Search Results

In this chapter we consider how the operation of attractor networks in the brain is influenced by noise in the brain produced by the random firing times of neurons for a given mean firing rate; how this can in fact be beneficial to the operation of the brain; and how the stability of these systems and how they are influenced by noise in the brain is relevant to understanding a number of mental disorders. The concept of noise in attractor networks is important to understanding decision-making, short-term memory, and depression and schizophrenia, and this is described in this Chapter. It is a key aim of this book to increase understanding of the brain that is relevant not only to its operation in health, but also in disease, and how it may be possible to ameliorate some of the effects found in these mental and other disorders.

This chapter shows how it is becoming possible to link quantitative neuronal network approaches to other techniques in neuroscience to develop quantitative theories about how brain systems involved in memory operate. It focuses on the hippocampus and nearby structures in the temporal lobe of the brain involved in learning about new facts or events. It discusses the functions performed by the hippocampus and evaluates the relation between spatial and non-spatial hippocampal function, episodic memory, and long-term semantic memory.

The introduction discusses the coverage of this book, which is about the computational neuroscience of vision. It introduces some of the background for understanding brain computation and discusses how some of the essential features of this can be captured by simple formalisms. The introduction also explores three neuronal network architectures, long-term potentiation and long-term depression, and the fine structure of the cerebral neocortex.

This chapter describes the so-called error-correcting neural networks that are capable of mapping a set of inputs to a set of required outputs using correction when errors are made. These networks do not use local learning rules, thus, their biological plausibility remains uncertain. However, it has been suggested that perceptron learning may be implemented by the special neuronal network architecture of the cerebellum. This chapter also discusses the delta rule and backpropagation of error in multilayer networks.

An introduction is provided to neurons; computation by biologically plausible networks of neurons; the representation of information in the brain; the functions of different brain areas; and the structure and connectivity of the cerebral neocortex.

This chapter examines the fundamental operation involved in pattern association memory. It provides a diagram showing the essential elements necessary for pattern association, forming what could be called a prototypical pattern associator network. It also discusses the importance of distributed representations for pattern associators, associative neuronal networks with non-linear neurons, and the implications of different types of coding for storage in pattern associators.

Despite tremendous progress, the most exciting times in dendrite research lie ahead of us. Much of the knowledge we have accumulated to date has been descriptive. Although we have amassed a detailed catalog of structural features and molecular machinery, and discovered a wide repertoire of functional effects imparted by the dendritic tree, how these features allow neuronal networks to generate behavior remains elusive. Two key questions that need to be addressed at this stage are: What computations does each neuron perform within its neuronal network? And which features of dendrites are most relevant to how the neuron performs these computations?

This chapter presents a longitudinal experiment that compares the effects of procedural vs explicit music teaching on brain networks. It shows that cortical activation during music processing reflects the auditory ‘learning biography’, the personal experiences accumulated over time. Listening to music, learning to play an instrument, formal instruction, and professional training result in multiple, in many instances multisensory, representations of music, which seem to be partly interchangeable and rapidly adaptive. Topics addressed include changing concepts in neuromusicology, concepts of perceptive modules and hierarchies, ‘music centres’ in the brain reflecting the auditory biography, and a tentative model on brain substrates of music processing. In summary, as soon as ‘real music’ is considered apart from laboratory experiments, one can expect individually formed and quickly adaptive brain substrates, including widely distributed neuronal networks in both hemispheres.

This chapter examines the neuronal mechanisms underlying syntactic structure and processing. In order to elucidate the neurobiology of syntax, it looks at syntactic encoding in language production. In particular, it explores processes that store and activate syntactic representations in neuronal networks, as well as how syntactic representations and processes relate to larger-scale neural features such as event-related potentials and areas of the brain determined functionally or anatomically. It also discusses the neural organization for processing syntax in the first spoken languages of adults, focusing on lateralization and regional specialization. The chapter considers other brain features involved in syntactic processing, including the receptor structure, before concluding with a discussion of neural organization for processing syntax in children and individuals with impaired development.

Long considered a mystery and still an object of worry among philosophers and neuroscientists for its stubborn refusal to die or just lie down and allow its own reduction to “objective” – third person – terms, consciousness is the explicit, sometimes outspoken side of the soul. Intentionality, the fact that so many of our thoughts are about something outside themselves, points back insistently to a subject with a distinctive (notably limited) perspective. Philosophers speak of qualia. But other facts about consciousness join in to attest to the soul’s active presence: the striking recursivity and implicit self-reference of thought (not least in the moral realm, where reflection begets conscience), our awareness of time and of others’ intentions, our ability to relate what we see to what we touch or hear and to claim both experiences as our own, the sense of connection that binds up our moments, and the simple (or not so simple) biology of selfhood that begins at the boundaries between one organism and the next but grows ever subtler and more versatile until consciousness openly claims one’s thoughts as one’s own and consciousness announces itself as the voice of an emergent person, of a soul.

This chapter starts by reviewing the literature on the architecture of the parser that arose from the sentence The horse raced past the barn fell. The author points out how research on ambiguity resolution evolved in the 1990s towards consideration of the processes occurring before the point of ambiguity. Focusing on Elman’s emergentist approach to language acquisition that demonstrated that structure could emerge through abstraction across experience, Altmann turns to analyzing how prediction based on previous experience with the language directly influences sentence processing. There is therefore little difference between non-ambiguous and ambiguous sentences, since the space of continuations is determined by past experience. Knowledge of language can be operationalized as the ability to predict on the basis of current and prior contexts. The author’s data show that what is being predicted includes representations of the changes in the real world that would constitute the event described by the language.