Generalization gradients in human predictive learning: Effects of discrimination training and within-subjects testing

Generalization gradients have been investigated widely in animal conditioning experiments, but much less so in human predictive learning tasks. Here, we apply the experimental design of a recent study on conditioned fear generalization in humans (Lissek et al., 2008) to a predictive learning task, and examine the effects of a number of relevant procedural parameters drawn from the generalization literature in animal conditioning. Experiment 1 shows that prior discrimination learning and steady-state testing procedures sharpen the gradient; Experiment 2 shows that within-subjects testing of the range of generalization stimuli also sharpens the gradient. In addition, Experiment 2 shows that, in case of very flat initial generalization, an orderly gradient can reveal itself through differential rates of extinction learning. Finally, Experiment 2 also evidenced an orderly gradient of generalization-of-extinction. These results suggest that discrimination processes have an important effect on the generalization of predictive learning in humans, and highlight behavioral analogies between animal conditioning and human predictive learning.     

Different parameters support generalization and discrimination learning in Drosophila at the flight simulator

We have used a genetically tractable model system, the fruit fly Drosophila melanogaster to study the interdependence between sensory processing and associative processing on learning performance. We investigated the influence of variations in the physical and predictive properties of color stimuli in several different operant-conditioning procedures on the subsequent learning performance. These procedures included context and stimulus generalization as well as color, compound, and conditional discrimination (colors and patterns). A surprisingly complex dependence of the learning performance on the colors' physical and predictive properties emerged, which was clarified by taking into account the fly-subjective perception of the color stimuli. Based on estimates of the stimuli's color and brightness values, we propose that the different tasks are supported by different parameters of the color stimuli; generalization occurs only if the chromaticity is sufficiently similar, whereas discrimination learning relies on brightness differences.     

Competition between antecedent and between subsequent stimuli in causal judgments

In the analysis of stimulus competition in causal judgment, 4 variables have been frequently confounded with respect to the conditions necessary for stimuli to compete: causal status of the competing stimuli (causes vs. effects), temporal order of the competing stimuli (antecedent vs. subsequent) relative to the noncompeting stimulus, directionality of training (predictive vs. diagnostic), and directionality of testing (predictive vs. diagnostic). In a factorial study using an overshadowing preparation, the authors isolated the role of each of these variables and their interactions. The results indicate that competition may be obtained in all conditions. Although some of the results are compatible with various theories of learning, the observation of stimulus competition in all conditions calls for a less restrictive reformulation of current learning theories that allows similar processing of antecedent and subsequent events, as well as of causes and effects.     

Judging interevent relations: From cause to effect and from effect to cause

Stimulus competition was studied in college students' correlational judgments in a medical decision-making setting. In accord with prior findings, subjects making cause-to-effect (predictive) judgments discounted a stimulus event that was moderately correlated with a target event when rival stimuli were more highly correlated with the effect. However, subjects making effect-to-cause (diagnostic) judgments were not at all disposed to discount a stimulus event which was moderately correlated with a target event when rival stimuli were more highly correlated with the cause. The theoretical implications of these results are considered in connection with associative and mentalistic models of causal attribution.     

Event Timing and age deficits in higher-order sequence learning

Recent studies have reported age deficits in learning sequences that contain subtle sequential regularities (e.g., Curran (1997) Psychological Research, 60(1-2), 24; D. V. Howard et al. (2004) Psychology and Aging, 19(1), 79; Howard, J. H. Jr, & Howard, D. V. (1997). Psychology and Aging, 12(4), 634). This finding is of potential theoretical interest, but the contribution of sequence event timing to this deficit has not been investigated. This study used an alternating serial reaction time task to examine implicit sequence learning in young adults when event timing mimicked that experienced by older adults in previous research. We varied the response-to-stimulus interval directly in Experiment 1 and indirectly by degrading the stimuli to influence response time in Experiment 2. Results indicate that these "aged" young adults learned the higher-order sequence structure implicitly, but they learned less than young controls and more than old adults on some measures of implicit learning in both experiments. In addition, these two different experimental manipulations produced distinct patterns of deficits despite having nearly identical effects on event sequence timing. These findings suggest that event timing alone cannot explain the age deficits observed in high-order implicit sequence learning.     

Whatever Gave You That Idea? False Memories Following Equivalence Training: A Behavioral Account of the Misinformation Effect

The misinformation effect is a term used in the cognitive psychological literature to describe both experimental and real-world instances in which misleading information is incorporated into an account of an historical event. In many real-world situations, it is not possible to identify a distinct source of misinformation, and it appears that the witness may have inferred a false memory by integrating information from a variety of sources. In a stimulus equivalence task, a small number of trained relations between some members of a class of arbitrary stimuli result in a large number of untrained, or emergent relations, between all members of the class. Misleading information was introduced into a simple memory task between a learning phase and a recognition test by means of a match-to-sample stimulus equivalence task that included both stimuli from the original learning task and novel stimuli. At the recognition test, participants given equivalence training were more likely to misidentify patterns than those who were not given such training. The misinformation effect was distinct from the effects of prior stimulus exposure, or partial stimulus control. In summary, stimulus equivalence processes may underlie some real-world manifestations of the misinformation effect.     

真实线和想象线的朝向辨别学习及其相互迁移

以往的朝向辨别学习普遍采用真实刺激作为材料,其特异性与迁移性是研究争论的焦点之一。该研究以实线和想象线为刺激,采用朝向辨别任务训练被试,系统考察两者朝向辨别学习的特点和相互关系。结果发现,训练实线和想象线任务的两组被试行为都有显著提高,并且这两种学习效果能部分地彼此迁移,且都能迁移到不同位置。该研究结果提示,不论是想象线的学习还是实线的学习都涉及高级脑区的参与,并且两者的学习有部分共同的脑机制。     

Disrupted attentional learning in high schizotypy: Evidence of aberrant salience

The relationship between learned variations in attention and schizotypy was examined in two experiments. In Experiment 1, participants low on a negative subscale of schizotypy exhibited an explicit bias in overt attention towards stimuli that were established as predictive of a trial outcome, relative to stimuli that were irrelevant. The same participants also showed a bias in learning about these stimuli when they presented in a novel context. Neither of these effects was observed in participants high in schizotypy. In Experiment 2, participants low on the negative subscale of schizotypy exhibited faster reaction times towards a target that was cued by a stimulus that had a history of predictive validity relative to a stimulus that had a history of irrelevance. Again, this effect was not present in participants high in schizotypy. These results imply a disruption in the normal allocation of attention to cues that have predictive significance in schizotypy.     

The relation between memory and expectancy as revealed by percentage and sequence of reward investigations

There is growing agreement that to explain instrumental learning properly, one should emphasize memory as well as expectancy. I call this approachmemory-expectancy theory. Amsel’s (1992) frustration theory is one variety of memory-expectancy theory. Capaldi’s (1994) sequential theory is another. In this report, I examine in considerable detail the effects of percentage and sequence of reward on extinction following different levels of acquisition training. These extinction findings, taken together with certain serial learning acquisition findings, seem to support a novel version of memory-expectancy theory, one that in some respects is similar to and in some respects is different from that suggested by Amsel. First, on the basis of this analysis, we may reject two ideas: that animals remember only the prior reward event and that animals anticipate only the reward event contingent upon the current response. Second, the analysis supports three salient propositions of the present memory-expectancy approach. Memories of reward events may serve as conditioned stimuli for expectancies of reward events. On any current trial, the animal may remember each of the reward events associated with one or more prior trials. On any current trial, the animal may anticipate not only the current reward event, but also reward events contingent upon subsequent trials. Essentially, according to this model, the stimuli that elicit expectancies, as well as the expectancies themselves, may change progressively over a series of learning trials.     

Comparing learned predictiveness effects within and across compound discriminations

In four human learning experiments, we examined the extent to which learned predictiveness depends upon direct comparison between relatively good and poor predictors. Participants initially solved (a) linear compound discriminations in which one or both of the stimuli in each compound were predictive of the correct outcome, (b) biconditional discriminations where only the configurations of the stimuli were predictive of the correct outcome, or (c) pseudodiscriminations in which no stimulus features were predictive. In each experiment, subsequent learning and test stages were used to assay changes in the associability of each stimulus brought about by its role in the initial discriminations. Although learned predictiveness effects were observed in all experiments (i.e., previously predictive cues were more readily associated with a new outcome than previously nonpredictive cues), the same changes in associability were observed regardless of whether the stimulus was initially learned about in the presence of an equally predictive, more predictive, or less predictive stimulus. The results suggest that learned associability is not controlled by competitive allocation of attention, but rather by the absolute predictiveness of each individual cue.     

The role of associative history in models of associative learning: A selective review and a hybrid model

Associative learning theories strive to capture the processes underlying and driving the change in strength of the associations between representations of stimuli that develop as a result of experience of the predictive relationships between those stimuli. Historically, formal models of associative learning have focused on two potential factors underlying associative change, namely processing of the conditioned stimulus (in terms of changes in associability) and processing of the unconditioned stimulus (in terms of changes in error). This review constitutes an analysis of the proper role of these two factors, specifically with regard to the way in which they are influenced by associative history (the prior training undergone by cues). A novel “hybrid” model of associative learning is proposed and is shown to provide a more satisfactory account of the effects of associative history on subsequent learning than any previous single-process theory.     

Aging, working memory, and discrimination learning

Older adults easily learn probabilistic relationships between cues and outcomes when the predictive event is the occurrence of a cue, but have greater difficulty when the predictive event is the nonoccurrence of a cue (Mutter & Pliske, 1996; Mutter & Plumlee, 2004; Mutter & Williams, 2004). This study explored whether this age-related deficit occurs in a simpler learning context and whether it might be related to working memory (WM) decline. We gave younger and older adults simultaneous discrimination tasks that allowed us to compare their ability to learn deterministic relationships when either the occurrence (feature positive; FP) or the nonoccurrence (feature negative; FN) of a distinctive feature predicted reinforcement. We also included a group of younger adults who performed the discrimination tasks under a concurrent WM load. Both age and WM load had a detrimental effect on initial FP and FN discrimination; however, these effects persisted only in FN discrimination after additional learning experience. Learning predictive relationships requires inductive reasoning processes that apparently do not operate as efficiently in individuals with reduced WM capacity. The impact of WM decline may ultimately be greater for negative cue-outcome relationships because learning these relationships requires more difficult inductive reasoning processes, which place greater demands on WM.     

An ecological approach to learning

This paper introduces a special journal issue devoted to the study of learning in ecological and developmental contexts. In the past, the dominant approach to the study of learning has depended upon the choice of arbitrary problems, stimuli, and responses to guarantee the generality of principles discovered in the laboratory. Recently, though, there has been considerable interest in integrating an ecological (functional) account of learning with the strengths of the classic laboratory approach. One form of integration is to use the technology of the laboratory to look for the operation of general learning laws in ecological problems. The approach favored here is to treat learning as a biological phenomenon by first placing it within a functional system of behavior, and then by analyzing where and how learning modifies the operation of that system. Because the results of such analyses are defined with respect to functioning systems rather than procedural paradigms, the ecological approach readily makes contact with issues in evolution, development, and physiology, an ability not completely shared by the classic general-process approach. Concerns about laboratory versus field, function versus mechanism, generality of results, and adaptive “storytelling” can be resolved or further clarified by the present approach. The papers in this issue represent a cross-section of research stemming from an ecological approach to learning, and provide specific analyses of how learning modifies and is expressed in functional systems of behavior.     

Aging,working memory,and discrimination learning

Older adults easily learn probabilistic relationships between cues and outcomes when the predictive event is the occurrence of a cue, but have greater difficulty when the predictive event is the nonoccurrence of a cue (Mutter & Pliske, 1996; Mutter & Plumlee, 2004; Mutter & Williams, 2004). This study explored whether this age-related deficit occurs in a simpler learning context and whether it might be related to working memory (WM) decline. We gave younger and older adults simultaneous discrimination tasks that allowed us to compare their ability to learn deterministic relationships when either the occurrence (feature positive; FP) or the nonoccurrence (feature negative; FN) of a distinctive feature predicted reinforcement. We also included a group of younger adults who performed the discrimination tasks under a concurrent WM load. Both age and WM load had a detrimental effect on initial FP and FN discrimination; however, these effects persisted only in FN discrimination after additional learning experience. Learning predictive relationships requires inductive reasoning processes that apparently do not operate as efficiently in individuals with reduced WM capacity. The impact of WM decline may ultimately be greater for negative cue–outcome relationships because learning these relationships requires more difficult inductive reasoning processes, which place greater demands on WM.     

Opioid receptors mediate direct predictive fear learning: evidence from one-trial blocking

Pavlovian fear learning depends on predictive error, so that fear learning occurs when the actual outcome of a conditioning trial exceeds the expected outcome. Previous research has shown that opioid receptors, including μ-opioid receptors in the ventrolateral quadrant of the midbrain periaqueductal gray (vlPAG), mediate such predictive fear learning. Four experiments reported here used a within-subject one-trial blocking design to study whether opioid receptors mediate a direct or indirect action of predictive error on Pavlovian association formation. In Stage I, rats were trained to fear conditioned stimulus (CS) A by pairing it with shock. In Stage II, CSA and CSB were co-presented once and co-terminated with shock. Two novel stimuli, CSC and CSD, were also co-presented once and co-terminated with shock in Stage II. The results showed one-trial blocking of fear learning (Experiment 1) as well as one-trial unblocking of fear learning when Stage II training employed a higher intensity footshock than was used in Stage I (Experiment 2). Systemic administrations of the opioid receptor antagonist naloxone (Experiment 3) or intra-vlPAG administrations of the selective μ-opioid receptor antagonist CTAP (Experiment 4) prior to Stage II training prevented one-trial blocking. These results show that opioid receptors mediate the direct actions of predictive error on Pavlovian association formation.     

Backward inhibitory learning in honeybees: a behavioral analysis of reinforcement processing

One class of theoretical accounts of associative learning suggests that reinforcers are processed according to learning rules that minimize the predictive error between the expected strength of future reinforcement and its actual strength. The omission of reinforcement in a situation where it is expected leads to inhibitory learning of stimuli indicative for such a violation of the prediction. There are, however, results indicating that inhibitory learning can also be induced by other mechanisms. Here, we present data from olfactory reward conditioning in honeybees that show that (1) one- and multiple-trial backward conditioning results in conditioned inhibition (CI); (2) the inhibition is maximal for a 15-sec interval between US and CS; (3) there is a nonmonotonic dependency on the degree of CI from the US-CS interval during backward pairing; and (4) the prior association of context stimuli with reinforcement is not necessary for the development of CI. These results cannot be explained by models that only minimize a prediction error. Rather, they are consistent with models of associative learning that, in addition, assume that learning depends on the temporal overlap of a CS with two processes, a fast excitatory and a slow inhibitory one, both evoked by a reinforcer. The fmdings from this behavioral analysis of reinforcement processing are compared with the known properties of an individual, identified neuron involved in reinforcement processing in the bee brain, to further understand the mechanisms underlying predictive reward learning.     

When obvious covariations are not even learned implicitly

The present experiments examine the impact of covariation type and attention on implicit covariation learning. Experiments 1 and 2 compare learning of stimulus–stimulus (s–s) and stimulus–response (s–r) covariations. Although stimuli and responses were predicted by a distinct feature of the display, implicit learning neither of s–s nor of s–r covariations was observed. Experiments 3–5 explore the impact of attention on implicit learning of an s–r covariation. Distinct features either of the targets or of an incidental but centrally presented object predicted the responses. Implicit covariation learning was restricted to predictors that were part of to-be-attended targets. Finally, Experiment 6 shows implicit learning of a partial s–r covariation in which only one of two responses is predicted by target features. It is argued that implicit learning is based on the formation of associations between simultaneously activated distinct representations. Because only attended stimuli seem to reach a sufficient level of mental distinctiveness, attending the predictive information most likely is an indispensable prerequisite for implicit covariation learning to occur.     

Protection from premature habituation requires functional mushroom bodies in Drosophila

Diminished responses to stimuli defined as habituation can serve as a gating mechanism for repetitive environmental cues with little predictive value and importance. We demonstrate that wild-type animals diminish their responses to electric shock stimuli with properties characteristic of short- and long-term habituation. We used spatially restricted abrogation of neurotransmission to identify brain areas involved in this behavioral response. We find that the mushroom bodies and, in particular, the α/β lobes appear to guard against habituating prematurely to repetitive electric shock stimuli. In addition to protection from premature habituation, the mushroom bodies are essential for spontaneous recovery and dishabituation. These results reveal a novel modulatory role of the mushroom bodies on responses to repetitive stimuli in agreement with and complementary to their established roles in olfactory learning and memory.