This book is based on a relatively small and coherent set of mechanistic principles, which are introduced in part I of the text, and then applied in part II to understanding a range of different cognitive phenomena. These principles are implemented in the Leabra algorithm for the exploration simulations. These explorations are woven throughout the chapters where the issues they address are discussed, and form an integral part of the text. To allow readers to get as much as possible out of the book without doing the simulations, we have included many figures and have carefully separated the procedural aspects from the content using special typesetting.
Because this book emphasizes the linkages and interactions between biology, computational principles, and a wide variety of human cognitive phenomena, we cannot provide exhaustive detail on all potentially relevant aspects of neuroscience, computation, or cognition. We do attempt to provide references for deeper exploration, however. Relatedly, all of the existing supporting arguments and details are not presented for each idea in this book, because in many cases the student would likely find this tedious and relatively uninformative. Thus, we expect that expert neuroscientists, computational/mathematical researchers, and cognitive psychologists may find this book insufficiently detailed in their area of expertise. Nevertheless, we provide a framework that spans these areas and is consistent with well-established facts in each domain.
Thus, the book should provide a useful means for experts in these various domains to bridge their knowledge into the other domains. Areas of current debate in which we are forced to make a choice are presented as such, and relevant arguments and data are presented. We strive above all to paint a coherent and clear picture at a pace that moves along rapidly enough to maintain the interest (and fit within the working memory span) of the reader. As the frames of a movie must follow in rapid enough succession to enable the viewer to perceive motion, the ideas in this book must proceed cleanly and rapidly from neurobiology to cognition for the coherence of the overall picture to emerge, instead of leaving the reader swimming in a sea of unrelated facts.
Among the many tradeoffs we must make in accomplishing our goals, one is that we cannot cover much of the large space of existing neural network algorithms. Fortunately, numerous other texts cover a range of computational algorithms, and we provide references for the interested reader to pursue. Many such algorithms are variants on ideas covered here, but others represent distinct frameworks that may potentially provide important principles for cognition and/or neurobiology. As we said before, it would be a mistake to conclude that the principles we focus on are in any way considered final and immutable -- they are inevitably just a rough draft that covers the domain to some level of satisfaction at the present time.
As the historical context (section 1.3) and overview of our approach (section 1.4) sections made clear, the Leabra algorithm used in this book incorporates many of the important ideas that have shaped the history of neural network algorithm development. Throughout the book, these principles are introduced in as simple and clear a manner as possible, making explicit the historical development of the ideas. When we implement and explore these ideas through simulations, the Leabra implementation is used for coherence and consistency. Thus, readers acquire a knowledge of many of the standard algorithms from a unified and integrated perspective, which helps to understand their relationship to one another. Meanwhile, readers avoid the difficulties of learning to work with the various implementations of all these different algorithms, in favor of investing effort into fully understanding one integrated algorithm at a practical hands-on level. Only algebra and simple calculus concepts, which are reviewed where necessary, are required to understand the algorithm, so it should be accessible to a wide audience.
As appropriate for our focus on cognition (we consider perception to be a form of cognition), we emphasize processing that takes place in the human or mammalian neocortex, which is typically referred to simply as the cortex. This large, thin, wrinkled sheet of neurons comprising the outermost part of the brain plays a disproportionally important role in cognition. It also has the interesting property of being relatively homogeneous from area to area, with the same basic types of neurons present in the same basic types of connectivity patterns. This is principally what allows us to use a single type of algorithm to explain such a wide range of cognitive phenomena.
Interactive, graphical computer simulations are used throughout to illustrate the relevant principles and how they interact to produce important features of human cognition. Detailed, step-by-step instructions for exploring these simulations are provided, together with a set of exercises for the student that can be used for evaluation purposes (an answer key is available from the publisher). Even if you are not required to provide a written answer to these questions, it is a good idea to look them over and consider what your answer might be, because they do raise important issues. Also, the reader is strongly encouraged to go beyond the step-by-step instructions to explore further aspects of the model's behavior.
In terms of the detailed organization, part I covers Basic Neural Computational Mechanisms across five chapters (Individual Neurons, Networks of Neurons, and three chapters on Learning Mechanisms), and part II covers Large-Scale Brain Area Organization and Cognitive Phenomena across five chapters (Perception and Attention, Memory, Language, and Higher-Level Cognition, with an introductory chapter on Large-Scale Brain Area Functional Organization). Each chapter begins with a detailed table of contents and an introductory overview of its contents, to let the reader know the scope of the material covered. When key words are defined or first used extensively, they are highlighted in bold font for easy searching, and can always be found in the index. Simulation terms are in the font as shown.
Procedural steps to be taken in the explorations are formatted like this, so it is easy to see exactly what you have to do, and allows readers who are not running the model to skip over them.
Summaries of the chapters appear at the end of each one (this chapter excluded), which encapsulate and interrelate the contents of what was just read. After that, a list of references for further reading is provided. We hope you enjoy your explorations!