–40 s→food trials) is more similar to the memory of B that was associatively provoked by Y, than is the short-lived trace of A (associated with food on A–10 s→food trials) to the memory of A that was associatively provoked by X. These conclusions were reinforced by the effects of a neural manipulation that disrupted discrimination learning involving the short traces of stimuli but not the long traces of the same stimuli. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Analysis of the unique cue in configural discriminations.

Investigated the basis of configural discriminations, using an autoshaping procedure, in 4 experiments with 68 female Carneaux pigeons. The elements of both a negative patterning (A+, B+, AB–) and a conditional discrimination (AC+, BD+, AD–, BC–) were paired in a 2nd-order procedure with 2 new key lights, X and Y. Responding was then tested to X and Y presented in compound in each other and with A and B. The pattern of responding to compounds containing X and Y was like the pattern of responding to compounds containing their associates, A and B. This suggests that A and B can be replaced by their associates without disrupting responding to their compounds. Because X and Y are physically different from A and B, this in turn suggests that any unique cue controlling responding to their compounds does not depend on the physical presence of the component stimuli. Instead the unique stimulus appears to arise from the joint activation of memory representations. (13 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)