Varieties of perceptual learning

Although most studies of perceptual learning in human participants have concentrated on the changes in perception assumed to be occurring, studies of nonhuman animals necessarily measure discrimination learning and generalization and remain agnostic on the question of whether changes in behavior reflect changes in perception. On the other hand, animal studies do make it easier to draw a distinction between supervised and unsupervised learning. Differential reinforcement will surely teach animals to attend to some features of a stimulus array rather than to others. But it is an open question as to whether such changes in attention underlie the enhanced discrimination seen after unreinforced exposure to such an array. I argue that most instances of unsupervised perceptual learning observed in animals (and at least some in human animals) are better explained by appeal to well-established principles and phenomena of associative learning theory: excitatory and inhibitory associations between stimulus elements, latent inhibition, and habituation.     

Social learning about predators: a review and prospectus

In comparison with social learning about food, social learning about predators has received little attention. Yet such research is of potential interest to students of animal cognition and conservation biologists. I summarize evidence for social learning about predators by fish, birds, eutherian mammals, and marsupials. I consider the proposal that this phenomenon is a case of S-S classical conditioning and suggest that evolution may have modified some of the properties of learning to accommodate for the requirements of learning socially about danger. I discuss some between-species differences in the properties of socially acquired predator avoidance and suggest that learning may be faster and more robust in species in which alarm behavior reliably predicts high predatory threat. Finally, I highlight how studies of socially acquired predator avoidance can inform the design of prerelease antipredator training programs for endangered species.     

Social learning strategies

In most studies of social learning in animals, no attempt has been made to examine the nature of the strategy adopted by animals when they copy others. Researchers have expended considerable effort in exploring the psychological processes that underlie social learning and amassed extensive data banks recording purported social learning in the field, but the contexts under which animals copy others remain unexplored. Yet, theoretical models used to investigate the adaptive advantages of social learning lead to the conclusion that social learning cannot be indiscriminate and that individuals should adopt strategies that dictate the circumstances under which they copy others and from whom they learn. In this article, I discuss a number of possible strategies that are predicted by theoretical analyses, includingcopy when uncertain,copy the majority, andcopy if better, and consider the empirical evidence in support of each, drawing from both the animal and human social learning literature. Reliance on social learning strategies may be organized hierarchically, their being employed by animals when unlearned and asocially learned strategies prove ineffective but before animals take recourse in innovation.     

Human and animal perceptual learning: Some common and some unique features

A selective summary of the four contributions to this special issue of Learning & Behavior on perceptual learning is presented. Mackintosh and Hall propose an associative analysis of perceptual learning. It is argued that Tsushima and Watanabe’s psychophysical evidence and Fiser’s Bayesian-modeling work represent (in different ways) challenges to the associative approach. Some tentative suggestions are explored with regard to how animal learning theorists might meet these challenges. Finally, the role of awareness in perceptual learning is briefly examined.     

Perceptual learning in human and nonhuman animals: A search for common ground

Perceptual learning has been extensively studied in both human and nonhuman animals, but the two lines of research have, for the most part, developed independently, addressing seemingly rather different issues by rather different methods. It has been argued, however, that analysis of the disparate phenomena studied in experiments on perceptual learning reveals that in all the studies, the essential feature is that appropriate training allows behavior to come to be controlled by the unique features, rather than by the common features, of similar stimuli. It has further been argued that experiments with nonhuman animals have established the existence of a range of learning processes that allow this to occur, and that these processes have general relevance, applying to humans as well as to animals.     

Sketching,not representational competence,predicts improved science learning

Representational competence is a target of novel learning environments given the assumption that improved representational competence improves learning in science. There exists little evidence, however, that improving representational competence is positively correlated with learning outcomes across science disciplines. In this report, we argue that the previously reported weak relationships between representational competence and science learning outcomes have resulted from designs that do not explicitly analyze the discipline‐specific skills related to the representational competence construct. Here, we demonstrate through a detailed analysis of students' representation use that at least two demonstrated skills comprising representational competence (e.g., construction and selection) are not strongly related to improved conceptual understanding in the domain. We discuss the implications of these results for the design of future learning environments that aim to improve learning through improved representational competence.     

Flattening generalization gradients, context, and perceptual learning

The present research investigated the effects of physical context change and perceptual learning on generalization. In a video game, participants learned to suppress their mouse-clicking behavior in the presence of one stimulus (AX). Generalization was observed between the AX stimulus and another stimulus (BX) that was designed to be similar. When testing was conducted in a context different from that in which AX was used in training, responding to AX was attenuated, and responding to BX was enhanced. That is, the generalization gradient flattened. The latter effect was only evident in groups for which generalization had been reduced through a preexposure manipulation believed to produce perceptual learning. Experiment 2 demonstrated that the increase in generalization observed in the first experiment was due to the context change between the preexposure and test rather than to a change between the conditioning and test contexts. Implications for flattening generalization gradients and mechanisms of perceptual learning are discussed.     

On the nature of CS and US representations in Pavlovian learning

A significant problem in the study of Pavlovian conditioning is characterizing the nature of the representations of events that enter into learning. This issue has been explored extensively with regard to the question of what features of the unconditioned stimulus enter into learning, but considerably less work has been directed to the question of characterizing the nature of the conditioned stimulus. This article introduces a multilayered connectionist network approach to understanding how “perceptual” or “conceptual” representations of the conditioned stimulus might emerge from conditioning and participate in various learning phenomena. The model is applied to acquired equivalence/distinctiveness of cue effects, as well as a variety of conditional discrimination learning tasks (patterning, biconditional, ambiguous occasion setting, feature discriminations). In addition, studies that have examined what aspects of the unconditioned stimulus enter into learning are also reviewed. Ultimately, it is concluded that adopting a multilayered connectionist network perspective of Pavlovian learning provides us with a richer way in which to view basic learning processes, but a number of key theoretical problems remain to be solved, particularly as they relate to the integration of what we know about the nature of the representations of conditioned and unconditioned stimuli.     

Remembering places: Student reliance on place in timed essays

This is the story of a research journey that follows the trail of a novel evaluand—place. I examine place as mentioned by rising juniors in timed exams. Using a hybridized methodology—the qualitative approach of a hermeneutic dialectic process as described by Guba and Lincoln (1989), and the quantitative evidence of place mention—I query student writing to uncover motivations for reliance on place-based experience despite the absence of an explicit mention of place in the writing prompt. Finally, I consider implications for using generative prompts to elicit evidence of student learning.     

For a learnable mathematics in the digital cultures

I begin with some general remarks concerning the co-evolution of representational forms and mathematical meanings. I then discuss the changed roles of mathematics and novel representations that emerge from the ubiquity of computational models, and briefly consider the implications for learning mathematics. I contend that a central component of knowledge required in modern societies involves the development of a meta-epistemological stance – i.e. developing a sense of mechanism for the models that underpin social and professional discourses. I illustrate this point in relation to recent research in which I am investigating the mathematical epistemology of engineering practice. Finally, I map out one implication for the design of future mathematical learning environments with reference to some data from the `Playground Project. This revised version was published online in July 2006 with corrections to the Cover Date.     

The Teaching Brain and the End of the Empty Vessel

I am excited to present this special section that explores the teaching brain. The goal of the series is to facilitate a transition in the lens on teaching from an empty vessel to a phenomenon as dynamic, variable, and context‐dependent as learning. This transformation will likely push all of us to reevaluate our understanding and research on teaching. Over the coming year, each issue will provide several articles that seek to shed light on a different aspect of this burgeoning new area of research. This issue opens the series with a piece designed to lay out the conceptual framework and evidence base for a new way to think about teaching: the teaching brain. Next, Michael Chazan gives an archeological grounding for the existence of teaching in the earliest ancestors of Homo sapiens. Sidney Strauss and Margalit Ziv then describe how teaching is a fundamental human cognitive ability. Together, these articles begin to create a paradigm shift in the definition of teaching. We look forward to an exciting journey.     

Learning about associatively activated stimulus representations: Implications for acquired equivalence and perceptual learning

Pavlovian conditioning is taken to reflect the formation of links between the central representations of stimuli. A link will be formed when presentation of the relevant stimuli is scheduled in a way that ensures that two representations are activated concurrently. Once this has occurred, a representation can be activated not only by the occurrence of the appropriate stimulus but also by way of the link. Evidence is reviewed to suggest that activation produced by this second route is, in some ways, functionally equivalent to direct activation; in particular, an associatively activated representation (animage) appears capable of forming further associative links with other event representations. Learning about associatively activated stimulus representations may play a role in a range of phenomena. Its contribution to the following is discussed: sensory preconditioning, second-order conditioning, acquired equivalence and distinctiveness, equivalence class formation, and the perceptual learning effect. Finally, consideration is given to the way in which existing theories of associative learning might be modified in order to accommodate this process.     

Perceptual learning transfer in an appetitive Pavlovian task

In two experiments, rats were given intermixed or blocked preexposure to two similar compound stimuli, AX and BX. Following preexposure, conditioning trials took place in which AX (Experiment 1) or a novel compound stimulus NX (Experiment 2) was paired with a food-unconditioned stimulus in an appetitive Pavlovian preparation. Animals that were given alternated preexposure showed lower generalization from AX to BX (Experiment 1) and from NX to a new compound, ZX (Experiment 2), than animals that were given blocked preexposure, a perceptual learning and a perceptual learning transfer effect, respectively.     

Making connections among multiple graphical representations of fractions: sense-making competencies enhance perceptual fluency,but not vice versa

Prior research shows that representational competencies that enable students to use graphical representations to reason and solve tasks is key to learning in many science, technology, engineering, and mathematics domains. We focus on two types of representational competencies: (1) sense making of connections by verbally explaining how different representations map to one another, and (2) perceptual fluency that allows students to fast and effortlessly use perceptual features to make connections among representations. Because these different competencies are acquired via different types of learning processes, they require different types of instructional support: sense-making activities and fluency-building activities. In a prior experiment, we showed benefits for combining sense-making activities and fluency-building activities. In the current work, we test how to combine these two forms of instructional support, specifically, whether students should first work on sense-making activities or on fluency-building activities. This comparison allows us to investigate whether sense-making competencies enhance students’ acquisition of perceptual fluency (sense-making-first hypothesis) or whether perceptual fluency enhances students’ acquisition of sense-making competencies (fluency-first hypothesis). We conducted a lab experiment with 74 students from grades 3–5 working with an intelligent tutoring system for fractions. We assessed learning processes and learning outcomes related to representational competencies and domain knowledge. Overall, our results support the sense-making-first hypothesis, but not the fluency-first hypothesis.     

Perceptual learning with tactile stimuli in rodents: Shaping the somatosensory system

The animal kingdom contains species with a wide variety of sensory systems that have been selected to function in different environmental niches, but that are also subject to modification by experience during an organism’s lifetime. The modification of such systems by experience is often called perceptual learning. In rodents, the classic example of perceptual learning is the observation that simple preexposure to two visual stimuli facilitates a subsequent (reinforced) discrimination between them. However, until recently very little behavioral research had investigated perceptual learning with tactile stimuli in rodents, in marked contrast to the wealth of information about plasticity in the rodent somatosensory system. Here we present a selective review of behavioral analyses of perceptual learning with tactile stimuli, alongside evidence concerning the potential bases of such effects within the somatosensory system.     

An elemental model of associative learning: I. Latent inhibition and perceptual learning

This paper presents a brief, informal outline followed by a formal statement of an elemental associative learning model first described by McLaren, Kaye, and Mackintosh (1989). The model assumes representation of stimuli by sets of elements (i.e., microfeatures) and a set of associative algorithms that incorporate the following: real-time simulation of learning; an error-correcting learning rule; weight decay that distinguishes between transient and permanent associations; and modulation of associative learning that gives high salience to and, hence, promotes rapid learning with novel, unpredicted stimuli and reduces the salience for a stimulus as its error term declines. The model is applied in outline fashion to some of the basic phenomena of simple conditioning and, in greater detail, to the phenomena of latent inhibition and perceptual learning. A detailed account of generalization and discrimination will be provided in a later paper.     

Testing the translational-symmetry hypothesis of abstract-concept learning in pigeons

The translational-symmetry hypothesis of abstract-concept learning was tested in a same/different (S/D) task with pairs of pictures. The translational-symmetry hypothesis proposes that subjects discriminate same trials by the simultaneous repetition of features in the two pictures (and different trials by the lack of feature repetition). Pigeons that had learned a simultaneous S/D task were tested with delays between the two pictures to remove emergent perceptual cues. In Experiment 1, we tested delays of 0 and 1 sec. The results did not show the accuracy decrease expected according to the translational-symmetry hypothesis. In Experiment 2, we expanded the delays to 2 and 6 sec. Even at the longest delay, there was no evidence of the precipitous performance decline or default strategy that would be predicted by translational symmetry. The results provide evidence against translational symmetry (or other perceptually emergent features) that might control these pigeons’ performance in our two-item S/D task.     

Effects of representational guidance during computer-supported collaborative learning

This research investigates the role of representational guidance by comparing the effects of two different representational tools. We used a design with two different groups defined by the type of argumentative diagram students co-constructed while working in a computer-supported collaborative learning (CSCL) environment. The Graphical Debate-tool offered different representational guidance than the Textual Debate-tool. The results show that groups that worked with the Graphical Debate-tool constructed representations of higher quality and wrote essays that were better in terms of grounds quality. Furthermore, working with the Graphical Debate-tool was found to have a positive effect on students’ learning as measured by a knowledge post-test. In contrast to our expectations however, there was little difference between the two conditions regarding the online collaboration process. It can be concluded that representational guidance has an impact on group and individual performance and should therefore be taken into account during instructional design.     

Effect of extended training on generalization of latent inhibition: an instance of perceptual learning

Four experiments examined generalization of latent inhibition (LI) as a function of the length of preexposure in a conditioned taste aversion procedure with rats. Experiment 1 showed that one or four nonreinforced presentations of a flavor compound (BX) retarded subsequent conditioning to another compound (AX). However, after eight presentations of BX, conditioning to AX occurred at the same rate as with no preexposure. These results indicate that generalization of LI decreased as the length of preexposure to BX increased. Experiment 2 replicated this effect of reducing generalization, as well as demonstrating that LI actually increased as the length of preexposure to AX increased. Experiment 3 extended the generality of the effect to a procedure in which both BX and AX were preexposed. Experiment 4 demonstrated a similar reducing-generalization effect when generalization of LI from BX to X was assessed. All of these data are consistent with the notion that prolonged preexposure to BX enhances its discriminability. Different learning mechanisms that might be responsible for this perceptual learning effect are discussed.