Akira Miyake and Priti Shah (eds)
Models of working memory:
Mechanisms of active maintenance and executive control

New York, NY: Cambridge University Press, 1999

  CogSci index  |  CogWeb
Abstract of the Introduction   Contents

The authors believe that the field of working memory has made significant progress during the past 25 years and has reached a critical point at which detailed comparisons of different theoretical proposals are not only possible but would be tremendously beneficial for further theoretical development of working memory research. The central rational behind this volume is to provide such a forum for systematic comparisons of existing models and theories of working memory. This strong theoretical focus reflects a current need in the field. Whereas existing models of working memory each provide a sophisticated account of certain aspects of working memory processes and functions, different models have different theoretical emphases and tend to leave some other aspects of working memory relatively unspecified.


Contents  top

List of contributors.

  1. Models of working memory: An introduction. Priti Shah and Akira Miyake.
  2. Working memory: The multiple-component model. Alan D. Baddeley and Robert H. Logie.
  3. An embedded-processes model of working memory. Nelson Cowan.
  4. Individual differences in working memory capacity and what they tell us about controlled attention, general fluid intelligence, and functions of the prefrontal cortex. Randall W. Engle, Michael J. Kane and Stephen W. Tuholski.
  5. Modeling working memory in a unified architecture: An ACT-R perspective. Marsha C. Lovett, Lynne M. Reder and Christian Lebiere.
  6. Insights into working memory from the perspective of the EPIC architecture for modeling skilled perceptual-motor and cognitive human performance. David E. Kieras, David E. Meyer, Shane Mueller and Travis Seymour.
  7. The soar cognitive architecture and human working memory. Richard M. Young and Richard L. Lewis.
  8. Long-term working memory as an alternative to capacity models of working memory in everyday skilled performance. K. Anders Ericsson and Peter F. Delaney.
  9. Interacting cognitive subsystems: Modeling working memory phenomena within a multiprocessor architecture. Philip J. Barnard.
  10. Working memory in a multilevel hybrid connectionist control architecture (CAP2). Walter Schneider.
  11. A biologically based computational model of working memory. Randall C. O'Reilly, Todd S. Braver and Jonathan D. Cohen.
  12. Models of working memory: Eight questions and some general issues. Walter Kintsch, Alice F. Healy, Mary Hegarty, Bruce F. Pennington and Timothy A. Salthouse.
  13. Toward unified theories of working memory: Emerging general consensus, unresolved theoretical issues, and future research directions. Akira Miyake and Priti Shah.

Name index.
Subject index.

Chapter Abstracts

2. Baddeley, Alan D.; Logie, Robert H. Working memory: The multiple-component model. pp. 28-61. Contents

Abstract: The authors' own definition of working memory is that it comprises those functional components of cognition that allow humans to comprehend and mentally represent their immediate environment, to retain information about their immediate past experience, to support the acquisition of new knowledge, to solve problems, and to formulate, relate, and act on current goals. Their theoretical approach has developed in the framework of working memory comprising multiple specialized subcomponents of cognition. Although the Baddeley-Logie model maintains the original tripartite structure proposed by A. D. Baddeley and G. J. Hitch (1974), it has undergone a number of important changes, particularly in regard to specifying functions of the central executive.

3. Cowan, Nelson. An Embedded-Processes Model of working memory. pp. 62-101. Contents

Abstract: This chapter presents the author's Embedded-Processes Model, a broad-scope information processing framework originally developed to synthesize a vast array of findings on attention and memory. The mnemonic functions preserving information that can be used to do the necessary work collectively make up working memory. This is a functional definition in that any processing mechanism contributing to the desired outcome are said to participate in the working memory system. In contrast, some researchers appear to prefer to define working memory according to the mechanism themselves. Though my framework has much in common with those of other researchers, a functional definition of working memory seems more likely to encourage a consideration of diverse relevant mechanisms. Some theories of working memory equate it to the focus of attention and awareness and some equate it to the sum of activated information. Often, the distinction between activation and awareness is left unclear, but I argue that the distinction is important and that working memory must involve both, and some long-term memory information as well.

4. Engle, Randall W.; Kane, Michael J.; Tuholski, Stephen W. Individual differences in working memory capacity and what they tell us about controlled attention, general fluid intelligence, and functions of the prefrontal cortex. pp. 102-134. Contents

Abstract: The authors describe their theoretical perspective on working memory, which could be called the "controlled attention" framework. Recent cognitive neuroscience findings on the role of prefrontal cortex in executive control and attentional processes provided an inspiration for this approach, and they discuss a series to illustrate the role of "controlled attention" in working memory and complex cognition. The authors describe an attempt by their lab to determine construct validity for short-term and working memory measures. The authors proposed that individual differences on measures of working memory capacity primarily reflect differences in capability for controlled processing and, thus, will be reflected only in situations that either encourage or demand controlled attention. We propose that working memory capacity reflects the ability to apply activation to memory representations, to either bring them into focus or maintain them in focus, particularly in the face of interference or distraction..

5. Lovett, Marsha C.; Reder, Lynne M.; Lebiere, Christian. Modeling working memory in a unified architecture: An ACT-R perspective. pp. 135-182. Contents

Abstract: We describe a model of working memory that is developed within the ACT-R cognitive architecture. Some of the main features are derived from the basic features of ACT-R: 1) processing depends on the current goal of the system; 2) the accessibility of declarative and procedural knowledge varies with experience; 3) there is limited attentional resource, focused on the current goal, that increases the accessibility of goal-relevant knowledge relative to other knowledge; 4) in more complex and memory-demanding tasks, this limited resource is spread more thinly thus impairing retrieval of goal-relevant items; and 5) the "capacity" of this attentional resource may vary from person to person, influencing the ability to access goal-relevant information across domains.... In this chapter, we describe the basic features of the ACT-R theory and relate them to several working memory issues (the designated questions). We present an ACT-R model that predicts working memory results at the aggregate level and then go on to report our work in progress--developing ACT-R models of individual's working memory performance. We summarize some of the encouraging results we have obtained thus far and discuss their implications for future work.

6. Kieras, David E.; Meyer, David E.; Mueller, Shane; Seymour, Travis. Insights into working memory from the perspective of the EPIC architecture for modeling skilled perceptual-motor and cognitive human performance. pp. 183-223. Contents

Abstract: Computational modeling of human perceptual-motor and cognitive performance based on a comprehensive detailed information-processing architecture leads to new insights about the components of working memory. To illustrate how such insights can be achieved, a precise production-system model that uses verbal working memory for performing a serial memory span task through a strategic phonological loop has been constructed with the Executive-Process/Interactive-Control (EPIC) architecture of D. M. Kieras and D. E. Meyer.... More specifically, this chapter considers working memory from the perspective of a particular architecture for characterizing the human information-processing system. Such architectures are essential to construct because they provide theoretical foundations and sets of mechanisms for human cognition and action, though which veridical computational models of performance can be formulated for specific tasks.

7. Young, Richard M.; Lewis, Richard L. The Soar cognitive architecture and human working memory. pp. 224-256. Contents

Abstract: We examine the various phenomena of human working memory (WM) from a cognitive architecture of broad scope, the Soar architecture, which focuses on the functional capabilities needed for a memory system to support performance in a range of cognitive tasks. We argue for and demonstrate 3 points concerning the limitations of WM. 1. We show how the cognitive system, even with a limited-capacity short-term store, can handle complex tasks that require large quantities of information, by relying heavily on recognition-based long-term memory working in concert with the external environment. 2. We argue that limitations on WM arise even in purely functional cognitive systems built without preset capacity constraints, and hence that empirically demonstrated limitations of effective WM do not necessarily imply a capacity-constrained underlying memory system. 3. We show how a specific mechanism of similarity-based interference can act as a resource constraint on the cognitive system and offer a coherent account of a wide range of psycholinguistic phenomena.

8. Ericsson, K. Anders; Delaney, Peter F. Long-term working memory as an alternative to capacity models of working memory in everyday skilled performance. pp. 257-297. Contents

Abstract: We use K. A. Ericsson and W. Kintsch's (1995) terminology, where the temporary storage in working memory (WM) proposed by A. D. Baddeley (1992) is called short-term working memory (ST-WM), distinguishing it from other possible types of working memory, such as maintained access to information in long-term working memory (LT-WM). In fact, we advocate the broadest possible description of working memory and its possible mechanisms. At the most general level, the essence of the concept of WM (or the part of memory that works) is that only a minute fraction of all the knowledge, skills, and information stored in subjects' vast long-term memory is influencing the subjects' behavior and thought processes at a specific instant of time. Hence, the phenomenon of working memory includes all those mechanisms that maintain selective access to the information and the procedures that are necessary for a subject to complete one or more specific concurrent tasks. By adopting this broad description of WM, we are now in a position to compare and contrast more specific mechanisms and conceptions and their distinctive research approaches within a common framework.

9. Barnard, Philip J. Interacting Cognitive Subsystems: Modeling working memory phenomena within a multiprocessor architecture. pp. 298-339. Contents

Abstract: This chapter describes one approach to developing a macro-theory of the organization and functioning of the complete mental mechanism, the Interacting Cognitive Subsystems (ICS). At a systemic level it defines what the resources are, how they are organized, what they do, and how they interoperate. To control the potential problem of rapid proliferation of unwarranted assumptions, a layered approach is adopted. Then the chapter illustrates how ICS could be applied to the analysis and explanation of performance in working memory.... The first part of the chapter describes the basic mechanisms and representations of the ICS framework in a series of stages. The architecture is introduced first in outline form, together with a basic description of the types of information handled by each subsystem. The internal organization of subsystem resources and the core principles governing their operation are then defined. The full architecture is subsequently elaborated, followed by concrete illustrations of the wider patterns of information flow that it permits. The second part includes a brief overview of some current research driven by the approach.

10. Schneider, Walter. Working memory in a multilevel hybrid connectionist control architecture (CAP2). pp. 340-374. Contents

Abstract: In a connectionist control network, working memory is implemented via short-term activation and connection changes that support cognitive operations. The Controlled Automatic Processing version 2 (CAP2) approach is a model of skilled processing and learning. When applied to working memory the model instantiates multiple forms and mechanisms of working memory.... The chapter provides an interpretation of working memory based on biological, information-processing, and behavioral constraints. The chapter interrelates previously published themes, which include the distinction between controlled and automatic processing, the role of multileveled connectionist control structure in working memory, the use of that control structure to enable learning by instruction, attentional control, and the role of consciousness in the effective control of processing. Topics include (a) a brief review of physiological themes and mechanisms underlying working memory; (b) a detailed description of the CAP2; (c) some brief brain-imaging results about the changes that occur as skills is acquired; and (d) the conclusion with a discussion of the specific working memory questions, as well as a commentary about the other models presented in this volume.

11. O'Reilly, Randall C.; Braver, Todd S.; Cohen, Jonathan D. A biologically based computational model of working memory. pp. 375-411. Contents

Abstract: This chapter presents a biologically based model of working memory. The authors' connectionist framework represents an attempt to start developing an explicit computational model of working memory and executive control that is biologically plausible and is firmly rooted in the principles of cognitive processing in the brain. This chapter brings studies of working memory into closer alignment with our rapidly expanding knowledge of its underlying biological and neural basis. We define working memory as controlled processing involving active maintenance and/or rapid learning, where controlled processing is an emergent property of the dynamic interactions of multiple brain systems, but the prefrontal cortex (PFC) and hippocampus (HCMP) are especially influential owing to their specialized processing abilities and their privileged locations within the processing hierarchy (both the PFC and HCMP are well connected with a wide range of brain areas, allowing them to influence behavior at a global level).





Maintained by Francis F. Steen, Communication Studies, University of California Los Angeles