Timing the Future: The Case for a Time-based Prospective Memory

timing the future the case for a time-based prospective memory

This page intentionally left blank

timing the future the case for a time-based prospective memory

edited by

joseph glicksohn, PhD michael s myslobodsky, MD, DSc

World Scientific NEW JERSEY

•

LONDON

•

SINGAPORE

•

BEIJING

•

SHANGHAI

•

HONG KONG

•

TA I P E I

•

CHENNAI

Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

TIMING THE FUTURE The Case for a Time-Based Prospective Memory Copyright © 2006 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN 981-256-497-7

Typeset by Stallion Press Email: [email protected]

Printed in Singapore.

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future fm

Contents Preface vii

CHAPTER 1

Time Perception and Time-Based Prospective Memory 1 Peter Graf and Simon Grondin

CHAPTER 2

Prospective Remembering Involves Time Estimation and Memory Processes 25 Richard A. Block and Dan Zakay

CHAPTER 3

Dynamic Attending and Prospective Memory for Time 51 Mari Riess Jones

CHAPTER 4

Representing Times of the Past, Present and Future in the Brain 87 Wim A. van de Grind

CHAPTER 5

At the Crossroads of Time and Action: A Temporal Discounting Primer for Prospective Memory Researchers 117 Thomas S. Critchfield and Gregory J. Madden

CHAPTER 6

Time Management 143 Jan Francis-Smythe v

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future fm

vi

Contents

CHAPTER 7

Transcending the Now: Time as a Dimension of Psychological Distance 171 Cheryl J. Wakslak, Yaacov Trope and Nira Liberman

CHAPTER 8

Time Monitoring and Executive Functioning: Individual and Developmental Differences 191 Timo Mäntylä and Maria-Grazia Carelli

CHAPTER 9

The Neural Correlates of Timing Functions 213 Katya Rubia

CHAPTER 10

The Neurology and Neuropsychology of Time-Based Prospective Memory 239 Janet Cockburn

CHAPTER 11

What it Takes to Remember the Future 263 Joseph Glicksohn and Michael S. Myslobodsky

Index 307

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future fm

Preface

S

ince millennia, people have pondered over the puzzle of time, its nature and its meaning. Among numerous Merriam Webster definitions of the word, one alludes to “events which succeed one another from past through present to future.” That definition alludes to the recurrent ebb and flow of events that are ever new, in the Heraclitic sense, to some intrinsic stereotypy of the bodily internal rhythms as well as those of nature when a change or a spatial distance between events is defined as time. The word “tide” (from Old Germanic “tídiz”) must have evolved into contemporary “time” applied also to evolutionarily programmed biochemical events, such as feast–famine and physical activity–rest cycles. The remnants of this tradition are in our calendar with the cycles of death (in the Fall) and recovery (in Spring). As in ancient Rome, our year begins with Januarius (the sacred month of Janus, the two-faced patron of beginnings and endings who saw the past and had his other face, turned to the future; he was also the deity-protector of portals, which as we know guide both ways). Nothing seemed more vital than having a glimpse of the future to cushion disquiet when awaiting the unknown. Timekeeping devices were produced as a way of organizing the world by planning a linear succession of affairs, predicting and thus hoping for controlling “an hour-glass and a scythe.” Whatever we do (often labeled in the language of orienting response by “what” it is) is also defined spatially (as “where” we do it), but these two questions only become relevant when they are provided with a time-tag of “when.” We often fall victims of mishandling the “when” regardless of how well the “what” and “where” are located. That alone explains the practical value of studies of the mechanisms of time-cued intentions.

vii

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future fm

viii

Preface

In neuropsychology, the capacity for memory of future intentions is designated as prospective memory. Memory is considered prospective when it is provided from the outset with an explicit and mostly transitory retrieval “address” (“prospective trigger”) either in a time-based format or as contextual or event-related cues. That specific reference determines how prospective memory is organized and how it is related to other cognitive functions. While remembering to implement intentions at a certain point in time is believed to merit an independent status in brain research, the pace of investigations in the area was surprisingly slow. In 2000, Ellis and Kvavilashvili asserted: “Although the first experimental study on prospective memory within cognitive psychology was conducted nearly 30 years ago . . . subsequent research in this new field grew steadily but somewhat slowly due to the efforts of only a handful of researchers.” Five years later, in their Welcome to the homepage of the 2nd International Conference on Prospective Memory, its organizers write more enthusiastically that “since the 1st Conference in 2000, the field of prospective memory research has been booming.” Our literature search of published papers appearing after 2000 in the Web of Science using the term “prospective memory” witnessed a steep proliferation of reports in this field reaching about 200 papers. Hardly a mark of a renaissance, this list assures that no longer is prospective memory area an emerging field struggling to achieve greater scientific respectability and independence. The nature of time-cued prospective memory has received relatively little attention. In spite of its worth, it remains a mostly descriptive area of prospective memory research. Part of it belongs to the difficulty of the notion of time. The word is both an everyday patinated cliché and one of those grandiloquent terms that still ignites debates of physicists and philosophers. Neuropsychologists refrain from the transcendental aspects of timing the future in favor of a more pragmatic, task-related stance. These tasks are examined by the students of diverse disciplines of cognitive neurosciences, mostly those of time perception, memory for time and prospective memory. That is why the chapters in this collection deal with time perception and explore the ways time is used as a cue of prospective memory paradigms. Surprisingly, the tradition of studying the effects of temporal distance from decision to outcome has strong methodological roots in economics (or behavioral economics) (e.g. “temporal discounting”). Temporal discounting scenarios increasingly inspire cognitive psychologists as well as health professionals whereas social psychologists, rather recently, model its more human face (e.g.

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future fm

Preface

ix

“temporal construal theory”). In the past, each group exhibited an attitude of “benign neglect” towards the other. That attitude has changed, also in large part because of the greater visibility of the prospective memory studies, so that it has become more feasible to provide psychological and neurobiological bridges between these areas. In designing the book, we consciously omitted all “applied” illustrations of prospective memory in the conviction that the discipline is genuinely practical. In its different forms, prospective memory keeps our lives glued to the future when we need to shop, keep appointments, order prescribed drugs, meet a partner on the tennis court, pay the bills or need to come “on time” for a party. It is a part of our self-testing, error-monitoring, plan evaluation and other problemsolving behaviors. Total evasion of planning heralds a perilous withdrawal from reality — more so because prospective goal is a state, but prospective memory is the act of getting there. It says a lot about people’s resolve, being in command of events, self-command, and self-restraint. Adequate prospective memory delays the prospects of cognitive disability or the need to move an elderly individual to a nursing home. That explains why the readership of this volume easily includes researchers and practitioners dealing with development and aging, social competence, pregnancy, manipulation of consumers’ intentions, forecast of sales, marketing and advertising, and even those examining performance in a multitask environment of traffic operators. We hope that this publication will stimulate the reader’s interest in the processes whereby people time the future. It is a pleasure and a privilege for us to recognize all those who helped to produce this volume. Elaine Tham (now at Springer) was an enthusiastic midwife of the book. Terry Goldberg (now at Albert Einstein College of Medicine) gave it a formal welcome. Simon Goodman (Penn University), Richard Coppola (NIHM), and Leslie Hicks (Howard University), as well as our students on both sides of the Atlantic prompted stimulating discussions in the early phase of this project. The last chapter owes much to the generous comments of Arnold Wilkins (University of Essex, UK). Lia Kvavilashvili (University of Hertfordshire, UK) shared with us a sample of her slide presentation at the 2nd Meeting on Prospective Memory ahead of publication. Francois Lalonde helped in producing illustration materials. Alexandra Parmet-Myslobodsky proposed the design of the cover that was handsomely amended by Ian Seldrup our editor at World Scientific Publishing/Imperial College Press, who provided his hand and counsel all the way through. The project was partially supported by a sabbatical year from Bar-Ilan University for Joseph Glicksohn and a 2005

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future fm

x

Preface

Howard University Excellence Award to Michael Myslobodsky. We acknowledge help from all those who aided in internal peer-review of each submission selected for the volume. This is also an opportunity to recognize the support of Dan Weinberger (NIMH, Bethesda, USA) whose hospitality and friendship facilitated the execution of this project. Last, but not the least, Luzian Barr’s moving story was an unwitting inspiration and a reminder of the powers of prospective memory. During WW2, he was deported to Germany from Poland. “If we are to remain alive,” he heard his father mutter before the family was scattered in different concentration camps, “let us gather in the municipality of Łódz.” Freed from the camp after the war — still a teenager — he made an arduous journey through the devastated postwar terrain of the two countries to reach, finally, his native city of Łódz. There, on the municipality steps he found the second survivor of his family, his older brother.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

1 Time Perception and Time-Based Prospective Memory Peter Graf ∗ and Simon Grondin†

Introduction In this chapter, we review experimental psychology research in two domains: time perception and time-based prospective memory (ProM). Intuition suggests that these domains are connected, that they involve at least some of the same high-level cognitive processes or mechanisms. In view of this intuition, it is surprising that only a small number of empirical investigations have focused directly on the processes or mechanisms that link time perception and time-based ProM. Why? In order to answer this question, in the first part of this chapter, we summarize recent empirical and theoretical work on time perception, and on how this ability changes across the adult lifespan. In the second part, we review empirical and theoretical work on time-based ProM and on how this cognitive function changes across the adult lifespan. In addition, we examine the manner in which time- and event-based ProM tasks have been defined, in order to identify where — under what kinds of study/testing conditions — time-related processes might be recruited in support of performance on time-based tasks.

∗ University of British Columbia Department of Psychology, Vancouver, BC, V6T 1Z4, Canada; e-mail: [email protected] † Université Laval École de Psychologie, Canada; e-mail: [email protected]

1

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

2

P. Graf & S. Grondin

This chapter is a true team effort that was motivated by the desire to discover and delineate cognitive processes that are involved in both time perception and time-based ProM, and by the hope that it will lay the foundation for new collaborative research between these domains.

Time Perception What are the major empirical and theoretical questions that motivate research on psychological time and time perception? To answer this question, we begin this chapter section with some observations on conceptual and method issues related to research on time perception. Next, we describe the dominant theoretical model of time perception, the internal-clock model, focusing especially on a recent information-processing version of it. Then, we use this model to guide the presentation of significant findings that have emerged from recent research, including from research on age related changes on temporal judgments.

Conceptual and Method Issues The study of memory involves a retrospective component and a prospective component, and similarly, the study of time involves two components or research areas, one concerned with retrospective timing and the other with prospective timing. The distinction between prospective and retrospective timing concerns, respectively, situations where subjects/observers are informed in advance that they will have to make a time-related judgment versus situations where subjects/observers receive no prior warning about the need to make a time/duration judgment.1 In memory research, a vastly greater number of empirical and theoretical investigations have focused on retrospective memory than on ProM. By contrast, it is prospective timing that has received the most attention from time perception researchers in the past 30 years. Generally speaking, timing models developed to account for prospective judgments attempt to capture two fundamental features of temporal performance. One is related to the accuracy, or validity, of the time estimates provided by subjects, that is, it asks how closely related to physical time is subjective or perceived time. The other feature of performance concerns the variability of the perceived time estimates that have been obtained from a large number of trials. Research is often centered on one or both or these two aspects of performance (i.e. accuracy and variability). The dependent variables used when addressing

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

3

specific questions about accuracy or variability tend to be given different names, depending on the experimental method that is employed (e.g. verbal estimates, categorization, production and reproductions of intervals) as well as on the index adopted for expressing variability. The classical emphasis of time perception research has concerned the analysis of the ratio of the variability of estimated time to physical time (Weber fraction) or to the mean of the time-estimates (coefficient of variation). Throughout the history of research on the psychology of time, a number of different independent variables have been targeted.2 The most prominent among these are: the duration (length of time interval) under investigation,3 the sensory modality used for marking time,4 the nature of the cognitive demands made on subjects during an interval to be estimated, and the influence of participants’ age.5 Below, we will briefly review the research that bears either directly or indirectly on the last of these variables.

Theoretical Models Some theoretical models of psychological time are based on the concept of a clock process, but others do not presuppose this type of construct.3 Investigations of retrospective timing have been led by researchers with a traditional cognitive background. They held the view that subjective time is mediated by cognitive mechanisms. One classical example is Ornstein’s model,6 which deals with intervals longer than 10 s. This model postulates that the amount of storage space that needs to be allocated in memory for the purpose of estimating time varies directly with subjective duration. The availability of memory storage space is assumed to be determined by the number and complexity of stimuli to be processed during a given time period. By contrast Block and Reed7 and Zakay and Block8 argued that it is the number of contextual changes encoded into memory that determines the retrospective impression of duration. A large number of different theoretical models have been proposed to account for subjects’ performance on prospective timing tasks. Some models rely strictly on cognitive concepts without assuming the existence of a clock. For instance, Thomas and Weaver9 describe time estimation in terms of an attention-based model. They assume that the number of stimuli to be processed during a given time period is inversely correlated with subjective duration because increasing attention to these stimuli leaves temporal processing with fewer, and possibly insufficient, attentional resources.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

4

P. Graf & S. Grondin

A very different way of thinking about time perception was introduced by M. Jones and collaborators (see Chap. 3 for a more complete description of their theoretical approach and how it may apply to prospective timing).10,11 Jones and Boltz proposed a dynamic attending model.12 In the context of ProM, this model is most interesting because it emphasizes the fact that sensitivity to the occurrence of future events might depend on the properties of past events. The occurrence of physical regularities within the flow of events in the environment is assumed to mark non-arbitrary (or coherent) beginnings and endings of several succeeding time spans which offer temporal predictability for forthcoming events. This predictability sets within an observer an attending attitude called a future-oriented attending mode. The accuracy of temporal judgments was assumed to depend on temporal coherence and on the capacity to synchronize the internal rhythmicity of attending, called attunement, with the appropriate external rhythm afforded by the environment. When sequences of events in the environment do not provide temporal coherence, an observer is forced to adopt an internal strategy, called an analytic attending mode, for dealing with such unpredictable event occurrences. Before turning to the description of probably the most popular version of an internal clock, the pacemaker-accumulator device, the reader should note that animal timing and neuroscience offer many other timing models. For instance, Staddon and Higa proposed a pacemaker-free model where a cascade of interval timers is assumed to exist and where memory-strength decay determines specific time periods.13,14 And most popular are pacemaker-free models that emphasize a neural network description or some oscillatory process.15–17 A Pacemaker-Accumulator Device. A long tradition in research on time perception has proceeded on the assumption that prospective timing is mediated by a unique or dedicated internal clock. This clock, often described as a pacemakercounter or pacemaker-accumulator device,18 is at the foundation of many theoretical models.3 In general, these models assume that the pacemaker emits pulses that are accumulated in a counter, and the number of pulses that have been counted determines the perceived length of an interval. According to this type of model, how does one explain the occurrence of errors in judging time? One central cause of error is often assumed to be the reliability of the pacemaker, i.e. errors are thought to be a property of the pulse emitter device. The mode of pulse distribution can be deterministic or stochastic, and the pacemaker rate of responding/signaling over a long time period may be fixed or variable. Differences in models are related to properties of the

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

5

pacemaker.3 Errors in timing may also occur because of variability in the latency of the onset or offset stages of responding, and in matching the internal signals with the physical dimensions of the intervals to be judged.4 This source of error is more likely to have a small impact when intervals to be timed are relatively brief. Other properties of the counter might also be a major source of timing error. Killeen and Taylor proposed the existence of a cascade of counters. If counting is hierarchical, as it is when decimal or binary systems are used, dropped counts can become increasingly costly when larger numbers are counted.19 Killeen and Taylor noted that there should be a disproportionate error in timing each time the next stage in the counter must be set. These authors have demonstrated that the mean count registered should grow approximately as a power function of the duration of the to-be-timed interval. An Information-Processing Theory. Probably the most frequently cited contemporary theoretical account that builds on the idea of a pacemakeraccumulator device is called the Scalar Expectancy Theory (SET).20,21 Although it was developed primarily in order to explain animal timing data, this theory has been successfully applied to human time perception.22 One very important feature of SET is that it acknowledges that sources of variance other than at the clock level exert a major influence on temporal performances.23 The pacemakercounter device is embedded in a larger information processing system, and thus is subject to errors that may be caused not only by the clock processes described above, but also by memory and decisional processes (see Chap. 2 for a discussion of the latter processes). In this version of the clock, the accumulation of the pacemaker’s pulses into the counter is reported to be under the control of a switch mechanism, whose functioning is influenced by the amount of attention devoted to time processing. SET has two fundamental properties. The first is that the mean representation of time for a series of temporal judgments equals real time. In other words, in the long run, subject produced estimates of target durations converge on the actual duration of targets. The second critical feature of SET is that the variability — often expressed as a one standard deviation unit — of time estimates or judgments increases linearly with the mean representation of time. The constant proportion between variability and the mean is said to be scalar, which is essentially known in psychophysics as Weber’s law (i.e. the ratio of variability to mean time is a Weber fraction). The availability of attentional resources is assumed to have a critical influence on the functioning of the switch mechanism.24–26 Its role is central in accounting

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

6

P. Graf & S. Grondin

for the variability of time estimates and is most commonly invoked in order to explain the findings of investigations on perceived duration.27,28 The dual-task strategy, classical in cognitive psychology, has been employed in multiple timing experiments. This strategy builds on the assumption that attention is a limitedcapacity system. Therefore, if two tasks need to be carried out simultaneously, less attention will be available for each one. Brown and West showed that, in conditions where subjects were required to process multiple sources of temporal information, increasing the number of sources that had to be attended decreased the accuracy of timing.29,30 Somewhat along the same lines, the critical influence of attention on temporal information processing during the interval to be timed was demonstrated by Macar, Grondin and Casini. Their procedure was based on that used for analyzing attention-operating characteristics. Before each trial, a participant is asked to allocate a percentage of attention to each of two tasks to be performed simultaneously: a temporal task, which is to discriminate the length of the sensory signal, and a non-temporal task, which is to discriminate the intensity of the signal. When more attention was allocated to the temporal task, perceived duration was longer and better performance was observed in duration discrimination.31,32 These attentional effects can be readily accommodated by a pacemakeraccumulator model, if we assume the existence of a switch component that determines the access of pulses to the accumulator. The switch would be under the control of attention, with less attention to time resulting in a smaller transmission of pulses and in more variability.

Time Perception and Aging Although there has been a great deal of research on time perception, only a relatively small number of investigations have focused on how this high-level ability (i.e. time perception) is affected by aging, and consequently, many important questions remain unanswered, especially questions on age-related changes in the variability of time estimates.5,33 Because timing is so central to many simple tasks that need to be performed everyday (e.g. driving a car, carrying out a series of planned tasks), and because timing, as noted above, is so closely linked to memory and attention mechanisms (both of which are known to decline with aging), this section is dedicated to research on aging and time perception. Overall, aging is accompanied by a decrease in the accuracy of estimating time, but this decrease appears to depend on the method adopted for conducting

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

7

the investigation as well as on the range of the to-be-timed durations that are under scrutiny. For very brief intervals (circa 50 ms), Rammsayer, Lima and Vogel reported no difference between age groups (mean ages = 25.1, 45.5 and 64.6 years old) in the ability to discriminate the relative duration of intervals marked by two brief auditory signals.34 The mean difference threshold was about 17 ms in this experiment. However, Rammsayer reported that the discrimination of intervals of 1 s duration was poorer by older adults (70.4 years old) than by younger adults.35 As well, in the same study, the reproduction of 1 s intervals, but not that of 15 s intervals, by older adults was longer than the reproductions by young adults. In a task that required subjects to categorize a series of six tones on the basis of their duration (from 250 to 622 ms or from 622 to 1548 ms), McCormack, Brown, Maylor, Richardson and Darby reported that older adults (74.1 years old) performed significantly worse than young adults (19.5 years old).36 The same authors reported that for a similar task involving 9 tones varying from 250 to 2039 ms, older adults (70.5 years old) made fewer correct responses than young adults, and the pattern of errors was different between the groups. When the pitch of nine tones had to be categorized, older adults (68.7 years old) made fewer correct responses than young adults, but both groups showed a similar pattern of errors. Based on these data, McCormack et al. concluded that older adults have a distorted memory representation for duration information. The effects of age have also been examined in the categorization of intervals lasting between 3 to 6 s, and marked by auditory or visual signals.37 In this experiment, the level of attention was manipulated: there were trials with only one stimulus presented in either modality (i.e. in the full attention condition), and trials where two stimuli, of different lengths and different modalities, were presented (i.e. in the divided attention condition). This manipulation was conducted in the morning (9 am) for half of the participants, and in the afternoon (4 pm) for the other half. The older adults (69.3 years old) showed larger effects due to the modality and attention manipulation than the young adults (20.1 years old): visually marked intervals were perceived as much shorter than auditory marked intervals, and sensitivity to time decreased in the divided attention condition. Moreover, sensitivity to time was higher and the modality effect was smaller when testing occurred in the afternoon rather than in the morning in both age groups, except that in the full attention condition, older adults tested in the morning showed better sensitivity to time for intervals marked by visual signals.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

8

P. Graf & S. Grondin

In a task involving the reproduction of intervals lasting 6, 8 or 10 s, Vanneste and Pouthas showed that older adults (65.3 years old) were more sensitive than younger adults (20.2 years old) to a manipulation of attention to time.38 The participants had to pay attention to one, two or three stimuli that marked time. The results showed that in the 2- and 3-target conditions, the older participants were more variable and less accurate than the younger ones. This age-related effect on timing performance might be due to an age-related reduction of attentional resources. However, in a study by Craik and Hay that employed intervals of 30, 60 or 120 s during which subjects were occupied with a perceptual judgment task, older participants (72.2 years old) gave shorter verbal estimates and produced longer intervals than younger adults (22.2 years old), but the level of complexity of the task had a negligible effect on performance.39 As underscored by this brief review of the relevant research, it is difficult to reach compelling conclusions about the effect of aging on temporal performance. Nevertheless, for very brief intervals (<100 ms), it appears that the clock function is intact but when judging longer intervals (up to 10 s), experimental manipulations that focus on attentional resources and on memory processes are likely to show larger effects in groups composed of older than younger participants.

Time-Based Prospective Memory Prospective memory (ProMa ) is the ability to formulate plans and intentions, to retain them, and to execute them upon the occurrence of the appropriate cues.40,41 Like retrospective memory, ProM covers a vast domain. To navigate this domain and to facilitate communication, researchers have distinguished among different types of tests or situations that require the use of the ProM function. The most prominent of these distinctions has focused on how the retrieval of a previously formed plan or intention is cued or triggered.42,43 Tests or situations where retrieval of a plan is signaled by a specific event (e.g. the occurrence of a specific sensory stimulus, the completion of a specified activity)

a We

use the label ProM rather than PM for a number of different reasons, most importantly, because in mainstream memory research (i.e. a research area dominated by questions about retrospective memory), PM has long been used as an abbreviation for Primary Memory. In addition, we eschew using PM because future work may show that primary memory is a critical cognitive function that is required for both retrospective and prospective memory tasks.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

9

are said to be event-based or event-cued tests. By contrast, time-based or timecued tests are those where plan retrieval is signaled by a specified clock time (e.g. meet me today at 5 pm) or by a specified amount of elapsed time (e.g. call me back in 20 min). The focus of this section is on the latter tests. Although the label time-based evokes a direct link between time and memory processes, a review of the literature indicates that this link has not been a major focus of research on time-based ProM or on how this aspect of memory changes across the adult lifespan. Instead, the bulk of time-based ProM research has targeted other factors, most notably, the role of attention-resources and how changes in their availability across the adult lifespan affects overall task success rate under different conditions of testing. Our objective in this chapter section is, first, to review the research on age-related changes on time-based ProM tasks, second, to highlight factors that might have prevented or at least discouraged investigations of time-related processes in this domain, and third, possibly to identify a sub-domain of time-based ProM where future research may be more successful in revealing an important link between time and memory processes.

Basic Properties of Time-Based Prospective Memory By far the best-known experiment on time-based ProM is the classical 1985 study by Ceci and Bronfenbrenner.44 Although this study did not focus on agerelated performance changes, it is a convenient vehicle for portraying the general method that is typically used for investigating time-based ProM, for previewing the pattern of age-related changes that tend to occur in these investigations, and for introducing the basic questions that continue to motivate this kind of research. Ceci and Bronfenbrenner’s experiment explored the development of timebased ProM in 10- and 14-year old children. For one part of their first experiment, the children’s main task was to remove cupcakes from the oven, after a delay of 30 minutes, when they would be properly baked. The children had to carry out this task either in the familiar context of their own home or in the unfamiliar context of a typical psychology research laboratory. While waiting for the cupcakes to be baked, the children played a popular video game in an adjoining room that was furnished with a clock they could use for monitoring time. Of interest to the experimenters was to find out, first, whether or not the children would remove the cupcakes from the oven at the appropriate time (i.e. will he/she succeed in carrying out the task according to the instructions),

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

10

P. Graf & S. Grondin

and second, how often — according to what schedule — they would check the clock during the retention interval (i.e. the 30 min required for baking the cupcakes). The main findings reported by Ceci and Bronfenbrenner were that in the unfamiliar laboratory context, only one child failed to remove the cupcakes from the oven at the appropriate time whereas 42% of the children either failed or were late on this task when they had to do it at home. There was also a clear and surprising developmental effect on the baking-task success rate. In the familiar home context, failures were three times more common among the 10-year olds than among the 14-year olds. In addition, as highlighted by the results in Fig. 1, the children used different schedules for clock checking in the home and laboratory contexts. When tested in the context of their own homes children showed a U-shaped pattern of clock checking, whereas in the laboratory the frequency of clock checking started low, but then increased steadily with the approach of the end of the baking period. Overall, the children checked the clock more often in the laboratory context, where task success rate was higher (at the ceiling), than in the home context. However, the number of clock checks did not predict task success rate, as evidenced by the finding that in the home context, the younger children made more clock checks than the older children and yet task success rate was higher in the older children. As emphasized by Ceci and Bronfenbrenner, task success is predicted not by the number of clock checks but by their effective and strategic allocation toward the end of the baking period. To the extent that clock-checking is efficient and skillful (i.e. allocated in the most informative manner), it leaves more time and attention resources that can be harnessed for other ongoing activities, such as the video game the children were playing while waiting for the cupcakes to be baked. Ceci and Bronfenbrenner’s44 main focus was on the childhood development of clock checking strategies, and on the contextual determinants of these strategies. Similar questions about clock checking strategies have been targeted by a small number of more recent investigations (reviewed by Mäntylä and Carelli in Chap. 8 of this volume), but the vast majority of them have focused on other questions about age-related changes in time-based ProM. The most important of these other questions concerns the overall rate of task successes and failures rather than clock checking strategies; it asks whether aging has a larger or smaller effect on time- versus event-based ProM tasks. The second most often asked question, following on the heels of Ceci and Bronfenbrenner’s44 work, asks about contextual influences on ProM task performance.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

11

Mean Number of Clock Checks

4 3.5

14-year olds 3 2.5

10-year olds 2 1.5 1 0.5 0 1

2

3

4

5

6

5

6

Home Context

Mean Number of Clock Checks

4

14-year olds 3.5 3

10-year olds 2.5 2 1.5 1 0.5 0 1

2

3

4

Laboratory Context

Fig. 1. The figure, adapted from Ceci and Bronfenbrenner,44 shows the mean number of clock checks during each 5-minute period of the retention interval. The figure highlights the different schedules of clock checking used by children who were tested in the context of their own homes versus in the laboratory, as well as the fact that the younger children checked the clock more often than the older children.

Age-Related Changes in Time-Based Prospective Memory One of the foundational assumptions implicit in Ceci and Bronfenbrenner’s44 work is the notion that strategic (e.g. skillful, calculated) clock checking is adaptive, that it increases ProM task success rate, requires less frequent clock checking and thereby releases time and resources for other simultaneously occurring activities. This basic assumption is consistent with the broad theoretical claims that have motivated the vast majority of recent investigations of age-related changes in time-based ProM. One of these claims stems from

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

12

P. Graf & S. Grondin

Craik’s45 influential proposal that all remembering tests or situations can be arranged along a continuum that marks the extent to which retrieval performance depends on self-initiated, subjects controlled or attention-resource demanding processes as opposed to being dependent on environmentally-driven, automatic processes. Craik further proposed that of all memory tests, performance on ProM tests is most dependent on self-initiated, attention-resource demanding processes. Therefore, consistent with the widespread view that aging is accompanied by a decline in the attention-resources that are available for processing informatiom,46–49 he predicted that age-related changes in performance would be larger on ProM tests than on any other kind of memory tests. For a number of reasons to be discussed later in this chapter, Einstein and McDaniel43,50 went one step beyond Craik and argued that performance on time-based ProM tests is more dependent on self-initiated resource demanding processes than performance on event-based tasks, and consequently, they predicted that age-related performance declines would be larger and more common on the former than latter tests. The pattern of results predicted by Einstein and McDaniel has been found, for example, by Einstein, McDaniel, Richardson, Guynn and Cunfer51 and by Park, Hertzog, Kidder, Morrell and Mayhorn.52 However, a number of other studies have shown that older adults consistently tend to outperform their younger counterparts on time-based ProM tasks, but only when testing occurs in the context of their everyday life.53–58 The results of a meta-analysis by Birt40 highlight these different patterns of age-related changes in performance on ProM tests. She sampled a total of 25 different articles published in peer reviewed journals that reported on 34 different experiments. Altogether, these experiments included a total of 2,695 different participants, and Birt was able to compute 96 different age effect sizes (Cohen’s d), each defined according to Cohen59,60 as the difference between the mean performance for the young group and for the old group divided by the pooled population standard deviation. Of the 96 effect sizes, 71 (74%) came from experiments that focused on event-based ProM and 25 (26%) came from experiments on time-based ProM. By contrast to the prediction of Einstein and McDaniel,43 the overall results of the analysis showed larger age effects on event-based tasks than on time-based tasks, with mean weighted effect sizes of d = 0.59 and d = −0.05, respectively. This general summary of the results is misleading, however, because it ignores the context in which memory performance was assessed. In part

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

13

motivated by the classic study of Ceci and Bronfenbrenner,44 time- and eventbased ProM have been investigated under naturalistic field conditions, that is, in the context of subjects’ familiar home and everyday life, as well as under the relatively unfamiliar and artificial conditions that prevail in a typical psychology research laboratory. Birt’s40 meta-analysis included, for time-based ProM, 12 age-effect sizes from data that were collected under artificial laboratory conditions and 13 effect sizes based on data obtained under naturalistic field conditions, and for event-based ProM, 39 and 6 age-effect sizes came from laboratory and naturalistic field studies, respectively. The results of the metaanalysis, summarized in Table 1, showed a larger mean age-effect for time- than event-based tasks for the laboratory studies (note, however, there is overlap in the 95% confidence intervals). This outcome is consistent with the prediction by Einstein and McDaniel.43 However, in stark contrast to this prediction, the results from the field studies show a reversed age-effect, that is, compelling evidence that older adults are more successful than young adults on time-based tasks if those tasks had to be performed in the context of their everyday life. The finding that age-effects on time-based ProM tasks are dependent on contexts raises intriguing questions about the factors that determine performance. To explain their findings, Ceci and Bronfenbrenner44 assumed that their context manipulation affected performance because young children are more familiar with the home than laboratory environment, and consequently, they were more anxious in the laboratory. They further speculated that the heightened anxiety level may have led the children to check the clock more frequently, thereby increasing task success rate. Another possibility is that Ceci and Bronfenbrenner’s children participants viewed the ProM task as more important, relative

Table 1. The table shows age-effect sizes (Cohen’s d means and 95% confidence intervals) obtained on event- and time-based ProM tasks that had to be carried out either under the relatively artificial conditions of a typical psychology research laboratory or under the naturalistic field conditions of subjects’ own home and every day life. The data are adapted from the PhD dissertation of Angela Birt.40 Event-Based Tasks STUDY TYPE Laboratory Field

Time-Based Tasks

Mean d

95% CI

Mean d

95% CI

0.7 0.34

0.62/0.79 0.16/0.53

0.99 −0.6

0.76/1.21 −0.77/−0.43

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

14

P. Graf & S. Grondin

to playing the video game, in the laboratory context than in the home context. Previous work has shown that performance is higher on ProM tasks that are designated (i.e. by the experimenter) or self-evaluated (i.e. by participants) as more important.61 The results in Table 1 do not show the overall level of performance in the different contexts (instead, they focus on age-related performance effects that occur in different contexts) and thus are not directly comparable to the context effects reported by Ceci and Bronfenbrenner.44 Nevertheless, consistent with Ceci and Bronfenbrenner, it might be argued that the findings in Table 1 are, at least in part, a reflection of subjects’ degree of familiarity with the contexts of laboratory and field studies. The young adults who typically participate in memory experiments tend to be undergraduate students, whereas the older participants tend to be community living individuals, and thus, it seems likely that the former would be more familiar with psychology laboratories and more at ease with the personnel and equipment that are involved in laboratory studies. However, rather than emphasizing how context familiarity might affect subjects’ level of anxiety, as did Ceci and Bronfenbrenner,44 the researchers who collected the evidence summarized in Table 1 tended to focus on the contextdependent availability and use of strategic knowledge. They have noted that older adults, to a greater extent than younger adults, have spent a lifetime developing specific skills and reminder systems for time-based ProM tasks, and the use of this strategic knowledge facilitates their task performance.55,62 We can understand older adults’ poor performance under laboratory conditions on the additional assumption that their reminder systems are context specific (e.g. connected with the context of home), and not easily tuned to or adapted for the relatively stark, barren environment of the typical research laboratory. Moreover, it may be that the very attempt to employ their well-practiced reminder system prevents the older participants from discovering or relying on the more abstract, context-independent clock-checking strategies that may be more appropriate in the laboratory context. In order to decide among such possibilities, future research will need to explore age-related changes in the strategies used under different testing conditions, as well as the factors that trigger the deployment of different strategies (e.g. testing contexts, the availability of resources). The finding of a typical age-related decline in time-based ProM task performance under laboratory conditions versus a reversed age-effect under field conditions is well known, but the debate about what causes this age-effect pattern continues. What is more important for this chapter is that even though the

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

15

findings in Table 1 have focused attention on the strategies that are engaged for time-based ProM tasks under different testing conditions by subjects from different age groups, only a very small number of investigations have attempted to explore these strategies (for a review, see Chap. 8 of this volume by Mäntylä and Carelli). And despite the celebrated work by Ceci and Bronfenbrenner,44 even fewer studies have examined age-related changes in clock-checking strategies and in time-related processes. What accounts for the dearth of research on these issues? We believe the answer to this question is intimately connected with the very nature of timebased ProM, with what is unique about this memory function, and with how it tends to be employed. In the remaining parts of this chapter, we shall attempt to identify what is special about time-based ProM. Specifically, we shall argue that time-based ProM is composed of several distinguishable components or functions, and argue that clock-checking strategies and time-related processes are likely to be critically involved in only some of them.

What, if Anything, is Special about Time-Based Prospective Memory? The quick and most common answer to this question is that time-based ProM tasks are different from event-based tasks by virtue of the fact that while specific external cues are available to signal when a previously formed plan is to be retrieved for the latter tasks, no external cues are provided for the former tasks. However, this basis for distinguishing between tasks is readily dismissed by the argument that a particular clock reading (e.g. when both hands of the clock point straight up, when the number 12 appears on the face of a digital clock) is as much an external cue as, for example, the appearance of a colleague in the hallway. Moreover, there is considerable empirical evidence showing that not all retrieval cues are equally effective.43,63–65 Cues that are larger or louder are more effective than cues that are smaller or softer. Cues that are presented in the visual-fovea are more effective than peripherally presented cues. Cues that are perceptually distinctive are more effective than those that are non-distinctive. Consistent with this type of evidence, therefore, it may be that the specific cues used in connection with time- and event-based prospective tasks differ in terms of potentially important perceptual properties (e.g. intrusiveness, loudness). If so, future research ought to explore the nature of these properties and their specific role in determining performance on time- and event-based ProM tasks.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

16

P. Graf & S. Grondin

More important, however, this line of reasoning suggests that the difference between time- and event-based prospective tasks is not defined or definable by the presence versus absence of specific retrieval cues. A more compelling case in support of a distinction between time- and eventbased ProM tasks focuses on the predictability or calculable-proximity of cues or situations that signal when it is appropriate or necessary to retrieve a previously formed plan. To highlight this potentially important feature of time- and eventbased ProM tasks, consider the following two examples: making a phone call tomorrow at 9 am and giving a message to a colleague on the next encounter with him/her. For the first task, it is possible at any time during the retention interval to compute and therefore know how close we are to the situation where a plan needs to be retrieved and executed. However, we may have no idea when and where the next encounter with a colleague might occur. Therefore, consistent with these examples, it is possible that performance differences between timeand event-based ProM tasks occur because only the former provide warning signals, an opportunity for monitoring the approach or proximity of retrieval cues, and consequently, for suspending competing attentional demands and for preparing to execute a planned activity. However, by contrast to the core assumption implicit in the foregoing paragraph, many events such as the next encounter with a colleague or the location of the supermarket on our way home are reasonably predictable and thus permit a sort of proximity calculation. The power to make predictions in these cases comes from knowledge of our colleague’s habits, familiarity with the location of the supermarket in our neighborhood, perhaps even shared cultural knowledge about the typical location of mailboxes on our city streets. Consequently, to the extent that at least some event-based tasks permit some type of cue proximity calculations, the core difference between time- and event-based ProM tasks is not the presence versus absence of retrieval cue predictability. Rather, the critical difference may be the nature of the dimension (e.g. time versus memory for habits or for spatial location) that is available for making cue proximity calculations. It is possible that the between-task difference is the relative prominence of the dimension(s) available for making proximity calculations (note: the time dimension is highlighted by the description of time-based tasks whereas the description of event-based tasks does not identify possible dimensions for making predictions and thus relevant dimensions need to be inferred or discovered), the reliability of the information provided by each dimension, or the users’ familiarity with each dimension. To our knowledge, to date there has

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

17

been no systematic investigation either of the possibility that time- and eventbased tasks implicate different dimensions for estimating the proximity of plan retrieval cues, or of any other such variable that might be used to define the difference between the two task types (i.e. time- and event-based tasks). An additional factor that complicates efforts to define the difference between time- and event-based tasks stems from the fact that subjects may translate or transform one type of task into the other. Qualitative evidence for this type of translation comes from interviews on the strategies subjects employ for different types of ProM tasks.66 When required to describe the strategies they would use for a typical time-based task, such as a doctor’s appointment at 2 pm on Thursday, subjects tend to link the planned task with other activities or events scheduled for that day (e.g. I plan to leave immediately after my yoga class). To the extent that subjects engage in this type of translation activity, the nominal difference between time- and event-based tasks ceases to exist. Recent work by Cook, Marsh and Hicks67 showed that when this type of timeevent linking occurs, it may serve either to facilitate or inhibit performance on the target task. Cook et al. conducted an experiment in which subjects were required to make pleasantness ratings about words in Phase 1, complete a demographics questionnaire in Phase 2 and carry out a syllable counting task in Phase 3. In addition to completing these tasks, subjects were also required to carry out a nominal time-based ProM task, specifically to press a target key on the keyboard after a delay of 6 minutes. Between subjects, Cook et al. manipulated the length of Phase 1 such that it took about 3.5 min in one condition compared to 7 minutes in the other condition. More importantly, they also manipulated subjects’ expectations about when the time-based prospective task would have to be carried out. They told half of their subjects that the prospective task would most likely have to be executed in Phase 3 of the experiment, during the syllable counting tasks; the remaining half of the subjects were not provided this additional information. Because of the manipulation of the length of Phase 1, the additional information given to half of the subjects was valid for those who received the short Phase 1 task, but it was misleading for those who received the long Phase 1 task. The data in Fig. 2, adapted from a table in Cook, Marsh and Hicks,67 show that under control conditions when the subjects did not have any specific expectations about the context in which they had to carry out the planned task, their performance was about the same (∼52%) with the short and long Phase 1 task, that is, performance was not significantly affected by the nature of the context

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

18

P. Graf & S. Grondin

Task Performance

1 0.8 0.6 0.4 0.2 Phase 1 Short

Phase 1 Long

No Context Expectation

Phase 1 Short

Phase 1 Long

Specific Context Expectation

Fig. 2. The bars show the mean proportions of successful performance on a time-based ProM task under conditions where subjects either did or did not have a specific expectations about the context in which a planned task had to be executed (figure constructed on the basis of data reported in Cook, Marsh and Hicks67 ).

in which the planned task had to be carried out (i.e. whether the ongoing task required making pleasantness ratings or counting syllables). By contrast, in the conditions where subjects did have a specific expectation about the context where the planned task would need to be executed, performance was facilitated when subjects’ expectations were valid (i.e. if the task had to be executed in the context where they expected it), but it was impaired when their expectations were invalid. Until these findings have been replicated, it is premature to draw strong inferences from them. However, they do suggest that when both time and event cues are available for a ProM task, performance is more strongly influenced by event cues. What lessons follow from these reflections on the similarities and difference(s) between time- and event-based ProM tasks? In the foregoing paragraphs, we have acknowledged potential differences between the two task types. However, we have emphasized the similarities between them, and suggested that especially in the rich context of everyday life, the nominal differences between time- and event-based tasks might be minimized or absent because of the specific strategies subjects employ to manage commitments to future plans and intentions within the regular and predictable circumstances of their daily life. Consistent with this suggestion, we encourage more investigations of the different strategies subjects employ for time- and event-based ProM tasks, especially

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

19

investigations of the factors that cause the age-effect pattern of performance shown by the results in Table 1. One obvious implication of the suggestion that the differences between timeand event-based ProM tasks are minimized by subjects’ reliance on knowledge about the regular and predictable circumstances of their daily life is that differences between these task types would be more evident, magnified, when testing occurs in the barren, unfamiliar and unpredictable context of the psychology research laboratory. It is tempting to regard the evidence in Table 1 as supporting this possibility, but this use and interpretation of the data would be inappropriate. To our knowledge, only one study, conducted by Logie, Maylor, Della Sala and Smith68 has directly compared time- and event-based ProM task performance under the exact same controlled laboratory conditions (i.e. in the absence of obvious confounding variables), but the results of this study are marred by performance ceiling effects in several conditions. The final lesson we draw from our reflections on the similarities and difference(s) between time- and event-based ProM tasks is that even in the barren context of the research laboratory, clock-checking strategies and time-related processes are more likely to be critically involved when the retention interval is relatively short. Consistent with the research finding that time-related processes are attention demanding, it seems that extensive reliance on these kinds of processes would be adaptive only for activities to be carried out in the immediate future. When the retention interval is longer than a few minutes, it would most likely be filled by other activities (e.g. I might check my email if I have 5 min before the next meeting), and our experience-driven ability to estimate the duration of these activities would provide an alternative basis for calculating/monitoring when to execute a planned task. It seems likely that the impact on task performance due to clock-checking and time-related processes would be reduced by the extent to which subjects rely on such alternative monitoring strategies. The suggestion that different processes might mediate performance on timebased tasks with short versus long retention intervals is consistent with the proposal that ProM encompasses a number of functionally distinct components,42 the most prominent of which are: monitoring, episodic ProM and habitual ProM. The monitoring function, analogous to retrospective’s short-term or working memory, is engaged for short-term tasks such as pressing the record button when a movie resumes or turning off the kettle after the water has come to a boil. Distinct about tasks that require monitoring is the fact that a plan, once

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

20

P. Graf & S. Grondin

formed, remains active and perhaps dominant in consciousness throughout the retention interval. Episodic ProM41,69 is analogous to episodic retrospective memory, and it differs from monitoring primarily because plans and intentions do not remain active in consciousness through the retention interval. Remembering to stop for groceries on the way home from work or to convey a message to a colleague the next time we meet are everyday examples of episodic ProM tasks. Finally, episodic ProM seems different from habitual ProM primarily because the former function is engaged for one-off situations, whereas the latter is used for repeated tasks, for example, for taking medication according to a prescribed schedule.42 Future research may reveal that clock-checking and time-related processes are critical for performance in situations that require monitoring, but not in situations that involve episodic ProM or habitual ProM.

Conclusions Do time-based ProM tasks involve a timing mechanism that exerts a significant influence on performance? Is this mechanism the kind of pacemaker-counter device that has been targeted by research on time perception? Are age-related changes in performance of time-based ProM tasks caused by changes in memory processes, in timing processes or in both? Current research on time-based ProM and on age-related changes in time-based ProM does not provide clear answers to such questions, mostly because by and large, they have not yet been the focus of systematic investigation. In this chapter, we underscored the fact that the words time-based are nothing but a label, a common way of describing the manner in which the retrieval of previously formed plans and promises is signaled for some ProM tasks. We have argued that although this label suggests that time-based ProM involves timerelated processes, these processes are likely to be implicated only for some special purposes, specifically, for situations that require monitoring. Monitoring is the short-term function of ProM that is analogous to the short-term component of retrospective memory. It may be that in monitoring situations, the most effective strategy is one that involves clock-checking and depends on a pacemaker-counter device, either because alternative strategies are not available or their implementation is more resource demanding. The time-perception models reviewed earlier in this chapter seem compatible with the proposal that time-related processes are invoked for time-based monitoring tasks. Monitoring tasks have been described as dual task situations42 where attention needs to be allocated to at least two sources, for example, an

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

21

ongoing activity (e.g. cleaning up the office desk) and the flow of time (e.g. is it time to leave yet?). The claim that in time-based monitoring task subjects are aware of time is consistent with the pacemaker-counter models described earlier in this chapter as well as in the next chapter. More important, if time-based monitoring tasks require some type of pacemaker-counter, we expect future research to demonstrate a strong predictive link between performance on time-estimation tasks and performance on time-based monitoring tasks.

References 1. Hicks R, Miller G, Kinsbourne M. Prospective and retrospective judgments of time as a function of amount of information processed. American Journal of Psychology 1976; 89: 719–730. 2. Roeckelein J E. The Concept of Time in Psychology (A Resource Book and Annotated Bibliography). Westport, Ct: Greenwood, 2000. 3. Grondin S. From physical time to the first and second moments of psychological time. Psychological Bulletin 2001; 127: 22–44. 4. Grondin S. Sensory modalities and temporal processing. In Helfrich H, ed. Time and Mind II: Information Processing Perspectives. Göttingen, Germany: Hogrefe & Huber, 2003: 75–92. 5. Block R A, Zakay D, Hancock P A. Human aging and duration judgments: A meta-analytic review. Psychology and Aging 1998; 13: 584–596. 6. Ornstein R. On the Experience of Time. New York: Penguin, 1969. 7. Block R A, Reed M A. Remembered duration: Evidence for a contextual-change hypothesis. Journal of Experimental Psychology: Human Learning and Memory 1978; 4: 656–665. 8. Zakay D, Block R A. Prospective and retrospective duration judgments: An executive-control perspective. Acta Neurobiologiae Experimentalis 2004; 64: 319–328. 9. Thomas E A C, Weaver W B. Cognitive processing in time perception. Perception & Psychophysics 1975; 17: 363–367. 10. Barnes R, Jones M R. Expectancy, attention, and time. Cognitive Psychology 2000; 41: 254–311. 11. Large E W, Jones M R. The dynamics of attending: How we track time varying events. Psychological Review 1999; 106: 119–159. 12. Jones M R, Boltz M G. Dynamic attending and responses to time. Psychological Review 1989; 96: 459–491. 13. Staddon J E R, Higa J J. Time and memory: Towards a pacemaker-free theory of interval timing. Journal of the Experimental Analysis of Behavior 1999; 71: 215–251. 14. Block R A. Psychological timing without a timer: The roles of attention and memory. In Helfrich H, ed. Time and Mind II: Information Processing Perspectives. Göttingen, Germany: Hogrefe & Huber, 2003: 41–59. 15. Church R M, Broadbent H A. Alternative representations of time, number, and rate. Cognition 1990; 37: 55–81. 16. Grossberg S, Schmajuk N A. Neural dynamics of adaptive timing and temporal discrimination during associative learning. Neural Networks 1989; 2: 79–102.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

22

P. Graf & S. Grondin

17. Treisman M, Faulkner A, Naish P L N, Brogan D. The internal clock: Evidence for a temporal oscillation underlying time perception with some estimates of its characteristic frequency. Perception 1990; 19: 705–743. 18. Killeen P R, Weiss N A. Optimal timing and the Weber function. Psychological Review 1987; 94: 455–468. 19. Killeen P R, Taylor T. How the propagation of error through stochastic counters affects time discrimination and other psychophysical judgments. Psychological Review 2000; 107: 430–459. 20. Gibbon J. Ubiquity of scalar timing with a Poisson clock. Journal of Mathematical Psychology 1992; 36: 283–293. 21. Meck W H. (Ed.) Functional and Neural Mechanisms of Internal Timing. Boca Raton, Fl: CRC Press, 2003. 22. Wearden J. Applying the scalar timing model to human time psychology: Progress and challenges. In Helfrich H, ed. Time and Mind II: Information Processing Perspectives. Göttingen, Germany: Hogrefe & Huber, 2003: 21–39. 23. Gibbon J, Church R M. Sources of variance in an information processing theory of timing. In Roitblat H L, Bever T G, Terrace H S, eds. Animal Cognition. Hillsdale, NJ: Erlbaum, 1984: 465–488. 24. Brown S W, Boltz M. Attentional processes in time perception: Effects of mental workload and event structure. Journal of Experimental Psychology: Human Perception and Performance 2002; 28: 600–615. 25. Burle B, Casini L. Dissociation between activation and attention effects in time estimation: Implications for internal clock models. Journal of Experimental Psychology: Human Perception and Performance 2001; 27: 195–205. 26. Grondin S. Time psychophysics and attention. Psychologica 2001; 28: 177–191. 27. Block R A, Zakay D. Retrospective and prospective timing: Memory, attention and consciousness. In Hoerl C, McCormack T, eds. Time and Memory: Issues in Philosophy and Psychology. Oxford: Clarendon Press, 2001: 59–76. 28. Glicksohn J. Temporal cognition and the phenomenology of time: A multiplicative function for apparent duration. Consciousness and Cognition 2001; 10: 1–25. 29. Brown S W, Stubbs D A, West A N. Attention, multiple timing, and psychophysical scaling of temporal judgments. In Macar F, Pouthas V, Fiedman W, eds. Time, Action, Cognition: Towards Bridging the Gap. Dordrecht Netherlands: Kluwer, 1992: 129–140. 30. Brown S W, West A N. Multiple timing and the allocation of attention. Acta Psychologica 1990; 75: 103–121. 31. Macar F, Grondin S, Casini L. Controlled attention sharing time estimation. Memory & Cognition 1994; 22: 673–686. 32. Grondin S, Macar F. Dividing attention between temporal and nontemporal tasks: A performance operating characteristic — POC — analysis. In Macar F, Pouthas V, Friedman W, eds. Time, Action, Cognition: Towards Bridging the Gap. Dordrecht, Netherlands: Kluwer, 1992: 119–128. 33. Lustig C. Grandfather’s clock: Attention and interval timing in older adults. In Meck W H, ed. Functional and Neural Mechanisms of Internal Timing. Boca Raton, Fl: CRC, 2003: 261–293.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

Time Perception and Time-Based Prospective Memory

23

34. Rammsayer T H, Lima S D, Vogel W H. Aging and temporal discrimination of brief auditory intervals. Psychological Research 1993; 55: 15–19. 35. Rammsayer T H. Ageing and temporal processing of durations within the psychological present. European Journal of Cognitive Psychology 2001; 13: 549–565. 36. McCormack T, Brown G D A, Maylor E A, Richardson L B N, Darby E R J. Effect of aging on absolute identification of duration. Psychology and Aging 2002; 17: 363–378. 37. Lustig C, Meck W H. Paying attention to time as one gets older. Psychological Science 2001; 12: 478–484. 38. Vanneste S, Pouthas V. Timing and aging: The role of attention. Experimental Aging Research 1999; 25: 49–67. 39. Craik F I M, Hay J F. Aging and judgments of duration: Effects of task complexity and method of estimation. Perception & Psychophysics 1999; 61: 549–560. 40. Birt A R. Prospective memory: A distinct form of remembering? Evidence from task comparisons and normal aging. Unpublished doctoral dissertation, University of British Columbia, Vancouver, BC, Canada, 2001. 41. Graf P. Prospective memory retrieval revisited. In Ohta N, MacLeod C M, Uttl B, eds. Dynamic Cognitive Processes. Tokyo: Springer, 2005: 305–332. 42. Graf P, Uttl B. Prospective memory: A new focus for research. Consciousness and Cognition 2001; 10: 437–450. 43. Einstein G O, McDaniel M A. Normal aging and prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 1990; 16: 717–726. 44. Ceci S J, Bronfenbrenner U. Don’t forget to take the cupcakes out of the oven: Prospective memory, time-monitoring and context. Child Development 1985; 56: 152–164. 45. Craik F I M. A functional account of age differences in memory. In Klix F, Hagendorf H, eds. Human Memory and Cognitive Capabilities, Mechanisms, and Performances. Amsterdam: Elsevier, 1986: 409–422. 46. Hasher L, Zacks R T. Automatic and effortful processes in memory. Journal of Experimental Psychology: General 1979; 108: 356–388. 47. Jorm A F. Controlled and automatic information processing in senile dementia: A review. Psychological Medicine 1986; 16: 77–78. 48. Salthouse T A. A theory of cognitive aging. Amsterdam: North Holland, 1985. 49. Salthouse T A. Resource-reduction interpretations of cognitive aging. Developmental Review 1988; 8: 238–272. 50. Einstein G O, McDaniel M A. Retrieval processes in prospective memory: Theoretical approaches and some new empirical findings. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Application. Mahwah, NJ: Erlbaum, 1996: 115–141. 51. Einstein G O, McDaniel M A, Richardson S L, Guynn M J, Cunfer A R. Aging and prospective memory: Examining the influences of self-initiated retrieval processes. Journal of Experimental Psychology: Learning, Memory, and Cognition 1995; 21: 996–1007. 52. Park D, Hertzog C, Kidder D P, Morrell R W, Mayhorn C B. Effect of age on event-based and time-based prospective memory. Psychology and Aging 1997; 12: 314–327. 53. Martin M. Aging and patterns of change in everyday memory and cognition. Human Learning 1986; 5: 63–74.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch01

24

P. Graf & S. Grondin

54. Maylor E A. Age and prospective memory. Quarterly Journal of Experimental Psychology 1990; 42A: 471–493. 55. Moscovitch M. A neuropsychological approach to memory and perception in normal and pathological aging. In Craik F I M, Trehub S, eds. Aging and Cognitive Processes. New York, NY: Plenum, 1982: 55–78. 56. Patton G W, Meit M. Effect of aging on prospective and incidental memory. Experimental Aging Research 1993; 19: 165–176. 57. Rendell P G, Craik F I M. Virtual week and actual week: Age-related differences in prospective memory. Applied Cognitive Psychology 2000; 14: S43–S62. 58. West R L. Prospective memory and aging. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory: Current Research and Issues (Vol. 2) Chichester, England: Wiley, 1988: 199–125. 59. Cohen J. Statistical power analysis for the behavioral sciences (1st edn.). New York: Academic Press, 1969. 60. Cohen J. Statistical power analysis for the behavioral sciences (revised edn.). New York: Academic Press, 1977. 61. Kliegel M, Martin M, McDaniel M A, Einstein G O. Importance effects on performance in event-based prospective memory tasks. Memory 2004; 12: 553–561 62. Maylor E A. Does prospective memory change with age? In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Application. Mahwah, NJ: Erlbaum, 1996: 173–197. 63. Brandimonte M A, Passolunghi M C. The effect of cue-familiarity, cue-distinctiveness, and retention interval on prospective remembering. Quarterly Journal of Experimental Psychology 1994; 47A: 565–587. 64. Cohen A, Dixon R A, Lindsay D S, Masson M E J. The effect of perceptual distinctiveness on the prospective and retrospective components of prospective memory in young and old adults. Canadian Journal of Experimental Psychology 2003; 57: 274–289. 65. McDaniel M A, Einstein G O. The importance of cue familiarity and cue distinctiveness in prospective memory. Memory 1993; 1: 23–41. 66. Siu D, Graf P. Plans for success: Improving prospective memory task performance. Presented at the 2nd International Conference on Prospective Memory, Zurich, Switzerland, 2005. 67. Cook G I, Marsh R L, Hicks J L. Associating a time-based prospective memory task with an expected context can improve or impair intention completion. Applied Cognitive Psychology 2005; 19: 345–360. 68. Logie R H, Maylor E A, Della Sala S, Smith G. Working memory in event- and timebased prospective memory tasks: Effects of secondary demand and age. European Journal of Cognitive Psychology 2004; 16: 441–456. 69. Kvavilashvili L, Ellis J. Varieties of intentions: Some distinctions and classifications. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Application. Mahwah, NJ: Erlbaum, 1996: 23–51.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

2 Prospective Remembering Involves Time Estimation and Memory Processes Richard A. Block∗ and Dan Zakay†

Introduction Successful models of prospective rememberinga must be parsimonious, elegant, and plausible. In addition, they should not focus solely on prospective remembering per se. They should follow from well-established findings and principles concerning other kinds of memory and cognitive processes, such as those involved in retrospective remembering. Thus, we think that prospective remembering does not involve any special cognitive or memory systems. Instead, prospective remembering relies on the functioning of well-known attention and memory systems. We do not deny that some additional unique abilities may also be involved in prospective remembering.2,3 However, it is useful to begin with models that relate prospective remembering to findings and models that are well established in other, non-prospective (e.g. retrospective) memory situations. ∗ Department of Psychology Montana State University, PO Box 173440, Bozeman, MT 59717-3440, USA; e-mail: [email protected] † Department of Psychology Tel-Aviv University, Ramat-Aviv 69972, Israel; e-mail: [email protected] a We prefer the term prospective remembering because it suggests the dynamic nature of the processes involved (as in the term prospective timing). In addition, prospective remembering involves more than just memory.1

25

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

26

R. A. Block & D. Zakay

Researchers have mainly studied two types of prospective remembering situations, called time-based and event-based. One way to view this distinction is that time-based prospective remembering is more self-initiated, whereas event-based prospective remembering is more environmentally cued. Graf and Grondin (this volume) argued that this distinction is not very useful. In some naturalistic, time-based prospective remembering situations, a person has access to external chronometers, which may yield event-based cues (e.g. seeing a clock reading). Keeping this in mind, we find the distinction between time-based and event-based prospective remembering to be useful, and we explain why it is useful. We also discuss situations that may involve a mixture of time-based and event-based processes. Other researchers have discussed a third type of prospective remembering situation, called activity-based. Controversy remains about whether or not activitybased intentions are distinct from event-based intentions (see Kvavilashvili & Ellis4 for discussion). If one agrees that the distinction is needed, perhaps a slight elaboration of our model of event-based prospective remembering may also apply to activity-based prospective remembering. In this chapter, we review relevant research and theories on time-based and event-based prospective remembering. We propose and describe two models, one that explains time-based prospective remembering (the attentionalgate model) and one that explains event-based prospective remembering (the recursive-reminding model). We show how these models are able to account for some of the major findings in the literature, as well as to guide future research. A consideration of how the two kinds of processes may interact in some mixed time-based and event-based situations is also included. We conclude by mentioning a few unusual applications in altered states of consciousness.

Time-Based Prospective Remembering In a situation requiring time-based prospective remembering, a person forms a self-generated intention or is given (as by an experimenter) an other-generated intention to perform a specific action at a specific future time. The future time may be targeted either as a specific clock time (e.g. “at 8:30 today”) or as a specific interval (e.g. “5 minutes from now”). State-of-the art explanations of time-based prospective remembering have typically relied on older theories instead of more recent theories from the extensive literature on time estimation. For example, Cook, Marsh, and Hicks5 said that “there is no existing

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

27

theory about how time-based intentions are successfully completed, except one slightly dated theory based mainly on intuition” (p. 346): Harris and Wilkins’s6 test-wait-test-exit model, which is a variant of Miller, Galanter, and Pribram’s7 test-operate-test-exit model. Einstein and McDaniel8 also discussed this kind of model. It proposes a process in which the person loops through test-wait cycles until another test seems needed. When a test finally indicates a time at which it is appropriate to respond, the loop is exited, and the person makes the response. The test-wait-test-exit model is inadequate in that it does not explicitly address several questions: What is being tested? What happens while a person is waiting, and how does the person decide that another test is needed? How does any test indicate whether or not it is an appropriate time to respond? The fact that the test-wait-test-exit model is viewed as a state-of-the-art model reveals that most prospective remembering researchers are not aware of more recent time-estimation models, as well as that time-estimation researchers have not yet discussed the connections between their research and the topic of prospective remembering. In this chapter, we remedy these problems by proposing a more explicit model of time-based prospective remembering, the attentional-gate model. The attentional-gate model retains some of the features of the test-wait-test-exit model that have enabled that model successfully to explain extant research findings. In addition, the attentional-gate model adds some more explicitly described components that enable it to explain the same kinds of findings that the test-wait-test-exit model explains, but also to explain other extant findings and to predict future findings. Before considering our model, consider three of the most commonly obtained findings concerning time-based prospective remembering, those concerning (a) secondary-task attentional demands, (b) age-related changes, and (c) interval length. Secondary-task attentional demands. Time-based prospective remembering is adversely affected by the attentional, or workload, demands of any nontemporal (secondary) task.8 If a person is performing an attention-demanding secondary task during the retention interval (between forming an intention and the target time for the action), prospective remembering is inversely related to the difficulty of that task. Task difficulty may be assessed in terms of demands on attention, working memory, or both. Prospective remembering may be measured in terms of probability of responding, latency of responding, and similar measures.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

28

R. A. Block & D. Zakay

Age-related changes. Any variable that is correlated with attentional resource allocation also affects time-based prospective remembering. For example, there are medium-to-large effects of normal aging on time-based prospective remembering, with older adults showing decreased probability of responding and increased latency, especially under conditions in which there are relatively high secondary-task attentional demands.8,9 As we discuss later, time-based tasks depend heavily on self-initiated monitoring (executive) processes, and processes that involve attentional resource allocation tend to show age-related declines.10,11 However, with increasing age, people show an increased tendency to rely on external aids and other time-management strategies (Francis-Smythe, this volume; see also Maylor12 ). Perhaps as a result of the use of strategies in naturalistic conditions (which may not be available in laboratory conditions), older adults’ prospective-remembering performance may actually be better than younger adults’ performance in those situations.9 Interval length. Although researchers have not systematically investigated large ranges of the interval between the formation of a time-based intention and the target time, some evidence suggests that time-based prospective remembering is better at shorter intervals than at longer intervals.13 In addition, people attend more to time toward the end of the interval than at the start of it.6,8,14 This is expected on the basis of a models like our attentional-gate model (see later; see also Church15 ).

Prospective Duration Judgment The time dimension is always embedded in any human experience or activity and is an inseparable part of it.16 However, the relevance and importance of time is not constant but varies depending on the meaning assigned to a certain situation. Consider, for example, a person who is relaxing on a beautiful beach on the first day of vacation, reading a novel. Not having any obligations, deadlines, or scheduled meetings for the next two weeks, time is probably not an important issue for that person. A typical result of such a situation is that when the person becomes aware of clock time, he or she is amazed to learn that the subjective duration that was felt is much shorter than the objective time that elapsed since coming to the beach. In other words, subjective time was advancing slower than objective time. Now, consider a situation in which the person is waiting for a date with an attractive person encountered earlier but is not sure whether or not the attractive person will arrive. When the objective time of the date is exceeded by a few minutes, a typical behavior of the person waiting will be to look again

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

29

and again at a watch and with each glance at the watch to discover that objective time did not advance much since the last glance. In this situation, subjective time was advancing much faster than objective time. The discrepancy between the two situations indicates that duration judgments in each case were based on different processes. Although in the first case retrospective duration judgment processes were mainly involved, prospective duration judgment processes were involved in the second case. This distinction is related to the different nature of the cognitive processes underlying retrospective duration judgment and prospective duration judgment. Retrospective duration judgment is inferred on the basis of information retrieved from memory that reflects the amount of change in cognitive context which occurred during a target interval.17,18 Prospective duration judgment, on the other hand, is based on an attentional process and reflects the amount of attentional resources allocated for temporal information processing.19 This difference led Block20 to refer to retrospective duration judgment processes as remembered duration and to refer to prospective duration judgment processes as experienced duration. The differences between the cognitive processes underlying retrospective duration judgment and prospective duration judgment have received strong empirical support.21,22 In order to understand better the conditions under which retrospective or prospective duration judgment processes are initiated, Zakay23 introduced the concepts of temporal relevance and temporal uncertainty. Temporal relevance refers to the significance of time in a certain situation in terms of reaching optimal behavior. For example, if performing a task requires accurate timing, temporal relevance is high; however, if timing has no impact on task performance, temporal relevance is low. Temporal uncertainty refers to the degree to which the duration of a to-be-performed task is known or can be accurately estimated. For example, while performing a routine, well-known task, temporal uncertainty is low; but if an unexpected obstacle prevents the completion of the task and it is not known when the obstacle will be removed, temporal uncertainty is high. When both temporal relevance and temporal uncertainty are high, most available attentional resources will be allocated for temporal information processing (e.g. the example of waiting for an attractive person to arrive), and prospective duration judgment processes will be initiated. However, when both temporal relevance and temporal uncertainty are low (e.g. reading a novel while on vacation), few attentional resources will be allocated for temporal information processing. If the situation leads the person to estimate the

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

30

R. A. Block & D. Zakay

duration of the past time period, retrospective duration judgment processes will be initiated. The process of attentional resource allocation is continuous and under the control of executive functions that monitor the person’s current resource allocation strategy,24 as we explain later. This allocation is flexible25 and reflects the relative strength of temporal relevance and temporal uncertainty at a given moment. Because the allocation process is continuous, so is the shift between prospective duration judgment and retrospective duration judgment. For example, the person reading on the beach is not interested in time, and both temporal relevance and temporal uncertainty are low. However, if suddenly the person receives a message that something has happened and that the person must come to a certain place as quickly as possible, then both temporal relevance and temporal uncertainty become high, many attentional resources are allocated for temporal information processing, and a process of prospective duration judgment is initiated.

Attentional-Gate Model Considering that prospective duration judgment processes depend on the amount of attentional resources allocated for temporal information processing, the nature of such temporal information processing needs to be explained. A basic assumption regarding prospective duration judgment processes is that they almost always occur under dual-task conditions because there is almost always some nontemporal task that should be performed simultaneously with the timing task. Imagine an extreme case in which timing is certainly the most important task, such as waiting impatiently for some event to occur. In such a case, one is usually thinking about possible reasons for the delay and about consequences of the delay. Thus, resources are divided between timing and the nontemporal task.26 This competition over shared resources is resolved by the person’s resource allocation strategy. But what is the nature of the temporal information processing itself? We proposed an attentional-gate model in order to provide an explanation for this issue.27,28 The attentional-gate model is an elaboration of a so-called scalartiming model, an information-processing version of scalar expectancy theory, which was originally proposed to explain processes underlying animals’ temporal behavior.29 The typical scalar-timing model is composed of a pacemaker that emits signals at a constant rate, a switch, an accumulator (which is also called a counter), and a decision-making process. The literature on animal and human

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

31

timing also contains descriptions of variants of this basic pacemaker-counter kind of model.30 The additional component that we added to the scalar-timing model in order to create the attentional-gate model is an attentional gate. The idea underlying the addition of this component is that whereas all other components developed in the early stages of the evolution of timing mechanisms in animals, attentional control is probably unique to primates in that it is associated with brain areas that are not highly developed in earlier mammals. One of these, for example, is the anterior cingulate gyrus, which is intimately involved in the executive control of attention.31 The attentional gate is therefore assumed in order to explain the influence of a person’s attentional resource allocation on prospective timing. As modified to account for time-based prospective remembering, the attentional-gate model operates in the following way (see Fig. 1): (i) A pacemaker emits signals (pulses) at a fairly constant rate that is only slightly changed as a result of changes in arousal level. Although the origin of these relatively constant signals is not fully understood, they may become manifest as synchronized neural firings in specialized neural networks. (ii) The flow of signals passes through an attentional gate, which is controlled by the executive functions that determine a person’s resource allocation policy. The more resources are allocated for timing, the wider (metaphorically speaking) the gate is opened, thus allowing for more signals in a time unit to pass through and enter the accumulator. Thus, the number of signals that enter the accumulator is determined by the amount of attentional resources allocated for timing. To the extent that a person needs to perform a concurrent nontemporal (external stimulus) informationprocessing task, fewer attentional resources are available to attend to time. (iii) The meaning assigned to a situation influences a switch. When the meaning implies a beginning of a target interval that should be timed, the switch opens, enabling the flow of signals from the pacemaker to the accumulator. (In the literature, this condition is often described as causing the switch to be closed, using a metaphor of electrical conductivity. We prefer to use a metaphor of a flow and to speak about switch opening.) When the meaning of a situation implies the end of a target interval, the switch closes again, thus preventing further flow of signals.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

32

R. A. Block & D. Zakay

Fig. 1. An adaptation of the attentional-gate model of prospective timing to processes involved in time-based prospective remembering.

(iv) An accumulator stores the number of signals that passed through the gate from the start of a target interval. When the target interval ends, the switch closes, and the number in the accumulator is a representation of the duration of the target interval. This number is then transferred to a working memory component. A representation of a target interval can be encoded in reference memory directly from long-term memory, such as when one has to produce an interval defined in seconds and minutes. In this kind of situation, one can retrieve from long-term memory a respective representation and store it in reference memory. (v) When a target interval has to be produced or reproduced, the same processes occur, but in this case the number of signals that have entered the accumulator is compared on a constant basis with a representation stored in reference memory. This process, a cognitive comparison, continues until a decision is made that a close match is obtained, upon which the process stops. The person then retrieves the representation of the intended response, which was previously encoded in long-term memory, and makes the target response.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

33

Based on the assumptions of the attentional-gate model, the nature of temporal information processing can be understood as the process of counting the signals in the accumulator as well as the process of making the decision, a process that also demands attentional resources. The strength of the attentional-gate model lies in its ability to provide coherent explanations of most phenomena that characterize prospective duration judgment and to define temporal information processing in a parsimonious and plausible way in terms of the functioning of the central nervous system. Predictions stemming from the attentional-gate model have been empirically supported in several experiments.32,33 From a structural and functional point of view, however, the attentional-gate model should be treated as a hypothetical construct waiting to be validated further, mainly by brain research.33 We also note that other models have been suggested in the literature, including timing-without-a-timer models (i.e. models that do not propose a pacemaker-counter, or internal clock, kind of process).34 Whether the attentional-gate model or competing models better explain time-based prospective remembering is an empirical issue, as well as a theoretical one.

Prospective Timing and Executive Functions Time-based prospective remembering can be considered to be a high-level executive function that requires monitoring and controlling the execution of activities at future times. From this perspective, it is of interest to show that the attentionalgate model can explain prospective timing in relation to the activity of high-level executive functions. This is also in line with Brown’s32 argument that prospective timing consumes resources associated with the executive control of working memory. Zakay and Block35 conducted two studies in which prospective timing was requested simultaneously with tasks that must be monitored and controlled by high-level executive functions. In the first study, participants were required to time the duration of reading sentences that contained syntactic ambiguity, a task that requires high-level executive functions. Duration judgments (reproduced durations) were compared to duration judgments during which reading unambiguous sentences was required. Because resolving syntactic ambiguity demands more resources than regular reading, fewer attentional resources can be allocated for timing in the first condition than in the second condition. The result should be shorter reproductions in the semantic-ambiguity condition than in the no-ambiguity condition, a prediction that was supported by the findings. In a second study, Zakay and Block tested the impact of switching between

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

34

R. A. Block & D. Zakay

tasks on prospective timing. The task-switching paradigm is a robust manipulation of executive control. As predicted by the attentional-gate model, objective intervals were prospectively judged to be shorter for the task-switching condition than for the no-switching condition. Findings from both studies support the attentional-gate model.

Attentional Distractions, the Asymmetric Interference Effect, and the Attentional-Gate Model The dependency of prospective duration judgment processes on the allocation of attention for temporal information processing makes it vulnerable to attentional distractions by competing stimuli and by nontemporal tasks. When a distractor appears, the attentional gate narrows, reflecting the reduction of resources allocated for timing. The result will be a reduced accuracy of timing (in the direction of underestimating the objective time that has elapsed), as well as an increased variability. The increase in variability can be explained by the relative shortage of attentional resources. Brown32 also discussed a related phenomenon, the asymmetric interference effect. The asymmetric interference effect is found when timing competes with the performance of a concurrent nontemporal task. In most cases, unless the nontemporal task inherently includes counting, the timing task suffers more than the nontemporal task. Zakay and Bibi36 argued that the asymmetric interference effect reflects a natural tendency to treat timing as the secondary task and to treat the nontemporal task as the primary one. They found that the asymmetric interference effect is eliminated when timing is treated as the primary task. Another condition under which the asymmetric interference effect disappears is when the nontemporal task is relatively automatic.37,38 The attentional-gate model can explain these findings. For example, in the later case, the nontemporal task is not competing directly with timing and the attentional gate can reflect a resource allocation policy according to which timing gets all of the available resources.

Attentional-Gate Model: Successful and Future Predictions for Time-Based Prospective Remembering In a time-based prospective remembering task, temporal uncertainty is low but temporal relevance is very high. As a result of both, a significant amount of attentional resources are allocated for timing, and a prospective duration judgment

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

35

process is initiated. If the target time for the action is in the range of seconds and minutes in the future, the processes described by the attentional-gate model provide a necessary and sufficient description for the prospective remembering process. In such a case, the representation of the target duration is probably taken from long-term memory and stored in reference memory. When a match between this representation and the ongoing count of signals is obtained, the person performs the action (e.g. turns on the television in order not to miss the news headlines). If this is the case, then it is expected that a person will find it difficult to be accurate and will probably be checking a watch or clock before the target time, because the typical finding is that in prospective situations target durations are underestimated. (A similar phenomenon occurs in the negativeasynchrony task, as described by Zakay and Block.39 ) Another potential mistake is a failure of prospective remembering attributable to distraction, a failure that is also explained by the attentional-gate model. If the target time associated with the prospective task is far in the future (e.g. hours, days, or weeks), a person will probably divide it into several shorter intervals until the target time is near. The reason for this is the difficulty of continuing to allocate attention while timing a long period. The error of missing the target time because of attentional distraction is expected to be greater in a long-duration condition than in a short-duration condition. Research on prospectively made productions has mostly used intervals on the order of seconds, whereas research on prospective remembering has also used intervals on the order of minutes, hours, days, and weeks. Different processes may apply if the target time is hours, days, or weeks in the future. Perhaps the nominally time-based situation becomes more like an event-based situation as the target time is extended further into the future. At the very least, researchers should systematically explore the effects of interval length. Finally, if a person suffers from relatively low attentional resources, or in the relative inability to divide attention between competing tasks, time-based prospective remembering may be impaired. For example, compared to younger adults, older adults show a decreased ability to divide attentional resources.11,40 In a time-based prospective remembering situation, older adults also tend to perform the specified action at a relatively late time, to be more variable in the timing process, and more frequently to fail to perform the action at all.14,41,42 This kind of finding is an additional one that the attentional-gate model can easily explain: Older adults tend to make more variable time productions than do younger adults.43

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

36

R. A. Block & D. Zakay

Event-Based Prospective Remembering In a situation requiring event-based prospective remembering, a person forms a self-generated intention or is given (as by an experimenter) an other-generated intention to perform a specific action when a specific event occurs in the future, with the exact time-of-occurrence usually being somewhat vaguely defined. The future event may be understood or described in terms of a specific spatial location (e.g. “when I pass the post office”), a specific object or person (e.g. “when I next see my friend Mary”), and in other such ways.44 Researchers are undecided about which of several models can best explain event-based prospective remembering (for a discussion of some of them, see Einstein and McDaniel8 ). Before considering our model, consider three of the most commonly obtained findings concerning time-based prospective remembering: (a) secondary-task attentional demands, (b) age-related changes, and (c) contextual changes. Secondary-task attentional demands. In contrast to time-based prospective remembering, event-based prospective remembering is not affected much, if at all, by secondary-task attentional demands, or workload.45 If a person is performing a task during the retention interval (between forming the intention and the target time for an action), event-based prospective remembering is apparently adversely affected only if the relevant cue is outside of focal attention.46,47 Assuming that the person attends fully to the retrieval cue, there is little or no evidence that event-based prospective remembering requires the availability of executive processes.48 Age-related changes. Just as older adults tend to perform worse than young adults on time-based tasks, they also tend to perform worse on event-based tasks. However, the literature is somewhat inconsistent. In two experiments, Einstein and McDaniel49 found no difference between younger and older adults on a task that involved pressing a key whenever a specific target word appeared. This kind of finding contrasts with the typical finding that there are usually medium-to-large age-related deficits in the performance of time-based prospective remembering tasks.41,45 However, meta-analytic findings reveal that there are medium-size age-related deficits in the performance of event-based prospective remembering tasks.9,50,51 Contextual changes. Contextual changes are important in event-based prospective remembering (see, for example, Marsh and colleagues52 ). If the

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

37

target event does not occur in the environmental context expected at the time the intention was formed, event-based prospective remembering performance is impaired.8

Four Relevant Stages At least four major stages must be understood in order fully to explain eventbased prospective remembering: encoding, retention, retrieval, and decision. First, a person forms an intention to perform a specific action at some future time, an occasion during which some target event is expected to occur, thereby encoding a memory trace of the intention (which we will refer to as event-plusaction). Second, this information must be retained during the interval separating the time of encoding and the time of occurrence of the event. Third, the person encounters the target event and may (or may not) retrieve the memory trace that was encoded earlier. Finally, the person may (or may not) decide to perform the intended action. Prospective memory researchers have studied some of the variables that affect these four stages, although to our knowledge they have not yet studied all of the variables that may be expected to influence prospective remembering of event-based intentions. Encoding. A person may form an event-based intention, and thereby encode it into memory, in one of two ways. First, a person may perceive an external stimulus, or cue, and encode an intention about a future action to be taken at some future time when the same external stimulus occurs again. This is what often happens in laboratory studies, in which an experimenter tells a person what action to perform when he or she encounters a specific stimulus in the future. Second, a person may simply imagine performing a future action in a future context, thereby encoding a memory trace of the action, along with contextual associations related to the expected future context. For example, a person may think about telling a colleague something at the next opportunity, which is likely to be at his or her office. The colleague-in-office context is encoded into memory in somewhat the same way that it would be if an instructional stimulus had been provided (within limits to be determined experimentally). Several major encoding-related variables may influence the likelihood that an intention is retrieved later, when the target event actually occurs. To our knowledge, researchers have not adequately investigated all relevant variables. Based on well-established principles of retrospective memory, these are nevertheless

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

38

R. A. Block & D. Zakay

the kinds of variables (among others) that are expected to influence the likelihood that the encoding will lead to successful future action: (i) The person may encode the event-plus-action memory trace on more than one occasion, and those repeated encoding episodes may be separated by various temporal spacings, or lags. The likelihood that the encoded information will be retrieved when the event occurs is expected to increase as a function of the number of repetitions of the encoding, as well as the spacing of them.53 Encoding an event-plus-action memory trace on several occasions that differ in context prevailing during encoding may also enhance later retrieval.54 (ii) The person may encode the event-plus-action memory trace in a verbal (propositional) code, an imaginal code, or both. To the extent that the event-plus-action memory trace is encoded in more than one type of code, performance is expected to improve.55 Similarly, the person may encode the event-plus-action memory trace at various levels of processing, ranging from shallow to deep. Deeper encodings are likely to form relatively more durable memory traces.56 (iii) At the time of encoding, the person may not know the future context in which the target event will occur. To the extent that the event-plus-action memory trace is encoded with few or no contextual associations, it is relatively impoverished, and retrieval of the encoded memory trace may fail. Similarly, knowing when a future event is likely to occur will lead to an encoded memory trace that is more likely to match the actual temporal context when the event occurs. If an event occurs in a temporal or environmental context that is different from what was encoded, retrieval may fail. For example, someone may intend to tell a colleague something when he or she is next encountered, which usually occurs in an office setting. If the colleague is encountered in another setting prior to the office setting, the person may fail to retrieve the encoded intention. Retention. The length of the retention interval between the time of encoding and the future event is expected to affect the likelihood of successful prospective remembering according to well-known principles of forgetting. As the retention interval lengthens, the encoded intention is more likely to be forgotten, just as there are effects of retention interval on retrospective remembering. In addition, there may be interference effects, both proactive and retroactive. A person may encode different intended actions that concern the same contextually

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

39

defined event, and whether or not any particular event-plus-action association is retained and retrieved may depend on the number of similar intentions encoded in memory. For example, a person may encode an intention to tell a friend about a new movie when the friend is encountered next, then encode an intention to tell the same friend about some new software, and then encode an intention to tell the same friend about a past dinner engagement. These three event-plusaction associations share the same context (i.e. the next occasion on which the friend is encountered), and as a result they may suffer from interference effects. This influence on forgetting, which increases as a function of the similarity of memory traces, is well known in the memory literature, even though researchers may not have discussed or studied it much (but see Taylor and colleagues47 ). Retrieval. The third stage is perhaps the most critical, as well as the least understood in the literature. Researchers have been relatively silent on a theoretical understanding of this process. A notable exception is the theorizing of Graf.57 He discussed several basic steps involved at the time of successful retrieval: cue noticing, cue identification, and plan recollection. Here, we propose a different view of the processes involved. Some retrospective memory literature clearly reveals a likely process that may underlie retrieval of an event-plus-action memory trace. In the original mention of this kind of process, it was called study-phase retrieval; more recently, it has been called recursive reminding.58–60 The basic finding underlying the recursive-reminding model is that when an event occurs more than once, memory traces of previous occurrences of the event are retrieved in a relatively automatic way, along with associated contextual information (e.g. the approximate time, or temporal context, of the earlier event, as well as the place, or environmental context, in which the event occurred). Although this process does not occur if two events are completely unrelated,61 it is likely that the events do not have to be identical. Hence, when a person experiences an event, perceiving that event may result in retrieving a memory trace of an earlier intention concerning that event, which was encoded earlier. The notion that the retrieval of a previously encoded intention (event-plusaction memory trace) is often relatively automatic may be useful in clarifying some findings. Specifically, some researchers have found that event-based prospective remembering is not affected much, if at all, by whether or not a person is performing an attention-demanding secondary task at the time the target event occurs.62

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

40

R. A. Block & D. Zakay

Even though the retrieval of a previously encoded event-plus-action memory trace may occur in a relatively automatic way if the event receives attention, several variables may nevertheless influence the success or failure of event-based prospective remembering: (i) The memory trace encoded earlier may not be retrieved later because the context in which the event was expected to occur does not match the actual context in which the event actually occurred. Consider the previous example, in which a person thinks about telling a colleague something at the next opportunity, which is likely to be at his or her office. However, the person may encounter the colleague at a grocery store before going to the office. The actual colleague-in-store context inadequately matches the encoded colleague-in-office context, and the person does not retrieve the previously encoded intention and therefore does not perform the desired action. (ii) An event-plus-action memory trace may be encoded at a time during which the person is directly perceiving the target event or at a time during which a person is merely imagining the target event. In the latter case, failure to retrieve automatically the event-plus-action memory trace may be a result of a failure of the recursive-reminding process attributable to the fact that the two occurrences of the event were not similar enough to lead to automatic retrieval of the encoded intention. (iii) Failure of a process called reality monitoring may influence whether or not an intended action is performed. This failure refers to the occasional inability of people to distinguish between internal thoughts and external events (see Mitchell and Johnson63 for a recent review). If a person vividly imagines that he or she is performing a future action, when the person encounters the target event later, he or she may decide that the action has already been performed and, for that reason, may not perform the action. Decision. The fourth stage, decision, is relevant in everyday prospective remembering situations, although it is probably of relatively minor importance in laboratory studies. If an event-plus-action memory trace is successfully retrieved when the target event occurs, a person may nevertheless decide not to perform the action. This may occur if circumstances have changed since the intention was encoded, and the person decides that the action either is no longer necessary or is undesirable.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

41

Recursive-Reminding Model: Successful and Future Predictions for Event-Based Prospective Remembering The recursive-reminding model proposes that when a repeated event occurs, memory traces of previous occurrences of the event are retrieved in a relatively automatic way, along with associated contextual information. This model successfully predicts several typical findings. Because the underlying retrieval processes are usually relatively automatic, event-based prospective remembering is not affected much by attentional demands (workload) during the retention interval. However, if competing attentional demands or a distracting event prevent a person from attending fully to the target event, the recursive-reminding process may not occur, and prospective remembering may fail.57 With this major exception, there is little evidence that event-based prospective remembering requires executive processes that are involved in attentional resource allocation.48,57 Somewhat inconsistent evidence reveals that event-based prospective remembering shows medium-size effects of variables that are correlated with attentional resource allocation, such as normal aging.9,50,51 At first glance, this evidence seems at odds with the recursive-reminding model, which says that retrieval of an event-based intention involves relatively automatic processes that are not expected to show much age-related decline.64,65 Older adults have relatively limited attentional resources than do younger adults. (However, there may be some event-based situations in which secondary-task attentional demands may influence performance; see Marsh and colleagues.66 ) In addition, when secondary-task or aging effects are found in event-based situations, other factors may be involved, such as a general slowing of cognitive functioning that is typical in older adults.67 Graf 57 recently suggested that when age-related differences are found, they may be a result of differences in encoding, not in retrieval. He concluded that there is “little support for the assumption that substantial attentional resources are required for the recollection of previously formed plans” (p. 321). In a similar way, he concluded that most age-related declines on retrospective memory tasks are attributable to the encoding phase, and fewer declines are attributable to the retrieval phase. Although the recursive-reminding model focuses heavily on the retrieval phase, it must be remembered that encoding and retention phases are also critically involved in event-based prospective remembering. The contextual match or mismatch between the expected event and the actual event also affects event-based performance: If the event does not occur in the

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

42

R. A. Block & D. Zakay

context expected at the time the intention was formed, event-based prospective remembering is impaired.8 For example, Cook and colleagues5 found that “a correct expectation of the context one will be in during the [temporally defined] response window improves time-based memory performance” (p. 352).

Situations Involving Mixed Time-Based and Event-Based Prospective Remembering In this section, we focus on situations that involve mixed time-based and eventbased prospective remembering — those in which a temporally cued future action and an environmentally cued future action may co-occur in an interacting combination. The distinction here is similar to Ellis’s44 distinction between “pure and combined retrieval contexts” (p. 5). Some combined retrieval contexts may involve what we will call an OR rule, whereas other combined retrieval contexts may involve what we will call an AND rule. As an example of a situation involving an OR rule, suppose that it is 8:00, and a person encodes the intention to perform an action at 9:00 (i.e. after 60 minutes has elapsed). The person remembers that a clock on a local building chimes at 9:00 and also encodes the intention to perform the action when the clock chimes. In this situation, prospective remembering could be based on either time-based processes (timing 60 minutes from now) or on event-based processes (hearing the clock chime). In this example, researchers can probably tell which process was the one that was actually used by measuring the time of the action relative to the target time (9:00). If the action was performed at 8:54 (i.e. before the clock chimed), then prospective remembering must have involved the timebased attentional-gate process. On the other hand, if it was performed at 9:00:03 (i.e. 3 sec after the clock chimed), the attentional-gate process is not that precise, and prospective remembering was undoubtedly controlled by the event-based recursive-reminding process. As an example of a situation involving an AND rule, suppose that a person intends to remember to purchase an item at a store when he or she drives by the store, but only after a week has elapsed because that is when the item will be discounted in price. In this case, the dual requirements of an environmental event and a temporal interval must be met. Although these mixed kinds of situations are not uncommon in everyday prospective remembering, only a few researchers have studied them or commented on them. Most research has mainly investigated situations in which an action is to be performed either when a

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

43

specified event occurred or when a specified duration had elapsed. Additional research on AND-rule situations is needed. A few researchers have compared event-based and time-based prospective remembering in the same experiment, although very rarely have researchers explored a temporally defined future action and an environmentally defined future action in some interacting combination. For example, in one recent experiment, younger and older adults were studied.68 Some were asked to indicate whenever an animal appeared in a film that they were viewing, whereas others were asked to respond whenever they judged that three minutes had elapsed. In this experiment, it would have been interesting to add a condition in which participants were asked to respond either when an animal appeared or when three minutes had elapsed. Findings of these kinds of experiments should reveal effects that can be explained by a combination of the two models we described here, the attentional-gate model (for time-based processes) and the recursivereminding model (for event-based processes).

Prospective Remembering in Altered States of Consciousness Mainstream prospective remembering researchers have devoted little or no attention to relevant evidence from studies of people in various states of consciousness. This kind of evidence may clarify prospective remembering in general, as well as the two models proposed here (the attentional-gate and recursive-reminding models). Here we give three interesting examples.

Ordinary Sleep and Time-Based Prospective Remembering A few researchers have investigated the claim that some people seem to be able to awaken at a preselected (experimenter-defined) time during nocturnal sleep, such as 1:23 (see, for example, Tart69 ). This is, of course, isomorphic to ordinary time-based prospective remembering; the target action in these cases is to awaken from sleep at a target time. Some researchers have reported data suggesting that such prospective remembering may occur, possibly with accuracy approaching or equaling that of ordinary (awake) time-based prospective remembering.70 However, this evidence does not necessarily contradict the attentional-gate model (with its consciously controlled attention to time). In particular, participants under such conditions may awaken several times prior to the target time,69 and this could entail several consciously controlled openings of the attentional gate. This evidence, as well as potential processes that may

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

44

R. A. Block & D. Zakay

underlie successful time-based prospective remembering during sleep, needs to be clarified by additional research.71

Lucid Dreaming and Event-Based Prospective Remembering Several researchers have investigated lucid dreaming, a relatively unusual state in which a dreaming person becomes aware that he or she is dreaming. In order to investigate lucid dreaming, a person may be instructed before going to sleep that if he or she becomes lucid during a dream, a certain action should be performed. This is, of course, isomorphic to ordinary event-based prospective remembering; the target action in these cases is to make a specific response when the target event (becoming aware of lucid dreaming) occurs. Evidence reveals that trained lucid dreamers can remember to perform the action (e.g. move one’s eyes three times in a vertical direction, or clench one’s fist three times) near the onset of the lucid dream period.72,73

Hypnosis and Event-Based Prospective Remembering Hypnosis researchers have been fascinated by the possibility that people who are hypnotized and given a suggestion that at a later time (after being brought out of hypnosis) they will perform an action when a specific event occurs. For example, hypnotized people could be told that when they later hear the word experiment, they will automatically rub an earlobe.74 Some research indicates that people tend to perform the action when they receive the post-hypnotic cue, which seems to be an interesting case of event-based prospective remembering. Although there is controversy surrounding the issue of the relevance of the hypnotic state,75 this evidence may nevertheless further support the view that recursive reminding (our model of event-based prospective remembering) may occur in a relatively automatic way.

Summary and Conclusions We proposed two models of prospective remembering, one for time-based remembering and one for event-based prospective remembering. An attentionalgate model is needed to explain time-based prospective remembering; it makes contact with relevant research on time-estimating processes. A recursivereminding model is needed to explain event-based prospective remembering; it makes contact with relevant research retrospective-remembering processes. We

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

45

described these models in detail, contrasting them to other less explicit models in the literature on prospective remembering. We also argued why they are both needed, as well as how they may interact in situations that may involve some mixture of time-based and event-based prospective remembering. Although we disagree with Crowder’s76 de-emphasis of the term remember when referring to prospective remembering, we agree with him that “performing delayed intentions often depends on automatic interruptions of activities in progress” (p. 145). We have emphasized this automaticity in situations involving event-based prospective remembering. However, in situations involving timebased prospective remembering, the interruption of an activity in progress is not automatic but is instead subject to controlled processes (involving dividing attention between nontemporal and temporal information processing). We also agree with Crowder that “memory for intentions plays a role in . . . prospective situations” (p. 146). We have emphasized ways in which attention and memory are involved in the attentional-gate process and the recursive-reminding process. We have not discussed other cognitive processes involved in prospective remembering, and we agree with Crowder these processes are worthy of additional research.

Acknowledgements Support for Dan Zakay was provided by a grant from the Israeli Academy of Sciences. We thank Peter Graf, Simon Grondin, and Michael Myslobodsky for very helpful comments on a previous draft of this chapter.

References 1. Dobbs A R, Reeves M B. Prospective memory: More than memory. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Erlbaum, 1996: 199–221. 2. Graf P, Uttl B. Prospective memory: A new focus for research. Consciousness and Cognition 2001; 10(4): 437–450. 3. Salthouse T A, Berish D E, Siedlech K L. Construct validity and age sensitivity of prospective memory. Memory & Cognition 2004; 32(7): 1133–1148. 4. Kvavilashvili L, Ellis J. Varieties of intention: Some distinctions and classifications. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Erlbaum, 1996: 23–51. 5. Cook G I, Marsh R L, Hicks J L. Associating a time-based prospective memory task with an expected context can improve or impair intention completion. Applied Cognitive Psychology 2005; 19(3): 345–360.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

46

R. A. Block & D. Zakay

6. Harris J E, Wilkins A J. Remembering to do things: A theoretical framework and an illustrative experiment. Human Learning 1982; 1: 123–136. 7. Miller G A, Galanter E, Pribram K. Plans and the Structure of Behavior. New York: Holt, Rinehart, & Winston, 1960. 8. Einstein G O, McDaniel M A. Retrieval processes in prospective memory: Theoretical approaches and some new empirical findings. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Erlbaum, 1996: 115–141. 9. Henry J D, MacLeod M S, Phillips L H, Crawford J R. A meta-analytic review of prospective memory and aging. Psychology and Aging 2004; 19(1): 27–39. 10. McDowd J M, Shaw R J. Attention and aging: A functional perspective. In Craik F I M, Salthouse T A, eds. The Handbook of Aging and Cognition (2nd edn). Mahwah, NJ: Erlbaum, 2000: 221–292. 11. Anderson N D, Craik F I M, Naveh-Benjamin M. The attentional demands of encoding and retrieval in younger and older adults: I. Evidence from divided attention costs. Psychology and Aging 1998; 13(3): 405–423. 12. Maylor E A. Older people’s memory for the past and the future. Psychologist 1996; 9(10): 456–459. 13. Nigro G, Senese V P, Natullo O, Sergi I. Preliminary remarks on type of task and delay in children’s prospective memory. Perceptual and Motor Skills 2002; 95(2): 515–519. 14. Einstein G O, McDaniel M A, Richardson S L, Guynn M J, Cunfer A R. Aging and prospective memory: Examining the influences of self-initiated retrieval processes. Journal of Experimental Psychology: Learning, Memory, and Cognition 1995; 21(4): 996–1007. 15. Church R M. A concise introduction to scalar timing theory. In Meck W H, ed. Functional and Neural Mechanisms of Interval Timing. Boca Raton, FL: CRC Press, 2003: 3–22. 16. Flaherty M G, Meer M D. How time flies: Age, memory, and temporal compression. Sociological Quarterly 1994; 35(4): 705–721. 17. Block R A. Temporal judgments and contextual change. Journal of Experimental Psychology: Learning, Memory, and Cognition 1982; 8(6): 530–544. 18. Block R A. Prospective and retrospective duration judgment: The role of information processing and memory. In Macar F, Pouthas V, Friedman W J, eds. Time, Action and Cognition: Towards Bridging the Gap. Dordrecht, Netherlands: Kluwer Academic, 1992: 141–152. 19. Zakay D. Attention allocation policy influences prospective timing. Psychonomic Bulletin & Review 1998; 5(1): 114–118. 20. Block R A. Experiencing and remembering time: Affordances, context, and cognition. In Levin I, Zakay D, eds. Time and Human Cognition: A Life-Span Perspective. Amsterdam: North-Holland, 1989: 333–363. 21. Block R A, Zakay D. Prospective and retrospective duration judgments: A meta-analytic review. Psychonomic Bulletin & Review 1997; 4(2): 184–197. 22. Zakay D, Tsal Y, Moses M, Shahar I. The role of segmentation in prospective and retrospective time estimation processes. Memory & Cognition 1994; 22(3): 344–351.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

47

23. Zakay D. On prospective time estimation, temporal relevance and temporal uncertainty. In Macar F, Pouthas V, Friedman W J, eds. Time, Action and Cognition: Towards Bridging the Gap. Dordrecht, Netherlands: Kluwer Academic, 1992: 109–117. 24. Gopher D. Attention control: Explorations of the work of an executive controller. Cognitive Brain Research 1996; 5(1/2): 23–38. 25. Macar F, Grondin S, Casini L. Controlled attention sharing influences time estimation. Memory & Cognition 1994; 22(6): 673–686. 26. Kahneman D. Attention and Effort. Englewood Cliffs, NJ: Prentice-Hall, 1973. 27. Block R A, Zakay D. Models of psychological time revisited. In Helfrich H, ed. Time and Mind. Kirkland, WA: Hogrefe & Huber, 1996: 171–195. 28. Zakay D, Block R A. The role of attention in time estimation processes. In Pastor M A, Artieda J, eds. Time, Internal Clocks and Movement. Amsterdam: North-Holland/Elsevier Science, 1996: 143–164. 29. Gibbon J, Church R M, Meck W H. Scalar timing in memory. In Gibbon J, Allan L G, eds. Timing and Time Perception. Annals of the New York Academy of Sciences: Vol. 423. New York: New York Academy of Sciences, 1984: 52–77. 30. Grondin S. From physical time to the first and second moments of psychological time. Psychological Bulletin 2001; 127(1): 22–44. 31. Posner M I, Raichle M E. Images of Mind. San Francisco, CA: Freeman, 1994. 32. Brown S W. Attentional resources in timing: Interference effects in concurrent temporal and nontemporal working memory tasks. Perception & Psychophysics 1997; 59(7): 1118–1140. 33. Coull J T, Vidal F, Nazarian B, Macar F. Functional anatomy of the attentional modulation of time estimation. Science 2004; 303(5663): 1506–1508. 34. Block R A. Psychological timing without a timer: The roles of attention and memory. In Helfrich H, ed. Time and Mind II: Information Processing Perspectives. Göttingen, Germany: Hogrefe & Huber, 2003: 41–59. 35. Zakay D, Block R A. Prospective and retrospective duration judgments: An executive-control perspective. Acta Neurobiologiae Experimentalis 2004; 64: 319–328. 36. Zakay D, Bibi A. Temporal judgments, selective attention and attention allocation policy. Unpublished manuscript, Tel-Aviv University, Israel 2002. 37. Block R A, Zakay D, Cloninger M A. Face and Time: Evidence for Encoding Interference. New Orleans, LA: Psychonomic Society, 2000. 38. Block R A, Zakay D, Richmond A S. Automatic and Intentional Processes in Face Encoding. Vancouver, Canada: Psychonomic Society, 2003. 39. Zakay D, Block R A. The negative asynchrony phenomenon: A prospective timing perspective: Commentary on Müller et al. In Aschersleben G, Bachmann T, Müsseler J, eds. Cognitive Contributions to the Perception of Spatial and Temporal Events. Amsterdam: North-Holland/Elsevier Science, 1999: 251–257. 40. McDowd J M, Craik F I. Effects of aging and task difficulty on divided attention performance. Journal of Experimental Psychology: Human Perception and Performance 1988; 14(2): 267–280. 41. Park D C, Hertzog C, Kidder D P, Morrell R W, Mayhorn C B. Effect of age on event-based and time-based prospective memory. Psychology and Aging 1997; 12(2): 314–327.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

48

R. A. Block & D. Zakay

42. Rendell P G, Thomson D M. Aging and prospective memory: Differences between naturalistic and laboratory tasks. Journal of Gerontology 1999; 54(4): B256–B269. 43. Block R A, Zakay D, Hancock P A. Human aging and duration judgments: A meta-analytic review. Psychology and Aging 1998; 13(4): 584–596. 44. Ellis J. Prospective memory or the realization of delayed intentions: A conceptual framework for research. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Erlbaum, 1996: 1–22. 45. Bastin C, Meulemans T. Are time-based and event-based prospective memory affected by normal aging in the same way? Current Psychology Letters: Behaviour, Brain & Cognition 2002; 7: 105–121. 46. Hicks J L, Cook G I, Marsh R L. Detecting event-based prospective memory cues occurring within and outside the focus of attention. American Journal of Psychology 2005; 118(1): 1–11. 47. Taylor R S, Marsh R L, Hicks J L, Hancock T W. The influence of partial-match cues on event-based prospective memory. Memory 2004; 12(2): 203–213. 48. van den Berg S M, Aarts H, Midden C, Verplanken B. The role of executive processes in prospective memory tasks. European Journal of Cognitive Psychology 2004; 16(4): 511–533. 49. Einstein G O, McDaniel M A. Normal aging and prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 1990; 16(4): 717–726. 50. Uttl B. Age-related changes in event cued prospective memory proper. In Ohta N, MacLeod C M, Uttl B, eds. Dynamic Cognitive Processes. Tokyo: Springer, 2005: 273–303. 51. Uttl B. Age-related changes in event cued visual and auditory prospective memory proper. In Aging, Neuropsychology, and Cognition in press. 52. Marsh R L, Hicks J L, Hancock T W, Munsayac K. Investigating the output monitoring component of event-based prospective memory performance. Memory & Cognition 2002; 30(2): 302–311. 53. Hintzman D L. Repetition and memory. In Bower G H, ed. Psychology of Learning and Motivation, Vol. 10. New York: Academic Press, 1976: 47–91. 54. Hintzman D L, Stern L D. Contextual variability and memory for frequency. Journal of Experimental Psychology: Human Learning and Memory 1978; 4(5): 539–549. 55. Paivio A. Dual coding theory: Retrospect and current status. Canadian Journal of Psychology 1991; 45: 255–287. 56. Craik F I M, Lockhart R S. Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior 1972; 11: 671–684. 57. Graf P. Prospective memory retrieval revisited. In Ohta N, MacLeod C M, Uttl B, eds. Dynamic Cognitive Processes. Tokyo: Springer, 2005: 305–332. 58. Hintzman D L, Summers J J, Block R A. Spacing judgments as an index of study-phase retrieval. Journal of Experimental Psychology: Human Learning and Memory 1975; 1(1): 31–40. 59. Tzeng O J, Lee A T, Wetzel C D. Temporal coding in verbal information processing. Journal of Experimental Psychology: Human Learning and Memory 1979; 5(1): 52–64.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

Prospective Remembering Involves Time Estimation and Memory Processes

49

60. Hintzman D L. Judgment of frequency versus recognition confidence: Repetition and recursive reminding. Memory & Cognition 2004; 32(2): 336–350. 61. Hintzman D L, Block R A. Memory for the spacing of repetitions. Journal of Experimental Psychology 1973; 99(1): 70–74. 62. Brandimonte M A, Ferrante D, Delbello R. Event-based prospective memory is insensitive to short-term memory load: Some observations on automaticity and monitoring in prospective remembering. International Journal of Cognitive Technology 2001; 6(2): 33–40. 63. Mitchell K J, Johnson M K. Source monitoring: Attributing mental experiences. In Tulving E, Craik F I M, eds. The Oxford Handbook of Memory. New York: Oxford University Press, 2000: 179–195. 64. Hasher L, Zacks R T. Automatic and effortful processes in memory. Journal of Experimental Psychology: General 1979; 108(3): 356–388. 65. Light L L. Memory and aging: Four hypotheses in search of data. Annual Review of Psychology 1991; 42: 333–376. 66. Marsh R L, Hancock T W, Hicks J L. The demands of an ongoing activity influence the success of event-based prospective memory. Psychonomic Bulletin & Review 2002; 9(3): 604–610. 67. d’Ydewalle G, Bouckaert D, Brunfaut E. Age-related differences and complexity of ongoing activities in time- and event-based prospective memory. American Journal of Psychology 2001; 114(3): 411–423. 68. Logie R H, Maylor E A, Della Sala S, Smith G. Working memory in event- and timebased prospective memory tasks: Effects of secondary demand and age. European Journal of Cognitive Psychology 2004; 16(3): 441–456. 69. Tart C T. Waking from sleep at a preselected time. Journal of the American Society of Psychosomatic Dentistry and Medicine 1970; 17: 3–15. 70. Zung W W K, Wilson W P. Time estimation during sleep. Biological Psychiatry 1971; 3(2): 159–164. 71. Block R A. Time and consciousness. In Underwood G, Stevens R, eds. Aspects of Consciousness: Vol. 1. Psychological Issues. London: Academic Press, 1979: 179–217. 72. Erlacher D, Schredl M, LaBerge S. Motor area activation during dreamed hand clenching: A pilot study on EEG alpha band. Sleep & Hypnosis 2003; 5(4): 182–187. 73. LaBerge S, Levitan L, Dement W C. Lucid dreaming: Physiological correlates of consciousness during REM sleep. Journal of Mind & Behavior 1986; 7(2&3): 251–258. 74. Barnier A J, McConkey K M. Hypnotic and posthypnotic suggestion: Finding meaning in the message of the hypnotist. International Journal of Clinical & Experimental Hypnosis 1999; 47(3): 192–208. 75. McCue P A. Is a trance state associated with the execution of posthypnotic suggestions? Contemporary Hypnosis 1992; 9(1): 43–52. 76. Crowder R G. Commentary: The trouble with prospective memory: A provocation. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Erlbaum, 1996: 143–147.

April 7, 2006 13:26 WSPC/SPI-B378 Timing the Future ch02

This page intentionally left blank

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

3 Dynamic Attending and Prospective Memory for Time Mari Riess Jones∗

Introduction: Attending in Time This chapter is about attending and its effects on our sense of time. By inference, it is also about memory because in a real sense memory depends on the original act of attending. When psychologists study attention, it is often examined through the lens of a particular experimental task. Indeed, attention is often operationally defined in terms of various classic attention tasks (e.g. selective versus divided attending designs), each of which may or may not require people to sustain their attending over a span of time.1 Of course, operational definitions of attention tend to circumscribe the range of possible experimental outcomes and limit the generalizability about the construct of attention, which at its core is a very general one. Nevertheless, defining the construct of attending is notoriously difficult. In fact, definitions of attention quickly turn into theories of attending. Finally, the operational definitions that we initially use to describe our tasks ultimately wind up shaping the theory we develop to define the broader construct of attending. In the mid-twentieth century, the attention tasks in vogue were ones that relied heavily upon dichotic listening, a task often popularized by generalizing its findings to the ecologically valid scenario known as the cocktail party ∗ The Ohio State University Department of Psychology, Townshend Hall Columbus, OH 43210, USA; e-mail:

[email protected]

51

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

52

M. R. Jones

phenomenon. Two auditory messages are simultaneously delivered to a listener, one to each ear mimicking the experience a listener has when two people are speaking at or near her during a crowded cocktail party. Variants of this task required that listeners either selectively monitor the conversation at one ear (a selective attention task) or attend to conversation in both ears (a divided attention task). These tasks tended to require that people sustain their attending activities over seconds or minutes to detect a target sound (sustained attending). The accompanying attention theories of that day were reinforced by research along these lines and were built upon the famous channel theory of Broadbent,2 where in its simplest application each ear functions as a single information channel. Filters arguably associated with channels then could explain selective attention to one or the other ear (channel). Nowadays, the most common tasks for studying attention have shifted. In contrast with earlier research, which employed auditory events, this shift has been accompanied by increased reliance on visual stimuli. In visual search tasks, task-relevant targets or features are identified in advance so that people can selectively search visual arrays containing, for example, red items or attend selectively to a particular point in space.3 In such tasks, attention is sustained over the length of search, meaning that search time measures the time to monitor an array and locate a target. Indeed current research on attention rests heavily on presentations of brief static visual scenes which viewers must rapidly search in order to locate and/or identify a designated target item. Instead of channels, contemporary attention theories (of which there are many) are mainly concerned with location and feature maps of various sorts.4 Theoretical questions concern the way attentional resources may be allocated to certain locations and/or features to enable binding of features together to form experienced visual objects.1,5

Attending and Processing Time In light of the contemporary concentration on visual search tasks and the tendency to define attention in terms of these kinds of tasks it is pertinent to raise questions about the relevance of such approaches to people’s response to time. Time plays an important role in many contemporary theories of attending, but in some respects, this role is a confining one, as I shall argue. Many current theories interpret time mainly as processing time. In this form, time looms large as a dependent variable seen, for instance, in reaction time (RT) or search time. Allowable processing time also figures as a design constraint that determines the temporal structure of attention tasks. The stipulated time intervals between

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

53

stimulus elements to which one must respond are typically designed to be of sufficient absolute durations to permit processing of the individual elements. Thus, whereas fast reaction times may reflect efficient processing of a presented stimulus, a task designed with short exposure times may reduce processing accuracy of a stimulus element. A constraint that emerges when time is strictly treated as processing time is that one function of time, as it figures in task structure, is ignored. For instance, it is commonly assumed that the only function of various time intervals within (or between) trials in a task is to provide for ample processing. Accordingly, this tacitly assumes that the material presented within a single trial is processed per interval and independently of other time intervals within the same trial. For example, in a visual attention task, one typically presents a fixation point, then a cuing element, then a target, possibly a post-target distractor, and then a pause for the next trial. Time spans between the onsets of each element in such a sequence are often selected independently under the assumption that processing time depends on each element separately.5 In other words, the temporal context within (and between) trials is considered irrelevant to the way a person may attend in such tasks, except insofar as each isolated time interval, independently, allows sufficient processing. However, taken together the whole sequence of element onsets and their related time spans forms a temporal profile that, in itself, may affect performance. The function of this temporal context on attending is not captured when temporal intervals are interpreted strictly as opportunities for processing preceding stimulus items. Although it is undeniable that a psychological activity like processing happens within each time interval, it also possible that time plays another role as well.

Attending and Relative Time The preceding account leads to these questions: Might the temporal profiles of stimulus elements within a trial, i.e. the temporal context, affect the processing of a subsequent target element, that is, speed up or slow down processing time? Are the required time intervals that combine to yield a within-trial temporal profile really independent? What if the time intervals (of sufficient absolute durations for processing) that separate successive elements within (and between) trials are randomly selected? Would people respond to them independently? The answer is currently unknown because the time structure of tasks is rarely a subject of study.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

54

M. R. Jones

The view of attending and timing that we describe in this chapter sparks such questions. It is a view of attending that differs from other, widely accepted, interpretations of attending. I assume that the time structure of a task contributes to a guiding of attending. Time structure refers to the relative timing of successive time spans that are respectively marked by onsets of stimulus items (signal tones, lights, and so forth). I claim that the relative timing among a series of stimulus elements can more or less effectively pace attending. In this, attending has an important temporal component that is responsive to relative timing in the external environment which includes stimulus events and the structure of a task within which such events are embedded. Humans have evolved in a manner that exploits this component because such sensitivity enables avoidance of dangers and also facilitates appropriate intra-species communications in voice and gesture. Attending is an inherently dynamic activity. This is evident in our sensitivities to stimulus timing and in the way attending is guided in time by the relative timing of an environmental context. Temporal aspects of attentional activities come to synchronize (i.e. lock-into) corresponding temporal aspects of an environment. Attentional pacing allows an attender to heighten attending in advance so as to distribute attending energy over the “right” neighborhood in time. This is not to say that temporal pacing necessarily results in effective or beneficial performance in a given task. Indeed, attentional pacing may be responsible for various kinds of mistakes in certain situations because it has the potential for misguiding as well as correctly guiding attending in time. For instance, when conversing with a particularly ponderous speaker, whose utterances unfold with a slow and hypnotizing pace, we might find ourselves asking the speaker to repeat a phrase because our attention was not focused at the correct time scale and/or point in time to appropriately lock-into his comment. In this account of attending, it is clear that relative timing among successive intervals is as important as the absolute durations of these intervals. So, if people are indeed sensitive to the relative time structure of tasks and events, just how does this work? Moreover, what does this have to do with time perception? In the following sections, I develop this alternative view of attending and attempt to show its connection with time perception and time estimation. I begin by contrasting the Dynamic Attending Theory (DAT) with other approaches to time, and then outlining some of its basic assumptions. Next, I consider some applications of DAT to tasks that require prospective time judgments.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

55

Dynamic Attending Theory (DAT): Periodicities in Time In a famous paper, James Gibson argued that time is not perceivable but events are.6 Oddly enough, although I study time perception, I agree with the spirit of this claim. For this reason, I often use stimuli that approximate real world events in order to show that time is inseparable (i.e. integral) from event structure. This is most evident in rhythmic patterns associated with natural events because the rhythm of a sequence depends upon the relative timing of non-temporal markers, namely, on discrete changes along some dimension such as space, frequency, intensity, etc. Generally, I consider an event to be a natural happening that has structure over and in time.7 Gibson stressed that events have structure over time by which he meant that changes along the spatial dimension are not simply static differences in space, but rather they are perceived as spatial relations because the changes involved transpire over time; they are manifest as spatial change over time in the form of velocity gradients, for example. In his view, the transformation of a non-temporal property (such as a spatial change) with respect to time reveals some invariant event relationship such as a gradient ratio, namely, a spatial ratio that is “picked up” by a perceiver. Whereas this eventspecific invariant is perceived, the total time span of a transformation is not necessarily perceived as such. A time span merely supports the pickup of an invariant that persists over the transformational interval, according to Gibson. Indeed, in our everyday activities we usually do not pay explicit attention to such time intervals; rather we concentrate on various properties that allow us to navigate in the world; our first impulse is not to ask ourselves: “How long did that take?” Unless forewarned (prospectively), we may be fairly poor at answering questions about event time. For instance, it might be difficult for a watchful fan to accurately report the total time of an action that involves a ball player dribbling a ball, who moves the ball down the basketball court in the midst of an active game when suddenly queried after this action was completed (a retrospective time judgment). All of this reinforces Gibson’s original claim that we are not normally oriented for perceiving a time interval as separate from the event itself. This, perhaps, explains the ubiquity of wristwatches. But I suggest that time spans themselves are an integral part of events and that their relationships within these events are important. Consider the basketball example. It is interesting in part because the events involved exhibit structure in time as well as over time. When we examine event structure more closely we find something that Gibson failed to mention: namely, a rather compelling series

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

56

M. R. Jones

of transformations in time that describe relative time properties of the dance conveyed by a dribbling basketball player. Invariants associated with timing (i.e. time ratios) were not of central interest to Gibson because he focused on visual percepts arising from events created by changes in space that mainly transpire over a single time span. However, many events that are important to us comprise multiple time spans, and in these events we can show that invariant properties relate to the extended structure in time. Let us strip down the basketball example to consider a single bouncing ball. An immediate response to a bouncing ball is not to abstract a single time span at a time scale that covers the total duration from first to last bounce. Nor are we inclined to abstract out or remember one or more absolute time intervals that are marked by successive, discrete, bounces; such time spans contribute to a lower time scale within the bouncing ball event. Normally we do not even think about absolute durations at either time scale. What we do notice, I submit, is the relative timing of bounces. We implicitly sense that there are recurrent and connected time spans that convey a (possibly changing) periodic pattern; as each bounce occurs we come to anticipate the next. Critical temporal properties in a bouncing event depend intrinsically on the ball’s non-temporal properties such as its material, its volume, the surface on which it bounces and so on. Because these non-temporal properties effectively determine the relative amplitudes and relative durations of successive bounces, we can say that the temporal properties in such events depend upon non-temporal properties. Of course, as we experience (either visually or acoustically) a bouncing ball we are often tacitly mesmerized by the resulting rhythm. In fact, I maintain that it is the rhythm, specifically the relative time structure, which qualifies this as an event having certain invariant properties. Most of the events that I study do not conform to the brief static visual arrays, common in many attention tasks, nor can they be categorized as extended dynamic visual scenes that Gibson was fond of. Instead they are acoustic events that simulate musical patterns; often, they are quite simple relative to artful music. An example of one such event appears in Fig. 1 where a sequence of tones of different pitches mark out a series of time intervals (inter-onset-intervals or IOIs). In auditory events, such as speech and music, temporal relationships and rhythmic invariants stand out. Indeed, an intriguing aspect of auditory patterns is their rhythmical richness. Many of the auditory events common to our daily lives manifest this richness through multiple time scales. In speech and music, various levels of event structure exist that tend to covary with time scales; thus in speech

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

57

A sequence of two pitches High

Pitch

Total Duration 7 IOIs

Low

2 IOI

IOI

Time Fig. 1. A sequence of alternating tones with high (hatched) and lower (solid) pitches separated by a single IOI at the lowest time level, forming an isochronous rhythm. High tones are separated by two IOIs at the next time scale. At the highest time scale of total duration, seven IOIs separate the onset of the first tone from the onset of the last one.

phonemic structure transpires over smaller time scales which are embedded within larger time scales of syllables and words. Similarly, the time spans among adjacent onsets form a lower level time scale, as indicated in Fig. 1. These lower level IOIs are nested within longer time spans (higher time scales) that relate non-adjacent events. Different time scales are often distinctively marked by nontemporal acoustic markers, such as pronounced frequency or intensity changes. Thus, the total duration of this event embeds at least two other time scales; one corresponds to the time intervals of higher pitched tones alone; the other corresponds to single IOIs marked simply by intensity changes at the lowest level of event structure. Thus, as with the event of a bouncing ball, we find that the non-temporal properties (here tone frequency and/or intensity changes in tones) of an event affect the temporal properties. In this respect, the structure in time found in auditory events reveals one way in which the event’s rhythm emerges from both time spans and the non-temporal properties which mark these time spans.

The Function of Event Time Structure What Gibson did not claim, but I do, is that to perceive events we tacitly use the time structure within a given event to attentionally track its course in real time.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

58

M. R. Jones

DAT assumes that for a listener, one function of event time structure is to provide a basis for guiding attending in time at some time scale(s). To be sure, when confronted with dynamically changing events, whether visual or auditory, we do not perceive abstracted time intervals, as Gibson correctly implied. Rather, what we perceive is the integrated dynamic event in which non-temporal and temporal relationships jointly delineate relative time invariants. I suggest that an event’s emergent rhythmic pattern contributes to how we attend dynamically in time and to our experiences of event timing, allowing us to render judgments about an event’s time structure. According to DAT, the judgments we make about time, regardless of their accuracy, will always be a byproduct of how we attended to and perceived an event. Having said this, it is a tall order to explain precisely how various structural (temporal and non-temporal) aspects of an event combine to dynamically guide attending in various tasks, including time judgment tasks. The jury is still out on these complicated issues. But to begin to tackle them, we can consider the possibility that one important aspect of a listener’s response to such events entails an implicit use of the event’s own time structure to pace attending to it. Naturally, this proposal raises additional issues surrounding how this might occur and about what it means in experimental tasks where listeners must judge event time at one or another time scale.

Theoretical Assumptions of DAT This brings me to some basic theoretical assumptions about internal timing and attention. Three assumptions are basic to DAT.8 The first assumption I have already discussed; it relates to event structure. I assume that all events have structure in time where this structure incorporates aspects of non-temporal rhythmic markers. However, it is important to acknowledge that such event time structures include rhythms that are highly incoherent as well as those with great rhythmic coherence. The second basic assumption concerns the basis of attending as involving internal timing. I have long assumed that the internal time structure of living things comprises a variety of nested internal attending periodicities (i.e. multiple time scales) in the form of neural oscillations. These oscillators carry attending energy, hence they are the vehicles of attending. At a more general level, internal oscillations are an important part of our biological structure; my colleagues. I have proposed that oscillations that underlie attending are realized by neural oscillations capable of entrainment.9 This means that although attending

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

59

oscillations manifest periodic structure in time, they are not rigidly periodic (i.e. as in beat-based models). Rather, entraining oscillators incorporate (within limits) adaptable periods. Recent evidence for the role of neural oscillations in attending has been reported by Large and colleagues.10– 13 Although I focus upon neural (internal) oscillations, I assume that they determine external counterparts observable in various overt behaviors, such as locomotion patterns, body gestures, and communication patterns produced by speakers and musicians, among other natural activities. The third assumption speaks most directly to how attending dynamics work. Acts of attending involve a stimulus event as well as activation of an internal oscillation. This last assumption posits how an inherently temporal attender relates to an intrinsically temporal event as it unfolds in real time. DAT conveys a specific story about the interaction of these two kinds of time structures. It assumes this involves a dynamic coordination described by principles of entrainment. The biological oscillations in an attender are capable of adaptive temporal coordinations with external rhythmic structure of events, thereby leading to a new relative time expression: dyadic relative time. Dyadic relative time refers to the emergent time structure that develops when internal neural oscillations (of an attender) coordinate with the external time structure (of an event). Effectively, an attender comes to synchronize with an event at certain time levels or scales. Although such synchrony can emerge from automatic and/or voluntary components of an entrainment activity, the important point is that, at a given time scale within an event, attending ultimately involves a synchronization of some internal oscillation, of a given period, with that event time scale. In some simple tasks, people may automatically pace attending to a lower level event time scale. Thus, for instance, by default, people may attend successively to each tone in Fig. 1 by simply “tuning into” relevant time spans between tones at the lowest event time scale (IOIs). In other tasks that, for instance, involve selective attending to complex events, people may learn to voluntarily synchronize their attending to a different task-relevant time scale within that event. For the Fig. 1 melody, if explicitly told to attend only to higher tones, they learn to rely instead on a higher time scale (2 × IOIs) or even its total duration (7 × IOIs). Dyadic interactions characterize these various types of attender-event interactions. In sum, people may effectively run on “automatic pilot” in tracking an event at one or more fixed time scale(s), or they may voluntarily shift attending energies to lock-in certain neural oscillations to specific time scales within an event.14

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

60

M. R. Jones

Applications of Dynamic Attending Theory In this section, I first describe general features of DAT. Next, I describe applications of DAT to different tasks. These include a conventional attending task where people must identify certain non-temporal properties of an event. I incorporate this application to illustrate that DAT is conceived as a general approach to attending in which timing plays a novel role. Finally, I extend DAT applications to time perception. In the latter sections, I address differences between retrospective and prospective designs and present recent research on prospective research designs.

Attending in Real Time I begin by illustrating the way in which general attending activities are conceived by DAT. This involves internal oscillations that allow us to tacitly track an extended event and to perceive different aspects of it. Assume that a listener hears the auditory pattern shown in Fig. 2 in a task requiring pitch monitoring. In this case, the illustrated pattern is more representative of real musical events than that shown in Fig. 1. This nine tone sequence forms a pleasant melody

Distance in pitch space

One nine-tone sequence: F#5 F5 E5 D#5 D5 C#5 C5 B4 A#4 A4 G#4 G4 F#4 F4 E4 D#4 D4 C#4 C4 B3 A#3 A3 G#3

Target

Time Fig. 2. A nine tone melody with specific pitches indicated on the ordinate and time intervals on the abscissa. Circles indicate successive (similar) tone groups. In pitch monitoring tasks, the eighth tone (target), embedded in the third tone group, could change in pitch.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

61

that could readily fit into a number of compositions we find in Western music. Although several time scales are evident, the relevant time structure remains fairly simple in this tune. At the lowest level (IOI), the time scale forms an isochronous rhythm that is marked out by intensity changes associated with tone onsets following silences (i.e. IOIs, are identical at 600 ms). At a higher time scale, it can be seen that the melodic structure identifies three groups of tones (circled in Fig. 2). At this time level, although the rhythm is slower it is also isochronous. At each level, the changes that mark component time spans are coherently distributed in time to outline distinct isochronous rhythms of respectively different rates. A question we address here is this: “Does a listener use the resulting time structure to pace attending to this event, and if so, how?” To address these questions, we return to the entraining properties of neural oscillations. Figure 3 illustrates one of the simplest ways to model a neural oscillatory process at a single time scale. The oscillators of interest here exhibit two important features, namely period and phase; period refers to the recurrent cycle of attending energy whereas phase refers to a point within a cycle where attending energy is maximal, reflecting heightened attention. Many biological oscillations have an ability to entrain, that is, to synchronize with quasi-periodic stimulation. These entraining oscillations typically have special properties in that their periods are resilient to disturbances. Although the phase of such an oscillator may shift with a disturbance from an external stimulating force, the oscillator’s period may only briefly change to accommodate this disturbance. That is, stabilizing properties of an entraining oscillator mean that, in the limit,

Fig. 3. An illustration of entrainment of an internal neural oscillation that carries attending energy in time to track a sequence of stimulus IOIs. Two important aspects of this model involve the oscillator’s period and its phase.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

62

M. R. Jones

the oscillator’s period will return its favorite cycle time. Such oscillators are termed limit cycle oscillators and they offer great explanatory power. This is because oscillators having a wide variety of natural periods exist; it is well established that a range of biological oscillations exist in humans that exhibit periods ranging from fractions of milliseconds to years.19 In this approach, we assume that the entraining neural oscillations which supporting attending operate with periods ranges from milliseconds to hours.8 As limit cycle oscillations, each oscillator can express a persistent and steady ebb and flow of attending energy, over its intrinsic periodicity, where a pulse of energy is assumed to be concentrated at an expected phase point in time.15,16 Entrainment is illustrated in Fig. 3. This figure effectively shows how a dyadic system, comprising an external event and an internal oscillation, may work. Thus, let us assume that the period of an active oscillator approximates the time scale of the lower level rhythm (rather than a higher time scale) within the melody of Fig. 2. At this stimulus IOI level, pulses of attentional energy are carried, as recurrent bursts; each burst (i.e. pulse peak) is assumed to reflect a temporal expectancy for a tone onset. As an event unfolds in real time, pulse peaks come to phase align with successive tone onsets such that the system anticipates, in time, future points of change. In other words, the dyadic system is drawn to phase synchrony. This describes the process of entrainment. Effectively, entrainment offers a basis for modeling dyadic relative time. Entrainment is a natural activity that is found in most animals and many plants. We are most familiar with circadian entrainment, which involves periods of 24 hours, but biological entrainment has been observed on many different time scales. It is a natural vehicle for explaining how dyadic timing works if we assume that entrainment realizes the function that event time plays for an attender, namely it facilitates synchronicity of attending with an event. Entrainment has a potential for insuring synchronicity. Therefore, in DAT entrainment is the fundamental mechanism because it enables moment-to-moment synchronicities in attending at the time scales we rely upon to communicate efficiently through speech and music. It describes how we can so quickly, indeed seemingly effortlessly, lock-into the ongoing dynamics of a variety of events in the world that our species uses to communicate in gesture and sound. All such behaviors depend upon simultaneous coordination of attending energies with aspects of the structure of various events. It is possible to formalize entrainment activities. However, the mathematics underlying this formalization depend on the type of oscillators involved,

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

63

among other things; moreover the mathematics of entrainment can become complex.17–19 The gist of the basic principles involved can be sketched if we focus upon the entrainment of a single oscillator to a single stimulus time scale, as portrayed in Fig. 3. One such model, based upon the sine circle map equation, has been developed by Edward Large.15,20 A key feature of entrainment, evident in this model, concerns an oscillator’s adaptivity. Entrainment entails momentary, adaptive, changes in the phase and period of an entraining oscillator in response to event timing. The period of an oscillator can adjust its length (within limits) to match recurrent time spans within an event (e.g. here a repeating stimulus IOI). Not only can an entraining oscillator adapt the period length of a neural oscillation, shrinking or expanding it to match event time spans, it can also temporally shift the phase location of the pulse peak to align it in time with new tone onsets. We have interpreted the pulse peak as an internal reflection of an expected point in time; thus, the shift of this peak toward a temporally unexpected tone onset (or change marker) is the oscillator’s reaction to an expectancy violation. In other words, if a rhythmically surprising marker occurs within some event, the oscillator can (after the fact) adjust to this unexpected onset. Essentially, the phase of the pulse peak shifts in the direction (in time) of an ill-timed tone; this process of phase correction governs a person’s ability to reactively attend to an expectancy violation. Certain DAT parameters determine how effectively adjustments in phase and period can occur. The values of these parameters are important to understanding how entrainment works and how it describes attentional tracking in real time.16 Thus, values of two parameters determine how well the oscillator of Fig. 3 adjusts its period and phase to align with successive sound onsets. The adaptive process is also driven by event rhythm. Therefore, another major determinant of efficient entrainment (and hence attentional synchrony) is the coherence of the event rhythm itself. A very regular rhythm (e.g. a coherent event such as an isochronous rhythm) results in efficient and speedy entrainment whereas a very irregular rhythm (e.g. a rhythm with ill-timed markers, hence low coherence) results in many occasions for period and/or phase corrections and this slows entrainment. In sum, entrainment is the theoretical vehicle for describing dyadic interaction of an attender with an event. In this dynamic interplay, an event’s time structure functions to continually pace attending, keeping it “in tune.” Effectively, attending is paced by the event time structure such that each adjustment of an oscillator to the timing of a non-temporal marker moves the oscillator closer

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

64

M. R. Jones

to synchrony with the overall event. This analysis suggests that event perception involves a continuous cycle of adaptive attending and perceiving that is driven by a goal of synchronization.

Dynamics of Attending to Pitch The general description of attending and perception that I have just outlined is the launching point for understanding how people respond to various aspects of event structure including both non-temporal markers and temporal intervals. Because DAT assumes that event timing contributes, in a general way, to attending, it is important to illustrate that event timing plays a significant role in performance with a conventional attentional monitoring task. Some evidence already exists to suggest that the relative timing of event markers influences people’s ability to respond to these markers.21–24 Here I briefly describe a recent set of studies to make this point. We presented average listeners with melodies similar to those depicted in Fig. 2. Each melody occurred three times. Listeners were asked to pay attention to the pitch of target tone (the penultimate tone in each recurrent pattern). Thus, this was not a time perception task as listeners were instructed to attend to pitch and not time. In the third repetition of a melody, we sometimes changed the target’s pitch by +/− one semitone. Listeners had to determine if the target tone was the same pitch as in the preceding presentations of the melody or if it was a higher or lower pitch. In the first experiment, the melody was isochronous (i.e. all IOIs were 600 ms); therefore, changes in a target, when they occurred, were always at a rhythmically expected time. In this study, listeners were fairly good at this task. However, in a second experiment, we varied the relative time of the target tone at the IOI level; equally often the target tone occurred early (by 30% of the IOI), on-time, or late (by 30% of the IOI). These time changes affected only the lower time scale; the higher order rhythm remained isochronous. In this second experiment, we found that all listeners were much poorer in identifying the pitch of the target when it was early or late than when it was rhythmically on time.25 Using various melodies (pitch patterns), we found this outcome in events with different melodic structures; however it was somewhat more pronounced in sequences with more coherent melodies. A typical outcome pattern for the simpler melodies is shown in Fig. 4. Sensitivity to a target pitch change (a d  measure) was greatest for on-time targets and significantly lower for the same targets when they occurred early or late in these rhythmic contexts. We think such performance profiles arise because listeners’ attention was entrained to the

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

d'

Dynamic Attending and Prospective Memory for Time

65

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 Early

On-time

Late

Relative Time of Pitch Target

Fig. 4. Results of a pitch monitoring task in which the onset time of a target pitch (see Fig. 2) could be early, on-time, or late relative to a preceding rhythm. Pitch identification performance (measured as d  ) is plotted on the ordinate.

time scale most relevant to this pitch monitoring task, namely the lower (i.e. the basic IOI) level. At this level, entrainment appears to be relatively efficient in that more attending energy was targeted to the rhythmically expected targets than to those occurring early or late. This research leads to several conclusions. First, it is clear that perceiving the pitch of melodic tones may be affected by a rhythmic context. We interpret this to mean that listeners’ peak of attentional energy seems to be directed toward points in time (within an IOI) that are consistent with the established rhythm, leading to best DAT performance in pitch identification at rhythmically expected time points. Second, because the higher order time scale (over groups of three tones) remained unchanged with our variations of relative timing at the level of IOIs (early, on-time, late) we infer that in this tone monitoring task, listeners were relying upon the lower time scale (not a higher level one) to the drive expectancies about “when” future targets should occur. Jones and Boltz14 have labeled such attending to relatively lower levels of event structure analytic attending. Third, irregularities among time spans marked by non-temporal changes function as expectancy violations, and hence enhance difficulty evident in selectively poor performance in judging ill-timed target pitches. Fourth, and finally, by illustrating that rhythmic manipulations affect pitch judgments these data suggest an inter-dependency of non-temporal (e.g. pitch) and temporal relationships in event perception. This inter-dependency is consistent with the idea that people perceive events as integral.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

66

M. R. Jones

Comments on Tasks, Time Scales and Attending Modes The preceding section invites the inference that listeners who perform in a pitch monitoring task are influenced by both task demands and event time structure. This inference deserves some commentary to pave the way to a discussion on time judgments (next section). In DAT, it is assumed that sets of oscillators can be automatically activated by stimulus onsets and related time spans, namely by an event’s time structure. However, the degree of such activation can vary. Contributing to the strength with which a given oscillator is active are instructions and task demands, as well as preceding temporal contexts. That is, in addition to the stimulating activity of an event’s time structure, the task itself can positively reinforce the activity of certain task-relevant oscillators thereby heightening a listener’s reliance on one or another time event/oscillator dyadic pair. Thus, in some cases, oscillators with relatively small periods will facilitate a listener’s ability to “tune into” fast moving parts of a sequence whereas in other cases oscillators with relatively long periods facilitate a selective attending mode to slower moving parts of the sequence. In this way, event structure and task-relevance combine to encourage different attending modes (see Jones and Boltz14 for a full discussion). To illustrate, a pitch monitoring task may engage only one attending oscillation to track a given task-relevant time scale. That is, it requires that people track onsets of successive tones to anticipate a target’s onset time and to judge its pitch. Given the finding that a singular violation of a coherent lower-order rhythm can systematically affect performance in this task, our conclusion that the task-relevant time scale for this monitoring requirement is the lower-level time scale seems a reasonable one. In this case, the appropriate theoretical description is one that involves analytic attending associated with a single oscillator with a period around 600 ms IOI. However, for argument’s sake, now imagine a different task that involves the same melodic event. For instance, if people are told to “pay attention to tone groupings” it is likely they will rely upon time spans longer than the lower order IOIs, namely the time spans that cover successive three-tone groups. Such instructions can contribute to activation of oscillators that are congruent with the time span of a three tone group. To explain attending over various event time spans will require a different, perhaps more complicated, dyadic system. For instance, assuming that task-relevance encourages a listener to use a longer time span, one equal to the duration of a three-tone group, we might enlist a second oscillator with an appropriately longer period. This second oscillator may

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

67

operate in conjunction with the lower order oscillator.1 In this case, if the higher order oscillator is the more active oscillator, then people will be selectively attending in time over higher order time spans. Jones and Boltz14 defined this mode of attending as future-oriented attending. Activation of oscillators with relatively long time periods allows people to anticipate the future course of an ongoing event. With the introduction of future oriented attending, I am bringing into play the concept of attending modes that was developed by Jones & Boltz.14 In 1989, Marilyn Boltz and I argued that two different modes of attending operate in responding to events with many time levels; depending on task relevance, a listener may engage in either future-oriented or analytic attending. Future-oriented attending involves attending selectively over the relatively high time levels of an event whereas analytic attending occurs in tasks where people selectively attend over relatively low time levels of an event. Future-oriented attending occurs if people attend to three-tone groups, whereas analytic attending occurs if people attend to individual tones, as in the pitch monitoring task. Analytic attending means that people only need to anticipate “when” each of the series of tones will occur. In this case, they can rely simply on lower level event time structure. Finally, if one must shift between time levels, then at least two oscillators (one associated with higher event time spans and one associated with lower level event IOIs) are involved. Such two oscillator models have been successfully enlisted to describe tracking certain music-like events (see Experiment 3 of Large and Jones15 ). Finally, although we have seen that the pitch monitoring task is an example of a task in which analytic attending is encouraged, other tasks encourage futureoriented attending as when people attend to three-tone groups. But it is also the case that future-oriented attending will be involved if the active oscillator is still longer than that tuned to three-tone groups. For instance, an oscillator with a period “tuned to” the total duration of a single nine-tone melody will also involve future-oriented attending. In the experiment that I just described, these nine-tone melodies recur; therefore an oscillator would reflect a listener’s ability to anticipate the ending time of each presentation of the melody. What kind of tasks might activate such an oscillator? The answer is certain time judgment tasks, namely ones that require people to judge the time span of an entire event. This is a higher order event time span which can also be considered a long time interval that is filled with a melodic event. Many retrospective and prospective time judgment tasks require that people estimate an event’s total duration; to

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

68

M. R. Jones

the degree that an oscillator relative to an event’s total duration is active when the event occurs, the operation of such a higher order oscillation means that future-oriented attending will be involved in judgments about the total duration of filled time intervals. Often a single oscillator entraining to one or another level of event time structure is sufficient and parsimonious. Mathematical complexities inevitably accompany multiple oscillator models and descriptions of shifting attending modes (Large and Jones15 discuss the N oscillator.a But in some cases, two or three oscillators may be useful despite their added complexity. One advantage of multiple oscillations is that they allow for the modeling of more flexible attending strategies. In these scenarios, we can incorporate combinations of voluntary and automatic attending strategies, namely analytic attending (to lower time scales) and future-oriented attending (to higher event time scales). Examples are given by Jones and Boltz14 (see also Large and Jones15 ).

Time Perception and DAT A theory of attending that is centered upon an analysis of dyadic time holds implications for time perception. Modeling a dyadic system not only addresses performance in conventional attending tasks (e.g. pitch monitoring), but also performance in time judgment tasks. In both kinds of dynamic interactions, a person’s internal timing (biological oscillations) is assumed to lock into and track event time structure (event rate and rhythm) at certain task-relevant time scales. I will suggest that whereas pitch monitoring tasks tend to encourage reliance on lower time scales, and hence on analytic attending, many popular time judgment tasks engage reliance on higher order event time scales, and thus they emphasize future-oriented attending. Of special interest is the role of event time structure in determining judgments about time itself. This approach does not imply that time estimations in a given task will necessarily be veridical. Rather, DAT predictions about over- and under-estimates of time depend on oscillator entrainments to a given time scale within a to-be-judged event. Moreover, as already indicated, people normally do not explicitly judge timing because they use it as part of the event structure that guides attending. In our daily lives, we often find ourselves caught up in a In multiple oscillator systems, two or more oscillators when combined as both relevant to a common task, will by virtue of common activation patterns, come to couple, or entrain to each other, to form a dynamic internal rhythm scheme.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

69

events: We anticipate “when” to reply to a question, “when” to swing a bat, “when” to stop talking and so forth, indicating that our focus is directed toward the question, the bat, another speaker and so forth and not upon preceding time intervals that lead to these timed responses. This also contributes to the fragility of time estimates especially in retrospective (versus prospective) time judgment situations. Indeed, as I indicated earlier, we rarely accurately estimate such time spans; and perhaps for this reason we equip ourselves with watches.

General Comments on Prospective Tasks Versus Retrospective Strategies I have maintained that a default attentional orientation is one in which people use time structure to anticipate various kinds of non-temporal information. Nevertheless, it is useful to examine how listeners render explicit judgments about event time itself. This is what people do in prospective time estimation tasks. In prospective time estimation tasks people are told in advance that they will be asked to judge the duration of a certain (standard) time event time interval. Because of this, a prospective task permits an assessment of people’s reliance on certain time intervals in tasks where the experimenter can specify the time spans pertinent to the task. It offers clear identification of standard and comparison time spans for subjects. Thus, one advantage of the prospective time task is that it affords important experimental control over the task relevant time spans that people may use to render time estimates. In contrast to prospective time estimation, in retrospective paradigms people are not told in advance that they will render a time judgment. Instead, they initially encounter a to-be-judged event (as a standard duration) when they are unaware that they will later judge its total duration. During the standard time interval they often perform a decoy task with information that fills this interval. I suggest that prospective and retrospective designs send, respectively, different messages to listeners about how to use event time structure when they must respond to the initial presentation of an event (i.e. the standard duration) versus its second presentation at recall/recognition (i.e. the comparison duration). The prospective paradigm correctly signals the time span over which people should attend, the memory for which is evaluated in comparative tests using the comparison interval. By contrast, the retrospective paradigm commonly carries an implicit assumption that event time, with its various levels, has little impact on people’s attending during the decoy task. I think this is a debatable assumption.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

70

M. R. Jones

It can be argued that a listener is inevitably attending over certain time levels when performing a decoy task. Moreover, the time levels relevant to a decoy task within a retrospective paradigm are likely not to be those most relevant for a later time judgment. In a typical decoy task, people are busy tacitly “using” whatever time structure, or time scale, is beneficial to performing the assigned (decoy) task. In other words, I suggest that in the retrospective paradigm, people do not ignore time in the decoy task; rather they rely on different temporal referents (time scales) in the decoy versus the time judgment task. To illustrate this point, imagine a retrospective time judgment task in which pitch monitoring (described previously) is the decoy task. Let’s say that this nine-tone melody (Fig. 2), which lasts about five seconds, functions in a decoy task for a retrospective time judgment task. In the decoy task the melody repeats once and listeners are told to detect a change in target pitch in the second presentation of the melody. Next, following the decoy by some retention interval, a retrospective time judgment task is presented wherein the listener is told to determine whether a presented time interval (filled or empty) is shorter, the same, or longer than the duration of the decoy interval (using a rating scale). Note that the time judgment task, as described here, focuses a listener on the total event duration, a higher-level time scale and one that was irrelevant to the performance in the pitch detection (decoy) task. As my earlier discussion of pitch monitoring reveals, from a DAT perspective, the neural oscillations with periodicities congruent with IOIs of 600 ms are typically activated in pitch monitoring, leading to analytic attending in the decoy task. By contrast, different neural oscillators are relevant to the time judgment task where estimates of longer (higher-order) time spans are involved. In this case, task invites future-oriented attending over the total time of the comparison interval. Thus, we find a mismatch between the task-relevant time levels and attending modes. Accordingly, the estimates of total event duration that might be reported by listeners in this retrospective situation may very well be distorted in the direction of the value of the (briefer) period of the neural oscillations used to respond to the standard interval during the decoy task. This leads to a prediction that retrospective judgments about overall event durations will be relatively poor. Moreover, estimation errors will generally be in the direction of underestimation of the true value of a filled duration. Finally, this comparative analysis of retrospective versus prospective paradigms differs from conventional analyses of these tasks. The most familiar theoretical approaches to time estimation share a common assumption that nontemporal and temporal event properties are treated by attenders as independent

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

71

aspects of some event. Therefore, when a to-be-judged time interval is filled with an event, such as a melody or a bouncing ball, the processing of temporal and non-temporal information is considered to independently compete for a common pool of attentional resources. Moreover, in a retrospective format, it is often assumed that people can ignore temporal information when they are responding to non-temporal information in an initial decoy task. Subsequent time judgments about the total duration of the event filling the decoy time interval (i.e. the standard interval) then depend heavily upon the amount of non-temporal (i.e. attended) information remembered at the time of a comparative time judgment. Indeed, it has been successfully argued that various non-temporal event properties, such as the number of elements (e.g. tones) or number of changes (i.e. non-temporal change markers) and so forth, strongly influence time.26–29 By contrast, in prospective tasks, when people are told in advance that they will be judging durations of filled intervals, they may attend to time, but their attention may also be drawn to filler material.30,31 Thus, in prospective tasks as well as retrospective ones, intervals filled with more complex material can lengthen perceived time.32,33 However, with explicit dual task scenarios in which people must simultaneously judge total duration and overtly perform a concurrent (e.g. decoy) task, interference from the concurrent task is often observed such that responding to non-temporal filler information produces more interference with complex material than with less complex fillers; in turn, there is greater underestimation of total durations in the former case than in the latter.31,34 A variety of explanations have been offered for this complicated pattern of findings. One approach posits that the number of non-temporal changes experienced by people affects perception/memory for a time span.26,27,32,35,36 Other prominent positions incorporate a clock-counter to explain how people comply with instructions to judge time as such.37 The clock comprises a pacemaker-pluscounter which fills a to-be-estimated interval with accumulated ticks (providing a numerical code for the total duration). Combinations of these approaches are also found. Thus, the extent to which the event and/or the concurrent task encourages attending to non-temporal information filling an interval dictates (presumably) how much attentional resources are drawn from a clock-counter. Specifically, when resources are sapped, then clock counts are lost and the event’s total duration is underestimated.28,38,39 Finally, DAT offers a still different interpretation, one which places greater emphasis on time itself than other approaches. DAT does not assume that time is converted into non-temporal codes based upon either the number of

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

72

M. R. Jones

non-temporal changes experienced or on the number of clock ticks counted. Rather, this theory attempts to provide a description of event time perception which is compatible with an individual’s real-time attentional monitoring of natural events. Moreover, in light of its emphasis on the relative time structure of events, it places greater emphasis on the role of temporal coherence and the temporal distribution of change markers within an event than upon the number of changes alone. Essentially an entrainment view invites the following questions: “How coherently are event markers distributed in time within events that fill the to-be-judged time spans?” And “What level(s) of time is (are) used in attending to standard and comparison time intervals?”

Prospective Time Judgments With prospective time judgment tasks, DAT entrainment models stress the role of attentional synchrony in responding to events to explain how people judge time. For comparative judgments, DAT assumes that attending and accompanying synchronies in attending will be associated with time levels that are task relevant. Specifically, time judgments are determined by violations of temporal expectancies involving task-relevant time spans. The temporal expectancies at play depend both on the task and upon the time structure of the event. This idea captures the intuition that if something (a non-temporal marker either within or terminating an event) arrives sooner than anticipated, the relevant time span will be experienced as comparatively short. Typically, a standard interval supplies the time level (or levels) upon which temporal expectancies about a comparison time interval are based. Thus, when people are told to judge the total duration of a standard interval filled with some event, the relevant time span is a higher-order time span of that event. In tasks where people must judge whether a comparison time interval is shorter or longer than a standard (or shorter, the same, longer), their performance will be based on temporal expectancy violations provided by a comparison’s ending time taken relative to an expected ending based upon the preceding standard (all other things being equal). In other words, the asynchrony associated with the difference between an expected ending time and an observed ending time of comparison time interval systematically influences time judgments about the duration of a comparison: comparison time intervals that end late relative to an expected ending time are judged long relative to the standard whereas those that end prior to an expected ending are judged as shorter in overall duration. Differences in performance due to task and instructions, among other things, are reflected in values of parameters (e.g. period, phase

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

73

parameters) associated with oscillators of different periodicities. This reasoning applies generally to both filled and unfilled time intervals in prospective time estimation tasks. (1) Judging total durations of filled intervals. Questions which usually arise in judging durations of filled intervals concern the influence of material filling an interval on judged durations. As we have seen, people often systematically err in judging two identical time intervals as a function of what fills these intervals. This is related to the filled interval effect, which roughly states that filled intervals are judged to be longer than unfilled intervals. Although this is often true, a corollary of the filled interval effect is more problematic. It states that filled intervals containing more information (complexity, number of changes, etc.) are judged longer than those with less information. Although there is evidence favoring this position, mixed results have also been reported.40 A problem comes with operationalizing “more information.” For instance, Marilyn Boltz has shown that time estimates of rhythmically coherent events differ significantly from those of rhythmically incoherent ones.40,41 Thus, for natural events, it appears that the number of non-temporal changes is less critical with regard to event complexity than the relative timing of non-temporal changes. More complex events are ones that exhibit more expectancy violations. Any ill-timed, non-temporal change within an event provides an expectancy violation, which, in turn, can yield a perceived shortening (if it arrives unexpectedly early) or lengthening (unexpectedly late) of time spans. Thus, careful analysis of event structure and its possible impact on an attender is critical to predictions about filled intervals in this approach. To simplify matters, in the present discussion I focus only upon the role of the most important temporal expectancy violations, namely those which happen at the end of some event which fills an interval. This was manipulated in a study that Marilyn Boltz and I undertook together. We filled identical time intervals with folk tunes and asked listeners to judge their relation (total) durations. Some melodies were rhythmically coherent and others were not. Nevertheless, standard and comparison intervals contained the same number of items (tones) within a melody, complexity (number and type of melodic groups), and number of non-temporal changes in well-matched pairs of melodies. None of these factors explained performance as well as an account that emphasized people’s expectancies about when a comparison melody might end (in time) given a preceding standard melody. Systematic differences emerged in judged durations of these events due to rhythmic coherence and listener’s ensuing expectancies

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

74

M. R. Jones

about “when” a comparison melody should end. A coherent melody that seemed to end earlier than expected was incorrectly judged shorter than a similar comparison tune that ended on time.14 We assumed that depending on the task and the coherence of the folk tunes, listeners would use event structure to rely on either higher event time scales (i.e. future-oriented attending), or on lower time scales (i.e. analytic attending). When a rhythmically coherent melody filled the time interval people would readily generate expectancies about “when” a tune would end based on future oriented attending. In these melodies we predicted that performance would depend on a perceived difference between expected and observed ending times. The basic idea is that the temporal distribution of information within an interval, together with the assigned task, can induce a listener to anticipate (incorrectly) that a comparison event will end early or late relative to a standard time interval. This in fact was the case. With rhythmically incoherent melodies, analytic attending was more likely; people were less capable of anticipating an ending time and hence less likely to “tune into” the higher order time span and hence to judge total duration on the basis of perceived differences in ending times. In subsequent research, Jim Klein, Marilyn Boltz and I used much simpler tone patterns. These were sequences comprising four melodic phrases that differed in length; the beginnings and ends of phrases were clearly marked by recurrent pitch changes.42 Listeners estimated the relative duration of standard and comparison pairs (of total durations). We used different combinations of melodies in which pairs were either truly equal in total duration or in which the two melodies veridically differed in total duration. Again, we found that even when two melodies were of exactly the same total duration, people erroneously judged them to differ based upon melodic phrasing involving the final tone. The magnitude and direction of an expectancy violation based on the final phrase length predicted performance. For instance, if the last phrase of a four phrase sequence was longer than preceding ones, listeners judged the whole melody to be longer than a standard sequence of exactly the same total duration that ended with a shorter final phrase. This underscores the role of event structure in guiding attending during a filled interval and it highlights the importance of implied ending times for time estimations. This research also offers converging evidence that people can generate expectancies about “when” an event ends. In some experiments listeners were presented with melodies that were truncated in various ways (e.g. by eliminating several final tones) and they were asked to estimate the ending time, given the

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

75

preceding pattern. In general, they were quite good at this even projecting over relatively long silent intervals. Thus, although people had to delay their response for long silent intervals (ranging up to 22 seconds), nonetheless, they could fairly accurately pinpoint an implied ending time. The fact that they could do this suggests that listeners were directly sensitive to the time structure associated with longer (higher order) time scales within these events (i.e. as in future-oriented attending) and were not attending on a tone-by-tone basis (analytic attending). (2) Judging durations of unfilled time intervals. A critical question that arises with unfilled intervals concerns how people judge empty time intervals. By definition no concurrent or interfering task/material fills the relevant time spans. But again comparative time estimation is involved such that a pair of empty time intervals is usually presented for comparison. In this case, one theoretical issue concerns whether or not people use a clock that enables a numerical code (based on accumulated ticks) for each interval, thus converting time into a nontemporal code. This is the explanatory mechanism embraced by the majority of clock-counter theorists. Alternatively, it is possible that people do not enlist a pacemaker but rather directly engage with the unfolding series of time spans. This is the approach implied by DAT. Objectively, in experiments that have employed prospective judgment tasks to study perception of empty time intervals, the intervals involved tend to be shorter, on average (e.g. under 2 seconds) than time intervals that appear in experiments designed to study perception of filled time intervals where the time spans contain material such as melodies or tasks or some other activity; the latter tend to longer time spans (e.g. well over 2 seconds). In light of this, we may speculate that findings derived from the filled interval paradigms, where people must judge total event duration, are most relevant to judgments about the highest time levels of everyday events, whereas findings emerging from research using (short) empty intervals may be more relevant to our understanding of how people assess the lowest time levels of everyday events. With this in mind, many of the experiments that I report consider how people judge a standard versus a comparison time interval when these follow a rhythmic context. In this section I concentrate on recent work on prospective time judgments that I undertook with Ralph Barnes and Devin McAuley. We asked people to judge the duration of individual comparison time intervals relative to a preceding standard time interval where both intervals where unfilled. The intervals were relatively brief ranging from around 300 ms to 800 ms.43 The standard time interval may be either a solitary time interval, marked by two brief tones of the same pitch, or it could be the last time interval in a series of empty time intervals

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

76

M. R. Jones

that form a rhythmic event. In either case, the standard interval always preceded a comparison time interval that was yoked to it. Of course, among the most common experimental situations is one that is least likely to characterize real world events; it occurs when no serial context is given prior to a standard/comparison pair. We consider this case as well as others involving various context rhythms. (I should mention that in most of these studies the comparison time interval is either the same as the standard or shorter or longer than the standard by a supra-threshold time difference,
t; that is,
t/IOI = 0.10 to 0.12 where the threshold in such situations is known to be around 0.04.) Our main interest centered on the role of serial context on time judgments. In particular, from a DAT perspective we wondered whether a rhythmic context would systematically distort time judgments. According to this theory, time judgments will be systematically distorted whenever a standard does not “fit” with a preceding rhythmic pattern. The basic designs we used are outlined here. We either varied the ending time of a standard time interval (Fig. 5a), given some rhythmic context, or we varied the onset time of a comparison interval (Fig. 5b). The former is more directly relevant to time judgments about the standard and so I focus here on these experiments; generally speaking manipulations of the onset time of a comparison (Fig. 5b) only harmed performance when these changes were large, suggesting the need for significant phase adjustments of the entraining oscillation. In most experiments we followed an isochronous rhythmic context (a recurrent IOI of 600 ms) with a to-be-judged standard/comparison pair of time

Fig. 5. Two common paradigms for manipulating both context rhythms and relative timing of standard and comparison IOIs in time estimation tasks.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

77

intervals. The trick here is that the standard interval, to which a comparison is always yoked, is either identical to the prevailing IOI of the preceding rhythm (i.e. a standard of 600 ms) or it is shorter or longer than 600 ms (by suprathreshold amounts). In some experiments, three different standard intervals were used; in others, five standards were involved. In either case, only one of these intervals “fit” the prior serial context, i.e. conformed to rhythmic expectancies based on the context rhythm. People were always told to ignore the context rhythm and focus on the task-relevant standard time interval. The other four standards violated temporal expectancies in different ways, that is, ended earlier or later than expected. The comparison interval that followed a standard was always yoked to that standard time interval. In most tasks, people judged whether a comparison was the Same, Shorter, or Longer than the preceding standard. Thus, the aim was to assess how well people could remember a rhythmically unexpected standard time interval (i.e. relative to a rhythmically expected one). The most basic question was: Does the presence of a to-be-ignored serial context systematically affect time judgments? The answer appears to be “Yes.” Ralph Barnes and I found that people have trouble determining durations of rhythmically unexpected time intervals.44 Thus, given a simple isochronous rhythmic context, we found that listeners were best at judging standard/comparisons that involved rhythmically expected standards (given the context) and worst at judging those standard/comparison pairs that involved rhythmically unexpected standards. Figure 6 shows performance in Observed Expectancy Profiles Proportion Correct, PC

0.9 0.8 0.7 0.6 0.5 0.4 0.3 Very Early

Early

Expected

Late

Very Late

Standard Ending Time

Fig. 6. Results from Barnes & Jones (2000) indicating accuracy (PC) in time judgments of expected and unexpected standards, given a preceding rhythmic context.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

78

M. R. Jones

an experiment involving five different standard durations which ranged from from a very short standard duration of 524 ms to a very long one of 676 ms; the expected standard was always 600 ms. The quadratic trend of accuracy scores (Proportion Correct, PC) as a function of standard duration is significant; we refer to this trend as an expectancy profile. This profile strongly suggests that the rhythmic context induces expectancies that, when violated, promote systematic distortions in prospective time judgments. Moreover, it suggests that the sequence of inducing time intervals encourages analytic attending at the level of individual IOIs. It is possible that the resulting distortions of standard time intervals are simply due to a range effect associated with the set of different standards/comparisons encountered in a session and not the result of temporal expectancies induced by the preceding rhythmic context. However, this explanation is ruled out by a control study. The control study44 removed the serial (rhythmic) context which preceded each standard time interval but left the range of standard/comparison pairs intact. This design resembles the two interval (standard/comparison) condition most common to many psychophysical designs where no rhythmic context is present. In this context-less condition, we found a small, very weak, quadratic trend that could be attributed to the range of the standard durations; although statistically reliable, this trend component was significantly flatter than the expectancy profile observed with the rhythmic context present. This reinforced our conclusion that rhythmic context is a major determinant of expectancy profiles in these prospective time judgment tasks. Another question was: Does rhythmic context provide evidence for underlying periodicities proposed to operate in entrainment? Again we found an affirmative answer. This is revealed in the research with Ralph Barnes44 and Devin McAuley.16 Barnes and Jones presented isochronous rhythms of three different rates to, respectively, three different groups of listeners who received sequence rates based on IOIs of 300, 500 or 600 ms. However, all three groups received the same set of three different standard durations (524 ms, 600 ms, and 679 ms) in the same time judgment task as described above. In this task, people heard only comparisons that were shorter or longer than a standard and they were asked to respond Shorter or Longer; we calculated a non-parametric measure of ROC, Ag.b

b Ag

is a non-parametric measure that combines evaluation of confidence ratings with detection in two alternative forced choice tasks. It parallels A , but is used with non-parametric data.46

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

79

According to DAT, if the initial IOIs of a context sequence activate in listeners certain neural oscillations that have periods that approximate these IOIs, then different attending (neural) oscillators will be used by listeners to track events in the three different rate conditions. Moreover, if such internal periodicities entrain to the rhythm of each serial context, then DAT predicts quadratic expectancy profiles in the 300 ms rhythm as well as the 600 ms rhythm but not in the 500 ms rhythm condition. This is because the temporal offset of a 600 ms standard can be conceived as “expected” for both an oscillator with a period of 600 ms and for an entrained oscillator with a rate that is twice as fast (1/2 of 600 ms as in the 300 ms IOI group). However, this argument does not hold for an oscillator of a 500 ms period. The Ag data for this experiment are shown in Fig. 7. As predicted, significant quadratic trends appeared for the two groups of listeners whose rhythmic contexts should have induced internal periods congruent with 600 ms standards, but not for the group with a rhythmic context based on IOIs of 500 ms. This finding was replicated for the 600 and 300 ms conditions by McAuley and Jones.16 We note that our finding is difficult for a clock-counter approach to explain because its predictions are based on interval (numerical) time codes. According to clock-pacemaker theories, similarity among rates is gauged on an interval not a ratio time scale; therefore, in these accounts the group of listeners receiving the 500 ms IOIs should perform more similarly to the 600 ms group than to listeners experiencing the 300 ms IOIs. Related discussions of ratio versus interval time scales are found in the literature.8,14,16,44,45

Observed Expectancy Profiles Three Time Judgment Conditions 1

300 500 600

0.9

Ag

0.8 0.7 0.6 0.5 0.4 524

600

676

Duration of Standard Fig. 7. Results for Barnes & Jones (2000) indicated sensitivity, Ag, to time changes (standard versus comparison) when a standard ended either as expected or was unexpected given one of three different (isochronous) context rates.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

80

M. R. Jones

A simple entrainment model, developed by Devin McAuley, fit the data of Barnes and Jones and related findings quite well. It predicted that time judgments in these tasks are based on the magnitude and directional discrepancy of an expectancy violation that is experienced when a comparison time interval ends. Basically, given a rhythmic context which includes a standard time interval, whenever a standard interval ends prior to an expected point in time (based on a continuation of the prevailing rhythm) then people will be biased to judge it shorter than an equivalent comparison time interval. Similarly, if the standard ends after the expected ending time, then listeners will judge it “longer” than an equivalent comparison. An interesting aspect of this account is that it implies that attentional monitoring, set in motion by certain rhythmic patterns, can grossly distort a person’s memory of the standard. Thus, the rhythm preceding a standard systematically biases comparative time judgments. In the McAuley and Jones16 experiments, we found that these biases persisted even with context rhythms containing a relatively long empty time interval (an interval equal to two IOIs) prior to the standard. This suggests a lawful persistence of attending periodicities that are induced by rhythmic invariants in the opening segments of the context rhythm. Clearly, DAT offers a different approach to how we perceive and remember event timing than what is found in traditional clock models. For DAT, the internal expression of a duration is a periodic oscillation whereas for clock-pacemaker models, event durations are quantized internally as interval codes.16 Moreover, practically speaking most clock models address situations where people must judge two time intervals (standard and comparison) whereas entrainment models address situations where people respond to a series of time intervals. Judgments about one or two time intervals from the perspective of entrainment theory are always referenced to the impact of some surrounding temporal context on a listener. Although it has not been clear that a clock-accumulator model can readily explain such context effects, McAuley and Jones developed several plausible versions of the clock model in an attempt to discover if clock models could be extended to address observed effects of serial context on time judgments. These extensions involved various averaging algorithms of the numerical codes associated with context IOIs. With additional parameters such extensions of the clock model approach performed adequately. Nevertheless, in a final accounting all of the clock model adaptations performed more poorly than a simpler entrainment model. In sum, analytic attending and related expectancies are induced by a series of empty time intervals. These expectancies lead to systematic distortions of

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

81

certain deviant time intervals in ways that suggest the operation of a persisting internal oscillation that entrains to sequential events. Because the entrainment model places a heavy explanatory burden on attending and expectancies, one might wonder how memory for the standard can be addressed in such a view. I end this chapter with a discussion about how memory, which is a crucial element of both retrospective and prospective paradigms, is conceived in DAT.

Memory in DAT A common way of thinking about memory involves some sort of storage and retrieval mechanism. This approach is typically applied in interpretations of retrospective and prospective time judgment tasks in that what is stored in memory is either an accumulation of the non-temporal information (in retrospective paradigms) and/or an accumulation of clock ticks (in prospective paradigms). In both cases, what is remembered is coded such that time qua time is lost and replaced by static code. Judgments ultimately depend on such memory codes for standard and comparison time intervals, which are retrieved and compared during a decision stage. At this point, a comparative judgment entails an abstract re-conversion of these numerical codes (and their difference) back into the temporal terminology of time judgment tasks (e.g. “shorter,” “same,” or “longer”). But finally, the important theoretical point is that in both retrospective and prospective memory situations, the stored memory traces are inherently non-temporal. Event time is converted into a non-temporal code for storage purposes. This clock-counter approach presents a conceptual contrast with the way in which memory is conceived in an entrainment view. From the perspective of DAT, event time does not need to be converted into a non-temporal code for storage purposes because the attender is inherently temporal. In fact, the oscillator effectively carries along a memory, albeit not always an accurate memory, of the standard time interval. In prospective tasks, we have proposed that time judgments at the point of making a decision depend on an individual’s momentary attentional state: Did the comparison end as expected or not? Although all comparative time judgment tasks require some sort of memory of the standard, it is not necessary to assume that this memory is abstracted from a dynamic context or that a memory for duration must be timeless. Although DAT engages a memory process, in entrainment models, the memory for a time interval is nestled in the period of an entraining oscillator. This memory is quite dynamic. It modulates in real time, changing from moment to moment as the oscillator adapts to new time intervals in an event.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

82

M. R. Jones

This dynamic approach to prospective memory has been illustrated by McAuley and Jones.16 We assumed that the period of an oscillator, entrained at the level of recurrent IOI in our task, could adapt to changes in the rhythm of a serial context that included a standard IOI, but that period adaptation was gradual. By gradual we mean that if an isochronous rhythm is expressed by a series of 4 or 5 IOIs of 600 ms and then followed by a standard IOI of 540 ms, an entraining oscillator tuned to 600 ms would only adjust its period slightly downward after its initial encounter with a single IOI of 540; if more intervals of 540 ms occurred, presumably the period would come somewhat closer to this value, but very slowly. This period adaptation is somewhat sluggish. This means that an entraining oscillator could expand or contract its period a bit, but not much. A sluggish entraining oscillator thus may take many trials to match its period to a fixed, recurrent, cycle within an event. The degree of adaptivity (i.e. sluggish versus speedy) is governed by a period adaptation parameter of the model. This adaptive property is important because it determines an oscillator’s functional memory of past event cycles. In large measure, an oscillator with sluggish period adaptivity is tied to the past. Because it is limited in its adjustment to new time intervals, a sluggish oscillator will not remember the most recent event cycles very well. Alternatively, if period adaptation were very rapid (i.e. speedy period adaptation), then this would imply efficient period adjustment by the oscillator to any time interval that violates a rhythmic expectancy. In this second scenario, memory is weighted in favor of the more recent time spans and less weight is given to prior rhythmic context. In fact, using one version of this type of DAT model, we were able to show that very rapid period adaptation predicts that rhythmic context will have little effect on people’s memory for a standard time interval.16 This provided evidence for our claim that performance in tasks that offer no rhythmic context and require single interval time estimates can be described as a special case of DAT. However, in coherent rhythmic contexts, our data appear to cast doubt on the speedy period adaptation scenario. In several different research projects we have found that it is fairly difficult for people to truly ignore such a rhythmic pattern.15,16,44 Instead, the rhythm that precedes a standard often has a substantial effect on how well listeners remember that standard. For this reason, we assume that the oscillator is sluggish in period correction following an unexpected time interval (i.e. an unexpected standard). Effectively, the oscillator is “remembering” earlier time intervals in a sequence better than the most recent one. This assumption leads to predictions that are consistent with the data that

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

83

I have reported, namely that people are more accurate when a standard time interval is similar to the IOIs of the rhythmic induction sequence than when it is very different from induction IOIs. In conclusion, the DAT approach to memory reflects a general feature of this approach to timing, namely that attention, perception and memory are inextricably bound together.8 Attending can be anticipatory, thereby affecting what is perceived and how it is perceived; and memory for event time is based on those aspects of the original act of attending/entrainment to that event which persist over a retention interval. Clearly, memory is not independent of attention and this connection is realized in DAT. As a result, there are no requirements to translate an experience with an event into a non-temporal code, to retrieve a stored code and finally to reverse the translation of this code to arrive at a time judgment. Instead, oscillations simply persist as memories of a past attending/perception episode; as such they continue in a more or less veridical form to affect one’s judgments about a later event. At a general level, this type of approach implies that memory and attending are both inherently temporal.

Acknowledgements The author is indebted to Robert Ellis, Joseph Glicksohn, and Katy Rubia who commented on earlier versions of the chapter. Research reported in this chapter has been supported by a grant from the National Science Foundation to Mari Riess Jones (BCS 980446).

References 1. 2. 3. 4. 5.

6. 7. 8.

Johnston W, Dark V. Selective attention. Annual Review of Psychology 1986; 37: 43–75. Broadbent D E. Perception and Communication. New York: Pergamon, 1958. Egeth H, Yantis S. Visual attention. Annual Review of Psychology 1997; 48: 269–297. Treisman A, Gelade G. A feature integration theory of attention. Cognitive Psychology 1980; 12: 97–136. Jones M R. Temporal expectancies, capture and timing in auditory sequences. In Gibson C F B, ed. Attraction, Distraction, and Action: Multiple Perspectives on Attentional Capture. Elsevier Science, B.V., 2001. Gibson J. Events are perceivable, but time is not. In Fraser J, Lawrence N, eds. The Study of Time II. New York: Springer-Verlag, 1975: 295–301. Fraisse P, ed. Time and Rhythm Perception. New York: Academic Press, 1978. Jones M R. Time, our lost dimension: Toward a new theory of perception, attention, and memory. Psychological Review 1976; 83(5): 323–355.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

84

M. R. Jones

9. McAuley D J. Perception of time phase: Toward an adaptive oscillator model of rhythmic pattern processing. Unpublished dissertation. Indiana University, 1995. 10. Snyder J S, Large E W. Gamma-band activity reflects the metric structure of rhythmic tone sequences. Cognitive Brain Research 2005; 24(1): 117–126. 11. Snyder J S, Large E W. Tempo dependence of middle- and long-latency auditory responses: Power and phase modulation of the EEG at multiple time-scales. Clinical Neurophysiology 2004; 115(8): 1885–1895. 12. Large E W, Crawford J D. Auditory temporal computation: Interval selectivity based on post-inhibitory rebound. Journal of Computational Neuroscience 2002; 13(2): 125–142. 13. Large E W, Fink P, Kelso J A. Tracking simple and complex sequences. Psychological Research 2002; 66(1): 3–17. 14. Jones M R, Boltz M. Dynamic attending and responses to time. Psychological Review 1989; 96: 459–491. 15. Large E W, Jones M R. The dynamics of attending: How people track time-varying events. Psychological Review 1999; 106(1): 119–159. 16. McAuley J D, Jones M R. Modeling effects of rhythmic context on perceived duration: A comparison of interval and entrainment approaches to short-interval timing. Journal of Experimental Psychology: Human Perception and Performance 2003; 29: 1102–1125. 17. Rinzel J, Terman D, Wang X, Ermentrout B. Propagating activity patterns in large-scale inhibitory neuronal networks. Science 1998; 279(5355): 1351–1355. 18. Ermentrout G B, Rinzel J. Beyond a pacemaker’s entrainment limit: Phase walk-through. American Journal of Physiology 1984; 246: R102–106. 19. Winfree A T. The Geometry of Biological Time. Second edn. New York: Springer-Verlag, 2000. 20. Large E W, Kolen J. Resonance and the perception of musical meter. Connection Science 1995; 6: 177–208. 21. Jones M R, Boltz M, Kidd G. Controlled attending as a function of melodic and temporal context. Perception & Psychophysics 1982; 32: 211–218. 22. Jones M R, Pfordresher P Q. Tracking musical patterns using Joint Accent Structure. Canadian Journal of Psychology 1997; 51: 271–290. 23. Jones M R, Moynihan H, MacKenzie N, Puente J. Temporal aspects of stimulus-driven attending in dynamic arrays. Psychological Science 2002; 13: 313–319. 24. Dowling W J, Lung K M, Herrbold S. Aiming attention in pitch and time in the perception of interleaved melodies. Perception & Psychophysics 1987; 41(6): 642–656. 25. Jones M R, Moynihan H, Puente J. Effects of auditory pattern structure on anticipatory and reactive attending. Cognitive Psychology in press. 26. Block R A. Temporal judgments and contextual change. Journal of Experimental Psychology: Learning, Memory & Cognition 1982; 8: 530–544. 27. Block R A. Remembered duration: Effects of event and sequence complexity. Memory & Cognition 1978; 6: 320–326. 28. Block R A. Cognitive models of psychological time. In Block R A, ed. Models of Psychological Time. Hillsdale, N.J.: Erlbaum, 1990: 1–35. 29. Ornstein R E. On the Experience of Time. New York: Penguin Books, 1969.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

Dynamic Attending and Prospective Memory for Time

85

30. Brown S W. Time perception and attention: The effects of prospective versus retrospective paradigms and task demands on perceived duration. Perception & Psychophysics 1985; 38: 115–124. 31. Brown S W. Attentional resources in timing: Interference effects in concurrent temporal and non temporal working memory tasks. Perception & Psychophysics 1997; 59: 1118–1140. 32. Brown S W. Time, change, and motion: The effects of stimulus movement on temporal perception. Perception & Psychophysics 1995; 57: 105–117. 33. Poynter W D, Homa D. Duration judgment and the experience of change. Perception & Psychophysics 1983; 33: 548–560. 34. Brown S W. Automaticity versus timesharing in timing and tracking dual-task performance. Psychological Research 1998; 61: 71–81. 35. Block R A. Psychological timing without a timer: The roles of attention and memory. In Hede H, ed. Time and Mind II: Information Processing Perspectives. Ashland, OH: Hogrefe & Huber, 2003. 36. Michon J A. Processing of temporal information and the cognitive theory of time experience. In Fraser J T, Haber F C, Muller G W, eds. The Study of Time. New York: Springer-Verlag, 1972. 37. Gibbon J, Church R M, Meck W H. Scalar timing in memory. In Gibbon J, Allan L G, eds. Annals of the New York Academy of Sciences. New York: New York Academy of Sciences, 1984: 52–77. 38. Zakay D, Block R A. New perspective on prospective time estimation. In De Keyser V, d’Ydewalle G, eds. Time and the Dynamic Control of Behavior. Ashland OH: Hogrefe & Huber, 1998. 39. Block R A, Zakay D, Hancock P A. Human aging and duration judgments: A meta-analytic review. Psychology and Aging 1998; 13(4): 584–596. 40. Boltz M G. The processing of temporal and nontemporal information in the remembering of event durations and musical structure. Journal of Experimental Psychology: Human Perception and Performance 1998; 24: 1087–1104. 41. Boltz M G. Effects of event structure on retrospective duration judgments. Perception & Psychophysics 1995; 57(7): 1080–1096. 42. Jones M R, Boltz M, Klein J M. Duration judgments and expected endings. Memory & Cognition 1993; 21: 646–665. 43. McAuley D J, Kidd G R. Effect of deviations from temporal expectations on tempo discrimination of isochronous tone sequences. Journal of Experimental Psychology: Human Perception and Performance 1998; 24: 1786–1800. 44. Barnes R, Jones M R. Expectancy, attention, and time. Cognitive Psychology 2000; 41: 254–311. 45. Gilden D L, Schmuckler M A, Clayton K. The perception of natural contour. Psychological Review 1993; 100(3): 460–478. 46. Davison T C B, Jagacinski R J. Nonparametric analysis of signal detection confidence ratings. Behavior Research Methods & Instrumentation 1977; 9: 545–546.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch03

This page intentionally left blank

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

4 Representing Times of the Past, Present and Future in the Brain Wim A. van de Grind∗

Introduction and Overview People have often wondered about the timing of their experiences relative to events in the physical world and more recently also relative to brain events. How can we play ball with a brain full of sluggish neurons? How can we consciously initiate actions if we first become aware of them when they are already in progress, as some reports (Libet14,15 ) have us believe? Do we have neural stopwatches to note and remember durations of events, and how are these durations coded in the brain? Here, I present a systematic approach to such timing problems. Timing must be based on two ordering relations (simultaneity detection and temporal order judgement) and a metric. I will concentrate on simultaneity detection and a duration metric. In an analysis of the so-called flash-lag effect we will see that it is sometimes possible to use information on distance and speed to time experiences. The idea will be made concrete with a simple network model. Next it will be argued that prospectively planned actions can be largely automatic, despite the fact that they are voluntary. This eliminates Libet’s anomaly of a voluntary act without a preceding conscious decision to act. A central problem in theories on memory of time is the coding of time. ∗ Functional

Neurobiology of Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands; e-mail: [email protected]

87

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

88

W. A. van de Grind

I suggest that a labelled line code would be a proper choice and present a simple network model based on such a code to remember and reproduce durations.

Prospective Memory and Time Prospective memory (ProM) allows us to execute a planned action A at some future moment of time. In a world without clocks and calendars the specification of this future moment of time will usually be some cue C, some event or state, such as a full moon or after dinner. This relies on the capacity of animals to delineate and remember experienced events in time and space.1 The association of cue C and planned action A can be described as a temporary sensorimotor program of the form “if C, then A.” Usually, the perceptual program that allows us to recognize C will already exist, because C is a concept or class of events that we know from experience. Therefore, C will be available in explicit memory. One might think that cued forms of ProM could work almost without any time sense of the organism. However, concepts like “after dinner” or “at sunrise” imply the capacity to make temporal order judgments (TOJ) and simultaneity judgments, respectively. These are two of the three basic requirements for a time measurement system. Longer intervals than those between natural events must have been important in evolution. We marry after four full moons, for example, means the introduction of a metric for time-intervals, which is the third and final requirement for setting up a timing system. Cue-counting is a form of metric for timing. Such a counting process also underlies our modern time metric, the cesium clock, but other options exist. A metric allows us to also measure long durations (see section on “Timing at the Output Side”). Fortunately there are several physiological clocks to facilitate survival and help us with the longer durations. If a certain time of day is most favorable for fishing in a certain lake or sea, experienced fishers — be they human or other animals — will appear around that time. This can be done without a technical time piece. Pigeons in the lab can learn to peck a key every morning at a fixed time, and we can learn to wake up at a fixed time every morning. All this can be accomplished with clock-like processes in the brain, allowing ProM without explicit perceptual cues or external clocks. To make this possible, our brains must and do handle the basics of time measurement: (a) simultaneity judgments, (b) TOJs, and (c) duration estimation (a metric).

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

89

The Problems of Neural Space-Codes, Time-Codes and Temporal Binding I will concentrate on the timing of our plans, percepts, volitions and actions. Below in “Space, Time and Motion” we will take up the most basic problems, those concerning possible neural representations and person-level experiences of time information. Everybody agrees that a red ball is not coded as a red spherical activity pattern in the brain, in other words experienced colors or forms are not coded as colors or forms in the nervous system. Why then do so many neuroscientists expect that physical positions of stimuli map onto the position of neurons, and physical timing of stimuli onto physical time patterns of brain activity? In electronic equipment, functional interactions of harnessed electromagnetic waves do the work. The position of wires and nodes in an electronic circuit are only relevant to the designer, who tries to minimize parasitic effects, not to the intended functions of the working system. Electronic systems do not observe their own layout, only external observers can do that. Why should that be different for the nervous system? A “somatotopic” mapping might be useful during ontogenesis, to keep connections short and genetic specification simple, but there is no homunculus to read those maps. This is Lotze’s local sign problem.2 An excellent theoretical treatment and solution to the problem is available.3–5 However, because neuroscience has totally neglected the problem, we can only speculate on how the brain handles local signs. Position perception in vision can be dissociated from the retinotopic map in area V1, the primary visual cortex.6 This shows that the map position of an activated neuron is not a reliable perceptual position label. At the other end of the brain, in the prefrontal cortex, the activity of certain cells or cell groups has been found to stand for a remembered visual position.7–9 It follows, that the brain has at least one systematic representation of perceived spatial positions, a local sign map, that is, a neural correlate of consciously perceived positions. I will therefore take the idea for granted, that the brain has an internal, activitycoded, representation of positions and spatial relations, that is not tied in any fixed way to somatotopic maps. Local sign is a feature just like color or pitch. Of course, if this internal representation is in some sense veridical and somatotopic mappings are also in some sense veridical, one can expect a good correlation between the two in many circumstances. However, this is because of a common source, the physical spatial order of our living space to which we are tuned

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

90

W. A. van de Grind

during ontogenesis. It is not because some central homunculus is checking the positions in the brain of active neurons. The analogous problem of time coding is even less well recognized. Most scientists assume that neural timing directly represents or equals physical timing, and that experiential timing mirrors neural timing. We will see in a later section, that this is certainly wrong. For vision, we know that there are special motion sensors, a finding that started Gestalt psychology. When timing our actions, the brain could therefore in principle use spatial and speed information, rather than time. Let us look at an example. Lee and Reddish10 showed that plummeting gannets use “time-to-contact” (TTC) information to choose the right moment to fold their wings just before hitting the water surface. David Lee has shown in dozens of papers that TTC information is also used to time all sorts of actions in humans, from crossing the road to athletic competitions. TTC can be determined from the ratio of object size to image expansion speed. Therefore, one can design a TTC-network model11 that generates the gannet’s wingfolding command at some TTC-threshold value, just by tracking image size and expansion speed. Classical physics describes speed as distance over time, but there is no reason why our brain could not consider time as distance over speed. The visual system, which I will use as a paradigm perceptual system, is known to have a strongly parallel structure, with all sorts of features processed at different stages in different processing streams. This must lead to a significant temporal dispersion of related information. There are, for example, separate cortical processing areas for color (area V4) and motion (area V5). Suppose we look at a visual scene with a striped flag moving irregularly in the wind, then how do V4 and V5 “agree” which color goes with which stripe at any given time and how they temporally synchronize their changes? This is the notorious binding problem, which is entwined with the space and time coding problems. We are able to judge simultaneity versus asynchrony of separate features of visual images (see the section on “Timing of Perceptual Experiences”), although we know that these features are represented in different brain areas. Therefore, we must have at least one brain program12 that analyzes temporal relations between activity patterns in various brain areas, and makes the results reportable (conscious) at the person level. The relevance of such timing comparisons across features and senses will be discussed in more detail in a later section.

The Timing of Volition and Actions Prospective memory connects two separate person-level activities. One occurs at time t1 and is called a “willed” decision to act in a certain way at a future time.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

91

The second is the planned action itself, performed at some later time t2 . This raises first of all the problem of (free?) will: How are plans generated and how are they chosen? But there is an interesting second problem. Can the voluntary action at t2 be of a more automatic kind or do we always need a new conscious act of will to start the planned action or even to start each of its separate motor components? The latter seems to be the default assumption, but I will argue later on, that it is unlikely. For example, if I decide tonight to walk to the bank tomorrow to cash a check, my walking tomorrow is voluntary. It is part of a plan I made prospectively and I do not need an act of will before each step or before entering the bank. I just do all those things unthinkingly, but if asked I will confirm that I do them voluntarily. I do not walk at gun point, but as a consequence of my own decision from yesterday. Acts of will are scarce and they are not directed at individual movements or their details. Most of the things we do voluntarily follow from older prospectively “willed” plans that are still valid as long-term projects. Many experiments on the timing of voluntary actions are in fact studies of ProM, although they mistakenly search for acts of will during the processes at time t2 . Acts of will are sometimes even expected to precede each separate detailed action, like raising a finger or twitching a voluntary muscle. For example, early work by Libet13–15 was interpreted to show that an unconscious decision to act is taken some 500 ms before the movement, because that is when a cortical wave, the “readiness potential” (RP) starts. The subject first reported awareness of the upcoming movement around 200 ms before the action, just in time to be able to consciously choose either to veto the action or not. To Libet, free will is just this option to control an upcoming (nonconsciously generated) action, so it is not involved in generating the action. Of course, the veto decision must also come from somewhere, so if it suddenly pops up in consciousness it should have a nonconscious decision as source. Libet does not accept this idea and says that conscious control occurring after the decision to act has become conscious “can appear without prior initiation by unconscious cerebral processes.” It appears then that he wants to postulate a cause-free kind of consciousness. The underlying implicit idea, that the concept of free will should force us to expect a separate act of will preceding each and every voluntary act, has hardly ever been challenged explicitly. Yet, it is precisely the falsification of this expectation that led to the demise of the concept of free will in neuroscience (often diagnosed to be an illusion16,17 ). From a ProM point of view, it seems likely that we can plan in advance to let our finger or hand move at random

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

92

W. A. van de Grind

during a later experiment, and leave detailed decisions on when to move to a prospective brain program that runs autonomously. In that case the RP does not precede a conscious decision or plan to act at all. It only betrays the preparation of a finger movement in an autonomous (but monitored) motor program, which causes randomly timed movements. Largely unsupervised random generators certainly exist in the nervous system and can be seen in action when we view ambiguous figures or during binocular rivalry. In these cases the percept flips spontaneously at random intervals. A ProM interpretation of Libet’s findings is quite a reasonable alternative (see the section on “Timing at the Output Side”).

Linking Hypotheses and Consciousness Hypotheses to conceptually link person-level to brain-level processes are necessary in an evolving brain science, and the assumption of a neuronal correlate of consciousness (NCC) is one such useful link. A long series of papers by Crick and Koch on this topic was recently summarized by Koch.18 He defines an NCC as the minimal set of neuronal mechanisms and events jointly sufficient for a specific conscious percept or experience. It is an empirical matter then to discover such neuronal correlates and/or to sensibly partition our percepts and experiences. Bennett and Hacker19 warn against the mereological fallacy, the careless habit of assigning properties or behavior of a whole system to some of its parts. In the case of an NCC we only think of a correlation between the neuronal and person levels of study, so this is not a mereological fallacy yet. However, as soon as we start to think in identities, for example, if we assume that cortical area V4 embodies some form of color consciousness, we have gone too far. Of course V4 cannot be conscious, not even a little bit, because it is no person. When I briefly glance out of the window, many colored patches no doubt excite my V4 cells without ever becoming reportable at the person level. Other brain areas need to cooperate to get a reportable (= conscious) experience. Some of the activity in V4 can be an NCC for color, but not all of it all of the time. During neurophysiological experiments on awake task-performing animals, or in clinical trials with humans, reportability is used to determine what the subjects are conscious of. Being conscious of something is another term for being able to honestly report on it, using a natural repertoire of communication, anything from language to music to body language to button presses in the laboratory. Baars and coworkers have extensively argued the case for consciousness as reportability.20– 23 Hacker24 nicely explains why we need not spend time on more esoteric matters, like what-it-is-like-to-be-an-x questions. It is a fact of

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

93

life, that we cannot report what is going on in our memory systems, in our motor areas of the brain, in our prefrontal cortex, in area V4 and the like. At the person-level we can report on the behavioral meaning of some of these brain activities. A person does not experience the neuronal activity called NCC, but what it stands for, what it represents. According to Bennett and Hacker, we need to make an effort to keep the person level and neuronal levels conceptually separated. This is sound advice that should be heeded. Yet, it is not always possible when speculating about the functional partitioning of labor in our brains.

Space, Time and Motion in Physics and Perception Comparison of Physical, Neural and Experiential Timing Einstein’s innovative relativity theory was based on the following intuitively simple and reasonable assumptions, which led to counter-intuitive and complex consequences: (i) Every action or interaction takes time; (ii) therefore, there must be a maximum speed for the propagation of physical influences; (iii) this must be c, the maximum speed of light (electromagnetic waves). The first two of these assumptions are directly and literally relevant to our present topic! The third assumption is of metaphorical value only, but its force will be felt later where I suggest that motion can be a reference for perceptual space-time measurements. One counter-intuitive consequence of these assumptions is that the time interval between a given pair of events need not be equal for every observer. Imagine Einstein flying face-forward with the speed of light, holding a mirror in front of his face. Even then reflected light would still travel with the speed of light from the mirror to his face, so he would not lose the sight of his face in the mirror, as classical physics would predict. Moreover, the same light would be measured to have the same speed by an earth-bound observer who saw Einstein zip past. The earth-bound observer would measure a different distance between Einstein’s face and the mirror and would measure a different travel time for the light from face to mirror and back, but the ratio of these two would still be c. It is nontrivial in this thought experiment to synchronize Einstein’s clock and that of an earth-bound observer.25 Einstein emphasized that all time measurements are based on simultaneity detection. We could in principle determine the duration of an event with subnanosecond precision if only we could ascertain the simultaneity of its start and stop with different individual cesium clock ticks. Timing at the Olympic Games has a temporal resolution that is more than a hundred million times lower. We

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

94

W. A. van de Grind

cannot ascertain the simultaneity of start and finish events with clock events any better. Two physical events that are not simultaneous can be ordered (e.g. B came after A) and one can count the number of clock ticks between A and B (interval duration). Here we immediately see a gulf between the physical domain and the domain of conscious experiences. Pöppel26 reports that the asynchronythreshold, that is, the time-shift between two stimuli that just allows us to detect that they are non-simultaneous, is about 5 ms for two auditory clicks, 10 ms for two touches, and 20–30 ms for two flashes. However, the threshold for a TOJ is about 30–40 ms in all three cases! This points to a significant and remarkable principled difference between physical timing and mental (experiential) timing! If you cannot ascertain the order of two mental events A and B, but you can perceive them as non-simultaneous, this means in algebraic terms that A  = B, but that neither A < B, nor B < A. Hence, perceptual timing does not have the same mathematical structure as physical timing, meaning that we should not naively apply one (physical timing) to the other. During single-neuron recordings we can directly refer stimulus timing and response timing to the same external clock. However, neurons are not in the business of mirroring stimulus properties, they work at an interpretation by performing transformations, they are at various intermediate conceptual levels between the physical world and our experiences. Some temporal properties of some neurons might influence some temporal properties of our experiences. Yet, many temporal properties of neuronal activity patterns are irrelevant to the timing of experiences, because they are translated into some other perceptual quality than time. There are many examples of this. A small delay between clicks stimulating the left and right ear is perceived as an egocentric direction of the single click that we hear: time difference is translated into azimuthal direction. If we give two light flashes close enough together in space and time, we see one motion step rather than two flashes. Here physical time and space differences are translated into perceived motion and the timing of the individual flashes is lost. If we change the distance or time between the flashes (within certain limits) we change perceived speed, not perceived timing. In the Pulfrich effect27 we watch a fronto-parallel swinging motion and see it move along an elliptical path in depth. A grey-filter in front of one eye delays its neural responses, and neural time differences translate into perceived depth.27,28 Brains of echolocating bats translate the cry-to-echo time into forward distance. In short, measured neural responses can always be related to timing in the external world, but their timing properties can be totally irrelevant to the experiences

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

95

they evoke, if any. Neural activity timing is no straightforward intermediate between physical and experiential timing. We cannot analyze experiential timing directly either, because we always have to make the experience measurable first, and that introduces an unknown delay. It is helpful to first analyze the relatively simple case of conscious perception of a visual flash, because it brings the pitfalls of timing conscious experiences into sharp focus.

How a Paradigm of Simplicity (Latency of Flash Perception) Ends in Illusion Can we measure the latency of conscious flash perception? Originally, reaction times were used to estimate the speed of perception29 or “Wahrnehmungszeit” (W-time) of flashes. Reaction times decrease with increasing flash luminance until they appear to reach some asymptote, the minimum W-time plus a motor reaction time. By guestimating and subtracting the latter we get the W-time as a function of flash luminance. Because this is a very imprecise method, Hazelhoff developed another approach in his thesis of 1923, reported by Hazelhoff and Wiersma.30 They moved a dark bar of 5 mm width, which the subject had to follow with her eyes, across a screen with visible markings. At position x in the middle of the track, a narrow light bar of 1 mm width was briefly flashed centered on the moving dark bar. The observer had to report where she saw the moving dark bar at the W-time of the light flash, which proved to be at a more advanced position along the track, say x +
x. If the dark bar moved at a speed V , the W-time was surmised to be
t =
x/ V . For a flash luminance of 200 times threshold-luminance Hazelhoff and Wiersma found an average W-time for their 40 subjects of 104 ms. The maximum W-time near threshold luminance was on the order of 250 ms. Unfortunately, there might be a problem with the logic of this approach. Motion is used to measure latency of the flash, but motion itself must also have a latency. If a moving slit of light hidden behind a screen-with-window suddenly enters the window, we see it pop up at some distance from the window’s edge. Fröhlich31 discovered this effect. It provides an opportunity to determine the W-time of motion, by measuring at which distance from the edge a moving slit is first seen. Depending on the slit’s luminance Fröhlich found values ranging from 40 (high luminance) to 150 ms. Metzger32 concluded that a relative mislocalization of a flash, as reported by Hazelhoff and Wiersma, should be explained from a difference in latency between the compared moving and

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

96

W. A. van de Grind

flashed stimuli. Nijhawan33 discovered a new version of the Hazelhoff–Wiersma effect and called it the flash-lag effect (FLE). Nijhawan33,34 and Khurana and Nijhawan35 also added an interesting twist to the theory. They hypothesized that we see the position of moving stimuli veridically. At the W-time for motion, when the physical stimulus has already advanced a little relative to the sampled position, this requires some form of extrapolation to the present actual physical position. A predictive visual motion system would facilitate motor interactions with moving objects, such as catching a ball.33 Critics of perceptual extrapolation pointed out, however, that it suffices if the motor system predicts the target’s future position. In response to this, Nijhawan and Kirschfeld36 had subjects move an unseen metal rod by hand along a trajectory and presented a flash in the middle of the track. The flash position lagged relative to the felt position of the unseen moving rod and hand. The motor system indeed appears to predict the position of moving objects (like the rod), thus causing an FLE. However, if that is accepted, why then reject the same explanation for the highly similar purely visual FLE? Should we assume that the motor system corrects for all the different latencies of the various sensory systems that can control it? This is certainly not more sensible a priori than to assume that each system corrects its own latencies by prediction. Nijhawan’s hypothesis gives visual motion a reference status analogous to the speed of light in Einstein’s theory. It makes motion the golden standard for measuring or calibrating experienced time and space! Since motion is detected by dedicated motion sensors, while we do not seem to have separate space or time sensors, the idea of motion primacy is attractive and deserves scrutiny.

Perceived Motion as a Possible Reference for Visual Space-Time Experiences Figure 1a schematizes a so-called bilocal motion sensor presumed to be active in the primary visual cortex V1. Input from the left retinal receptive field RF1 is delayed T units and then multiplied with output activity of the right-hand receptive field RF2 to give motion output M, which is sent to subsequent motion processing regions, such as V5. If an object travels in T s from the left- to the right-hand receptive field, a motion output is generated from this sensor, which is optimally tuned to speed V = span/delay. Motion sensors are sensitive to a small range of speeds and our brain is presumed to have millions of them for

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

97

Fig. 1. (a) Bilocal motion sensor in V1, with two retinal receptive fields RF1 and RF2, multiplicatively combined via a delay T . (b) If both receptive fields are stimulated in synchrony the upper multiplication detects this flicker or uniformity, sends it to output FU and inhibits the motion output. (c) A series of motion sensors with recruitment coupling, a shifted (extrapolated) position output P, and circuitry to block extrapolation where the motion ends.78 Explanation in the text.

various speeds and motion directions. If both receptive fields fire in synchrony the stimulus flickers or is uniform and static (detected by the FU-unit in Fig. 1b). Therefore, the FU-signal is used to inhibit the motion output to prevent confusion of flicker and motion. We have sensors for motion to the right (as sketched), but also for all hexagonal directions away from any given retinal receptive field. Motion sensors have been shown to recruit their neighbors in the motion direction37–39 which is modeled by chaining delay units, as sketched in Fig. 1c (second row from the top). A motion sensor, say at position 2 in Fig. 1c (bottom row), gets a signal from the left via a delay and after multiplication with its direct input it produces a motion output M. The signal from 1 also continues through the chained delay (second row) towards position 3. This spill-over can explain recruitment. At each motion output we also get a position signal output P, which is sent to an internal position map (mentioned in the introduction), but the mapping is shifted a distance z = V · t relative to the unit’s retinal input. Here t is the motion processing latency and V the tuning speed of the considered units. The shift-value z is presumed to be a result of tuning during

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

98

W. A. van de Grind

initial ontogenetic interactions with the world. No magic tricks are necessary, just a smart remapping. Eagleman and Sejnowski40 interpreted Nijhawan’s hypothesis to mean extrapolating into the future. But, no temporal operation is implied by the model of Fig. 1, just a position-shifted mapping. Eagleman and Sejnowski inferred that extrapolation means an unchanged FLE if the motion stops at the flash, tested this prediction and showed it to be wrong. This same apparent falsification had been presented before using a motion reversal at the flash,41,42 which does not result in an FLE either. However, overshoot is not a necessary consequence of prediction, it would simply signify bad design. The network of Fig. 1c uses FU-signals to block motion outputs and extrapolation as soon as motion stops. The FU-signal is supposed to spread a certain distance in both directions, such that it covers the extrapolation range. FU-spread and extrapolation distance are tuned together during ontogenesis. At motion onset the activation in a bilocal sensor has to travel the span before a motion signal can be generated. This can explain the Fröhlich effect, which should depend on the span and extrapolation shift, a prediction that has not yet been tested. We have confirmed the related prediction that a component of the reaction time to motion onset is proportional to the delay of activated motion sensors.43 Above 12–20 deg/s bilocal sensors have larger spans when tuned to higher speeds.44,45 From this we can expect an FLE-increase with speed, which was reported by Krekelberg and Lappe.39,46 The veridicality of motion-position must be limited to an average daytime luminance level, otherwise one would need a separate mapping for every possible luminance or adaptation level. The perceived position of the moving stimulus will therefore start to lag its physical position at decreasing luminance. In that case the perceived motion has not yet reached the flash position when the (undimmed) flash goes off. Motionlag due to dimming will decrease the FLE or even reverse its sign at very low luminances. This has indeed been reported.39,47 The Hazelhoff–Wiersma effect depends on a visible background.32 If we pursue an object against a homogeneous background or in the dark, no FLE is experienced.48 The foveal flash-image will simply follow the eye movement. With a continuously visible static reference (which moves across the retina of the pursuing eye) we get an FLE,48 similar to the FLE with fixating eye.33 After the fact we can therefore equate the FLE and the Hazelhoff–Wiersma effect. The FLE has become a typical widely ramifying theme for specialized vision research34,49 and studies of visually guided movements.48,50 Reviews are

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

99

available.51–53 What are the lessons of present concern? First, the problem of timing conscious percepts has not yet been solved definitively, not even for a single flash! Second, it is hard to synchronize the experiential clock and the physical clock. Third, to do so we need a common reference, as suggested for motion in the case of daylight vision. We will also see that the FLE plays a role in timing studies of voluntary actions (see “Timing at the Output Side”).

Timing Perceptual Experiences Libet’s Problem for Conscious Perception The classical problem of timing our experiences and voluntary actions was almost forgotten when Benjamin Libet revitalized it in the second half of the previous century. That is why I proposed54 to call it Libet’s problem. Here I discuss the perceptual side of Libet’s problem; in “Timing at the Output Side” we look at the timing of will and action. Libet13 concluded that our conscious perception lags about half a second behind its physical causes, but that the experiences are referred back to the physical time of their causes. Back-referral in time would be possible in principle if neural signals got a time tag when they arise in receptor sheets. These time tags could act like dates on a letter and be read by a central neural network to solve the problem of temporal dispersion. The problem is that no indications of such front-end time tags have ever been found, and it is hard to see why back-referral would have survival value. Because these matters have been spelled out before,13,54–60 I will just explain why Libet’s conclusions do not follow from the experimental results. Libet and colleagues applied electrical stimulation to the somatosensory cortex (C), peripheral skin (P) of the hand, and the medial lemniscus (LM). For C-stimulation a single strong pulse did not reliably give a conscious sensation, whereas a long train of weaker pulses did. At liminal intensity and pulse frequencies above 15 pulses per second perceptual latency was about 500 ms. This was accepted as a typical latency for conscious experiences at low intensities. Gomes56 argued that one can interpret most of the pulse train of 500 ms as a necessary buildup of an internal neural potential. If this potential passes a threshold value, an NCC is activated with a latency TC . In Fig. 2b I schematize the reasoning of Gomes, and in Fig. 2a of Libet. The depicted situation concerns C-stimulation felt in one hand and P-stimulation of the other hand, so that the observer only had to judge which came first, left or right.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

100

W. A. van de Grind

Fig. 2. (a) Libet’s explanation of his results of comparing the perceived timing of cortical (C) and Peripheral (P) stimulation. Indices: E = experienced, N = neuronal, S = stimulus. (b) An illustration of an interpretation by Gomes (1998)56 of the same results. Explanation in the text.

Libet et al. assumed that their minimum train duration to evoke a conscious percept with cortical stimulation, “mtd” in Fig. 2a, equalled the perceptual onset latency Tcon . They expected the peripheral skin stimulation P to have the same latency, so that it would be experienced simultaneously for an SOA (stimulus onset asynchrony) of zero. To their surprise they found that P had to be given much later (up to 450 ms) to get a simultaneous experience, as indicated in Fig. 2a by PS . However, on their equal-latency assumption this would predict a much later onset of a P-experience, as indicated by the dotted rectangle in Fig. 2a, trace PE . Because it was in fact felt to be simultaneous with the cortical stimulus, Libet and colleagues concluded that PE must be back-referred to the interrupted vertical line in Fig. 2a. They were certain that the simultaneity of CE and PE for this case could not be explained by a very short latency of PE , because another series of (masking) experiments had shown that the latency for peripheral stimulation was 125–300 ms. The P stimulus evokes a primary cortical potential in the EEG after some 15 ms (the small wave in trace PN in Fig. 2), so the idea arose that the back-referral process could use this wave as a time marker. Gomes56 postulates a buildup process, like a leaky integrator (see also van de Grind54 ), that takes about the same time as the minimum train duration, mtd. He also assumes that the buildup process sends a signal to another mechanism (an NCC), which reacts with a latency TC , as indicated in Fig. 2b. The peripheral stimulus evokes an experience after its normal latency of 125–300 ms (TP in Fig. 2b). Gomes has meticulously analyzed the other experiments as well (e.g. the LM and P comparison) and given similar explanations that fit the results

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

101

without back-referral. Moreover, the latencies for conscious experience (TP and TC ) are much shorter than in Libet’s explanation, and more in line with the W-time for flashes and motion (as discussed above). The experiments of Libet and coworkers are still highly interesting, but their explanations are untenable.

Asynchrony Detection as a Method to Time Experiences Dennett and Kinsbourne61 argue that our subjective sense of sequence and simultaneity must be determined by something other than a unique order of arrival at some internal finish line or “Cartesian theatre.” Multiple NCCs in combination with reportability through multiple channels is, of course, an option, but we need not a priori exclude that the number will be small. What about empirical data? Let us look at an ancient perceptual timing problem, segregation of the auditory rumble and visual flash of a thunderstorm. We now know of the different propagation speeds of sound and light, but different latencies in vision and hearing should also be taken into account. The perceptual latency for sound is about 40 ms shorter than for vision. In terms of a common neuronal finish-line this means that a simultaneous auditory and visual experience is expected if a sound-producing visible object is placed at 12 m distance.26 It was found by several groups that multisensory neurons in the superior colliculus of mammals produce a strongly enhanced response if their auditory and visual input signals arrive simultaneously. The decisive factor is that the auditory and visual input signals overlap in time.62 At this finish line for multisensory integration we have a complex situation, as schematized in Fig. 3. The time course of spike-arrival probability is sketched for visual (V) and auditory (A) activity. Rectangles symbolize the physical stimuli at the eyes and ears. Multisensory integration might just be possible if the auditory signal arrives at t4 (trace A1), which corresponds to stimulus time t1 . Both ontransients coincide if the auditory signal arrives at t5 (trace A2), there is still full overlap when the auditory signal arrives at t6 (trace A3), and maybe just enough if it arrives at t7 (trace A4). In a psychophysical experiment where we give the auditory stimulus at a variable time t after the visual stimulus, this state of affairs is mirrored in the probability of experiencing the two stimuli as simultaneous p(A ∼ V), where A ∼ V stands for “A perceived as simultaneous with V”: Fig. 3b. We think of a forced-choice experiment here, where you can choose between the responses “simultaneous” or “asynchronous” and the presentations are suitably randomized.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

102

W. A. van de Grind

Fig. 3. (a) The visual stimulus V (hatched rectangle) is followed after a latency TV by a response of duration dV . The auditory stimulus A is presented in the next four traces with increasing stimulus onset asynchrony (SOA), leading to auditory activity at an audio-visual neuron at the times t4 –t7 in the traces A1–A4, respectively. (b) Corresponding psychophysical results, representing the probability of experiencing the visual and auditory stimuli simultaneously as a function of SOA.

At t2 in Fig. 3a we expect to find that we virtually always see the stimuli as simultaneous, so it corresponds to the “point of subjective simultaneity (PSS)” , indicated in Fig. 3b. Points θ1 and θ2 are the asynchrony-thresholds, time shiftvalues for which the observer sees asynchrony in 75% of the presentations. The PSS is expected at t2 ≈ (TV − TA ) + 0.5(dV − dA ), so that it is not possible to deduce a differential latency from the PSS if you do not know the response durations. The window of simultaneity is on the order of the sum of the longer and twice the shorter duration. It is not a good idea to use TOJs instead of p(A ∼ V), because we do not understand yet why TOJs differ from the more basic asynchrony thresholds. One possibility is that TOJs are mediated by another neural network after the signals have been dispersed more in time than at low-level multisensory neurons. Asynchrony thresholds and p(A ∼ V) as a function of SOA are easier to interpret. For audiovisual integration the window of simultaneity (Fig. 3b) is rather broad (80–100 ms), so we can enjoy apparent simultaneity for a reasonable distance range, say from 0.1 to 50 m.63 Despite reports to that effect64 no compensation for the sound travel-time is necessary to explain this range.65 The idea of a common audio-visual finish-line in the brain makes sense, but even in this case there must be several others. For example, looking at a talking head we

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

103

do not see asynchrony until sound is delayed by some 250 ms. Speech movements of the face and sound of spoken language are no doubt first analyzed in sufficient detail at later processing levels than thunder. Their finish-line must be more central. This might also hold in many instances of the ventriloquist effect, where we see a sound source at its visual position, despite large spatial shifts. Vision might be the space-specialist and sound the time-specialist.66 In accordance with this idea it is possible to shift W-times of visual stimuli, such as flashes, with near synchronous auditory stimuli, a phenomenon called temporal ventriloquism.67 Let us consider asynchronies within a single perceptual system, vision. Zeki and coworkers68−73 proposed that a processing site where certain features are made explicit, such as V4 for color and V5 for motion, is itself a conscious perception site. Moutoussis and Zeki74,75 and Zeki and Moutoussis76 did a series of experiments to quantify asynchronies of (among other things) color and motion. In one experiment,74 a random pattern of 30 small squares moved up and down with a period duration of 716 ms or 537 ms. Their color was changed in the same rhythm, but with a variable phase. Observers saw the color and motion in phase when motion changed direction about 70–80 ms before the color change. This suggests a differential latency of 70–80 ms, with color leading. The results were interpreted to show that V4 (color) and V5 (motion) have independent and asynchronous access to consciousness. The type of experiment is no different from the classical experiments on multisensory integration, including the finding of a PSS-value that is unequal to zero, possibly signifying a differential latency. In these periodic presentations without pauses, the duration of a given motion direction signal (say up) and a given color signal (say red) is forced to be equal, namely half a cycle (268 or 358 ms in this case). According to Fig. 3, this suggests that the PSS-value reported by Moutoussis and Zeki74 is indeed a differential latency. However, unlike the situation of Fig. 3, we cannot now expect that lowlevel combinatory neurons fuse the color and motion signals. On the contrary, color is probably bound to form or motion by a common position label at a rather high processing level, as in a coloring book for children. Color is a “don’t-care” condition for objects and motion, in the sense that almost any color can go with any given object, like (parts of) a shirt or house. In the Moutoussis–Zeki experiment, the observer has to report on a maximum temporal overlap of the population response of (say) upward-tuned neurons in V5 and (say) red-tuned color neurons. This comparison takes time and the responsible brain program might not be able to follow the rate of about four population changes per second.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

104

W. A. van de Grind

Results by Nishida and Johnston77 show that the PSS shifts towards zero for longer presentation periods, reaching zero when the inter-change period (halfcycle duration) is about 2 s. The activity-patterns in V4 and V5 are obviously compared somewhere in the brain, otherwise the experiment would not have worked. Yet, it seems doubtful that the PSS reported by Moutoussis and Zeki can be interpreted as a generic latency difference between single unit activities in V4 and V5. In fact recordings in these areas suggest the reverse, that V5 is faster than V4. There is an additional problem with the experiments by Moutoussis and Zeki. If we compare the timing of two light flashes of unequal luminance, we also get a PSS unequal to zero (differential latency), but the PSS moves towards zero with decreasing difference in “strength.” How could we equate the “strength” of a color change and a motion change? To circumvent such problems, we78 compared differently-directed rocking (to-and-fro moving) random dot patterns presented in static windows or flashed (single frame) versions of such patterns. In that case we look at population responses in the same cortical area, V5. Moreover, we studied the onset-offset synchrony rather than overlap synchrony. The window of simultaneity proves to be narrow for motions in the same direction (about 43 ms), but very broad for a mutually orthogonal motion pair (about 84 ms). The PSS is about zero in these cases. Observers interpreted stimuli of the same motion direction as one large sheet of random dots moving behind an occluder with two windows cut out. This causal coupling was lost for relatively small asynchronies, by far the smallest of all direction combinations. It is plausible that the coupling is done by neurons in MST, a region receiving direct input from V5, where the receptive fields are very large. Cells in MST covering both windows are strongly activated for synchronous motion, but less so for increasing asynchronies. In this way we have a sharp finish-line in single higher-order cells, which does not occur for arbitrary direction combinations. We found that it is also possible to time flashes and motions. Flashes and equal luminance motion were found to have a PSS of about zero, supporting the idea of an earlier section, that the FLE is a consequence of position (not time) extrapolation. A change of luminance of only the motion stimulus increased its latency by about double the amount found if only the luminance of the flash was changed. This suggests a temporal comparison between different way stations, V1 for the flash and V5 for motion. At least in this case there appears to be one brain program that compares signals from different processing levels at a common finish line.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

105

The conclusions for our present theme are as follows. First, asynchrony studies comparing different perceptual features are easy to do, but can be difficult to interpret in the absence of physiological data suggesting the whereabouts of possible common finish-lines. In the second place, the default assumption must be that there are many such finish-lines for different feature-pairs, but it might be possible to find out experimentally whether or not groups of these finish-lines are served by common analysis programs. Thirdly, not all temporal comparisons need to have a bearing on “binding,” because binding of features through a common set of local signs need not be very time critical if it occurs in the form of working memory or global workspace. Fourth, narrow windows of simultaneity probably signify an underlying causal coupling or Gestalt formation operation.

Timing at the Output Side Timing Voluntary Decisions and the Partial Liberation of Our Will Libet and coworkers14,15 reported that a cortical “Readiness Potential” (RP) precedes and predicts voluntary hand movements. The RP appeared to start some 500 ms before the movement, but it was not until the last 200 ms that subjects reported becoming aware of it, just in time to veto the upcoming movement if they so wished. Keller and Heckhausen79 later showed that also nonconscious movements are preceded by RPs, so it seems more likely that the RP is a relatively neutral sign of a movement preparation. In any case RPs do not seem to precede the veto-moment by the full 300 ms reported by Libet. The subjects had to notice the position of a rotating dial on a scale at the veto-moment, and are thus presumed to be able to use the veto-moment to trigger a visual percept. But then we would expect a FLE, meaning that we see the moving pointer shifted 80–100 ms forward in time, relative to the number read at the veto-moment. The veto-moment might therefore already occur 200 rather than 300 ms after the start of the RP. Even these 200 ms are probably an over-estimate.54–59 Taking everything into account it becomes questionable whether there really is a significant asynchrony between the start of an RP and the veto-moment,54 but let us assume there is. This would not be a problem for those who think that free will or willed actions are illusions, tricks used by our brain to give us the comfortable feeling of being in charge.16,17 However, such a position makes it hard to see what survival value our conscious experiences might have. It seems more reasonable to assume that both involuntary and voluntary actions

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

106

W. A. van de Grind

are caused by preceding brain activity, but that some of the antecedents of voluntary actions are conscious and the basis for our feeling that we have a free will. As argued in the introduction, there are more voluntary actions than acts of will, because most of our voluntary actions are consequences of “prospective” plans (P-plans), “willed” a significant while back. The conscious processes leading to a P-plan are experienced as a deliberation process, an interaction between several players (brain programs), each of them representing one of our sub-functions, for example a creative plan-spinner, a conscientious censor and a wise evaluator. The feeling of deliberation on which we can report is a major part of our free will, because we are (among other things) the plan-spinner, censor and evaluator, it is our deliberation, and we can report on the things that are exchanged between the participating brain programs. This is so because “we” stands for the reportable (experiential) part of these brain processes. The deliberation process is plastic, “learns” from failures and successes and is constantly restructured, which adds to the feeling of freedom, of autonomy of evolvability. We are not a playball of fixed brain processes reacting in fixed ways to a given situation. We use our way of choosing from our type of plans, which fit into our way of life. That is why we feel responsible for our choices, and we are. We have several P-plans controlling our voluntary activities, but these voluntary activities can consist of largely automatic processes, like my walking to the bank mentioned in the introduction. Binocular rivalry80 shows an automatic random process at work, that we can start or stop at will. If you look at a horizontally striped pattern with one eye, and a vertically striped pattern with the other (e.g. with the help of a stereoscope) the two incompatible patterns “fight” for conscious access. Now you see one, then the other pattern, and this alternation repeats itself as long as you decide to watch, with intervals between perceptual flips drawn from a gamma distribution. Observers have a modest conscious influence on the flip intervals.81 The process is voluntary in the sense that you decided consciously to sit down and watch, the rest is automatic. The flip moments can be predicted by an external observer monitoring neural activity in the temporal cortex.82 Of course the parallel with Libet’s experiment is not perfect, but good enough to appreciate that we have the option to let random generators do their job. They act in the service of a process we can start and stop at will, but do not normally control decision-by-decision.

Duration Coding To measure the duration of intervals between pairs of events, we need a “timing mechanism” (TM), to (i) detect the interval’s start, (ii) read a clock-parameter

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

107

at the interval’s start, (iii) detect the interval’s end, (iv) read the clock-parameter at the interval’s end, (v) calculate the duration from these two readings, and (vi) translate the duration value into a useable code. The clock-parameter is an internal variable of a clock. It can be any discrete or continuous quantity that changes with time. The TM and its clock will together be called a “timer,” a device that is analogous to a stopwatch plus its user. While using a stopwatch we are part of the TM, starting, stopping and translating clock ticks in a time code that can be communicated. Our brain does not have homunculi to read its stopwatches, so we need to spell out how its timers are controlled, used and read out. The clock-parameter has drawn a lot of attention, the code in which time is stored or used for transactions with other brain modules has had virtually no attention in the literature. The theory of simultaneity detection, as discussed above, is relevant for tasks (i)–(iv) of the TM, regardless of the durations to be quantified. For example, for an oscillating clock signal, the TM needs to determine the temporal coincidence of an external start or stop event and the closest clock tick. The width of the window of simultaneity therefore co-determines the resolution of duration measurements. Neural events used to control a timer can be either intrinsic, e.g. stored in memory, or perceptual events. But what are the codes? For labelled line codes or rate-codes, it is relatively easy to design a TM, but it is hard to see how population codes could be handled. In any case time codes in memory must take a structural form. Neural activity is too costly to use as carrier in a long-term storage medium.83 In addition, the structural changes have to be minor in order not to limit storage capacity too much or cause interference. There is abundant information to show that most motor programs in all animals have inherent rhythm generators, delays and timers, ensuring automatic execution, even of such complex patterns as walking, swimming or flying.84 Coordination or synchronization to external rhythms is provided by higher centers such as the cerebellum, supplementary motor areas (SMA), the basal ganglia and many more.85,86 Sophisticated models have been developed to explain skillful actions like drumming, based either on a nonlinear dynamic systems approach or competing information processing approaches.87 Most of the information processing models use combinations of pacemakers, counters, and comparators,85,88,89 which implies that event counts represent durations. Staddon and Higa90 and Staddon,91 discuss shortcomings of some of the counter-based models and propose the strength of a memory trace as an alternative clock-parameter. Their model needs two stages of leaky integrators, one with a short and the next with a longer time constant. According to Staddon91

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

108

W. A. van de Grind

it would be possible with a series of 10–12 slower and slower serial stages, to model long-term forgetting, which can be described with a sum of exponentials. One problem with this proposal is that leaky integrators with time constants of days to years have never been found in the nervous system. An activity-based, long-term storage in leaky integrators would also be energetically too costly. Lewis and Walsh92 and Lewis and Miall93 have argued for a multiple timer hypothesis. They propose that timers in the prefrontal and parietal cortex cover the longer durations that are important in cognition and voluntary actions. The question then arises whether all these timers use the same metric and the same code for storage and retrieval. Can interval durations be added and subtracted mentally? Walsh94 has recently proposed that the brain uses a common magnitude measure for numbers, space and time. This revives an interesting classic idea of the Gestalt school, most notably worked out by Erich M. von Hornbostel.95 Our language uses spatial phrases to express time (a long wait, a short interval, time flies), but that might be a fluke of history. Moreover, space is three-dimensional, whereas time is a one-dimensional scalar quantity. Rosenbaum96 proposed to represent time as path-length, thus equating the number of dimensions. Pathlength is represented by delay-lines in his proposal, so, as in our FLE-model,78 speed along certain paths is the common reference again. Figure 4 presents a timer inspired by these ideas. It can measure a duration, store and reproduce it. In principle one could have timers like the one in Fig. 4 for various duration ranges, but there is a practical limit to delay values. Using neurons it seems hardly possible to make delays longer than a few dozen seconds, but if chemical or genetic cascades could replace the delay lines, anything up to a lifetime’s duration is possible. The circuit can be so simple, because I used a labelled line code to store durations. This simple coding principle is universally used in the nervous system. If you stimulate auditory axons you hear something, for visual axons you see something, etc. This may be a generic trick of the nervous system, and it can be used for time information as well. The bottom row of modules in Fig. 4 are delays (rectangles D) traversed by a start signal (input a) from left to right. When the interval ends, the undelayed stop signal (input b) will coincide at a coincidence gate (C-gate) that is just then reached by the delayed start signal, say C-gate 5. This gate is opened and changes the state of memory element 5 in the FF-row from state 0 to 1, whereas all other memory elements remain in state 0. Somewhat later we want to reproduce the interval. A readout signal (input c) is given to all memory elements, but only the

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

109

Fig. 4. Model of storing and reproducing an interval of a given duration. Inputs at the left bottom are from neurons detecting the start and stop event of the presented interval. D are delay stages, C are coincidence gates, which only respond if the two inputs per gate are stimulated simultaneously or within a short interval. FF are the storage elements, flip-flops set from state 0 to state 1 by the black arrow input, reset by the open arrow input. During reset they produce an output signal to an output delay line. The remembered duration depends on the position of the FF that is set and later reset. See text for further explanation.

element in state 1 will now change state. It is assumed that a state-change from 1 to 0 causes output activity arriving, in our example, at point 5 of the output delays (top row in Fig. 4). This signal travels to the left in Fig. 4, and when it reaches output d of the upper delay line it has reproduced the stored duration. Speed is the common (but hidden) reference, and path-length (along delay lines) represents time. The memory elements can be flip-flop-like pairs of neurons, or even single neurons acting as flip-flops, which is known to be possible.97 In long-term memory one would need molecular flip-flops, and in that case readout should not be accompanied by a reset, or the read-out duration should be re-stored during readout. A combined read-out and re-storage process has been proposed for human memory systems.98 It would be easy to implement in the circuit of Fig. 4. We only need to feed the readout signal both into inputs c and a, and send the output d also to b in Fig. 4. An obvious strength of this model is the simplicity of its code, a kind of look-up table for intervals. Delay-line storage also has the characteristics necessary for time-based ProM. Take the output delay-line (upper row) in Fig. 4, detached from the rest, as an example. If the brain gated a start signal to point 5, a do-it-now signal would arise at

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

110

W. A. van de Grind

the output d of the delay line after 5 D-units. Thus, if a D-unit could delay a propagating signal long enough, any warning signal timing would be possible. For longer durations this might require molecular processes as delays. Because the circadian clock is ultimately based on the interaction of a small number of molecules, it does not seem too far-fetched to envision long-delay molecular processes.

References 1. Zacks J M, Tversky B. Event structure in perception and conception. Psychological Bulletin 2001; 127(1): 3–21. 2. Lotze H. Mikrokosmos. Leipzig: Hirzel Verlag, 1884. 3. Koenderink J J. The concept of local sign. In van Doorn A J, van de Grind W A, Koenderink J J, eds. Limits in Perception, Utrecht: VNU Science Press, 1984: 495–547. 4. Koenderink J J. Simultaneous order in nervous nets from a functional standpoint. Biological Cybernetics 1984; 50: 35–41. 5. Koenderink J J. Geometrical structures determined by the functional order in nervous nets. Biological Cybernetics 1984; 50: 43–50. 6. Whitney D, Goltz H C, Thomas C G, Gati J S, Menon R S, Goodale M A. Flexible retinotopy: Motion-dependent position coding in the visual cortex. Science 2003; 302: 878–881. 7. Constantinidis C, Franowicz M N, Goldman–Rakic P S. The sensory nature of mnemonic representation in the primate prefrontal cortex. Nature Neuroscience 2001; 4: 311–316. 8. Goldman–Rakic P S. Cellular basis of working memory. Neuron 1995; 14: 477–485. 9. Goldman–Rakic P S. The prefrontal landscape: Implications of functional architecture for understanding human mentation and the central executive. Philos T Roy Soc Lond B 1996; 351: 1445–1453. 10. Lee D N, Reddish P E. Plummeting gannets: A paradigm of ecological optics. Nature 1981; 293: 293–294. 11. van de Grind W A. Smart mechanisms for the visual evaluation and control of self-motion. In Warren R, Wertheim A H, eds. Perception and Control of Self Motion. Hillsdale N.J.: Lawrence Erlbaum Associates, 1990; 357–398. 12. Young J Z. Programs of the Brain. Oxford UK: Oxford University Press, 1978. 13. Libet B. Unconscious cerebral initiative and the role of conscious will in voluntary action. The Behavioral and Brain Sciences 1985; 8: 529–566. 14. Libet B, Wright E W, Gleason C A. Readiness-potentials preceding unrestricted ‘spontaneous’ vs pre-planned voluntary acts. Electroencephalography and Clinical Neurophysiology 1982; 54: 322–335. 15. Libet B, Gleason C A, Wright E W, Pearl D K. Time of conscious intention to act in relation to onset of cerebral activity (Readiness-potential): The unconscious initiation of a freely voluntary act. Brain 1983; 106: 623–642. 16. Wegner D M. The mind’s best trick: How we experience conscious will. Trends in Cognitive Sciences 2003; 7(2): 65–69.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

111

17. Wegner D M. Précis of the illusion of conscious will. Behavioral and Brain Sciences 2004; 27: 649–692. 18. Koch C. The quest for consciousness. A neurobiological approach. Englewood: Roberts and Company, 2004. 19. Bennett M R, Hacker P M S. Philosophical Foundations of Neuroscience. Oxford UK: Blackwell Publishing Ltd, 2003. 20. Baars B J. Understanding subjectivity: Global workspace theory and the resurrection of the observing self. Journal of Consciousness Studies 1996; 3(3): 211–216. 21. Baars B J, Ramsøy T Z, Laureys S. Brain, conscious experience and the observing self. Trends in Neuroscience 2003; 26(12): 671–675. 22. Seth A K, Baars B J, Edelman D B. Criteria for consciousness in humans and other mammals. Consciousness and Cognition 2005; 14(1): 146–167. 23. Edelman D B, Baars B J, Seth A K. Identifying hallmarks of consciousness in nonmammalian species. Consciousness and Cognition 2005; 14(1): 168–186. 24. Hacker P M S. Is there anything it is like to be a bat? Philosophy 2002; 77: 157–174. 25. Reichenbach H. The Philosophy of Space and Time. New York: Dover Publications Inc, 1958. (Original German version appeared in 1928.) 26. Pöppel E. Grenzen des Bewusstseins (Limits of Consciousness). Frankfurt a.M. & Leipzig: Insel Verlag, 2000. 27. Pulfrich C. Die Stereoskopie im Dienste der isochromen und heterochromen Photometrie, Die Naturwissenschaften 1922; 25: 553–564; 26: 569–574; 27: 596–601; 33: 714–722; 34: 735–743; 35: 751–761. 28. Howard I P, Rogers B J. Binocular Vision and Stereopsis. New York, Oxford: Oxford University Press–Clarendon Press, 1995. 29. Donders F C. On the speed of mental processes. 1868. Reproduced in Acta Psychologica 1969; 30: 412–431. 30. Hazelhoff F F, Wiersma H. Die Wahrnehmungszeit. Zeitschrift für Psychologie 1925; 96: 171–188; 97: 174–190. 31. Fröhlich F W. Über die Messung der Empfindungszeit. Zeitschrift für Sinnesphysiologie 1923; 54: 58–78. 32. Metzger W. Versuch einer gemeinsamen Theorie der Phänomene Fröhlichs und Hazelhoffs und Kritik ihrer Verfahren zur Messung der Empfindungszeit. Psychologische Forschung 1932; 16: 176–200. 33. Nijhawan R. Motion extrapolation in catching. Nature 1994; 370: 256–257. 34. Nijhawan R. Visual decomposition of colour through motion extrapolation. Nature 1997; 386: 66–69. 35. Khurana B, Nijhawan R. Extrapolation or attention shift? Reply. Nature 1995; 378: 566. 36. Nijhawan R, Kirschfeld K. Analogous mechanisms compensate for neural delays in the sensory and the motor pathways: Evidence from motor flash-lag. Current Biology 2003; 13: 749–753. 37. van de Grind W A, van Doorn A J, Koenderink J J. Detection of coherent movement in peripherally viewed random-dot patterns. Journal of the Optical Society of America 1983; 73: 1674–1683.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

112

W. A. van de Grind

38. Watamaniuk S N, McKee S P, Grzywacz N M. Detecting a trajectory embedded in randomdirection motion noise. Vision Research 1995; 35: 65–77. 39. Krekelberg B, Lappe M. Temporal recruitment along the trajectory of moving objects and the perception of position. Vision Research 1999; 39: 2669–2679. 40. Eagleman D M, Sejnowski T J. Motion integration and prediction in visual awareness. Science 2000; 287: 2036–2038. 41. Whitney D, Murakami I. Latency difference, not spatial extrapolation. Nature Neuroscience 1998; 1: 656–657. 42. Whitney D, Murakami I, Cavanagh P. Illusory spatial offset of a flash relative to a moving stimulus is caused by differential latencies for moving and flashed stimuli. Vision Research 2000; 40: 137–149. 43. van den Berg A V, van de Grind W A. Reaction times to motion onset and motion detection thresholds reflect the properties of bilocal motion detectors. Vision Research 1989; 29(9): 1261–1266. 44. van Doorn A J, Koenderink J J, van de Grind W A. Limits in spatio-temporal correlation and the perception of visual movement. In van Doorn A J, van de Grind W A, Koenderink J J, eds. Limits in Perception. Utrecht: VNU Science Press, 1984: Ch. 8, 203–234. 45. van de Grind W A, Koenderink J J, van Doorn A J. The distribution of human motion detector properties in the monocular visual field. Vision Research 1986; 26(5): 797–810. 46. Krekelberg B, Lappe M. A model of the perceived relative positions of moving objects based upon a slow averaging process. Vision Research 2000; 40: 201–215. 47. Purushothaman G, Patel S S, Bedell E, Ogmen H. Moving ahead through differential visual latency. Nature 1998; 396: 424. 48. Nijhawan R. The flash-lag phenomenon: Object and eye movements. Perception 2001; 30: 263–282. 49. Sheth B R, Nijhawan R, Shimojo S. Changing objects lead briefly flashed ones. Nature Neuroscience 2000; 3: 489–495. 50. Schlag J, Schlag-Rey M. Through the eye, slowly: Delays and localization errors in the visual system. Nature Reviews Neuroscience 2002; 3: 191–200. 51. Nijhawan R. Neural delays, visual motion and the flash-lag effect. Trends in Cognitive Sciences 2002; 6(9): 387–393. 52. Ögmen H, Patel S S, Bedell H E, Camuz K. Differential latencies and the dynamics of the position computation process for moving targets, assessed with the flash-lag effect. Vision Research 2004; 44: 2109–2128. 53. Krekelberg B, Lappe M. Neuronal latencies and the position of moving objects. Trends in Neuroscience 2001; 24: 335–339. 54. van de Grind W A. Physical, neural and mental timing. Consciousness and Cognition 2002; 11: 241–264. 55. Libet B. Time factors in conscious processes: Reply to Gilberto Gomes. Consciousness and Cognition 2000; 9: 1–12. 56. Gomes G. The timing of conscious experience: A critical review and reinterpretation of Libet’s research. Consciousness and Cognition 1998; 7: 559–595.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

113

57. Gomes G. Problems in the timing of conscious experience. Consciousness and Cognition 2002; 11: 191–197. 58. Pocket S. On subjective back-referral and how long it takes to become conscious of a stimulus. A reinterpretation of Libet’s data. Consciousness and Cognition 2002; 11: 144–161. 59. Trevena J A, Miller J. Cortical movement preparation before and after a conscious decision to move. Consciousness and Cognition 2002; 11: 162–190. 60. Klein S A. Libet’s temporal anomalies: A reassessment of the data. Consciousness and Cognition 2002; 11: 198–214. 61. Dennett D C, Kinsbourne M. Time and the observer. The where and when of consciousness in the brain. Behavioral and Brain Sciences 1992; 15: 183–247. 62. Stein B E, Meredith M A. The Merging of the Senses. Cambridge Mass.: Bradford Book, MIT Press, 1993. 63. Lewald J, Guski R. Cross-modal perception integration of spatially and temporally disparate auditory and visual stimuli. Cognitive Brain Research 2003; 16: 468–478. 64. Sugita Y, Suzuki Y. Implicit estimation of sound-arrival time. Nature 2003; 421: 911. 65. Lewald J, Guski R. Auditory-visual temporal integration as a function of distance: No compensation for sound-transmission time in human perception. Neuroscience Letters 2004; 357: 119–122. 66. Welch R B, Warren D H. Immediate perceptual response to intersensory discrepancy. Psychological Bulletin 1980; 88: 638–667. 67. Vroomen J, de Gelder B. Temporal ventriloquism: Sound modulates the flash-lag effect. J Exp Psychol Human Perc & Perf 2004; 30(3): 513–518. 68. Zeki S. Localization and globalization in conscious vision. Annu Rev Neurosci 2001; 24: 57–86. 69. Zeki S. The disunity of consciousness. Trends in Cognitive Sciences 2003; 7(5): 214–218. 70. Zeki S, Bartels A. The asynchrony of consciousness. Proc R Soc Lond B 1998; 265: 1583–1585. 71. Zeki S, Bartels A. The autonomy of the visual systems and the modularity of conscious vision. Phil Trans R Soc Lond B 1998; 353: 1911–1914. 72. Zeki S, Bartels A. Toward a theory of visual consciousness. Consciousness and Cognition 1999; 8: 225–259. 73. Bartels A, Zeki S. The theory of multistage integration in the visual brain. Proc R Soc Lond B 1998; 265: 2327–2332. 74. Moutoussis K, Zeki S. A direct demonstration of perceptual asynchrony in vision. Proc R Soc Lond B 1997; 264(1380): 393–399. 75. Moutoussis K, Zeki S. Functional segregation and temporal hierarchy of the visual perceptive systems. Proc R Soc Lond B 1997; 264: 1407–1414. 76. Zeki S, Moutoussis K. Temporal hierarchy of the visual perceptive systems in the Mondrian world. Proc R Soc Lond B 1997; 264: 1415–1419.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

114

W. A. van de Grind

77. Nishida S, Johnston A. Marker correspondence, not processing latency, determines temporal binding of visual attributes. Current Biology 2002; 12: 359–368. 78. Lankheet M J M, van de Grind W A. Simultaneity versu asynchrony of visual motion and luminance changes. In Nijhawan R, ed. Problems of Space and Time in Perception and Action. Cambridge UK: Cambridge University Press, in press. 79. Keller I, Heckhausen H. Readiness potentials preceding spontaneous motor acts: Voluntary vs involuntary control. Electroencephalography and Clinical Neurophysiology 1990; 76: 351–361. 80. Blake R, Logothetis N K. Visual competition. Nature Reviews Neuroscience 2001; 3: 1–14. 81. van Ee R, van Dam L C J, Brouwer G J. Voluntary control and the dynamics of perceptual bi-stability. Vision Research 2005; 45: 41–55. 82. Logothetis N K. Single units and conscious vision. Phil Trans R Soc Lond 1998; 353: 1801–1818. 83. Lennie P. The cost of cortical computation. Current Biology 2003; 13: 493–497. 84. Orlovsky G N, Deliagina T G, Grillner S. Neuronal Control of Locomotion. From Mollusc to Man. Oxford UK: Oxford University Press, 1999. 85. Ivry R B, Richardson T C. Temporal control and coordination: The multiple timer model. Brain and Cognition 2002; 48: 117–132. 86. Macar F, Lejeune H, Bonnet M, Ferrera A, Pouthas V, Vidal F, Maquet P. Activation of the supplementary motor area and of attentional networks during temporal processing. Exp Brain Res 2002; 142: 475–485. 87. Krampe R T, Engbert R, Kliegl R. Representational models and nonlinear dynamics: Irreconcilable approaches to human movement timing and coordination or two sides of the same coin? Introduction to the special issue on movement timing and coordination. Brain and Cognition 2002; 48: 1–6. 88. Gibbon J, Malapani C, Dale C L, Gallistel C R. Toward a neurobiology of temporal cognition: Advances and challenges. Current Opinion in Neurobiology 1997; 7: 170–184. 89. Wing A M. Voluntary timing and brain function: An information processing approach. Brain and Cognition 2002; 48: 7–30. 90. Staddon J E R, Higa J J. Time and memory: Towards a pacemaker-free theory of interval timing. J Exp Anal Behav 1999; 71: 215–251. 91. Staddon J E R. Interval timing: Memory, not a clock. Trends in Cognitive Sciences 2005; 9(7): 312–314. 92. Lewis P A, Walsh V. Neuropsychology: Time out of mind. Current Biology 2002; 12: R9–R11. 93. Lewis P A, Miall R C. Distinct systems for automatic and cognitively controlled time measurement: Evidence from neuroimaging. Current Opinion in Neurobiology 2003; 13: 250–255. 94. Walsh V. A theory of magnitude: Common cortical metrics of time, space and quantity. Trends in Cognitive Sciences 2003; 7(11): 483–488. 95. von Hornbostel E M. Die Einheit der Sinne. Melos Zeitschr f Musik 1927; 4: 290–297. (Translation: The unity of the senses. Psyche 1927; 7(28): 290–297.)

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

Representing Times of the Past, Present and Future in the Brain

115

96. Rosenbaum D A. Time, space, and short-term memory. Brain and Cognition 2002; 48: 52–65. 97. McCormick D A. Neural networks: Flip-flops in the brain. Current Biology 2005; 15(8): R294–R296. 98. Nader K. Memory traces unbound. Trends in Neurosciences 2003; 26(3): 65–72.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch04

This page intentionally left blank

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

5 At the Crossroads of Time and Action: A Temporal Discounting Primer for Prospective Memory Researchers Thomas S. Critchfield∗ and Gregory J. Madden†

Prospective memory (ProM) researchers are concerned with abilities that lie at the crossroads of time and action, but ProM is not the only phenomenon in which time and action interact. This essay describes one such phenomenon known as temporal discounting (TD). While ProM is defined by a delay between the decision to act and the execution of the required action, TD is defined by a delay between action and the consequences that make the action worth deciding upon. Any complete account of the psychological importance of time must consider both phenomena and, as we will argue in our concluding section, there may exist opportunities for cross pollination of these two research areas. The purpose of the present essay, then, is to promote interest in the relationship between TD and ProM. Toward this end, we will introduce some TD concepts, describe how TD is studied, and summarize key TD principles.

∗ Illinois

State University, Department of Psychology, Normal, IL 61790, USA; e-mail: [email protected] of Kansas, Department of Applied Behavioral Science, Lawrence, KS 66045, USA; e-mail: [email protected] † University

117

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

118

T. S. Critchfield & G. J. Madden

Because memory researchers are concerned with the ecological validity of laboratory research,1,2 throughout the essay we illustrate how TD principles are useful in understanding one type of socially-important behavior, drug abuse. TD research is noteworthy in the extent to which a unified conceptual framework both accounts for laboratory performance and makes important predictions about performance outside of the laboratory.3 We conclude the essay by speculating about how TD may intersect with ProM tasks, placing special emphasis on competition (interference) between ProM and other tasks, on encoding as a function of task importance, and on the effects of delay on task importance. Before proceeding with the essay, two brief notes are in order. First, we acknowledge the perils of attempting to draw connections between disparate research traditions. We trust the reader to accommodate for our limited knowledge of the ProM literature, and we hope that idiosyncracies of language do not distract too much from a discussion of general concepts. In the latter case, for example, note that when we speak of “actions” we mean anything an individual does, not only the physical movements that are so often the focus of study in behavioral psychology but also cognitive events such as encoding and attending. When we speak of consequences, we refer not simply to reinforcement or punishment mediated by other people but also to the benefits and costs that derive naturally from actions. Second, TD is a component of behavioral choice theory, which focuses on the consequence-based competition between concurrent courses of action.3–6 Because it is popular currently to dismiss behavioral psychology as irrelevant to cognitive science,7 we expect that our attempt to link ProM to TD will be met with skepticism. The reader may note, however, that TD is one research area in which behavioral and cognitive scientists have interacted somewhat comfortably.8,9 The acid test for any speculation, of course, is empirical, and only future studies can tell whether TD principles predict ProM effects of practical or theoretical importance.

Brief Review of Temporal Discounting Impulsiveness and “Temporal Myopia” Name a problem involving lack of “willpower” and you are likely to invoke common behaviors such as cigarette smoking, overeating, substance abuse, unprotected sex, procrastination, compulsive buying, and gambling. In all of these cases, individuals appear to act for immediate gratification at the expense

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

119

of their long-term best interests. Research on temporal discounting (TD) asks why they do this. Delay-specific erosion in the psychological impact of consequences is the defining feature of TD and appears to be central to many socially important impulsive behaviors.3 For example, a shopper may know that today’s extravagant purchase will mean financial hardship at the end of the month, and a substance abuser may realize that today’s drug use will compromise future health, relationship, and employment prospects. Yet this knowledge often has little impact on action.10 TD research asks why by examining the conditions under which time affects the functional relation between behavior and its consequences.

Impulsiveness and the Battle of Selves Impulsive acts are perplexing partly because for each there exists at least one alternative course of action that would yield superior long-term well-being. The shopper could increase her funds through investing, for example, and the substance abuser could improve his lot by working to bolster health, relationship, or employment prospects rather than to obtain and consume drugs. In this sense, it is as if each individual hosts two selves, one who lives for the moment and one who must live with regret for the sins of the past. The conflict between these hypothetical selves feels as real to us as it did to Thomas Aquinas (and Sigmund Freud), who saw the forces of good (superego) and evil (id) engaged in an eternal struggle for our overt actions. But impulsive acts are behavior, not the dictates of homunculi, and no account of behavior is complete without identifying its situational determinants. When we act impulsively — when the self that lives in the moment wins — we demonstrate an apparently illogical preference for smaller benefits over larger ones. Under these circumstances, “smaller, sooner reinforcers” (SSRs) associated with impulsive acts have superior subjective value to “larger later reinforcers” (LLRs). That is, SSRs sometimes are psychologically more influential at the moment when action is dictated. Delay-based devaluation of LLRs would help to explain how this could occur, and devaluation may be a vestige of species adaptation to environments in which resources are scarce and life spans are short.11 Under such conditions, individuals who immediately exploit available resources have a survival advantage over those who, figuratively speaking, hold out for a better deal that may never be forthcoming. Consistent with an evolutionary perspective, recent research

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

120

T. S. Critchfield & G. J. Madden

suggests that different neural systems are activated in the consideration if immediate versus delayed consequences.12 Whatever the reason, humans clearly are often impulsive, and in TD research impulsivity is examined as intertemporal choice pitting actions that lead to SSRs against actions that lead to LLRs. Two core ideas underpin this research area: first, that action tendencies are driven by consequences (the outcomes that result from actions); and second, that as the amount of time between behavior and consequences increases, the capacity of consequences to influence behavior degrades. Research and theory in behavior analysis provide extensive support for, and elaboration upon, both of these ideas.13,14

How TD is Studied Laboratory research on TD can involve real consequences and delays, but we will focus on a commonly-used procedure in which responses to hypothetical choice scenarios are assessed. Although the two approaches tend to yield similar findings,15,16 hypothetical scenarios allow the study of a much broader range of delays and outcome values than can be examined otherwise.3 In many studies, the consequences under consideration are amounts of money, the nominal value of which is easy to quantify, but procedures have been developed to examine discounting of other kinds of valuable outcomes such as health, vacation opportunities, or, for drug users, access to drugs.17–19 Data typically are collected individually with the goal of depicting each individual’s tendency to discount delayed outcomes via an empirical discounting function. On each trial, subjects choose between a SSR and a LLR. Across many trials, the delay of the LLR and the amount of the SSR are manipulated, typically in accord with psychophysical scaling principles.20 The goal is to identify the current subjective value of outcomes of various delays, with subjective value defined as the magnitude of SSR that generates indifference in a choice against the LLR. This subjective value suggests the extent to which a LLR has been discounted because of delay. To illustrate, in one experimental condition the size and delay of the LLR may remain constant (e.g. $1000 to be given in 1 year) while SSR size is manipulated, across trials, from near zero to equal that of the LLR. If SSR values are presented in ascending sequence, then initially the LLR is likely to be preferred (“Which would you rather have: $1 now or $1000 in 1 year?”). As the SSR grows larger across trials (“Which would you rather have: $990 now or $1000 in 1 year?”),

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

121

preference shifts to the LLR. The SSR value at which the switch occurs helps to define subjective value. Because discounting implies a loss of value, subjective value is almost always less than the nominal value of the LLR. When the subjective value has been estimated for the same LLR at several delays — often ranging from as little as a week to as long as 25 years20 — a negatively decelerating discounting function is derived for individual subjects. A discounting function shows the extent to which subjective value changes under delay. TD data typically are displayed as per the top panel of Fig. 1. Were a LLR to occur without delay, its subjective value (data points scaled on the left ordinate) presumably would equal its nominal value (bar, scaled on the right ordinate). Both hypothetical functions show that as delay increases LLR subjective value decreases. As will be explained, the rate at which subjective value decreases under delay varies across individuals and situations. To illustrate in the abstract, function B (steep discounting = large k value in Eq. (1) below) reflects greater impulsivity than function A (shallow discounting function = small k value). The bottom panel of Fig. 1 shows the same effects viewed through the lens of prospective action planning. For all individuals, the more temporally distant an expected consequence, the weaker its impact on current behavior.

A B

A B

Time

Fig. 1. Top: Hypothetical temporal discounting data, for two hypothetical cases differing in slope of the discounting function. Subjective value of a consequence is shown as a function of delay. Bottom: Same functions depicted prospectively. See text for details.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

122

T. S. Critchfield & G. J. Madden

Form of the Discounting Function Just as memory researchers have sought to accurately describe the shape of forgetting functions,21 much effort has been expended toward determining the precise form of the TD function.22 Early accounts assumed the function to be exponential,23 as per forgetting functions, but it is now clear that a hyperboliclike model22 provides a superior fit, typically accounting for more than 90% of the variance in most individual-subject functions: SV = A/(1 + k D)s .

(1)

Here, the subjective value (SV) of a consequence is a positive function of its size (amount, A) and a negative function of its delay (D). The nonlinear scaling parameter s, consistent with psychophysical principles, represents discrepancies between actual and perceived values of time and reward magnitude. The parameter k describes the rate at which subjective value decreases as a function of delay (see Fig. 1) and typically serves as the dependent measure of primary concern in TD studies.

Individual Differences in Discounting Functions Considerable individual differences exist in the slope of the discounting function.8,24 All individuals discount delayed outcomes, but to differing degrees, and many individual differences are systematic in ways that correspond to lay assumptions about impulsivity. For example, children discount more steeply than adults, and Americans discount more steeply than Chinese or Japanese.25– 27 Consistent with the view that successful social interactions require tolerance for temporary imbalances in personal benefits,28 individuals who steeply discount delayed rewards do poorly in selected social situations.24,29,30 Individual differences in TD also correlate with clinical problems that incorporate excessive impulsivity. For example, gamblers and persons with Attention Deficit Hyperactivity Disorder discount delayed rewards more steeply than normal controls,31,32 as do abusers of opioids, nicotine, cocaine, and alcohol.19,33– 36 Thus, individuals regarded by society as impulsive tend to show especially steep discounting. A TD analysis predicts that impulsive individuals will be at risk for substance abuse. Figure 2 illustrates how the decision to use drugs may be conceived as choosing SSR (desirable drug effects) over LLR (the variety of health, occupational, and social benefits of sustained drug abstinence). Because the

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

123

Fig. 2. Preference for the SSR depends on discounting rates. See text for explanation.

psychoactive effects of drugs ensue quickly following drug-taking, Time T1 is always imminent when drugs are available. By contrast, the benefits of drug abstinence tend to be more remote.4–6 The ability to resist drug temptation thus is related to the rate of discounting. For individuals who discount LLRs very little (top panel), the subjective value of abstinence-related benefits may always outweigh the value of getting high. For individuals who discount LLRs more steeply, however, the value of drug-taking at least sometimes outweighs the discounted value of LLRs.

Failure of the Will: Preference Reversal The study of systematic individual differences implies a trait-like quality to TD, but it is not accurate to imply that some individuals are impulsive while others are “self-controlled.” All individuals are “self-controlled” at certain times and impulsive at others. Figure 3 illustrates by summarizing one commonly observed type of “failure of the will,” in which an individual commits to a long-term goal (e.g. a New Year’s resolution to stop smoking) but then yields to immediate gratification when it becomes available (a friend offers an afterdinner cigarette). Nominal values of the LLR and an SSR are scaled on the right ordinate and time is scaled on the abscissa as per Fig. 1 (bottom panel). The unlabelled (gray) function shows discounting of the SSR, while functions A and B show two different rates of discounting of the LLR, as per Fig. 1. At

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

124

T. S. Critchfield & G. J. Madden

most points on the time scale leading to the occurrence of the SSR, the same predictions apply regardless of how the LLR is discounted. When the SSR is temporally distant, the discounted value of the LLR exceeds that of the SSR. At Time T1, when the SSR is imminent, its value exceeds that of the LLR. Thus, at some point between Time T3 and Time T1, a preference reversal will occur.37–39 A resolution to stop smoking, for example, may be declared sincerely at Time T3 because of the greater relative subjective value of LLRs associated with abstinence. When a cigarette is offered, however, SSRs associated with nicotine consumption predominate. Thus, impulsiveness is situational. Additionally, within individuals, the rate of TD itself varies across situations depending on details of the consequences involved.

Three Situational Influences on TD The magnitude effect. Within individuals, larger rewards are discounted proportionally less steeply than smaller ones.8,37 In Fig. 3, the magnitude effect is illustrated in function A, in which the LLR is discounted at a lower rate than the SSR, while function B shows the LLR discounted at the same rate as the SSR (a violation of the magnitude effect). The magnitude effect has two important implications for understanding preference reversals. First, with magnitude-specific discounting, relative values of LLR and SSR are more discrepant at Time T3 with magnitude-specific discounting (function A) than otherwise (function B). Thus, the magnitude effect makes preference for the LLR, and perhaps an associated sense of having “good intentions,” especially likely when the SSR is remote. Second, the magnitude effect speaks to the timing of the preference reversal. Given the magnitude effect, the LLR is preferred at Time T2 because its subjective value exceeds that of the discounted SSR. If,

Fig. 3. Preference reversal in intertemporal choice. See text for explanation.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

125

however, the SSR and LLR are discounted at the same rate, the SSR is preferred. Thus, the magnitude effect makes impulsive decisions unlikely except when the SSR is immediately available. Pre-commitment. Most generally, the preceding illustrates that a complication of intertemporal choice is the possibility of changing one’s mind at any time. Yet it is possible, in a sense, to choose not to change one’s mind later, as long as this decision is made when the subjective value of the SSR is small. Laboratory models have shown that pigeons, like people, when faced with choices between an LLR and an SSR, often begin working for the LLR (at Time T3 in Fig. 3) and then switch to working for the SSR as it becomes more immediately available (Time T1).38 When, however, initial choices for the LLR are irrevocable — that is, they eliminate the future opportunity to work for the SSR — pigeons often will continue to choose to work for the LLR.39 Analogously, weight-loss support groups often instruct members to purchase food when they are not hungry.40 In the terms of Fig. 3, this is Time T3, when the delayed benefits of continued dieting (the LLR) outweigh the discounted benefits of a high-calorie snack (the SSR). Stocking the pantry with healthy foods at Time T3 assures that, when hunger strikes, only healthy foods are readily available and money that might have been spent on obtaining unhealthy snacks has already been spent. This “pre-commitment” decision, which helps to lock the individual into a non-impulsive course of action, is possible only prior to the preference reversal point illustrated in Fig. 3. Decisions made when the SSR is imminent (Time T1 in Fig. 3) can be expected to be impulsive. The domain effect. Outcomes in qualitatively different domains (e.g. money versus health or access to drugs) may be discounted at different rates by the same individuals.41 For example, one study found money outcomes to be more steeply discounted than health outcomes.17 Thus, it is not accurate to simply refer to an individual as “impulsive” or “nonimpulsive.” These patterns appear to vary somewhat with the type of outcome involved, possibly suggesting an experiential basis of discounting patterns. Special instances of the domain effect are relevant to substance abuse, as drugdependent individuals appear to be especially susceptible to impulsive decision making when the outcomes involve drugs. Thus, for example, heroin users discount heroin LLRs more steeply than money LLRs and similar effects obtain for cigarette smokers making choices involving delayed access to cigarettes.34 The sign effect. Some evidence suggests that unpleasant outcomes (e.g. money losses) are more resistant to discounting than pleasant ones (e.g. money gains).17,42 This finding is consistent with the assumption that unpleasant events

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

126

T. S. Critchfield & G. J. Madden

are more psychologically potent than pleasant ones,43 but note that Green and Myerson have proposed an interaction between sign and magnitude such that discounting is more pronounced for gains than for losses at small, but not at large, nominal values.8 Details aside, the sign effect serves as a reminder that associated with impulsive acts are delayed aversive consequences. For example, drug abuse involves not only forgoing LLRs associated with drug abstinence but also ignoring the delayed negative outcomes associated with chronic drug-taking. Like delayed rewards, delayed aversive events lose their psychological impact as they become more temporally distant. For example, the prospect of suffering from lung cancer tomorrow would undoubtedly decrease cigarette smoking more than the same health problem anticipated 40 years in the future. As Fig. 4 shows, the subjective value of delayed aversive events is discounted hyperbolically.44 The bar extending downward (scaled on the right abscissa) shows the nominal negative value of delayed aversive consequences associated with chronic drug use (e.g. loss of friends, money, health, etc.). When discounting (scaled on the left abscissa) is shallow (Curve B), these consequences have tangible impact at the moment when drugs are immediately available. When discounting is steep (Curve A), delayed aversive consequences have minimal impact at the same point in time. Cigarette smokers have been found to discount the value of delayed health losses more than a matched group of individuals who had never smoked. In a reversal of the typical sign effect, smokers (but not nonsmokers) discounted health losses more steeply than they discounted health gains.45 The implications of such findings are straightforward. The aggregate value of drug taking should be the sum of the immediate positive value of the drug and the discounted negative value of the drug’s delayed aversive effects.46 Individuals

More impulsive

Less impulsive

Time

Fig. 4. Temporal discounting of aversive consequences. See text for details.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

127

who steeply discount delayed aversive outcomes thus are insulated, in decisionmaking terms, from the aversive consequences of drug taking. For these individuals, the positive value of drugs is decreased very little by the delayed aversive consequences of Fig. 4, and drug-taking easily competes with the heavily discounted benefits of drug abstinence.

Two Unanswered Questions about TD In what way does time matter in TD? TD researchers share with ProM researchers an interest in determining whether time is a primary variable or a marker for other, more fundamental variables.47 In TD research, one type of concern is raised by the finding that similar hyperbolic and hyperbolic-like functions describe both TD and the discounting of outcomes based on their probability of occurrence.48 Thus, the function linking subjective value and the odds against an event’s occurrence also is a negatively decelerating hyperbola. Consequently, theorists argue over the primacy of time versus probability in discounting. On the one hand, time may be fundamental in both temporal and probability discounting. From this perspective, delay directly alters the psychological impact of events and changes in outcome frequency or probability are experienced as a change in expected wait until outcome occurrence.20 On the other hand, probability, or risk of nonoccurrence, may be fundamental. From this perspective, probability directly alters the psychological impact of events and changes in delay are experienced as changes in expected frequency.22 Theoretical debate is both fueled and complicated by the fact that not all experimental findings are parallel for TD and probability discounting — most notably, the magnitude effect is reversed for probability discounting (i.e. greater discounting of larger outcomes). We refer the reader to other sources for a thorough examination of this debate.8 Cause and effect. Because much of the evidence linking TD to substance abuse and other problems of impulsivity is correlational, there exists no consensus on whether pre-existing TD tendencies put some individuals at risk of drug use, or drug use (occurring for whatever reason) enhances impulsivity. Some evidence suggests that pre-existing individual differences in TD put individuals at risk for later problems. For example, children who show early signs of steep discounting are at increased risk of social, academic, and behavioral difficulties 10 years later.49 Perhaps analogously, the discounting rate of adults correlates with self-reported age of first alcohol, marijuana, and tobacco use and with level of contemporary drug consumption.50

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

128

T. S. Critchfield & G. J. Madden

At least two studies using animal models have shown individual differences in TD to precede drug abuse. In one study, TD for food rewards was assessed in rats that subsequently were allowed free access to two concurrently available sweetened solutions, one of which contained alcohol. The most impulsive rats showed especially strong preference for the alcohol solution.51 In another study, impulsive rats learned especially quickly to press a lever to obtain intravenous cocaine injections.52 Note that rearing and other experiences were identical for the impulsive and “self-controlled” rats, fueling speculation about genetic origins of individual differences. To our knowledge, however, no studies have selectively bred animals for TD differences or isolated genes responsible for TD. It is possible that, consistent with lay conceptions, drug use sometimes increases impulsitivity. One study found that morphine produces transient dosedependent increases in TD and that injections of naltrexone, which antagonizes morphine effects, reverse this effect.53 In general, however, studies examining the effects on TD of acute exposure to other drugs have yielded mixed results.54–56 Apparently, the withdrawal symptoms associated with chronic drug use affect TD. In a special form of domain effect, opioid withdrawal increases the degree to which delayed heroin rewards are discounted, and nicotine withdrawal increases the degree to which delayed cigarette rewards are discounted.57,58 Because withdrawal symptoms tend to be aversive, and TD has been shown to increase under acute (non-drug-related) discomfort, these effects might simply represent a generic effect of discomfort on TD. The possibility remains, however, that under conditions of withdrawal impulsive individuals may become especially so with respect to drug reinforcers.

TD and Drug-Treatment Approaches We have shown that TD principles shed light on everyday problems like substance abuse. A strong test of TD-based interpretations is the extent to which they promote the development of successful interventions. An important series of correlational studies has shown that changes in quality-of-life variables (related to relationships, employment, and so forth) predict natural recovery, help-seeking, and relapse following treatment in alcohol abusers.59–62 Although these results cannot be attributed exclusively to TD, they can be interpreted as showing that changes in the magnitude and immediacy of non-drug-related consequences alters the subjective value of drug reinforcers.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

129

This suggests that manipulating the immediate benefits of actions normally associated with non-drug-related LLRs should, through competition, reduce the likelihood of impulsive, drug-related behaviors.5,6 Similarly, anything that lowers the value of drug-related SSRs should reduce the likelihood of drug-taking. Increasing the value of non-drug LLRs. TD principles predict that if substance abusers could experience the benefits of sustained drug abstinence immediately, drug use would be curtailed. Under normal circumstances, unfortunately, these benefits are delayed, discounted, and therefore weakened in the competition with drug taking and drug SSRs. One relatively successful approach to treating substance abuse, the Community Reinforcement approach, attempts to correct this competitive imbalance by making a host of non-drug reinforcers (e.g. occupational and relationship successes) more immediately available. Substance abusers are taught the skills needed to enhance non-drug reinforcement through services such as vocational training and marital or family therapy.63 In the seminal study with alcohol abusers, treatment outcomes for Community Reinforcement therapy were superior to those of a standard therapy.64 A problem with the Community Reinforcement Approach is that acquiring new skills takes time and therefore the benefits derived from these skills are delayed. More recently, Contingency Management Therapy has combined the Community Reinforcement approach with strategies for providing immediate reinforcement for drug-free behaviors.65 In one classic study, cocaine abusers were given a daily voucher (exchangeable for desired goods) for providing a drug-free urine sample. Treatment successes, defined as periods of cocaine abstinence, were nearly twice as likely with vouchers as without.66 Numerous large-scale clinical trials show that similar techniques can be applied successfully in community-based clinics and produce sustained periods of drug abstinence.67 Reducing the value of SSRs. Medications are available that either make drugtaking aversive (e.g. disulfiram for alcohol abuse) or neutralize the drug’s reinforcing effects (e.g. naltrexone for opioid abuse).68,69 In general, however, these medications suffer from the same shortcoming as non-drug LLRs: The benefits are deferred, so there is little immediate incentive for taking them. Because the medications typically act over a short time frame, they must be taken near the point in time where drug SSRs are imminent. Under these conditions, not surprisingly, patients often choose not to take the medication.68.69 The preceding is fully anticipated by “failures of the will” illustrated in Fig. 3. Many drug abusers have renounced drugs only to see good intentions

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

130

T. S. Critchfield & G. J. Madden

(e.g. to take a medication like naltrexone) falter when drugs are immediately available. The trick, it would seem, is to find a means for making irreversible commitments to a course of self-control (the pre-commitment effect). Sustainedrelease medications that block the reinforcing effects of drugs of abuse (depotantagonists) show promise as one such technique.70 In the terms of Fig. 3, the choice to take the medication can be made at Time T3, while the effects of the medication last through Time T1.

Connections to Prospective Memory The clear relevance of TD principles to socially-important problems such as drug abuse bolsters confidence that these principles describe something fundamental and general about behavior. Temporal discounting shares with ProM an emphasis on how thinking about temporally distant events influences thought and action, and there may be synergies between the two processes. TD is about consequences, and remembering has been shown, in a variety of ways, to be a function of its consequences. It follows, therefore, that TD processes can be expected to intersect with those of prospective remembering.

Consequences and Memory Remembering normally is an effortful process47 that is most likely to occur when it serves a purpose. Memory researchers often describe “purpose” in terms of achieving a goal, but in the language of behavior analysis, inherent in every goal is some consequence, a change in circumstances that is functionally related to behavior. Outside the laboratory, goals of ProM might include paying a bill or watering a valued houseplant on time. Importantly, these actions are not an end unto themselves, but rather gain importance because of their consequences. Paying the telephone bill enables future calling opportunities. Watering the plant causes it to grow in ways that may be pleasing to behold. Within experiments, the consequences of remembering are not always well defined. Subjects may be instructed to “remember as many items as possible,” and this may harness the “good subject role” in which participants apparently perceive social benefits in doing what the experimenter asks.71 Sometimes the benefits of remembering are explicit, as when monetary payment is contingent upon task performance. Typically, however, consequences are not systematically manipulated in memory research.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

131

Nevertheless, several lines of research converge upon the notion that retrospective memory, at least, is a consequence-driven process. Studies on directed forgetting, for example, suggest that when the incentive to remember is removed, remembering is impaired.82 Other studies show that the accuracy of animal working memory is a clear function of reinforcer size, delay, and probability.73–75 A few studies explicitly link human remembering to “need odds,” or the probability that the to-be-remembered information will be required to meet some environmental challenge.76,77 A role for consequences in ProM is implied by the finding that ProM may be superior for tasks that are “important” compared to tasks that are not.78 Appraisal of the importance of the planned action is thought to affect encoding.79 Yet task importance has not always been clearly operationalized, as when subjects are instructed that a ProM task is more critical to perform than some competing task.78 Presumably, in this instance, subjects perceived social benefits in doing as the investigator asks.71 In other instances, the practical basis of task importance is clearly specified, as when subjects are asked to perform actions with clear personal benefits versus actions without such benefits, or are offered payment for completing some actions but not others.80,81 Because consequences matter in remembering, TD principles suggest that delay should mediate the extent to which they matter. Such a conjunction of memory and temporal discounting was illustrated in a study by Sargisson and White showing recognition memory to be well described by a model derived as the product of an exponential forgetting function and a hyperbolic TD function.82 Figure 4 shows pigeon recognition memory performance (expressed as log d, a modified accuracy measure) as a function of both retention interval and delay to reinforcement following the recognition response. In Fig. 4, data points represent individual performance and the concave grid shows the prediction of the following model: log d =

ae−bt . 1 + h D[1 + j t]

(2)

Equation (2), which accounted for 81% to 93% of variance in individual performance, indicates that exponential forgetting is affected by hyperbolic temporal discounting. The numerator is based on a familiar negative exponential forgetting function, with t as retention interval, a as intercept, and b as slope.21 In the denominator, consistent with the hyperbolic-like Equation (1), D is the delay

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

132

T. S. Critchfield & G. J. Madden

Fig. 5. Combined effects of retention interval and reinforcer delay on recognition memory performance.82 The former follows an exponential forgetting function and the latter follows a hyperbolic temporal discounting function. Reproduced by permission of author and publisher; copyright (2003) Society for the Experimental Analysis of Behavior, Inc. See text for additional explanation.

to reinforcement and h is the slope of the discounting curve. The fitted parameter j scales the interaction between retention interval and reinforcement delay. Equation (2) thus provides an example of an integrated model of (retrospective) remembering and discounting. It would be premature, however, to propose an analogous model for ProM, which may be functionally distinct from retrospective memory in several ways.83 For now, a TD perspective is useful in promoting general observations about ProM that may help to guide future research questions.

TD and Prospective Memory A TD perspective on ProM incorporates two assumptions: first, that ProM should be influenced by task importance and, second, that ProM should be challenged increasingly as the wait until execution of planned actions increases. ProM and task importance. Based on the assumption that task importance bears on the magnitude of the consequences of remembering, three extensions

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

133

of TD principles into ProM can be proposed. First, prospective remembering should be more likely to occur when it supports the avoidance of unpleasant outcomes than the acquisition of pleasant ones (the sign effect), and that this effect interacts with size of the consequences involved (the magnitude effect). Second, with sign held constant, prospective remembering by the same individual should vary as a function of the type of consequence that promotes remembering (the domain effect). Third, the strength of prospective remembering should vary as a function of the size or quality of the consequence that supports remembering, as task importance effects suggest, but because the rate of TD depends on consequence size (Fig. 3), whether small or large consequences promote better remembering should depend on the relative delays associated with them (the magnitude effect). ProM and delay (wait time). The assumption that wait time matters bears directly on the delay between encoding and the consequences of doing so, and therefore, because of discounting, indirectly on the subjective magnitude of consequences for remembering. In procedural terms, the delay until a planned activity will be performed incorporates features of both retention interval and delay to reinforcement for initial encoding. In the terms of Equation (2), ProM tasks conflate t and D, although it is possible to partially disentangle these factors in procedures that involve additional delay to reinforcement after completion of the planned action (Fig. 3), a manipulation with precedent in the ProM literature.84 Any prediction of delay-specific effects must contend with the finding from a number of studies that ProM responding did not vary with wait time.47,81,85–88 Here, we offer three observations. The first observation is that the preponderance of contemporary ProM experiments employ laboratory procedures involving rather brief delays and tasks of limited importance. A major stimulus to innovation in TD research was the development of procedures for the consideration of large time frames and very salient consequences, because some key effects (e.g. the magnitude effect) become obvious only in this larger scale of analysis.8 A proper consideration of the role of TD in ProM may await the development of long-delay, big-consequence procedures. A second observation is that, according to TD principles, subjective appraisals of target task importance in ProM should vary as a function of delay until the task can be accomplished and associated consequences contacted. Indirect evidence for this view comes from time-based ProM procedures in which a target task must be performed after a specified delay, and subjects may check

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

134

T. S. Critchfield & G. J. Madden

a clock to monitor the passage of time. Because monitoring responses require effort and attention, their occurrence should correlate with the value (importance) of the target task. If, in turn, task importance is static — not affected by time — then monitoring responses might be distributed equally across the delay. Figure 6 shows that this is not the case. Monitoring responses are most common as the delay expires (left panel)89 — exactly the point at which consequences of task performance are imminent and therefore essentially undiscounted. The right panel shows that the rate of monitoring responses also is affected by task importance manipulations and that, consistent with the magnitude effect in TD, the “discounting” function is steeper for high importance (magnitude) tasks than for low importance (magnitude) tasks.78 The data in Fig. 6 are anticipated in part by extensive research on fixedinterval reinforcement schedules, in which the first response occurring after a fixed period of time elapses is reinforced.90 Fixed-interval tasks are directly analogous to ProM tasks in that reinforcement depends on responding once a specified interval elapses. Under these conditions, responding is temporally distributed in a positively accelerating pattern, with effort concentrated near the end of the interval, in a sense suggesting — consistent with TD principles — that the reinforcer matters most when it is imminent. Also consistent with TD principles and with Fig. 6, the degree of positive acceleration depends on the magnitude of the reinforcer, with larger reinforcers yielding steeper slopes.91

4

10 8

3

6

High importance Low importance

2 4 1

2

0

0 1

2

3

4

5

5-min Intervals

6

1

2

3

4

30-s Intervals

Fig. 6. Number of clock-monitoring responses in successive intervals of the wait prior to responding in time-based ProM tasks. Both panels redrawn from original figures.88,99

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

135

Our third observation is related to the view that in ProM time is a secondary variable that marks the opportunity for interference by intervening activities.47 Recall that, as Fig. 3 illustrates, TD occurs in the context of competition between concurrent action tendencies. The phenomenon is interesting primarily because of preference reversals between LLR and SSR. We suggest that the same is true in many instances of ProM, with the target task and interfering tasks associated with LLR and SSR, respectively. From this perspective, some lapses of ProM are uninteresting, as, for example, when the intention to wash the car is superceded by that to obtain an emergency appendectomy. Here, the value of the competing task obviously exceeds that of the ProM task. More intriguing are cases in which a planned task of distinct importance is not executed because of interference by a relatively unimportant one — as, for example, when one fails to take heart medication after pausing a moment to answer the telephone. Such cases appear to be an inspiration for much research on ProM,2 although often it is difficult to tell for certain because task importance is vaguely defined, particularly for the cover task that is programmed into many experiments to compete with the target task. In some cases, no cover task is programmed, leaving the type and amount of competition a matter of pure speculation. During a delay to execute the target task, the point at which monitoring responses become frequent (Fig. 6) should be approximately the point at which the delay-mediated preference reversal shown in Fig. 3 occurs, but to our knowledge no existing experiment permits a direct test of this interpretation. For now, only suggestive findings can be cited. For instance, Kliegel and colleagues described competition in terms of errors committed in a cover task. Although cover-task error rate was similar overall for high- and low-importance target task conditions, in the brief period close in time to target task performance, more cover task errors were made in the high-importance condition.78 This outcome is broadly consistent with the framework suggested by Fig. 3. In general, according to a TD perspective, competition between target task and cover task is not static and the temporal dynamics of this competition should depend on consequence magnitudes. TD, attention and event-based ProM. ProM investigators often distinguish between tasks that are time-based (the intended action must occur after a given delay) versus those that are event-based (the intended action must occur following some expected event), because the two have been found to be functionally distinct in some ways.78 Upon casual inspection the present discussion appears to apply only to time-based ProM, but we caution against a rush to judge.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

136

T. S. Critchfield & G. J. Madden

Success is possible on event-based ProM tasks to the extent that the key event is noticed, and attention often is effortful. Except where processes of attention operate automatically, some form of monitoring, reminiscent of that illustrated for time-based ProM in Fig. 6, presumably is required.78 Thus, even event-based ProM may require division of attention between the target activity and potentially interfering cover tasks, but to say this begs the question of why attention is divided in a specific way. Divided attention, like memory, can be consequencedriven,92,93 so under at least some circumstances ProM interference should be mediated by the consequences of attending. TD principles merely define one of the dimensions along which consequences may vary in their impact on attention. On everyday ProM. In the everyday world, competition likely exists not only between ProM and cover tasks but also between multiple ProM tasks that occur in overlapping time frames. Thus, the effort and attention devoted to one ProM task may come at the expense of others. We are not aware of studies that examine this possibility, but note that a TD conceptual framework should apply equally to both kinds of competition. Competition between two ProM tasks could be modeled in the laboratory with minor changes in traditional procedures. Cover tasks are usually assigned minor, but immediate, importance (e.g. through instructions). It would be a simple matter to link cover-task importance to quantifiable consequences, delay to which is manipulable. The analysis of Fig. 3 predicts that the impact on a target ProM task of a small-importance cover task should be maximized when the consequences are immediate, and diminished as consequences become deferred. In light of the preference reversal illustrated in Fig. 3 and the pre-commitment phenomenon, ProM should be enhanced under conditions in which the decision to perform the target task can be accompanied by steps to limit the importance or occurrence of competing tasks. Such steps normally are not permitted in laboratory ProM procedures, but they may well be an important component of successful ProM outside of the laboratory.

Conclusions: Behavioral Choice Theory and Prospective Memory TD principles are a component of behavioral choice theory which, as Fig. 3 suggests, focuses on the consequence-based competition between concurrent courses of action.5,6,94 Consistent with this perspective, a variety of studies show that prospective remembering may be impaired by competing task

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

137

demands.47,78,87 In both cases, it is reasonable to assume that competing action plans always exist. While cognitive scientists often have regarded the importance of competing tasks as a qualitative variable (one task is more important than another), over the past few decades behavioral psychologists have developed sophisticated quantitative models of relative “importance” as defined by the consequences of competing courses of action.94–96 These models have, however, been applied to a limited variety of phenomena. The present discussion suggests that behavior scientists and cognitive scientists, who typically work independently, may find it fruitful to collaborate on the study of competition, with ProM being one phenomenon about which both may have much to say.

Author Note We are grateful to Dawn M. McBride for a critical reading of a draft of this chapter.

References 1. Neisser U, Hyman I E. Memory Observed: Remembering in Natural Contexts, 2nd ed. New York: Worth, 2000. 2. Cohen G. Memory in the Everyday World, 2nd ed. East Sussex, UK: Psychology Press, 1996. 3. Critchfield T S, Kollins S H. Temporal discounting: Basic research and the analysis of socially important behavior. J Applied Behavior Analysis 2001; 34: 101–122. 4. Vuchinich R E. Alcohol abuse as molar choice: An update of a 1982 proposal. Psychology of Addictive Behaviors 1995; 9: 223–235. 5. Vuchinich R E, Tucker J A. Behavioral theories of choice as a framework for studying drinking behavior. J Abnormal Psychology 1983; 92: 408–416. 6. Vuchinich R E, Tucker J A. Contributions from behavioral theories of choice to an analysis of alcohol abuse. J Abnormal Psychology 1988; 97: 181–197. 7. Baars B J. The Cognitive Revolution in Psychology. New York: Guilford, 1986. 8. Green L, Myerson J. A discounting framework for choice with delayed and probabilistic rewards. Psychological Bulletin 2004; 130: 769–792. 9. Rachlin H. Judgment, Decision, and Choice: A Cognitive/Behavioral Synthesis. New York: Freeman, 1989. 10. Logue A W, Anderson Y. Higher-education administrators: When the future does not make a difference. Psychological Science 2001; 12: 276–281. 11. Logue A W. Self-Control: Waiting Until Tomorrow for What You Want Today. Upper Saddle River (NJ): Prentice-Hall, 1995. 12. McClure S M, Laibson D I, Loewenstein G, Cohen J D. Separate neural systems value immediate and delayed monetary rewards. Science 2004; 306: 503–507.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

138

T. S. Critchfield & G. J. Madden

13. Lattal K A, Chase P N, eds. Behavior Theory and Philosophy. New York: Kluwer, 2003. 14. Mazur J E. Learning and Behavior. 5th edn. Upper Saddle River, NJ: Prentice Hall, 2002. 15. Baker F, Johnson M W, Bickel W K. Delay discounting in current and never-before cigarette smokers: Similarities and differences across commodity, sign, and magnitude. J Abnormal Psychology 2003; 112: 382–392. 16. Madden G J, Begotka A M, Raiff B R, Kastern L L. Delay discounting of real and hypothetical rewards. Experimental and Clinical Psychopharmacology 2003; 11: 139–145. 17. Chapman G B. Temporal discounting and utility for health and money. J Experimental Psychology: Learning, Memory and Cognition 1996; 22: 771–791. 18. Raineri A, Rachlin H. The effects of temporal constraints on the value of money and other commodities. J Behavioral Decision Making 1993; 6: 77–94. 19. Madden G J, Petry N M, Badger G K, Bickel, W K. Impulsive and self-control choices in opioid-dependent patients and non-drug-using control participants: Drug and monetary rewards. Experimental and Clinical Psychopharmacology 1997; 5: 256–262. 20. Rachlin H, Raineri A, Cross D. Subjective probability and delay. J Experimental Analysis of Behavior 1991; 55: 233–244. 21. Rubin D C, Wenzel A E. One hundred years of forgetting: A quantitative description of retention. Psychological Review 1996; 103: 734–760. 22. Myerson J, Green L. Discounting of delayed rewards: Models of individual choice. J Experimental Analysis of Behavior 1995; 64: 263–276. 23. Koopmans T C. Stationary ordinal utility and impatience. Econometrica 1960; 28: 287–309. 24. Critchfield T S, Atteberry T. Temporal discounting predicts individual competititve success in a human analogue of group foraging. Behavioural Processes 2003; 64: 315–331. 25. Du W, Green L, Myerson J. Cross-cultural comparisons of discounting delayed and probabilistic rewards. Psychological Record 2002; 52: 479–492. 26. Green L, Fry A F, Myerson J. Discounting of delayed rewards: A life-span comparison. Psychological Science 1994; 5: 33–36. 27. Green L, Myerson J, Ostaszewski P. Discounting of delayed rewards across the lifespan: Age differences in individual discounting functions. Behavioural Processes 1999; 46: 89–96. 28. Kelley H H, Thibaut J. Interpersonal Relations: A Theory of Interdependence. New York: Wiley, 1978. 29. Harris A C, Madden G J. Delay discounting and performance on the prisoner’s dilemma game. Psychological Record 2002; 52: 429–440. 30. Stevens J R, Hauser M D. Why be nice? Psychological constraints on the evolution of cooperation. Trends in Cognitive Sciences 2004; 8: 60–65. 31. Barkley R A, Edwards G, Laneri M, Fletcher M, Metevia L. Executive functioning, temporal discounting, and sense of time in adolescents with Attention Deficit Hyperactivity Disorder (ADHD and Oppositional Defiant Disorder (ODD). J Abnormal Child Psychology 2001; 6: 541–556. 32. Petry N M. Pathological gamblers, with and without substance abuse disorders, discount delayed rewards at high rates. J Abnormal Psychology 2001; 110: 482–487.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

139

33. Coffey S F, Gudleski G D, Saladin M E, Brady K T. Impulsivity and rapid discounting of delayed hypothetical rewards in cocaine-dependent individuals. Experimental and Clinical Psychopharmacology 2003; 11: 18–25. 34. Madden G J, Bickel W K, Jacobs E A. Discounting of delayed rewards in opioid-dependent outpatients: Exponential or hyperbolic discounting functions? Experimental and Clinical Psychopharmacology 1999; 7: 284–293. 35. Vuchinich R E, Simpson C A. Hyperbolic temporal discounting in social drinkers and problem drinkers. Experimental and Clinical Psychopharmacology 1998; 6: 292–305. 36. Bickel W K, Odum A O, Madden G J. Impulsivity and cigarette smoking: Delay discounting in currently, never, and ex-smokers. Psychopharmacology 1999; 147: 447–454. 37. Green L, Fristoe N, Myerson J. Temporal discounting and preference reversals between delayed outcomes. Psychonomic Bulletin & Review 1994; 1: 383–389. 38. Rachlin H, Green L. Commitment, choice and self-control. J Experimental Analysis of Behavior 1972; 17: 15–22. 39. Siegel E, Rachlin H. Soft commitment: Self-control achieved by response persistence. J Experimental Analysis of Behavior 1995; 64: 117–128. 40. Stuart R B, Davis B. Slim Chance in a Fat World: Behavioral Control of Obesity. Champaign, IL: Research Press, 1974. 41. Schoenfelter T E, Hantula D A. A job with a future? Delay discounting, magnitude effects, and domain independence of utility for career decisions. J Vocational Behavior 2003; 62: 43–55. 42. Murphy J G, Vuchinich R E, Simpson C A. Delayed reward and cost discounting. Psychological Record 2001; 51: 571–588. 43. Baumeister R F, Bratslavsky E, Finkenauer C, Vohs K D. Bad is stronger than good. Review of General Psychology 2001; 5: 323–370. 44. Lowenstein G, Prelec D. Anomalies in intertemporal choice: Evidence and an interpretation. Quarterly J Economics 1992; 107: 573–597. 45. Odum A L, Madden G J, Bickel W K. Discounting of delayed health gains and losses by current, never- and ex-smokers of cigarettes. Nicotine & Tobacco Research 2002; 4: 295–303. 46. Ho M Y, Mobini S, Chiang T J, Bradshaw C M, Szabadi E. Theory and method in the quantitative analysis of “impulsive choice” behaviour: Implications for psychopharmacology. Psychopharmacology 1999; 146: 362–372. 47. Einstein G O, McDaniel M A, Williford C L, Pagan J L, Dismukes R K. Forgetting of intentions in demanding situations is rapid. J Experimental Psychology: Applied 2003; 9: 147–162. 48. Green L, Myerson J, Ostaszewski P. Amount of reward has opposite effects on discounting of delayed and probabilistic outcomes. J Experimental Psychology: Learning, Memory, and Cognition 1999; 25: 418–427. 49. Mischel W, Shoda Y, Peake P K. The nature of adolescent competencies predicted by preschool delay of gratification. J Personality and Social Psychology 1988; 54: 687–696. 50. Kollins S H. Delay discounting is associated with substance abuse in college students. Addictive Behaviors 2003; 28: 1167–1173.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

140

T. S. Critchfield & G. J. Madden

51. Poulos C X, Le A D, Parker J L. Impulsivity predicts individual susceptibility to high levels of alcohol self-administration. Behavioural Pharmacology 1995; 6: 810–814. 52. Perry J L, Larson E B, German J P, Madden G J, Carroll M E. Impulsivity (delay discounting) as a predictor of acquisition of IV cocaine self-administration in female rats. Psychopharmacology 2005; 178: 193–201. 53. Kieres A K, Hausknecht K A, Farrar A M, Acheson A, de Wit H, Richards J B. Effects of morphine and naltrexone on impulsive decision making in rats. Psychopharmacology 2004; 173: 167–174. 54. Hellemans K G C, Nobrega J N, Olmstead M C. Early environmental experience alters baseline and ethanol-induced cognitive impulsivity: Relationship to forebrain 5-HT-sub(1A) receptor binding. Behavioural Brain Research 2005; 159: 207–220. 55. Ortner C N M, MacDonald T K, Olmstead M C. Alcohol intoxication reduces impulsivity in the delay-discounting paradigm. Alcohol and Alcoholism 2003; 38: 151–156. 56. Richards J B, Zhang L, Mitchell S H, de Wit H. Delay or probability discounting in a model of impulsive behavior: Effect of alcohol. J Experimental Analysis of Behavior 1999; 71: 121–143. 57. Giordano L A, Bickel W K, Loewenstein G, Jacobs E A, Marsch L, Badger, G J. Mild opioid deprivation increases the degree that opioid-dependent outpatients discount delayed heroin and money. Psychopharmacology 2002; 163: 174–182. 58. Mitchell S H. Effects of short-term nicotine deprivation on decision-making: Delay, uncertainty and effort discounting. Nicotine & Tobacco Research 2004; 6: 819–828. 59. Murphy J G, Correia C J, Colby S M, Vuchinich R E. Using behavioral theories of choice to predict drinking outcomes following a brief intervention. Experimental and Clinical Psychopharmacology 2005; 13: 93–101. 60. Tucker J A. Predictors of help-seeking and the temporal relationship of help to recovery among treated and untreated recovered problem drinkers. Addiction 1995; 90: 805–809. 61. Tucker J A, Vuchinich R E, Rippens P D. Predicting natural resolution of alcohol-related problems: A prospective behavioral economic analysis. Experimental and Clinical Psychopharmacology 2002; 10: 248–257. 62. Vuchinich R E, Tucker J A. Alcoholic relapse, life events, and behavioral theories of choice: A prospective analysis. Experimental and Clinical Psychopharmacology 1996; 4: 19–28. 63. Meyers R J, Smith J E. Clinical Guide to Alcohol Treatment: The Community Reinforcement Approach. New York: Guilford, 1995. 64. Hunt G M, Azrin N. A community reinforcement approach to alcoholism. Behaviour Research and Therapy 1973; 11: 91–104. 65. Higgins S T, Silverman K. Motivating Behavior Change in Illicit-Drug Abusers: Research on Contingency Management Interventions. Washington (DC): American Psychological Association, 1999. 66. Higgins S T, Delaney D D, Budney A J, Bickel W K, Hughes J R, Foerg F, et al. A behavioral approach to achieving initial cocaine abstinence. American J Psychiatry 1991; 148: 1218–1224.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

A Temporal Discounting Primer for ProM Researchers

141

67. Silverman K, Higgins S T, Brooner R K, Montoya I D, Cone E J, Schuster C R, et al. Sustained cocaine abstinence in methadone maintenance patients through voucher-based reinforcement therapy. Archives of General Psychiatry 1996; 53: 409–415. 68. Fuller R K, Branchley L, Brightwell D R, Derman R M, Emrick C D, Iber F L, et al. Disulfiram treatment of alcoholism: A Veterans Administration Cooperative study. J American Medical Association 1986; 256: 1449–1455. 69. Fals-Stewart W, O’Farrell T J. Behavioral family counseling and naltrexone for male opioiddependent patients. J Consulting and Clinical Psychology 2003; 71: 432–442. 70. Sobel BF, Sigmon SC, Walsh SL, Johnson RE, Liebson IA, Nuwayser ES, et al. Open-label trial of an injection depot formulation of buprenorphine in opioid detoxification. Drug and Alcohol Dependence 2004; 73: 11–22. 71. Rosenthal R. Experimenter Effects in Behavioral Research. New York: Irvington; 1966. 72. Roper K L, Zentall T R. Directed forgetting in animals. Psychological Bulletin 1993; 113: 513–532. 73. Jones B M, White K G, Alsop B. On two effects of signaling the consequences for remembering. Animal Learning & Behavior 1995; 23: 256–272. 74. McCarthy D, Davison M. The interaction between stimulus and reinforcer control on remembering. J Experimental Analysis of Behavior 1991; 56: 51–66. 75. Odum A L, Shahan T A, Nevin J A. Resistance to change of forgetting functions and response rates. J Experimental Analysis of Behavior 2005; 84: 65–75. 76. Anderson J R, Schooler L J. Reflections of the environment in memory. Psychological Science 1991; 2: 396–408. 77. Anderson RB, Tweney RB, Rivardo M, Duncan S. Need probability affects retention: A direct replication. Memory & Cognition 1997; 25: 867–872. 78. Kliegel M, Martin M, McDaniel M A, Einstein G O. Varying the importance of a prospective memory task: Differential effects across time- and event-based prospective memory. Memory 2001; 9: 1–11. 79. Kvavilashvili L, Ellis J. Varieties of intention: Some distinctions and classifications. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Erlbaum, 1996: 23–51. 80. Somerville S C, Wellman H M, Cultice J C. Young children’s deliberate reminding. J Genetic Psychology 1983; 143: 87–96. 81. Meacham J A, Singer J. Incentive effects in prospective remembering. J Psychology 1977; 97: 191–197. 82. Sargisson R J, White K G. The effect of reinforcer delays on the form of the forgetting function. J Experimental Analysis of Behavior 2003; 80: 77–94. 83. Einstein G O, McDaniel M A. Normal aging and prospective memory. J Experimental Psychology: Learning, Memory, and Cognition 1990; 16: 717–726. 84. McDaniel M A, Einstein G O, Stout A C, Morgan Z. Aging and maintaining intentions over delays: Do it or lose it. Psychology and Aging 2003; 18: 823–835. 85. Einstein G O, McDaniel M A, Manzi M, Cochran B, Baker M. Prospective memory and aging: Forgetting intentions over short delays. Psychology and Aging 2000; 15: 671–683.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch05

142

T. S. Critchfield & G. J. Madden

86. Hicks J L, Marsh R L, Russell E J. The properties of retention intervals and their affect on retaining prospective memories. J Experimental Psychology: Learning, Memory, and Cognition 2000; 26: 1160–1169. 87. McDaniel M A, Einstein G O, Graham T, Rall E. Delaying execution of intentions: Overcoming the cost of interruptions. Applied Cognitive Psychology 2004; 18: 533–547. 88. Nigro G, Cicogna P C. Does delay affect prospective memory performance? European Psychologist 2000; 5: 228–233. 89. Ceci S J, Baker J, Bronfenbrenner. Prospective memory: Temporal calibration and context. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory: Current Research and Issues. Chichester, UK: Wiley, 1988: 360–365. 90. Ferster C B, Skinner B F. Schedules of Reinforcement. New York: Appleton-Century-Crofts, 1957. 91. Critchfield T S, Haley R, Sabo B, Colbert J, Macropoulis G. A half century of scalloping in the work habits of the United States Congress. J Applied Behavior Analysis 2003; 36: 465–486. 92. Baron A, Myerson J, Hale S. An integrated analysis of the structure and function of behavior: Aging and the cost of dividing attention. In Davey G, Cullen C, eds. Human Operant Conditioning and Behavior Modification. Oxford: Wiley, 1988: 139–166. 93. Nevin J A, Davison M, Shahan T A. A theory of attending and reinforcement in conditional discriminations. J Experimental Analysis of Behavior 2005; 84: 281–303. 94. Herrnstein R J. On the law of effect. J Experimental Analysis of Behavior 1970; 13: 243–266. 95. Baum W M. Matching, undermatching, and overmatching in studies of choice. J Experimental Analysis of Behavior 1979; 32: 269–281. 96. Davison M, Nevin J A. Stimuli, reinforcers, and behavior: An integration. J Experimental Analysis of Behavior 1999; 71: 439–482.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

6 Time Management Jan Francis-Smythe∗

Introduction This chapter explores the relationship between time-based prospective memory (ProM) and the popular notion of time management. Time management is proposed to be the behavioral manifestation of a Time Personality,1 which in itself is the result of the interaction between inherent predisposition and environmental influence. Time-based ProM is the recall of an action to be performed at an appropriate time,2 perhaps more easily conceptualized as the ability to remember to perform intended actions in the future, for example, go for your dentist appointment at 2 pm next Thursday. In time-based prospective remembering, the action to be recalled is likely to be unrelated to the activity currently being performed and is therefore a self-initiated process. Much past research has shown age differences in ProM where older peoples’ recall is impaired relative to younger peoples.3 However, it has been found that in real-world tasks, where people are free to use their own system for remembering, often termed compensatory strategies, there are generally few age-related impairments.4 Time management may be one such compensatory strategy. This chapter proposes that time management is also used as an aiding strategy to ProM in situations where there is no failing or impairment of ProM but simply a desire to optimize ProM ability or to divert ∗ Business

School, University of Worcester, Henwick Grove WR2 6AJ, UK; e-mail: [email protected]

143

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

144

J. Francis-Smythe

cognitive resources elsewhere. This chapter therefore proposes that the use of time management is both a compensatory and an aiding strategy for ProM. Through a review of the psychological literature it explores the nature of time management, the factors which may affect its use as a compensatory or aiding strategy for ProM and the potential outcomes of its use.

The Nature of Time Management Time management is a term that is used routinely in everyday language. We hear comments like “She’s late again — if only she could manage her time better!” or “Yes, I have a slot between 2.30 and 3.00 when I could fit you in,” but what is time management and how has the academic literature defined it? According to Peeters and Rutte5 “up till now a commonly accepted scientific definition of time management is lacking” (p. 65). One of the earliest references to time management was in 1973, when according to Lakein,6 it meant determining needs, setting goals to achieve the needs, prioritizing the tasks required and matching tasks to time and resources through planning, scheduling and making lists. In 1991, Britton and Tesser7 added “carrying out the tasks” to the list of activities but it was not until 1999 that FrancisSmythe and Robertson8 drew attention to the fact that whilst all of these activities address an essential first stage in time management, namely effective planning and action, there is also an additional and perhaps even more crucial stage, that of keeping to the schedule, which is as much about monitoring as “doing.” Whilst the most recent definition5 includes monitoring progress, it omits scheduling (the planning of “when” the task(s) will be completed) and more importantly the monitoring of the schedule. In effect then, it might be suggested that their definition is more about project management than time management. Perhaps a better definition might be: Time management means setting and prioritizing goals, planning and scheduling tasks, and monitoring progress both against the schedule and of task completion. Monitoring is a key point for this chapter, which is exploring the relationship between time management and ProM (remembering to do something at a particular time). Introducing this element allows for the fact that plans may need to change, maybe as a result of changing priorities or unexpected resource limitations, reflecting the thoughts of Kleiner9 that successful time management is about “juggling the many tasks at hand at any one time” (p. 24). The definition thus needs to reflect this iterative process: Time management means setting and

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

145

prioritizing goals, planning and scheduling tasks, and monitoring progress both against the schedule and of task completion in an iterative process, in order to accommodate changing goals and priorities. The dominant view has been that people engage in time management in order to be more efficient and to enable them to better achieve their objectives. The notion that this “efficiency” is subjective and that time management may be difficult is alluded to by Koch and Kleinmann10 “time management is the self-controlled attempt to use time in a subjectively efficient way to achieve outcomes” (p. 201). Incorporating this into the definition we have: Time management is the self-controlled attempt to use time in a subjectively efficient way to achieve outcomes through setting and prioritizing goals, planning and scheduling tasks, and monitoring progress both against the schedule and of task completion, in an iterative process, in order to accommodate changing goals and priorities. The terms “self-controlled attempt” and “subjectively” are important as they reflect the very individualistic nature of time management; we don’t always succeed with our attempts and one-way doesn’t suit all.

Measuring Time Management Time management has traditionally been measured through self-report. The following three instruments are those most frequently cited in the literature (i) The Time Structure Questionnaire (TSQ11) contains 26 items that measure the degree to which people perceive their use of time as structured and purposive, measured through five scales: Sense of Purpose (SP) — related to having a sense of purpose in life, Structured Routine (SR) — related to routine and planning, Present Orientation (PO) — related to having a tendency not to think about missed opportunities or about the future, Effective Organization (EO) — related to organization, motivation and activity patterns and Persistence (P) — related to persistence and the maintenance of activities. Alphas have been reported as follows12 : SP = 0.75; SR = 0.76; PO = 0.55; EO = 0.75; P = 0.75. (ii) The Time Management Questionnaire (TMQ7 ) with three sub-scales of Short-Range Planning (extent of daily and weekly planning), Time Attitudes (extent to which time is perceived as used constructively and perception of control over time) and Long-Range Planning (perceived ability to plan ahead for several weeks and avoid procrastination). (iii) The Time Management Behavior Scale (TMBS13) containing four sub-scales of Setting Goals and Priorities (setting of goals and prioritizing of tasks to achieve them, coefficient alpha = 0.83),

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

146

J. Francis-Smythe

Mechanics-Planning and Scheduling (behaviors associated with managing time such as making lists and planning, coefficient alpha = 0.62), Perceived Control of Time (extent to which a person believes they can affect how time is spent, coefficient alpha = 0.69) and Preference for Dis-Organization (preference for disorganization in one’s workspace and approach to projects, coefficient alpha = 0.60). The TMBS has been shown to correlate significantly with the TSQ12 (r = 0.69, p < 0.001). An appraisal of each of these shows that they focus predominantly on the planning, prioritizing and scheduling activities of time management and less on the monitoring or flexibility aspects of the newly proposed definition. A revised measure of time management is therefore now called for to accommodate this.

Time Management and Prospective Memory The purpose of time management then is to achieve outcomes. These outcomes may be many and varied, for example, from completing several major projects with competing deadlines on time, to remembering to post a birthday card so it arrives either on or before the due date or to be on time for your dentist appointment. Whilst achieving the first (completing the projects) may require some use of ProM, it is likely to require the additional planning, prioritizing, scheduling and monitoring behaviors of time management as well. Achieving the second (posting the birthday card) may rely totally on ProM by simply going to buy the card and posting it two days before the actual birthday, or it may be aided by employing simple or complex time management behaviors. For example, [simple] making a note in a diary two days before the birthday to buy and post the card, [complex] buying all your birthday cards for the next three months in one visit to the shop, writing them in advance, labelling them with to-be-posted dates, filing them in a calendared to-do file, checking the file each day and actioning its contents for that day — in this case posting the card. (Dare I admit that as a time management obsessive, this is my own preferred method but with a slight modification — instead of a calendared to-do file they sit prominently date-sequenced on the kitchen window-sill!). Achieving the third (dentist appointment) may again rely totally on ProM (we all know someone who says they never need to write appointments down; they just remember them!). However, one may use a diary as a back-up to ProM or in place of ProM (aiding strategy — the busy Chief Executive only knows he or she has a dentist appointment because their time planner says so!). Time management then is much more than “remembering to do something on time”; however

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

147

for the purpose of this chapter, whilst each of the important different aspects of the literature related to time management will be considered, they will be discussed with the specific focus of how they contribute to our understanding of the relationship between time management and ProM and the extent to which they might support the case for a time-based ProM.

Time Management and External Aids The most obvious aspect of time management that is related to ProM is the use of external aids (e.g. alarm, diary, calendared to-do file, time planner) as a reminder to do something at a certain time. The effectiveness of external memory aids for improving ProM has been reported in a number of studies, e.g. Long et al.14 Burt and Forsyth15 suggest people use these external aids to avoid the stress and anxiety associated with the failures of managing their day-to-day activities; typically they show how different formats of planners can have implications for the effectiveness with which they can do this. Whilst past research on the use of such aids has shown that with increasing age there is a growing tendency to use them,16 with the advent of the new technologies such as mobile phones and “BlackBerries” (the mini-hand held computer that makes calls, accesses emails and the internet and is used so extensively by younger people) this may well no longer be the case. By way of example, my daughter, a 27 year-old accountant recalls how despite the fact that she is a “good” time manager and has “never missed” an appointment she always sets her mobile phone alarm on “vibrate” whilst she is in the office to signal important impending appointments. For the even more technically minded personal, digital assistants and Blackberries can serve the same purpose. The impact of the availability and use of these technologies on our inherent ability to estimate time will be an interesting avenue for future research. We might ask if the use of these external aids is only associated with preventing failures of ProM, that is as a compensatory strategy? The answer is no. As well as aiding recall, they provide a means of using time more efficiently, of allowing others to plan one’s time and of controlling potential ProM overload. In other words, use of an external aid (as a means of time management) is also an aiding strategy to ProM.

Time Management — Inherited and/or Learnt? Time management as discussed here has been considered as a process, a set of behaviors or, in contemporary terms, as a “competency” where “competencies”

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

148

J. Francis-Smythe

are the behaviors that a person carries out to achieve results. Whilst there have been competing explanations over the years as to what determines behavior (trait theorists, e.g. Allport,17 Staw and Ross18 ; situational theorists, e.g. Mischel19 ), it is now generally accepted that behavior is determined jointly by individual characteristics and the situation (interactionism, e.g. Lewin,20 Schneider21 ). From this interactionist perspective, whilst “competency potential” are those individual attributes necessary for someone to produce the desired behaviors,22 the manifestation of those behaviors may depend on the situation. The extent to which time management is used and the types of behaviors/styles adopted will be heavily influenced by our individual time-related characteristics but ultimately the situation will determine the behavior, that is, if the boss says that he or she wants something now and it does not fit with your schedule — you’ll still do it! The suggestion, therefore, is that we have developed a range of timemanagement behaviors (here considered as a compensatory or aiding strategy to ProM) that are founded on our own individual time-related characteristics, which we may or may not use in different situations. The extent to which time management behavior is malleable has been questioned by a number of researchers, e.g. Shahani, Weiner and Street,12 Calabresi and Cohen,23 Landy et al.,24 Macan,25 and Wessman.26 This has major implications for the time management training industry. Research that has evaluated the efficacy of time management training might be expected to go some way towards addressing this. The effects of increased use of time management on reduction of stress and increased performance is well documented (see later section), however, according to Claessens, Van Eerde, Rutte and Roe27 there is limited empirical evidence to show that time management training increases time management behaviors (for example, in Refs. 28–33). So, what might these individual time-related characteristics that precede and contribute to time management behavior be and what situations might facilitate or inhibit their display?

Factors Which May Affect the Use of Time Management as a Compensatory or Aiding Strategy for Prospective Memory Ability to Estimate Time Accurately If ProM is the ability to remember to perform intended actions in the future, when someone does not turn up for an appointment at 2 pm today, do we assume this is a ProM failure? Not necessarily. It could be that they simply did not recall

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

149

the intended action (ProM failure), or it could be that they recalled it but decided not to go (no longer an intended action), or it could be that they recalled it at some point before the scheduled time, intended to go but were then unsuccessful in managing their time between the point of recall and the time of required action (time management error) and in fact turned up late. When someone “misses” an appointment, how often do we attribute the cause to each of the above, and on what basis do we make those attributions? If it’s my mother, who is 85, then I might assume it is an age-related ProM failure, if it’s my colleague who is highly achievement-orientated and maximizes the use of every minute, then I will assume she is late because she has mis-managed her time and “over-run,” if it’s my boss then I might assume she no longer thought it important enough. What might seem at the outset a ProM failure may or may not be one. In the case of my mother, time management might be useful as a compensatory strategy for ProM in which she checked her calendar each day for appointments or set an alarm clock to remind her when to do things. In the case of my colleague, time management may have replaced ProM, she has become reliant on diaries and to-do lists and yet that still fails. Why is this? Could it be because even with an effective ProM or time management as an aiding strategy, we still need to be able to estimate time? By way of example, I remember this morning that I have to go to the dentist at 2 pm — how will I actually achieve this? By setting an alarm to go off at 1.50 pm to tell me to leave the office, or simply roughly monitoring the time passing this morning and then more accurately monitoring the time passing after say 1 pm? Doing it at the right time (unless an external aid like an alarm is used) requires us to be able to monitor the passing of time and, just as importantly, be able to estimate the duration of our preceding planned activities so that we do not over-run. Are some people better at estimating time durations than others and what relationship is there, if any, with time management or ProM? Francis-Smythe and Robertson8 proposed that “good” time management requires the ability both to plan a schedule and keep to it and that this will involve an ability to predict in advance how long a task will take (expected duration estimate), to estimate time in passing (prospective duration estimate) and to estimate retrospectively how long the task (or subcomponents of the task) have taken (retrospective duration estimate) (for a review of these three time estimation paradigms see Block34 ). Remembering to do something on time such as turning up for an appointment will also require the same degree of estimating and monitoring of time.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

150

J. Francis-Smythe

The Planning Fallacy Faced with the task of scheduling one or a list of tasks for completion, it is obviously necessary to be able to predict in advance how long each task will take to complete (expected duration estimate) irrespective of whether they are to be scheduled consecutively or simultaneously. There is a whole body of literature based around the “planning fallacy” — a tendency of people to believe that they will complete a future task significantly sooner than they actually do.35,36 There is considerable evidence that people under-estimate the time it will take to complete a task in a wide-range of activities from novel laboratory tasks to large-scale industrial projects (for review see Sagristano et al.37 ) and indeed that this optimistic prediction does not lessen as the deadline approaches.38 We might therefore expect that most people will under-estimate the duration of their preceding activities, over-run and be late for the appointment. But is this the case? Kruger and Evans37 suggest that there are a variety of complementary explanations for the planning fallacy including (a) inside/outside accounts, (b) motivation accounts and (c) unpacking, each of which is an important contributor to the effect. The inside/outside account (original explanation by Kahneman and Tversky39 ) suggests that when people consider how long it will take them to complete a task, they adopt an “inside” or “singular” perspective, focusing on the specific aspects of the task and a scenario of how it will be completed as opposed to an “outside” or “distributional” perspective based on how long similar tasks have taken in the past. A number of findings have supported this explanation, e.g. Buehler, Griffin and Ross35,40 and Vallone et al.42 The second explanation, failure to unpack, suggests that peoples’ errors arise because they do not consider each of the sub-components of a task. Koole and Spijker43 describe this as a practical application of Gollwitzer’s44 concept of implementation intentions, that proposes that, to ensure an intention is carried out one needs to form a plan (implementation intention) that sets out the where, when and how of the behavior so that the deadline is met. Gollwitzer44 describes the function of implementation intentions as “passing the control of one’s behavior on to the environment” (p. 173). The formation of implementation intentions results in a strategic switch from conscious and effortful goal-directed behavior to being automatically controlled by selected situational cues and has been shown to increase a variety of behaviors, including the likelihood of breast self-examination45 ; taking vitamin pills46 ; using public transport.47 Interestingly then, this is in effect moving the ProM requirement

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

151

from time-based to event-based. Koole and Spijker43 showed that students who formed such intentions could reduce the planning fallacy. They asked students to write a story about a particular day in the following two weeks. Half the participants were asked to form implementation intentions concerning where and when they would write the story and then to predict when it would be completed. The rest of the participants were simply asked to predict when they would finish the story without formulating implementation intentions. The group that furnished their goals with implementation intentions made more optimistic predictions but, importantly, was more likely to complete within their estimated time than the group without such intentions, thus reducing the planning fallacy. Williams48 has shown that implementation intentions work best for those low in strategic control (a measure of the amount of control people perceive themselves to have in their environment; a combination of measures of time management behavior, procrastination, order and self-discipline). Motivational explanations have also been suggested.8,49,50 Byram50 showed that participants’ predictions were more influenced by their motivations than by their cognitions. In a study in which participants were encouraged to evaluate multiple scenarios (optimistic, best guess and pessimistic) and to decompose tasks into sub-tasks, the planning fallacy was still evident and was in fact more affected by a manipulation of incentives. Under-estimation of the time it would take to complete tasks and sub-tasks became even greater when financial incentives were introduced. These findings are supported by others51,52 who suggest rather than examining cognitive processes alone examination of explicit motivations (internal and external) may be an important consideration in judgment and prediction research. Burt and Kemp49 and Francis-Smythe and Robertson8 similarly suggest that the accuracy of peoples’ estimations of the time it will take them to complete a task will be driven by the estimation strategy they use. Burt and Kemp49 asked students to estimate the duration of ten activities (e.g. going for a specified walk, completing a form, sorting a pack of cards) and then to complete five of the activities of their own choosing under timed conditions. They suggest that in their study, where participants were free once one task was completed to move on to the next task, participants used a safe-estimation strategy, whereby unused time was used to begin the next activity sooner. Francis-Smythe and Robertson8 asked participants to estimate how long it would take them to spell-check three pages of text, and then subsequently carry out the task, prior to moving on to the next task. All participants were aware they would progress together from

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

152

J. Francis-Smythe

the spell-check task to the next task irrespective of when they finished the spellcheck. In other words, free unused time was not available at the end of the spell-check task for use in beginning the next task. It was suggested that in this context, where there was no motivation for over-estimating the expected duration of the task because they could not begin the next task any sooner, participants employed a strategy aiming to maximize accuracy. Therefore, if predicting duration of a major project, where winning the contract may be determined by ability to complete soonest, an optimistic estimation strategy may be used (i.e. the planning fallacy), if predicting the duration of a series of tasks with no external constraints or when being on time for an appointment is involved, then a safe-estimation strategy may be used. Francis-Smythe and Robertson8 also suggest that the choice of strategy is likely to be affected by individual differences, typically those high on anxiety may use safe-estimation strategies, over-estimating the time a task will take with the result that they have un-allocated time gaps and are always early for tasks/appointments, whereas a person high on time urgency who needs to achieve more and more in less time will be more likely to use an accurate-estimation strategy to ensure that lost “waiting” time is minimized — even to the extent that they will risk arriving late for appointments rather than be early and waste time. It is likely then that those who succumb to the planning fallacy are perhaps in most danger of not doing something on time, as their previous task is likely to over-run. Traditionally, attempts to reduce the planning fallacy have focused on encouraging more pessimistic or realistic predictions. But the optimistic outlook can have positive effects on motivation and goal-achievement53,54 so more recently attention has turned to looking at ways to reduce the planning fallacy by encouraging behaviors to ensure the project is completed on time, e.g. Kruger and Evans,37 Aarts, Dikjerhuis and Midden55 and Gollwitzer and Brandstatter.56

Time Management and the Planning Fallacy Is there any evidence to suggest that “good” time managers are any better at estimating expected durations, that is, that they are accurate at estimating how long a task will take to complete and hence are less prone to the planning fallacy? Burt and Kemp49 carried out a study in which students completed the Time Structure Questionnaire (TSQ11 ) and were asked to estimate the expected duration of five activities (e.g. going to the library and checking out a book,

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

153

writing a one-page letter, etc.) on a 60 minute time-line. They were then asked to complete the activities under timed conditions without a watch and then to verbally estimate how long they thought the activities had actually taken (retrospective). They showed that participants who felt they were capable of managing their time were, in fact, quite poor at estimating how long it would take to perform the activities. However, Francis-Smythe and Robertson8 in a similar study utilizing tasks of spell-checking a page of text, watching a video and completing a cross-word puzzle (but where the more specific time-management scale, the TMBS13 — Time Management Behavior Scale, was used) showed that people who perceived themselves as “good” time managers were in fact more accurate at estimating the duration of a future task. Of those who did not perceive themselves as “good” time managers, some grossly over-estimated but many under-estimated the duration of an up-coming task to quite a considerable extent. This latter finding of under-estimation by many of the self-reported “poor” time managers is supportive of the planning fallacy. These findings suggest that “good” time managers are better at estimating how long tasks will take and hence should therefore not “over-run” to the same extent as “poor” time managers on tasks preceding a scheduled intended action; in other words it might be expected that they will be more likely to remember and carry out a future action on time.

Monitoring Time Keeping to a schedule once set is likely to involve monitoring time as it passes — if I have scheduled a meeting to run between 2 and 4 pm with another one to begin at 4 pm, then being on time for the second meeting requires me, with or without the use of an external aid such as a clock, to monitor time in passing. This might be by estimating time as it passes (prospective estimates) or by reflecting at certain key points how much time has just passed (retrospective estimates) and keeping a running total. In a study where participants were required to perform a series of actions at precise moments while watching a film57 and in another study where participants were given a word-learning task and asked to push a button at a pre-defined moment58 it was shown that the time checks gradually increased as the critical moment approached and that this behavior enhanced ProM (recall) performance. Costermans and Desmette59 carried out a study which required participants to carry out actions indicated by cards at specific times whilst viewing a film. This work showed that whilst overall participants checked the time more frequently as the specific times approached

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

154

J. Francis-Smythe

there were individual differences in this behavior. Those showing stronger rate increases (number of times checked per minute) performed better on the ProM task, that is, those with weaker strategies produced delayed responses. This suggests, for time-based ProM, that time monitoring is vital and that time monitoring acts as an intermittent cue in ongoing behavior to help keep in mind future required actions, and that there are variations in peoples ability/motivation to do this. In each of these studies, a clock was available for the participants to check the time. In the Francis-Smythe and Robertson8 study, referred to earlier, participants were asked to estimate time prospectively and retrospectively without sight of a clock. This work showed that those who perceive themselves as “good” time managers tend to experience time as flowing more rapidly — thus over-estimating the duration of the time that had passed. For example, 10 actual minutes passing might seem like 20 minutes and be estimated as such. This is suggested as a motivational strategy designed to enhance a sense of control over time as it ensures they will always be on time for the next task, or alternatively it may have been a cautious reaction to the lack of an external time check such as a watch which they suggest “good” time managers may be more dependent on. In summary then, these findings suggest that those who perceive themselves as good at, and utilizing, time management behaviors do seem to be more accurate at estimating future task durations (less prone to the planning fallacy) and at keeping to a schedule by over-estimating time in passing. This then might suggest that “good” time managers are more likely to complete a task on time and as such be effective in their use of time management as an aid to ProM. Interest and emotion. Other factors are known to affect perceived duration, for example, interest in task (for high interest tasks, time appears to pass more quickly8,60 ), and emotion or affect (positive emotional slides appear to pass more quickly than negative emotional ones61 ). The role of time management as a possible moderator of these effects appears not to have been explored. It is an interesting question whether time management behavior might reduce or suppress interest and emotion effects. Are those who are “good” time managers less prone to interest and emotion effects?

Time-Related Personality Characteristics A review of the literature shows that an interest in time and individual differences spans a period from the early 1900s to the present day across a range of disciplines including psychology, management and organization studies, marketing,

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

155

consumer behavior and sociology. However, research in each of these areas has proceeded along quite disparate paths with the result that in the literature, there exists a number of individual attributes related to time. These may be attempts to measure attitudes towards time, thoughts or feelings about time or time-related behaviors. Some of the more prominent constructs include time orientation, punctuality, time urgency, polychronicity and procrastination (for a review see Francis-Smythe and Robertson1 ). More recently, Francis-Smythe and Robertson1 sought to integrate much of the previous work into the identification of a time personality, a multi-dimensional construct which takes account of individual behaviors, cognition and affect and which is measured through the Time Personality Indicator (TPI62 ). This five-factor indicator comprises Leisure Time Awareness (an awareness of actual clock time and how time is being spent, Cronbach alpha = 0.71, example item “I generally prefer not to be aware of what time it is on holiday”), Punctuality (attitude towards being on time — at the level of both minutes and days, Cronbach alpha = 0.71, example item “I prefer not to be late for social appointments”), Planning (attitude towards planning and sequencing tasks in advance, Cronbach alpha = 0.70, example item “At work, I like writing lists to help me sequence my activities”), Polychronicity (preference for doing more than one thing at a time, Cronbach alpha = 0.63, “At work, I don’t mind having to have several things on the go at the same time”) and Impatience (tendency to want to complete task in hand quickly, Cronbach alpha = 0.65, “At work, I frequently feel like hurrying other people up”). Of the myriad of time-related constructs that exist, those of time urgency, procrastination and polychronicity appear to have been the most extensively researched to-date. These traits will now be considered in more detail and from the perspective of being predictors or precursors of time management behaviors, the final behaviors being the result of the interaction of any number or combination of these (and other) traits and the situation. More specifically, for this chapter, the interest will be on the way in which these characteristics may affect the use of time management behaviors as a compensatory or aiding strategy for ProM — remembering to do something on time.

Time Urgency Time urgency is considered to be a multi-dimensional construct, one of the components of a Type A behavior pattern, composed of time awareness, scheduling, list making, eating behavior and deadline control.63 Koslowsky64 describes

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

156

J. Francis-Smythe

how people high on time urgency feel the pressure of time on all occasions and during all activities, whether work, recreation or leisure. Time urgency has been shown to be an important predictor of both negative health and positive performance outcomes, typically coronary heart disease, hypertension, sleep, respiratory and digestive problems, higher classroom performance, better work-related attitudes and punctuality at work.65–69 Conte, Schwenneker, Dew and Romano,66 in a study with undergraduate students, showed that those who perceived themselves as high in time urgency estimated time as passing faster than it actually does. Dishon-Berkovitz and Koslowsky67 showed that employees high on time urgency had significantly less recorded incidents of lateness for work. From this it might be suggested that high time urgency might predispose people to use effective time management behaviors as either compensatory or aiding strategies to ProM to “remember to do things on time.”

Procrastination Procrastination has been defined as a self-regulation style that involves delay in the start and/or completion of a task70 and has been shown to result in negative mental and physical health consequences such as anxiety and depression and fewer wellness behaviors such as healthy eating and exercise.71,72 Both trait and situational explanations of procrastination appear in the literature, such as low self-esteem, low self-efficacy, perfectionism, fear of failure, boredom proneness, low conscientiousness and task aversiveness (for review see Vodanovich and Rupp72 ). Typically those low in self-esteem or perceived self-competence in carrying out a task are more likely to procrastinate. Sirois71 has shown in a study of health-related activities that the lowered self-efficacy of procrastinators led to lower intentions which in turn led to less healthy behaviors. However, previous work with college students and academic tasks shows only differences in actual behaviors of procrastinators and non-procrastinators, not in their intentions. Time Discounting. Behavioral decision-making explanations based on time discounting have also been proposed. Positive time discounting (a universal phenomenon that people value delayed outcomes less than immediate outcomes) is exhibited more by procrastinators than non-procrastinators.73 Koch and Kleinmann10 claim positive time discounting is witnessed daily with respect to time management behaviors. Typically, we see people completing all their small, non-important tasks before starting on a new major project that, although

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

157

more important, has a more distant deadline. The reward for the major project is a long way off; the satisfaction of completing the small non-important tasks now is immediate. Refusing to allow yourself to read your emails on a day that you have allocated to writing that paper is another way to minimize the bias, that is, putting ourselves into an environment where there are few alternatives when we need to work on a longer-term less urgent goal. In contrast, non-procrastinators are more likely to engage in negative time-discounting “saving the best till last,” getting rid of lots of little unrewarding tasks to then look forward to and enjoy the larger more rewarding one (a preference for improvement). Konig and Kleinmann74 showed that negative time-discounting was facilitated the closer the tasks were to each other in time and suggest that this has an implication for interventions for procrastinators — if a series of tasks can be broken down into smaller ones which are close to each other in time (i.e. the use of implementation intentions as described previously), then there is more chance that negative time discounting will be used and the smaller less rewarding tasks completed first in a preference for improvement mode, thereby decreasing procrastination. Puffer75 also relates the setting of task priorities to the emotional reaction to a task rather than the goal of efficiency (“the urgent but unimportant issue”) and the satisfaction of completing lots of small tasks. Similarly, more recent work by Tice, Baumeister and Zhang76 suggests that procrastination occurs as a result of affect regulation taking priority over other programs of self-regulation (the priority hypothesis). Tice et al.’s76 work shows that when people think doing something immediately will make them “feel” better, they opt for the small reward now rather than the larger reward later; in other words emotional distress undermines self-regulation, and procrastination is often caused by the immediate desire/need to make oneself “feel better.” Positive procrastination. Van Eerde77 proposes that whilst much of the literature to-date presents procrastination as dysfunctional, there may be certain circumstances where this is not the case, typically in creative work where the time gained serves to incubate ideas, where the time pressure creates a challenge leading to actually completing the task faster, as a temporary relief from stress or as a strategic effort to repair a bad mood. Whilst there appear to be no studies in the literature that have explored procrastination and time-based ProM, this review suggests that the extent to which someone is a procrastinator may well affect their ability to carry out an activity on time; whilst the procrastinator may remember something needs doing, their

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

158

J. Francis-Smythe

intentions may be weaker and the task does not get completed on time for any one of the explanations given previously. For those high on procrastination, the use of time management as a compensatory or aiding strategy to ProM is recommended in the literature (e.g. use of implementation intentions) to increase the likelihood that something gets “done on time.”

Polychronicity Of each of the individual difference constructs mentioned so far, polychronicity is the one that has perhaps been most extensively studied with respect to time management. The term polychronic time use is said to have originated with Hall’s78 work in anthropology, but has been more recently defined as the extent to which people prefer to engage in two or more tasks or events simultaneously.79,80 Bluedorn, Kaufman and Lane79 discuss how this is not as simple as moving between say three different tasks, focusing on each one monochronically (one at a time), but that the polychronic is likely to dovetail the tasks so that in the move between tasks there is a period when both tasks are being worked on or thought about at the same time; the more switching there is, the more polchronicity. Slocombe and Bluedorn81 draw a distinction between polychronicity and time urgency, saying that polychronicity is about the preferred pattern of activity (i.e. how work is done), whereas time urgency is about the rate of activity (i.e. how much work is done). Kaufman-Scarborough and Lindquist82 allude to the fact that the traditional view of time management up to the late 80s was founded on assumptions of monochronicity — orderly behavior where time is used for one purpose within a given clock block, that activities are sequenced, and that time is measured objectively in minutes and hours. In earlier studies, polychronicity was linked to time pressure where people were forced to tolerate interruptions and combine tasks, it was not perceived as a desirable means of time use. More recent research, however, acknowledges the positive nature of polychronicity and its role in many work settings, especially where they involve relationship building or people facing roles. Indeed, Slocombe83 suggests the decision to operate mono or polychronically may reflect a decision based on priority of relationships over task. Managed polychronicity. Whether the situation demands it or it is an individual preference, polychronicity can be accommodated in a time management behavior repertoire, such as blocking out time to be focused on one task, allowing time for interruptions, in other words, polychronics can still time manage — they just do it in a different way. This is perhaps best demonstrated in

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

159

Kleiner’s9 definition of time management, “juggling the many tasks at hand at any one time.” Indeed, in a study of householders Kaufman-Scarborough and Lindquist82 showed that polychronics reported using time planners to the same extent as monochronics, and pocket planners more often. They also updated them more often, leading them to the conclusion that both mono and polychronics engage in time management but in different ways. Interestingly, and conversely, a study by Frei, Racicot and Travagline84 showed that whilst academic faculty members’ involvement in multiple projects gave the impression they were polychronic, they had a preference for monochronicity. They handled this by dividing their time into smaller and smaller chunks so that within these chunks they could operate on one project at a time monochronically. The situation demanded polychronicity and so they developed an approach which best accommodated both the situational need and their own preference to maximize organizational and individual outcomes and personal health. Interestingly, polychronics tend to procrastinate less.82 It has been suggested that because they can handle more than one task at a time, pursuing each in bite-size chunks towards their independent goals for completion, whereas monochronic persons will delay starting a task until they know they have a free run to finish it. This might also explain the “urgent but unimportant” behavior noted by Puffer75 where, for example, we deal with our supposedly “urgent” easy emails first, but put off getting on with (i.e. procrastinate over) the important report we really need to start writing. In the same way as procrastination might enhance/facilitate creativity, so too might polychronicity. Creativity is having the ability to integrate diverse ideas and information.85 This requires polychronic thought but Persing86 shows us how this does not mean that creative workers necessarily prefer to work in polychronic environments. The key is volition and whether the individual is choosing to engage in the multiple tasks — in this case multiple thoughts. Van Eerde77 proposed that procrastination in creative work gave extra time to incubate ideas. If extra time to incubate ideas is likely to enhance creativity then maybe being polychronic and making a start on a number of projects together rather than making good progress with just one will increase the incubation time, and hence creative output for each of them. Polychronicity then may be seen as being both functional and sometimes dysfunctional dependent on a person’s preference, the imposed situation and the compatibility between the two. Contrary to earlier views, polychronicity can form an effective mode of working within a time management framework.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

160

J. Francis-Smythe

Again, whilst there have been no specific studies reported on polychronicity and ProM, it might be suggested that as polychronics are at ease with several tasks/thoughts on the go at the same time they may well have little problem “remembering to do something on time” either using ProM or time management as a compensatory or aiding strategy. Perhaps this is what is meant by the phrase “If you want something done ask a busy person!” There is a general acceptance in the individual-difference literature that the time-related personality constructs reviewed here do indeed represent traits — predispositions to respond in certain ways.66,77,87 There is little empirical evidence to suggest whether these time-related predispositions can be attributed to nature (hereditary characteristics) and/or nurture (environmental influences) and what little there is appears to be focused around Type A behavior in general and not specifically the time urgency element of it.24,88 Anecdotally, whatever the origin, there must be few amongst us who have not witnessed the similarities in families in time-related characteristics. To what extent will these traits guide behavior? From an interactionist perspective, ultimately the situation determines the behavior but one might expect a high degree of stability in temporal behaviors across time when individuals are presented with identical situations. Whilst there is some such evidence in the literature for stability of constructs (e.g. for time urgency63; time structure11; procrastination89; time personality1 ), it must be acknowledged that this is through test-retest reliability data typically collected over intervals of 1–2 months which is limited in terms of both time span and the fact that it can only be truly taken as indicative of the stability of the measures as opposed to the traits themselves. However, further tentative support comes from significant differences in time personality having been shown to exist between incumbents in different jobs, for example, teachers, managers and professionals score higher than students, careworkers and manual workers.1 Whether this reflects the dispositional nature of the construct and that people have selected themselves into the “best-fit” organization,90 or whether in fact the construct is malleable and people have adapted to the requirements of the job, is not to be determined from cross-sectional data alone. Taken together, it is here suggested that varying combinations of differing time-related constructs or traits guide the time management behaviors that an individual uses in different situations. The importance of interactions between traits is as relevant here as with other aspects of personality. For example, whilst my partner is a chronic procrastinator, he is also high on punctuality. The combination of these traits means that he experiences extreme stress near a deadline, since he has put off beginning the task for as long as possible but is then compelled to deliver “on time.” Contrast this with

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

161

a chronic procrastinator who is low on punctuality, the task is put off but there is no compulsion to deliver on time; there is little resultant stress as the two traits are quite compatible. Much more research is needed to explore the role of individual time-related traits such as time urgency, polychronicity, procrastination and time personality in the manifestation of different time management behaviors under varying circumstances.

Situational Effects So far, the effects of individual factors such as time estimation ability and timerelated personality characteristics have been considered in relation to the use of time management as a compensatory or aiding strategy for ProM. But what do we know about the extent to which the use of time management is determined by the situation? From an interactionist perspective, we believe that whilst individuals will have differing preferences and abilities related to time, ultimately it is the situation which will determine behavior. The extent to which the situation can support or tolerate these individual predispositions will determine which time management behaviors are manifest. “Situations” which impose time-related constraints on individuals are varied, from personal and family relationships to leisure, holiday and work activities. Based on P-E (person-environment) fit theory91 it has generally been proposed that “congruity” or “fit” between the individual and the situation is the ideal, leading to enhanced performance and well-being (for organizations see Francis-Smythe and Robertson92 for review; Waller, Conte, Gibson and Carpenter93 for teams; Adams and Jex,94 for families). Typically, a job which imposes the need for punctuality, such as a train driver, might not be best suited to someone who scores low on Punctuality in the Time Personality Indicator62 or a job on a production line might not best suit someone high in polychronicity. Interestingly, however, in a study of 277 parcel delivery drivers92 it was shown that time personality per se as opposed to “fit” was a better predictor of affective well-being. This was explained by suggesting that being punctual, organized, flexible and meeting deadlines serves to enhance the quality of workplace interactions and relationships with colleagues, clients and line-managers and hence affective well-being. The extent to which this might also apply in personal relationships would be an interesting avenue for future research. Another “situational” influence on time-related behavior has been proposed by Slocombe83 in the form of “others beliefs” where it is suggested the Theory of Planned Behavior95 can be used to explain polychronic behavior. Typically,

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

162

J. Francis-Smythe

behavior is determined by intention, which in turn is determined by belief (is polychronicity the best way for success?), attitude (is it good or bad for me to operate polychronically/how do I feel about it?) and subjective norm (what do others close to me think about me behaving polychronically?). Situations may impose a need for us to behave in time-related ways which are not in line with our preferences and abilities, this may include “remembering to do something on time” and in these instances we may adopt time management as either a compensatory or aiding strategy to ProM to achieve positive outcomes. Importantly however, the extent and form of time management we adopt will likely be the one that best suits our individual preferences and abilities for that particular situation. For some this may mean adopting a more focused time monitoring strategy on the day of an appointment; for others it may mean setting an alarm clock to signal when the appointment is due. For my son, when challenged with cooking the Sunday roast, it means writing a schedule of when to put each of the different components into the oven to ensure it’s delivered to perfection! Whilst we have a preference for particular behaviors/strategies, we are likely to have an individual repertoire of time management behaviors, which we apply in different situations.

The Potential Outcomes of the Use of Time Management as a Compensatory or Aiding Strategy for Prospective Memory A number of studies have demonstrated the positive effects of time management. These studies have mostly been conducted in the work and education context. Typically, researchers working with employees and students have found time management correlates positively with job satisfaction, self-reported job performance, grade point average, health and perceived control of time and negatively with job-induced and somatic tensions, strain and role ambiguity (for review see Claessens et al.27; Peeters and Rutte5 ). In 1994 Macan29 proposed that the effects of time management on each of the outcome variables was mediated by perceived control of time, that is, time management behavior gave an enhanced sense of control over one’s time which in turn produced the positive effects on performance and health. Claessens et al.,27 in a review of studies since 1994, claim that there is limited support for the notion that the effects of time management on performance and health are wholly mediated by perceived control of time and instead propose that time management behavior has both direct and indirect effects on outcomes, arguing “engaging in time management behaviors

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

163

may lead to a better temporal fit between personal resources and demands allowing one to distribute energy and attention more effectively, thereby helping to avoid or reduce delays and overload” (p. 939). Francis-Smythe and Robertson92 also showed that time personality had a direct effect on job-related affective well-being, proposing that being punctual, organized, meeting deadlines, and being flexible all serve to enhance the quality of workplace interactions and relationships with colleagues, clients and line-managers, which in turn impacts on individual well-being. Other work has focused attention on the fit between an individual’s time management behaviors and/or personality (time urgency) and those of an organization demonstrating positive outcomes on organizational commitment, performance, satisfaction and well-being.68,81,96,97 In addition, some work appears to have been carried out in non-educational/occupational contexts, typically time-management strategies have been associated with reduced levels of stress and an increase in marital adjustment among married couples94,98 and a reduction in work-family conflict.94 The extent to which these findings wholly or partially apply to the narrower and more explicit concept of the use of time management as a compensatory or aiding strategy for ProM (i.e. remembering to do something on time) is not, as far as the author is aware, documented in the literature. However, if one assumes time management is being used as an effective compensatory strategy for ProM, then it would be assumed that it should alleviate or remove the negative effects of ProM failure and where it is being used as an aiding strategy it might be expected, as a result of “doing things on time” to result in at least some of the positive outcomes described above, dependent on the context.

Summary This chapter has suggested that time management can serve as a compensatory or aiding strategy for ProM and that time-estimation ability, time-related personality characteristics and situational effects may all play a role in determining the extent and nature of time management behaviors displayed. It has been shown that it is generally accepted that effective ProM and/or time management enhances achievement of outcomes and promotes well-being by reducing negative outcomes. A number of researchers have discussed time management within the context of the stressor/strain literature and as such have proposed a number of competing models. Typically, Macan29 suggests that the relationship

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

164

J. Francis-Smythe

between time management and strain is mediated by perceived control of time; use of time management gives an enhanced sense of control over time, which in turn reduces strain. Claessens et al.27 support this view but claim that in addition, time management has a direct effect on strain, more time management less strain. In a similar vein, Koslowsky64 proposes that time management itself acts as a mediator in the stress-strain process, and is a specific coping strategy allowing the individual to control several time-related components of the stressor. He also suggests that this relationship is moderated by time urgency, people high in time urgency who engage in time management behaviors are likely to experience less strain, but those low in time urgency may show little effect. Peeters and Rutte5 however, suggest that time management is a moderator in the stress/strain process and that high time management results in less strain. Whilst the studies cited above have explored a number of stressor and strain variables, findings from these and the discussion in the preceding sections on variables which may have an impact on the use of time management as a compensatory or aiding strategy to ProM (time estimation ability, time-related personality characteristics and situational effects) might lead us to suggest a possible exploratory framework/model for further research on the role of ProM and/or time management in a stressor/strain process, where the stressor is “scheduledactivity” and the strain or outcome variables are reduced performance and well-being. Whether ProM ability/time management are mediators (as in Fig. 1. Model A — where they account directly for the effects of the scheduled-activity overload on strain — “mediators speak to how or why such effects occur”99 ) or

SITN PERS TE

ScheduledActivity Overload

TM or PM

PCT

Performance and Well-being

Fig. 1. Model A — Prospective memory/time management as mediators in stress/strain process. SITN = situational effects, PERS = time-related personality characteristics, TE = time estimation ability, TM = time management, PM = prospective memory, PCT = perceived control of time.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

SITN

Scheduledactivity Overload

PERS

TE

165

TM or PM

PCT

Performance and Well-being

Fig. 2. Model B — Prospective memory/time management as moderators in stress/strain process. SITN = situational effects, PERS = time-related personality characteristics, TE = time estimation ability, TM = time management, PM = prospective memory, PCT = perceived control of time.

moderators (as in Fig. 2. Model B — where the effect of the overload is dependent on the effectiveness of ProM/time management) has yet to be established. Through a review of recent work on time management, time estimation and time-related individual characteristics, this chapter has supported the case for a time-based ProM by demonstrating how time management can be used as a compensatory or aiding strategy to ProM.

References 1. Francis-Smythe J, Robertson I. Time-related individual differences. Time and Society 1999a; 8(2): 273–307. 2. Wilkins A J, Baddeley A D. Remembering to recall in everyday life: An approach to absentmindedness. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory. London: Academic Press, 1978: 27–34. 3. Cockburn J, Smith P T. Effects of age and intelligence on everyday memory tasks. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory: Current Research and Issues. Vol. 2: Clinical and Educational Implications. Chichester: John Wiley & Sons, 1988: 132–136. 4. Schmidt I W, Berg I J, Deelman B G. Prospective memory training in older adults. Educational Gerontology 2001; 27: 455–478. 5. Peeters M A G, Rutte C G. Time management behavior as a moderator for the job demand control interaction. Journal of Occupational Health Psychology 2005; 10(1): 64–75. 6. Lakein A. How to Get Control of Your Time and Your Life. New York: Signet, 1973. 7. Britton B K, Tesser A. Effects of time-management practices on college grades. Journal of Educational Psychology 1991; 83(3): 405–410. 8. Francis-Smythe J A, Robertson I T. On the relationship between time management and time estimation. British Journal of Psychology 1999b; 90: 333–347.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

166

J. Francis-Smythe

9. Kleiner B H. The art of handling many things at once. Managerial Auditing Journal 1992; 7(6): 24–29. 10. Koch C J, Kleinmann M. A stitch in time saves nine: Behavioural decision making explanations for time management problems. European Journal of Work and Organizational Psychology 2002; 11(2): 199–217. 11. Bond M J, Feather N T. Some correlates of structure and purpose in the use of time. Journal of Personality and Social Psychology 1988; 55(2): 321–329. 12. Shahani C, Weiner R, Street M K. An investigation of the dispositional nature of the time management construct. Anxiety, Stress & Coping 1993; 6: 231–243. 13. Macan T, Shahani C, Dipboye R L, Phillips A. College students’ time management: Correlations with academic performance and stress. Journal of Educational Psychology 1990; 82(4): 760–768. 14. Long T E, Cameron K A, Harju B L, Lutz J, Means L W. Women and middle-aged individuals report using more prospective memory aids. Psychological Reports 1999; 85(3/2): 1139–1153. 15. Burt C D B, Forsyth D K. Designing materials for efficient time management: Segmentation and planning space. International Journal of Cognitive Technology 1999; 4(1): 11–18. 16. Maylor E A. Older people’s memory for the past and the future. Psychologist 1996; 9(10): 456–459. 17. Allport G. Personality: A Psychological Interpretation. New York: Holt, 1937. 18. Staw B, Ross J. Stability in the midst of change: A dispositional approach to job attitudes. Journal of Applied Psychology 1985; 70: 469–480. 19. Mischel W. Personality and Assessment. New York: Wiley, 1968. 20. Lewin K. Field Theory in Social Science. New York: Harper, 1951. 21. Schneider B. Interactional psychology and organisational behavior. In Staw B M, Cummings L L, eds. Research in Organisational Behavior. Greenwich CT: JAI Press, 1983. 22. Robertson I T, Callinan M, Bartram D. Organizational Effectiveness. Chichester: Wiley, 2002. 23. Calabresi R, Cohen J. Personality and time attitudes. Journal of Abnormal Psychology 1968; 73: 431–439. 24. Landy F J, Rastgary H, Thayer J, Colvin C. Time urgency: The construct and its measurement. Journal of Applied Psychology 1991; 76(5): 644–657. 25. Macan T M. Time management: Correlational examination with employees’ stress, satisfaction, and performance. Unpublished Manuscript, 1992. 26. Wessman A E. Personality and the subjective experience of time. Journal of Personality Assessment 1973; 37: 103–114. 27. Claessens B J, Van Eerde W, Rutte C G, Roe R A. Planning behavior and perceived control of time at work. Journal of Organizational Behavior 2004; 25: 937–950. 28. King A C, Winett R A, Lovett S B. Enhancing coping behaviors in at-risk populations: The effects of time-management instruction and social support in women from dual-earner families. Behavior Therapy 1986; 17: 57–66. 29. Macan T. Time management: Test of a process model. Journal of Applied Psychology 1994; 79(3): 381–391.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

167

30. Macan T H. Time-management training: Effects on time behaviors, attitudes and job performance. Journal of Psychology 1996; 130: 229–236. 31. Orpen C. The effect of time management training on employee attitudes and behavior: A field experiment. The Journal of Psychology 1994; 128(4): 393–396. 32. Slaven G, Totterdell P. Time management training: Does it transfer to the workplace? Journal of Managerial Psychology 1993; 8: 20–28. 33. Woolfolk A E, Woolfolk R L. Time management: An experimental investigation. The Journal of School Psychology 1986; 24: 267–275. 34. Block R A. Experiencing and remembering time: Affordances, context and cognition. In Levin I, Zakay D, eds. Time and Human Cognition: A Life Span Perspective. Amsterdam, North Holland, 1989: 333–363. 35. Buehler R, Griffin D, Ross L. Exploring the planning fallacy — Why people underestimate their task completion times. Journal of Personality and Social Psychology 1994; 67(3): 336–381. 36. Tversky A, Kahneman D. Judgment under uncertainty: Heuristics and biases. Science 1974; 185: 1123–1131. 37. Kruger J, Evans M. If you don’t want to be late, enumerate: Unpacking reduces the planning fallacy. Journal of Experimental Social Psychology 2004; 40: 586–598. 38. Williams S E, Francis-Smythe J. Strategic control, thought focus, and the planning fallacy: Addressing individual differences in the time estimation and behaviour. Manuscript in Preparation, 2005. 39. Kahneman D, Tversky A. On the psychology of prediction. Psychological Review 1973; 80(4): 237–251. 40. Buehler R, Griffin D, Ross M. Inside the planning fallacy: On the causes and consequences of optimistic time predictions. In Gilovich T, Griffin D, Kahneman D, eds. Heuristics and Biases: The Psychology of Intuitive Judgment. Cambridge, MA: Cambridge University Press, 2002: 250–270. 41. Dunning D, Griffin D W, Milojkovic J D, Ross L. The overconfidence effect in social prediction. Journal of Personality and Social Psychology 1990; 58: 568–581. 42. Vallone R P, Griffin D W, Lin S, Ross L. The overconfident prediction of future actions and outcomes by self and others. Journal of Personality and Social Psychology 1990; 58: 582–592. 43. Koole S, Spijker M. Overcoming the planning fallacy through willpower: Effects of implementation intentions on actual and predicted task-completion times. European Journal of Social Psychology 2000; 30: 873–888. 44. Gollwitzer P M. Goal achievement: The role of intentions. In Stroebe W, Hewstone M, eds. European Review of Social Psychology. Chichester, England: Wiley, 1993: 141–185. 45. Orbell S, Hodgkins S, Sheeran P. Implementation intentions and the theory of planned behaviour. Personality and Social Psychology Bulletin 1997; 23: 945–954. 46. Sheeran P, Orbell S. Implementation intentions and repeated behaviour: Augmenting the predictive validity of the theory of planned behaviour. European Journal of Social Psychology 1999; 29: 349–369.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

168

J. Francis-Smythe

47. Bamberg S. The promotion of new behavior by forming an implementation intention: Results of a field experiment in the domain of travel mode choice. Journal of Applied Social Psychology 2000; 30(9): 1903–1922. 48. Williams S E. Stand by your plan: A new approach to the planning fallacy. Unpublished doctoral dissertation, University of Sussex, UK. 2002. 49. Burt C D B, Kemp S. Construction of activity duration and time management potential. Applied Cognitive Psychology 1994; 8: 155–168. 50. Byram S J. Cognitive and motivational factors influencing time prediction. Journal of Experimental Psychology: Applied 1997; 3(3): 216–239. 51. Henry R A. The effects of choice and incentives on the over-estimation of future performance. Organizational Behavior and Human Decision Processes 1994; 57: 210–225. 52. Henry R A, Sniezek J A. Situational factors affecting judgments of future performance. Organizational Behavior and Human Decision Processes 1993; 54: 104–132. 53. Buehler R, Griffin D, MacDonald H. The role of motivated reasoning in optimistic time predictions. Personality and Social Psychology Bulletin 1997; 23: 238–247. 54. Sherman S J. On the self-erasing nature of errors in prediction. Journal of Personality and Social Psychology 1980; 39(2): 211–221. 55. Aarts H, Dikjerhuis A P, Midden C. To plan or not to plan? Goal achievement or interrupting the performance of mundane behaviors. European Journal of Social Psychology 1999; 29: 971–979. 56. Gollwitzer P M, Brandstatter V. Implementation intentions and effective goal pursuit. Journal of Personality and Social Psychology 1997; 73: 186–199. 57. Harris J E, Wilkins A J. Remembering to do things: A theoretical framework and an illustrative experiment. Human Learning 1982; 1: 123–136. 58. Einstein G O, McDaniel M A, Richardson S L, Guynn M J, Cunfer A R. Aging and prospective memory: Examining the influences of self-initiated retrieval processes. Journal of Experimental Psychology: Learning, Memory, and Cognition 1995; 21: 996–1007. 59. Costermans J, Desmette D. A method for describing time-monitoring strategies in a prospective memory setting. Current Psychology of Coginition 1999; 18(3): 289–306. 60. Zakay D. Relative and absolute duration judgments under prospective and retrospective paradigms. Perception & Psychophysics 1993; 54: 656–664. 61. Angrilli A, Cherubini P, Pavese A, Manfredini S. The influence of affective factors on time perception. Perception & Psychophysics 1997; 59(6): 972–982. 62. Francis-Smythe J A, Robertson I T. The Time Personality Indicator (TPI) Manual. Manchester, UK: Robertson-Cooper Ltd. Manual in Preparation, 2005. http://www. robertsoncooper.com/company/homepage.aspx 63. Conte J M, Landy F J, Mathieu J E. Time urgency: Conceptual and construct development. Journal of Applied Psychology 1995; 80(1): 178–185. 64. Koslowsky M. Some new organizational perspectives on moderators and mediators in the stress-strain process: Time urgency, management, and worker control. In Erez M, Kleinbeck U, eds. Work Motivation in the Context of a Globalizing Economy. Mahwah, NJ: USum Associates, 2001: 313–327. 65. Conte J M, Mathieu J E, Landy F J. The nomological and predictive validity of time urgency. Journal of Organizational Behavior 1998; 19: 1–13.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

Time Management

169

66. Conte J M, Schwenneker H H, Dew A F, Romano D M. Incremental validity of time urgency and other Type A subcomponents in predicting behavioral and health criteria. Journal of Applied Social Psychology 2001; 31(8): 1727–1748. 67. Dishon-Berkovits M, Koslowsky M. Determinants of employee punctuality. The Journal of Social Psychology 2002; 142(6): 723–739. 68. Greenberg J. Time urgency and job performance: Field evidence of an interactionist perspective. Journal of Applied Social Psychology 2002; 32(9): 1964–1973. 69. Wright L, McCurdy S, Rogoll G. The TUPA scale: A self-report measure for the Type A subcomponent of time urgency and perceptual activation. Psychological Assessment 1992; 4(3): 352–356. 70. Ferrari J R, Tice D M. Procrastination as a self-handicap for men and women: A task avoidance strategy in a laboratory setting. Journal of Research in Personality 2000; 34: 73–83. 71. Sirois F M. Procrastination and intentions to perform health behaviors: The role of selfefficacy and the consideration of future consequences. Personality & Individual Differences 2004; 37(1): 115–128. 72. Vodanovich S J, Rupp D E. Are procrastinators prone to boredom? Social Behavior and Personality 1999; 27(1): 11–16. 73. Schouwenburg H C, Groenewoud J T. Study motivation under social temptation: Effects of trait procrastination. Personality and Individual Differences 2001; 30: 229–240. 74. Konig C J, Kleinmann M. Business before pleasure: No strategy for procrastinators? Personality and Individual Differences 2004; 37: 1045–1057. 75. Puffer S. Task completion schedules: Determinants and consequences for performance. Human Relations 1989; 42(10): 937–955. 76. Tice D M, Baumeister R F, Zhang L. The role of emotion in self-regulation: Differing role of positive and negative emotions. In Philippot P, Feldman R S, eds. The regulation of emotion. Mahwah, NJ: USum Associates, 2004: 213–226. 77. Van Eerde W. A meta-analytically derived nomological network of procrastination. Personality & Individual Differences 2003; 35(6): 1410–1418. 78. Hall E T. The Silent Language. Garden City, New York: Doubleday, 1959. 79. Bluedorn A C, Kaufman C, Lane P M. How many things do you like to do at once? An introduction to monochronic and polychronic time. Academy of Management Executive 1992; 6(4): 17–26. 80. Kaufman C J, Lane P M, Lindquist J D. Exploring more than 24 hours a day: A preliminary investigation of polychronic time use. Journal of Consumer Research 1991a; 18(December): 392–401. 81. Slocombe T E, Bluedorn A C. Organizational behavior implications of the congruence between preferred polychronicity and experienced work-unit polychronicity. Journal of Organizational Behavior 1999; 20: 75–99. 82. Kaufman-Scarborough C, Lindquist J D. Time management and polychronicity: Comparisons, contrasts and insights for the workplace. Journal of Managerial Psychology 1999; 14(3/4): 288–312. 83. Slocombe T E. Applying the theory of reasoned action to the analysis of an individual’s polychronicity. Journal of Managerial Psychology 1999; 14(3/4): 313–322.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch06

170

J. Francis-Smythe

84. Frei R L, Racicot B, Travagline A. The impact of monochronic and Type A behavior patterns on research productivity and stress. Journal of Managerial Psychology 1999; 14(5): 374–387. 85. Sethia N K. The shaping of creativity in organisations. In Hoy F, ed. Academy of Management Best Papers Proceedings 1989. Atlanta, GA: Academy of Management, 1989: 224–228. 86. Persing D L. Managing in polychronic times: Exploring individual creativity and performance in intellectually intensive venues. Journal of Managerial Psychology 1998; 14(5): 358–373. 87. Ganster D C, Schaubroeck J, Sime W E, Mayes B T. The nomological validity of the Type A personality among employed adults. Journal of Applied Psychology 1991; 76: 143–168. 88. Wright L. The Type A behavior pattern and coronary artery disease. American Psychologist 1988; 43: 2–14. 89. Lay C H. At last, my research article on procrastination. Journal of Research in Personality 1986; 20: 474–495. 90. Holland J L. Making Vocational Choices. 2nd edn. Englewood Cliffs, NJ: Prentice Hall, 1985. 91. Edwards J R. Person-job fit: A conceptual integration, literature review, and methodological critique. In Cooper C L, Robertson I T, eds. International Review of Industrial and Organisational Psychology. Chichester: Wiley, 1991: 283–357. 92. Francis-Smythe J A, Robertson I T. The importance of time congruity in the organisation. Applied Psychology: An International Review 2003; 52(2): 298–321. 93. Waller M J, Conte J M, Gibson C B, Carpenter M A. The effect of individual perceptions on deadlines on team performance. Academy of Management Executive 2001; 26(4): 586–600. 94. Adams G A, Jex S M. Relationships between time management, control, work-family conflict and strain. Journal of Occupational Health Psychology 1999; 4(1): 72–77. 95. Ajzen I, ed. From Intentions to Actions: A Theory of Planned Behavior. New York: Springer, 1985. 96. Kaufman C, Lane P M, Lindquist J D. Time congruity in the organization: A proposed quality of-life framework. Journal of Business and Psychology 1991b; 6(1): 79–106. 97. Schriber J B, Gutek B A. Some time dimensions of work: Measurement of an underlying aspect of organisation culture. Journal of Applied Psychology 1987; 72(4): 642–650. 98. McLaughlin M, Cormier L S, Cormier W J. Relationship between coping strategies and distress, stress, and marital adjustment of multiple-role women. Journal of Counseling Psychology 1988; 35(2): 187–193. 99. Baron R M, Kenny D A. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Applied Psychology 1986; 51: 1173–1182.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

7 Transcending the Now: Time as a Dimension of Psychological Distance Cheryl J. Wakslak∗ , Yaacov Trope† and Nira Liberman‡

Introduction The evaluation of events can differ dramatically based upon whether the event is considered in the near or distant future. Construal Level Theory1 suggests that temporal shifts in evaluation are driven by differences in the mental construal of near and distant events. In particular, CLT posits that distant events are represented by their essential, abstract, and global features (high-level construals), whereas near events are represented by their peripheral, concrete, and local features (low-level construals). In the current chapter, we use this framework to examine temporal shifts in representation, prediction, evaluation and behavior. We begin by discussing the concept of level of construal, as well as the theoretical rationale behind CLT’s core assumption. Next, we review empirical research on differences in the consideration of near and distant events. Finally, we conclude by placing research on temporal distance within the context of a more general approach to understanding psychological distance.

∗ New

York University Department of Psychology, New York, USA; e-mail: [email protected] New York University Department of Psychology, New York, USA; www.psych.nyu.edu/trope ‡ Tel Aviv University Department of Psychology, Israel. †

171

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

172

C. J. Wakslak, Y. Trope & N. Liberman

Level of Construal The concept of level of construal begins with the basic notion that objects and events are classified into categories despite the fact that no two objects or events are identical. Thus, categorization requires that one disregard unique features of the given object, and therefore involves an implicit decision about which features are central to the object and which are more incidental. Take, for example, the categorization of an object as a book. This categorization highlights the object’s similarity to other books and involves the decision that the function of the book is its central property, disregarding other properties such as the book’s color and size. Of course, a concrete representation can lend itself to multiple abstract representations. An abstract representation is selected among different possible abstractions according to context-defined relevance, which, in turn, may be affected by one’s goals. For example, for someone interested in reading, the categorization of an object as a book would be relevant; for someone interested in sorting their recycling, “paper object” might be a more relevant abstract conceptualization of the same object. Regardless of the particular abstract representation that has been chosen, moving to the abstract level involves omitting the features that are perceived to be less important to the abstract construct in question while retaining those considered more central or important. Because abstract representations necessarily impose one of possibly many alternative interpretations, and because irrelevant or inconsistent details are omitted from the abstract representation or assimilated to it, abstract representations may be expected to be simpler, more structured, and less ambiguous, than concrete representations.2,3 Abstraction thus involves moving to a more schematic, simple and coherent representation. The process of abstraction is not an all-or-none phenomenon. Representations become more abstract and schematic the more unique, incidental features that are omitted. Thus, object categorization may be thought of as organized hierarchically (e.g. convertible, car, and vehicle) with representations that are higher in the hierarchy more inclusive and less concrete.4 In the same way, traits form hierarchies (e.g. an excellent guitarist, musical, talented) such that more abstract traits are less detailed about the behaviors, objects, circumstances, and people they involve.5 Further, categories are often structured around goals (e.g. the goal to lose weight). In these cases, goal relevant features (e.g. the number of calories in a bag of chips) are more central than goal irrelevant features (e.g. the crunchiness of the chips). Goal directed actions likewise form hierarchies of abstractness, as goals could be translated into more abstract, superodinate

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

173

goals.6–8 In such hierarchies, each action (e.g. study for an exam) has a superodinate, abstract level, which answers the question of why the action is performed (e.g. do well) and a subordinate, concrete level, which supplies the details of how the action is to be performed (e.g. read a textbook). Operationally, one way to distinguish between higher- and lower-level features of an object or event is by asking how much difference it would make if the particular feature were altered or removed. Altering a high-level feature should produce a more substantial change in the concept in question than altering a low level feature. For example, consider the difference between changing the content of an exam and the format of an exam. The exam would be altogether a different thing if the content were changed; this would be much less the case if it were the format that were changed. Thus, the content of the exam comprises a higher-level feature of the exam than the formatting. To repeat, the basic premise of Construal Level Theory is that distant events are mentally represented in a schematic and abstract manner that includes the central, superordinate features of the event, whereas near events are mentally represented in a less schematic, more concrete manner that allows for the inclusion of incidental, less important features. Why might this be the case? One possible reason for this association stems from the relationship between direct experience and information about an event. Typically, as an event becomes removed from direct experience (e.g. as an event is placed farther into the future), information about the event becomes less available or reliable, leading individuals to form a more abstract and schematic representation of the event. CLT assumes that an association thus forms between psychological distance and level of construal, and that this association is over-generalized, causing individuals to form high-level construals of distant events and low-level construals of near events even in cases when the amount and reliability of information is constant. Now that we have laid out the basic theoretical propositions of Construal Level Theory, we examine recent empirical evidence in support of it. We begin by considering the effects of temporal distance on level of mental representation, and then look at implications of temporal shifts in construal for prediction, evaluation, and choice.

The Effect of Temporal Distance on Mental Representation Construal Level Theory suggests that the representation of distant future events will be more abstract, broad, and structured than the representation of near future

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

174

C. J. Wakslak, Y. Trope & N. Liberman

events. Research has examined this claim as it applies to the categorization of objects, actions, traits, and value driven behavior. For example, in one study, Liberman, Sagristano, and Trope9 asked participants to imagine an event (e.g. camping trip) to take place either in the upcoming weekend or a weekend a few months later and to classify a given set of 38 objects related to the event (e.g. tent, toothbrush) into as many mutually exclusive and exhaustive groups as they deemed appropriate. If participants formed more abstract representations of distant future events, then they should generate fewer, more broad categories for distant future events than for near future events. Indeed, this was the case. In a second study, Liberman et al.9 examined the structure underlying people’s preferences for near and distant events. As expected, multidimensional scaling showed that preferences for distant events were organized around simpler structures than those for near events. Thus, it was more difficult to reduce near future preferences to general underlying dimensions than it was to do so for distant future preferences. Studies examining the construal of actions have similarly found evidence that more distant actions are construed at a higher level. Liberman and Trope10 asked participants to imagine themselves engaging in various activities (e.g. reading a science fiction book) either tomorrow or next year and describe these activities. As expected, people used more superordinate, high-level descriptions of distant activities (e.g. “getting entertained”) and low level descriptions for near future activities (e.g. “flipping pages”). A related, forced-choice study used an adapted version of Vallacher and Wegner’s11 Level of Personal Agency questionnaire that was originally designed to assess stable individual differences in action identification. The questionnaire presents 19 activities (e.g. “locking a door”), each followed by two restatements, one corresponding to the why (high-level) aspects of the activity (e.g. “putting a key in the lock”) and the other corresponding to the how (low-level) aspects of the activity (e.g. “securing the house”). As predicted, participants chose significantly more high-level, why restatements when the activities were described as occurring in the distant-future than when the same activities were described as occurring in the near-future. Eyal12 extended this analysis by focusing in particular on the construal of events in terms of abstract, value-laden principles. For example, in one study, participants read about distant future and near future situations that involved an abstract principle or a dilemma (e.g. “In a few days/in a few years, the University will set to increase the number of minority students”), and were instructed to choose a description of this situation either in terms of a global principle or

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

175

in terms of a lower level action, devoid of moral implications (e.g. “endorsing affirmative action” versus “making changes to admission lists”). Distant future situations were perceived in terms of high level principles more than near future situations. Presumably, principles more easily apply to the distant future, but as the situation gets closer in time, morals and ideologies seem to lose their relevance. Extending beyond identification measures of construal, Day and Bartels13 examined the extent to which differences in the construal of events influenced ratings of similarity, a construct that has been widely implicated in a variety of fundamental cognitive processes such as retrieval, categorization, and inference. Similarity ratings were collected for pairs of events sharing either high-level or low-level commonalities, and described as occurring in either the near or distant future. Because information that is more salient in a representation is assumed to be given more weight in judgments of similarity, temporal shifts in representation should influence similarity ratings. Thus, for distant future events, similarity ratings should be largely driven by abstract, structuring information such as goals, causes and relationships, while for near future events similarity ratings should be driven by commonalities in low-level, concrete aspects of the situations. Consistent with predictions, an interaction was observed between temporal distance and commonality level, such that pairs with high-level commonalities were perceived as somewhat more similar in the distant than the near future, while pairs sharing low-level features were perceived as more similar in the near future than the distant future. Like identifying an event as goal or value driven, ascribing a behavior to an abstract, decontextualized personal disposition involves a high-level construal of behavior.14–16 Thus, CLT predicts that individuals will be more likely to demonstrate the correspondence bias, i.e. the tendency to attribute situationallyconstrained behavior to the corresponding personal disposition,17,18 when the social target is more distal. Nussbaum, Trope and Liberman19 used the Jones and Harris20 attitude attribution paradigm to test this hypothesis with respect to temporal distance. Student participants from Tel-Aviv University read an essay arguing in favor of Israel’s withdrawal from Lebanon. (The study was conducted a few months before Israel’s withdrawal from Lebanon in June, 2000.) They were told that the essay was written by a student who had been instructed either to express her own opinion (unconstrained condition) or to argue in favor of withdrawal (situationally-constrained condition). Participants were asked to estimate the likelihood that the writer would express pro-withdrawal attitudes

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

176

C. J. Wakslak, Y. Trope & N. Liberman

in a variety of near future (next day) or distant future (a year later) situations (e.g. express pro-withdrawal attitudes in a conversation with friends, attend a pro-withdrawal rally). The results showed that the judged likelihoods of essayconsistent (pro-withdrawal) behavior in the near future were moderated in view of the situational constraints, whereas the judged likelihoods for the more distant future were high regardless of situational constraints. These findings demonstrate that the correspondence bias, the tendency to underweight low level, situational constraints on observed behavior, is more pronounced when this behavior is used for predicting the distant future than the near future. If distant behaviors are more closely linked to traits, then people should expect others to behave more consistently across different situations in the distant future than in the near future. Nussbaum et al.19 tested this hypothesis by asking participants to predict an acquaintance’s behavior in four different situations (e.g. a birthday party, waiting in line in the supermarket) in either the near future or the distant future. Participants predicted the extent to which their acquaintances would display 15 traits (e.g. behave in a friendly versus unfriendly manner) representative of the Big Five personality dimensions. As hypothesized, the results showed that participants expected others to behave more consistently across distant future situations than across near future situations. This was manifested in both lower cross-situational variance and higher cross-situational correlations across predicted distant future behaviors in the four situations than across predicted near future behaviors. The structured representation of temporally distant events, indicative of highlevel construals, is not limited to verbal tasks. Förster, Friedman, and Liberman21 examined performance on visual tasks that require abstraction of coherent images from fragmented or “noisy” visual input (e.g. the Snowy Picture Test and the Gestalt Completion Test22 ). If the representation of distant events is more structured, then participants should be better able to abstract the structure from these visual tasks when they are placed in a temporally distant context. Participants were asked to imagine their lives tomorrow (near future perspective) or on a day one year from now (distant future perspective) and to imagine working on the experimental task on that forthcoming day. As expected, performance on both the Snowy Picture Test and the Gestalt Completion Test was higher in the temporally distant condition than the temporally near condition. Thus, individuals’ ability to detect structure in a visual image was facilitated when the task was distanced in time, suggesting that construal effects have implications for perceptual processes as well as language based ones.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

177

Finally, temporal distance may not only affect construal, but may be affected by construal. We argued earlier that an association forms between distance and construal that is over-generalized beyond situations in which distance entails less knowledge. Through this association, the relationship between distance and construal may have become bi-directional, such that events construed at higher level are perceived to be more distant. Liberman, Trope, Macrae and Sherman23 examined the effect of construal level on the temporal distance of activity enactment. In one of their studies, participants were first asked to indicate either “why” (i.e. high-level construal) or “how” (i.e. low-level construal) a person would perform an activity (e.g. “Ron is considering opening a bank account. Why (How) would Ron do that?”), and were then asked to estimate how much time from now the person would do the activity. As predicted, participants indicated more distant enactment times after a high-level, “why” construal than after low-level “how” construal. The authors found similar effects with other manipulations of level of construal, and with participants’ estimates of the enactment time of their own activities. In sum, extensive research conducted within the framework of Construal Level Theory demonstrates that future temporal distance enhances the level of construal of objects, actions, situations, and people. What are the implications of these differences in construal? What predictions does CLT offer about the manner in which temporal distance impacts prediction, evaluation, and choice? We turn now to examine this set of questions.

The Effect of Temporal Distance on Predictions, Evaluations, and Behavior Prediction According to Construal Level Theory, predictions about a distant future event should be based on the implications of high-level rather than low-level construals. Theories are, by definition, abstract constructions of schematic relations among entities in an idealized, noise-free world. When tested empirically, however, theoretical predictions may fail to replicate due to nonsystematic influences of the specific conditions and circumstances of the test situation. Focusing on theories (or high-level construals of experiments) may therefore enhance confidence in theoretical predictions, whereas focusing on nonsystematic factors (low-level construals of experiments) may reduce confidence. For example, economic theory posits that increasing interest rates causes the stock market to

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

178

C. J. Wakslak, Y. Trope & N. Liberman

decline. The theory acknowledges that other factors might also affect the stock market, but treats them as noise. According to CLT, then, economists should be more confident in predicting that the stock market will fall if interest rates are raised when considering the more distant future. Normatively, this should not be the case; predictions about more distant entities should allow for greater uncertainty regarding unknown factors, and should therefore be made with less confidence because less is usually known about them. The logic of CLT is consistent with the view that errors in prediction may stem from individuals’ over-reliance on schematic models of future situations and neglect of background contextual factors.24– 27 However, while past research compares predictions to actual outcomes, CLT extends this reasoning in order to compare near future and distant future predictions. For example, in one study28 participants imagined replicating five classic findings in psychology. Half of the participants read a short description of the theory that gave rise to the predicted results, whereas the other half did not receive information about the theory, but rather read only the description of the study and the prediction. Participants in both experimental conditions imagined themselves conducting an experiment either tomorrow morning (near future condition) or a year from now (distant future condition) and indicated how confident they were that the predicted effect would be found in their experiment on a 0% to 100% scale. Results indicated that temporal distance increased participants’ confidence only when they were provided with a theoretical basis for the predictions. These results support the assumption of CLT that temporal distance enhances confidence in prediction only when this confidence derives from high level constructs. Several other studies assessed confidence in predicting one’s own performance on a general knowledge quiz.28 An earlier study empirically established that the knowledge domain of a quiz question is perceived as central and the format of the quiz question as peripheral. Participants expected to take a general knowledge quiz either on the same day or two months later. The quiz consisted of the same set of questions, drawn from a variety of knowledge domains, asked in either a relatively easy or hard format. Specifically, in one study, the quiz consisted of either multiple-choice questions (relatively easy format) or open-ended questions (relatively hard format). In another study, the quiz consisted of questions with either two response alternatives (relatively easy) or four response alternatives (relatively hard). Participants’ perceived ability in each knowledge domain (e.g. how knowledgeable you are in geography, history etc.) was also measured. The results showed that difficult question format appropriately

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

179

reduced confidence in near future performance, but failed to reduce confidence in distant future performance. Further, participants’ beliefs about their general knowledge in different domains predicted their confidence in that domain in the distant future better than in the near future. Thus, consistent with CLT, the low level aspect of the quiz (question format) affected confidence in near future outcomes more than in distant future outcomes, while the high-level aspect of the quiz (domain knowledge) affected confidence in distant future outcomes more than near future outcomes. The Nussbaum et al.19 studies on the effect of future temporal distance on dispositional attribution that were discussed in the first section of this chapter as an example of the effect of future temporal distance on level of construal are also relevant to prediction. They show that people base their predictions of others’ more distant future behavior more on high-level, dispositional attributions and less on low level, situational attributions. Evaluation and Behavior How do people evaluate and make choices about distant future outcomes, as opposed to near future outcomes? An assumption shared by a variety of behavioral scientists is that the value of an outcome is discounted or diminished as temporal distance from the outcomes increases.29−32 Contrary to the claim of overall time discounting, CLT proposes that the effect of temporal distance on the attractiveness of an option will depend upon the value associated with the high-level construal of the option (high-level value) and the value associated with the low-level construal of the option (low-level value). Temporal distance should increase the weight of high-level value and decrease the weight of lowlevel value. As a result, temporal distance should shift the overall attractiveness of an option closer to its high-level value than to its low-level value. When the low-level value of an option is more positive than its high-level value, the option should be more attractive in the near future (time discounting). However, when the high-level value of an option is more positive, the option should be more attractive in the distant future (time augmentation). This hypothesis was examined with different manipulations of high versus low levels of construal: primary, goal related versus secondary, goal irrelevant sources of value; feasibility versus desirability and expectancy versus value in gambles; arguments in favor versus arguments against an action; abstract and primary attitudes and values versus concrete and secondary attitudes and values; and high versus low priority issues in a multi-issue context. We briefly review

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

180

C. J. Wakslak, Y. Trope & N. Liberman

this literature here, and refer the reader to more extensive reviews elsewhere for more detailed information.1,10,33 Primary versus secondary features of objects. Consider the purchase of a product with multiple features, some of which are related to its primary function and others which are not. According to CLT, as the product becomes removed in time, the weight placed on the primary features relative to secondary features in determining the value of the product should increase. In a test of this prediction, Trope and Liberman33 asked participants to indicate how satisfied they would be in either the near or distant future with the purchase of one of two radio sets: one that had good sound but a poor built-in clock, and one that had poor sound but a good built-in clock. Given that listening to programs was described as one’s goal in buying the set, sound quality is more central than the quality of the clock, and should be emphasized in a higher level construal of the radio set. CLT therefore predicts that the advantage in ratings of the “high quality sound/poor clock” radio set over the “low quality sound/good clock” radio set should be stronger in the distant future than the near future, with ratings of the clock with good sound increasing and ratings of the clock with bad sound decreasing with temporal distance. The results confirmed these predictions. Further, in additional studies the same temporal changes in preference were found for evaluations of experimental sessions with interesting and boring main and filler tasks as well as experimental sessions with both affective and cognitive goal-relevant and goal-irrelevant features. Desirability and feasibility. An important distinction made when considering goal directed action is between desirability and feasibility concerns. Given that desirability refers to the value of an action’s end state (and thus pertains to why an activity is performed), and feasibility refers to the ease or difficulty of reaching the end state (and thus pertains to how an activity is performed), higher level construals of actions should emphasize desirability over feasibility concerns to a greater degree than lower level contruals of the same actions. CLT therefore predicts that distant future preferences will be guided by desirability concerns over feasibility concerns to a greater extent than near future preferences. Liberman and Trope10 tested these predictions in a series of hypothetical and realistic choice scenarios. For example, in one study conducted in a field setting, university students choose between four assignments to be performed in either the near future (given immediately to be due in one week) or the distant future (given nine weeks later to be due one week afterward). Participants

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

181

stated how much they would like to do each of the four assignments, which varied in interest (desirability) and ease (feasibility). They were told that the assignments would be distributed according to their preferences. Consistent with CLT, the preference for the easy but uninteresting assignment decreased over time, whereas the preference for the hard but interesting assignment increased over time. Thus, in selecting a near-future assignment, students were willing to sacrifice interest (desirability) for the sake of ease (feasibility). In contrast, in selecting a distant-future assignment, students were willing to sacrifice ease for the sake of interest, thus committing themselves to a desirable but less feasible task. A similar temporal pattern was obtained with various other options.10 Feedback seeking is another important decision that often pits feasibility against desirability concerns. Freitas, Salovey and Liberman34 reasoned that feedback seeking involves a conflict between the goal of gaining information about oneself (a desirability consideration) and the difficulty of going through the process of self evaluation (a feasibility consideration). They therefore predicted and found that distant future feedback preferences depended on the accuracy of the offered feedback, whereas near future feedback preferences depended on the evaluative implications of the feedback. Informative but unflattering feedback was preferred for the distant future, whereas uninformative but flattering feedback was preferred for the near future. An interesting implication of CLT’s view on feasibility and desirability concerns the effect of temporal distance on planning. Liberman and Trope10 conceptualized time constraints as a feasibility aspect of an activity and investigated the role of time constraints and desirability of activities in near and distant future planning. They showed that plans for the distant future tend to reflect desirability of activities and disregard time constraints, thus creating a tendency to over-commit. It appears that in making distant future plans individuals consider each activity in isolation and fail to take into account that each activity they plan comes at the expense of some other activities they may want to engage in at the same time. The desirability/feasibility distinction may also be used to characterize dimensions of positive gambles in which there exists an opportunity to win a desirable prize. In this situation, the prize can be conceptualized as the desirability, end-state dimension, whereas the probability of winning is a subordinate consideration having to do with the random mechanism that determines the feasibility of winning the prize. In this view, the prize value pertains to the high-level construal of a gamble, while the probability pertains to the low-level construal.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

182

C. J. Wakslak, Y. Trope & N. Liberman

(Indeed, participants have empirically supported the notion that probability in such gambles is subordinated to the payoff.35 ) According to CLT, then, people will assign more weight to payoffs and less weight to probabilities in deciding for the more distant future. Sagristano, Trope and Liberman35 found support for this prediction. They invited participants to state the amount of money they were willing to bet on a set of 20 gambles they expected to play on that same day or two months later. The bets varied in probability of winning and expected value. As expected, payoffs predicted the bids participants placed on distant future bets more than bids on near future bets. Probabilities, in contrast, were a better predictor of near future bets than of distant future bets. These findings extend CLT to uncontrollable, random outcomes where the feasibility consideration cannot be overcome by increased effort. Arguments in favor and against an action. Much as feasibility is subordinate to desirability, in deciding whether to undertake an action cons are subordinate to pros. This is because the subjective importance of cons depends on whether or not pros are present more than the subjective importance of pros depends on whether or not cons are present. Consider, for example, a decision to take a medication. We would only inquire about potential side effects if we know that the medication has some health benefit; otherwise, we would decide against taking it without inquiring into side effects. In contrast, we would inquire about potential benefits of the medicine regardless of the existence of side effects; if it does not have side effects information about benefits will tell us if it worth taking, while if it does have some side effects information about benefits will tell us if they outweigh the side effects. After establishing these subordination relations in a series of studies, Eyal, Liberman, Trope and Walther36 examined the implication that follows from CLT: If cons are subordinate to pros, then pros should become more salient as temporal distance from the action increases, whereas cons should become less salient as temporal distance from the action increases. Participants generated arguments in favor and against new (i.e. non-routine) near future or distant future actions. As predicted, participants generated relatively more pro arguments and fewer con arguments when the actions were to take place in the more distant future. The proposed action involved new exam procedures (e.g. switching to open ended questions instead of multiple choice questions; Study 2), social policies (e.g. restricting private cars in the city center; Study 3), and a variety of personal and interpersonal behaviors (e.g. approaching a fellow student and offering to write an assignment together, Studies 4–6). In all the studies, participants generated more pros and less cons as temporal distance from the actions increased.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

183

Predicting behavioral intentions from attitudes and values. Earlier in this chapter we argued that values and general attitudes are part of high-level construals and therefore more likely to be evoked to describe a distant future situation. This would imply that values and general attitudes should be more readily applied to and guide choice in situations that are psychologically distant. Sagristano, Eyal, Trope, and Liberman37 conducted a two-stage study to examine this prediction. In the first experimental session, participants indicated their general attitudes toward blood donation, volunteering for psychology experiments, and physical fitness advisement. In a second, purportedly unrelated session, participants were offered an opportunity to engage in those activities either in the next two days or several weeks later, and their behavioral intentions were assessed. As expected, participants’ general attitudes better predicted their intention for the distant future than for the near future. Another set of studies examining the formation of value-consistent intentions used Schwartz’s38 value questionnaire to assess the importance participants assign to a wide range of values (e.g. power, benevolence, hedonism). For example, one study asked participants to imagine 30 behaviors (e.g. rest as much as I can) and to indicate the likelihood of performing each behavior either in the near future or in the distant future. The researchers then correlated the rated importance of each value and the mean likelihood for performing the behaviors corresponding to that value. As predicted, these correlations were higher when the behaviors were planned for the distant future than when they were planned for the near future.37 It is also possible to distinguish between values that are central to an individual, and more peripheral, secondary values. When a situation is related to a number of different values, the individual’s central values are more likely to guide choice from a psychologically distant than proximal perspective, whereas the individual’s secondary values are more likely to guide their choice from the psychologically proximal than distant perspective. Investigating this issue, Eyal, Liberman, Sagristano and Trope39 measured or manipulated the centrality of values and examined how they predict behavioral intentions. For example, one study assessed the relative centrality of achievement versus altruism values, and examined near and distant intentions of solving a dilemma between getting ahead by working extra hours or helping a friend. Results indicated that people who were predominantly achievement oriented planned to be achieving in the distant future more than in the near future whereas people who were predominantly altruistic, planned to be more cooperative in the distant future than in the near future. In other words, participants solved the conflict in favor

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

184

C. J. Wakslak, Y. Trope & N. Liberman

of their personally more central value in the distant future more than in the near future. These results imply that distant future decisions reflect predominant values whereas in near future decisions secondary values are also brought into consideration. Primary and secondary concerns within an interpersonal context. As with evaluating a single product with multiple features, issues within the context of a single interpersonal negotiation can differ in their centrality and worth. From a CLT perspective, negotiators would be expected to focus more on central concerns and less on peripheral concerns as psychological distance increases. To investigate this idea, Henderson, Trope and Carnevale40 examined behavior within a live negotiation. They found that while 91% of dyads with a temporally distant perspective reached a fully logrolling agreement in which lowest and highest priority issues were fully traded off, only 50% of dyads with a temporally near perspective did so. Results also provided evidence that participants approached the negotiation in a more global, structured manner when it was distanced; participants in the distant future condition were more likely to make integrative, multi-issue offers, than those in the near future condition. Finally, this effect resulted in an increase both in individual and joint outcome for the distant as opposed to near condition participants. In summary, the studies reported here support the CLT claim that high-level information will receive more weight in decisions about distant future events and low-level information will receive more weight in decisions about near future events. This occurred for primary and secondary features of products, desirability and feasibility aspects of assignments and feedback, payoffs and probabilities of gambles, pros and cons of action alternatives, central and peripheral values, and high and low priority issues in a negotiation context. The effects occurred despite the equal information available for the near and distant future options and despite the equally irreversible nature of the decision in many of the situations. Further, the studies do not show that participants are simply uncertain or indifferent about their choices for the future. Rather, choices for the future seem to discriminate more clearly between alternatives. We believe this is the case because they are based upon higher level of construal, which are more structured and schematic than construals of near future events. In the final section, having established the utility of the construal level approach to temporal distance, we situate this work within a more general theory of psychological distance.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

185

Temporal Distance as a Dimension of Psychological Distance In the current chapter, we have argued that because less information is typically known about temporally distant events than temporally proximal events, an association forms between distance and level of abstraction. This association is then over-generalized to situations where equal information is available, leading to the variety of construal effects we have documented. According to the logic of this account, temporal distance should not be alone in this association. As an event becomes farther removed from direct experience on any one of a number of dimensions, it should be represented in a more highlevel manner. Recent research on a variety of distance dimensions, including spatial distance,41 social distance,42 and hypotheticality43 has supported this claim, finding that the more spatially distant, socially distant, or improbable an event, the more it is represented in an abstract, superordinate manner. Further, using an implicit association test paradigm, Bar-Anan, Liberman and Trope44 found evidence for implicit associations between each of these distance dimensions and level of construal (i.e. shorter response times when distance related words were paired with high level construal stimuli and proximal related words were paired with low level construal stimuli than with incongruent pairings). It may thus be useful to conceptualize various forms of distance dimensions within a unified theory of psychological distance wherein similar principles of construal apply across the different dimensions and the formation of abstract construals is involved in transcending the proximal on all of these dimensions. One implication of this framework would be that the distance dimensions themselves should be interrelated. In fact, recent research has found evidence for this at both the explicit and implicit level. In a series of studies using politeness as an indicator of social distance,45 Stephan46 and Reichman and Ben Arie47 found that both temporal distance (writing instructions for a person expected to read them in either the near or distant future) and spatial distance (writing notes for a person in a class in another city or classroom versus the same classroom) produced a corresponding increase in social distance. Further, communication characterized by socially distant language was expected to be enacted at a later time than communication characterized by socially close language; likewise, two speakers speaking in a socially distant manner were perceived to be at a greater physical distance from one another than speakers communicating in a socially close manner.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

186

C. J. Wakslak, Y. Trope & N. Liberman

Evidence for the automatic association of distance dimensions has been found as well. Bar-Anan, Liberman, Trope, and Algom48 used a picture-word version of the Stroop paradigm49 in which participants discriminated between cues of one psychological distance dimension while ignoring cues of another psychological distance dimension. They reasoned that if psychological distance is a shared meaning of various distance dimensions it would be easier to perform the task when the relevant and the irrelevant cues were congruent in psychological distance, than when the relevant and irrelevant cues were incongruent in terms of psychological distance. Using perspective pictures, Bar-Anan et al. placed an arrow on the picture pointing either to a proximal or a distal point on the landscape shown in the picture, and a word denoting a psychologically proximal entity (“tomorrow,” “friend,” “we” or “sure”) or a psychologically distal entity (“year,” “enemy,” “others” or “maybe”) was printed on the arrow. In some of the experiments, the task was spatial discrimination, namely, participants indicated whether the arrow pointed to a spatially proximal or distal location. In other experiments, the task was semantic discrimination, namely, participants indicated whether the word on the arrow was, for example, “we” or “others.” In both types of tasks, and across all four dimensions of distance, participants were faster in responding to distance-congruent than to distance-incongruent stimuli. These results demonstrate that dimensions of psychological distance not directly relevant to the current task are activated, and that these various distances share a common aspect of meaning.

Conclusion According to Construal Level Theory, distant future events are represented in an abstract, schematic manner that emphasizes central and superordinate features (high-level construals), whereas near future events are represented in a concrete, less schematic manner, that includes incidental and subordinate features (low-level construals). Research on the mental representation of near and distant future events supports these assumptions, and research on future prediction, preferences and behaviors shows the range of implications that a construal framework entails. Further, research on other psychological distance dimensions, including spatial distance, social distance, and hypotheticality, indicates that high level construals are associated with psychological distance on any of these dimensions. Thus, while not denying the uniqueness of each of these dimensions, the current analysis suggests that they have something important in

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

187

common. Conceptualized as dimensions of psychological distance, each related to mental construal, they create a unifying theoretical framework for understanding a range of seemingly unrelated psychological phenomena, a framework that may allow us to capture a fundamental aspect of meaning.

References 1. Trope Y, Liberman N. Temporal construal. Psychological Review 2003; 110(3): 403–421. 2. Fiske S T, Taylor S E. Social Cognition. 2nd edn. New York, NY: McGraw-Hill, 1991. 3. Smith E R. Mental representation and memory. In Gilbert D T, Fiske S T, Gardner L eds. The Handbook of Social Psychology. 4th edn. New York, NY: McGraw-Hill, 1998: 391–445. 4. Rosch E, Lloyd B B, eds. Cognition and Categorization. Oxford, England: Lawrence Erlbaum, 1978. 5. Hampson S E, John O P, Goldberg L R. Category breadth and hierarchical structure in personality: Studies of asymmetries in judgments of trait implications. Journal of Personality and Social Psychology 1986; 51(1): 37–54. 6. Carver C S, Scheier M F. Themes and issues in the self-regulation of behavior. In Wyer R S, Jr., ed. Perspectives on Behavioral Self-regulation: Advances in Social Cognition, Vol. XII. Mahwah, NJ: Erlbaum, 1999: 1–105. 7. Miller G A, Galanter E, Pribram K H. Plans and the Structure of Behavior. New York, NY: Adams-Bennister-Cox, 1986. 8. Vallacher R R, Wegner D M. What do people think they’re doing? Action identification and human behavior. Psychological Review 1987; 94(1): 3–15. 9. Liberman N, Sagristano M D, Trope Y. The effect of temporal distance on level of mental construal. Journal of Experimental Social Psychology 2002; 38(6): 523–534. 10. Liberman N, Trope Y. The role of feasibility and desirability considerations in near and distant future decisions: A test of temporal construal theory. Journal of Personality and Social Psychology 1998; 75(1): 5–18. 11. Vallacher R R, Wegner D M. Levels of personal agency: Individual variation in action identification. Journal of Personality and Social Psychology 1989; 57(4): 660–671. 12. Eyal T. Time-dependent changes in values. Unpublished manuscript. Tel Aviv University, 2004. 13. Day S B B, D. M. Temporal Distance, Event Representation, and Similarity. 26th Annual Meeting of the Cognitive Science Society (CogSci2004). 2004, Hillsdale, NJ. 14. Fiedler K, Semin G R, Finkenauer C, Berkel I. Actor-observer bias in close relationships: The role of self-knowledge and self-related language. Personality and Social Psychology Bulletin 1995; 21(5): 525–538. 15. Semin G R, Fiedler K. The cognitive functions of linguistic categories in describing persons: Social cognition and language. Journal of Personality and Social Psychology 1988; 54: 558–568. 16. Semin G R, Smith E R. Revisiting the past and back to the future: Memory systems and the linguistic representation of social events. Journal of Personality and Social Psychology 1999; 76(6): 877–892.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

188

C. J. Wakslak, Y. Trope & N. Liberman

17. Gilbert D T, Malone P S. The correspondence bias. Psychological Bulletin 1995; 117(1): 21–38. 18. Jones E E. The rocky road from acts to dispositions. American Psychologist 1979; 34(2): 107–117. 19. Nussbaum S, Trope Y, Liberman N. Creeping dispositionism: The temporal dynamics of behavior prediction. Journal of Personality and Social Psychology 2003; 84(3): 485–497. 20. Jones E E, Harris V A. The attribution of attitudes. Journal of Experimental Social Psychology 1967; 3: 1–24. 21. Förster J, Friedman R S, Liberman N. Temporal construal effects on abstract and concrete thinking: Consequences for insight and creative cognition. Journal of Personality and Social Psychology 2004; 87(2): 177–189. 22. Ekstrom R B, French J W, Harman H H, Dermen D. Manual for Kit of Factor-Referenced Cognitive Tests. Princeton, NJ: Educational Testing Service, 1976. 23. Liberman N, Macrae S, Sherman S, Trope Y. The Effect of Level of Construal on Temporal Distance. Unpublished manuscript. Tel Aviv University, 2004. 24. Buehler R, Griffin D, Ross M. Exploring the “planning fallacy”: Why people underestimate their task completion times. Journal of Personality and Social Psychology 1994; 67(3): 366–381. 25. Gilbert D T, Wilson T D. Miswanting: Some problems in the forecasting of future affective states. In Forgas J P, ed. Feeling and Thinking: The Role of Affect in Social Cognition. Studies in Emotion and Social Interaction, second series. New York, NY: Cambridge University Press, 2000: 178–197. 26. Griffin D W, Dunning D, Ross L. The role of construal processes in overconfident predictions about the self and others. Journal of Personality and Social Psychology 1990; 59(6): 1128–1139. 27. Wilson T D, Wheatley T, Meyers J M, Gilbert D T, Axsom D. Focalism: A source of durability bias in affective forecasting. Journal of Personality and Social Psychology 2000; 78(5): 821–836. 28. Nussbaum S, Liberman N, Trope Y. Predicting the near and distant future. Unpublished manuscript. Tel Aviv University, 2005. 29. Ainslie G, Haslam N. Hyperbolic discounting. In Elster G L J, ed. Choice Over Time. New York: Russell Sage Foundation, 1992: 57–92. 30. Elster J, Loewenstein G. Utility from memory and anticipation. In Elster G L J, ed. Choice Over Time. New York: Russell Sage Foundation, 1992: 213–234. 31. Mischel W, Shoda Y, Rodriguez M L. Delay of gratification in children. Science 1989; 244: 933–938. 32. Read D, Loewenstein G. Time and decision: Introduction to the special issue. Journal of Behavioral Decision Making 2000; 13: 141–144. 33. Trope Y, Liberman N. Temporal construal and time-dependent changes in preference. Journal of Personality and Social Psychology 2000; 79(6): 876–889. 34. Freitas A L, Salovey P, Liberman N. Abstract and concrete self-evaluative goals. Journal of Personality and Social Psychology 2001; 80(3): 410–424.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

Transcending the Now: Time as a Dimension of Psychological Distance

189

35. Sagristano M D, Trope Y, Liberman N. Time-dependent gambling: Odds now, money later. Journal of Experimental Psychology: General 2002; 131(3): 364–376. 36. Eyal T, Liberman N, Trope Y, Walther E. The pros and cons of temporally near and distant action. Journal of Personality and Social Psychology 2004; 86(6): 781–795. 37. Sagristano M D, Trope Y, Eyal T, Liberman N. How temporal distance affects attitudebehavior correspondence. Unpublished manuscript. Florida Atlantic University, 2004. 38. Schwartz S H. Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In Zanna M P, ed. Advances in Experimental Social Psychology, Vol. 25. New York, NY: Academic Press, 1992: 1–65. 39. Eyal T, Liberman N, Sagristano M D, Trope Y. Resolving value conflicts in planning the future. Unpublished manuscript. Tel Aviv University, 2005. 40. Henderson M D, Trope Y, Carnevale P J. Negotiation from a near and distant time perspective. Unpublished manuscript. New York University, 2005. 41. Fujita K F, Henderson M D, Eng J, Trope Y, Liberman N. Spatial distance and mental construal of social events. Psychological Science in press. 42. Smith P, Trope Y. You focus on the forest when you’re in charge of the trees: The effect of power priming on information processing. Journal of Personality and Social Psychology in press. 43. Wakslak C J, Trope Y, Liberman N, Alony R. Seeing the forest when entry is unlikely: Probability and the mental representation of events. Unpublished manuscript. New York University, 2005. 44. Bar-Anan Y. Automatic associations between dimensions of psychological distance. Unpublished manuscript. Tel Aviv University, 2004. 45. Brown P, Levinson S C. Politeness: Some Universals in Language Usage. Cambridge, England: Cambridge University Press, 1987. 46. Stephan E. Social distance and its relation to level of construal, temporal distance and physical distance. Unpublished manuscript. Tel Aviv University, 2004. 47. Reichman N, Ben Arie Y. The effect of spatial distance on politeness: Evidence for the effect of spatial distance on social distance. Unpublished manuscript. Tel Aviv University, 2004. 48. Bar-Anan Y, Liberman N, Trope Y, Algom. The automatic processing of psychological distance cues: Evidence from a stroop task. Unpublished manuscript. Tel Aviv University, 2005. 49. Stroop J R. Studies of interference in serial verbal reactions. Journal of Experimental Psychology: General 1935; 18: 643–662.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch07

This page intentionally left blank

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

8 Time Monitoring and Executive Functioning: Individual and Developmental Differences Timo Mäntylä∗ and Maria-Grazia Carelli†

Time Monitoring and Cognitive Control in Children and Adults Most goal-directed activities require temporal integration and monitoring of action sequences.1 For example, memory for future intentions involves a monitoring phase during which the individual has to pay attention to the target event among other events (e.g. “take the pill every six hours”). In general terms, monitoring is the process by which agents assess their environments, and involves activities such as checking the progress of initiated plans, finding out what time it is, and anticipating obstacles.1,2 In this chapter, we summarize our own and others’ work on time monitoring in human agents. Our primary aim was to investigate how children and adults check the progress of an initiated plan while being involved in a variety of ongoing activities. Our focus is on monitoring strategies in the context of timebased prospective memory (ProM) tasks, rather than on tasks of vigilance or goal-directed problem solving. We will also examine time monitoring in relation

∗

Department of Psychology, Umeå University, 90187 Umeå, Sweden; e-mail: [email protected] of Psychology, Umeå University, 90187 Umeå, Sweden.

† Department

191

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

192

T. Mäntylä & M.-G. Carelli

to individual and developmental differences in prefrontally-mediated executive control functions. According to this view, individuals with problems in executive control functions use less efficient monitoring strategies than individuals with better functioning control mechanisms.

Characteristics of Strategic Monitoring Efficient monitoring requires a strategy, or a scheme, for scheduling actions (i.e. when and how to monitor). However, in most everyday situations, this strategy must balance the cost of monitoring against the cost of having inaccurate information about the environment. In general terms, monitoring can be considered as a simple decision task in which the individual (or agent) has to balance the cost of not having inaccurate information about the environment and the cost of monitoring actions. However, deciding between these costs can be a complicated optimization problem. Atkin and Cohen3,4 categorized monitoring strategies on the basis of general features of agent architectures, tasks, and environments. Atkin and Cohen concluded that most monitoring strategies can be divided into two general categories, namely, periodic monitoring and interval reduction. They also proved analytically that interval (proportional) reduction outperforms periodic monitoring in most conditions: “Only when an agent has no goal, or monitoring provides no information about progress towards this goal, should periodic monitoring be chosen” (Atkin & Cohen,3 p. 22). Atkin and Cohen also suggested that a combination of the two approaches might be an optimal strategy in many situations: “Use proportional reduction to get close to the goal, and then periodic for the final few steps” (p. 32). Although monitoring is a necessary task for any agent, including humans, insects and robots, few studies have investigated how these agents actually behave while the goal or deadline is approaching. Yet, the monitoring concept has close connections to several areas, including operant conditioning,5 process control,6,7 and some areas of the AI domain. For example, early work on reinforcement schedules is closely related to monitoring problems. In the fixed interval (FI) schedules, the subject receives reinforcement for its first response occurring at more than T time units since its previous reinforcement. A typical finding is that the subject waits a certain period of time T before responding again, and then accelerates responding as the T approaches. This fixed interval scallop is very similar to an interval reduction strategy discussed above.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

193

Although the monitoring concept can be related to these and other areas (e.g. optimal foraging behavior in ethology, see Pyke,8 for a review), research on time-based ProM is informative for understanding the underlying mechanisms of strategic monitoring. Studies on time-based ProM tasks suggest certain regularities of behavior prior to successful performance of a task. After being instructed to perform a specific action (e.g. reminding the experimenter after 20 min), most participants first test the time by looking at a clock, and then wait for a period of time until another clock checking appears appropriate. These check-wait cycles are repeated until the critical period to respond arrives, at which point they perform the action and then stop monitoring. In the following section of the chapter, we will take a closer look at studies examining how people monitor future-oriented intentions while being involved in a variety of activities.

Time Monitoring Children and Adults Based on the Test-Operate-Test-Exit (TOTE) framework by Miller, Gallanter, and Pribram,9 Harris and Wilkins10 proposed a specific model for strategic monitoring in time-based ProM tasks. In their Test-Wait-Test-Exit (TWTE) model, monitoring is assumed to involve a series of test-wait cycles until a final test is made during a critical period. To test this model, Harris and Wilkins10 asked participants to hold up a series of cards after 3 or 9 min, while they were watching a movie. For example, if the cards indicated 3 min, 9 min, 3 min, respectively, participants were to wait 3 min and hold up the first card, wait 9 min and hold up the second card, and so on. Participants monitored the time by turning to look at the clock on the wall behind them, and the experimenter recorded the number of clock checks. Harris and Wilkins found that monitoring of the clock was closely related to the latency of responding, with shorter latencies associated with a greater rate of monitoring, especially during the period immediately preceding the target time. Consistent with the TWTE model, the overall pattern of clock checking frequency showed a J -shaped function. According to Harris and Wilkins, participants first checked the clock frequently to synchronize their internal clocks with the external clock. During the middle of the period they could rely on their internal clocks to keep track of the time. Finally, as the response time approached, they relied frequently on the external clock to ensure they were not late in performing the ProM task. Ceci and Bronfenbrenner11 conducted a seminal study of time-based ProM in school-aged children. In their study, children ages 10 and 14 years were instructed to remove cupcakes from the oven in exactly 30 min to avoid burning

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

194

T. Mäntylä & M.-G. Carelli

them. In another condition, the children charged a battery and were instructed to turn off the charger after 30 min to prevent overcharging. During the 30 min interval, the children played a video game in a separate room (either at home or in a laboratory setting). The clock was placed behind the child, so the experimenter (sibling) could easily see when the child turned around to determine how much cooking or charging time remained. However, this checking was associated with a cost in that the act of monitoring was a distraction from the game. Ceci and Bronfenbrenner11 found that all children checked the clock frequently during the first 10 minutes of the waiting period and then engaged in very little clock checking until the final moments of the waiting period. Specifically, older children in both settings and younger children in the home setting reduced the frequency of monitoring actions during the middle period (from 10 to 25 min) of the task interval. When younger children were tested in the unfamiliar laboratory setting (and with unknown experimenter) they maintained the frequency of clock checking at the same high level also during the mid-phase of the task. Kerns12 extended the findings of Ceci and Bronfenbrenner11 by investigating the development of ProM across a wider age range. Children ages 6–12 years played a computer game (“Cyber Cruiser”) that involved driving a vehicle on a road. In addition to the primary task of driving (and earning points when not hitting other vehicles), the children were instructed to monitor the level of available fuel. Using two different buttons on the joystick, they could check the fuel level and refuel when the tank was less than 1/4 full. The duration of the game was 5 min, with the car running out of gas after one minute of play without filling. If participants ran out of gas, the gas gauge was automatically refilled and the game re-started, with zero points. Kerns12 found age-related differences in ProM performance in that younger children ran out of gas more frequently than older children. Concerning monitoring behavior both younger and older children showed similar patterns of gas checks. Specifically, children demonstrated a J -shaped, rather than U shaped, distribution of monitoring actions (see also Cayenberghs, De Bruycker & d‘Ydewalle,13 for similar findings with younger adults). The studies of Ceci and Bronfenbrenner11 and Kerns12 suggest that even young school-aged children monitor deadlines strategically in that they accelerated clock checking when the deadline is approaching. Our own work on time monitoring in children and adults is consistent with this general pattern. Mäntylä, Carelli and Forman14 instructed school-aged children and young adults to indicate the passing of time while watching a video. Time monitoring was

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

195

based on a computerized task in which participants were instructed to press a designated response button every 5 minutes. The experimenter explained that the button should be pressed when the clock on the computer monitor showed 05:00, 10:00, 15:00 and so on, without being informed about the length of the video. The experimenter also clarified that the (red) button should be used only for indicating the passing of every 5 min, but that the green button on the response box could be used to check the clock any time during the waiting period. As expected, there were clear age differences in monitoring frequency and monitoring accuracy. Compared with adults, children checked the clock more frequently and produced more late responses (i.e. ProM failures). This finding is hardly surprising considering the age range, but the same pattern of results was observed when the children’s monitoring data was analyzed separately. Furthermore, adults showed a negative correlation between monitoring frequency and accuracy, whereas children’s clock checking frequency was not related to accuracy. In other words, children who checked the clock frequently were not more likely to remember on time than children with lower monitoring frequency. Concerning the overall pattern of clock checking frequency, both age groups showed signs of strategic monitoring in that clock checking increased until the critical time occurred. Figure 1 illustrates the monitoring data for younger and older children. Consistent with earlier work, both age groups (and adults) increased clock checking when the 5-minute target time was approaching.

4

Mean Frequency

8-9 years 10-11 years

3

2

1

0 1

2

3

4

5

6

7

8

9

10

Task Duration (in min)

Fig. 1. Mean number of clock checks for younger and older children in Mäntylä et al.14

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

196

T. Mäntylä & M.-G. Carelli

However, inconsistent with the studies of Ceci and Bronfenbrenner11 and Kerns,12 both children and adults showed relatively flat monitoring functions during the early and middle phases of each 5-minute task interval (including the first interval). We also examined individual and developmental differences in time monitoring in relation to time estimation. Specifically, following the notion that the U -shaped monitoring function in Ceci and Bronfenbrenner’s study11 reflected developmental differences in the sense of time, a positive correlation between early clock checking and time estimation error would be expected. To test this notion, participants also completed a time reproduction task. In this task, a “smiley” appeared on a computer screen for a duration varying between 4 to 32 seconds, and participants reproduced the corresponding stimulus duration by pressing a designated key on the computer keyboard. Consistent with earlier work, timing error (i.e. absolute discrepancies) increased as a function of target duration for both age groups. However, there were no reliable age effects in that both groups produced comparable timing errors, measured in terms of absolute and relative discrepancies. Furthermore, time reproduction performance was not related to monitoring frequency or accuracy. In other words, participants who made great errors in the reproduction task (the mean absolute discrepancies varying between 500 ms and 2500 ms) did not show higher rates of clock checking during the first 5-minute interval than individuals with more accurate time estimation performance. Similarly, monitoring accuracy (i.e. time-based ProM performance) was not related to timing error. We also examined these data separately for short (≤8 s) and long (≥12 s) stimulus durations, but both analyses showed non-significant correlations for both age groups. Taken together, these studies suggest that school-aged children and younger adults use similar monitoring strategies in that the rate of monitoring is increased as the deadline approaches (i.e. interval reduction). However, the pattern is less clear when it comes to children’s monitoring behavior during the early stages of a task interval. The few existing studies show both U -shaped11,15 and J -shaped12 monitoring functions, and our findings suggest rather linearly increasing clock checking behavior for children and adults. Although our time estimation data should be interpreted with caution, they suggest that individual differences in duration judgments were not related to monitoring frequency or response accuracy.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

197

Time Monitoring in Younger and Older Adults The overall shape of a monitoring function is probably mediated by a number of factors, and it is reasonable to assume that both task-specific factors and individual difference variables contribute to the choice of a monitoring strategy. In this section, we examine younger and older adults’ monitoring behavior in the context of time-based ProM tasks. Taken together, these studies show agerelated differences in monitoring frequency and response accuracy. Typically, older adults are less likely to remember on time than younger participants, and these effects are closely related differences in monitoring behavior during the period immediately preceding the target time. Einstein et al.16 (see also Einstein & McDaniel17 ) examined event-based and time-based ProM performance in young adults, middle-aged adults, and older adults. In Experiment 1 of their study, participants were instructed to press a designated key after 5 minutes and 10 minutes had elapsed, while completing a continuous memory task. Einstein et al.16 observed age differences in ProM performance in that younger participants were more likely to remember on time than older participants. Furthermore, the number of monitoring responses was somewhat higher for younger participants relative to older participants, and this effect was accentuated during the last period. Specifically, younger participants checked the clock more frequently than older adults only during the last or proximal 2-minute period. In spite of age differences in monitoring frequency and response accuracy, the two groups showed similar monitoring functions. That is, the first three quartiles showed a flat or weakly increasing monitoring, followed by increased clock checking during the final quartile. Thus, inconsistent with the findings of Wilkins and Harris10 and Ceci and Bronfenbrenner,11 these data did not show high levels of early monitoring. Park et al.18 examined event-based and time-based ProM performance in younger and older adults. In addition to ProM performance, they examined clock-checking behavior during a 6-minute and 12-minute period. In their study, participants were instructed to indicate the passing of six times (every 2 minutes) or 12 times (every minute) by pressing a response button, while completing a working memory task. In the control condition, the ProM task was completed in the absence of working memory load. One of their findings was that the rate of clock checking increased during the last quarter of the monitoring interval (i.e. 15 s or 30 s), and that younger adults checked the clock more frequently than older adults did. Furthermore, these effects interacted with task load, so that younger adults in the no-load condition, as well as older adults in both control

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

198

T. Mäntylä & M.-G. Carelli

and memory load conditions, checked the clock relatively infrequently. Thus, monitoring in these conditions was rather periodic, with only a slight increase in clock-checking frequency. However, young adults in the memory load condition checked the clock more often from the second through fourth quartile than older adults, with a particularly steep increase in the time period immediately preceding the response window (i.e. a linearly increasing monitoring function). According to Park et al.18 young participants made more clock checks when they were also performing the primary task in order to maintain adequate time-based responding. Both Einstein et al.16 and Park et al.18 observed age deficits in time-based ProM performance, and these effects were attributed to reduced clock-checking behavior in older participants. That is, younger adults monitor more frequently than older adults (who have fewer resources for self-initiated activities), and as a result they are more likely to remember on time than older adults. However, the findings of Maylor, Smith, DellaSala & Logie19 and Logie, Maylor, DellaSala & Smith20 suggest a rather different pattern if the clock-checking behavior prior to success and failure is considered separately. Maylor et al.19 investigated the effects of normal aging and dementia on ProM performance. In Experiment 1 of their study, participants viewed a film for a later recognition memory task. They were instructed either to say “animal” when an animal appeared in the film (event-based ProM task) or to stop a clock every 3 minutes (time-based ProM task). In both tasks, young participants were more successful than older participants, who were, in turn, more successful than patients with Alzheimer’s disease (AD). In the time-based task, the frequency of clock checking was related to accuracy. For successful responses, older participants and AD patients checked the clock more often than young participants did, and clock checking was most frequent (and very similar across the three groups) in the final period of the 3-minute time interval (see also Logie et al.20 ). In contrast, unsuccessful responses showed flat monitoring functions even in the final period for both older adults and AD (all young adults were omitted because they never failed). According to Maylor et al.,19 this pattern is consistent with the view that momentary lapses of attention are primarily responsible for time-based ProM deficits in normal aging and dementia. Mäntylä and Carelli21 examined time monitoring across the adult life span. Younger adults, middle-aged adults and older adults were instructed to indicate the passage of time every 5 minutes, while listening to a 30-minute story (referred to as the low-load condition) or completing a series of cognitive tasks

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

199

(referred to as the high-load condition). The procedure was similar to that of Mäntylä et al.,14 except that participants were given more explicit monitoring instructions. Specifically, participants were informed that good performance in this task meant a combination of few clock checks while maintaining high accuracy (within 10 seconds of the target time). Figure 2a summarizes the monitoring data as a function of age group and task load. As can be seen, the rate of monitoring increased across age in the low-load condition, with older adults checking the clock more frequently than

(a)

2.50 No Load

Monitoring Frequency

Load 2.00

1.50

1.00

0.50

Young

Middle age

Old

(b) Monitoring Accuracy (Delay < 10s)

0.80

0.60

0.40

0.20

0.00

Young

Middle age

Old

Fig. 2. Monitoring frequency (a) and monitoring accuracy (b) as a function of age and task load in Mäntylä & Carelli.21

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

200

T. Mäntylä & M.-G. Carelli

younger adults. However, when the task demands were increased in the highload condition, the older participants also reduced their clock checking. This change in older adults’ monitoring strategy had clear effects on response accuracy (i.e. time-based ProM performance). As shown in Fig. 2b, the low-load condition showed minimal age differences in monitoring accuracy (i.e. proportion correct responses within 10 s). Thus, older participants obtained the same level of ProM performance as younger participants by checking the clock more frequently. As in the studies of Maylor et al.19 and Logie et al.20 (in which the clock-checking behavior prior to success and failure was considered separately), the low-load condition produced similar levels of ProM performance for all age groups, but older participants responded on time by checking the clock more often than younger participants. However, when the task demands were increased in the high-load condition, older participants were not able to check the clock frequently, which led to more late responses and age-related differences in ProM performance. Apart from age-related differences in overall monitoring frequency (in the low-load condition) and response accuracy (in the high-load condition), the three age groups showed similar monitoring functions. Figure 3 illustrates young and older adults clock checking behavior across the six 5-minute task intervals (middle-aged adults are not included in the figure, but they showed a very similar pattern). As can be seen in Fig. 3, both age groups made very few clock checks during the first two minutes of each 5-minute interval. This initial phase

1.25 Old Young

Mean Frequency

1.00 0.75 0.50 0.25 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Task Duration (in min)

Fig. 3. Mean number of clock checks for younger and older adults in Mäntylä & Carelli.21

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

201

was then followed by a linearly increasing clock checking rate for both age groups. It should be noted that clock checking was again minimized immediately after each 5-minute response, producing a sawtooth-like monitoring function. This apparent regularity of clock checking across a series of 5-minute intervals might suggest that participants synchronized their monitoring behavior to the overall time structure of the task. Because the participants were required to respond at equally spaced times (i.e. every 5 minutes), they somehow coordinated their clock checking with this predictable monitoring function. This regularity of monitoring is consistent with the dynamical attention theory (DAT) of Jones and colleagues.22,23 According to the DAT theory, the structure of world events offers temporal structures to which attenders can lock into, and this synchronization (or “entrainment”) in turn facilitates one’s ability to track changing events and to anticipate their course in real time (see also Chapter 3 in this volume). Taken together, the studies summarized here suggest rather similar monitoring behavior across the adult life span. In a variety of task conditions and intervals, most individuals monitor deadlines strategically in that they increase the rate of clock checking as the deadline approaches. Furthermore, most studies involving younger and older adults suggest that age-related differences in monitoring behavior are accentuated in the time period immediately preceding the response window. Although interval reduction is a common strategy among children and adults, a large body of evidence (see also Cicogna et al.24 ; Costermans & Desmette25 ) also shows a great deal of variability in monitoring behavior. This variability is not only age-related, but these differences can be observed within an age group. For example, our studies show large variability within a relatively homogenous group of university students, with some individuals showing very efficient monitoring performance, measured both in terms of monitoring behavior and time-based ProM performance. In the following section, we focus on one potential mechanism underlying individual and age-related differences in monitoring performance. Our primary focus is on the notion that monitoring performance is mediated by higher-order control functions, often referred to as executive functions or frontal lobe functions. According to this view, individuals with less efficient executive functions are expected to rely on less efficient monitoring strategies than individuals with better functioning control processes, because they have difficulties, for example, in inhibiting irrelevant thoughts or external distractors, maintaining and updating

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

202

T. Mäntylä & M.-G. Carelli

working memory contents, or shifting attention among different goal-directed activities.

Time Monitoring and Executive Control Functions The term executive function has a variety of interpretations, but most definitions refer to a collection of poorly defined cognitive activities such as planning, sequencing, monitoring, task switching, organization and inhibition (see e.g. Stuss & Knight26 ; Rabbitt27 ; Royal et al.28 for overviews). Although the concept of executive functioning is not easily defined and there is no available “gold standard” for measuring executive functioning,28– 30 there are several reasons to believe that time monitoring is closely related to executive functioning. First, most conceptions of monitoring emphasize its control nature. In robotic systems, a distinction is frequently made between sensing and monitoring. Sensing refers to the typically low-level data acquisition mechanisms needed to keep a world model up-to-date, and monitoring involves querying this world model. The distinction between sensing and monitoring is most common in artificial intelligence and robotics,31,32 but similar conceptualizations can be found in some models of metacognition33 and executive functioning34 (see also Fernandez-Duque et al.35 ). For example, in Nelson and Narens’ model,33 metacognitive regulation is considered as a metalevel system that modulates cognitive processes at the lower level. The metalevel contains a cognitive model of the object level, organized according to certain metacognitive principles. The metalevel is continuously updated by bottom-up information, and in return controls the object level by providing top-down input, initiating and terminating actions performed by the object level. Similarly, in Norman and Shallice’s model,34 the executive system modulates lower level schemas according to the subject’s intentions and goals, and in the absence of explicit control information is processed automatically by schemas. Second, most cognitive control functions, including planning, task initiation, updating and coordination are time-related in that they require compliance with temporal constraints.1,36 For example, Fuster1 proposed a general theory of prefrontal functioning in which temporal organization and integration of cognition and behavior plays a central role: “The enactment of a goal-directed sequence of actions is a continuous process of temporal integration. At the root of this process is the mediation of cross-temporal contingencies between the action plan, the goal, and the acts leading to the goal” (p. 96).

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

203

Consistent with this view, patient studies indicate that sense of time is closely related to higher cognitive control functions and their underlying neural mechanisms.37–40 For example, Barkley et al.37 asked school-aged children with attention deficit hyperactivity disorder (ADHD) to reproduce varying time durations. They found that, compared with healthy children, ADHD children’s reproductions were less accurate. These findings were consistent with Barkley’s41 model of ADHD and suggest that ADHD-related problems in inhibitory processes are related to biases in time perception (see also e.g. Capella et al.42 ; Meaux and Chelonis,43 for similar findings). It should also be noted that studies showing age-related differences in (timebased) ProM are consistent with the notion that executive control functions mediate strategic monitoring. Specifically, assuming that cognitive aging is associated with decrements in executive functioning, older adults (and children) would be expected to use less efficient monitoring strategies than younger adults. Past research on adult age differences in prospective memory, including the studies summarized in the previous section, is consistent with this notion, and shows age effects both in time-based and event-based tasks of ProM performance (i.e. monitoring accuracy). Furthermore, a number of studies provide direct evidence for the notion that ProM performance is mediated by executive functioning44– 49 (see also Salthouse et al.30 ). For example, in Kerns’ study,12 school-aged children completed four tasks of executive functioning, including two tasks of visuo-spatial working memory (delayed alternation-nonalternation and self-ordered pointing tasks) and two measures of inhibitory capacity (Stroop and go-no go tasks). Kerns12 hypothesized that accuracy on the ProM task (“Cyber Cruiser,” see also above) would be correlated with both measures of working memory and inhibitory control. The ProM measure (times out of gas) correlated significantly with three of the six executive function measures (the Stroop interference score, total number of errors on the self-ordered pointing task, and total number of errors on the nonalternation component of the delayed alternation task, respectively). Monitoring frequency in terms of number of gas checks correlated with one measure of executive functioning (omissions on the go-no go task), with more gas checks being related to a higher number of omission errors. However, when age was controlled for, there were no significant correlations between the number of checks and the measures of executive function, suggesting “that a failure of prospective memory (running out of gas), but not the frequency of checking per se, is related to other indicators of executive control” (Kerns,12 p. 68).

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

204

T. Mäntylä & M.-G. Carelli

Cayenberghs et al.31 examined time-based ProM performance in younger adults by using three complex tests of executive functioning (Wisconsin Card Card Sorting Task, Tower of London Sorting Task, and Controlled Oral Word, respectively) and three more basic tasks that were assumed to tap the inhibition, updating and shifting components of executive functions (see also Miyake et al.29 ). One of their findings was that inhibition (and to some extent shifting) was related to ProM performance (times out of gas), whereas the more complex tasks of executive functioning were not related to monitoring accuracy. Furthermore, none of the executive functioning measures, including inhibition, were related to monitoring frequency. These findings and those of Kerns12 suggest that also time-based ProM performance is related to individual differences in executive functioning in schoolaged children and younger adults. However, individual (and developmental) differences in executive functioning show only a weak relation or no relation to monitoring frequency. Thus, although past research suggests that executive functioning is related to (event-based) ProM performance and that monitoring frequency and accuracy are positively correlated in most (time-based) ProM tasks (summarized in the previous sections), the few existing studies suggest that monitoring frequency is not related to individual differences in executive functions. Our own studies on time-based ProM performance in school-aged children14 and younger and older adults21 provided more direct support for the notion that both monitoring frequency and accuracy are related to individual differences in some aspects of executive functioning. In the former study, executive functioning was based on three basic constructs of cognitive control, namely, mental set shifting, updating, and inhibition of prepotent responses, respectively (see also Miyake et al.29 ; Salthouse et al.30 ). Each construct was measured by two experimental tasks that were assumed to tap a specific target function (see Table 1). Separate exploratory factor analyses of the executive functioning data yielded a 2-factor solution for both age groups. As shown in Table 1, the updating and inhibition tasks constituted one factor and the two shifting tasks the second factor, with similar patterns for children and adults. These two factors were differentially related to monitoring performance, measured in terms of number of clock checks and late responses. Specifically, children with low performance in the inhibition and updating tasks (i.e. high Factor 1 scores) showed a higher rate of clock checking and more late responses than better functioning children.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

205

Table 1. Factor loadings for the principal factor analysis of the inhibition, updating, and shifting tasks in Mäntylä et al.14 Adults Factor 1

Children

Construct

Task

Factor 2

Factor 1

Factor 2

Inhibition

Stroop Stop Signal

0.73 0.74

0.38 −0.03

0.67 0.68

0.20 −0.30

Updating

Matrix Monitoring N-back

0.59 0.56

0.07 −0.17

0.71 0.71

−0.04 0.05

Shifting

Connections Category Fluency

−0.03 0.09

0.83 0.64

0.07 −0.07

0.77 0.76

Younger children (8–9 years) showed lower performance in most tasks of executive functioning than older children (10–12), but the correlation between the Factor 1 scores and monitoring frequency was significant even when the effect of age was statistically controlled. By contrast, individual differences in shifting (Factor 2) were not related to monitoring performance. The adult participants showed a similar pattern of results in that individuals with poor performance in the inhibition/updating tasks checked the clock more frequently than individuals with more efficient inhibition/updating functions. Monitoring accuracy was indirectly related to executive functioning in that frequency and accuracy were negatively correlated. Again, the shifting component was unrelated to monitoring frequency and accuracy. Mäntylä & Carelli21 extended these findings across the adult life span. In this study, young adults, middle-aged adults and older adults completed three tasks of response inhibition and three updating tasks. The younger participants were university students between 20–30 years, whereas the middle-aged (40–56 years) and older (64–81 years) participants were based on a population-based sample of adults. The six tasks of inhibition and updating were significantly correlated with each other, and factor analysis yielded a one-factor solution. As expected, executive functioning was significantly correlated with age, with older adults showing greater problems in executive functioning than younger adults. Furthermore, participants with low performance in the inhibition and updating tasks showed less efficient monitoring performance than individuals with better performance in the executive functioning tasks. Considering that age and executive functioning were highly correlated it is reasonable to expect that the observed relation between monitoring frequency

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

206

T. Mäntylä & M.-G. Carelli

and executive functioning was mediated by age. For older adults, the correlation between executive functioning and monitoring frequency was eliminated after controlling for age. However, younger and middle-aged adults showed a significant correlation between frequency and executive functioning even when age was taken into account. Taken together, these findings suggest that timebased ProM performance, both in terms of monitoring frequency and accuracy, is closely related to individual (and age-related) differences in executive control functions.

Conclusions The aim of this chapter was to examine strategic monitoring in the context of time-based ProM tasks. Taken together, the studies summarized here show regularities in time monitoring behavior in that both children and adults accentuate monitoring as the deadline approaches, and failures of monitoring during that critical period have direct consequences for response accuracy. This pattern suggests that interval reduction is a general and efficient monitoring strategy in a variety of task situations. These findings are also consistent with the Test-Wait-Test-Exit model of monitoring10 in that most participants seem to first test the time by looking at a clock, and then wait for a period of time until another clock checking appears appropriate. These check-wait cycles are repeated until the critical period to respond arrives, at which point they perform the action and then stop monitoring. However, as noted by Harris and Wilkins,10 the TWTE model is a descriptive account of monitoring behavior, and in most tasks of ProM (rather than vigilance) the waiting period between the intention and its goal is not a continuous sequence of test-wait cycles. Instead, monitoring occurs as a background task with one or more ongoing foreground tasks, and the individual has to balance the cost of interrupting the foreground task (e.g. playing the computer game in the studies of Ceci & Bronfenbrenner11; Kerns12 ) and having accurate information about the background task (e.g. gas level). From that perspective, an important question concerns the conditions that direct attention from its actual focus to postponed intentions. According to one prominent view of prospective memory, the cognitive system relatively automatically responds to the occurrence of target events in the environment.50– 52 This spontaneous retrieval theory proposes that people rely

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

207

on spontaneous memory-based and/or attentional processes to retrieve intentions when ProM targets are encountered. The primary assumption of this account is that participants do not necessarily monitor the environment for target events but that remembering occurs when the presence of the target event initiates successful retrieval processes. Thus, retrieval is considered as spontaneous in the sense that it can occur without the involvement of attentional (executive) resources. An alternative approach to understanding ProM retrieval is to assume that an executive attentional system explicitly monitors the environment for target events. According to this monitoring account, retrieval occurs through the capacity-demanding attentional process of monitoring the environment for the target events. When a target event is encountered, the executive attentional system interrupts the ongoing activity and initiates the processes necessary for performing the intended action. Smith53 presented a strong version of this view, and argued that “retrieval of an intention will never be automatic, because nonautomatic preparatory processes must be engaged during the performance interval, or the time in which the opportunity to carry out the action is likely to occur, but before the occurrence of the target event” (p. 349). Although a number of studies have shown that dividing attention during retrieval decreases event-based ProM performance,18,54,56 the issue concerning the nature of ProM retrieval is open to debate.50,54,57 Even if one assumes that a variety of cognitive processes can be recruited to support ProM retrieval,51 it is reasonable to argue that time-based tasks rely more heavily on self-initiated processes, whereas event-based tasks may rely on more spontaneous or automatic processing.58,59 In most time-based tasks of ProM, intentions are triggered by time-related cues that can be mediated by external (e.g. noticing a clock on the wall) or internal (e.g. time-related associations, internal clock) factors. However, compared to event-based ProM tasks, these cues are more “invisible,” and self-initiated thoughts and monitoring are critical to successful performance in most time-based tasks of prospective memory.

References 1. Fuster J. Frontal lobes. Current Opinion in Neurobiology 1993; 3: 160–165. 2. Fuster J. Physiology of executive functions: The perception-action cycle. In Stuss D T, Knight R T, eds. Principles of Frontal Lobe Function. New York: Oxford University Press, 2002.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

208

T. Mäntylä & M.-G. Carelli

3. Atkin M S, Cohen P R. Monitoring strategies for embedded agents: Experiments and analysis. Journal of Adaptive Behavior 1996; 4: 125–172. 4. Cohen P R, Atkin M S, Hansen E A. The interval reduction strategy for monitoring cupcake problems. Proceedings of the Third International Conference on the Simulation of Adaptive Behavior 1994: 82–90. 5. Ferster C B, Skinner B F. Schedules of Reinforcement. New York: Appleton-Century-Crofts, 1957. 6. Moray N. Monitoring behavior and supervisory control. In Boff K R, Kaufman L, Thomas J P, eds. Handbook of Perception and Human Performance, Vol. II, Chapter 40. New York: Wiley, 1986. 7. Senders J W. Visual Scanning Processes. Hillsdale, NJ: Erlbaum, 1983. 8. Pyke G H. Optimal foraging: A critical review. Annual Review of Ecology and Systematics 1984; 15: 523–575. 9. Miller G, Galanter E, Pribram K H. Plans and the Structure of Behavior. New York: Holt, Rinehart, Winston, 1960. 10. Harris J E, Wilkins A J. Remembering to do things: A theoretical framework and an illustrative experiment. Human Learning 1982; 1: 123–136. 11. Ceci S J, Bronfenbrenner U. “Don’t forget to take the cupcakes out of the oven”: Prospective memory, strategic time-monitoring, and context. Child Development 1985; 56: 152–164. 12. Kerns K A. The cybercruiser: An investigation of development of prospective memory in children. Journal of the International Neuropsychological Society 2000; 6: 62–70. 13. Cayenberghs K, DeBruycker W, Helsen L, d’Ydewalle G. The fractionation of executive functioning in prospective memory: The effect of task complexity. In The 2nd International Conference on Prospective Memory. Zurich, Switzerland, 2005. 14. Mäntylä T, Carelli M G, Forman H. Time control in children and adults. Manuscript submitted for publication, 2005. 15. Ceci S J, Baker J G, Bronfenbrenner U. Prospective remembering, temporal calibration, and context. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory: Current Research and Issues. London: Wiley, 1988: 360–365. 16. Einstein G O, McDaniel M A, Richardson S L, Guynn M J, Cunfer A R. Aging and prospective memory: Examining the influence of self-initiated retrieval processes. Journal of Experimental Psychology: Learning, Memory and Cognition 1995; 21: 996–1007. 17. Einstein G O, McDaniel M A. Normal aging and prospective memory. Journal of Experimental Psychology: Learning, Memory and Cognition 1990; 16: 717–726. 18. Park D C, Hertzog C, Kidder D P, Morrell R W, Mayhorn C B. Effect of age on event-based and time-based prospective memory. Psychology and Aging 1997; 12: 314–327. 19. Maylor E A, Smith G, DellaSala S, Logie R H. Prospective and retrospective memory in normal aging and dementia: An experimental study. Memory and Cognition 2002; 20: 871–884. 20. Logie R H, Maylor E A, Della Sala S, Smith G. Working memory in event- and timebased prospective memory tasks: Effects of secondary demand and age. European Journal of Cognitive Psychology 2004; 16: 441–456.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

209

21. Mäntylä T, Carelli M G. Time monitoring and cognitive control in young and old adults. In 2nd International Conference on Prospective Memory, Zurich, Switzerland, 2005. 22. Jones M R, Boltz M. Dynamic attending and responses to time. Psychological Review 1989; 96: 459–491. 23. Large E W, Jones M R. The dynamics of attending: How we track time-varying events. Psychological Review 1999; 106: 119–159. 24. Cicogna P C, Nigro G, Occhionero M, Ésposito M J. Time-based prospective remembering: Interference and facilitation in a dual task. European Journal of Cognitive Psychology 2005; 17: 221–240. 25. Costermans J, Desmette D. A method for describing time monitoring strategies in a prospective memory setting. Current Psychology of Cognition 1999; 18: 289–306. 26. Stuss D T, Knight R T, eds. Principles of Frontal Lobe Function. New York: Oxford University Press, 2002. 27. Rabbitt P. Methodology of Frontal and Executive Function. Hove, UK: Psychology Press, 1997. 28. Royall D R, Lauterbach E C, Cummings J L, Reeve A, Rummans T A, Kaufer D I, LaFrance Jr W, Coffey C E. Executive control function: A review of its promise and challenges for clinical research. Journal of Neuropsychiatry and Clinical Neuroscience 2002; 14: 377–405. 29. Miyake A, Friedman N P, Emerson M J, Witzki A H, Howerter A, Wager T D. The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology 2000; 41: 49–100. 30. Salthouse T A, Atkinson T M, Berish D E. Executive functioning as a potential mediator of age-related cognitive decline in normal adults. Journal of Experimental Psychology: General 2003; 132: 566–594. 31. Firby R J. An investigation into reactive planning in complex domains. In Proceedings of the Sixth National Conference on Artificial Intelligence 1987: 202–206. 32. McDermott D. Robot planning. AI Magazine 1992; 13: 55–79. 33. Nelson T O, Narens L. Metamemory: A theoretical framework and new findings. In Bower G, ed. The Psychology of Learning and Motivation, Vol. 26. New York: Academic Press, 1990. 34. Norman D A, Shallice T. Attention to action. Willed and automatic control of behavior. In Davidson R J, Schwartz G E, Shapiro D, eds. Consciousness and Self Regulation (pp. 1–17). New York: Plenum, 1986. 35. Fernandez-Duque D, Baird J, Posner M. Executive attention and metacognitive regulation. Consciousness and Cognition 2000; 9: 288–307. 36. Ingvar D H. Memory of the future: An essay on the temporal organization of conscious awareness. Human Neurobiologia 1985; 31: 503–524. 37. Barkley R A, Koplowitz S, Anderson T, McMurray M B. Sense of time in children with ADHD: Effects of duration, distraction, and stimulant medication. Journal of the International Neuropsychological Society 1997; 3: 359–369. 38. Janowsky J S, Shimamura A P, Squire L R. Memory and metamemory: Comparisons between patients with frontal lobe lesions and amnesic patients. Psychobiology 1989; 17: 3–11.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

210

T. Mäntylä & M.-G. Carelli

39. Parkin A J, Hunkin N M, Walter B M. Relationships between normal aging, frontal lobe function, and memory for temporal and spatial information. Neuropsychology 1995; 9: 304–312. 40. Shallice T, Burgess P W. Deficits in strategy and application following frontal lobe damage in man. Brain 1991; 114: 727–741. 41. Barkley R A. Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD. Psychological Bulletin 1997; 121: 65–94. 42. Capella B, Gentile J R, Juliano D B. Time estimation by hyperactive and normal children. Perceptual and Motor Skills 1977; 44: 787–790. 43. Meaux J B, Chelonis J J. Time perception differences in children with and without ADHD. Journal of Pediatric Health Care 2003; 17: 64–71. 44. Burgess P W, Veitch E, de Lacy Costello A, Shallice T. The cognitive and neuroanatomical correlates of multitasking. Neuropsychologia 2000; 38: 848–863. 45. Glisky E L. Prospective memory and frontal lobes. In Brandimonte M A, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications (pp. 297–317). Hillsdale, NJ: Erlbaum, 1996. 46. Martin M, Kliegel M, McDaniel M, The involvement of executive functions in prospective memory performance of adults. International Journal of Psychology 2003; 38: 195–206. 47. McDaniel M A, Glisky E L, Rubin S R, Guynn M J, Routhieaux B C. Prospective memory: A neuropsychological study. Neuropsychology 1999; 13: 103–110. 48. Mäntylä T. Assessing absentmindedness: Prospective memory complaint and impairment in middle-aged adults. Memory & Cognition 2003; 31: 15–25. 49. Mäntylä T, Nilsson L-G. Remembering to remember in adulthood: A population-based study. Aging, Neuropsychology and Cognition 1997; 4: 81–92. 50. Einstein G O, McDaniel M A, Thomas R, Mayfield S, Shank H, Morrisette N, Breneiser J. Multiple processes in prospective memory retrieval: Factors determining monitoring versus spontaneous retrieval. Journal of Experimental Psychology: General 2005; 134: 327–342. 51. McDaniel M A, Einstein G O. Strategic and automatic processes in prospective memory retrieval. Applied Cognitive Psychology 2000; 14: 127–144. 52. McDaniel M A, Guynn M J, Einstein G O, Breneiser J. Cue-focused and reflexive-associative processes in prospective memory retrieval. Journal of Experimental Psychology: Learning, Memory and Cognition 2004; 30: 605–614. 53. Smith R E. The cost of remembering to remember in event-based prospective memory: Investigating the capacity demands of delayed intention performance. Journal of Experimental Psychology: Learning, Memory and Cognition 2003; 29: 347–361. 54. Einstein G O, Smith R E, McDaniel M A, Shaw P. Aging and prospective memory: The influence of increased task demands at encoding and retrieval. Psychology and Aging 1997; 12: 479–488. 55. Marsh R L, Hicks J L. Event-based prospective memory and executive control of working memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 1998; 24: 336–349. 56. McDaniel M A, Robinson-Riegler B, Einstein G O. Prospective remembering: Perceptually driven or conceptually driven processes? Memory & Cognition 1998; 26: 121–134.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

Time Monitoring and Executive Functioning

211

57. Marsh R L, Hicks J L, Cook G I. On the relationship between effort toward an ongoing task and cue detection in event-based prospective memory. Journal of Experimental Psychology: Learning, Memory and Cognition 2005; 31: 68–75. 58. Craik F I M. A functional account of age differences in memory. In Klix F, Hangendorf H, eds. Human Memory and Cognitive Capabilities: Mechanisms and Performances (pp. 409–422). Amsterdam: Elsevier, 1986. 59. Einstein G O, McDaniel M A. Retrieval processes in prospective memory: Theoretical approaches and some new empirical findings. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications (pp. 115–142). Mahwah, NJ: Erlbaum, 1996.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch08

This page intentionally left blank

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

9 The Neural Correlates of Timing Functions Katya Rubia∗

Introduction Human behavior is conducted in time and space, which makes good timing skills essential to human function. Behavior needs to be planned beforehand by considering future implications; the onset of the behavior needs to be timed to the optimal moment; and for an accurate temporal adjustment, the temporal interval between the intention of the behavior and its physical execution needs to be estimated. Timing of behavior therefore needs to be orchestrated by the triad of functions of temporal foresight, time estimation and motor timing. All three functions are likely to be closely intertwined in order to provide appropriately timed behavior in the motor, speech and cognitive domains. Neurobiological evidence suggests that they are mediated by similar brain regions although specific differences have emerged in their mediating networks. In the timing literature, the distinction has been made between motor timing and time perception.1 Motor timing refers to the timing aspects of the output of behavior such as the temporal organization of motor, speech or cognitive acts. Time perception refers to the more passive and perceptive aspects of cognitive time management such as perceiving temporal intervals and the ability to estimate temporal delays. In laboratory settings, motor timing has so far ∗ Department

of Child and Adolescent Psychiatry Institute of Psychiatry, King’s College London, London SE5 8AF, UK; e-mail: [email protected]

213

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

214

K. Rubia

been measured in tasks of finger tapping, rhythm production, rhythmic finger movements, sensorimotor synchronization, and the temporal organization of movements. The time range used with these methods ranges from milliseconds to seconds and minutes. Time estimation has been measured in tasks of temporal estimation, where temporal intervals from milliseconds to minutes or even hours need to be estimated; in tasks of temporal production or reproduction, where subjects are told to (re)produce a time interval given to them in conventional time units; in time discrimination tasks, where two different temporal intervals need to be discriminated; or in rhythm discrimination tasks. A third timing function that covers usually longer temporal intervals up to days and years is the ability to consider the consequences of one’s acts in the future, also called temporal foresight or temporal bridging.2 Compared to motor timing and time estimation functions, this function has been measured in a more indirect way in tasks usually implicating reward choice such as tasks of gambling or of temporal discounting. In these tasks, subjects have to choose between an immediate smaller reward or a bigger, but temporally delayed, reward. The reward value of the immediate option is discounted against the reward value of the delayed option in proportion to the temporal delay that separates the two rewards. Performance on this task requires reward choice in addition to temporal bridging abilities: the ability to cognitively bridge the temporal interval between the two reward options in order to decide on the preferred option. This task is therefore a measure of the subjective value of the temporal delay or, inversely, the degree of delay aversion. A preference for the smaller, immediate rewards is generally considered as an indicator of “temporal myopia,” or “myopia for the future,” a poor ability to consider the future consequences of one’s acts. Impulsive subjects, for example, who suffer from delay aversion, are more likely to choose the immediate reward over a delayed reward and thus show a steeper temporal discounting function. The hypothetical time intervals that separate the choices in tasks of temporal discounting typically range from weeks to years and are targeting therefore long-term temporal bridging abilities. Abnormalities in any of these three timing functions will result in abnormal behaviors. Impulsiveness, in particular, has been related to poor timing skills, with impulsive subjects showing deficits in all three timing domains, in motor timing, time perception and temporal foresight. Children with Attention Deficit Hyperactivity Disorder (ADHD), for example, where impulsiveness can be considered the main feature3 have shown deficits in tasks of motor

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

215

timing,4–8 time perception7,9,10 and temporal discounting.11 Problems with temporal myopia as measured in a steeper discounting of delayed rewards in temporal discounting tasks have been observed most consistently in subjects with substance abuse,12–16 who have also shown deficits in time estimation17–19 and motor timing.20 Timing functions are also impaired in a wide range of other, non-impulsive psychopathologies. Deficits in time estimation have thus been observed in patients with brain lesions,21,22 antisocial personality disorder,23 dyslexia and dysphasia,24–26 schizophrenia,27– 29 depression,30– 31 schizophrenia,32 and Parkinson’s disease.33–35 Motor timing functions have been found to be abnormal in several psychopathologies such as dyslexia,36–37 Parkinson’s disease38,39 and alcohol abuse.20 This chapter is a review of the neural correlates of these three different timing functions and a discussion of the overlap and differences of the mediating neural networks. Over the last decades, brain lesion and imaging studies using a wide range of timing tasks, from simple motor tapping to higher complex tasks of temporal foresight, have attempted to specify the neural correlates associated with the various functions of motor and cognitive time management. Several regions in the frontal lobes such as dorsolateral and inferior prefrontal cortices and the supplementary motor area (SMA), but also non-frontal cortical regions such as the parietal lobes and subcortical brain areas including the cerebellum, the basal ganglia and the anterior cingulate, have been found to be implicated in timing functions. An important distinction in the timing literature is to be made between different temporal domains in which timing functions are being measured. This is particularly relevant as different cognitive functions are being co-measured in different temporal domains. For example, time estimation or reproduction of several seconds or minutes requires the ability to sustain attention to time as an underlying basis function. Patients with problems in sustaining attention will necessarily be impaired in long-term time estimations of seconds or minutes. Likewise, tasks where time intervals of several seconds have to be reproduced or compared, require good working memory skills, as the first interval to be reproduced or compared with needs to be held in working memory. The load on working memory in these tasks increases with the length of the time interval. Temporal discrimination tasks in the milliseconds range, for example, will therefore have a lower load on working memory than temporal reproduction tasks of several minutes. In reviewing the literature, therefore, the time range that has been tested by several studies will be clearly indicated.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

216

K. Rubia

Neural Correlates of Motor Timing The Frontal Lobes Regions of the frontal lobes have in particular been related to motor timing, in line with the known role of the frontal lobes for regulating the output of behavior.1 Within the frontal lobes, in particular the lateral prefrontal cortex and the supplementary motor area seem to be crucial for mediating temporal adjustment. Subcortical brain regions that have been associated with motor timing are the basal ganglia, the cerebellum and the anterior cingulate gyrus. Within the frontal lobes, in particular the dorsolateral prefrontal cortex (DLPFC) has been shown to be activated during sensorimotor synchronization of hundreds of milliseconds2,42 and of several seconds.2,40–42 In some studies, the focus of the lateral prefrontal activation was in a more inferior frontal location during finger tapping,43 rhythmic finger movement,44 rhythm reproduction,45 and sensorimotor synchronization40 or was found in both the DLPFC and the inferior prefrontal cortex.42 When synchronization of longer intervals of seconds was compared to synchronization of milliseconds, only the longer interval timing functions were mediated by the DLPFC and the inferior prefrontal cortex (IFC), while the shorter ones were mediated by more motor frontal areas such as the supplementary motor cortex (SMA) or the anterior cingulate gyrus (ACG).42 In an elegant attempt to disentangle the involvement of different prefrontal brain regions in different timing aspects, Brunia et al.46 attributed IFC activation to the execution of an anticipated timed movement (the production of a 3 s interval) based on feedback on previous performance, while the DLPFC appeared to use internal cues for temporal programing of the motor output. Another important brain region that has been implicated in motor timing functions is the SMA. The SMA forms part of fronto-striatal pathways; it has projections to and from the basal ganglia via the thalamus, and is also connected to frontal and parietal cortical attention areas.47 Focal lesions in the SMA have been shown to produce deficits in the timing of movements as tested in rhythm reproduction.48 Activation of the SMA has consistently been found in motor timing tasks, including tasks of finger tapping and rhythm tapping of hundreds of milliseconds as well as motor preparation and temporal synchronization of several seconds.2,41–43,45,46,49,50 The closely adjacent anterior cingulate gyrus (ACG) has also been found to be activated in motor timing tasks such as sensorimotor synchronization of hundreds of milliseconds2,40 and of seconds.2,42 There is thus evidence for a role of the SMA and the ACG for fine-temporal adjustment of the motor output.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

217

The Cerebellum The cerebellum and the basal ganglia are two subcortical brain structures that have been most crucially related to motor timing. The importance of the cerebellum in timing processes was postulated a long time ago51 and is now fairly well established.52 Lesion studies have shown that patients with cerebellar lesions display poor performance on both motor tapping and time estimation tasks in the range of hundreds of milliseconds.53–55 The cerebellum has also been found to be activated in functional imaging studies of sensorimotor synchronization of short intervals in the milliseconds range43,45,56,57 and of longer time intervals of several seconds.40,50 Transcranial magnetic stimulation of the cerebellum has shown to increase the variability of tapping of millisecond intervals.58

The Basal Ganglia Other subcortical regions that have consistently been related to motor timing are the basal ganglia. The left and right putamen40,42 and left putamen, globus pallidum and caudate nucleus41,50 have been found to be activated during sensorimotor synchronization tasks of several seconds. Motor timing in the milliseconds/seconds range has also been shown to be mediated by the basal ganglia. The left putamen has been found to be activated during a finger tapping task of hundreds of milliseconds43,57 and the left and right putamen have been found to be activated during rhythm reproduction in the milliseconds range.45 The role of the cerebellum and the basal ganglia in motor timing is not surprising given the important role these two structures have in fine modulation of the behavioral output and of movement. Both the basal ganglia and the cerebellum have important reciprocal connections with motor areas of the frontal lobes,59–61 and also receive input from sensory brain regions such as the parietal lobes. Their role in fine modulation of the motor and cognitive output makes them well suited to regulating the temporal adjustment of behavior, probably via fronto-cerebellar and fronto-striatal connections.

Neural Correlates of Time Estimation The Frontal Lobes The prefrontal cortex was one of the first brain regions to be related to time perception, based on animal and lesion studies of an involvement of the prefrontal cortex in planning of behavior and the perception of time.1 Lesion studies have confirmed the involvement of the frontal lobes in time estimation. Patients with

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

218

K. Rubia

lesions of right and left frontal brain regions appear to be impaired in their ability to estimate temporal durations of milliseconds, seconds and minutes.21,22,62–64 In some of these lesion studies, in particular the integrity of the right DLPFC and the right inferior parietal lobe has been shown to be critical for time discrimination and estimation deficits of several seconds.21,22,63,65,66 Modern functional imaging studies using functional magnetic resonance imaging (fMRI) or positron emission tomography (PET) have confirmed the role of particularly the right hemispheric DLPFC and IFC in time estimation of several seconds66–71 and in time discrimination in the milliseconds’ range72−74 (see Fig. 1). Differences have been found between the neural correlates of long-term and short-term time estimations. Mangels et al.63 found that damage in lateral prefrontal cortex impaired the discrimination of long (4 secs) but not short temporal durations (400 msecs). A contrast of sensorimotor synchronization of several

(a) Motor timing of several seconds.

(b) Time discrimination of several hundreds of milliseconds.

Fig. 1. The figure shows activation of identical regions of the right dorsolateral and inferior prefrontal cortex during (a) motor timing and (b) time discrimination. (a) Group brain activation of 8 healthy adults in the right dorsolateral and inferior prefrontal cortex during synchronization of a motor response to a 5 s temporal interval when contrasted with synchronization to a 0.6 s interval. The long event rate condition imposes a greater load on time estimation than the short event rate condition (for further details see Rubia et al.42 ). (b) Group brain activation of 20 healthy adults during temporal discrimination of second intervals that differed by hundreds of milliseconds, when contrasted with a temporal order task. Shown are right dorsolateral and inferior prefrontal cortices and the SMA (for details see Smith et al.74 ).

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

219

seconds with that of hundreds of milliseconds in fMRI showed activation of lateral dorsolateral and inferior prefrontal brain regions2,42 (see Fig. 1). Similar findings of an implication of the DLPFC during the estimation of several seconds, but not milliseconds, has been made by Jones et al.69 These findings may suggest that regions of the prefrontal cortex have the function of a hypothetical accumulator within an internal clock model, which is required only with durations of more than several seconds. Indeed, prefrontal activation in timing tasks of durations of several seconds has often been related to other underlying functions besides pure timing processes, such as sustained attention to the time interval or working memory components71–73 based on the well-known role of the DLPFC in working memory75,76 and sustained attention.77,78 In fact, in some of the studies, DLPFC activation was not only related to temporal discrimination but also to the attentional control conditions.40,79,80 However, other studies have suggested that DLPFC may have a more primary role in time estimation processes.2,42,68,81,82 It has been argued that DLPFC activation often observed during working memory tasks such as the delayed response task (where a response is requested after a certain temporal delay period) could in fact reflect underlying timing processes such as bridging the temporal gaps involved in these tasks or timing of the motor response.2,42 This argument is based on the fact that studies using delay tasks with minimal working memory load have observed strong DLPFC and IFC activation2,42 (see Fig. 1). In support of this, an fMRI study found increasing activation in the DLPFC with increasing delays in a working memory task, but not with increasing working memory load.83 A study by Pochon et al.84 comparing a delayed matching task with a delayed response preparation task found that right-sided DLPFC activation was stronger for the response preparation than for the working memory task. Single cell recordings in the prefrontal cortex in monkeys have shown that specific neurons in the DLPFC of monkeys seem to be mediating mnemonic content, while others were specifically firing during temporal processes.81,85 It has therefore been argued that specific neural circuits in DLPFC could act as cortical oscillators and form the neural basis of a central clock mechanism.82 Further support for the involvement of the DLPFC in specific time estimation processes comes from the fact that the DLPFC has been found to be involved also in shorter time estimation processes in the milliseconds range, where sustained attention and working memory functions are less relevant.57,72–74,86 The most likely explanation for the overlap of DLPFC regions for both time estimation and working memory is that different regions within the DLPFC mediate timing and working

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

220

K. Rubia

memory functions.2,87,88 An alternative view could be that working memory, that is, holding the temporal interval online, is an important inherent cognitive component of time estimation processes which would be more crucially involved during longer intervals of several seconds. The DLPFC could then be thought to act as an “accumulator,” storing information about a passing time interval and making it the working memory component of a hypothesized internal clock.63,89 This theory would be supported by findings that concurrent tasks of working memory seem to impair processes of time reproduction and time estimation, probably due to shared working memory processes90,91 and that midazolam, which is known to affect memory processes, affects temporal discrimination of long intervals of seconds, but not short intervals of milliseconds.92 Not only the right DLPFC, but also the adjacent right inferior prefrontal cortex (IFC) has commonly been found to be activated during time estimation processes. Thus, the IFC has been shown to be activated in perceptive timing paradigms such as temporal discrimination of hundreds of milliseconds,66,72,74,93,94 attention to synchrony/asynchrony,89 rhythm perception,95 timed counting of hundreds of milliseconds86 and temporal production of several seconds.46 A combination of event related potentials (ERPs) with PET showed increased activation in the right IFC and ACG during time discrimination trials; furthermore, the timing of the ERPs in the right inferior prefrontal regions were aligned with the durations themselves.94 As mentioned above, we observed right IFC and DLPFC activation in a motor delay task, where subjects had to adjust the motor response to a stimulus appearing every 5 seconds, but not during synchronization of milliseconds, suggesting a role of both the DLPFC and IFC in time estimation functions42,66 (see Fig. 1). Very similar locations of right IFC and DLPFC activation were observed during a temporal discrimination task of hundreds of milliseconds74 (see Fig. 1). An fMRI study using a parametric design found that increasing the load on attention to time intervals increased activation in the right DLPFC and IFC.68 Gruber et al.89 and Schubotz et al.95 found activation in the IFC where subjects were instructed simply to attend to rhythm and where no movement was required. In conclusion, there is thus sufficient evidence that besides the right DLPFC, the adjacent IFC also plays an important role in purely perceptive time estimation processes. Although the SMA has traditionally been suggested to be involved in purely motor functions, including motor timing, recent evidence has associated the SMA also in functions of pure perceptive timing. Thus, some studies have observed increased SMA activation during estimation of longer time intervals

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

221

of seconds as opposed to milliseconds94,96,97 and in time production of several seconds.70 SMA activation has also been found in imaging studies using tasks of discrimination of short intervals in the milliseconds range,73 of discriminination and reproduction of seconds71,74,98 (see Fig. 1). It has furthermore been implicated in rhythm discrimination involving milliseconds,89,95 timed counting86 and in temporal orienting to brief temporal intervals of hundreds of milliseconds.79 Macar et al.71 found SMA activation in both short (milliseconds) and long (seconds) time interval discriminations. Manipulation of the load on purely perceptive attention to time intervals showed an increase in SMA activation.68 It thus appears that, while earlier studies have postulated a strong role of the SMA in motor timing processes, more recent studies have shown that the timing functions of the SMA also include purely perceptive timing. The ACG has been more consistently related to time estimation than to motor timing. The ACG is closely interconnected to the SMA, parietal and frontal lobes and their fronto-strio-thalamic pathways.61 The ACG has thus been found to be activated in studies of time estimation such as time production and reproduction of seconds70,71,99 and temporal discrimination72 and timed counting86 in the milliseconds range. However, we observed a biphasic activation in the ACG during sensorimotor synchronization of both hundreds of milliseconds and of several seconds,2,42 suggesting a role of this brain region for both motor timing and time estimation. It has been suggested that the ACG, rather than being specifically related to cognitive time management per se, might be related to motor attention functions. The ACG forms part of the midline attention system100 and has been attributed a role in attention to action as well as an evaluative comparator role assisting executive control,101,102 both important functions that are necessary for motor timing and distinguishing time intervals, respectively. The anterior cingulate has important anatomical and functional connections with the lateral prefrontal cortex, which it seems to assist in self-monitoring functions of cognitive control,103 making it well suited to assisting in cognitive time management functions.

The Cerebellum There is strong evidence for the implication of the cerebellum not only in motor timing, but particularly in time estimation. A lesion study showed that the poor performance of cerebellar patients on motor tapping and time discrimination contrasted with the performance of patients with cortical lesions, who showed deficits in finger tapping but not discrimination tasks, and patients with basal

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

222

K. Rubia

ganglia damage, whose performance did not differ from that of controls in either task.54,55 Since temporal discrimination is often thought to be the purest measure of time perception,5 this study was interpreted as evidence for a central role of the cerebellum in temporal perception. Other studies of cerebellar patients have shown them to be poor at time discrimination in both long (seconds) and short (hundreds of milliseconds) intervals62,63,104 and, in contrast to patients with prefrontal lesions, the temporal discrimination deficits of cerebellar patients were not alleviated by counting strategies and the use of short durations.63 The above evidence, derived from focal lesion studies is supported by functional imaging studies where cerebellar activation has been found during temporal discrimination of short intervals of hundreds of milliseconds,72–74,105,106 temporal orienting of under a second intervals,79 rhythm discrimination,95 rhythm reproduction of hundreds of milliseconds45 and time production of several seconds.70,80 Based on the findings in the literature of an involvement of the cerebellum in both motor timing and time perception tasks, it has been speculated that the cerebellum might be especially relevant to event timing, such as the prediction of the temporal occurrence of events including the temporal adjustments of behavior to the timing of these events.107 A study manipulating the temporal occurrence of events did in fact isolate lateral cerebellar activation for the timing of events as opposed to other parameters, suggesting that the cerebellum might be mediating the attempt to predict the timing of stimulus occurrence.108 Since the majority of imaging studies have found an involvement of the lateral portions of the cerebellar hemispheres in timing processes, it has been suggested that motor execution may be subserved by medial regions of the cerebellum, while internal clock processes or temporal management may be subserved by lateral regions of cerebellum.54 In line with this functional division is the difference in the connectivity of these two regions of cerebellum — the lateral cerebellum projects to the premotor cortex and DLPFC, important for motor and perceptive timing, while the medial cerebellum is connected with the spinal cord, affecting motor implementation.59,60 Apart from the lateral cerebellar hemispheres, evidence from PET studies also suggests that the vermis of the cerebellum might be crucial for time estimation as it has been found to be activated in temporal discrimination of hundreds of milliseconds.72,105

The Basal Ganglia Although some studies showed no deficits in patients with lesions of the basal ganglia in time discrimination or time estimation,55,109 there is evidence for the

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

223

involvement of the basal ganglia in time estimation. Thus, lesions in the right supralenticular white matter, presumably consisting of fronto-striatal pathways, have been found to be associated with impaired time estimation and production of several seconds in patients with brain lesions.21 The caudate and putamen have been found to be activated in time discrimination tasks in the milliseconds range,73,105,106 in rhythm discrimination of hundreds of milliseconds,95 and in time production of several seconds.70 We observed right putamen activity in a sensorimotor task of 5 s in healthy adults using fMRI, which was not observed during synchronization of milliseconds and is therefore more likely to be related to time estimation than to motor timing per se.42 The putamen in particular could be isolated when temporal discrimination in the milliseconds range were compared to frequency detection.110 The putamen has also been shown to be activated when attention was directed to temporal intervals, suggesting a role for the basal ganglia for attention to time.68 It is interesting to note that studies that have compared time estimations of long and short intervals, in the seconds and milliseconds range, respectively, have found that while right prefrontal brain regions seem to be activated stronger or exclusively for the long time intervals, the cerebellum and basal ganglia show activation during both short and long time intervals,69 suggesting that these two brain regions might be more crucially involved in hypothetical internal clock mechanisms, important for both shortand long-term timing, while the prefrontal lobes may mediate timing functions via their role in additional cognitive functions underlying time estimation such as working memory or sustained attention to time.

The Parietal Lobes A non-frontal cortical brain region that has commonly been associated with time estimation, but less with motor timing, is the inferior parietal cortex. Focal lesion studies have found time estimation deficits of several seconds in patients with predominantly right parieto-occipital brain lesions.22,111 Inferior parietal lobes have found to be activated during a sensorimotor synchronization task of several seconds, which involved both time estimation and time estimation functions,40,42 during rhythm reproduction of hundreds of milliseconds,45 during time estimation tasks of several seconds67,70,71 and in temporal discrimination72,73,93,106 rhythm discrimination95 and timed counting of hundreds of milliseconds.86 It has been argued that the role of the parietal lobes in time estimation tasks could be related to its known implication in sustaining attention, therefore providing sustained attention to time.86,112 This is in line with a study that compared

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

224

K. Rubia

attention to time with attention to sensory stimuli and found bilateral parietal lobe activation.68 In a time discrimination task that was controlled for sustained attention by a control task, we did not observe any parietal lobe activation.74 Sustained attention to time intervals is certainly a necessary basis function for time estimation processes. The inferior parietal lobes are interconnected with the frontal lobes, the basal ganglia and the cerebellum,113,114 all of which have shown to be importantly implicated in time estimation. The parietal lobes with their connections to fronto-striatal and fronto-cerebellar circuits are thus strategically well placed to support cognitive time management processes by assisting with sustained attention to time.

Neural Correlates of Temporal Foresight Compared to time estimation and motor timing, relatively few studies have investigated the neural basis of temporal foresight. One of the reasons is the fact that temporal foresight is difficult to measure directly in a laboratory situation and hard to disentangle from other cognitive functions as it involves relatively long temporal intervals. A rather indirect way of measuring temporal foresight has been by use of tasks of gambling or temporal discounting which involve the choice between a present, smaller and a delayed, larger reward. The tasks thus measure the subjective value of the temporal delay in terms of reward; performance on these tasks is thought to reflect at least partly the ability to cognitively bridge the temporal interval between the two reward options, that is, intertemporal competence. The choice of immediate rewards measures the impact of reward and has been shown to elicit activation in limbic areas such as lateral orbitofrontal and ventral striatal brain regions while the more reflective choice of delayed rewards has been suggested to measure brain regions associated with temporal foresight. Lesion studies have associated frontal brain regions with temporal foresight. Thus, lesions of the bilateral, but predominantly right ventromedial and dorsolateral frontal cortex have shown to impair the intertemporal choice, making patients choose the less advantageous immediate reward options.115–119 Further task manipulations suggested that the underlying reason for poor performance on this task, rather than reward hypersensitivity, was an insensitivity for future consequences, “temporal myopia.”117 Another lesion study using temporal discounting found that patients with lesions of the DLPFC and ventromedial frontal lobe did not differ from non-frontal lobe lesions in their rates of temporal discounting,120 but patients with lesions in the ventromedial frontal lobe had a foreshortening effect on the personal future time perspective as

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

225

measured on a questionnaire.120 Modern imaging studies have provided further insights into the neural correlates of temporal foresight and temporal myopia. Positron emission tomography on the Bechara gambling task found activation of the right orbitofrontal cortex and DLPFC for the choice of the long-term advantageous options, thus supporting the role of the right prefrontal cortex for temporal foresight.121 Interestingly, subjects with marijuana or cocaine abuse, who showed poor inter-temporal choice also showed a reduced activation in orbitofrontal cortex and DLPFC.122,123 A study using a temporal discounting task found that lateral prefrontal brain regions including the dorsolateral, ventrolateral and orbital frontal cortex and the parietal lobes showed increased activation when subjects chose longer term reward options as opposed to shorter term reward options, also implicating the lateral prefrontal cortex and, in addition, the parietal brain regions in the mediation of temporal foresight.124 Another fMRI study using a different decision making task tapping into intertemporal choice, where subjects learned to make decisions that obtained them larger future rewards but small immediate losses, showed activation for the long-term versus immediate reward options in the DLPFC and inferior parietal lobes, in line with the findings of the previous study.125 Other additional areas that were found to be related to the long-term reward option were the premotor cortex, insula, basal ganglia and cerebellum, areas that have also been associated with other timing functions, suggesting a fronto-striatal network is important for temporal foresight.125 In conclusion, studies that have investigated temporal foresight have related this function to similar right lateral prefrontal or fronto-striatal brain regions that have also been shown to be particularly involved in the estimation of longer time intervals of seconds and minutes, supporting the importance of lateral prefrontal brain regions for inter-temporal competence. Brain regions that have been found more specifically to be related to temporal foresight but not to time estimation are the ventromedial prefrontal brain areas.

General Conclusions In summary, several brain regions have been shown to be involved in timing functions, including the right dorsolateral, inferior/orbital prefrontal and parietal cortices, the anterior cingulate gyrus, SMA, basal ganglia and the cerebellum (see Table 1). Evidence from the literature suggests that each of these brain regions may contribute to specific aspects of timing functions. The cerebellum

Motor timing of sec

Time estimation of msecs (Time estimation, (re)production and discrimination)

Time estimation of secs and mins (Time estimation, (re)production and discrimination)

Inter-temporal bridging of long intervals (Tasks of inter-temporal choice such as gambling and temporal discounting)

Dorsolateral prefrontal cortex

Larsson et al.59

Lejeune et al.,40 Lewis et al.,41 Rubia et al.2,42

Maquet et al.,72 Rao et al.,73 Smith et al.,74 Ortuno et al.86

Harrington et al.,21 Kagerer et al.,65 Mangels et al.,63 Basso et al.,67 Coull et al.,68 Jones et al.,97 Lewis & Miall,63 Macar et al.71

Bechara et al.,115−117 Tranel et al.,118 Clark et al.,119 Bolla et al.,122 McClure et al.,124 Tanaka et al.125

Inferior prefrontal cortex

Rao et al.73

Rubia et al.,42 Lejeune et al.,40 Brunia et al.46

Maquet et al.,72 Pedersen et al.,93 Pouthas et al.,94 Smith et al.,74 Ortuno et al.86

Kagerer et al.,65 Harrington et al.,21 Mangels et al.63

McClure et al.124

Lateral orbitofrontal cortex

Bolla et al.,122 McClure et al.124

Ventromedial orbitofrontal cortex

Bechara et al.,115−117 Tranel et al.,118 Clark et al.119

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

Motor timing of msecs (Sensori motor synchronization and rhythm (re)production)

K. Rubia

Brain region

226

Table 1. Schematic overview of brain regions that have been found to be involved in the different timing functions of motor timing, time estimation and temporal foresight in lesion and functional imaging studies. The review is not exhaustive and there may be studies that have not been included.

Lejeune et al.,40 Lewis et al.,41 Rubia et al.,42 Riecker et al.,50 Rao et al.,73 Larsson et al.,59 Penhune et al.45

Cerebellum

Inui & Hatta,56 Larsson et al.,59 Penhune et al.,45 Rao et al.,43 Ivry & Diener et al.,53 Ivry et al.,54 Justus & Ivry et al.55

Supplementary Rubia et al.,2,42 Rao motor area et al.,43 Brunia et al.,46 Penhune et al.,45 Lang et al.,49 Riecker et al.50

Jueptner et al.,105 Rao et al.,59 Dupont et al.,106 Nenadic et al.,110 Coull et al.68

Rubia et al.,21 Lewis & Miall63

Lejeune et al.,40 Riecker et al.,50 Theoret et al.58

Ivry & Diener et al.,53 Ivry et al.,54 Justus & Ivry et al.,55 Casini & Ivry,62 Nichelli et al.,64 Mangles et al.,63 Dupont et al.,106 Jueptner et al.,105 Maquet et al.,72 Rao et al.,73 Mathiak et al.,137 Smith et al.74

Mangels et al.,63 Casini & Ivry,62 Nichelli et al.,64 Lewis & Miall,63 Tracy et al.80

Lewis et al.41

Rao et al.,73 Smith et al.,74 Ortuno et al.,86 Macar et al.,71

Pouthas et al.,94 Jones et al.,97 Ferrandez et al.,96 Lewis & Miall,63 Macar et al.,71,98 Coull et al.68

Anterior cingulate gyrus

Rubia et al.,2 Lejeune et al.40

Rubia et al.2,42

Ortuno et al.86

Macar et al.,71 Hinton et al.,99 Lewis & Miall63

Parietal lobes

Rubia et al.,2 Lejeune et al.40

Rubia et al.,2,42 Lejeune et al.40

Dupont et al.,106 Mathiak et al.,137 Maquet et al.,72 Rao et al.,73 Pederson et al.,93 Ortuno et al.86

Kagerer et al.,65 Harrington et al.,21 Petrovivi & Scheider,111 Basso et al.,67 Lewis & Miall,63 Macar et al.,71 Coull et al.68

The Neural Correlates of Timing Functions

Basal ganglia

McClure et al.,124 Tanaka et al.125

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

Table 1. (Continued )

227

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

228

K. Rubia

and the basal ganglia seem to be the brain regions that are closest in mediating pure timing mechanisms as they have been associated with timing at both short and long temporal domains, and may possibly form the neural basis of internal clock mechanisms. The role of the basal ganglia as the locus of a hypothetical internal timing device is particularly interesting in view of the known implication of the neurotransmitter dopamine in timing processes and of timing deficits in disorders of dopaminergic imbalance. Dopaminergic agents have thus been shown to have an effect on time estimation and motor timing functions in healthy subjects.126–128 Animal studies show disruptions in response timing after focal lesions or drugs targeting the dopaminergic functions in the basal ganglia (for a review, see Meck et al.129 ). There is also evidence for a dopaminergic involvement in temporal discounting. Dopaminergic agonists such as amphetamines have been shown to lead to shallower temporal discounting in humans,130 while tryptophan depletion have been shown to have no effect.131 The involvement of striatal dopamine in timing functions is further supported by evidence that patients with known striatal dopamine disturbance are impaired in timing. For example, patients with Parkinson’s disease who are known to suffer from striatal dopamine deficiencies show deficits in motor timing and time perception that can be ameliorated with dopamine-agonists.33– 35,38 Similarly, patients with ADHD who are known to have elevated striatal dopamine transporter levels,132 probably leading to reduced extracellular striatal dopamine release, have been shown to have deficits in time estimation, motor timing and temporal discounting,5,7,9,11 some of which have been shown to be ameliorated with the dopamine agonist Methylphenidate.7 Interestingly, patients with ADHD also show structural and functional deficits in other brain regions that mediate timing functions other than the basal ganglia such as the vermis of the cerebellum and the right prefrontal lobes.133–136 The right lateral prefrontal lobes seem to be particularly involved in those timing processes that require the temporal bridging of longer temporal intervals of seconds and minutes and intertemporal choice of hypothetical long temporal gaps exceeding minutes. The contribution of the right lateral prefrontal cortex to timing processes may therefore be closely linked to other functions that are known to be mediated by this brain region such as sustaining attention to time and working memory. The SMA has traditionally been related to motor timing, although recent evidence has also attributed a role to this area in pure perceptual time estimation,

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

229

which may be necessary for fine temporal adjustment of movement. The anterior cingulate gyrus has been suggested to have a more generic role in attentional components necessary for both motor timing (attention to action) and time estimation (evaluative comparator role). This area has been shown to be anatomically and functionally closely connected to the lateral prefrontal cortex and may thus contribute to timing functions as a comparator of temporal intervals in time estimation tasks and by assisting motor timing with allocation of motor attention. The parietal lobes with their close connections to fronto-striatal and frontocerebellar pathways seem to contribute to timing functions through allocation of sustaining attention to time. To summarize, this literature review shows that complex neural networks are involved in timing functions, including cortico-cortical, fronto-striatal, frontocerebellar and cortico-neocortical interconnections. Future research using functional imaging techniques in combination with sophisticated timing paradigms and using connectivity analyses will be crucial to further disentangle the specific contributions of these different brain regions to the specific functions necessary for optimal temporal behavior. It would furthermore be interesting to investigate the localization of timing in the future or time-based prospective memory, as involved for example in executing an action in a specific time period in the future. Timing in the future would involve holding the information online in memory for the time period of the delay, good intertemporal bridging abilities and adequate timing of the motor execution when the time period has finally elapsed. Presumably, this would involve all three timing functions, and most crucially intertemporal bridging of longer time periods, in addition to functions of working memory. Memory for the future might be mediated therefore in particular by lateral prefrontal brain regions as the mediators of both working memory and intertemporal bridging of longer temporal intervals, possibly with its interconnections to striatal and cerebellar brain regions, necessary for the execution of the timed behavior when the time has elapsed. It has been argued that timing functions are based on an hypothetical internal clock, which has been suggested to be localized either in the basal ganglia,69 the cerebellum51 or the dorsolateral prefrontal cortex.63,89 The fMRI literature seems to suggest that it is rather more likely that there is no unitary timing device such as an internal clock, but that timing is mediated by different, overlapping neural networks that work synergistically to provide different timing functions for specific temporal domains.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

230

K. Rubia

References 1. Fuster J M. The prefrontal cortex — An update: Time is of the essence. Neuron 2001; 30(2): 319–333. 2. Rubia K, Overmeyer S, Taylor E et al. Prefrontal involvement in “temporal bridging” and timing movement. Neuropsychologia 1998; 36(12): 1283–1293. 3. Rubia K. The dynamic approach to neurodevelopmental psychiatric disorders: Use of fMRI combined with neuropsychology to elucidate the dynamics of psychiatric disorders, exemplified in ADHD and schizophrenia. Behavioral Brain Research 2002; 130(1&2): 47–56. 4. Carte E T, Nigg J T, Hinshaw S P. Neuropsychological functioning, motor speed, and language processing in boys with and without ADHD. Journal of Abnormal Child Psychology 1996; 24(4): 481–498. 5. Rubia K, Taylor A, Taylor E, Sergeant J A. Synchronization, anticipation, and consistency in motor timing of children with dimensionally defined attention deficit hyperactivity behavior. Perceptual and Motor Skills 1999; 89(3): 1237–1258. 6. Rubia K, Taylor E, Smith A B, Oksannen H, Overmeyer S, Newman S. Neuropsychological analyses of impulsiveness in childhood hyperactivity. British Journal of Psychiatry 2001; 179: 138–143. 7. Rubia K, Noorloos J, Smith A, Gunning B, Sergeant J. Motor timing deficits in community and clinical boys with hyperactive behavior: The effect of methylphenidate on motor timing. Journal of Abnormal Child Psychology 2003; 31(3): 301–313. 8. Stevens D A, Stover C E, Backus J T. Hyperkinetic child: Effect of incentives on speed of rapid tapping. Journal of Consulting and Clinical Psychology 1970; 34(1): 56–59. 9. Smith A, Taylor E, Rogers J W, Newman S, Rubia K. Evidence for a pure time perception deficit in children with ADHD. Journal of Child Psychology and Psychiatry and Allied Disciplines 2002; 43(4): 529–542. 10. Toplak M E, Rucklidge J J, Hetherington R, John S C F, Tannock R. Time perception deficits in attention-deficit/hyperactivity disorder and comorbid reading difficulties in child and adolescent samples. Journal of Child Psychology and Psychiatry and Allied Disciplines 2003; 44(6): 888–903. 11. Barkley R A, Edwards G, Laneri M, Fletcher K, Metevia L. Executive functioning, temporal discounting, and sense of time in adolescents with attention deficit hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD). Journal of Abnormal Child Psychology 2001; 29(6): 541–556. 12. Bickel W K, Marsch L A. Toward a behavioral economic understanding of drug dependence: Delay discounting processes. Addiction 2001; 96(1): 73–86. 13. Baker F, Johnson M W, Bickel W K. Delay discounting in current and never-before cigarette smokers: Similarities and differences across commodity, sign, and magnitude. Journal of Abnormal Psychology 2003; 112(3): 382–392. 14. Coffey S F, Gudleski G D, Saladin M E, Brady K T. Impulsivity and rapid discounting of delayed hypothetical rewards in cocaine-dependent individuals. Experimental and Clinical Psychopharmacology 2003; 11(1): 18–25.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

231

15. Giordano L A, Bickel W K, Loewenstein G, Jacobs E A, Marsch L, Badger G J. Mild opioid deprivation increases the degree that opioid-dependent outpatients discount delayed heroin and money. Psychopharmacology 2002; 163(2): 174–182. 16. Kirby K N, Petry N M. Heroin and cocaine abusers have higher discount rates for delayed rewards than alcoholics or non-drug-using controls. Addiction 2004; 99(4): 461–471. 17. Mathew R J, Wilson W H, Turkington T G, Coleman R E. Cerebellar activity and disturbed time sense after THC. Brain Research 1998; 797(2): 183–189. 18. Mintzer M Z, Stitzer M L. Cognitive impairment in methadone maintenance patients. Drug and Alcohol Dependence 2002; 67(1): 41–51. 19. Solowij N, Stephens R S, Roffman R A et al. Cognitive functioning of long-term heavy cannabis users seeking treatment. Journal of the American Medical Association 2002; 287(9): 1123–1131. 20. Parks M H, Morgan V L, Pickens D R et al. Brain fMRI activation associated with self-paced finger tapping in chronic alcohol-dependent patients. Alcoholism-Clinical and Experimental Research 2003; 27(4): 704–711. 21. Rubia K, Schuri U, von Cramon D Y, Poeppel E. Time estimation as a neuronal network property: A lesion study. Neuroreport 1997; 8(5): 1273–1276. 22. Harrington D L, Haaland K Y, Knight R T. Cortical networks underlying mechanisms of time perception. Journal of Neuroscience 1998; 18(3): 1085–1095. 23. Bauer L O. Antisocial personality disorder and cocaine dependence: Their effects on behavioral and electroencephalographic measures of time estimation. Drug and Alcohol Dependence 2001; 63(1): 87–95. 24. May J G, Williams M C, Dunlap W P. Temporal-order judgements in good and poor readers. Neuropsychologia 1988; 26(6): 917–924. 25. Nicolson R I, Fawcett A J, Dean P. Time-estimation deficits in developmental dyslexia — Evidence of cerebellar involvement. Proceedings of the Royal Society of London Series B-Biological Sciences 1995; 259(1354): 43–47. 26. Tallal P, Sainburg R L, Jernigan T. The neuropathology of developmental dysphasia — Behavioral, morphological, and physiological evidence for a pervasive temporal processing disorder. Reading and Writing 1991; 3(3–4): 363–377. 27. Davalos D B, Kisley M A, Ross R G. Effects of interval duration on temporal processing in schizophrenia. Brain and Cognition 2003; 52(3): 295–301. 28. Ulferts J, Meyer-Lindenberg A, Gallhofer B. Time discrimination: A comparison between healthy controls, unmedicated schizophrenics, zotepine-treated schizophrenics and schizophrenics treated with conventional neuroleptics. Neuropsychiatrie 1999; 13(3): 133–138. 29. Volz H P, Nenadic I, Gaser C, Rammsayer T, Hager F, Sauer H. Time estimation in schizophrenia: An fMRI study at adjusted levels of difficulty. Neuroreport 2001; 12(2): 313–316. 30. Kuhs H, Hermann W, Kammer K, Tolle R. Time-estimation and the experience of time in endogenous-depression (Melancholia) — An experimental investigation. Psychopathology 1991; 24(1): 7–11.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

232

K. Rubia

31. Mundt C, Richter P, van Hees H, Stumpf T. Time-experience and time-estimation in depressive patients. Nervenarzt 1998; 69(1): 38–45. 32. Rammsayer T H. Temporal discrimination in schizophrenic and affective disorders: Evidence for a dopamine-dependent internal clock. International Journal of Neuroscience 1990; 53: 111–120. 33. Lange K W, Tucha O, Steup A, Gsell W, Naumann M. Subjective time estimation in Parkinson’s disease. Journal of Neural Transmission (Supplement) 1995: (46): 433–438. 34. Pastor M A, Artieda J, Jahanshahi M, Obeso J A. Time-estimation and reproduction is abnormal in Parkinsons disease. Brain 1992; 115: 211–225. 35. Riesen J M, Schnider A. Time estimation in Parkinson’s disease: Normal long duration estimation despite impaired short duration discrimination. Journal of Neurology 2001; 248(1): 27–35. 36. Denckla M B, Rudel R G, Chapman C, Krieger J. Motor proficiency in dyslexic-children with and without attentional disorders. Archives of Neurology 1985; 42(3): 228–231. 37. Waber D P, Weiler M D, Bellinger D C et al. Diminished motor timing control in children referred for diagnosis of learning problems. Developmental Neuropsychology 2000; 17(2): 181–197. 38. O’Boyle D J, Freeman J S, Cody F W J. The accuracy and precision of timing of self-paced, repetitive movements in subjects with Parkinson’s disease. Brain 1996; 119: 51–70. 39. Elsinger C L, Rao S M, Zimbelman J L, Reynolds N C, Blindauer K A, Hoffmann R G. Neural basis for impaired time reproduction in Parkinson’s disease: An fMRI study. Journal of the International Neuropsychological Society 2003; 9(7): 1088–1098. 40. Lejeune H, Maquet P, Bonnet M et al. The basic pattern of activation in motor and sensory temporal tasks: Positron emission tomography data. Neuroscience Letters 1997; 235(1–2): 21–24. 41. Lewis P A, Wing A M, Pope P A, Praamstra P, Miall R C. Brain activity correlates differentially with increasing temporal complexity of rhythms during initialisation, synchronisation, and continuation phases of paced finger tapping. Neuropsychologia 2004; 42(10): 1301–1312. 42. Rubia K, Overmeyer S, Taylor E et al. Functional frontalisation with age: Mapping neurodevelopmental trajectories with fMRI. Neuroscience and Biobehavioral Reviews 2000; 24(1): 13–19. 43. Rao S M, Harrington D L, Haaland K Y, Bobholz J A, Cox R W, Binder J R. Distributed neural systems underlying the timing of movements. Journal of Neuroscience 1997; 17(14): 5528–5535. 44. Kawashima R, Inoue K, Sugiura M, Okada K, Ogawa A, Fukuda H. A positron emission tomography study of self-paced finger movements at different frequencies. Neuroscience 1999; 92(1): 107–112. 45. Penhune V B, Zatorre R J, Evans A C. Cerebellar contributions to motor timing: A PET study of auditory and visual rhythm reproduction. Journal of Cognitive Neuroscience 1998; 10(6): 752–765. 46. Brunia C H M, de Jong B M, van den Berg-Lenssen M M C, Paans A M J. Visual feedback about time estimation is related to a right hemisphere activation measured by PET. Experimental Brain Research 2000; 130(3): 328–337.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

233

47. Schell G R, Strick P L. The origin of thalamic inputs to the arcuate premotor and supplementary motor areas. Journal of Neuroscience 1984; 4(2): 539–560. 48. Halsband U, Ito N, Tanji J, Freund H J. The role of premotor cortex and the supplementary motor area in the temporal control of movement in man. Brain 1993; 116: 243–266. 49. Lang W, Obrig H, Lindinger G, Cheyne D, Deecke L. Supplementary motor area activation while tapping bimanually different rhythms in musicians. Experimental Brain Research 1990; 79(3): 504–514. 50. Riecker A, Wildgruber D, Mathiak K, Grodd W, Ackermann H. Parametric analysis of ratedependent hemodynamic response functions of cortical and subcortical brain structures during auditorily cued finger tapping: A fMRI study. Neuroimage 2003; 18(3): 731–739. 51. Braitenberg V. Is the cerebellar cortex a biological clock in the millisecond range? Progress in Brain Research 1967; 25: 334–346. 52. Harrington D L, Haaland K Y. Neural underpinnings of temporal processing: A review of focal lesion, pharmacological, and functional imaging research. Reviews in the Neurosciences 1999; 10(2): 91–116. 53. Ivry R B, Diener H C. Impaired velocity perception in patients with lesions of the cerebellum. Journal of Cognitive Neuroscience 1991; 3(4): 355–366. 54. Ivry R B, Keele S W, Diener H C. Dissociation of the lateral and medial cerebellum in movement timing and movement execution. Experimental Brain Research 1988; 73(1): 167–180. 55. Justus T C, Ivry R B. The cognitive neuropsychology of the cerebellum. International Review of Psychiatry 2001; 13(4): 276–282. 56. Inui N, Hatta H. Effect of practice on force control and timing in bimanual finger tapping. Journal of Human Movement Studies 2003; 44(1): 79–91. 57. Larsson J, Gulyas B, Roland P E. Cortical representation of self-paced finger movement. Neuroreport 1996; 7(2): 463–468. 58. Theoret H, Haque J, Pascual-Leone A. Increased variability of paced finger tapping accuracy following repetitive magnetic stimulation of the cerebellum in humans. Neuroscience Letters 2001; 306(1–2): 29–32. 59. Middleton F A, Strick P L. Anatomical evidence for cerebellar and basal ganglia involvement in higher cognitive function. Science 1994; 266(5184): 458–461. 60. Middleton F A, Strick P L. Basal ganglia and cerebellar loops: Motor and cognitive circuits. Brain Research Reviews 2000; 31(2–3): 236–250. 61. Picard N, Strick P L. Motor areas of the medial wall: A review of their location and functional activation. Cerebral Cortex 1996; 6(3): 342–353. 62. Casini L, Ivry R B. Effects of divided attention on temporal processing in patients with lesions of the cerebellum or frontal lobe. Neuropsychology 1999; 13(1): 10–21. 63. Mangels J A, Ivry R B, Shimizu N. Dissociable contributions of the prefrontal and neocerebellar cortex to time perception. Cognitive Brain Research 1998; 7(1): 15–39. 64. Nichelli P, Clark K, Hollnagel C, Grafman J. Duration processing after frontal lobe lesions. Ann NY Acad Sci 1995; 15(769): 183–190. 65. Kagerer F A, Wittmann M, Szelag E, Von Steinbuchel N. Cortical involvement in temporal reproduction: Evidence for differential roles of the hemispheres. Neuropsychologia 2002; 40(3): 357–366.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

234

K. Rubia

66. Rubia K, Smith A. The neural correlates of cognitive time management: A review. Acta Neurobiologiae Experimentalis 2004; 64(3): 329–340. 67. Basso G, Nichelli P, Wharton C M, Peterson M, Grafman J. Distributed neural systems for temporal production: A functional MRI study. Brain Research Bulletin 2003; 59(5): 405–411. 68. Coull J T. fMRI studies of temporal attention: Allocating attention within, or towards, time. Cognitive Brain Research 2004; 21(2): 216–226. 69. Jones C R G, Rosenkranz K, Rothwell J C, Jahanshahi M. The right dorsolateral prefrontal cortex is essential in time reproduction: An investigation with repetitive transcranial magnetic stimulation. Experimental Brain Research 2004; 158(3): 366–372. 70. Lewis P A, Miall R C. Brain activity during non-automatic motor production of discrete multi-second intervals. Neuroreport 2002; 13(14): 1731–1735. 71. Macar F, Lejeune H, Bonnet M et al. Activation of the supplementary motor area and of attentional networks during temporal processing. Experimental Brain Research 2002; 142(4): 475–485. 72. Maquet P, Lejeune H, Pouthas V et al. Brain activation induced by estimation of duration: A PET study. Neuroimage 1996; 3(2): 119–126. 73. Rao S M, Mayer A R, Harrington D L. The evolution of brain activation during temporal processing. Nature Neuroscience 2001; 4(3): 317–323. 74. Smith A, Taylor E, Lidzba K, Rubia K. A right hemispheric frontocerebellar network for time discrimination of several hundreds of milliseconds. Neuroimage 2003; 20(1): 344–350. 75. Diwadkar V A, Carpenter P A, Just M A. Collaborative activity between parietal and dorso-lateral prefrontal cortex in dynamic spatial working memory revealed by fMRI. Neuroimage 2000; 12(1): 85–99. 76. Mull B R, Seyal M. Transcranial magnetic stimulation of left prefrontal cortex impairs working memory. Clinical Neurophysiology 2001; 112(9): 1672–1675. 77. Coull J T, Frackowiak R S J, Frith C D. Monitoring for target objects: Activation of right frontal and parietal cortices with increasing time on task. Neuropsychologia 1998; 36(12): 1325–1334. 78. Posner M I, Petersen S E. The attention system of the human brain. Annual Review of Neuroscience 1990; 13: 25–42. 79. Coull J T, Nobre A C. Where and when to pay attention: The neural systems for directing attention to spatial locations and to time intervals as revealed by both PET and fMRI. Journal of Neuroscience 1998; 18(18): 7426–7435. 80. Tracy J I, Faro S H, Mohamed F B, Pinsk M, Pinus A. Functional localization of a “time keeper” function separate from attentional resources and task strategy. Neuroimage 2000; 11(3): 228–242. 81. Constantinidis C, Williams G V, Goldman-Rakic P S. A role for inhibition in shaping the temporal flow of information in prefrontal cortex. Nature Neuroscience 2002; 5(2): 175–180. 82. Lewis P A. Finding the timer. Trends in Cognitive Sciences 2002; 6(5): 195–196.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

235

83. Braver T S, Cohen J D, Nystrom L E, Jonides J, Smith E E, Noll D C. A parametric study of prefrontal cortex involvement in human working memory. Neuroimage 1997; 5(1): 49–62. 84. Pochon J B, Levy R, Poline J B et al. The role of dorsolateral prefrontal cortex in the preparation of forthcoming actions: An fMRI study. Cerebral Cortex 2001; 11(3): 260–266. 85. Fuster J M. Unit-activity in prefrontal cortex during delayed-response performance — Neuronal correlates of transient memory. Journal of Neurophysiology 1973; 36(1): 61–78. 86. Ortuno F, Ojeda N, Arbizu J et al. Sustained attention in a counting task: Normal performance and functional neuroanatomy. Neuroimage 2002; 17(1): 411–420. 87. D’Esposito M, Ballard D, Zarahn E, Aguirre G K. The role of prefrontal cortex in sensory memory and motor preparation: An event-related fMRI study. Neuroimage 2000; 11(5): 400–408. 88. Zarahn E, Aguirre G, D’Esposito M. Replication and further studies of neural mechanisms of spatial mnemonic processing in humans. Cognitive Brain Research 2000; 9(1): 1–17. 89. Gruber O, Kleinschmidt A, Binkofski F, Steinmetz H, von Cramon D Y. Cerebral correlates of working memory for temporal information. Neuroreport 2000; 11(8): 1689–1693. 90. Fortin C, Couture E. Short-term memory and time estimation: Beyond the 2-second “critical” value. Canadian Journal of Experimental Psychology (Revue Canadienne De Psychologie Experimentale) 2002; 56(2): 120–127. 91. Neath I, Fortin C. Is the interference between memory processing and timing specific to the use of verbal material? Memory 2005; 13(3–4): 395–402. 92. Rammsayer T H. Neuropharmacological evidence for different timing mechanisms in humans. Quarterly Journal of Experimental Psychology Section B — Comparative and Physiological Psychology 1999; 52(3): 273–286. 93. Pedersen C B, Mirz F, Ovesen T et al. Cortical centres underlying auditory temporal processing in humans: A PET study. Audiology 2000; 39(1): 30–37. 94. Pouthas V, Garnero L, Ferrandez A M, Renault B. ERPs and PET analysis of time perception: Spatial and temporal brain mapping during visual discrimination tasks. Human Brain Mapping 2000; 10(2): 49–60. 95. Schubotz R I, Friederici A D, von Cramon D Y. Time perception and motor timing: A common cortical and subcortical basis revealed by fMRI. Neuroimage 2000; 11(1): 1–12. 96. Ferrandez A M, Hugueville L, Lehericy S, Poline J B, Marsault C, Pouthas V. Basal ganglia and supplementary motor area subtend duration perception: An fMRI study. Neuroimage 2003; 19(4): 1532–1544. 97. Jones C, Jahanshahi M, Dirnberger G, Frith C D. Estimation of long vs short intervals: The functional anatomy of time estimation studied with PET. Cognitive Neuroscience Society Annual Meeting 2000: 126–127 (abstract). 98. Macar F, Anton J L, Bonnet M, Vidal F. Timing functions of the supplementary motor area: An event-related fMRI study. Cognitive Brain Research 2004; 21(2): 206–215. 99. Hinton S C, Harrington D L, Binder J R, Durgerian S, Rao S M. Neural systems supporting timing and chronometric counting: An FMRI study. Cognitive Brain Research 2004; 21(2): 183–192.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

236

K. Rubia

100. Mesulam M M, Nobre A C, Kim Y H, Parrish T B, Gitelman D R. Heterogeneity of cingulate contributions to spatial attention. Neuroimage 2001; 13: 1065–1072. 101. Carter C S, Botvinick M M, Cohen J D. The contribution of the anterior cingulate cortex to executive processes in cognition. Reviews in the Neurosciences 1999; 10(1): 49–57. 102. van Veen V, Cohen J D, Botvinick M M, Stenger V A, Carter C S. Anterior cingulate cortex, conflict monitoring, and levels of processing. Neuroimage 2001; 14(6): 1302–1308. 103. Carter C S, Macdonald A M I, Stenger V A, Cohen J D. Dissociating the contributions of DLPFC and anterior cingulate to executive control: An event-related fMRI study. Brain and Cognition 2001; 47(1–2): 66–69. 104. Nichelli P, Alway D, Grafman J. Perceptual timing in cerebellar degeneration. Neuropsychologia 1996; 34(9): 863–871. 105. Jueptner M, Rijntjes M, Weiller C et al. Localization of a cerebellar timing process using pet. Neurology 1995; 45(8): 1540–1545. 106. Dupont P, Orban G A, Vogels R et al. Different perceptual tasks performed with the same visual stimulus attribute activate different regions of the human brain — A positron emission tomography study. Proceedings of the National Academy of Sciences of the United States of America 1993; 90(23): 10927–10931. 107. Ivry R B, Spencer R M, Zelaznik H N, Diedrichsen J. The cerebellum and event timing. Ann NY Acad Sci 2002; 978(1): 302–317. 108. Dreher J C, Grafman J. The roles of the cerebellum and basal ganglia in timing and error prediction. European Journal of Neuroscience 2002; 16(8): 1609–1619. 109. Aparicio P, Diedrichsen J, Ivry R B. Effects of focal basal ganglia lesions on timing and force control. Brain and Cognition 2005; 58(1): 62–74. 110. Nenadic I, Gaser C, Volz H P, Rammsayer T, Hager F, Sauer H. Processing of temporal information and the basal ganglia: New evidence from fMRI. Experimental Brain Research 2003; 148(2): 238–246. 111. Petrovici J N, Scheider G. Time experience in normal subjects and in brain-damaged patients. Fortschritte Der Neurologie Psychiatrie 1994; 62(7): 256–267. 112. Pardo J V, Fox P T, Raichle M E. Localization of a human system for sustained attention by positron emission tomography. Nature 1991; 349(6304): 61–64. 113. Cavada C, Goldmanrakic P S. Topographic segregation of corticostriatal projections from posterior parietal subdivisions in the macaque monkey. Neuroscience 1991; 42(3): 683–696. 114. Schmahmann J D, Pandya D N. Anatomical investigation of projections from thalamus to posterior parietal cortex in the Rhesus-monkey — A Wga-Hrp and fluorescent tracer study. Journal of Comparative Neurology 1990; 295(2): 299–326. 115. Bechara A, Damasio H, Tranel D, Damasio A R. Deciding advantageously before knowing the advantageous strategy. Science 1997; 275(5304): 1293–1295. 116. Bechara A, Damasio H, Tranel D, Anderson S W. Dissociation of working memory from decision making within the human prefrontal cortex. Journal of Neuroscience 1998; 18(1): 428–437.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

The Neural Correlates of Timing Functions

237

117. Bechara A, Tranel D, Damasio H. Characterization of the decision-making deficit of patients with ventromedial prefrontal cortex lesions. Brain 2000; 123: 2189–2202. 118. Tranel D, Bechara A, Denburg N L. Asymmetric functional roles of right and left ventromedial prefrontal cortices in social conduct, decision-making, and emotional processing. Cortex 2002; 38(4): 589–612. 119. Clark L, Manes F, Nagui A, Sahakian B J, Robbins T W. The contributions of lesion laterality and lesion volume to decision-making impairment following frontal lobe damage. Neuropsychologia 2003; 41(11): 1474–1483. 120. Fellows L K, Farah M J. Dissociable elements of human foresight: A role for the ventromedial frontal lobes in framing the future, but not in discounting future rewards. Neuropsychologia 2005; 43(8): 1214–1221. 121. Bolla K I, Eldreth D A, London E D et al. Orbitofrontal cortex dysfunction in abstinent cocaine abusers performing a decision-making task. Neuroimage 2003; 19(3): 1085–1094. 122. Bolla K I, Eldreth D A, Matochik J A, Cadet J L. Sex-related differences in a gambling task and its neurological correlates. Cerebral Cortex 2004; 14(11): 1226–1232. 123. Bolla K I, Eldreth D A, Matochik J A, Cadet J L. Neural substrates of faulty decisionmaking in abstinent marijuana users. Neuroimage 2005; 26(2): 480–492. 124. McClure S M, Laibson D I, Loewenstein G, Cohen J D. Separate neural systems value immediate and delayed monetary rewards. Science 2004; 306(5695): 503–507. 125. Tanaka S C, Doya K, Okada G, Ueda K, Okamoto Y, Yamawaki S. Prediction of immediate and future rewards differentially recruits cortico-basal ganglia loops. Nature Neuroscience 2004; 7(8): 887–893. 126. Rammsayer T H, Vogel W H. Pharmacological properties of the internal clock underlying time perception in humans. Neuropsychobiology 1992; 26(1–2): 71–80. 127. Rammsayer T H. On dopaminergic modulation of temporal information-processing. Biological Psychology 1993; 36(3): 209–222. 128. Rammsayer T H. Are there dissociable roles of the mesostriatal and mesolimbocortical dopamine systems on temporal information processing in humans? Neuropsychobiology 1997; 35(1): 36–45. 129. Meck W H. Neuropharmacology of timing and time perception. Cognitive Brain Research 1996; 3(3–4): 227–242. 130. de Wit H, Enggasser J L, Richards J B. Acute administration of d-amphetamine decreases impulsivity in healthy volunteers. Neuropsychopharmacology 2002; 27(5): 813–825. 131. Crean J, Richards J B, de Wit H. Effect of tryptophan depletion on impulsive behavior in men with or without a family history of alcoholism. Behavioral Brain Research 2002; 136(2): 349–357. 132. Spencer T J, Biederman J, Madras B K et al. In vivo neuroreceptor imaging in attentiondeficit/hyperactivity disorder: A focus on the dopamine transporter. Biological Psychiatry 2005; 57(11): 1293–1300. 133. Castellanos F X, Lee P P, Sharp W et al. Developmental trajectories of brain volume abnormalities in children and adolescents with attention-deficit/hyperactivity disorder. Journal of the American Medical Association 2002; 288(14): 1740–1748.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch09

238

K. Rubia

134. Rubia K, Overmeyer S, Taylor E et al. Hypofrontality in attention deficit hyperactivity disorder during higher-order motor control: A study with functional MRI. American Journal of Psychiatry 1999; 156(6): 891–896. 135. Rubia K, Smith A. Attention deficit-hyperactivity disorder: Current findings and treatment. Current Opinion in Psychiatry 2001; 14(4): 309–316. 136. Rubia K, Smith A B, Brammer M J, Toone B, Taylor E. Abnormal brain activation during inhibition and error detection in medication-naive adolescents with ADHD. American Journal of Psychiatry 2005; 162(6): 1067–1075. 137. Mathiak K, Hertrich I, Grodd W, Ackermann H. Discrimination of temporal information at the cerebellum: Functional magnetic resonance imaging of nonverbal auditory memory. Neuroimage 2004; 21(1): 154–162.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

10 The Neurology and Neuropsychology of Time-Based Prospective Memory Janet Cockburn∗

Introduction Although there is considerable anecdotal and observational support for the pervasiveness of prospective memory (ProM) impairment in neurological conditions,1,2 empirical evidence is limited and theoretical interpretation is constrained by clinical prominence given to effects on everyday functioning.3 In the light of these limitations, it is, perhaps, not entirely surprising that specific considerations of time-based ProM (ProM-T) are elusive. Moreover, comparison across studies is confounded by differences in participant aetiology and background as well as in material selection and experimental design. Nevertheless, as this chapter will show, studies of ProM in acquired neurological conditions, together with evidence from research into time perception, lend some support to hypothesized distinctions within ProM, notably between conditions selectively designed to use time- and event-based target cues, as well as between prospective and retrospective memory.

∗ University

of Reading, School for Psychology, UK; e-mail: [email protected]

239

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

240

J. Cockburn

Perception of Time and Prospective Memory: Neurological Evidence One frequently cited description of ProM is as the timely remembering of a planned action,4,5 indicating that time, in the context of timeliness, is a fundamental component. A selective form of ProM that is sensitive to time-based stimuli, is likely to be, in some way, associated with perception of time and may, therefore, be similarly affected by neurological damage to critical cortical or subcortical areas. Fuster6 argued that human activity was firmly anchored in time and that a cerebral organization capable of accessing past and present experiences, located in the frontal cortex, was integral to temporal order of planned behavior. Empirical support for this argument exists in that the ability to measure or estimate time is compromised in some neurological conditions that present with frontal impairment.7,8 Patients with selective orbito-frontal cortex (OFC) lesions have been reported to have a faster subjective sense of time than either patients with other cortical lesions (mainly in dorso-lateral prefrontal cortex (DLPFC)) or normal controls.9 This study suggests the possibility of one specific frontal cortical area being integral to accuracy of estimating or reproducing time intervals but also provides some indication that frontal areas involved in planning and monitoring action, such as the right DLPFC,10 may not be the same as those involved in perceiving or estimating the passage of time. Integrity of both areas may be necessary for efficient time-based ProM, but this possibility has not yet been explored systematically. Other research has indicated that the frontal lobes are not the sole repositories of time perception. Evidence for involvement of subcortical structures in timing operations has been found11 through comparison of temporal processing of patients with Parkinson’s disease (PD) and control participants. The relatively poorer performance of PD patients on both time perception and time production was interpreted as indicating involvement of basal ganglia and associated thalamo-cortical connections in temporal processing. Drane and colleagues12 found that, among patients with temporal lobe epilepsy, intact right hemisphere functioning enhanced performance on retrospective and prospective time estimation tasks. In particular, accuracy of estimations of intermediate temporal duration (between five minutes and one hour) was affected by amobarbital injections to the right hemisphere. Interhemispheric interaction appeared to be a prerequisite only for accurate retrospective

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

241

but not prospective judgements. They argued that the involvement of the right hemisphere in maintaining sustained attention and vigilance might underpin its role in time estimation since a continuous focus of attention would be needed to maintain awareness of the passage of time, to review and discriminate between events and to organize them into a temporal sequence in order to make accurate time judgements — functions that are also needed in successful ProM. One study has directly addressed the possibility of a relationship between time perception and prospective remembering.13 Participants in this study were long-term survivors of childhood posterior fossa tumors. A dissociation was identified between ability retrospectively to estimate short durations, which was impaired in patients, and long durations (30 and 60 minutes), at which patients were unimpaired compared to age-matched controls. They also investigated long-term prospective estimation and found that more patients than controls failed to produce the prospective estimation (telling the examiner when they believed 30 minutes of the session had elapsed) but that those who responded were no less accurate at time estimation than controls. Task failure, therefore, resulted from failure to activate the ProM component, retrieval of the intention to tell the examiner, and not from failure to maintain an accurate record of the passage of time. This result was interpreted as indicating interference from tumorrelated ProM deficits on ability to produce long-duration prospective temporal estimates. Successful time-based ProM may, therefore, depend on integration of neural mechanisms underpinning both memory and time perception. Taken together, evidence from these studies suggests that no single cortical, subcortical or cerebellar area is the sole, or even the main, repository of temporal processing functions. A similar conclusion was reached by Damasceno,14 who examined performance on eleven distinct time perception and estimation tasks by a large sample of patients with documented cerebral lesions and a control group. He noted temporal disorientation to be associated with limbic or diffuse lesions, while reproduction of duration was associated with frontal and temporal lesions. Therefore, complex neural networks may underpin the multi-faceted components of time perception, with damage in one or more of a number of areas influencing performance that may, itself, vary in demand according to specific task condition. Since ProM is also a complex activity that can be varied along a number of dimensions, a similar, provisional conclusion may be indicated. The ability to monitor ongoing time is a necessary component of maintaining the memory trace for an intended action, especially when no environmental support in the form of an overt cue is available.15 However, that alone is not sufficient

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

242

J. Cockburn

evidence for “time” to be a distinct and separable component of procedures involved in remembering prospectively.

Time as a Distinct Component of Prospective Memory Rabbitt16 argued for a distinction between “clock time” as often used in laboratory-based prospective memory (ProM) studies, in which participants are required to judge time units to an external standard, such as monitoring a clock in order to respond every two minutes,17 and the “living time” within which people constantly adjust their everyday behavior. He noted that because timekeeping entails monitoring external or internal processes during some other ongoing activity, it acts as a secondary task, placing additional demands on information-processing resources. Such resources are limited for everyone but may be especially compromised in neurological conditions, such as acquired brain injury or degenerative disorders. On that basis, the greater the need for secondary monitoring during an ongoing activity, the greater the likelihood of information-processing resources being overloaded, leading to failure of some component of the activity constellation. The ongoing need to monitor the passage of time in time-based prospective memory (ProM-T) tasks may make substantial demands on resources. In contrast, the need to identify and respond to the target event in event-based prospective memory (ProM-E) task, with which ProM-T is often contrasted,3 may make fewer demands on processing resources. However, some recent evidence18 suggests that the extent to which the need for controlled processing, as distinct from automatic, stimulus-driven processing, is imposed by the overall task conditions may be a more salient factor in ProM performance than a distinction between time- and event-based conditions.

Anecdotal and Observational Evidence: Questionnaire Studies Are distinctions between retrospective and ProM, or within ProM, influential in real life? In one of the earliest published studies to investigate self-report of everyday memory function and memory problems among adults with closed head injury,1 ProM was included as one of the six memory factors hypothesized to be identifiable in everyday behavior. Separate constructs within ProM were not specified. Moreover, analysis of responses to the 30-item questionnaire did not identify ProM as a distinct factor but linked it to attention, with items, such as “I miss appointments and meetings I’ve scheduled” loading on the same factor as items such as “I often have to ask for instructions or directions to be repeated

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

243

again.” Although respondents reported this composite attention/ProM factor as being the most problematic in their daily lives, it is not possible to extrapolate from the data that ProM is necessarily perceived as a distinct cognitive or behavioral factor in real world situations. Hannon, Adams, Harrington et al.19 developed a 52-item ProM questionnaire to explore self-ratings of young, older and brain-injured adults on a number of dimensions relating to ProM. The questions covered four types of situation, which appear to cut across a time- versus event-based dichotomy: long-term episodic, e.g. “I forgot to send a card for a birthday or anniversary”; short-term habitual, e.g. “I forgot to put a stamp on a letter before mailing it”; internally cued, e.g. “I was driving and temporally forgot where I was going”; and techniques to assist memory, e.g. “I rehearse things in my mind so I will not forget to do them.” They compared ratings across the three groups, finding a significant difference between groups only on the short-term habitual items, for which participants with brain injury reported themselves to perform significantly less well than either young or older adults. Self-report score on these items correlated weakly with performance on experimental tasks of short-term ProM. Although the short-term battery comprised two time-based and two event-based tasks, scores were combined to provide an overall measure of ProM performance and so it is not possible to identify any selective influence of ProM-T. There was no between-group difference in performance of the long-term task, which incorporated a time-based element in that participants were given two questionnaires to complete at home and return on a specified date. Accuracy of questionnaire return was equivalent across groups. Unequal group sizes (there were 114 younger adults, 27 older and 15 with brain injury) precluded more precise analysis of any relationship between effects of brain-injury and ProM self-report or performance. Smith, Della Sala, Logie & Maylor2 compared relative frequency of prospective and retrospective memory failures among older adults, with and without Alzheimer’s disease, and younger adults. Again, the questionnaire was not designed to elicit specific time-based responses but cue-types did distinguish between self-cued (e.g. “Do you decide to do something in a few minutes’ time and then forget to do it?”), which may be more amenable to a temporal target condition, and environmentally-cued (e.g. “Do you forget to buy something you planned to buy, like a birthday card, even when you see it in the shop?”). Within the Alzheimer patient group, errors in both retrospective and ProM were reported as occurring very frequently, with marginally higher incidence of ProM

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

244

J. Cockburn

errors. However, self-cued errors were reported significantly more frequently than environmentally-cued errors in responses to the retrospective but not the prospective questions. In contrast, the pattern across older and younger adult control groups indicated a substantially greater incidence of short-term selfcued errors in both prospective and retrospective memory situations. Interpretations of these differences may be complicated by ceiling effects in error ratings among Alzheimer patients, as reported by their carers, which may, in turn be influenced by the adverse impact of patients’ ProM failures on the everyday lives of their carers. Although not directly addressed in the study, the incidence of reported memory lapses in situations where self-cueing is required may present indirect support for effects of similar information-processing demands in tasks designed to elicit time-based processing. However, self-cued tasks are among the most difficult to simulate under experimental conditions, even if participants are encouraged to devise their own cues to facilitate timely retrieval, which presents difficulties for empirical testing. Groot, Wilson et al.20 examined responses to three questionnaires in association with the development of a test of ProM (described later in this chapter): The Everyday Memory Questionnaire (EMQ)21; the Cognitive Failures Questionnaire (CFQ)22 ; and the Dysexecutive Questionnaire (DEX).23 Their aim was to establish the cognitive and behavioral validity of items in a new test of ProM, not to identify specific components of the construct. Although none of the chosen questionnaires directly examines self- or other-reported ProM performance, all measure attributes associated with ProM and all contain items germane to understanding of one’s own ProM functioning in real world situations. Correlations between participants’ and carers’/others’ ratings on the questionnaires and performance on ProM tasks were low, on the whole, and did not differentiate between groups. Correlations with ProM tasks were stronger for EMQ and DEX than for CFQ, suggesting that memory and executive function may be more influential than more general cognitive factors. Extrapolating from arguments that ProM-T places greater demands on information processing resources than ProM-E,16 ProM-T tasks might be associated more strongly with executive questions of DEX and ProM-E tasks with memory questions of EMQ. However, since correlations were not examined separately in relation to ProM-T and ProM-E tasks, this remains speculation. None of the questionnaire studies reviewed above has specifically distinguished a ProM-T component. Therefore, although they present anecdotal and self-report indications that difficulties in remembering to perform future

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

245

activities, especially those with a substantial self-cuing requirement, cause considerable distress to individuals with neurological damage and their carers (cf. Mateer, Sohlberg & Crinean1 ), there is no direct evidence that everyday situations in which time is a distinct factor of a to-be-performed task, cause any more or less distress or result in more or less frequent failure than situations where, for example, ProM tasks are event-based.

Prospective Memory in Neurological Conditions In comparison with studies on normal ageing, relatively little research has been carried out on ProM among individuals with acquired brain injury or other neurological conditions3,24 . The evidence that has been published tends to suggest significant impairments of ProM, both in comparison to retrospective memory and relative to performance of control groups. Although there is little indication of specific theoretical consideration of distinct constituents of ProM, a number of neural processes have been identified as underpinning the general construct. These include medial temporal and diencephalic structures that support retrospective memory25 and also have connections with temporal organization.26 Relationships have been postulated between frontal lobes and aspects of memory function, especially those associated with planning, initiating and monitoring actions.27 Shimamura, Janowsky & Squire28 noted that memory disorders associated with frontal-lobe damage include memory for temporal order and source memory. They argued for distinct roles of frontal and medial temporaldiencephalic structures in memory, with the frontal lobes acting as the “working” memory, manipulating and organizing material for storage and subsequent access, and medial temporal lobes and diencephalic regions being essential for effective storage so that material can be retrieved as and when needed. They identified the memory disorders observed in patients with frontal-lobe lesions as denoting deficits in ProM, which they associated with the dysexecutive syndrome and with symptoms of disinhibition, in terms of shared characteristics, such as failure to access and utilize temporal or spatial features of context in which information is set but did not present empirical research to support their assumptions. Although these arguments do not discriminate directly within ProM, it is likely that successful performance of ProM-T tasks places greater reliance on self-initiation of access to stored information and utilization of temporal context identification than is needed for ProM-E tasks, but further research within clearly defined parameters is needed.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

246

J. Cockburn

Frontal areas that monitor or estimate the passage of time may be activated in ProM-T but evidence is, as yet, limited. Glisky29 speculated that ProM-T relies on the integrity of a range of functions: memory for temporal information; time estimation abilities; strategic planning; environmental monitoring; divided attention; all of which are mediated by some component part of the frontal cortex. She considered that the cognitive processes on which ProM relies depend differentially on task demands, which may themselves depend on different underlying brain structures. At least some of the processes would be likely to depend on the integrity of the frontal lobes but characteristics of individual tasks would dictate the extent and nature of frontal cortex involvement. Speculation that the frontal lobes act in concert with other cortical and subcortical circuits to subsume executive function has been advanced by Andres & Van der Linden.30 They presented evidence that questions the notion of the prefrontal cortex alone as the repository of the central executive and proposed a series of networks mediating central executive function. The prospective components of ProM may be differentially associated with specific areas of the frontal lobes as well as with other areas of the cerebral cortex. Burgess and colleagues31 have used functional imaging techniques to investigate the hypothesis that Brodman area 10, rostral prefrontal cortex, is activated in ProM and that medial and lateral aspects may be selectively sensitive to different processing demands.

Empirical Evidence Results from studies of ProM in neurological conditions are, in general, difficult to interpret in terms of specific components because they vary widely in purpose, design and population sampled, as well as in underlying theoretical assumptions. A number of studies have assumed some relationship between impaired executive function, and/or frontal lobe lesions and deficits in ProM after acquired brain injury. Kliegel and colleagues32 attempted to impose a theoretical framework on processes involved in the realization of delayed intentions (ProM), which they broke down into four consecutive components: intention formation; retention; re-instantiation; execution. They used this categorization to examine for similarities between ProM performance hypothesized to be associated with effects of executive function decline after traumatic brain injury (TBI) and ProM deficits seen in healthy older adults, which might be explained in terms of age-related executive dysfunction. They tested their predictions in a group of young adults with acquired brain injury who had normal retrospective memory

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

247

and mildly impaired executive functions, a group of healthy older and a group of healthy younger adults, using a complex ProM task that allowed separate assessment of performance on each of the four postulated phases. Although the study did not directly address ProM-T as a distinct construct, the task, which was broadly similar to the Six Elements Task,27 has a time-based component: participants had to devise and follow their own time-based schedule to complete the six tasks (execution phase). Results indicated that all groups performed well on the intention retention phase, but patients with executive function deficits and older participants performed less accurately than younger adults in the intention formation, re-instantiation and execution phases. There were no significant differences between patient and older adult groups. On that basis, they infer that executive skills are needed more for some phases of a ProM task than others, and that older adults and individuals after TBI may have particular difficulties in these phases. Although this study does not identify a precise function for ProMT, the execution phase, especially self-initiated task switching, may represent time-based conditions. Cockburn33 did address the possibility that different cognitive processes and different neuroanatomical structures underly distinct ProM components. She described a single case of a patient who had CT evidence of bilateral medial frontal infarcts following a sub-arachnoid haemorrhage (SAH) but no other area of infarction. Verbal recall memory was well preserved but recognition memory was poor and there were marked deficits in executive function, notably initiation and inhibition of an ongoing activity. She was tested extensively and showed consistent impairment on time- but not event-based tasks over a sixmonth period. Cockburn argued that the pattern of performance was consistent with a further division within self-initiated ProM tasks such that those where the background task needed to be interrupted in order to perform the ProM task represented the most demanding level of self-initiation combined with inhibition, whereas those with the prospective elements compatible with performance of the background activity may be more “data-driven” and so make fewer demands on processing resources. This may suggest that time is only one component in a hierarchy of difficulty in ProM tasks. A later group study34 attempted to clarify these divisions. Although results reinforced the hypothesized association between ProM-E and retrospective episodic memory, they failed to support arguments for executive function involvement in ProM-T. A larger sample might have yielded a stronger association with performance on tests of executive function since subsequent

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

248

J. Cockburn

calculation of relative effect size for these data18 identified greater impairment of ProM-T (r = 0.51) than ProM-E (r = 0.42), which suggests some cognitive construct other than episodic memory is involved in ProM-T. When Katai and colleagues35 compared performance of patients with Parkinson’s disease (PD) and healthy controls on event- and time-based ProM, using a similar design to that used by Einstein and colleagues with older adults,36,37 they found patients to be impaired on the event- but not the timebased task. The specific failure of the patients was inability to retrieve the content of the event-based instruction when the target cue word appeared before them. This was assumed not to be a memory failure because instructions were recalled successfully at the end of the task but was explained in terms of impoverished self-initiation associated with PD related frontal dysfunction. Contrary to the predictions of Glisky,29 patients appeared more able to initiate the required response in the time-based condition. Because this, however, only occurred twice in contrast to the four targets in the event-based condition, the tasks were not fully comparable in terms of task difficulty. It is possible to speculate that the design of the time-based task, requiring participants to monitor a clock placed behind them and respond at each of two times, permitted the establishment of a habitual action, thus reducing the need for self-initiation. Frequency of clock checking did not differ significantly between patients and controls, with both groups increasing responses in the minutes immediately preceding a target time. This behavior suggests that the ProM instruction was active to a greater or lesser extent throughout the 17 minutes of the task, thus potentially remaining in working memory. Katai et al. likened the process to the Test-Wait-Test-Exit (TWTE) model of Harris and Wilkins38 in which the loop through test-wait cycles is more or less continuous until the test is made during a critical period when “exit” is activated. In contrast, the event-based instruction to respond to each occurrence of a target word in a series of words does not require a plan to be held in ongoing activation and so may have led to the intention being submerged in the processes of the background task and having to be retrieved from long-term memory at the appropriate moment. Whereas it can be argued that, in the real world, remembering to do something at a predetermined time requires planning and organization, establishing a repetitive, sequential action in response to an arbitrary instruction, may place less demand on self-initiation. Evidence for involvement of specific frontal areas was not presented for this group, but the authors suggest that ProM components of time- and event-based tasks may be mediated by different neural

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

249

networks, with dissociation in PD being attributable to selective impairment of critical networks. Dopamine depletion in the head of the caudate nucleus and the frontal cortex, and also neuronal loss in the nucleus basalis of Meynert, which innervates the basal forebrain, have been noted to occur commonly in PD.39 One may, however, hypothesize that frontal areas associated with time estimation, such as ventro-medial prefrontal cortex, may be less affected in PD than dorso-lateral prefrontal cortex, which is implicated in multi-tasking and other executive functions.40 The patients in this study did not appear to show similar deficits in time perception or production to the PD patients reported by Harrington et al.,11 who postulated an association with basal ganglia and thalamo-cortical dysfunction, suggesting that relations between structure and function, within as well as across neurological conditions, are highly complex and merit further careful investigation. Kinsella and colleagues41 examined everyday memory after TBI, by investigating interrelationships between subjective memory reports, performance on traditional tests of retrospective memory and on tests of ProM based on those used in the Rivermead Behavioral Memory Test.42 Time-based ProM was not identified as a specific dimension of investigation but was implicitly addressed in the tasks devised. Participants had spontaneously to request a questionnaire at the end of the testing session and remember its purpose. They also had to return an evaluation form by mail after the session. Patients were significantly more likely than controls to forget to ask for the questionnaire, but no less likely to return the form on time. However, the difference did approach significance and Kinsella et al. suggested that motivation might have been a confounding factor. Perceived importance of a task has been found to have greater impact on ProM-T than ProM-E among undergraduates,17 but this has not yet been specifically addressed in a neurological population. Hypothesizing that the documented dysfunctions of executive and retrospective memory skills among people with multiple sclerosis (MS) would indicate vulnerability to performance of delayed intentions (ProM), Bravin and colleagues43 assessed retrospective memory, executive skills and performance on two experimental, “naturalistic” tests of ProM in a group of MS patients and an age-matched control group. Both tests were broadly time-based: one required the participant to remind the experimenter to repeat a pulse reading after one hour; the other to interrupt an ongoing clerical task after five minutes in order to perform another activity. Relative to controls, patients were more likely to fail on the retrospective memory component of the long delay task (remembering

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

250

J. Cockburn

what they had to say) than the prospective (remembering the appropriate time to give the reminder), but the two groups did not differ significantly on the shortinterval task. Expressing surprise that their results did not appear to demonstrate ProM failure associated with executive dysfunction, the authors, however, commented that the visual presence of a clock in the room might have acted as an external visual cue, thus translating the retrieval context of the ProM task from self-initiation into reliance on cued recall. Participants may have set up a TWTE strategy, similar to that hypothesized by Katai et al.35 in their study of PD patients. Comparison of results from these two studies suggests that the design of the test components, the context and interpretation of errors may all exert an overriding influence on the outcome of the investigation. One of the few studies to have attempted to equate task demands of ProM-T and ProM-E tasks is described by Shum and colleagues,44 who criticised earlier research for treating ProM as a unitary construct and for failing to present sufficient items to enable adequate discrimination between ProM-E and ProMT performance. Their overall aim was to examine effects of severe TBI on hypothesized subtypes of ProM, using a large enough range of response opportunities to provide a reliable differentiation between performance in ProM-T and ProM-E situations. In addition, assuming an impairment on ProM-T, they examined the relationship between time-monitoring behavior and accuracy of ProM responses, predicting that control participants would demonstrate a more strategic monitoring pattern of the passage of time between crucial target points than would patients. The study was carefully designed so that time- and eventbased response targets were embedded into matched multiple-choice general knowledge filler tasks and yielded the same number of response opportunities, five in each case, at approximately equivalent intervals throughout the filler task. Results indicated that, as predicted, the percentage of correct responses (within 20 seconds of the target time or appearance of the event cue) was lower for the time-based than the event-based task, for both patients and age-matched controls. However, there was no significant group by task interaction. Since none of the participants failed all response opportunities and all were able to recall the task instructions at the end of the study, Shum et al. hypothesized that incorrect responses were due more to difficulties in remembering to perform an action (ProM) than to retrospective memory failure to recall task contents. One difference between the two task conditions was revealed by analysis of types of failure, with more omissions being recorded for ProM-E tasks and more late responses for ProM-T tasks. This distinction was thought, though, to be a

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

251

function of task design, in that the target times in the ProM-T task were every five minutes and the clock reset itself after each five-minute period. Therefore, participants would be made aware of a failure when they next checked the clock. In contrast, they would not know if they had failed to respond to an appearance of the target event cue (the words “Prime Minister” in a filler question). Contrary to prediction, patients demonstrated a pattern of time monitoring over each five-minute period that did not differ statistically from that of controls — both groups monitoring more frequently in the final minute. Additionally, greater frequency of monitoring in the fifth minute among patients was associated with more timely responding. This finding was interpreted as indicating that poorer performance of TBI patients on a time-based task could not be explained by inadequate time monitoring, but might be associated with poorer ability at estimating the passage of time, a function thought to be mediated primarily within the frontal lobes.9,14 Patients in Shum et al.’s study had all sustained a severe head injury and were assumed to have frontal and temporal lobe damage but no specific lesion data were presented. Experimental design of a number of comparative investigations of ProM-T and ProM-E has been criticized by Carlesimo, Casadio & Caltagaroni.15 These authors noted that previous researchers had failed to distinguish between prospective and retrospective components of the to-be-performed task, thus impeding interpretation of the association between episodic and ProM skills. In an attempt to resolve both this confound and the controversy surrounding relative difficulty of ProM-E and ProM-T tasks after closed head injury, Carlesimo et al. designed their study so that the forgetting rate of prospective (remembering the intention to perform) and retrospective (remembering the sequence of actions to be performed) components could be separately measured within the same basic task format. Environmental support was also varied, being absent in the time-based condition and present, in the form of an audible time signal, in the event-based condition. Time monitoring throughout the retention intervals was recorded and showed that, in contrast to the pattern demonstrated in the Shum et al.44 study, which also recruited patients late after severe TBI, normal control participants not only monitored more frequently but also more strategically, increasing their checks towards the end of the retention period, whereas the patients maintained a more or less steady but low incidence of monitoring. Time monitoring was associated with task completion accuracy, with controls being more accurate overall than patients, and task accuracy within the patient group being associated with frequency of monitoring. On the whole, patients

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

252

J. Cockburn

appeared to perform less accurately on the time-based tasks, with poorest performance overall on time-based actions in the short delay condition. In contrast, they recorded higher levels of accuracy on functionally unrelated time-based actions than on functionally unrelated event-based actions in the long delay condition, hypothesized to be the most demanding of the task conditions. The correlation between episodic memory score and ProM performance was low suggesting that memory loss per se was a less likely explanation of poor ProM than the inability to maintain an active memory representation sufficiently long to perform the action sequence when required. Results from this study reinforce suggestions that, although timeliness may be an important component in successful ProM, it is likely to be only one of several influential factors. The development of a dedicated test of ProM20,45 addresses a number of the criticisms voiced by Shum et al.44 of limited utility and generalizability of ProM assessments in comparison with those developed to test retrospective memory. In the report by Groot et al.,20 a heterogeneous sample of people with acquired brain injury and a similar-sized control group were tested on a battery of four event-based and four time-based ProM tasks incorporated into a set of cognitively demanding filler activities. They were also given a range of standard neuropsychological tests of attention, memory, executive function and speed of processing. Two of the study’s aims directly address the themes of this chapter: investigation of neuropsychological mechanisms associated with ProM functioning; comparison of difficulty of time- and event-based ProM tasks. Results indicated that ProM performance was associated with cognitive functioning across all domains sampled rather than any one individually, but was poorer, overall, for patients than for controls. All participants, controls and patients, had more difficulty with ProM-T than ProM-E, which the authors interpreted in terms of ProM-T tasks making greater demands on inhibitory control mechanisms. Although note-taking improved the performance of patients, there is no evidence that it improved ProM-T more or less than ProM-E. These findings have since been replicated in an updated version of the test, using three instead of four of each type of ProM task.45 Whereas other studies have tended to assess ProM-T and ProM-E separately, here tasks were interleaved and so participants not only had to remember, with or without self-initiated aids, the intention and content components of each of the eight tasks, but they also had to activate and monitor a cancel function46 at appropriate intervals, thus adding to the overall cognitive processing load, which may have contributed to the relatively higher error rates on the ProM-T than on the ProM-E components of the battery. However, although the study design

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

253

allowed computation of a mathematical equivalence between scores on the two sets of tasks, it is harder to ensure functional comparability of attentional load between tasks with different demands. For example, a score of 1, denoting success at reminding the tester after 15 minutes not to forget her key (ProM-T), may not be fully comparable across all parameters of task difficulty with a score of 1 for remembering to change pens after completing seven filler tasks (ProM-E).

Evidence from “Everyday” Actions Attempts to examine ProM performance in a more realistic framework have been made by creating virtual environments in which participants can be assessed on ProM tasks that require planning and organizing skills.47,48 The basic task condition was a virtual bungalow on moving day. Participants were given a series of tasks to complete, incorporating time- and event-based activities, which were connected to labelling and preparing items for removal. These studies present data demonstrating impairments of time- and activity-based but not event-based ProM in patients with prefrontal lesions48 and lesser impairments in time than event-based tasks among stroke patients.47 No details are given of side or location of lesion among the stroke patients but it is possible that few had a prefrontal lesion. Moreover, although they were more impaired at the ProM tasks than on an immediate free recall task, 39% of them failed to recall all ProM task instructions immediately on completion of the overall activity, raising the possibility that they had some stroke-related long-term episodic storage and/or retrieval impairment. Brooks et al., however, noted that the time-based ProM task, to ask the experimenter to click on a button by the clock at five-minute intervals, might not be a valid simulation of a real time-based task because the interval between retrieval occasions was sufficiently brief that participants could maintain it in working memory, analogous to the task conditions presented by Katai et al.35 Nevertheless, in contrast to the stroke patients, control participants in this study did perform relatively less well on the time-based activity than on the event-based task of labelling certain items as “fragile.” This dissociation between stroke patients and controls possibly indicates that, when episodic memory is unimpaired, monitoring time makes the greater demand on processing resources. The frontal lobe patients reported by Morris et al.48 were also significantly late in responding as a group, implying that these patients did not set up a TWTE strategy despite having a clock constantly visible. Morris et al., noted that the requirement for internal generation of the appropriate activity, together with maintenance of continuous time monitoring, in the virtual

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

254

J. Cockburn

task mirrors everyday activities in which frontal patients are seen to fail. The paper, however, did not report everyday functional abilities of these patients, nor whether they were impaired on standard executive tests. Comparison between these two studies highlights the differences in evidence obtained from different neurological groups when task conditions are essentially the same and reinforces the need for comprehensive neuroanatomical and demographic information to be included in analyses of results. Another line of approach to examining structure and composition of ProM is to focus on its role among cognitive processes that underpin activities of daily living (ADL). The respective roles of ProM, strategic planning and working memory on ADL, were examined in patients with frontal lesions following head injury.49 Utilizing a script task that simulated aspects of preparing and presenting a meal, the authors hypothesized a division into macrostructure, representing the script as a whole and involving strategic planning and ProM, and microstructure, representing overlearnt sub-sequences of the script. Closed head-injured participants made significantly more errors than the control group on macrostructure components of serving the meal on time and with acceptable intervals between courses. This stage of the script sequence was identified as depending on ProM, and specifically time-based ProM, in that success was determined as timely sequential delivery of each separate course. However, the authors acknowledge that, without an independent measure of ProM, their interpretation can only be an assumption. Moreover, failure at other stages of the overall task appeared to be more closely related to strategic planning skills, and the correlation between success on the two stages was low, suggesting that planning actions and carrying them out effectively at the appropriate time are two distinct components of ADL, possibly associated with different cortical areas.10 Since there were no significant differences between patients and controls on standard neuropsychological tests of executive or memory function, the difficulties experienced by the patients in carrying out this simulation of a familiar task were argued to be an important contribution to interpreting and understanding impairments of everyday functioning in closed head injury patients.

Is There a Unique Role for Time in Prospective Memory? Martin and Schumann-Hengsteler50 found that performance of older adults on a ProM-T task was disproportionately affected by cognitive load of background task compared with younger adults. They interpreted poorer performance of

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

255

older adults on ProM-T tasks in terms of an age-by-task interaction. Older adults, having fewer cognitive resources, in particular a reduction in efficacy of working memory, were disproportionately penalized by the large amounts of cognitive load imposed by conditions in a time-based task. Since acquired brain injury often results in reduction of cognitive resources, it is plausible that performance on time-based ProM tasks would be similarly compromised. A recent meta-analysis18 supports the view that time-based ProM generally imposes greater demands than event-based on self-initiated retrieval processes. Eleven studies were identified in which performance of TBI patients and control participants on objective tests of ProM was compared. Calculations of effect sizes indicated ProM problems after TBI even when task demands were relatively minor. In the five studies that reported measures of ProM-T and ProM-E, average effect sizes, of r = 0.55 and r = 0.37 respectively, supported suggestions that ProM-T would be more impaired after TBI than ProM-E. Nevertheless, the wide variety of tasks and conditions utilized in the studies indicate that other factors may have as much, if not more, influence on resultant performance and, therefore, need to be considered.

Conclusions The distinction between ProM-E and ProM-T hypothesized in relation to studies of normal ageing (e.g. Einstein & McDaniel36 ) has often been adopted as a workable construct for investigating prospective memory failures in neurological conditions.3 To date, the balance of evidence appears to suggest that neurological patients may respond differently in ProM-T and ProM-E tasks (Table 1), with most findings indicating greater difficulty for ProM-T. Studies of patients with distinct neurological conditions may, therefore, yield evidence to support the case for time-based ProM having a distinct identity. The hypothesis that time-based tasks may be more dependent on intact frontal or executive function receives tangential support from findings of relatively greater impairment in event-based tasks among stroke patients,47 in whom substantial frontal damage may occur less frequently, and in patients with PD, in whom frontal dysfunction was hypothesized to affect retention of an intention rather than activation of its execution.35 However, much of the evidence accrued is equivocal because of design limitations of the materials and task conditions selected. Other factors, including task design, background conditions, patient aetiology and number and type of cues, all affect performance on experimental tasks and

Comparison Group

Cockburn34

18 mixed aetiology 18 age-matched brain injury controls

2 Between subjects-group. Within subjects-task

3

Patients < controls in both. Patients less accurate on ProM-T than ProM-E.

Shum et al.44

12 severe TBI patients

5 Between subjects-group. Within subjects-task

5

TBI < controls on ProM-E and ProM-T. Both groups < ProM-T than ProM-E.

Groot et al.20

36 mixed aetiology 28 controls brain injury

4 Between subjects-group. Within subjects-task

4

Between subjects-group

3 tasks, max score = 18

Patients < controls on ProM overall. Both groups < ProM-T than ProM-E. Overall, patients < controls. Trend for patients < ProM-T than ProM-E.

12 controls with no TBI or neurological problems

Wilson et al.45 72 mixed aetiology 214 controls brain injury

Design

Number of ProM-E Tasks/ Scoring Opportunities

3 tasks, max score = 18

Number of Results ProM-T Tasks/ Scoring Opportunities

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

Patient Sample

J. Cockburn

Author

256

Table 1. Studies that have published separate assessments of performance on tasks tapping ProM-T and ProM-E.

Comparison Group

Design

Morris et al.48 35 patients with prefrontal lesions

35 age-matched controls

Proportion of Difference Between subjects-group. Within “fragile” item between target time to labelled subjects-task open door and time door reached

Patients < controls on ProM-T and ProM-E. Relatively greater decrement for ProM-T.

Brooks et al.47 36 stroke patients

22 age-matched controls

3–6 retrieval Between subjects-group. Within occasions subjects-task

Overall, patients < controls. Patients < controls for ProM-E but only marginally < for ProM-T.

Katai et al.35

Patient Sample

20 patients with PD 20 age-matched controls

Number of ProM-E Tasks/ Scoring Opportunities

Number of Results ProM-T Tasks/ Scoring Opportunities

3–6 retrieval occasions

4 (each 2 (each PD patients < Between subjects-group. Within scored 1 or 0) scored 2 or 0) controls on ProM-E but NS difference subjects-task on ProM-T.

The Neurology and Neuropsychology of Time-Based Prospective Memory

Author

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

Table 1. (Continued)

257

Author

Patient Sample

Design

Number of ProM-E Tasks/ Scoring Opportunities

Number of Results ProM-T Tasks/ Scoring Opportunities

16 age-matched controls

Non-parametric analysis: Dependent variables — recall of intention; recall of to-be-performed actions. Between subjects-group. Within subjects-retention interval; task type (ProM-T or ProM-E)

8 action triplets across 2 time intervals

8 action triplets across 2 time intervals

Overall, patients < controls. Patients ProM-T < ProM-E for shorter time interval and functionally related actions. ProM-E < ProM-T for long-time interval functionally unrelated actions.

J. Cockburn

Carlesimo et al.15 16 severe CHI patients

Comparison Group

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

258

Table 1. (Continued)

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

259

these variables may cut across the time/event distinction, thus highlighting the difficulty in ascertaining the strength or even the existence of a specific construct of time-based ProM. Even in studies that have set out to compare performance on tasks tapping ProM-T and Prom-E,15,35,44 differences in results seem as likely to reflect differences in overall task construction as different influences of timeor event-based cue conditions. Unless tasks can be constructed so that failure is attributable solely to a time-based component, we cannot be sure that such an entity exists, nor will it be possible to identify salient distinguishing factors. Current evidence35,43,47 suggests there is some way to go in developing a reliable design for an analog of everyday ProM activity that necessitates a self-initiated time-based target for which the cue is not maintained in working memory. At the present time, evidence from neurology and neuropsychology offers hints that time-based ProM may be a distinct entity and be grounded in a perception of time but reaches no firm conclusion. There is much need for studies that marry precise identification of lesion location with cognitive abilities, especially in memory and executive function, and with carefully structured tasks of time perception and ProM. Such studies might bring us nearer to determining whether or not there is anything special about time-based ProM.

References 1. Mateer C A, Sohlberg M M, Crinean J. Focus on clinical research: Perceptions of memory function in individuals with closed-head injury. Journal of Head Trauma Rehabilitation 1987; 2: 74–84. 2. Smith G, Della Sala S, Logie R H, Maylor E A. Prospective and retrospective memory in normal ageing and dementia: A questionnaire study. Memory 2000; 8(5): 311–321. 3. Shum D, Fleming J, Neulinger K. Prospective memory and traumatic brain injury: A review. Brain Impairment 2002; 3(1): 1–16. 4. Winograd E. Some observations on prospective memory. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory. Chichester UK: Wiley, 1988: 348–383. 5. Kvavilashvili L. Remembering intention as a distinct form of memory. British Journal of Psychology 1987; 78: 507–518. 6. Fuster J M. Temporal organization of behavior. Human Neurobiology 1985; 4: 57–60. 7. Harrington D L, Haaland K Y, Knight R T. Cortical networks underlying mechanisms of time perception. J Neurosci 1998(a); 18(3): 1085–1095. 8. Harrington D L, Haaland K Y. Neural underpinnings of temporal processing: A review of focal lesion, pharmacological, and functional imaging research. Rev Neurosci 1999; 10(2): 91–116. 9. Berlin H A, Rolls E T, Kischka U. Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions. Brain 2004; 127(Pt 5): 1108–1126.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

260

J. Cockburn

10. Burgess P W, Veitch E, de Lacy Costello A, Shallice T. The cognitive and neuroanatomical correlates of multitasking. Neuropsychologia 2000; 38(6): 848–863. 11. Harrington D L, Haaland K Y, Hermanowicz N. Temporal processing in the basal ganglia. Neuropsychology 1998(b); 12(1): 3–12. 12. Drane D L, Lee G P, Loring D W, Meador K J. Time perception following unilateral amobarbital injection in patients with temporal lobe epilepsy. J Clin Exp Neuropsychol 1999; 21(3): 385–396. 13. Hetherington R, Dennis M, Spiegler B. Perception and estimation of time in long-term survivors of childhood posterior fossa tumors. J Int Neuropsychol Soc 2000; 6(6): 682–692. 14. Damasceno B P. Time perception as a complex functional system: Neuropsychological approach. Int J Neurosci 1996; 85(3–4): 237–262. 15. Carlesimo G A, Casadio P, Caltagirone C. Prospective and retrospective components in the memory for actions to be performed in patients with severe closed-head injury. J Int Neuropsychol Soc 2004; 10(5): 679–688. 16. Rabbitt P M A. Commentary: Why are studies of “prospective memory” planless? In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, N.J.: Lawrence Erlbaum Associates, 1996: 239–248. 17. Kliegel M, Martin M, McDaniel M A, Einstein G O. Varying the importance of a prospective memory task: Differential effects across time- and event-based prospective memory. Memory 2001; 9(1): 1–11. 18. Henry J D, Phillips L H, Crawford J R, Kliegel M, Theodorou G, Summers F. Traumatic brain injury and prospective memory: The influence of task complexity. Submitted. 19. Hannon R P A, Harrington S, Fries Das C, Gipson M T. Effects of brain injury and age on prospective memory self-rating and performance. Rehabilitation Psychology 1995; 40: 289–298. 20. Groot Y C, Wilson B A, Evans J, Watson P. Prospective memory functioning in people with and without brain injury. J Int Neuropsychol Soc 2002; 8(5): 645–654. 21. Sunderland A, Harris J E, Baddeley A D. Do laboratory tests predict everyday memory? A neuropsychological study. Journal of Verbal Learning and Verbal Behavior 1983; 22: 341–357. 22. Broadbent D E P C, Fitzgerald P, Parkes K R. The Cognitive Failures Questionnaire (CFQ) and its correlates. British Journal of Clinical Psychology 1982; 21: 1–18. 23. Burgess P W, Alderman N, Evans J, Emslie H, Wilson B A. The ecological validity of tests of executive function. J Int Neuropsychol Soc 1998; 4(6): 547–558. 24. Levin H S, Hanten G. Post-trauamtic amnesia and residual memory deficit after closed head injury. In Baddeley A, Wilson B A, Kopelman M, eds. A Handbook of Memory Disorders. 2nd edn. Chichester UK: Wiley, 2002: 381–411. 25. Squire L. Memory and Brain. New York: OUP, 1987. 26. Bisiacchi P. The neuropsychological approach in the study of prospective memory. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah NJ: Lawrence Erlbaum Associates, 1996: 297–318. 27. Shallice T, Burgess P W. Deficits in strategy application following frontal lobe damage in man. Brain 1991; 114(Pt 2): 727–741.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

The Neurology and Neuropsychology of Time-Based Prospective Memory

261

28. Shimamura A P, Janowsky J S, Squire L R. What is the role of frontal lobe damage in memory disorders? In Levin H S, Eisenberg H M, Benton A L, eds. Frontal Lobe Function and Dysfunction. New York: OUP, 1991. 29. Glisky E. Prospective memory and the frontal lobes. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah NJ: Lawrence Erlbaum Associates, 1996: 249–266. 30. Andres P. Supervisory attentional system in patients with focal frontal lesions. J Clin Exp Neuropsychol 2001; 23(2): 225–239. 31. Burgess P W, Simons J S, Dumontheil I, Gilbert S J. The gateway hypothesis of rostral prefrontal cortex (area 10) function. In Duncan J, Phillips L, McLeod P, eds. Measuring the Mind: Speed, Control and Age. Oxford, UK.: Oxford University Press, 2005: 217–248. 32. Kliegel M, Eschen A, Thone-Otto A I. Planning and realization of complex intentions in traumatic brain injury and normal aging. Brain Cogn 2004; 56(1): 43–54. 33. Cockburn J. Task interruption in prospective memory: A frontal lobe function? Cortex 1995; 31(1): 87–97. 34. Cockburn J. Failure of prospective memory after acquired brain damage: Preliminary investigation and suggestions for future directions. J Clin Exp Neuropsychol 1996; 18(2): 304–309. 35. Katai S, Maruyama T, Hashimoto T, Ikeda S. Event based and time based prospective memory in Parkinson’s disease. J Neurol Neurosurg Psychiatry 2003; 74(6): 704–709. 36. Einstein G O, McDaniel M A. Normal aging and prospective memory. J Exp Psychol Learn Mem Cogn 1990; 16(4): 717–726. 37. Einstein G O, McDaniel M A, Richardson S L, Guynn M J, Cunfer A R. Aging and prospective memory: Examining the influences of self-initiated retrieval processes. J Exp Psychol Learn Mem Cogn 1995; 21(4): 996–1007. 38. Harris J E, Wilkins A J. Remembering to do things: A theoretical framework and illustrative experiment. Human Learning 1982; 1: 1–14. 39. Brandt J, Munro C A. Memory disorders in subcortical dementia. In Baddeley A D, Wilson B A, Kopelman M, eds. A Handbook of Memory Disorders. 2nd edn. Chichester UK: Wiley, 2002: 591–614. 40. Burgess P W. Strategy application disorder: The role of the frontal lobes in human multitasking. Psychol Res 2000; 63(3–4): 279–288. 41. Kinsella G, Murtagh D, Landry A et al. Everyday memory following traumatic brain injury. Brain Inj 1996; 10(7): 499–507. 42. Wilson B A, Cockburn J, Baddeley A D. The Rivermead Behavioral Memory Test. Bury St. Edmunds UK: Thames Valley Test Co., 1985. 43. Bravin J H, Kinsella G J, Ong B, Vowels L. A study of performance of delayed intentions in multiple sclerosis. J Clin Exp Neuropsychol 2000; 22(3): 418–429. 44. Shum D, Valentine M, Cutmore T. Performance of individuals with severe long-term traumatic brain injury on time-, event-, and activity-based prospective memory tasks. J Clin Exp Neuropsychol 1999; 21(1): 49–58. 45. Wilson B A, Emslie H, Foley J, Shiel A, Watson P, Hawkins K, Groot Y, Evans J J. The Cambridge Prospective Memory Test (CAMPROMPT ). Oxford UK: Harcourt Assessment, The Psychological Corporation, 2005.

April 7, 2006 13:27 WSPC/SPI-B378 Timing the Future ch10

262

J. Cockburn

46. Ellis J A. Prospective memory or the realisation of delayed intentions. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah NJ: Lawrence Erlbaum Associates, 1996: 1–22. 47. Brooks B M, Rose F D, Potter J, Attree E A, Jayawardena S, Morling A. Assessing stroke patients’ ability to remember to perform actions in the future using virtual reality. In Proceedings of the 4th International Conference on Disability, Virtual Reality and Associative Technology. 2002, Veszprem, Hungary: 239–245. 48. Morris R G, Kotista M, Bramham J, Brooks B, Rose F D. Virtual reality investigation of strategy formation, rule breaking and prospective memory in patients with focal prefrontal neurosurgical lesions. In Proceedings of the 4th International Conference on Disability, Virtual Reality and Associative Technologies. 2002, Veszprem, Hungary: 101–108. 49. Fortin S, Godbout L, Braun C M. Cognitive structure of executive deficits in frontally lesioned head trauma patients performing activities of daily living. Cortex 2003; 39(2): 273–291. 50. Martin M, Schumann-Hengsteler, R. How task demands influence time-based prospective memory performance in young and older adults. International Journal of Behavioral Development 2001; 25: 386–391.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

11 What it Takes to Remember the Future Joseph Glicksohn∗ and Michael S. Myslobodsky†

Introduction Memory is considered prospective when this refers to an intention, achievement of which is kept in mind until finally accomplished, with no additional prompts.1 Admittedly, this definition may seem much too relaxed and general to be of practical use, since all intentions are retrospectively mnemonic in nature.2 Yet having good retrospective memory, say for a list of words, may not predict an equally good performance for some prospective memory (ProM) tasks, as was shown by Wilkins and Baddeley (cited in Wilkins3 ). “Remembering,” as Cole and Means4 posit, “is something one does in order to do something else; it is not the goal of activity” (p. 149). Because of the time lag, this “effort-goal” association is seldom clear. Since our early days, we are instructed to be ready to use much of our knowledge, but may spend a lifetime never having a chance to cite a line of Wordsworth, which we were asked to remember as children. In contrast, a time-cued ProM is an explicit example of this instrumental feature of memory, namely a self-initiated “directive” action and an “act of creation.” In the categorical decree of Alfred Adler, an “individual is either ‘prospective’ ∗ The Leslie and Susan Gonda (Goldschmied) Multidisciplinary Brain Research Center, and Department of Criminology, Bar-Ilan University, Ramat-Gan 52900, Israel; e-mail: [email protected] † Tel-Aviv University, Israel; Howard University, Washington, DC and Clinical Brain Disorders Branch NIMH/NIH, Bethesda, MD 20892-1379, USA; e-mail: [email protected]

263

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

264

J. Glicksohn & M. S. Myslobodsky

or he is psychically nonexistent”5 (p. 67). We might further elaborate that operationally, memory would be named as prospective when an intention (e.g. to phone someone at 9 o’clock in the evening) is deferred until some prompt (e.g. I’ve just seen the 8 o’clock evening news) makes this recalled. The word prospective (from pro- “forward” + specere “look at”) alludes to an effort to discern something far away from the viewer — in the present context, this being distant in time. Among its numerous synonyms to be found in Roget’s New MillenniumTM Thesaurus, over a dozen are sure to serve as possible qualifiers of ProM, such as anticipated, destined, eventual, expected, foreseen, forthcoming, future, hoped-for, impending, intended, likely, looked-for, planned, or possible. In a word, it is one of the instruments of pronoia. The proposition that ProM is a process spurred by the expectation of specific consequences is akin to what, in an older literature, was referred to as “incentive motivation,” or “anticipated reward value.”6 Any external stimulus that is expected to produce a desired outcome can be defined as being rewarding, regardless of whether it is has positive or negative connotation. Over the years, the notion of motivational salience or incentive has been used to describe these various stimuli without emphasizing the role of reward deferral as a backbone of ProM or to diminish the role of noncognitive factors in “willed plans.” In this concluding chapter, we will discuss what it takes to remember these deferred intentions in the context of the concepts covered by the contributors of the volume.

A Tribute to Lewin Perhaps the first study in experimental psychology addressing the role of intentions and ProM was that of Narziss Ach, a member of the Würzburg school of psychology,7 who already in 1905 had investigated “determining tendencies” in thought and action. In fact, given that Ach employed post-hypnotic suggestion to support the notion of such determining tendencies, his work is a clear forerunner of current research on this specific aspect of ProM (for reviews, see Hilgard8 and Orne et al.9 ). Ach’s work was well known by the time the Gestalt psychologists began their explorations of both memory and thought, in the second phase of their research program into the study of cognition. Kurt Lewin subsequently extended Ach’s work, which evolved into his interest in the notion of intentionality.10 Propitiously, Lewin and his students (the research team known as the Quasselstrippe) used to discuss research over

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

265

coffee, and met regularly at the Schwedische Cafe in the early 1920s. As Ash11 recalls: One evening someone expressed amazement at the cafe waiter’s apparent ability to remember what everyone had ordered without writing anything down. Some time after they had paid, Lewin called the waiter and asked what they had ordered. He replied indignantly that he no longer knew. This was the stimulus for Bluma Zeigarnik’s famous investigation of memory for completed and uncompleted tasks. (p. 271)

Bluma Wulfovna Zeigarnik (1900–1988), who conducted her research in Germany and later in Soviet Russia (initially as Vygotsky’s assistant), is known as a mere footnote to Lewin’s biography. This is an impressive footnote, nonetheless, referring as it does to the “Zeigarnik effect” that initiated the search for memorizing by “impersistence.” The work on the topic conducted by Zeigarnik and Lewin’s other students was published between 1927 and 1930 in the flagship journal of the Berlin school of Gestalt psychology, Psychologische Forschung (now known as Psychological Research — which continues to be inspired by the Zeigarnik effect12 ). Lewin, of course, referred to this work in his major publications,13,14 as did Köhler, when expounding on Gestalt psychology to his American and British audiences. Köhler15 introduces the problem of ProM as follows: The following experience is quite common: I have a task which, perhaps, I do not like, but which is urgent. In the course of the day, however, I find myself occupied with many other things. I talk with friends, I read a book, and so forth. But time and again something like a pressure makes itself felt in my interior, and upon examination this pressure proves to issue from that task. The pressure amounts to a persistent tendency of the task to be recalled, and thus to enter the present field of action. (p. 178, italics added)

Note Köhler’s judicious use of the term “field of action,” essentially wedding the intention with its future implementation (for pertinent discussions of the ecological variant of this field of action, see Kadar & Shaw16 and van Leeuwen & Stins17 ). Somewhat anticipating the coverage, it is wise to tribute this passage as a reminder of the sequential view of memory that decomposes intentionality into time-related steps.18,19 If we look back to Lewin of 1938, we find him even more specific here, suggesting that the results of these experiments on task interruption might well serve as “indirect measurements of the strength of the

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

266

J. Glicksohn & M. S. Myslobodsky

force to the goal,” noting (p. 170) that “within certain limits the tendency to resumption was greater if the distance from finishing the work was smaller.”14 As is well known, partial reinforcement augments persistence.20,21

Intentionality and Action Any new goal is scheduled so as to dovetail with other intentions that belong together, or else to reschedule them.22 One scenario of self-propelled long delayed tasks that are seldom discussed by students of ProM is that the wisdom of setting such goals may be doubted and their value is reduced (Chap. 5). Therefore, ProM must prioritize, moving a particular intention ahead of the queue, or specify its high-level priority, thereby establishing a sort of “dominant focus” of prepotent behavior designated as the “intention superiority effect.”23 The latter determines for how long a novel task would tolerate a delay before being consciously recalled.24 It seems that task importance is more demanding for a time-based ProM task, as opposed to an event-based one.25,a Yet, one might well ask whether the system which computes a timely act to be executed within a few minutes from now — as is currently investigated, and advocated for, in contemporary ProM research27 — can be conceived of as being similar to that which is needed for implementing goals over longer time scales,2 such that individuals might gain information, skills, or views that revise or question previous intentions (or, indeed, to acquire the necessary expertise needed28 ). In everyday life, the concluding path to a timed goal may require what is referred to as an instant “perceptual judgment,” or a process of checks, trials and further checks until confirmation has occurred,29 that is to say, the execution of a Test-Operate-Test-Exit (TOTE) plan.30 Note that while the TOTE has been employed to discuss the effective hammering in of a nail, it can also serve as a means by which the individual achieves some degree of tension reduction, and breathes a “sigh of relief,” as he or she gets closer to (or, actually arrives at) the goal. The Test-Wait-Test-Exit (TWTE) model of Harris and Wilkins18 is a variant of the TOTE developed specifically for a time-based ProM. This strategy places the intention in the background, with a subsequent series of increasingly shortened test-wait cycles (self-introduced nonlinearity), continuing until it is appropriate to act. This “J-shaped” pattern of responding to deadlines was a The subdivision of ProM into event- and time-cued variants does not refer to neuroanatomically or neurobiologically different memory systems, as we will try to show here, but to a set of tasks (using Graf and Schacter’s26 argument in a different context).

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

267

handsomely replicated across populations and ages (see Chap. 8). In his latest novel, A Tale of Love and Darkness, Amos Oz31 describes an amusing episode of his family planning to place a telephone call from Jerusalem to their folks in Tel-Aviv: As early as the Sunday before, my father would say to my mother, Fania, you haven’t forgotten that this is the week that we’re phoning to Tel-Aviv? On Monday my mother would say, Arieh, don’t be late home the day after tomorrow, don’t mess things up. And on Tuesday they would both say to me, Amos, just don’t make any surprises for us, you hear, just don’t be ill, you hear, don’t catch cold or fall over until after tomorrow afternoon . . . So they would build up the excitement . . . by three o’clock my father would say to my mother: “Don’t start anything new now, so you won’t be in a rush” “I’m perfectly OK, but what about you with your books, you might forget all about it.” “Me? Forget? I’m looking at the clock every few minutes . . .” (p. 8).

Note the tempo in which things happen in this example. The major point here is that for anxious individuals, an increased concern over overvalued goals coupled with the excessive pressure of urgency, may well cause them to overestimate the time required for task completion,32 and necessitate repeated time sampling in order to “exit” on time to the planned event. Such obsessive time monitoring behavior33 seems to act as an acceptable refresher button for reloading the working memory buffer store — what Robbins34 refers to as “primary memory” in the conviction of its being fundamental for the very perception of an event itself: . . . we now have a stack of stored past instants, all of which are retained and integrated or related somehow to form the whole event. A portion of this stack somehow “slides along” in time as the “present” or ongoing perception, some of the earlier instants falling off, others being added on as they arrive. This is a strange process . . . It relies on a memory that stores “instants” and supports the perception of events consisting of sets of instants, events clearly delimited in time, yet large enough in extent to support the perception of a whole event . . . (p. 779)

The accelerated responding close to the planned event is not necessarily a unique feature of timed behavior. In a naturalistic study, aimed at uncovering

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

268

J. Glicksohn & M. S. Myslobodsky

the relationship between thinking about and remembering intentions, individuals remembered better when asked to respond in a particular room (“place task”) than to respond every two hours (“time task”). Interestingly, in the former conditions, thoughts increased with the proximity to the target location, somewhat similar to the time-responding principle.35 Such a strategy will presumably vary with the initial strength of intention,10 one’s personality (Chap. 6), one’s expectations (Chap. 3), and the context within which testing is being made. These factors are particularly relevant (or perhaps more salient) the more the plans are delayed. To phrase this time-cued ProM dependence upon modulating variables somewhat axiomatically: The longer the temporal delay between the goal and its implementation, the less precise must be its temporal judgment.36

The Cinderella Problem The governing systems for timing the future may be subdivided in terms of outcome of planned events, into two categories, soft-wired and hard-wired time (Fig. 1). Hard-wired deadlines are mandatory and often require just scheduling rather than planning. By contrast, time-cued ProM which is governed by softtime systems permits rearrangements, even if it is not “subjective time” in that it may be objectively measurable. As an example, a delayed submission of a chapter does not corrupt its quality. It may delay the production of an edited book, but the chapter itself may only benefit from polishing, given another month. Thus, a soft-time system accepts laxity or a rescheduling of the timewindow. To piggyback on one of Dibble’s37 examples, when pressing a button to schedule an elevator, one does not expect to arrive upstairs at a prearranged time with an accuracy of the order of a second. “Life goes on,” he quipped, “after a missed deadline.” This attitude can be understood, but to a point. A patient in the emergency room would wish his nurse would depend on a hard-wired time system, such as not to neglect examining his vital signs at the right moment. In other words, a soft-wired time system differs from a hard-wired time system not only in terms of the priority of scheduling, but also in terms of the penalties to be incurred for violating the deadline. Cinderella needed a miracle to absolve her of the instruction to leave the merriment before the clock strikes 12 (see Chap. 6 for further discussion of this). Given that ProM is not maintained in consciousness, time sampling throughout the planned retention interval may be conceived of as a test of competence for switching on and off ongoing activity. That makes many ProM tasks intolerant

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

269 Penalty

Reward Goal Reached or Failed

Proceed to

Urgent

Hard-wired Planning

Achievement ideation

Wants

Cue retrieval

Proceed to

Time-cued prompt Delayed

Call off or Postpone

Soft-wired

Intentions Rescheduled Goal Failed

Goal Achieved

Fig. 1. Schematic representation of the fate of remembered intentions. The trigger for time-cued ProM is the appreciation of needs that are reactivated by the ideation. In hard-wired time ProM, scheduled plans are considered to be aborted if a deadline is not met, whereas in soft-wired time ProM, planned events can still be rescheduled if a deadline is missed. The schematic indicates that the longer the deadline is delayed, the more likely it is to depend on the soft-wired time system and thus more likely to undergo a revision of value (“discounting”)38 (Chap. 5). However, the endpoint of delayed tasks may contain elements of hard-wired performance (Myslobodsky & Goodman, unpublished).

of errors. Imagine remembering to take a look at an island from a window of a flying airplane while also performing some engaging reading, or taking an example from Wilkins3 of a patient who was about to take his medication when disturbed by the telephone ringing, and who did not “remember to remember” his intention.b It is a gamble to make an exit by a single bold time-assessment, for such an act may not be undone. These examples are much less dramatic than those portrayed by the Cinderella scenario, but missing any of these events by a fraction of time may well mean missing the planned event altogether. This time duality indicates that remembering time is not the only goal, for in ProM, the emphasis is on how the goals are accomplished, not on how well they b This

unforgiving hard-wired ProM is identical to a “count-down” task. Its garden variety is exemplified by a mechanical kitchen timer armed to a maximal time setting to alert one of something cooking on the stove or in the oven when time reaches “0.”

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

270

J. Glicksohn & M. S. Myslobodsky

were remembered. Although there are misgivings that ProM research suffers from being assessed in a binary way (“remember-act” distinction),39 as opposed to the traditional, quantity-oriented framework of retrospective memory, the binary way is justified by being consistent with the requirement for precision,40 and the accuracy-oriented approach to memory.41 These examples also indicate that the subdivision of these two time systems is somewhat artificial since the boundary between them is fuzzy. In an almost forgotten classic, “An Affair to Remember,” Carry Grant (as Nicky Ferrante) has a romantic affair with Deborah Kerr (as Terry McKay) while on a cruise from Europe to New York. Their ProM motivates the plot when the happy lovers agree in six months to join up at the top of the Empire State Building. A tragic accident, however, keeps Terry from the reunion. This aborted time-cued ProM illustrates the looming conflict between intentions and their implementation. Is Terry’s “slip” (defined as an appreciation of the need to proceed doing something intended, as she did, but a failure to go nonetheless on the relevant date) part of a time-based ProM?

Hard-Wired Timing and a ProM Clock The beauty of ProM is that it works as a processor within computer systems and is not visible to its user, nor is it visibly affected by factors such as gender or formal education. Its impressive feature is the ability to engage consciousness of planned intentions occasionally, otherwise doing the job inconspicuously in the background. This two-tier — foreground-background — division of responsibility is not unique to ProM. Wim A. van de Grind (see Chap. 4) used it as a paradigm of a temporal dispersion of sensory information and partitioning of labor in our brain, such as a stream for the processing of specific (visual) features needed for object identification (e.g. color and shape) in the anatomically separate (occipitotemporal) “ventral stream,” and a stream for the processing of spatial relationships and movement processed in the (occipitoparietal) “dorsal stream” (reviewed in Ungerleider & Haxby42 ). These have commonly become to be known as the “what” and “where” pathways.43 When an intention (“what”) is stated, another question to solve is “where,” as well as a question of “when.” Goodale and Milner43 refocused attention on the different ways in which visual information is transformed for different goals. They maintain that the dorsal stream mediates the required sensorimotor transformations for visually guided actions directed at objects of interest, and thus relabeled the dorsal stream as a “how” system. But perhaps there is also a time problem here. The dorsal or

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

271

the “where” pathway develops earlier,44 as though suggesting that the question of “where” must be solved first before the system turns to analyzing the problem. Ecologically (e.g. for a prey or a predator), a signal is resolved spatially (“where”) whereas the temporal answer of “now” is implicitly given by the orienting response (i.e. the response to a novel and a potentially threatening event). The orienting response may be erroneous about the danger, as we have all learned, but one is never in doubt that something is about to happen “now” and within a msec-seconds time range. When a prey encounters the predator, its orienting response mobilizes a hard-wired time program for instantaneous immobility or for flight. Only when fear and its autonomic markers subside or are succumbed to, does the need for “what” and “how” become important. Although such hard-wired orienting-type responses measured in seconds are studied in the context of ProM, and could possibly be likened to time judgment, time sense or polymodal sensory memory (if we are to match taxonomic borders of retrospective memory), they merely represent laboratory curiosity that is hardly relevant for time-cued ProM, as already mentioned. That leads to the “clock problem” in ProM. Our body (brain) has a set of realvalue measurable variables (e.g. neuronal, somatic, metabolic, and endocrine signals), which collectively are given a family name of the “clock.” Many genes, too, are expressed in a circadian fashion and recognized as clock-controlling genes. We know by now that timing the future is a complex process that requires a steady metabolic backing. One of the most elegant evolutionary designs of The Blind Watchmaker to secure this flow of energy is in the placing of a “master clock” in the suprachiasmatic nucleus in the base of the hypothalamus. This clock deals with the circadian cycle and mediates various aspects of motivated behavior. Its pacemaker neurons respond to visual-system input for clock resetting, but were not reported to transmit signals in response to cognitive demands. Some cerebral rhythms that fluctuate below the frequency of delta (1–4 Hz) described in the 60s45 seemed at the time a better candidate for a timer sensitive to autonomic input. Like metronomes, many such local “clocks show the tempo without ‘knowing time’ ”. The role of each clock component in the time-course of ProM remains unknown. The major problem in conceptualizing such a clock is not so much in adopting one of its features or accepting its models as a given (see Chaps. 1 and 2), but in knowing what dependent variables must be chosen as an index of a particular behavior. This issue is familiar to all scientists dealing with the pacemaker problem. These difficulties prompted some workers46 to suggest that a gradual decay

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

272

J. Glicksohn & M. S. Myslobodsky

of memory strength, in itself, may serve as a timer.47 Lewin10 was seemingly concerned about the erosion of memory and intention as a timing cue: “what role is played by the length of time elapsing between the act of intending and the consummatory action? Does the after-effect of the intention decrease progressively as associations do, according to the so-called curve of forgetting?” (p. 96). Yet, this “pacemaker-liberated” interval timing of Staddon46 would still necessitate facilities of its own for monitoring the degree of memory attrition and for comparing it with previous representations, such as possibly happening in the case of habituation. A metaphor of habituation suggested by Staddon46 provides a plausible mechanism for the “sinking” of ProM into the background (in conditions of monotony and a complete identity of environmental cues) and accounts for an instantaneous recovery of responding (dishabituation) with an intrusion of any sporadic event, cognitive or environmental. It also implies that intentions depend on ad hoc created clocks that are set in motion to monitor episodic recall (see Chap. 3 for one such option) with a pace of cognitive intrusions or environmental prompts, much as was apparently foreseen in Wilkins’ Random Walk Model.48 Wilkins’ idea is in keeping with the fact that ProM is always an ad hoc “intruder” into a network with a history of co-activated pathways that control tasks of dissimilar personal relevance. Their postponement, reevaluation of cues of changing relevance as well as sifting through evolving personal thoughts can be conceived of as a random-walk-like behavior. That is to say, that one way to improve timing may be by using some “resource” interim cues that act as pace-regulators. In a recent study by Kvavilashvili,49 only 9 to 14% of rehearsals were self-initiated in the time-based tasks, i.e. cued by thoughts about plans and activities. The majority of rehearsals were triggered either by incidental external or internal cues, or they simply popped into mind without any obvious triggers. An alternative timing device to think of would be a hard-wired system known in time-zone computing and electronic engineering as delta time. In the present context, it is conceived of as a rigid pacemaker, as in the scalar-timing model (Chap. 1), except that it does not have the decision-making capacity of the scalar-timing model (e.g. its capacity to store and retrieve from the accumulator the pulses on reward). Its example is the Delta Time-Interval Oscilloscopes that permit to place markers at a selected initial event and obtain time to the next desirable marker. Such place markers could be likened to a series of known or expected “progression events” (an example is that of disease progression events as staging tumor grades or their invasion into the normal tissues).

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

273

These events are sensible assuming that each stage develops once the previous one has been completed. They give certain clues for time approximation and are often used in medical practice (for predicting recovery or survival from one stage of disease phenotype to another). They depend, however, on the exactness of the history of events that needs to be plotted for time prediction and precision of time boundaries. Of course, behavioral or medical events do not have a precise built-in counter to depend completely on behavioral delta-time estimates. Our internal clocks are apparently capable of opening up to top-down regulation (see Chap. 2); they click either too slow or too fast, may be unexpectedly “glued” to some events of the past, or show irrational preoccupation with living in the present, or provide the disparity between time and duration in Henri Bergson’s language. Therefore, the general format of a delta-time counter (an offset from the current date and time to a future moment) as a candidate for a ProM clock has yet to be explored.

Personal Time and Time Discounting: The Next Frontier of Prospective Memory? Professor Higgins in My Fair Lady sets a goal for an eager Eliza Doolittle his ambitious intentions in soft-time boundaries: Eliza, you are to stay here for the next six months learning how to speak beautifully, like a lady in the florist’s shop. If you’re good and do whatever you’re told, you shall sleep in a proper bedroom and have lots to eat, and money to buy chocolates and take rides in taxis . . . At the end of six months you shall go to Buckingham Palace in a carriage, beautifully dressed.

Professor Higgins knows quite well from the outset that six months is an arbitrary target, so that reward of his bet may be obtained sooner or postponed. He explains this to a somewhat uneasy Colonel Pickering: “In six months — in three if she has a good ear and a quick tongue — I’ll take her anywhere and pass her off as anything.” Clearly, this kind of ProM is maintained by a sequence of precise cues, such as the linguistic achievements or good manners of Eliza Doolittle (or “low level construals”) rather than by a set point in time. Therefore, each return to “waiting” (in the “test-wait-exit” loop) for some extra time before Professor Higgins risks taking her to Buckingham Palace could be viewed as an aborted goal that has to be rescheduled or as only a partial reward and tentative fulfillment of the

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

274

J. Glicksohn & M. S. Myslobodsky

ambition. Unlike this ambitious goal, but as imprecisely timed, one may plan to meet a friend only when he “cools off,” or by proposing other consensual, cultural-based notions of extended time, such as “afternoon,” “lunch-time,” or “martini-time,” that are rarely studied. Plans based on “subjective” time experience or “personal” time,50 or on the change in mood or achievements of others, make time accuracy superfluous. An important aspect of this psychological time is its vulnerability to cognitive intrusion (“cognitive orienting reflexes”) and to secondary goals that have certain emotional or cognitive valence51–53 (and leave a trace in the form of body states evoked by prior experience, i.e., “somatic markers”54 ), whether or not we know of them or can explain them. Yet another mark, shared by all long deferred intentions, is that decisions temporally separated from their outcomes may become less effective in predicting effects, the longer the goals are delayed. Their unpredictability leads to an error in original goal values and a degree of their “markdown” associated with numerous personal and extrapersonal circumstances, a feature somewhat reminiscent of scalar timing. Economists already in the 30s of 19th century speculated on the psychological motives underlying the tradeoffs among costs and benefits occurring at different times in the future (variability intertemporal-choice behaviors) “drawing on little more than introspection and personal observation”.38, c Such gradual reinforcer devaluation labeled as time discounting refers to any reason for caring less about the future, from changing tastes to health attrition. (See Chaps. 5 and 6 for review of the time discounting literature.) Time discounting should have been a part of time-cued ProM in that it examines what is happening with our plans and when, except that many implications of the discounting idea could only be seen at reasonably protracted delays. Unlike ProM, the success of “discounting” under the roof of behavioral economics was not determined by a struggle for legitimacy or the lineage of the term; its viability was tested only by the accuracy of predictions. As Critchfield and Madden (Chap. 5) tell us, discounting evolved into a useful theoretical device for exploring normal and aberrant behaviors, as well as the problems commonly encountered in sociology, psychiatry and ethnopsychiatry. For example, individuals who steeply discount delayed rewards tend to manifest impulsivity-related behavior problems (e.g., some normal or hyperactive youth, adult gamblers, or c It is of interest that the notion of “myopia” for time that will be mentioned later was a long-standing concern of economists.55

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

275

drug addicts). Actually, similar to the reckless adolescents, the elderly (or chronically sick patients) even with risk aversion may manifest the same hyperbolic discounting of reward (gains) and a reduced sensitivity to delayed punishment (costs). That may be associated with fear, say, for not being around to collect the rewards; an effect that could not be easily detected in a typical ProM laboratory study. In this regard, discounting paradigms are an ideal answer to a common plea for ecological relevance in ProM research. More recently, the theory of “temporal construal”53 (Chap. 7) put in a different perspective on the theme of discounting. The latter theory draws attention to the fact that remote future situations are represented by some abstract, coherent and fundamental features (“high level construals”) as opposed to more concrete, contextual, and detailed elements of the near future circumstances (“low level construals”) (e.g. a goal to evolve into a “lady” as opposed to having “lots to eat, and money to buy chocolates and take rides in taxis”). The levels of construal reach beyond the valence of planned events to touch with their types governed by the individual representations of behavior, thoughts, ideations, planning, and predictions (Chap. 7). These are impossible to refute, but their comprehensive range has yet to provide specific predictions for planning neuropsychological investigations. The classical time-discounting theories are somewhat more “user friendly,” so to speak, in being anchored in affect, drives, motivations and rewards. Although difficulties that arise when attention turns toward probing the individual timing behaviors in the world outside of controlled academic settings are considerable; a recent study by McClure et al.56 is a hopeful sign that they can be contended with. In the latter, participants had to make a series of binary choices between smaller/earlier and larger/later money amounts while their brains were scanned using fMRI. It is a complex task, in particular because it is based on soliciting verbal responses to hypothetical future options. Nevertheless, it was possible to demonstrate that such choices recruit two separate systems: the system associated with the midbrain dopamine system that was preferentially activated by decisions involving immediately available rewards and the fronto-parietal activation that was recruited (perhaps through abstract reasoning or “simulation” with imagery) when participants choose longer-term options.56 Applauding this elegant study in the same issue of Science, George Ainslie and John Monterosso mentioned in their commentary that there are conditions when people and even nonhuman subjects tend to lock in their larger/later preference, and thus, we would add, the construal level theory must better the case.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

276

J. Glicksohn & M. S. Myslobodsky

Error Monitoring The greatest disruption in time accuracy is detected when people perform timeestimation tasks concurrently with secondary working memory tasks.57 It is because of this that time estimation may serve as a suitable index of mental workload.58,59 Arguments regarding the critical role of events and their density on the perception of time have received extensive treatment in the literature over the past thirty-five years.60–65 In particular, Poynter65 has noted that more segmented intervals are retrospectively judged to be longer, as are intervals containing more complex events. But, what do we know about prospective time estimations? According to James,66 “In general, a time filled with varied and interesting experiences seems short in passing, but long as we look back. On the other hand, a tract of time empty of experiences seems long in passing, but in retrospect short” (p. 624; italics in original). With respect to this, one of us67 has recently suggested that: . . . within a prospective paradigm, the more focused internally is one’s attention . . . the slower the rate of functioning of the cognitive timer . . . Time will therefore seem long in passing, as James commented (and as best assessed using the method of production — the expectation here is for longer productions). Time will also seem short on retrospective reflection, as James noted, because retrospective time estimation (e.g. verbal estimation — the expectation here being for smaller estimations) has a different attentional quality . . . (p. 11).

If this is the case, a cluttered extent of time will be underestimated (i.e. shorter time productions) — hence the need for more time checks, because even the most well-calibrated timer is being influenced by ongoing events, be this external tempo,68–72 environmental stimulation,73–76 contextual change,77,78 or event structure.58,79,d d Contra

the “filled-interval illusion,”80 perceived time flies when one is having fun and gets stretched out in boredom.81 Discussing time experience with Popper, Eccles82 made a narcissistic observation that under certain conditions, “say a very nice dinner party, the time for the whole meal has gone without us appreciating even perhaps the food too much! We have been so busily engaged in attractive conversation. On other occasions you can feel a dinner party lasts a very long time because no one talks to you . . .” (p. 529). Again, based on self-observation, Bergson83 believed that we hardly feel the passing of time unless we become impatient: “It coincides with my impatience, that is to say, with a certain portion of my own duration, which I cannot protract or contract as I like. It is no longer something thought, it is something lived. It is no longer a relation, it is an absolute” (p. 10). In social or natural crises, time may not necessarily have a tempo of its own, nor absolute inescapable velocity.84

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

277

In a similar vein, a time-based ProM task might also serve as an index of workload.85 High motivation was reported to improve accuracy in ProM tasks and augment time-monitoring frequency86 — a finding standing in marked contrast to that indicating that higher intrinsic motivation was associated with checking and thinking about time less often.87 On the other hand, a community sample of clinically depressed adults showed impaired time-based ProM relative to nondepressed individuals88 — a finding matching that indicating that depressed individuals are also impaired in time estimation.89,90 Mäntylä91 has suggested that ProM is maintained by a form of residual activation (an “unresolved goal tension” in Lewin’s language). While our understanding of cue-related ProM is considerable (piggybacking on the achievements in the retrospective memory area), our understanding of brain mechanisms responsible for time taking is much less clear. It is uncertain whether or not there are separate networks that code specific intervals for different neuronal populations and various modalities92 or whether specific frontostriatal and thalamocortical circuits are activated when one needs to shift attention from one temporal context to another93 (see Chap. 9). Even less is known about the temporal displacement of the future events and the conditions leading to misjudging (over- or underestimating) future time. The prefrontal cortex is an area of heterogeneous organization and function that, in addition, has it own error-detection system that plays an important role in monitoring performance.94 It was long held to be in charge of planning95 and “memory of the future.”96 As with any remembrance, memory for the future cannot be complete, foolproof, and lasting — thus its execution must be monitored, and the choices made for its implementation need to be updated and corrected “on the move.” Neuropsychology and cognitive neuroscience have made a number of contributions to the understanding of the foundation of this monitoring function. The orbitomedial prefrontal cortex along with the hippocampus may be involved in discrepancy monitoring of the internal milieu sensorium caused by “just-prior” representations and those held “on-line.” This attentional monitoring system, in a sense, is the working memory of autonomic viscero-emotional processing.97 Together with the amygdala and insula, it presumably generates autonomic responses that are termed “somatic markers.”98 Like the Jungian “somatic unconscious” becoming material so as to exert cognitive-emotional control of behavior, here too the prefrontal cortex exerts an influence on hippocampal neurons in a variety of ways and through diverse anatomical input to the hippocampus, but it is premature to speculate about these influences on timing future behaviors in humans.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

278

J. Glicksohn & M. S. Myslobodsky

As mentioned above, a time-cued task rests on the same judgment, as does any other decision. However, judgment does not amount to proof, and much effort has been exhausted in examining a “general” error-related response system (in the mesioprefrontal cortex, including anterior cingulate cortex, presupplementary motor area, as well as inferior parietal lobule) involved in error monitoring, possibly shared by a number of tasks.99,100 We shall not attempt here to do more than to name them.e Diverse approaches and experimental designs have been proposed to define the triggers of motor (praxis and action knowledge) or cognitive responses (e.g. verbal-visual semantic knowledge) when the assumed time to act has come (erroneously or not). Such triggers are so irresistible as to be subjectively experienced as an insight that “pops into mind.”91 For want of the underlying neurobiological mechanisms, these “popping up experiences” were described under a welter of names. This terminological issue may be unfairly dismissed as inconsequential, yet we see it as a crucial one. Kvavilashvili and Mandler103 (see also for review) “provisionally” call them involuntary semantic memories, and used self-report and questionnaire studies to explore their roles in regard to ProM. An important feature of “involuntary memories” was the (apparent) dearth of stimulus information, i.e. “the absence of easily identifiable cues in one’s immediate environment, ongoing activities and concurrent thoughts.” There is no recognized marker for time-internal attention arrest that is similar to, say, an environmentally-driven saccade dubbed “oculomotor capture.”104 It would be e

In the majority of such studies, reviewed by Botvinick and colleagues,101 the anterior cingulate cortex was activated in one of three behavioral contexts: (i) tasks that required the overriding of prepotent responses, (ii) tasks that required selection among a set of equally permissible responses (underdetermined responding), or (iii) tasks that involved the commission of errors. All are relevant for ProM, but the latter context is likely to prevail in hard-wired time-cued ProM. Electrophysiological studies using the error related negativity (ERN), a negative deflection in the event-related potential, to trace these processes, have indicated that the dorsal anterior cingulate cortex is particularly sensitive to the commission of errors. The role of ERN in predicting reward/punishment and performance feedback has been discussed in a number of studies. The ERN is augmented when participants are surprised by error feedback. Building on earlier dipole models locating the error feedback negativity generator in the anterior cingulate, a subsequent event-related fMRI study did not obtain error feedback-evoked cingulate activity. These findings suggest that the anterior cingulate may not be involved in the generation of this ERN component. Although these results seem to question the involvement of the cingulate in the detection of errors per se, error awareness (e.g. in unexpected unfavorable outcomes or when a negative reward prediction error promotes subsequent response switching) or uncertainty about outcome are the typical conflict features, and that is when the ERN is consistently augmented. In Box 3 of their paper, Botvinick et al.101 outlined several pending research questions. They did not tell us how “conflict” is being defined and how early its presence can be established. According to Tversky and Shafir,102 there is hardly a commonly accepted procedure for assessing conflict or for establishing its strength.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

279

interesting to use psychophysiological and fMRI measures in order to examine whether or not the “pop-up” phenomena, in spite of their diverse appearances, are related to cognitive orienting responses.105 Regardless of task designs, a trial-and-error process of timing requires that various participating networks for motor preparation and motor attention input their oscillations into a cortical area on the mesial surface of the superior frontal gyrus or Brodmann’s area 6, known as the supplementary motor area (SMA). The SMA is believed to be essential for higher-order motor planning and temporal organization of movements that demand retrieval of motor memory, especially in sequential performance of multiple motor acts (and by extension, in cognition — see, e.g. what has been termed the “motor theory” of cognition106 ). The SMA provides prompts for actions on external and internal cues and possibly for “self-generated” actions without awareness in response to conflict and/or to uncertainty. Deficits that involve area 6 (e.g. in Parkinson’s disease) contribute to a syndrome characterized by impaired initiation of volitional movements and inferior self-initiation of movement production (akin to dynamic akinesia), as well as to motor neglect. In addition, Parkinson’s patients exhibit defective timing.107 Surprisingly, a deficient SMA does not impair time-based ProM tasks (a tap in 10 minutes and another one in 15 minute) as opposed to event-based ProM. Also, like controls, Parkinson’s patients accelerated their time checking right before the target times.108 This finding has yet to be replicated, since Parkinson’s patients (cognitively unimpaired, medicated patients) were shown to have dysfunctional striato-thalamo-cortical loops underlying impaired motor timing.109 In addition, substantia nigra pars compacta (SNc), the major source of dopamine deficit in the disorder, is implicated in timing behavior.110

Are Errors all Bad? We know very little of the nature of the device that chooses when to “exit” as opposed to the one that “restrains” one to a time-confined response. But it must be related to the processing of reward and punishment in diverse neural systems recruited by the task. Let us assume that a time-cued ProM task makes reward payoffs contingent upon action. If the goal is recalled after its intended time, hard-wired ProM is considered to be a failure (somewhat akin to missing the time-to-contact with the ball when swinging a baseball bat). The same happens with cognitive impulses when the ability to monitor one’s “somatic state” is temporarily lost, which sets the threshold of a negative (?) emotional

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

280

J. Glicksohn & M. S. Myslobodsky

signal higher. Likewise, that act cannot be undone. Therefore, payoffs are maximized wherein individuals consciously or preconsciously are biased to choose an erroneous premature choice of time exit that may not offer greater immediate reward but does entail lower immediate punishment. This low payoff choice (assuming that the choice is theirs as opposed to what they were told to do111 ) might seem beneficial. That is, rather than making a risky single-step-error, they proceed in an impulsive choice test-correction routine by generating errors of commission, learning by detection and exploitation of errors, and then by overcoming them.18,112 The inevitability of repetitive time sampling when the task is delayed18 is an admission to such an agreed upon imprecision, especially when attentional capacity is delegated to other activities.113,f One wonders to what extent this “self-imposed” nonlinearity in timing engages top-down control mechanisms, and whether or not the premature time-sampling is launched by (or builds upon) the random “train of thought” posited in Wilkins’ Random Walk Model.48,g

“Mental Trip” in Timing Intentions To make a provisional summary of the foregoing, a ProM task is comprised of a goal, cue sampling, and plan updating through a completion of the task with an action (or its avoidance). In order to schedule an acceptable timetable, one must also possess initial state knowledge,118 and synthesize a copy of future actions creating a “model of the future”119 or perception-action schema.120 Let us assume that the cue sampling is a kind of one-dimensional continuous process defined in terms of a vertex set (V = {v1 , v2 , v3 , . . . , vn }). Its edges f In

this regard, impulsive choice is quite adaptive for it forms part of a continuum of inter-temporal choice behavior.38 This time-correcting mechanism could be described in terms of Cagan’s114 hypothesis of adaptive expectations, where each successive step increases the accuracy by making a better adjustment toward a final event. Let E¯t t represent the anticipation of planned event E held at time t − 1, and let E t be the actual event at time t. Then E¯ t+1 − E t = k(E t − E¯t ), where k is a number between 0 and 1. That is why lottery owners find it more profitable to propose several smaller prizes rather than offer a single lump sum. g The Random Walk Model was originally described in an unpublished paper entitled “Remembering to remember” communicated in 1979 to the Department of Experimental Psychology, Cambridge University. It is cited (albeit infrequently) from a brief reference made to it by Harris48 on p. 82 of his chapter. The powerful mathematics of its better known predecessors, such as the Brownian random walk or the random walk nature of price changes in financial markets, would have added a needed mathematic tool115,116 to ProM processes. Alas, as Wilkins himself fittingly mused, it was “something I have forgotten to do” due to “the pressure of competing commitments.”117

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

281

i

(a)

V1

V2

V3

Vn

Vn+1 Vn+2

G

E

i (b)

P

Time

Fig. 2. (a) A diagram of time-cued ProM defined in terms of a vertex set V and time edges (v1 v2 , v2 v3 , etc.). Given that time cannot be tested with respect to itself, these steps are defined somewhat tautologically as samples in the history of time sampling (“episodes”), or time-samples with respect to previous episodes and environmental events. Therefore, each vertex of the journey is created either by external cues or self-initiated time sampling and is the point when the future becomes past. In reality, the sites visited on the “walk” to the goal E along the x-axis will randomly cross the line GE a number of times (in the manner of a “drunkard’s walk”) before the intended event E is reached. Time edges are shown as shrinking towards intended outcome (“time shrinking”), so that accuracy is achieved by inaccurate installments toward a final event. This model permits one to conceive of time-cued ProM as a response not to a single delayed reward but to a succession of small reinforcements spaced over time. To avoid complexity, the graph does not indicate that the randomness of the walk would be obscured at short distances since the deviation from the path would be too close to the horizontal axis which departs more significantly from the straight axis at E. (b) A hypothetical “reminiscence curve” for time-cue memory. In the spirit of Wilkins’ model48 the graph shows an acceleration of cue acquisition probability (P) close to E, as the number of steps from the departing point G (“time complexity”) goes up. The hypothetical topography and slope of this “recollection swelling” is emphasized to compare it with the “reminiscence bump” in retrospective time recall.

(v1 v2 , v2 v3 , etc.) are time intervals. Then a diagram (Fig. 2) may be proposed to model time-cued ProM from a goal G with a finite number of vertices through the final planned outcome E (“event”). In reality, though, we do not necessarily proceed systematically to a goal as soon as the need has been appreciated or

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

282

J. Glicksohn & M. S. Myslobodsky

as soon as the goal was scheduled. In Albert Bandura’s language, we have — in the first place — to “reproduce” mentally the desired behaviors.121 In our mental representation, we instantaneously “paste” ourselves to the goal when events imagined in the future begin to affect the present — almost like Macbeth’s dagger (“A dagger of the mind, a false creation”). In 1955, Critchley122 noticed that the body-concept changes its location in outer space with reference to the actual body, depending on whether one thinks “forward” or “backward”: … when the individual switches from the past and looks into the future, in order to plan some activity — a voyage, a piece of work, a social occasion — then he probably sees himself as a participant, looking upon himself not from afar, but as an alter ego standing close by, perhaps to one side of the imagined actor in the future scene (p. 100).

Therefore, timing the journey comes in a sort of “backward causal” process123 that may be a central feature of timing strategy, albeit one that is only rarely made explicit. For that reason, bidirectional loops (i) in Fig. 2 describe this hypothetical “mental-trip.” The ideational trip refers to the prospective event itself (e.g. a motor act or spatial situation) since the time schedule cannot be imagined. Speculatively, the loop i may represent a function mediated by the prefrontal cortex (“pre-task activity”’ or “task-set”).124 Unlike a regular “forward counting” time-cued ProM model, this one could be designated as a “backward counting” model. This hunch appears in different reincarnations, all controversial, and not relevant for time-cued ProM. We know of no paradigm that has attempted to test this idea directly, other than in regard to a somewhat tangential quandary known as Libet’s problem125,126 (see Chap. 4). Disregarding those divisive ideas, one might only mention that the merit of “reproducing” future behavior in the mind’s eye might be in improving performance, as was shown in the domain of motor imagery.127 With a mental image of the goal, one may be handicapped in short-lasting ProM tasks. In a delayed ProM task, one would possibly be more prepared to handle situations when information about sensory cues is incomplete, when integration is required of this information with task rules and task goals, and when transformation of sensory input into a response code is mandatory. It may teach something about the goal itself (Chap. 7). That sequence to a desired response could account for sustained activity in the prefrontal cortex.128 We are tempted here to extend this notion of sustained activation, applying it to examples of a time-cued ProM, when the brain has a model of how each new

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

283

action is to be constructed, and resolves potential sensory problems or scheduling clashes by thinking of some “progression events” and reward expectancy when picking most plausible solutions. Operationally, the transformation from a sensory to a response code may require in the future the study of interactions between activity in the prefrontal cortex and activity in other brain regions of sustained activity such as the parietal cortex for the prospective use of that information in the form of behavior impervious to digression. We may need to learn what is contributed by task designs and what belongs to “backward causal” processes in ProM. It would be important to explore whether the superiority of the right parietal temporal-order judgment (“prior-entry hypothesis”) revealed after parietal injury129 is responsible for the hypothesized normal capacity to imagine intentions fulfilled.

Bridging Between ProM and Internal Timing ProM can be subdivided into analogues of both short- and long-term retrospective memory.39 Likewise, psychological time can be subdivided into both hardwired time of “present” (i.e. continuous present, lasting up to about 30–60 secs, hence within the province of short-term, or working memory) and “future” — be this with respect to time duration130 or time perspective.131 When asked to estimate a duration of time, especially one during which one has been involved with a number of activities, delayed or remote time reproductions tend to be longer.78,132 Furthermore, task interruption has been found to lengthen the verbal estimation of the time spent on the task,133 while if you put on the kettle to make yourself a cup of tea (the “watched pot” problem), then one is invariably surprised by how long this actually takes.60,134 We mention these findings as a prolegomena to a distinction drawn in the ProM literature between two kinds of time-based ProM activities: pulses and steps.135 Goals that need to be retrieved at a precise time in the future were defined as pulses (e.g. give me a call at 16:00), whereas “steps” are tasks that can be completed during a wider time period (e.g. don’t forget to phone me this evening). We shall adopt the meaning of hard-wired and soft-wired time for these terms here. The trend in ProM research has often been directed towards examining the “pulses” because they are easy to control in the laboratory setting. The advantage of brief time periods is that they can be studied together with and attributed to some microelectrical or macro- human brain-activity (EEG/ERP or MEG) phenomena. Internal timing in second-range tasks may be unwittingly confounded (unless controlled for) by covert counting during the duration.136 In a different context, Bolbecker and

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

284

J. Glicksohn & M. S. Myslobodsky

associates137 noticed that “mental activities might range from those which are brief (events) to those which extend over a period of time (processes).” This taxonomy is reminiscent of that proposed earlier by Buchwald and colleagues138 who conceived of the term “response set” as of a short-term biasing effect for the initiation and fine-tuning of movements, whereas the term “cognitive set” was chosen for a fairly long-term modulation of neuronal activity before a response was initiated, as would happen in situations requiring decision making. There is a problem that with brief time periods, “the hidden mental activity” triggered by a specific stimulus (e.g. task instruction) that would continue to be mulled over, would mediate behavior for an indefinite epoch thereafter. In theory, predictions based on some “pulse-type” intervals may not require “planning” and “expectation,” but perhaps some form of long-term duration production (see Chap. 2). As argued above, it is not even certain whether they belong to bona fide ProM. Marsh and Hicks139 found improved prospective memory when the retention interval was increased from 2.5 to 15 minutes. They cite evidence showing that pending intentions may be maintained in an activated state, and speculate that longer delays may provide opportunities to remember the intention during momentary pauses in the foreground task. The scale of “steps” or “processes” may be so wide, however, as to encroach onto psychological time, thus rather becoming more of a “calendar” than a “clock,” and thus requiring a shift to time discounting paradigms. Time assessment of such individual “calendars” cannot be a steady process but tends to happen in bouts, or as one might say, in Lewin’s “momentary sections” of time. Inevitably, a long-delayed time-cued ProM is Janus-faced for it could only proceed forward by remembering ongoing life’s events (episodic memory).140

The Environmental Modulation of Time-Based ProM In order to show that a time-based ProM should be environmentally supported, an effort was made141,142 to marry the TWTE model18 with topographically random cue-sensing tools. That goal would have been achieved within the context of Wilkins’ Random Walk Model.48 The latter has the potential to answer the question of how the time commitments are getting through, and how time corrections are introduced in a manner compatible with the environment (i.e. foreground and background activities) in which a person finds herself. The crucial role of environmental or self-generated cues is that they provide the sense of time continuity by reinforcing one’s motivational state and by increasing the

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

285

likelihood of an “exit” at a desired event, thereby maintaining one’s intentions and overcoming the randomness of the “walk.” They could reactivate a pursuit for outcome seeking behavior even when individuals were asked to abandon the search (e.g. in line with the “white bear effect”). Also, the environmental support admits the fact that the characteristics of changes of behavior when reaching a goal are themselves the product of cue assimilation. Alternatively, one might ask, do we use some preconscious signals for time self-guidance, much as the one postulated for “information pick-up” used to guide locomotion when it corresponds to the viewer’s destination?143 If we track the passage of time, in a manner akin to Gibson’s ecological model,144 do we work out the “time-to-contact” with the swiftly approaching time-to-exit in a delta-time fashion? The answer is that there is no time guide we know of to compute the tau (the reciprocal of the relative rate of dilation of the visual angle in response to looming objects) with the same dependability, though the hunt is still on.145 The best we can do is to make repeated glances at one’s watch — and this is the cardinal feature of a time-based ProM, as currently investigated146 (see also Chap. 8). Much as one might argue that a cognitive timer is essential for a time-based ProM (see Chaps. 1 and 2), one cannot escape the fact that in postulating a process of time checking18,19 lasting up until the time when it is appropriate to act, emphasis has been placed (explicitly or implicitly) on the environmental cue which terminates this process, since time-based tasks appear to be less selfinitiated than previously thought.49 Furthermore, given that the process might extend over a rather long period, one has to acknowledge that events occurring during this period will themselves have an influence on the perception of time. The notion that time perception is essentially event perception has a venerable past, going back to William James,66 and more recently back to James Gibson,143 who suggested that “There is no such thing as the perception of time, but only the perception of events and locomotion” (p. 295). Furthermore, in acknowledging the common dynamics underlying space perception, motion perception and time perception,67 the same type of field effects found in the spatial domain can be transposed to the temporal one — with implications to be drawn for the study of a time-based ProM. In the spirit of Kurt Lewin’s10 “psychological distance,” Trope and associates (see Chap. 7 for review) conceive of temporal distance from future events, as being governed by the same general principles held for other distance dimensions, such as temporal distance from past events, spatial distance, and even social distance.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

286

J. Glicksohn & M. S. Myslobodsky

Consider the fact that in the spatial domain, events are grouped together based on proximity, similarity and the other Gestalt laws of perceptual grouping.15,147 If the same cognitive process underlies both time and distance judgments,148 then the same type of perceptual grouping will be found in the temporal domain.149 If individuals remember grouped elements as being more aligned with each other, relative to a spatial reference frame, than they actually are,150,151 then the same type of perceptual alignment will be found for time estimation.152 If with respect to the spatial domain, our mental representation is hierarchically ordered,153 then so will the case with respect to the temporal domain.51– 53 And if event perception is critical for navigating our spatial and cognitive environments,144,154 then event perception will be critical for the temporal dimension and its navigation.143 This binding of spatial cues and time is believed to be executed by a neural mechanism — for example, the “aggregate predictor”155 that could also bind actions and their effects in conscious awareness.156 Thus, an inescapable conclusion is that a time-based ProM must work in unison with other cognitive processes, because at the end of the day our time perception is intricately related to event perception (see Chap. 3) and spatial memory. Environmental cues could provide the framework for a sequence of episodic memory snapshots89 when each previous event inputs its priority to a subsequent one. These ideas have specific neurophysiological implications in that they allude to the role of parietal circuits in a time-based ProM. The parietal cortex contains multimodal neurons (“attentional neurons”) maintaining the transmutations of activities in extrapersonal space into “egocentric” spatiotemporal coordinates. Therefore, the reliability of the parietal machinery might seem essential for timing events within the context of a space-time coordinate system.157,158 Parietal cortex with its multiple interconnections permits the solution of the time problem within the realm of spatial cognition, dealing under the “same roof,” so to speak, with non-temporal space-based attention, memory and motor intentions.159 Reduced functional competence of parietal regions is expected to differentiate individuals exhibiting time underestimation from those with normal time assessment on a time-based ProM task. Experimentally-created parietal dysfunction increased response time in all task contingencies.160−162 Parietal lesions in humans, particularly on the right side, have long been recognized to lead to a bizarre syndrome of withdrawal from the environment, designated as perceptual hemineglect.163 Regrettably, we do not use the clinical model of hemispatial neglect with the same enthusiasm in order to examine imbalanced intentionality and timing as we do the imbalance of spatial representations.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

287

When individuals with focal left or right hemisphere parietal lesions were compared to normal controls on time-perception tasks, only right-hemisphere lesion patients showed time-perception deficits.109 A patient (AF) with an acute right temporal-parietal stroke described by Snyder and Chatterjee129 was biased to judge ipsilesional stimuli as occurring before contralesional stimuli. For vertically aligned stimuli, AF judged more accurately the temporal order of successive ipsilesional than contralesional stimuli. These findings suggest that the multimodal network disabled by a parietal deficit ceased to provide the service of locating valid environmental cues for time checking (making the cue environment increasingly impoverished). Recall that parietal neurons have the capacity of cue-grouping that helps to preserve the continuity of the selected course, the programing of purposeful movements and control of their environmental trajectories.164,165 The integrity of the prefrontal cortex is also essential for maintaining posterior cortical representations in an active state over long time scales.166 There is a tightly regulated reciprocal interaction between the parietal and prefrontal and orbitofrontal machinery that can be gleaned from some cases of frontal deficit. The nature of this deficit was conceptualized in a number of ways, but Lhermitte167,168 had it right by elegantly demonstrating the hypersensitivity of the posterior cortex to environmental “noise” that is manifested as the “environmental dependency syndrome.” The afflicted patients behave as “though implicit in the environment was an order to respond to the situation in which they found themselves.” With the loss of an anterior cortical filter (a “dynamic filter”169 ), an overactive parietal component of the fronto-parietal push-pull system may conceivably create an excessively cluttered (cue-dense) environment.

Time-Based Prospective Memory as an Instrument of Reward and of Social Skill In the ProM context, reward could be defined as an event whereby the timepredicting system makes certain that the planned behavior stays the course. Neurobiologically, discussions on “wanting” and rewards170 at the outset of ProM, commonly involve a massive literature dealing with the release of dopamine in the nucleus accumbens, striatum, and prefrontal cortex so as to sustain goal pursuit. Single-unit studies have shown increased ventral tegmental activity to rewards and reward predictors and decreased activity when predicted rewards are withheld.171 Dopamine-transmitting neurons appear to label

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

288

J. Glicksohn & M. S. Myslobodsky

environmental stimuli with an appetitive “tag,” thereby predicting and detecting rewards (monitoring miscalculations) and signaling motivationally relevant events.172,173 In 1971, one anthropologist134 insightfully delegated “temporal orientation toward the future” to the category of “deferred gratification” behaviors. We know very little of how the capacity to defer gratification affects delayed timecued ProM in real-life situations.174 The neuronal network of such behaviors varies as a function of the psychological and socio-cultural context in which it is experienced; familiarity with reward, information to determine the appropriate course of action, requirements to suppress previously rewarded responses, to overcome salient irrelevant cues, the probability of penalties for errors in reward prediction, self-monitoring and others.175–178 For example, if the gambling task has it right,98 the midbrain and ventral striatal regions appear to be activated in response to financial rewards, whereas the hippocampus responded to financial penalties when healthy volunteers were submitted to fMRI while “laying a bet” in the magnet. However, the dynamic network spilled into the globus pallidus, thalamus, and cingulate, caudate, insula, and ventral prefrontal cortex in the context of increasing reward or in the context of increasing penalty.176 Parietal activity, in addition to all its other functions, is apparently also essential for reward appreciation. Activity in the parietal cortex appears to correlate with the relative subjective desirability of actions, irrespective of the specific combination of reward magnitude, reward probability, and odds of response associated with the task.179 Parietal neurons are also believed to participate in the appraisal of future actions,180 a mission tightly connected with incentive motivations. More recent studies emphasize the role of the anterior cingulate in reward expectation, reward-based decision-making, particularly in monitoring outcome and signal errors, as well as in selecting appropriate motor responses.181,182 Williams et al.182 conducted single-neuron recordings from human subjects who were scheduled for cingulotomy. In a task that required making specific movements in response to changing monetary rewards, many neurons appeared to be activated in the presence of a diminished reward. Following dorsal anterior cingulate ablation, patients made selectively more errors when they were required to change movement based on reward decline, thereby suggesting that anterior cingulate in humans plays an important role linking reward-related information with the movement ultimately made. One might argue that reward may not seem mandatory for time-based ProM, since “timing the future” is often neutral or associated with negative emotions. On several grounds though,

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

289

this is not a serious objection, since reward may be disguised or confounded by anticipation, tension, error detection and conflict monitoring. More importantly, though, “wanting” a particular timed outcome is akin to secondary reward and does not have to be identical to “liking” of food or mating following an environmental cue.181 Cues that better predict the likelihood of reaching the goal would have an incentive salience (and passionate “wanting”) through reactivation of the circuits in the prefrontal cortex and dorsal striatum that refresh the set goal regardless of its hedonic value (“liking”). That is why real-life choices, as well as risk-taking behaviors when timing the future, entail more complex contingencies, including aversive as well as positively reinforcing outcomes. We also know that reward can be obtained in a self-reflective way or through information about future reinforcements of the “observational learning” (vicarious conditioning) of Bandura,121 the field which today is monopolized by the term “resonance behaviors.”183 As such, ProM also requires a considerable shared awareness (“resonance memory”?) being dependent upon emotional signaling needed to adapt to the socially imposed tempo of the environment. Navigating through a complex space of social circumstances, one must conceivably be guided by “somatic markers” (personal affective and autonomic inputs) and the anticipation of future ones, much like the anticipatory electrodermal activity that is recorded before normal individuals made selections of a “disadvantageous” response.98 What has been termed “myopia of the future” is akin to “goal neglect,” a failure to fully allocate attention on task requirements across time.184,h A deficient frontal cortex would preclude the development of somatic markers,54 and by extension, disrupt regulatory processes needed for anticipating outcomes and sequential time sampling.186,187 Following damage to the ventromedial prefrontal cortex (as well as the amygdala or insular cortices, especially on the right side), individuals are maneuvered by immediate prospects and manifest defects in planning and real-life decision-making, even when having otherwise normal intellectual functions and the absence of psychopathology based on DSM-IV criteria.188,189 Fuster160,190 emphasized the importance of the prefrontal cortices in behaviors with “cross-temporal contingencies,” that is, motor actions, thinking or h We may guess that Damasio’s98

“myopia of the future” is an inverse of the earlier notion of “time spanning” as a “bridge to the future.” An example given by Eysenck185 is that of psychopaths who “know very well what the contingencies are which are involved in their actions; they just disregard them” (p. 143, italics in the original). As Eysenck clarifies, they are unable to maintain “time spanning” by being “at the mercy of the immediate gradient of reinforcement.”

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

290

J. Glicksohn & M. S. Myslobodsky

spoken language that is contingent on events or information from the past or the anticipation of events or information in the future. Neuronal ensembles within the dorsolateral prefrontal cortex compute the sequence of goals and retain their representations with regard to the initial state, thereby maintaining the orderly flow of goal-directed behaviors in which there is a temporal delay between the stimulus and the required response.190–192 Patients with radiologically documented frontal-lobe lesions following a mild to moderate closed head injury manifested distinct anomalies in temporal organization and the execution of tasks as basic as meal preparation. While short sequences of actions are easily produced, longer action sequences requiring procedural knowledge are not correctly executed.93 Such an exceptional difficulty in strategic planning in mundane ProM tasks, particularly of its time-based kind, must be an important underpinning of maladjustment in simple real-world chores. A charming absentmindedness would then evolve into a syndrome begging for a neurological label. Goldstein and colleagues193 described a 51-year-old, right-handed man who showed a marked dissociation between his intact performance on standard neuropsychological tests and his everyday behavior following left frontal lobectomy for a space-occupying lesion (mixed astrocytoma-oligodendroglioma). His behavior was inferior on a test that required goal articulation, plan specification, self-monitoring, and evaluation of outcomes, as well as the establishment of mental “markers” for eliciting specific responses. This syndrome was defined as a “strategy application disorder.”194 With it, a paradigm has been enunciated which has served to aid in simplifying and explaining otherwise unrelated indices of frontal-lobe dysfunction. There is a considerable contribution of mesiofrontal circuits in multitasking. A multitasking procedure submitted to 60 people with circumscribed cerebral lesions and 60 age- and IQ-matched controls suggested that lesions to the left posterior cingulate are associated with deficits on all measures except planning. Remembering task contingencies after a delay was affected by lesions in the region of the left anterior cingulate, and rule-breaking and failures of task switching were additionally found in people with lesions affecting the medial and more polar aspects of Brodmann’s areas 8, 9 and especially 10 as well as the right dorsolateral prefrontal cortex.195 The way a dysfunctional frontal lobe and a deficient resident control mechanism commonly identified as executive functions compromise ProM in general, but predominantly a time-based ProM, is discussed by Janet Cockburn in Chap. 10.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

291

Memory for Future Events: What Lies Ahead? The story of a time-cued ProM summarized in Fig. 1 may seem surprisingly sketchy, and indeed it is. Whereas memory for time is an active area of cognitive neuroscience, time-cued ProM remains in its chronic infancy. However, there are emerging areas of critical interest of what is to come. One might expect studies directed at the understanding of what makes a “point in time” such an important resident of the human mind. Doob’s134 conviction that “deferment and renunciation” of [immediate] gratification is the foundation of anticipation gives a somewhat different perspective on ProM. Anyone who plans far ahead must have realized that delayed goal is at risk of perpetual updating and may be deserted altogether by some more enticing goals explored on the way (“planning fallacy” discussed by Trope and Liberman53 ). One explanation for planning fallacy is the effects of reinforcers differing in size and delay. Chapter 5 reads as a call to behavior scientists and cognitive scientists to pool resources in a study of time-cued ProM together with time discounting. Although a time discounting perspective has so much to offer to ProM research, with few exceptions,196,197 the neuronal mechanisms of temporal discounting are seldom explored. An effort to examine neuronal process underlying temporal distance changes of people’s representations of future events within the context of the temporal construal theory (Chap. 7) is a far more ambitious goal. The more cognitive neuroscientists press forward to “testing the theory,” the sooner diverse research paradigms will fragment it into a mosaic of more concrete elements. We certainly need to learn more of time-cued ProM developmental trajectories and its decline (see Chap. 8). Because children typically acquire their time preferences, much like food preferences, from their parents, their timing pattern might become similar. A study by Mäntylä and Carelli (see Chap. 8) provides a stimulus for looking into the distinction between acquired and inherited (or genetic) “timing phenotypes.” As with any off-line activity, delayed tasks always have more room for freedom and thus a stamp of personality, imagery, and postponed reward utilization may be found at each turn of the road (Chaps. 6 and 7). They comprise components that are self-generated, environmentallymodulated, content (task, goal)-related that are often difficult to dissociate in human studies, inasmuch as delayed goals of “mental travel” depend upon selfawareness (episodic memory) and awareness of others; they are in need of forecast updating and estimations of their accuracy, that is, variables that are

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

292

J. Glicksohn & M. S. Myslobodsky

essentially impossible to model in experimental animals.198 We cannot pretend to be able to guess how each step of the ProM sequence is carried out. One might think of time-cued ProM in terms of reading a hypertext that may introduce nonlinear dynamics by engaging attention to its key words and then failing to disengage from appraisal of topics that were initially unintended.199 This metaphor suggests that the reader would carry on aimlessly meandering through various themes. Hitherto poorly understood interactions between prefrontal and posterior areas contain diversion-resistant circuits protecting against this meandering,200 thereby allowing intended actions to be carried out regardless of “interrupting, switching, conscious control processes.”39 In the behavioral literature, this factor is examined under the heading of sustained choice procedures which demands that a choice for a delayed reinforcer is sustained once the choice has been made without defecting to smaller, more immediate reinforcers encountered during the delay period. On the other hand, and consistent with “random walks” in other domains, Wilkins’ Random Walk Model posits that “with each entry into the part of the multidimensional space in which thoughts would cue the to-be-remembered task, the size of this part of the space would increase, thus ensuring that more and more tangential thoughts would on subsequent occasions succeed in cueing the to-be-remembered task.”117 One might wonder how this potential increment in randomness would provide “resource support” or environmental channeling. The popular metaphor for achieving a goal (e.g. in science) as filling in an elaborate crossword puzzle proposed by the philosopher Susan Haack201 suggests that the swelling background (of the “multidimensional space?”) makes the imbedded cue recognition (“solution”) more likely. There is no a priori judgment of whether the cues (sensory experience, social interactions, or physical activity) are used to time the future in the manner of “hypertext” (and thus always in need of control) or whether the cues enter into a rigid “crossword-type” programatic machine channeling to a planned event. The choice, at this stage, would depend on the readers’ interests and tastes. Yet clearly, these two orthogonal processes of timing could be tested using modern techniques. Perhaps in this way we could learn why “procrastination behavior” appears to be an innate human trend.202 Computer modeling of cue sampling strategies may help our understanding not only of the nature of this mysterious asymmetry in meeting the deadlines but could also possibly tell the planners how to modify it. Their exploration would help to understand ProM in those whose time drags due to paralysis of action;

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

293

who are excessively passive and apathetic; slow cognitively or motorically (“bradiphrenic,” “bradikinetic”) as opposed to the “delighted” and “restless ones” whose time flies. That is where psychological time may become warped, so that its progression may halt, reverse in reveries, or jump to the future. We are wondering whether or not the errors of “telescoping” would correlate with the systematic errors made in timing the future. To the extent that punctuality reflects a choice determined by expected benefit, any ProM design should be susceptible to confounding effects of timing instruction since the timing of experiences is translated into some other perceptual quality than time (see Chap. 4). A related practical problem is in exploring timing the future in the context of real events of universal significance (disasters, war traumas, epidemics, mass exodus, incarceration, or emigration) and their modulation by sickness, aging and drug addiction (see Chap. 5). We have yet to learn when timing errors would vary as a function of the (subjective) probabilities of untoward events, their salience, and individuals’ capacity for visualization or whether they are similar to systematic biases in judgment, regardless of the competence for timing. We think rather forgivingly of people who violate deadlines, but seldom do we feel that such “time neglect” is akin to some incapacitating neurological impediment, such as anosognosia (ignorance or incapacity), when people are not aware of their involuntary movements or of their hemiplegia. And in general, how is the awareness of “lived time”203 when altered pharmacologically or neurologically translated in timing the future? There have been no ProM distinctions based on neuropharmacology. Research on the effects of pharmacotherapy on time-cued ProM, particularly in psychologically and neurologically impaired individuals, is hitherto wanting and remains another area of research opportunity for the future. The underlying mechanisms of motivation, language, and social contingencies in timing the future remain poorly understood and call for far more data before many hypotheses acquired “solid legs” to stand on. That direction was briefly discussed during the Third International Conference on Memory, 16–20 July, 2001, Valencia (Spain) (http://www.psy.herts.ac.uk/pub/ l.kvavilashvili/symp_Valencia.html). Advances in cutting edge imaging technologies and a flurry of fMRI studies somewhat shifted attention away from classical ProM paradigms. Such advances have been influential in portraying the pictures of ProM in the shapes and locations of brain activation. Some of these structures are represented in Fig. 3.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

294

J. Glicksohn & M. S. Myslobodsky

Impulse to implement/ abort plan (SMA) Goal (re)scheduling Reward processing Inhibition of prepotent responses (PF) Inaccuracy recognition Conflict resolution (AC)

Reward processing (PA)

Tagging goals to social conditions (OF) Attentional/ mnemonic operations Cue maintenance (PA, PF, H)

Emotionally valenced thoughts processing (OF) Modulation of signals from amygdala and hypothalamus; Impulse control (OF)

Future time perspective ‘Ideational trip’ (PA, PF)

Cue-reward associations (OF) Wanting (VTA, NA)

Episodic memory (H, A)

Fig. 3. Reward-related structures involved in time-cued ProM. Schematic representation of localization of structures implicated in time-cued ProM from episodic memory and “wanting” to goal implementation. In order to facilitate visualization, tentative brain regions are portrayed as though repeatedly populating an unending “time coil” on different segments of time cycles. This recurrent recruitment is associated with a significant degree of functional heterogeneity of structures represented by the same anatomical label. For example, there is a dissociation of processing of emotionally valenced plans and signals within some of these brain areas (e.g. lateral and medial parietal and prefrontal cortex, anteriormedial and posterior-lateral orbitofrontal cortex, etc.). Their representation is not reflected in the schematic since we know little about their roles in timing the future. For identical reasons, the size and interconnections of sampled brain areas with numerous “time-keeping circuits” were omitted. Abbreviations in brackets: AC, anterior cingulate cortex; A, amygdala; H, hippocampus; OF, orbitofrontal cortex; PA, superior and inferior parietal cortex as well as mesial parietal cortex; PF, dorsolateral, dorsomedial and mesial prefrontal cortex; SMA, supplementary motor area.

However, at the risk of repeating familiar arguments, it may be worthwhile pointing out that the degree of involvement of different subdivisions of structures dealing with planning, the storing of goals, avoiding distracting cues and inhibiting undesirable motor responses, focusing attention on local pacemakers and the rhythm of partial reward, tagging the goals to social information,

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

295

monitoring errors and, finally, providing a decisive impulse to implement the plan, is unlikely to be identical for all types of intentions, time frames, and paradigms employed. The levels of neuronal activity as reflected in metabolic maps published thus far do not tell us whether the recruited areas respond to intentions in a predetermined order, and which participants of the network are engaged in parallel by the tasks and flexibly change their role to various environmental circumstances as a function of delay from intention to execution. Very little work has been done on the neurophysiological basis of timing the future. Further functional imaging studies, particularly quantitative electroencephalography (EEG) and MEG, are needed to profile the phase locking between diverse brain networks along the course of long delayed time-cued ProM. Thus far, trying to detail the road of time-cued ProM in the language of functional neuroanatomy is likely to elicit an image portrayed in Fig. 4. Nevertheless, there may be many wonders in store for us.

Prospective memory

Fig. 4. A portrayal of the functional anatomy of time-cued ProM is still confusing.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

296

J. Glicksohn & M. S. Myslobodsky

Acknowledgments The authors thank Shane Frederick for providing his unpublished materials that were helpful during the preparation of this chapter. Alexandra ParmetMyslobodsky is credited with drawing attention to an episode from Amos Oz, Macbeth’s ideation, and Professor Higgins’ delayed plans. Chanita Goodblatt of the Department of Foreign Literatures & Linguistics Laboratory for Cognitive Poetics (Ben-Gurion University) was our arbiter along the way.

References 1. Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Application. Mahwah, NJ: Lawrence Erlbaum Associates, 1996. 2. Seifert C M, Patalano A L. Opportunism in memory: Preparing for chance encounters. Current Directions in Psychological Science 2001; 10(6): 198–201. 3. Wilkins A J. Remembering to do things in the laboratory and everyday life. Acta Neurologica Scandinavica 1986; 74: 109–112. 4. Cole M, Means B. Comparative Studies of How People Think. Cambridge, MA: Harvard University Press, 1981. 5. Sperber M. Masks of Loneliness. Alfred Adler in Perspective. New York: Macmillan Publishing Co., Inc., 1974. 6. Spence K W. Behavior Theory and Learning. Englewood Cliffs, NJ: Prentice Hall, 1960. 7. Humphrey G. Thinking: An Introduction to Its Experimental Psychology. London: Methuen, 1953. 8. Hilgard E R. Divided Consciousness: Multiple Controls in Human Thought and Action. New York: Wiley, 1977. 9. Orne M T, Dinges D F, Bloom P B. Hypnosis. In Kaplan H, Sadock B, eds. Comprehensive Textbook of Psychiatry. 6th edn. Baltimore: Williams & Wilkins, 1995: 1807–1821. 10. Lewin K. Intention, will and need. In Rapaport D, ed. Organization and Pathology of Thought. New York: Columbia University Press, 1951: 95–153. 11. Ash M G. Gestalt Psychology in German Culture, 1890–1967: Holism and the Quest for Objectivity. Cambridge: Cambridge University Press, 1995. 12. Mäntylä T, Sgaramella T. Interrupting intentions: Zeigarnik-like effects in prospective memory. Psychological Research-Psychologische Forschung 1997; 60(3): 192–199. 13. Lewin K. A Dynamic Theory of Personality: Selected Papers. New York: McGraw-Hill, 1935. 14. Lewin K. The Conceptual Representation and the Measurement of Psychological Forces. New York: Harper, 1938. 15. Köhler W. Gestalt Psychology: An Introduction to New Concepts in Modern Psychology. New York: Liveright, 1947. 16. Kadar E E, Shaw R E. Toward an ecological field theory of perceptual control of locomotion. Ecological Psychology 2000; 12(2): 141–180.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

297

17. van Leeuwen C, Stins J. Perceivable information or the happy marriage between ecological psychology and Gestalt. Philosophical Psychology 1994; 7(2): 267–285. 18. Harris J E, Wilkins A J. Remembering to do things: A theoretical framework and an illustrative experiment. Human Learning 1982; 1: 123–136. 19. Ceci S J, Bronfenbrenner U. “Don’t forget to take the cupcakes out of the oven”: Prospective memory, strategic time-monitoring, and context. Child Development 1985; 56(1): 152–164. 20. Pittenger D J. The two paradigms of persistence. Genetic, Social and General Psychology Monographs 2002; 128(3): 237–268. 21. Svartdal F. Extinction after partial reinforcement: Predicted vs. judged persistence. Scandinavian Journal of Psychology 2003; 44(1): 55–64. 22. Carstensen L L, Isaacowitz D M, Charles S T. Taking time seriously: A theory of socioemotional selectivity. American Psychologist 1999; 54(3): 165–181. 23. Goschke T, Kuhl J. Representation of intentions: Persisting activation in memory. Journal of Experimental Psychology: Learning, Memory and Cognition 1993; 19(5): 1211–1226. 24. Patalano A L, Seifert C M. Opportunistic planning: Being reminded of pending goals. Cognitive Psychology 1997; 34(1): 1–36. 25. Kliegel M, Martin M, McDaniel M A, Einstein G O. Varying the importance of a prospective memory task: Differential effects across time- and event-based prospective memory. Memory 2001; 9(1): 1–11. 26. Graf P, Schacter D L. Implicit and explicit memory for new associations in normal and amnesic subjects. Journal of Experimental Psychology: Learning, Memory, and Cognition 1985; 11(3): 501–518. 27. Kvavilashvili L, Ellis J A. Varieties of intention: Some distinctions and classifications. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Lawrence Erlbaum Associates, 1996: 23–51. 28. Ericsson K A, Charness N. Expert performance: Its structure and acquisition. American Psychologist 1994; 49(8): 725–747. 29. Woodworth R S. Reinforcement of perception. American Journal of Psychology 1947; 60: 119–124. 30. Miller G A, Galanter E, Pribram K H. Plans and the Structure of Behavior. New York: Holt, Rinehart and Winston, 1960. 31. Oz A. A Tale of Love and Darkness. London: Ghatto & Windus, 2003. 32. Kelly W E. Anxiety and the prediction of task duration: A preliminary analysis. Journal of Psychology 2002; 136(1): 53–58. 33. Einstein G O, Richardson S L, Guynn M J, Cunfer A R, McDaniel M A. Aging and prospective memory: Examining the influences of self-initiated retrieval processes. Journal of Experimental Psychology: Learning, Memory and Cognition 1995; 21(4): 996–1007. 34. Robbins S E. Bergson, perception and Gibson. Journal of Consciousness Studies 2000; 7(5): 23–45. 35. Sellen A J, Louie G, Harris J E, Wilkins A J. What brings intentions to mind? An in situ study of prospective memory. Memory 1997; 5(4): 483–507. 36. Boltz M G. Task predictability and remembered duration. Perception & Psychophysics 1998; 60(5): 768–784.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

298

J. Glicksohn & M. S. Myslobodsky

37. Dibble P. Embedded Systems Programming: Deadline Scheduling, 2003. 38. Frederick S, Loewenstein G, O’Donoghue T. Time discounting and time preference: A critical review. Journal of Economic Literature 2002; 40(2): 351–401. 39. Graf P, Uttl B. Prospective memory: A new focus for research. Consciousness and Cognition 2001; 10(4): 437–450. 40. Kruglanski A W. The psychology of being right: The problem of accuracy in social perception and cognition. Psychological Bulletin 1989; 106(3): 395–409. 41. Koriat A, Goldsmith M, Pansky A. Toward a psychology of memory accuracy. Annual Review of Psychology 2000; 51: 481–537. 42. Ungerleider L G, Haxby J V. “What” and “where” in the human brain. Current Opinion in Neurobiology 1994; 4(2): 157–165. 43. Goodale M A, Milner A D. Separate visual pathways for perception and action. Trends in Neurosciences 1992; 15(1): 20–25. 44. Livingstone M, Hubel D. Segregation of form, color, movement, and depth: Anatomy, physiology, and perception. Science 1988; 240(4853): 740–749. 45. Aladzhalova N A. Slow Electrical Processes in the Brain. New York: Elsevier, 1964. 46. Staddon J E R. Interval timing: Memory, not a clock. Trends in Cognitive Sciences 2005; 9(7): 312–314. 47. Staddon J E R, Higa J J. Time and memory: Towards a pacemaker-free theory of interval timing. Journal of the Experimental Analysis of Behavior 1999; 71(2): 215–251. 48. Harris J E. Remembering to do things: A forgotten topic. In Harris J E, Morris P E, eds. Everyday Memory, Actions, and Absent-Mindedness. New York: Academic Press, 1984: 71–92. 49. Kvavilashvili L. Automatic or controlled? Rehearsal and retrieval processes in everyday time- and event-based prospective memory tasks. The 2nd Conference on Prospective Memory: 2005, Zurich. 50. Francis-Smythe J A, Robertson I T. On the relationship between time management and time estimation. British Journal of Psychology 1999; 90: 333–347. 51. Huttenlocher J, Hedges L V, Bradburn N M. Reports of elapsed time: Bounding and rounding processes in estimation. Journal of Experimental Psychology: Learning, Memory, and Cognition 1990; 16(2): 196–213. 52. Huttenlocher J, Hedges L, Prohaska V. Hierarchical organization in ordered domains: Estimating the dates of events. Psychological Review 1988; 95(4): 471–484. 53. Trope Y, Liberman N. Temporal construal. Psychological Review 2003; 110(3): 403–421. 54. Damasio A R. The somatic marker hypothesis and the possible functions of the prefrontal cortex. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 1996; 351(1346): 1413–1420. 55. Strotz R H. Myopia and inconsistency in dynamic utility maximization. Review of Economic Studies 1956; 23:165–180. 56. McClure S M, Laibson D I, Loewenstein G, Cohen J D. Separate neural systems value immediate and delayed monetary rewards. Science 2004; 306(5695): 503–507. 57. Venneri A, Pestell S, Nichelli P. A preliminary study of the cognitive mechanisms supporting time estimation. Perceptual and Motor Skills 2003; 96(3): 1093–1106.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

299

58. Brown S W, Boltz M G. Attentional processes in time perception: Effects of mental workload and event structure. Journal of Experimental Psychology: Human Perception and Performance 2002; 28(3): 600–615. 59. Zakay D, Block R A, Tsal Y. Prospective duration estimation and performance. In Gopher D, Koriat A, eds. Attention and Performance XVII. Cognitive Regulation of Performance: Interaction of Theory and Application. Cambridge, MA: MIT Press, 1999: 557–580. 60. Block R A. Experiencing and remembering time: Affordances, context, and cognition. In Zakay D, Levin I, eds. Time and Human Cognition: A Life-span Perspective. Amsterdam: Elsevier Science Publishers, 1989: 333–363. 61. Block R A. Models of psychological time. In Block R A, ed. Cognitive Models of Psychological Time. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 1990: 1–35. 62. Fraisse P. The Psychology of Time. Westport, Conn: Greenwood, 1975. 63. Fraisse P. Perception and estimation of time. Annual Review of Psychology 1984; 35: 1–36. 64. Ornstein R E. On the Experience of Time. Harmondsworth, Middlesex: Penguin, 1969. 65. Poynter D. Judging the duration of time intervals: A process of remembering segments of experience. In Levin I, Zakay D, eds. Time and Human Cognition: A Life-span Perspective. Amsterdam: Elsevier Science Publishers, 1989: 305–331. 66. James W. The Principles of Psychology. New York: Dover, 1890. 67. Glicksohn J. Temporal cognition and the phenomenology of time: A multiplicative function for apparent duration. Consciousness and Cognition 2001; 10(1): 1–25. 68. Droit-Volet S, Wearden J. Speeding up an internal clock in children? Effects of visual flicker on subjective duration. The Quarterly Journal of Experimental Psychology 2002; 55B(3): 193–211. 69. Frankenhaeuser M. Estimation of Time: An Experimental Study. Stolkholm: Almqvist & Wiksell, 1959. 70. Penton-Voak I S, Edwards H, Percival A, Wearden J H. Speeding up an internal clock in humans? Effects of click trains on subjective duration. Journal of Experimental Psychology: Animal Behavior Processes 1996; 22(3): 307–320. 71. Treisman M, Brogan D. Time perception and the internal clock: Effects of visual flicker on the temporal oscillator. European Journal of Cognitive Psychology 1992; 4(1): 41–70. 72. Zakay D, Nitzan D, Glicksohn J. The influence of task difficulty and external tempo on subjective time estimation. Perception & Psychophysics 1983; 34(5): 451–456. 73. Aschoff J, Daan S. Human time perception in temporal isolation: Effects of illumination intensity. Chronobiology International 1997; 14(6): 585–596. 74. Cook C M, Koren S A, Persinger M A. Subjective time estimation by humans is increased by counterclockwise but not clockwise circumcerebral rotations of phase-shifting magnetic pulses in the horizontal plane. Neuroscience Letters 1999; 268(2): 61–64. 75. Glicksohn J. Subjective time estimation in altered sensory environments. Environment and Behavior 1992; 24(5): 634–652. 76. Glicksohn J. Entering trait and context into a cognitive-timer model for time estimation. Journal of Environmental Psychology 1996; 16(4): 361–370.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

300

J. Glicksohn & M. S. Myslobodsky

77. Macar F. Temporal judgments on intervals containing stimuli of varying quantity, complexity and periodicity. Acta Psychologica 1996; 92(3): 297–308. 78. Martínez B B. The role of cognitive changes in immediate and remote prospective time estimations. Acta Psychologica 1994; 85(2): 99–121. 79. Boltz M G. Effects of event structure on retrospective duration judgments. Perception & Psychophysics 1995; 57(7): 1080–1096. 80. Burrardi L. Factors affecting the filled-duration illusion in the auditory, tactual and visual modalities. Perception & Psychophysics 1972; 10: 292–294. 81. Flaherty M G. A Watched Pot: How We Experience Time. New York: New York University Press, 1999. 82. Popper K, Eccles J. The Self and Its Brain. Berlin: Springer, 1977. 83. Bergson H. The Evolution of Life: Mechanism and Teleology. New York: Henry Holt and Company, 1911. 84. Dombrowsky W R. The social dimensions of warning and the transition from folk wisdom to laymanship. In Nemec J, Nigg J M, Siccardi F, eds. Prediction and Perception of Natural Hazards. Perugia, Italy: Kluwer Academic Publishers, 1993: 23–28. 85. Sauer J. Prospective memory: A secondary task with promise. Applied Ergonomics 2000; 31(2): 131–137. 86. Morita T. Determinants of time-monitoring on prospective memory tasks. Japanese Journal of Psychology 1998; 69(2): 137–142. 87. Conti R. Time flies: Investigating the connection between intrinsic motivation and the experience of time. Journal of Personality 2001; 69(1): 1–26. 88. Rude S S, Hertel P T, Jarrold W, Covich J, Hedlund S. Depression-related impairments in prospective memory. Cognition and Emotion 1999; 13(3): 267–276. 89. Atance C M, O’Neill D K. Episodic future thinking. Trends in Cognitive Sciences 2001; 5(12): 533–539. 90. Bschor T, Ising M, Bauer M et al. Time experience and time judgment in major depression, mania and healthy subjects. A controlled study of 93 subjects. Acta Psychiatrica Scandinavica 2004; 109(3): 222–229. 91. Mäntylä T. Activating actions and interrupting intentions: Mechanisms of retrieval sensitization in prospective memory. In Brandimonte M, Einstein G O, McDaniel M A, eds. Prospective Memory: Theory and Applications. Mahwah, NJ: Lawrence Erlbaum Associates, 1996: 93–113. 92. Buonomano D V, Karmarkar U R. How do we tell time? Neuroscientist 2002; 8(1): 42–51. 93. Fortin S, Godbout L, Braun C M J. Strategic sequence planning and prospective memory impairments in frontally lesioned head trauma patients performing activities of daily living. Brain and Cognition 2002; 48(2&3): 361–365. 94. Ullsperger M, von Cramon D Y. Subprocesses of performance monitoring: A dissociation of error processing and response competition revealed by event-related fMRI and ERPs. NeuroImage 2001; 14(6): 1387–1401. 95. Luria A R. Higher Cortical Functions in Man. London: Tavistock, 1966.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

301

96. Ingvar D H. Memory of the future: An essay on the temporal organization of conscious awareness. Human Neurobiology 1985; 4(3): 127–136. 97. Wall P M, Messier C. The hippocampal formation–orbitomedial prefrontal cortex circuit in the attentional control of active memory. Behavioural Brain Research 2001; 127(1&2): 99–117. 98. Damasio A R. Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Avon, 1994. 99. Hester R, Fassbender C, Garavan H. Individual differences in error processing: A review and reanalysis of three event-related fMRI studies using the GO/NOGO task. Cerebral Cortex 2004; 14(9): 986–994. 100. Rubia K, Smith A B, Brammer M J, Taylor E. Right inferior prefrontal cortex mediates response inhibition while mesial prefrontal cortex is responsible for error detection. NeuroImage 2003; 20(1): 351–358. 101. Botvinick M M, Cohen J D, Carter C S. Conflict monitoring and anterior cingulate cortex: An update. Trends in Cognitive Sciences 2004; 8(3): 539–546. 102. Tversky A, Shafir E. Choice under conflict: The dynamics of deferred decision. Psychological Science 1992; 3(6): 358–361. 103. Kvavilashvili L, Mandler G. Out of one’s mind: A study of involuntary semantic memories. Cognitive Psychology 2004; 48(1): 47–94. 104. Theeuwes J, De Vries G J, Godjin R. Attentional and oculomotor capture with static singletons. Perception & Psychophysics 2003; 65(5): 735–746. 105. Sokolov E N, Nezlina N, Polyansky V B, Evtikhin D V. Orienting reflex: “Targeting reaction” and “searchlight of attention.” Zhurnal Vysshei Nervnoi Deyatelnosti Imeni I P Pavlova 2001; 51: 421–437. 106. Cohen B H. The motor theory of voluntary thinking. In Davidson R J, Schwartz G E, Shapiro D, eds. Consciousness and Self-regulation: Advances in Research and Theory. New York: Plenum, 1986: 19–54. 107. Pastor M A, Artieda J, Jahanshahi M, Obeso J A. Time estimation and reproduction is abnormal in Parkinson’s disease. Brain 1992; 115: 211–225. 108. Katai S, Maruyama T, Hashimoto T, Ikeda S. Event based and time based prospective memory in Parkinson’s disease. Journal of Neurology, Neurosurgery & Psychiatry 2003; 74(6): 704–709. 109. Harrington D L, Haaland K Y, Knight R T. Cortical networks underlying mechanisms of time perception. Journal of Neuroscience 1998; 18(3): 1085–1095. 110. Meck W H. Affinity for the dopamine D2 receptor predicts neuroleptic potency in decreasing the speed of an internal clock. Pharmacology, Biochemistry and Behavior 1986; 25(6): 1185–1189. 111. Walton M E, Devlin J T, Rushworth M F S. Interactions between decision making and performance monitoring within prefrontal cortex. Nature Neuroscience 2004; 7(11): 1259–1265. 112. Costermans J, Desmette D. A method for describing time-monitoring strategies in a prospective memory setting. Cahiers De Psychologie Cognitive/Current Psychology of Cognition 1999; 18(3): 289–306.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

302

J. Glicksohn & M. S. Myslobodsky

113. Wilkins A J, Baddeley A. Remembering to recall in everyday life: An approach to absentmindedness. In Gruneberg M M, Morris C, Sykes R N, eds. Practical Aspects of Memory. San Diego, CA: Academic Press, 1978: 27–34. 114. Cagan P. The monetary dynamics of hyperinflation. In Friedman M, ed. Studies in the Quantity Theory of Money. Chicago: Chicago University Press, 1956: 233–269. 115. Bouchaud J P. The subtle nature of financial random walks. Chaos 2005; 15(2): 26–104. 116. Révész P. Random Walk in Random and Non-random Environments. London: World Scientific, 2005. 117. Wilkins A J. 2005 (personal communication). 118. Shallice T, Burgess P. The domain of supervisory processes and temporal organization of behaviour. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 1996; 351(1346): 1405–1411. 119. Bernstein N A. The Co-ordination and Building of Movements. New York: Pergamon, 1967. 120. Neisser U. Cognition and Reality. San Francisco: W. H. Freeman, 1976. 121. Bandura A. Human agency in social cognitive theory. American Psychologist 1989; 44(9): 1175–1184. 122. Critchley M. The Divine Banquet of the Brain and Other Essays. New York: Raven Press, 1955. 123. Dennett D C. Consciousness Explained. Boston: Little, Brown and Company, 1991. 124. Sakai K, Passingham R E. Prefrontal interactions reflect future task operations. Nature Neuroscience 2003; 6(1): 75–81. 125. Eagleman D M, Sejnowski T J. Motion integration and postdiction in visual awareness. Science 2000; 287(5460): 2036–2038. 126. Libet B, Gleason C A, Wright E W, Pearl D K. Time of conscious intention to act in relation to onset of cerebral activity (Readiness-Potential): The unconscious initiation of a freely voluntary act. Brain 1983; 106: 623–642. 127. Jeannerod M. Mental Imagery in the Motor Context. Oxford: Blackwell, 1997. 128. Passingham D, Sakai K. The prefrontal cortex and working memory: Physiology and brain imaging. Current Opinion in Neurobiology 2004; 14(2): 163–168. 129. Snyder J J, Chatterjee A. Spatial-temporal anisometries following right parietal damage. Neuropsychologia 2004; 42(12): 1703–1708. 130. Grondin S. From physical time to the first and second moments of psychological time. Psychological Bulletin 2001; 127(1): 22–44. 131. Keough K A, Zimbardo P G, Boyd J N. Who’s smoking, drinking, and using drugs? Time perspective as a predictor of substance use. Basic and Applied Social Psychology 1999; 21(2): 149–164. 132. Zakay D, Fallach E. Immediate and remote time estimation — A comparison. Acta Psychologica 1984; 57(1): 69–81. 133. Schiffman N, Greist-Bousquet S. The effect of task interruption and closure on perceived duration. Bulletin of the Psychonomic Society 1992; 30(1): 9–11. 134. Doob L W. Patterning of Time. New Haven, CT: Yale University Press, 1971.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

303

135. Ellis J A. Memory for future intentions: Investigating pulses and steps. In Gruneberg M M, Morris P E, Sykes R N, eds. Practical Aspects of Memory: Current Research and Issues. Chichester: John Wiley, 1988: 371–376. 136. Hinton S C, Meck W H. Frontal-striatal circuitry activated by human peak-interval timing in the supra-seconds range. Cognitive Brain Research 2004; 21(2): 171–182. 137. Bolbecker A R, Cheng Z X, Felsten G et al. Two asymmetries governing neural and mental timing. Consciousness and Cognition 2002; 11(2): 265–272. 138. Buchwald N A, Hull C D, Levine M S, Villablance J. The basal ganglia and the regulation of response and cognitive sets. In Brazier MAB, ed. Growth and Development of the Brain. New York: Raven Press, 1975: 171–189. 139. Marsh R L, Hicks J L. Event-based prospective memory and executive control of working memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 1998; 24(2): 336–349. 140. Tulving E, Markowitsch H J. Episodic and declarative memory: Role of the hippocampus. Hippocampus 1998; 8(3): 198–204. 141. Goodman S, Myslobodsky M. Dual-cue model of time-cued prospective remembering. Abstracts/Society for Neuroscience 2004, San Diego, CA. 142. Hacker W, Auerbach I, Hagendorf E, Rudolf M. Carrying out of delayed time-based tasks — Self-initiated use of environmental support? Zeitschrift Fur Experimentelle Psychologie 1999; 46(2): 81–106. 143. Gibson J J. Events are perceivable but time is not. In Fraser J T, Lawrence N, eds. The Study of Time II. New York: Springer-Verlag, 1975: 295–301. 144. Gibson J J. The Ecological Approach to Visual Perception. Boston: Houghton Mifflin, 1979. 145. Caljouw S R, van der Kamp J, Savelsbergh G J P. Catching optical information for the regulation of timing. Experimental Brain Research 2004; 155(4): 427–438. 146. Shapiro J K, Shapiro L R, Russell S, Alper J S. A role for executive processes in prospective remembering. Brain and Cognition 1998; 37(1): 175–177. 147. Robertson L C. From Gestalt to Neo-Gestalt. In Knapp T J, Robertson L C, eds. Approaches to Cognition: Contrasts and Controversies. Hillsdale, NJ: Erlbaum, 1986: 159–188. 148. Thorndyke P W. Distance estimation from cognitive maps. Cognitive Psychology 1981; 13(4): 526–550. 149. Kemp S. Gestalt grouping effects in locating past events on timelines. Acta Psychologica 1987; 64(2): 139–149. 150. Glicksohn J. Rotation, orientation and cognitive mapping. American Journal of Psychology 1994; 107(1): 39–52. 151. Tversky B, Schiano D J. Perceptual and conceptual factors in distortions in memory for graphs and maps. Journal of Experimental Psychology: General 1989; 118(4): 387–398. 152. Friedman W J. Memory for the time of past events. Psychological Bulletin 1993; 113(1): 44–66. 153. Huttenlocher J, Hedges L V, Duncan S. Categories and particulars: Prototype effects in estimating spatial location. Psychological Review 1991; 98(3): 352–376.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

304

J. Glicksohn & M. S. Myslobodsky

154. McCabe V, Balzano G J, eds. Event Cognition: An Ecological Perspective. Hillsdale, NJ: Erlbaum, 1986. 155. Schmajuk N A. Role of the hippocampus in temporal and spatial navigation: An adaptive neural network. Behavioural Brain Research 1990; 39(3): 205–229. 156. Haggard P, Clark S, Kalogeras J. Voluntary action and conscious awareness. Nature Neuroscience 2002; 5(4): 382–385. 157. Freyd J J. Dynamic mental representations. Psychological Review 1987; 94(4): 427–438. 158. Boroditsky L, Ramscar M. The roles of body and mind in abstract thought. Psychological Science 2002; 13(2): 185–189. 159. Shannon B J, Buckner R L. Functional-anatomic correlates of memory retrieval that suggest nontraditional processing roles for multiple distinct regions within posterior parietal cortex. Journal of Neuroscience 2004; 24(45): 10084–10092. 160. Fuster J M. Synopsis of function and dysfunction of the frontal lobe. Acta Psychiatrica Scandinavica 1999; 99: 51–57. 161. Quintana J, Fuster J M. Mnemonic and predictive functions of cortical neurons in a memory task. NeuroReport 1992; 3(8): 721–724. 162. Quintana J, Fuster J M. Spatial and temporal factors in the role of prefrontal and parietal cortex in visuomotor integration. Cerebral Cortex 1993; 3(2): 122–132. 163. Myslobodsky M. Awareness salvaged by cunning: Rehabilitation by deception in audiovisual neglect. In Myslobodsky M, ed. The Mythomanias: The Nature of Deception and Self-deception. Hillsdale, NJ: Lawrence Erlbaum Associates, 1997. 164. Mohl W, Pfurtscheller G. The role of the right parietal region in a movement time estimation task. NeuroReport 1991; 2(6): 309–312. 165. Coull J T. fMRI studies of temporal attention: Allocating attention within, or towards, time. Cognitive Brain Research 2004; 21(2): 216–226. 166. Levy R, Goldman-Rakic P S. Segregation of working memory functions within the dorsolateral prefrontal cortex. Experimental Brain Research 2000; 133(1): 23–32. 167. Lhermitte F. Human autonomy and the frontal lobes II. Patient behavior in complex and social situations: The “environmental dependency syndrome.” Annals of Neurology 1986; 19(4): 335–343. 168. Lhermitte F, Pillon B, Serdaru M. Human autonomy and the frontal lobes I. Imitation and utilization behavior: A neuropsychological study of 75 patients. Annals of Neurology 1986; 19(4): 326–334. 169. Shimamura A P. The role of the prefrontal cortex in dynamic filtering. Psychobiology 2000; 28(2): 207–218. 170. Berridge K C, Robinson T E. What is the role of dopamine in reward: Hedonic impact, reward learning, or incentive salience? Brain Research Reviews 1998; 28(3): 309–369. 171. Martin L E, Potts G F. Reward sensitivity in impulsivity. NeuroReport 2004; 15(9): 1519–1522. 172. Schultz W. Predictive reward signal of dopamine neurons. Journal of Neurophysiology 1998; 80(1): 1–27. 173. Schultz W. Reward signaling by dopamine neurons. Neuroscientist 2001; 7(4): 293–302.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

What it Takes to Remember the Future

305

174. Metcalfe J, Mischel W. A hot/cool-system analysis of delay of gratification: Dynamics of willpower. Psychological Review 1999; 106(1): 3–19. 175. Elliott R, Dolan R J, Frith C D. Dissociable functions in the medial and lateral orbitofrontal cortex: Evidence from human neuroimaging studies. Cerebral Cortex 2000; 10(3): 308–317. 176. Elliott R, Friston K J, Dolan R J. Dissociable neural responses in human reward systems. Journal of Neuroscience 2000; 20(16): 6159–6165. 177. O’Doherty J P. Reward representations and reward-related learning in the human brain: Insights from neuroimaging. Current Opinion in Neurobiology 2004; 14(6): 769–776. 178. Ullsperger M, von Cramon D Y. Error monitoring using external feedback: Specific roles of the habenular complex, the reward system, and the cingulate motor area revealed by functional magnetic resonance imaging. Journal of Neuroscience 2003; 23(10): 4308–4314. 179. Dorris M C, Glimcher P W. Activity in posterior parietal cortex is correlated with the relative subjective desirability of action. Neuron 2004; 44(2): 365–378. 180. Sugrue L P, Corrado G S, Newsome W T. Matching behavior and the representation of value in the parietal cortex. Science 2004; 304(5678): 1782–1787. 181. Bush G, Vogt B A, Holmes J et al. Dorsal anterior cingulate cortex: A role in reward-based decision making. Proceedings of the National Academy of Sciences of the United States of America 2002; 99(1): 523–528. 182. Williams Z M, Bush G, Rauch S L, Cosgrove G R, Eskandar E N. Human anterior cingulate neurons and the integration of monetary reward with motor responses. Nature Neuroscience 2004; 7(12): 1370–1375. 183. Rizzolatti G, Fadiga L, Fogassi L, Gallese V. Resonance behaviors and mirror neurons. Archives of Italian Biology 1999; 137: 85–100. 184. Duncan J, Emslie H, Williams P, Johnson R, Freer C. Intelligence and the frontal lobe: The organization of goal-directed behavior. Cognitive Psychology 1996; 30(3): 257–303. 185. Eysenck H J. Personality and the law of effect. In: Berlyne D E, Madsen K B, eds. Pleasure, Reward, Preference. New York: Academic Press, 1973: 133–166. 186. Cockburn J. Task interruption in prospective memory: A frontal lobe function? Cortex 1995; 31(1): 87–97. 187. West R L. An application of prefrontal cortex function theory to cognitive aging. Psychological Bulletin 1996; 120(2): 272–292. 188. Bar-On R, Tranel D, Denburg N L, Bechara A. Exploring the neurological substrate of emotional and social intelligence. Brain 2003; 126: 1790–1800. 189. Bechara A, Damasio A R, Damasio H, Anderson S W. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 1994; 50(1&3): 7–15. 190. Fuster J. The Prefrontal Cortex. New York: Raven Press, 1989. 191. Constantinidis C, Williams G V, Goldman-Rakic P S. A role for inhibition in shaping the temporal flow of information in prefrontal cortex. Nature Neuroscience 2002; 5(2): 175–180. 192. Meck W H, Benson A M. Dissecting the brain’s internal clock: How frontal-striatal circuitry keeps time and shifts attention. Brain and Cognition 2002; 48(1): 195–211.

April 12, 2006 11:42 WSPC/SPI-B378 Timing the Future ch11

306

J. Glicksohn & M. S. Myslobodsky

193. Goldstein L H, Bernard S, Fenwick P B C, Burgess P W, McNeil J. Unilateral frontal lobectomy can produce strategy application disorder. Journal of Neurology Neurosurgery and Psychiatry 1993; 56(3): 274–276. 194. Burgess P W. Strategy application disorder: The role of the frontal lobes in human multitasking. Psychological Research-Psychologische Forschung 2000; 63(3&4): 279–288. 195. Burgess P W, Veitch E, Costello A D, Shallice T. The cognitive and neuroanatomical correlates of multitasking. Neuropsychologia 2000; 38(6): 848–863. 196. Tanaka S C, Doya K, Okada G, Ueda K, Okamoto Y, Yamawaki S. Prediction of immediate and future rewards differentially recruits cortico-basal ganglia loops. Nature Neuroscience 2004; 7(8): 887–893. 197. Fellows L K, Farah M J. Dissociable elements of human foresight: A role for the ventromedial frontal lobes in framing the future, but not in discounting future rewards. Neuropsychologia 2005; 43(8): 1214–1221. 198. Suddendorf T, Corballis M C. Mental time travel and the evolution of the human mind. Genetic, Social and General Psychology Monographs 1997; 123(2): 133–167. 199. Parks T E, Strohman L K. False memories of having said the unsaid: On the importance of a prior intention to speak. American Journal of Psychology 2005; 118(1): 115–121. 200. Sakai K, Rowe J B, Passingham R E. Active maintenance in prefrontal area 46 creates distractor-resistant memory. Nature Neuroscience 2002; 5(5): 479–484. 201. Haack S. Evidence and Inquiry: Towards Reconstruction in Epistemology. Oxford: Blackwell, 1993. 202. Milgram N A. El retraso: Una enfermedad de los tempos modernos [Procrastination: A malady of modern time]. Boletin de Psicologia 1992; 35: 83–102. 203. Fouks L, Guibert S, Montot M. The notion of lived time in P. Janet. Annales MedicoPsychologiques 1988; 146(10): 941–952.

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future index

Index

abstract, 171 accumulator, 31, 219, 220 accuracy, 2 acquired neurological conditions, 239 act of will, 91, 106 adult lifespan, 1 age, 3, 28 aiding strategy, 143 Alzheimer’s disease, 243 amygdala, 277 AND rule, 42 anterior cingulate, 215, 221 anticipate, 67 anxiety, 13 apparent simultaneity, 102 asymmetric interference effect, 34 asynchrony-threshold, 94, 102 attention, 27, 134, 136 analytic, 4, 65 attentional control, 219 automatic, 68 distraction, 34, 35 divided attention, 52, 246 energy, 54, 62 gate, 26, 31 modes, 67 pacing, 54 resource allocation, 28 resources, 3, 29, 71 synchrony, 63 attention deficit hyperactivity disorder (ADHD), 214

attunement, 4 audio-visual finish-line, 102 audio-visual integration, 102 autonomic responses, 277 aversive consequences, 126, 127 background activities, 270, 284 back-referral, 99 backward causal process, 282 basal ganglia, 215 behavioral choice theory, 136 behavioral meaning, 93 bilocal motion sensor, 96 binding problem, 90 binocular rivalry, 106 brain program, 90 brain-injury, 243 brain-level, 92 buildup process, 100 C-gate, 108 Cartesian theatre, 101 categorization, 3 caudate nucleus, 217 causal coupling, 104 central clock, 219 cerebellum, 215, 217, 241 cerebral rhythms, 271 chaining delay units, 97 change in cognitive context, 29 Cinderella Problem, 268 cingulate, 221 307

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future index

308 clock checking, 10 clock-counter, 71 clock-pacemaker, 79 clock-parameter, 106 cocktail party phenomenon, 51 coding of time, 87 coefficient of variation, 3 cognitive control, 191 cognitive failures questionnaire (CFQ), 244 cognitive orienting reflexes, 274 cognitive psychology, 6 coincidence gate, 108 color (area V4), 90 color consciousness, 92 compensatory strategy, 143, 155 concrete, 171 conscious perception site, 103 consciousness, 26, 270 consciousness as reportability, 92 construal, 171 context, 37 contextual, 3, 10 cortex anterior cingulate, 278 interior prefrontal, 215, 220 orbitomedial prefrontal cortex, 277 parietal cortex, 286 prefrontal cortex, 277, 287 supplementary motor area, 278, 279 ventromedial frontal, 224 counter-based models, 107 counting, 221 cross-temporal contingencies, 289 decision, 40 deferred gratification, 288 delay-lines, 108 delayed response, 219 deliberation process, 106 delta time, 272 demands, 27 developmental differences, 196 diencephalic, 245 dishabituation, 272 disinhibition, 245 divided, 51 dopaminergic, 228

Index dorsal stream, 270 dorso-lateral prefrontal cortex (DLPFC), 215, 216, 219, 240 duration, 6 duration coding, 106 duration estimate, 149 duration judgment, 29, 196 duration metric, 87 dyadic relative time, 59 dynamic attending model, 4 dynamic attending theory (DAT), 54 dynamic filter, 287 dysexecutive, 244, 245 ecological, 265 empty, 70 encoding, 37, 118, 131, 133 entrainment, 58 environment, 4, 157 environmental dependency syndrome, 287 episodic, 20 error monitoring, 278 error related negativity, 278 error-detection system, 277 event markers, 64 event related potentials (ERP), 220 event structure, 55, 65 event-based, 9, 26, 207, 242, 266 event-cued, 9 event-plus-action, 37 executive control, 31 executive function, 30, 201, 252, 290 expectancy, 65, 264 profile, 78 violation, 63 expected duration estimate, 149 experiences, 87 external aid, 149 external rhythm, 4 extrapolation range, 98 extrapolation shift, 98 feedback, 216 feeling of deliberation, 106 filled duration, 70 filled intervals, 71 filled-interval illusion, 276

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future index

Index finish line, 101 fixed-interval, 134 flash perception, 95 flash-lag effect (FLE), 87, 96 foreground, 270, 284 forgetting, 39, 122, 131, 132 Fröhlich effect, 98 free will, 91 frontal lobe function, 201 frontal lobes, 216 fronto-cerebellar, 217 fronto-striatal, 216 functional imaging, 218, 246 functional magnetic resonance imaging (fMRI), 218 future-oriented attending, 4, 67 gambling, 214 Gibson’s ecological model, 285 global, 171 globus pallidum, 217 goal-directed behavior, 150 goals, 145 habituation, 272 hard-wired time, 268 Hazelhoff–Wiersma effect, 98 head injury, 254 hippocampus, 277 homunculus, 89 hypnosis, 44 implementation intention, 150 impulsive choice, 280 impulsivity, 120–122, 127, 128, 214 incentive motivation, 264 inferior frontal cortex (IFC), 218, 220 inferior parietal lobe, 218, 278 information processing, 5, 242 inhibition, 202 inside/outside accounts, 150 insula, 8, 225, 277 intention, 26, 264 intentionality, 264 intentions, 8 interactionism, 148 interference, 118, 135, 136

309 interhemispheric, 240 internal clock, 219, 229 internal-clock model, 2 intertemporal bridging, 229 intertemporal-choice behaviors, 274 interval interval codes, 80 interval length, 27, 28 interval reduction, 192 involuntary semantic memories, 278 isochronous, 61, 82 labelled line code, 107, 108 latency, 95, 102, 104 Libet’s problem, 99 limit cycle, 62 local sign map, 89 low level construals, 273 lucid dreaming, 44 master clock, 271 maximum W-time, 95 meaning, 28 memory, 25, 264 episodic memory, 291 everyday memory questionnaire (EMQ), 244 long-term memory, 32 reference memory, 32 retrospective memory, 131, 132, 270 sensory memory, 271 working memory, 27, 215 mental (experiential) timing, 94 mental travel, 291 mereological fallacy, 92 metric for time-intervals, 88 monitoring, 20, 134, 136, 144, 191, 240, 246 accuracy, 195 frequency, 195 monochronicity, 158 motion (area V5), 90 motion reversal, 98 motivation accounts, 150 multimodal neurons, 286 multiple timer hypothesis, 108 multitasking procedure, 290 myopia for the future, 214

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future index

310 NCC for color, 92 need odds, 131 negative-asynchrony task, 35 neural correlates, 216 neural oscillations, 58 neurological, 240 neurology, 259 neuronal correlate of consciousness (NCC), 92 neuropsychology, 259 noncognitive factors, 264 nonconscious decision, 91 nonconscious movements, 105 observational learning, 289 oculomotor capture, 278 onset-offset synchrony, 104 operant conditioning, 192 optimal foraging, 193 OR rule, 42 orbito-frontal cortex (OFC), 240 orienting response, 271 oscillators, 58 overlap in time, 101 pacemaker, 30 pacemaker-accumulator, 4 pacemaker-counter, 4 parietal, 215 Parkinson’s disease (PD), 215, 240, 279 perceived motion, 94 percept flips, 92 perceptual extrapolation, 96 latency, 101 position label, 89 space-time, 93 period adaptation, 82 periodic monitoring, 192 periodic structure, 59 peripheral, 171 person-level, 90, 92 physics and perception, 93 pitch monitoring, 67 pitfalls of timing, 95 place task, 268

Index plans, 8, 144, 240 P-plans, 106 planning fallacy, 150 “willed” plans, 91 point of subjective simultaneity (PSS), 102 polychronicity, 155, 158 population response, 103 positron emission tomography (PET), 218 predisposition, 143 prioritizing, 144 process, 282 process control, 192 processing, 71 procrastination, 156, 292 production, 3, 214 propagation speed, 101 prospective, 2, 69, 264 duration judgment, 28, 29 failure, 149 ProM-E, 242 ProM-T, 239 overload, 147 prospective memory (ProM), 8, 88, 91, 143, 240, 242 remembering, 25–27 timing, 33 psychological distance, 171 Pulfrich effect, 94 pulse, 4, 62 putamen, 217 random generators, 92 random walk model, 280 rate-codes, 107 re-storage process, 109 reaction times, 95 readiness potential, 91, 105 reality monitoring, 40 recruitment, 97 recruitment coupling, 97 recursive-reminding, 41 model, 26 reinforcers, 119, 128, 129, 134 relativity theory, 93 remembered visual position, 89 reportability, 92 representation, 171

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future index

Index reproduction, 3, 8, 214 residual activation, 277 resource allocation, 30 retention, 36, 37 interval, 27 retrieval, 8, 36, 37 contexts, 42 retrospective, 2, 25, 69, 242 reward, 157, 264 rhythm, 55, 214, 220 perception, 220 production, 214 rhythm generators, 107 rhythmic markers, 58 rhythmicity, 4 right hemisphere, 240 scalar expectancy theory (SET), 5 scalar-timing model, 30 schizophrenia, 215 secondary-task, 27, 28 selective attention, 51, 52 sense of time, 203 sensorimotor synchronization, 216 shifting, 202 simultaneity detection, 87, 93 simultaneity judgments, 88 single-neuron recordings, 94 sleep, 43 social skill, 287 soft-wired, 268 somatotopic maps, 89 space-codes, 89 span, 98 spontaneous retrieval theory, 207 stimulus onset asynchrony, 100 storage elements, 109 store durations, 108 strategic, 11 strategic monitoring, 192 strategies, 28 strategy application disorder, 290 stroke, 253 structure, 172 subcortical, 215, 240 subjective time, 3, 268 subjective value, 119–128

311 substance abuse, 122 superiority effect, 266 superodinate, 172 supplementary motor area (SMA), 215 sustained attention, 219 switch, 5, 31 synchronization, 214 synchronize, 4, 54, 59 task importance, 131–135 task structure, 53 telescoping, 293 temporal binding, 89 bridging, 214 coherence, 4 construal, 275 discounting (TD), 119, 132, 214 discrimination, 220, 222 disorientation, 241 distance, 171 dispersion, 90 expectancy, 72 foresight, 214, 224 information processing, 29 myopia, 214 order judgments, 88 organization, 213 performance, 2 predictability, 4 relevance, 29 resolution, 93 uncertainty, 29 tension reduction, 266 Test-Operate-Test-Exit (TOTE), 193, 266 Test-Wait-Test-Exit (TWTE), 27, 248, 266 thalamo-cortical, 240 thalamus, 216 time accuracy, 276 discounting, 156 discrimination, 214 estimation, 69, 214, 219 experiential timing, 90, 95 internal timing, 58 management, 143, 221 management behavior scale (TMBS), 145

April 7, 2006 13:28 WSPC/SPI-B378 Timing the Future index

312 management error, 149 management questionnaire (TMQ), 145 monitoring, 191 motor timing, 213 neural activity timing, 95 neural timing, 90 orientation, 155 perception, 54 personal time, 274 personality, 143 personality indicator (TPI), 155 physical time, 2, 90, 94 psychological time, 274, 283 relative time, 53, 54, 65 reproduction, 196 scales, 56 span, 55 structure questionnaire (TSQ), 145, 152 time discounting, 274 urgency, 155 time as path-length, 108 time estimation, 26 time marker, 100 time perception, 1 time sampling, 267 time task, 268 time-based, 9, 26 time-based intentions, 27 time-based ProM, 1, 191, 239, 266 time-codes, 89 time-related characteristics, 148

Index time-to-contact, 90 timer, 107 timing, 2, 4, 215, 240 conscious percepts, 99 mechanism, 106 of volition, 90 transcranial magnetic stimulation, 217 traumatic brain injury (TBI), 246 tryptophan, 228 unpacking, 150 updating, 202 variability, 2 ventral stream, 270 ventriloquist effect, 103 ventromedial frontal, 224 verbal estimates, 3 veto decision, 91 voluntary, 59, 91 voluntary actions, 105 W-time, 95 Wahrnehmungszeit, 95 waiting, 27 wanting, 289 Weber fraction, 3 “willed” decision, 90 window of simultaneity, 102 Zeigarnik effect, 265