Search Results

This chapter presents a case for the dual coding approach as well as for the constructive-empiricist philosophy of science that the approach represents. Dual coding theory and the convergent operational approach associated with it have been applied to a wide range of memory and learning tasks using different materials. The tasks include associative memory, item memory, sequential memory, and discrimination learning with materials ranging from pictures, to concrete and abstract words or larger units. The results provide strong support for the dual coding approach to aspects of episodic memory, and they strain the explanatory capacity of current prepositional or other single-code, representational-processing theories unless they are modified by post-hoc assumptions that essentially represent rephrasing of the core assumptions of dual coding theory.

This chapter begins with a review of behavioral evidence of associative learning in vision. It considers which types of associations can be learned, and then explores the properties and constraints of the mechanisms involved in such learning. It also discusses how visual processing can be facilitated by knowledge of associative relationships between objects. The second part of the chapter analyzes the neural mechanisms that support visual associative learning.

This chapter reviews the role of the medial temporal lobe in associative memory. Part I presents the results from behavioral neurophysiological studies focused on recording in both the perirhinal cortex and area TE during the performance of well-learned visual-visual paired associate (VPA) tasks. Part II reviews the neurophysiological studies in the medial temporal lobe that have examined the development of new associative representations. These studies reveal dynamic changes in both the perirhinal cortex and hippocampus during the acquisition of new associative memories.

The use of computational models has been invaluable for exploring the link between neurons and behavior, enabling hypothetical mechanisms to be defined precisely and examined quantitatively. This chapter reviews many of these models, including models of spatial functions, models of more general associative mnemonic functions, models that stress feedforward processing through the hippocampal system, and those stressing recurrent processing within it. Spatial models are reviewed first, as they are most firmly rooted in the known electrophysiology of the region. These models cover both the representation of the animal's spatial location and orientation and the use of this information in spatial navigation. The models of mnemonic function, specifically associative or episodic memory, follow from Marr's seminal 1971 model. This model is used as a generic framework in which to consider the various subsequent developments to it. Finally, the chapter reviews those models attempting to bring together the spatial and mnemonic functions of the hippocampus.

A Darwinian approach to decision-making mechanisms must focus on the representation of the options between which the animal decides. For example, in matching behavior, is the animal deciding between the different locations in which to forage or simply whether to leave its current location? A neurobiologically informed approach must be concerned with the mechanism of representation itself. In the computational theory of mind, options are represented by the symbols that carry information about them forward in time. In conventional computing machines, symbols reside in an addressable read–write memory. Current theorizing about the neurobiological mechanism of memory rejects this form of memory in favor of an associative memory. The problem is that the associative bond—and its neurobiological embodiment, the plastic synapse—is not suited to the function of carrying acquired information forward in time in a computationally accessible form. It is argued that this function is indispensable. Therefore, there must be such a mechanism in neural tissue, most probably realized at the molecular level.

Creativity paradoxically involves studying how someone pulls out of their brain something that was never put into it, something new and useful, or appropriate to the task at hand. Even if we determine which areas of the brain are active and how these areas interact during creative thought, we will not have explained how the brain generates useful novelty. However, since the representational capacity of neurons emerges at a level that is higher than that of individual neurons, the inner workings of neurons is too low a level to explain creativity. Thus, this chapter focuses at a level midway between gross brain regions and neurons. Since creativity involves combining concepts from different domains, or seeing ideas from new perspectives, we examine the neural mechanisms underlying the representation of concepts and ideas, and show how novelty arises through activation of new regions of overlap between distributed aggregates of neurons. Neurons that would be active during associative (or divergent) thought, but not during analytic (or convergent) thought, are referred to as ‘neurds’. We explain why neurds are key to understanding the neural mechanisms underlying creativity, and show with examples how creativity benefits from the capacity to shift between these modes of thought.

This chapter examines one of Dickens's most popular serial novels alongside the religious tracts that were a hallmark of Evangelical propagandizing. Even as it shares with Evangelical tracts certain common associationist assumptions about the effects of reading on the memory, The Old Curiosity Shop contests the cultural politics of the larger evangelical movement by subverting the rhetorical and ideological rationales behind didactic fictions, such as Hannah More's Cheap Repository Tracts and Legh Richmond's Annals of the Poor. Shaping an “anti-didactic” strategy for popular fiction, Dickens's moralizing tale about Little Nell's unjust death counteracts the cultural influence of evangelicalism by substituting a benevolent curiosity and activist sensibility in the place of Evangelical fiction's staging of pious deaths to motivate the reader's religious conversion and social deference.

The scope of the theory of neural networks has been expanding rapidly. The statistical-mechanical techniques stemming from the theory of spin glasses have been playing important roles in the analysis of model systems. This chapter summarizes basic concepts, such as neurons, synapses, axons, and studies the characteristics of networks with interneuron connections given by a specific prescription called the generalized Hebb rule. An important goal is to elucidate the structure of the phase diagram with various parameters as their axes, such as the temperature and number of memorized patterns. Related is the capacity of a network, which means the number of patterns that can be memorized. The problem of learning, where the connections gradually change according to some rules to achieve specified goals, is delegated to the next chapter.

Learning, and hence memory, is ubiquitous not only throughout the animal kingdom, but apparently throughout many regions of the brain. Is all learning reducible to a single common form? Neuropsychological dissociations suggest that the mammalian brain possesses a number of different and potentially independent memory systems, with different mechanisms and anatomical dispositions, some of which are neurally widely dispersed and others of which are narrowly organized. Among the types considered are: short-term memory; knowledge and skills; stable associative memory; event memory; and priming. As double or multiple dissociations do not lead to logically inevitable conclusions, it has been argued that an alternative to multiple memory systems is variable modes of processing. Multiple memory systems may possibly share some common cellular mechanisms, but such mechanisms do not define the separate properties at the systems level.

Rapid eye movement (REM) sleep is a stage of sleep that evolved in part to provide a privileged time in each day when the brain is disconnected from sensory input and freed of intentional, directed thought. The neurochemistry and neurophysiology of the brain during REM sleep is optimized for the exploration of normally ignored connections and associations within the brain’s vast repertoire of stored information. This includes changes in the activity of dorsolateral prefrontal, anterior cingulate, and medial orbital frontal cortices and the hippocampus, and reductions in norepinephrine and increases in acetylcholine in the cortex. This exploration of normally weak associations is critical to the creative process, and REM sleep can thus be considered a period of unbridled creativity. Much of this creative process is reflected in the content of dreams. Even without waking dream recall, changes within associative networks produced by the brain mechanisms of dream construction can leave these brain networks—and the individual—primed for reactivation at a later time, leading to the “discovery” of creative insights. Some, but not all, of these brain changes are also seen during periods of quiet rest with activation of the default mode network (DMN). When active, this network can likewise provide a state of enhanced creativity. Nevertheless, REM sleep and dreaming provide a protected two hours every day when creative processes run at full speed.

David Marr’s classic 1971 paper laid the foundation for a standard model of the hippocampus combining pattern separation (Marr’s codon hypothesis) and pattern completion. His paper discusses functions that became components of a standard theoretical model of hippocampal function within the field, and inspired further studies on more detailed analysis of pattern separation in the dentate gyrus and pattern completion by attractor dynamics in region CA3. This theoretical framework has been the focus of a number of experimental tests over the years, including neurophysiological studies of patterns of neuronal activity in behaving rats, tests of behavioural effects with selective genetic manipulations, and fMRI studies of hippocampal activity in humans. This chapter also reviews subsequent research related to the dynamics of pattern completion, as well as subsequent modeling that addressed network dynamics in the hippocampus including theta rhythm oscillations, theta phase precession and the patterns of grid cell firing in entorhinal cortex. Finally, this chapter addresses the need for future breakthroughs by researchers that combine sophisticated mathematical techniques with detailed knowledge of the anatomy and neurophysiology of cortical structures.

In this chapter, modality-specific imagery is considered in the context of emergent symbols for actual representation, by drawings and maps. It is argued that covert coding of movement output can mediate non-visual spatial representation in memory. It suggested that the invention and discovery means of representation do not necessarily depend on a long process of associative learning. Further, this chapter proposed that cognitive and modality-specific aspects of information are not mutually exclusive aspects of information processing. Finally, the relevance of temporary memory to spatial development on non-verbal representation is discussed extensively.

The ventral prefrontal cortex learns to associate objects, faces, and vocalizations, and its connectional fingerprint explains why it alone can do so. It receives visual inputs from the inferior temporal cortex and auditory ones from the superior temporal cortex. It combines these inputs with those from the orbital prefrontal (PF) cortex so as to specify the goal that is currently desirable. This is then transformed into the target of search via connections with the frontal eye field and the target for manual retrieval via connections with the premotor areas. The ventral PF cortex can also learn to form associations between objects, for example by linking them into categories. These can be retrieved from long-term memory via connections with the hippocampus.