Cambridge Fundamentals of Neuroscience in Psychology Cambridge Fundamentals oF neurosCienCe in PsyChology

Slot n ic k
“Quote.”
Name, Affiliation

“Quote.”
Name, Affiliation Cognitive
Neuroscience

Cognitive Neuroscience of Memory

of Memory
Within the last two decades, the field of cognitive neuroscience has begun to thrive
with technological advances that non-invasively measure human brain activity.
This is the first book to provide a comprehensive and up-to-date treatment on the
cognitive neuroscience of memory. Topics include cognitive neuroscience techniques
and human brain mechanisms underlying long-term memory success, long-term
memory failure, working memory, implicit memory, and memory and disease.
Cognitive Neuroscience of Memory highlights both spatial and temporal aspects of the
functioning human brain during memory. Each chapter is written in an accessible
Scot t D. Slot n ic k
style and includes background information and many figures. In his analysis,
Scott Slotnick questions popular views, rather than simply assuming they are
correct. In this way, science is depicted as open to question, evolving, and exciting.
9781107446267: Slotnick: Cover: C M Y K

ISBN 978-1-107-44626-7

Cover design: Andrew Ward

9 781 1 07 446267 >
Cognitive Neuroscience of Memory

Within the last two decades, the field of cognitive neuroscience has begun to
thrive with technological advances that non-invasively measure human brain
activity. This is the first book to provide a comprehensive and up-to-date
treatment on the cognitive neuroscience of memory. Topics include cognitive
neuroscience techniques and human brain mechanisms underlying long-term
memory success, long-term memory failure, working memory, implicit mem-
ory, and memory and disease. Cognitive Neuroscience of Memory highlights
both spatial and temporal aspects of the functioning human brain during
memory. Each chapter is written in an accessible style and includes background
information and many figures. In his analysis, Scott Slotnick questions popular
views, rather than simply assuming they are correct. In this way, science is
depicted as open to question, evolving, and exciting.

Scott D. Slotnick is an Associate Professor of Psychology at Boston College,

Editor-in-Chief of the journal Cognitive Neuroscience, and author of the
book Controversies in Cognitive Neuroscience. He employs multiple cogni-
tive neuroscience techniques to investigate the brain mechanisms underlying
memory including functional magnetic resonance imaging (fMRI), electro-
encephalography (EEG), and transcranial magnetic stimulation (TMS).
Cambridge Fundamentals of Neuroscience in Psychology
Developed in response to a growing need to make neuroscience acces-
sible to students and other non-specialist readers, the Cambridge
Fundamentals of Neuroscience in Psychology series provides brief intro-
ductions to key areas of neuroscience research across major domains of
psychology. Written by experts in cognitive, social, affective, develop-
mental, clinical, and applied neuroscience, these books will serve as ideal
primers for students and other readers seeking an entry point to the
challenging world of neuroscience.

Forthcoming Titles in the Series

The Neuroscience of Intelligence, by Richard J. Haier
The Neuroscience of Expertise, by Merim Bilalić
The Neuroscience of Adolescence, by Adriana Galván
The Neuroscience of Aging, by Angela Gutchess
The Neuroscience of Addiction, by Francesca Filbey
Cognitive Neuroscience
of Memory

Scott D. Slotnick
Boston College
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
4843/24, 2nd Floor, Ansari Road, Daryaganj, Delhi – 110002, India
79 Anson Road, #06–04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781107084353
10.1017/9781316026687
© Scott D. Slotnick 2017
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2017
Printed in the United States of America by Sheridan Books, Inc.
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging in Publication Data
Names: Slotnick, Scott D.
Title: Cognitive neuroscience of memory / Scott D. Slotnick, Boston College.
Description: Cambridge : Cambridge University Press, 2016. | Includes index.
Identifiers: LCCN 2016049342 | ISBN 9781107084353
Subjects: LCSH: Memory. | Memory – Physiological aspects. | Cognitive neuroscience.
Classification: LCC QP406 .S5945 2016 | DDC 612.8/23312–dc23
LC record available at https://lccn.loc.gov/2016049342
ISBN 978-1-107-08435-3 Hardback
ISBN 978-1-107-44626-7 Paperback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
This book is dedicated to my incredible daughter Sonya, for
dominating my hippocampal sharp-wave ripples these past
twelve years
As regards the question . . . what memory or remembering is . . . it is the state of
a presentation, related as a likeness to that of which it is a presentation; and as to the
question of which of the faculties within us memory is a function . . . it is a function of the
primary faculty of sense-perception, i.e. of that faculty whereby we perceive time.
(Aristotle, [350 BCE] 1941, p. 611)
 Contents

List of Figures page x

Preface xxi

1 Types of Memory and Brain Regions of Interest 1

1.1 Cognitive Neuroscience 2
1.2 Memory Types 3
1.3 Brain Anatomy 8
1.4 The Hippocampus and Long-Term Memory 12
1.5 Sensory Regions 13
1.6 Control Regions 18
1.7 The Organization of This Book 21

2 The Tools of Cognitive Neuroscience 24

2.1 Behavioral Measures 25
2.2 High Spatial Resolution Techniques 25
2.3 High Temporal Resolution Techniques 30
2.4 High Spatial and Temporal Resolution Techniques 34
2.5 Lesions and Temporary Cortical Disruption
Techniques 37
2.6 Method Comparisons 43

3 Brain Regions Associated with Long-Term Memory 46

3.1 Episodic Memory 47
3.2 Semantic Memory 51
3.3 Memory Consolidation 53
3.4 Consolidation and Sleep 56
3.5 Memory Encoding 59
3.6 Sex Differences 61
3.7 Superior Memory 64

4 Brain Timing Associated with Long-Term Memory 71

4.1 Timing of Activity 72
4.2 The FN400 Debate 76
4.3 Phase and Frequency of Activity 79
viii Contents

5 Long-Term Memory Failure 88

5.1 Typical Forgetting 89
5.2 Retrieval-Induced Forgetting 92
5.3 Motivated Forgetting 96
5.4 False Memories 97
5.5 Flashbulb Memories 103

6 Working Memory 108

6.1 The Contents of Working Memory 109
6.2 Working Memory and the Hippocampus 114
6.3 Working Memory and Brain Frequencies 119
6.4 Brain Plasticity and Working Memory Training 122

7 Implicit Memory 129

7.1 Brain Regions Associated with Implicit Memory 130
7.2 Brain Timing Associated with Implicit Memory 135
7.3 Models of Implicit Memory 138
7.4 Implicit Memory and the Hippocampus 141
7.5 Skill Learning 146

8 Memory and Other Cognitive Processes 150

8.1 Attention and Memory 151
8.2 Imagery and Memory 159
8.3 Language and Memory 164
8.4 Emotion and Memory 166

9 Explicit Memory and Disease 171

9.1 Amnestic Mild Cognitive Impairment 172
9.2 Alzheimer’s Disease 177
9.3 Mild Traumatic Brain Injury 179
9.4 Medial Temporal Lobe Epilepsy 186
9.5 Transient Global Amnesia 190

10 Long-Term Memory in Animals 196

10.1 The Medial Temporal Lobe 197
10.2 Long-Term Potentiation 200
10.3 Memory Replay 203
10.4 Time Cells 205
10.5 Episodic Memory 210
Contents ix

11 The Future of Memory Research 219

11.1 Phrenology and fMRI 220
11.2 fMRI versus ERPs 225
11.3 Brain Region Interactions 227
11.4 The Future of Cognitive Neuroscience 232
11.5 A Spotlight on the Fourth Dimension 234

Glossary 238
References 248
Author Index 270
Subject Index 276

Color plates are to be found between pp. 170 and 171

Figures

1.1 The relationships between the fields of cognitive psychology,

cognitive neuroscience, and behavioral neuroscience. page 2
1.2 Organization of memory types. 3
1.3 Probability of “remember” or “know” responses as a function
of confidence judgments. 8
1.4 Brain regions associated with memory. 9
1.5 Gyri and sulci in brain regions of interest. 10
1.6 Brodmann map. 11
1.7 Depiction of medial temporal lobe resection in patient H. M.
Reproduced from Journal of Neurology, Neurosurgery, &
Psychiatry, Loss of recent memory after bilateral hippocampal
lesions, William Beecher Scoville and Brenda Milner, Volume
20, Pages 11–21, Copyright (1957), with permission from BMJ
Publishing Group Ltd. 13
1.8 Sensory brain regions of interest. 14
1.9 Sensory fMRI activity associated with perception and memory. 17
1.10 Item memory and source memory paradigm and fMRI results.
Reprinted from Cognitive Brain Research, Volume 17, Scott
D. Slotnick, Lauren R. Moo, Jessica B. Segal, and John Hart,
Jr., Distinct prefrontal cortex activity associated with item
memory and source memory for visual shapes, Pages 75–82,
Copyright (2003), with permission from Elsevier. 19
2.1 MRI scanner and fMRI results. (A) Photo courtesy of Preston
Thakral. (B, C) Reprinted from Proceedings of the National
Academy of Sciences of the United States of America,
Volume 93, Randy L. Buckner, Peter A. Bandettini,
Kathleen M. O’Craven, Robert L. Savoy, Steven E. Petersen,
Marcus E. Raichle, and Bruce R. Rosen, Detection of cortical
activation during averaged single trials of a cognitive task using
functional magnetic resonance imaging, Pages 14878–14883,
Copyright (1996) National Academy of Sciences, USA. 27
2.2 ERP setup and results. (A) Photo courtesy of Scott Slotnick.
Reprinted from NeuroImage, Volume 39, Jeffrey D. Johnson,
Brian R. Minton, and Michael D. Rugg, Content dependence
List of Figures xi

of the electrophysiological correlates of recollection, Pages

406–416, Copyright (2008), with permission from Elsevier. 31
2.3 MEG setup. Photo courtesy of CTFMEG/MEG International
Services Ltd., Canada. 33
2.4 Hippocampal depth electrode placement and results.
Reprinted from Proceedings of the National Academy of
Sciences of the United States of America, Volume 112, Nanthia
A. Suthana, Neelroop N. Parikshak, Arne D. Ekstrom, Matias
J. Ison, Barbara J. Knowlton, Susan Y. Bookheimer, and
Itzhak Fried, Specific responses of human hippocampal
neurons are associated with better memory, Pages
10503–10508, Copyright (2015) National Academy of Sciences,
USA. 36
2.5 Hippocampal lesion and recognition memory results.
Reprinted from Neuron, Volume 37, Joseph R. Manns,
Ramona O. Hopkins, Jonathan M. Reed, Erin G. Kitchener,
and Larry R. Squire, Recognition memory and the human
hippocampus, Pages 171–180, Copyright (2003), with
permission from Elsevier. 38
2.6 TMS setup and fMRI guided TMS results. (A, B) Photos
courtesy of Scott Slotnick. Reprinted from NeuroImage,
Volume 55, Scott D. Slotnick and Preston P. Thakral, Memory
for motion and spatial location is mediated by contralateral and
ipsiliateral motion processing cortex, Pages 794–800, Copyright
(2011), with permission from Elsevier. 40
2.7 tDCS setup. Photo courtesy of Bryan Coppede. 42
2.8 Spatial resolution and temporal resolution for different
methods. 43
3.1 Regions of the brain associated with episodic memory.
Reprinted from Current Opinion in Neurobiology,
Volume 23(2), Michael D. Rugg and Kaia L. Vilberg, Brain
networks underlying episodic memory retrieval, Pages
255–260, Copyright (2013), with permission from Elsevier. 48
3.2 Model of medial temporal lobe sub-region function. Reprinted
from NeuroReport, Volume 24(12), Scott D. Slotnick,
The nature of recollection in behavior and the brain, Pages
663–670, Copyright (2013), with permission from Wolters
Kluwer. 49
3.3 Regions of the brain associated with semantic memory. Reprinted
from Neuroimage, Volume 63(1), Kimiko Domoto-Reilly, Daisy
Sapolsky, Michael Brickhouse, and Bradford C. Dickerson,
xii List of Figures

Naming Impairment in Alzheimer’s disease is associated

with left anterior temporal lobe atrophy, Pages 348–355,
Copyright (2012), with permission from Elsevier. 52
3.4 Autobiographical memory disruption for recent and remote
events in patients with hippocampal lesions. Reprinted from
Proceedings of the National Academy of Sciences of the United
States of America, Volume 108, Thorsten Bartsch, Juliane
Döhring, Axel Rohr, Olav Jansen, and Günther Deuschl, CA1
neurons in the human hippocampus are critical for
autobiographical memory, mental time travel, and autonoetic
consciousness, Pages 17562–17567, Copyright (2011) National
Academy of Sciences, USA. 55
3.5 Sleep stages and brain oscillations associated with slow wave
sleep and long-term memory consolidation. (A) Reprinted
from Trends in Neurosciences, Volume 28(8), Robert Stickgold
and Matthew P. Walker, Memory consolidation and
reconsolidation: What is the role of sleep, Pages 408–415,
Copyright (2005), with permission from Elsevier. (B) Reprinted
from Psychological Research, Volume 76(2), Jan Born and
Ines Wilhelm, System consolidation of memory during sleep,
Pages 192–203, Copyright (2012), with permission from
Springer. 57
3.6 Regions of the brain associated with subsequent memory
effects. Reprinted from NeuroImage, Volume 54(3),
Hongkeun Kim, Neural activity that predicts subsequent
memory and forgetting: A meta-analysis of 74 fMRI studies,
Pages 2446–2461, Copyright (2011), with permission from
Elsevier. 60
3.7 Object–location virtual environment and hippocampal
laterality results. Reprinted from NeuroReport, Volume 17(4),
Lars Frings, Kathrin Wagner, Josef Unterrainer, Joachim
Spreer, Ulrike Halsband, and Andreas Schulze-Bonhage,
Gender-related differences in lateralization of hippocampal
activation and cognitive strategy, Pages 417–421, Copyright
(2006), with permission from Wolters Kluwer. 63
3.8 Change in the size of the posterior hippocampus as a function
of time as a London taxi driver. Reprinted from Proceedings of
the National Academy of Sciences of the United States of
America, Volume 97, Eleanor A. Maguire, David G. Gadian,
Ingrid S. Johnsrude, Catriona D. Good, John Ashburner,
Richard S. J. Frackowiak, and Christopher D. Frith,
List of Figures xiii

Navigation-related structural change in the hippocampi of taxi

drivers, Pages 4398–4403, Copyright (2000) National Academy
of Sciences, USA. 65
4.1 ERP activity associated with recollection and familiarity.
Reprinted from Brain Research, Volume 1122(1), Kaia
L. Vilberg, Rana F. Moosavi, and Michael D. Rugg,
The relationship between electrophysiological correlates of
recollection and amount of information retrieved, Pages
161–170, Copyright (2006), with permission from Elsevier. 73
4.2 ERP activity associated with conceptual repetition priming.
Reprinted from NeuroImage, Volume 49(3), Joel L. Voss,
Haline E. Schendan, and Ken A. Paller, Finding meaning in
novel geometric shapes influences electrophysiological
correlates of repetition and dissociates perceptual and
conceptual priming, Pages 2879–2889, Copyright (2010), with
permission from Elsevier. 77
4.3 Topographic maps illustrating the conceptual priming effect
and the mid-frontal old–new effect. Reprinted from
NeuroImage, Volume 63(3), Emma K. Bridger, Regine Bader,
Olga Kriukova, Kerstin Unger, and Axel Mecklinger,
The FN400 is functionally distinct from the N400, Pages
1334–1342, Copyright (2012), with permission from Elsevier. 78
4.4 Topographic maps and activation timecourses illustrating
spatial memory effects. Reprinted from Brain Research,
Volume 1330, Scott D. Slotnick, Synchronous retinotopic
frontal-temporal activity during long-term memory for spatial
location, Pages 89–100, Copyright (2010), with permission from
Elsevier. 81
4.5 EEG frequency band activity associated with subsequently
remembered and forgotten items. Reprinted from NeuroImage,
Volume 66, Uwe Friese, Moritz Köster, Uwe Hassler, Ulla
Martens, Nelson Trujillo-Barreto, and Thomas Gruber,
Successful memory encoding is associated with increased cross-
frequency coupling between frontal theta and posterior gamma
oscillations in human scalp-recorded EEG, Pages 642–647,
Copyright (2013), with permission from Elsevier. 83
5.1 Subsequent forgetting fMRI activity and default network fMRI
activity. (A) Reprinted from NeuroImage, Volume 54(3),
Hongkeun Kim, Neural activity that predicts subsequent
memory and forgetting: A meta-analysis of 74 fMRI studies,
Pages 2446–2461, Copyright (2011), with permission from
xiv List of Figures

Elsevier. (B) Reprinted from Annals of the New York Academy

of Sciences, Volume 1124, Randy L. Buckner, Jessica
R. Andrews-Hanna, and Daniel L. Schacter, The Brain’s
Default Network, Pages 1–38, Copyright (2008), with
permission from John Wiley and Sons. 91
5.2 Retrieval-inducted forgetting paradigm, behavioral
performance, and fMRI activity. Reprinted from Wimber et al.,
The Journal of Neuroscience: The official journal of the Society
for Neuroscience, Copyright (2008), Reproduced with
permission of the Society for Neuroscience. 93
5.3 Retrieval-induced forgetting EEG activity. Reprinted from
Staudigl et al., The Journal of Neuroscience: The official journal
of the Society for Neuroscience, Copyright (2010), Reproduced
with permission of the Society for Neuroscience. 95
5.4 Regions of the brain commonly and differentially associated
with true memory and related false memory. Reprinted from
Nature Neuroscience, Volume 7(6), Scott D. Slotnick and
Daniel L. Schacter, A sensory signature that distinguishes true
from false memories, Pages 664–672, Copyright (2004). 99
5.5 Brain activity associated with unrelated false memory. Rachel
J. Garoff-Eaton, Scott D. Slotnick, and Daniel L. Schacter, Not
all false memories are created equal: The neural basis of false
recognition, Cerebral Cortex, 2006, 16(11), 1645–1652, by
permission of Oxford University Press. 102
6.1 Object or location working memory paradigm and fMRI
results. Reprinted from Neuropsychologia, Volume 41(3),
Joseph B. Sala, Pia Rämä, and Susan M. Courtney, Functional
topography of a distributed neural system for spatial and
nonspatial information maintenance in working memory,
Pages 341–356, Copyright (2003), with permission from
Elsevier. 110
6.2 Sustained working memory fMRI activity in the dorsolateral
prefrontal cortex. Reprinted from Trends in Cognitive Sciences,
Volume 7(9), Clayton E. Curtis and Mark D’Esposito,
Persistent activity in the prefrontal cortex during working
memory, Pages 415–423, Copyright (2003), with permission
from Elsevier. 111
6.3 Color and/or location working memory paradigms and medial
temporal lobe lesion results. Reprinted from
Neuropsychologia, Volume 46(2), Carsten Finke, Mischa
Braun, Florian Ostendorf, Thomas-Nicolas Lehmann, Karl-
List of Figures xv

Titus Hoffiman, Ute Kopp, and Christoph J. Ploner,

The human hippocampal formation mediates short-term
memory of colour-location associations, Pages 614–623,
Copyright (2008), with permission from Elsevier. 118
6.4 Color working memory paradigm and EEG results. Reprinted
from Current Biology, Volume 19(21), Paul Sauseng, Wolfgang
Klimesch, Kirstin F. Heise, Walter R. Gruber, Elisa Holz,
Ahmed A. Karim, Mark Glennon, Christian Gerloff, Niels
Birbaumer, and Friedhelm C. Hummel, Brain oscillatory
substrates of visual short-term memory capacity, Pages
1846–1852, Copyright (2009), with permission from Elsevier. 120
6.5 Behavioral effects and brain effects of working memory
training. Reprinted from NeuroImage, Volume 52(2), Dietsje
D. Jolles, Meike J. Grol, Mark A. Van Buchem, Serge
A. R. B. Rombouts, and Eveline A. Crone, Practice effects in
the brain: Changes in cerebral activation after working
memory practice depends on task demands, Pages 658–668,
Copyright (2010), with permission from Elsevier. 124
7.1 Repetition priming paradigm and fMRI results. Reprinted
from Neuropsychologia, Volume 39(2), Wilma Koutstaal,
Anthony D. Wagner, Michael Rotte, Anat Maril, Randy
L. Buckner, and Daniel L. Schacter, Perceptual specificity in
visual object priming: Functional magnetic resonance imaging
evidence for a laterality difference in fusiform cortex, Pages
184–199, Copyright (2001), with permission from Elsevier. 132
7.2 Review of cortical repetition priming effects. Reprinted from
Current Opinion in Neurobiology, Volume 17(2), Daniel
L. Schacter, Gagan S. Wig, and W. Dale Stevens, Reductions in
cortical activity during priming, Pages 171–176, Copyright
(2007), with permission from Elsevier. 134
7.3 Repetition priming EEG and MEG results. (A) Reprinted
from Fiebach et al., The Journal of Neuroscience: The official
journal of the Society for Neuroscience, Copyright (2005),
Reproduced with permission of the Society for Neuroscience.
(B) Reprinted from Frontiers in Human Neuroscience, 2010,
Volume 4, Article 30, Jessica R. Gilbert, Stephen J. Gotts,
Frederick W. Carver, and Alex Martin, Object repetition leads
to local increases in the temporal coordination of neural
responses. 137
7.4 Models of repetition priming. Reprinted from Trends in
Cognitive Sciences, Volume 10(1), Kalanit Grill-Spector,
xvi List of Figures

Richard Henson, and Alex Martin, Repetition and the brain:

Neural models of stimulus-specific effects, Pages 14–23,
Copyright (2006), with permission from Elsevier. 139
7.5 Contextual cuing stimulus display. 143
7.6 Skill learning behavioral results and fMRI results. Reprinted
from Brain Research, Volume 1318, Liangsuo Ma, Binquan
Wang, Shalini Narayana, Eliot Hazeltine, Xiying Chen, Donald
A. Robin, Peter T. Fox, Jinhu Xiong, Changes in regional
activity are accompanied with changes in inter-regional
connectivity during 4 weeks motor learning, Pages 64–76,
Copyright (2010), with permission from Elsevier. 146
8.1 Spatial attention paradigm and fMRI results. (B) Reprinted
from Neuropsychologia, Volume 39(12), Joseph B. Hopfinger,
Marty G. Woldorff, Evan M. Fletcher, and George R. Mangun,
Dissociating top-down attentional control from selective
perception and action, Pages 1277–1291, Copyright (2001),
with permission from Elsevier. (C) Reprinted from Brain
Research, Volume 1302, Preston P. Thakral and Scott
D. Slotnick, The role of parietal cortex during sustained visual
spatial attention, Pages 157–166, Copyright (2009), with
permission from Elsevier. 153
8.2 Spatial memory fMRI and ERP results. Reprinted from Brain
Research, Volume 1268, Scott D. Slotnick, Rapid retinotopic
reactivation during spatial memory, Pages 97–111, Copyright
(2009), with permission from Elsevier. 156
8.3 Meta-analysis of control region activity associated with
attention, working memory, and episodic memory retrieval.
Reprinted from Consciousness and Cognition, Volume 14(2),
Hamid R. Naghavi and Lars Nyberg, Common fronto-parietal
activity in attention, memory, and consciousness: Shared
demands on integration? Pages 390–425, Copyright (2005),
with permission from Elsevier. 158
8.4 Visual perception, imagery, and attention paradigms and fMRI
results. Scott D. Slotnick, William L. Thompson, and
Stephen M. Kosslyn, Visual mental imagery induces
retinotopically organized activation of early visual areas,
Cerebral Cortex, 2005, 15(10), 1570–1583, by permission of
Oxford University Press. 160
8.5 Language processing regions. Reprinted from Journal of
Anatomy, Volume 197, Cathy J. Price, The anatomy of
language: Contributions from functional neuroimaging, Pages
List of Figures xvii

335–359, Copyright (2000), with permission from

John Wiley & Sons, Inc. 164
8.6 The amygdala and the hippocampus. Reprinted from Current
Opinion in Neurobiology, Volume 14(2), Elizabeth A. Phelps,
Human emotion and memory: Interactions of the amygdala
and hippocampal complex, Pages 198–202, Copyright (2004),
with permission from Elsevier. 167
9.1 Hippocampus and entorhinal cortex segmentation and
volumes of these regions in control participants and amnestic
mild cognitive impairment patients. Reprinted from
Proceedings of the National Academy of Sciences of the United
States of America, Volume 103, Travis R. Stoub, Leyla
deToledo-Morrell, Glenn T. Stebbins, Sue Leurgans, David
A. Bennett, and Raj C. Shah, Hippocampal disconnection
contributes to memory dysfunction in individuals at risk for
Alzheimer’s disease, Pages 10041–10045, Copyright (2006)
National Academy of Sciences, USA. 173
9.2 Pattern separation paradigm, behavioral results, and fMRI
results for control participants and aMCI patients. Reprinted
from NeuroImage, Volume 51(3), Michael A. Yassa,
Shauna M. Stark, Arnold Bakker, Marilyn S. Albert, Michela
Gallagher, and Craig E. L. Stark, High-resolution structural
and functional MRI of hippocampal CA3 and dentate gyrus in
patients with amnestic mild cognitive impairment, Pages
1242–1252, Copyright (2010), with permission from Elsevier. 175
9.3 Relationship between exercise engagement and Alzeimer’s
disease biomarkers in older adults. Reprinted from Annals of
Neurology, Volume 68, Kelvin Y. Liang, Mark A. Mintun,
Anne M. Fagan, Alison M. Goate, Julie M. Bugg,
David M. Holtzman, John C. Morris, and Denise Head,
Exercise and Alzheimer’s disease biomarkers in cognitive
normal older adults, Pages 311–318, Copyright (2010), with
permission from John Wiley & Sons, Inc. 180
9.4 N-back paradigm, behavioral results, and fMRI results for mild
traumatic brain injury patients and control participants.
Reprinted from NeuroImage, Volume 14(5), Thomas
W. McAllister, Molly B. Sparling, Laura A. Flashman, Stephen
J. Guerin, Alexander C. Mamourian, and Andrew J. Saykin,
Differential working memory load effects after mild traumatic
brain injury, Pages 1004–1012, Copyright (2001), with
permission from Elsevier. 182
xviii List of Figures

9.5 Stimuli and behavioral results for control participants and

medial temporal lobe epilepsy patients following removal of
left or right medial temporal lobe regions. Reprinted from
Neuropsychologia, Volume 24(5), Marilyn Jones-Gotman,
Right hippocampal excision impairs learning and recall of a list
of abstract designs, Pages 659–670, Copyright (1986), with
permission from Elsevier. 188
9.6 Brain images of transient global amnesia patients. Reprinted
from the Journal of Clinical Neurology, Volume 4(2),
YoungSoon Yang, SangYun Kim, and Jae Hyoung Kim,
Ischemic evidence of transient global amnesia: Location of
the lesion in the hippocampus, Pages 59–66, Copyright (2008). 192
10.1 Spontaneous object recognition task. Reprinted from
Neuroscience and Biobehavioral Reviews, Volume 32, Boyer
D. Winters, Lisa M. Saksida, and Timothy J. Bussey, Object
recognition memory: Neurobiological mechanisms of
encoding, consolidation and retrieval, Pages 1055–1070,
Copyright (2008), with permission from Elsevier. 198
10.2 Medial temporal lobe organization and phylogenic tree of
mammals. Reprinted from Hippocampus, Volume 16, Joseph
R. Manns and Howard Eichenbaum, Evolution of declarative
memory, Pages 795–808, Copyright (2006), with permission
from John Wiley & Sons, Inc. 200
10.3 Long-term potentiation experimental setup and results.
Reprinted from The Journal of Physiology, Volume 232,
T. V. P. Bliss and T. Lømo, Long-lasting potentiation of
synaptic transmission in the dentate area of the anaesthetized
rabbit following stimulation of the perforant path, Pages
331–356, Copyright (1973), with permission from
John Wiley & Sons, Inc. 202
10.4 Memory replay in the rat. Reprinted from Current Opinion
in Neurobiology, Volume 21, Gabrielle Girardeau and Michaël
Zugaro, Hippocampal ripples and memory consolidation,
Pages 452–459, Copyright (2011), with permission from
Elsevier. 204
10.5 Time cell behavioral apparatus and neural activity. Reprinted
from Neuron, Volume 78, Benjamin J. Kraus, Robert
J. Robinson II, John A. White, Howard Eichenbaum, and
Michael E. Hasselmo, Hippocampal “time cells”: Time versus
path integration, Pages 1090–1101, Copyright (2013), with
permission from Elsevier. 207
List of Figures xix

10.6 Time delay memory task and behavioral results. Reprinted

from Current Biology, Volume 16, Stephanie J. Babb and
Jonathan D. Crystal, Episodic-like memory in the rat, Pages
1317–1321, Copyright (2006), with permission from Elsevier. 211
10.7 Hippocampal anatomy in mammals. (A) Reprinted from
Hippocampus, Volume 16, J. R. Manns and H. Eichenbaum,
Evolution of declarative memory, Pages 795–808, Copyright
(2006), with permission from John Wiley & Sons, Inc. (B) With
kind permission from Springer Science + Business Media:
Brain Structure and Function, Organization and chemical
neuroanatomy of the African elephant (Loxodonta africana)
hippocampus, 219(5), 2014, 1587–1601, Nina Patzke,
Olatunbosun Olaleye, Mark Haagensen, Patrick R. Hof,
Amadi O. Ihunwo, and Paul R. Manger, Figure 2. 214
11.1 Past phrenology map and present brain map. Reprinted from
Proceedings of the National Academy of Sciences of the
United States of America, Volume 107, Nancy Kanwisher,
Functional specificity in the human brain: A window into the
functional architecture of the mind, Pages 11163–11170,
Copyright (2010) National Academy of Sciences, USA. 221
11.2 Face processing and shape processing fMRI activity.
Reprinted from NeuroImage, Volume 83, Scott D. Slotnick
and Rachel C. White, The fusiform face area responds
equivalently to faces and abstract shapes in the left and
central visual fields, Pages 408–417, Copyright (2013), with
permission from Elsevier. 223
11.3 Number of fMRI and ERP articles in the highest-impact
cognitive neuroscience journals. 225
11.4 Brain region interaction TMS target sites and fMRI visual
sensory effects during perception. Reprinted from Current
Biology, Volume 16, Christian C. Ruff, Felix
Blankenburg, Otto Bjoertomt, Sven Bestmann, Elliot
Freeman, John-Dylan Haynes, Geraint Rees, Oliver Josephs,
Ralf Deichmann, and John Driver, Concurrent TMS-fMRI
and psychophysics reveal frontal influences on human
retinotopic visual cortex, Pages 1479–1488, Copyright (2006),
with permission from Elsevier. 228
11.5 Brain region interaction TMS target site, visual sensory regions
of interest, and fMRI effects during working memory.
Reprinted from Proceedings of the National Academy of
Sciences of the United States of America, Volume 108, Eva
xx List of Figures

Feredoes, Klaartje Heinen, Nikolaus Weiskopf, Christian Ruff,

and John Driver, Causal evidence for frontal involvement in
memory target maintenance by posterior brain areas during
distractor interference of visual working memory, Pages
17510–17515, Copyright (2011) National Academy of
Sciences, USA. 230
11.6 The relationships between the fields of cognitive psychology,
cognitive neuroscience, and behavioral neuroscience in the
past and in the future. 233
Preface

The human brain and memory are two of the most complex and
fascinating systems in existence. Within the last two decades, the cogni-
tive neuroscience of memory has begun to thrive with the advent of
techniques that can non-invasively measure human brain activity with
high spatial resolution and high temporal resolution.
This is the first book to provide a comprehensive treatment of the
cognitive neuroscience of memory. It is related to three classes of other
books. First, textbooks on cognitive psychology or cognition provide
broad overviews of the cognitive psychology of memory and therefore
only consider a small fraction of the work on the cognitive neuroscience
of memory. Second, textbooks on cognitive neuroscience provide
broad overviews of the entire field and also consider only a small frac-
tion of the work on memory. Third, more specialized books on memory
focus on the cognitive psychology, the behavioral neuroscience, or the
computational modeling of memory rather than the cognitive neu-
roscience of memory.
This book highlights temporal processing in the brain. Cognitive
neuroscientists predominantly use functional magnetic resonance ima-
ging (fMRI) to identify the brain regions associated with a cognitive
process. Although fMRI has excellent spatial resolution, this method
provides little if any information about the time at which brain regions
are active or the way in which different brain regions interact.
By emphasizing both spatial and temporal aspects of brain processing,
this book provides a complete overview of the cognitive neuroscience of
memory and aims to guide the future of memory research.
Each chapter is written in an accessible style and includes background
information and many figures. Debated topics are discussed throughout
the text. The most popular view is routinely questioned rather than
simply assumed to be correct, as is done in the vast majority of textbooks.
In this way, science is depicted as open to question, evolving, and
exciting.
The audience for this book is educated lay people interested in the
cognitive neuroscience of memory and undergraduate students, graduate
students, and scientists who are interested in a comprehensive up-to-date
treatment of this topic. Each chapter includes learning objectives, an
introduction, sections on key topics, a summary, review questions, and
xxii Preface

recommended scientific articles. At a college or university, this book

could serve as a supplemental textbook in lower-level courses (for
instructors who desire a comprehensive treatment on this topic) or as
a main text in an intermediate-level undergraduate course, an advanced-
level undergraduate course, or a graduate seminar (with instructor
lectures, student presentations, and discussions of the recommended
scientific articles).
Many individuals significantly improved the quality of this book. First
and foremost, I thank Matthew Bennet, my editor. Without his vision,
guidance, and support, this book would not exist. I am grateful to Jessica
Karanian, Brittany Jeye, and two anonymous reviewers for providing
invaluable comments and suggestions on the entire book. I thank
Elizabeth Chua for her expert comments on the transcranial direct
current stimulation section (and for providing a photograph illustrating
this technique) and Lauren Moo for her insightful comments on the
explicit memory and disease chapter. Finally, I thank Jacqueline French
for her skilled copy editing and appreciate all of the professionals at
Cambridge University Press, including Valerie Appleby, Brianda
Reyes, Srilakshmi Gobidass, and Maree Williams-Smith, who made the
production of this book a smooth process.
CHAPTER ONE

Types of Memory and Brain Regions of Interest

Learning Objectives
• To understand each of the memory types.
• To list the brain regions that have been associated with memory.
• To describe the effects of removing the medial temporal lobes.
• To pinpoint the visual sensory regions in the brain.
• To identify the control regions in the brain.

Memory enables us to have skills, to communicate with others, to make

intelligent decisions, to remember our loved ones, and to know who we
are. Although human memory has been studied for over two centuries
(Aristotle, 350 BCE), the cognitive neuroscience of memory has only
been studied for the last two decades. Section 1.1 of this chapter gives
a brief overview of the field of cognitive neuroscience. Cognitive neuros-
cientists employ techniques that non-invasively track the functioning
human brain. Section 1.2 details the fourteen different types of memory.
In section 1.3, an overview of human brain anatomy is provided.
Commonly known anatomic distinctions such as the frontal lobe, the
parietal lobe, the temporal lobe, and the occipital lobe are reviewed and
then more detailed anatomy is discussed. Section 1.4 highlights the impor-
tance of the medial temporal lobe in memory, which was discovered in the
1950s when this region was surgically removed from one unfortunate
individual. In section 1.5, an overview of brain sensory regions is pro-
vided, such as the regions associated with visual perception and auditory
perception. When a person remembers detailed information, such as the
room they stayed in on their last vacation, the corresponding sensory
regions of their brain are reactivated. In section 1.6, the regions of the
brain that control memory retrieval are considered, which include part of
the frontal cortex, the parietal cortex, and the medial temporal lobe.
The final section, 1.7, provides an overview of the organization of this
book. This book identifies the brain regions associated with different
types of memory and details how activity in these regions changes over
time. After the current evidence on the cognitive neuroscience of memory
has been reviewed, the final chapter discusses the future of memory
2 Types of Memory and Brain Regions of Interest

research. In the last decade, there have been many advances in under-
standing the brain mechanisms underlying human memory, but there is
much to learn and the next decade promises to be even more exciting.

1.1 Cognitive Neuroscience

Cognitive psychology is the study of human mental processes such as
perception, attention, imagery, memory, language, and decision making.
Cognitive psychologists dissect these general processes into more specific
processes by identifying behavioral measures that differ between these
processes, such as accuracy or reaction time (see Chapter 2). Behavioral
neuroscience is the study of the brain mechanisms underlying behavior
in animals (see Chapter 10). Behavioral neuroscientists use invasive
methods that can only be used with non-human animals, but they are
ultimately interested in how their findings contribute to the understand-
ing of brain processing in humans. As shown in Figure 1.1, cognitive
neuroscience lies at the intersection of cognitive psychology and beha-
vioral neuroscience. Cognitive neuroscience is the study of the brain
mechanisms underlying human mental processing. Before delving into
the brain regions that have been associated with memory, the specific
types of memory need to be defined.

Figure 1.1 The relationships between the fields of cognitive psychology, cognitive
neuroscience, and behavioral neuroscience.
1.2 Memory Types 3

1.2 Memory Types

In everyday life, the term memory typically refers to consciously retriev-
ing previously experienced information, such as where someone left their
sunglasses before leaving home on a sunny day. However, many different
types of memory are investigated in cognitive neuroscience. To put the
scientific findings that are detailed in this book in the proper framework,
it is necessary to understand each type of memory and how it is related to
the other types of memory.
Figure 1.2 shows the different types of memory and how they are
related to one another. The number of memory types may appear
daunting, but there are major distinctions that divide these into six
pairs of memory types (with each pair of memory types listed at the
same vertical level in the figure). The fact that nearly all memory types
are in pairs indicates that scientists in the field of memory favor
dichotomies. A brief description of each memory type, and how it is
distinct from its paired type, will be provided in this section. A more
thorough description of each memory type will be provided in the
relevant sections of the book.
The first pair of memory types is explicit memory and implicit
memory, which refer to conscious memory and nonconscious

Memory

Explicit Memory Implicit Memory

Long-term Memory Working Memory

Episodic Memory Semantic Memory

Skills

Context Memory Item Memory

Repetition Priming

“Remembering” “Knowing”

Recollection Familiarity

Figure 1.2 Organization of memory types.

4 Types of Memory and Brain Regions of Interest

memory, respectively. That is, all forms of explicit memory are

associated with conscious experience/awareness of the previously
experienced information, whereas all forms of implicit memory are
associated with a lack of conscious experience/awareness of the pre-
viously experienced information. There are many types of explicit
memory, which are described below. Skills constitute one type of
implicit memory. After a skill is learned, performance of that skill
reflects nonconscious memory. For example, after a person learns to
ride a bike, they don’t think about rotating the pedals, steering,
braking, or balancing. Instead, their conscious experience is domi-
nated by where they want to ride or whatever else they happen to be
thinking about. Repetition priming is another category of implicit
memory that refers to more efficient or fluent processing of an item
when it is repeated. For example, when a television commercial is
repeated, that information is processed more efficiently (and when
the item from the commercial is seen again while shopping, implicit
memory presumably increases the chance that it will be purchased).
Skill learning can be assumed to be based on repetition priming (i.e.,
more efficient processing after a lot of practice), which illustrates that
these types of implicit memory are not independent.
The rest of the memory types are kinds of explicit memory.
The second pair of memory types is long-term memory and working
memory. Working memory is also referred to as short-term memory.
A typical explicit memory experiment will be detailed first to help
make the distinction between long-term memory and working mem-
ory. During the study phase of both long-term memory and working
memory paradigms, items such as words or objects are presented. After
the study phase, there is a delay period that can last a variable amount
of time. During the test phase, old items from the study phase and
new items are presented, and participants make an “old” or “new”
judgment for each item, which is referred to as old–new recognition.
Accurate memory is indicated by a greater proportion of “old”
responses to old items than “old” responses to new items. Long-term
memory and working memory differ with regard to whether or not
information is kept in mind during the delay period. In long-term
memory experiments, there are typically many items in the study
phase and the delay period is relatively long (e.g., minutes to hours –
hence the name of this memory type). Participants do not actively
maintain information from the study phase in their mind during the
delay period. In working memory experiments, there are typically a few
items in the study phase, the delay period is in seconds, and participants
1.2 Memory Types 5

are instructed to actively maintain information from the study phase in

their mind (which is working during the delay period, and hence the
name of this memory type). Although explicit memory refers to both
long-term memory and working memory, explicit memory is often used
to refer to only long-term memory. In this book, the terms will be used
according to the definitions provided in this section.
The third pair of memory types is episodic memory and semantic
memory. Episodic memory refers to the detailed retrieval of a previous
episode, such as what occurred, where it occurred, and when it
occurred. For example, when a person remembers the last time they
saw their parents, this is an example of an episodic memory. Semantic
memory refers to retrieval of factual information that is learned over
a long period of time, typically years, such as the definition of a word.
Semantic memories do not involve any memory for the previous learn-
ing episode. For instance, the definition of the word ‘sailboat’ simply
comes to mind without having to think back to when its meaning was
learned. If any information is retrieved from the previous experience,
this would constitute an episodic memory rather than a semantic mem-
ory. As mentioned above, cognitive neuroscience long-term memory
experiments generally consist of a study phase, a delay phase, and a test
phase. Although semantic memory is a type of long-term memory, it is
typically acquired over a period of years. This makes semantic memory
unique and related to language processing (see Chapter 8). As such,
unless otherwise specified, when the term long-term memory is used in
this book, it will refer to all the types of long-term memory except
semantic memory.
The fourth pair of memory types is context memory and item memory.
These are straightforward terms that refer to different kinds of memory
that operate during context memory experiments. During the study
phase of such experiments, items are presented in one of two contexts,
such as on the left or right side of the screen or in red or green. During the
test phase, old items and new items are presented and participants
make an “old”–“new” recognition judgment for each item, and for
items classified as “old” they also make a “context 1” or “context 2”
judgment (e.g., “left” or “right”). It is notable that the second judgment is
based on recall of previous contextual information rather than recogni-
tion, which is almost always the case for context memory judgments.
Recall refers to retrieval of information based on an associated memory
cue (e.g., recalling the context of an old item). Item memory refers to
accurate recognition of old items versus new items, while context mem-
ory refers to accurate retrieval of context information. Context memory
6 Types of Memory and Brain Regions of Interest

is also referred to as source memory, as a particular context can also be

considered a source of information. In addition, associative memory,
which refers to memory for an association between two items, is similar
to context memory in that one item can be considered the context for the
other item.
The fifth pair of memory types is “remembering” and “knowing.”
“Remembering” refers to the subjective experience corresponding to
detailed retrieval, while “knowing” refers to the subjective experience
corresponding to the lack of detailed retrieval. The quotes around
these terms and other behavioral responses that reflect subjective
experience (e.g., “old” and “new”) will be used throughout this
book. “Remembering” corresponds to the subjective mental experi-
ence of retrieving details from the previous experience, such as some-
one retrieving where they parked their car in a parking lot. If any
details are recalled from a previous experience, this constitutes
“remembering”. “Knowing” is defined by the lack of memory for
details from a previous experience, such as when someone is confident
they have seen someone before but not where or when they saw them.
“Remembering” is typically assumed to be related to context memory,
as it is thought to occur whenever contextual information is retrieved.
“Knowing” is typically assumed to be related to item memory and
semantic memory, which is why these memory types are connected in
the figure.
The sixth and last pair of memory types is recollection and
familiarity. The terms recollection and familiarity can refer to math-
ematical models of these two kinds of memory (Slotnick & Dodson,
2005; Wixted, 2007) but more commonly refer to all the forms of
detailed memory (i.e., episodic memory, context memory, and
“remembering”) and non-detailed memory (i.e., semantic memory,
item memory, and “knowing”), respectively. It may be useful to
think of context memory and item memory as measures of task
performance, “remembering” and “knowing” as measures of subjec-
tive experience, and recollection and familiarity as general terms
that describe strong memory and weak memory, respectively.
In one classic paper by Endel Tulving, a world-renowned cognitive
psychologist and cognitive neuroscientist, it was hypothesized that there
was a distinction between “remembering” and “knowing” (Tulving,
1985). This hypothesis stemmed from scientific evidence, as it was
based, in part, on a patient with a brain lesion who had no detailed
memory of the past (i.e., he could not “remember”) but could define
words. Tulving’s hypothesis was also based on introspection, as this was
1.2 Memory Types 7

Box 1.1: The power of introspection

William James, who has been referred to as the father of American
psychology, defined introspection as “the looking into our own minds
and reporting what we there discover” (James, 1890, p. 185). Basically,
introspection means the examination of your own mental processes.
Introspection has proven to be invaluable in cognitive psychology
and cognitive neuroscience and can be used to predict which type(s) of
memory operate during a particular task. Introspection can also be used to
identify which kind(s) of memory may be associated with a particular
event. To illustrate, item memory is a form of long-term memory that is
commonly assumed to reflect “knowing”/familiarity (see Figure 1.2).
However, item memory can also be detailed, which means this event
type can also be associated with “remembering”/recollection (and illus-
trates that the dichotomies in Figure 1.2 are not fixed). Despite the
potential power of introspection, it can lead to problems. It is based on
the experience of the person who is introspecting and can devalue the
experience of others or experimental findings. Thus, in practice, predict-
ing the type(s) of memory involved during a particular task or event
involves a balance between introspection, the insight of others, and data.

a novel proposal and it is clear throughout the paper that his arguments
were based on personal reflection as well as evidence. As discussed in
Box 1.1, introspection is a powerful way for scientists to understand
mental processing. Tulving ran behavioral experiments to test the
hypothesis that “remember” responses and “know” responses were dis-
tinct. During one experiment, words were presented during the study
phase, and then during the test phase old words and new words were
presented and participants made “old”–“new” recognition judgments.
For old items correctly classified as “old,” participants also made
a “remember”–“know” judgment and a confidence-rating judgment
(ranging from 1 to 3 corresponding to low confidence, intermediate
confidence, and high confidence). As shown in Figure 1.3, the probability
of “remember” responses increased with increasing confidence, while the
probability of “know” responses was maximal at the intermediate
confidence rating. These distinct response profiles provide behavioral
evidence in support of Tulving’s hypothesis that “remembering” and
“knowing” are distinct types of memory. A large body of research has
subsequently accumulated showing that “remembering” and “knowing”
are also associated with distinct regions of the brain (see Chapter 3).
8 Types of Memory and Brain Regions of Interest

Figure 1.3 Probability of “remember” or “know” responses as a function of confidence

judgements (key at the top right). Generated using data from Tulving (1985).

1.3 Brain Anatomy

The brain is composed of the occipital lobe, the temporal lobe, the
parietal lobe, and the frontal lobe. Each lobe has gray matter on the
cortical surface, which primarily consists of cell bodies, and white matter
below the surface, which primarily consists of cell axons that connect
different cortical regions. The occipital lobe is associated with visual
processing, the temporal lobe is associated with visual processing and
language processing, the parietal lobe is associated with visual processing
and attention, and the frontal lobe is associated with many cognitive
processes. Over half of the human brain is associated with visual proces-
sing. This illustrates that we are visual animals and is also the reason that
the vast majority of memory studies use visual items as stimuli (e.g.,
written words or pictures of objects).
Figure 1.4 shows the regions of the brain that are of relevance to
memory, which include the occipital cortex, the temporal cortex,
the parietal cortex, the dorsolateral prefrontal cortex, and the medial
temporal lobe. The cortex is folded with gyri protruding out (shown in
light gray) and sulci folding in (shown in dark gray). Figure 1.4A shows
1.3 Brain Anatomy 9

Figure 1.4 Brain regions associated with memory. Each region is shown within red ovals
and labeled. (A) Lateral view of the right hemisphere oriented with the occipital pole to the
left. Cortical surface gyri and sulci in this figure and all subsequent figures are shown in light
and dark gray. (B) Coronal view corresponding to the position in the lateral view indicated by
the dashed vertical line. (C) Axial view corresponding to the position in the lateral view
indicated by the dashed horizontal line. (A black and white version of this figure will appear
in some formats. For the color version, please refer to the plate section.)

a lateral view, as if viewing the brain from the side with the most poster-
ior/back of the brain (i.e., the occipital pole) to the left. The terms super-
ior view and inferior view refer to viewing the brain from directly above
(i.e., a bird’s-eye view) and viewing the brain from directly below (i.e.,
a worm’s-eye view), respectively. Figure 1.4B shows a coronal view, as if
viewing a thin slice of brain that is approximately parallel to the face
(indicated by the vertical dashed line in Figure 1.4A). Figure 1.4C shows
an axial view, as if viewing a thin slice of brain that is approximately
parallel to the ears and nose (indicated by the horizontal dashed line in
10 Types of Memory and Brain Regions of Interest

Figure 1.5 Gyri and sulci in brain regions of interest. Left, lateral view of the left hemisphere
(occipital pole to the right). Right, inferior view of the left hemisphere (occipital pole at
the bottom).

Figure 1.4A) with the occipital pole to the left. The medial temporal lobe
of each hemisphere consists of the hippocampus (labeled in
Figure 1.4B) and the immediately surrounding cortex. The dorsolat-
eral prefrontal cortex (shown in Figures 1.4A and 1.4C) is a large part
of the frontal cortex that consists of the dorsal and lateral surface
that is anterior to the motor processing regions, which are in the
posterior frontal cortex (described in the next paragraph). Cognitive
neuroscience brain activation results are usually shown on a cortical
surface (such as Figure 1.4A) and/or on a slice through the cortex (such
as Figures 1.4B and 1.4C).
In scientific articles, such as the recommended readings at the end of
each chapter in this book, a brain activation is almost always localized to
a specific gyrus or sulcus. Figure 1.5 shows the names of gyri and sulci that
are of particular relevance in the field of memory. Only the left hemi-
sphere is shown, as both hemispheres have the same organization. Many
of the names are straightforward such as the superior frontal gyrus, the
middle frontal gyrus, and the inferior frontal gyrus, which refer to their
respective spatial locations (i.e., the upper, middle, and lower parts of
the frontal lobe). Note that the superior frontal sulcus is between the
superior frontal gyrus and the middle frontal gyrus, and the inferior
1.3 Brain Anatomy 11

4 3 1
6
2
8 5
7
9

19
46
40
10 39 18

45 44 43 41
42
47 22
17
11
21
38 18
37 19
20

Figure 1.6 Brodmann map (1909). The left hemisphere with Brodmann areas labeled
(lateral view, occipital pole to the right).

frontal sulcus is between the middle frontal gyrus and the inferior frontal
gyrus. The central sulcus separates the frontal lobe from the parietal lobe.
The motor processing regions in the posterior frontal cortex include the
anterior bank of the central sulcus, the precentral gyrus, and the precen-
tral sulcus. The inferior parietal lobule, just below the intraparietal
sulcus, consists of the supramarginal gyrus and the angular gyrus.
The lateral sulcus is also called the Sylvian fissure. The first visual sensory
processing region, V1, lies within the calcarine sulcus, which runs along
the middle of the medial surface of the occipital lobe (the medial surface
is the flat part of the brain along the left of the inferior view shown to the
right in the figure). Related to this, a medial view refers to viewing
a hemisphere from the opposite direction as a lateral view.
A brain activation is also often localized to specific Brodmann area
(BA). Figure 1.6 shows the Brodmann area map. Korbinian Brodmann
created this map over a century ago based on different anatomic
characteristics within each region such as cell shape, layering, and density
(Brodmann, 1909). Such anatomic difference can be assumed to reflect
functional differences, which means that each Brodmann area might be
associated with a particular cognitive process. In reality, brain processing
is very complex and each brain region is associated with multiple
12 Types of Memory and Brain Regions of Interest

cognitive processes and multiple brain regions interact during each cog-
nitive process (see Chapter 11). However, there is still some degree of
functional specialization within each brain region. There are regions in
common between the Brodmann map and the gyri/sulci map. BA17 is the
same as V1, which lies within the calcarine sulcus. BA39 and BA40
correspond to the angular gyrus and the supramarginal gyrus, respec-
tively. The lateral part of BA7 (i.e., the part that can be seen in the figure)
corresponds to the superior parietal lobule. The medial part of BA7 (i.e.,
the part that cannot be seen in the figure) corresponds to the precuneus.
Both the superior parietal lobule and the precuneus have been associated
with memory. BA4 and BA6 are motor processing regions.
All scientific studies report the gyri/sulci and/or Brodmann area(s)
associated with each brain activation. Although this level of anatomic
detail is not emphasized in this book, the suggested readings report very
specific results. The gyri/sulci map and the Brodmann map (Figures 1.5
and 1.6) can be referred to as needed.

1.4 The Hippocampus and Long-Term Memory

In the 1950s, a radical surgical procedure was conducted in an effort to
relieve the epileptic seizures of a 29-year-old man named Henry
Molaison, who was referred to until recently as patient H. M. (Scoville
& Milner, 1957). As shown in Figure 1.7, this patient had the hippocam-
pus and the surrounding cortical regions removed. The medial temporal
lobe is shown intact in one hemisphere to illustrate the resected region in
the other hemisphere, but the medial temporal lobe was actually
removed in both hemispheres. The surgery did not affect his intelligence
or personality, but it did cause a severe deficit in long-term memory,
which is referred to as amnesia (semantic memory was intact; see
the second section of this chapter). In particular, he had almost no
memory of events that occurred a few years before the surgery (i.e.,
retrograde amnesia) and had no memory for events that occurred after
the surgery (i.e., anterograde amnesia), but his memory for earlier events
appeared to be normal. For example, 10 months before the surgery he
and his family moved to a new house a few blocks away from their old
house. After the surgery, he had no memory for his new address, he could
not find his way to the new home, and he did not know where objects
were kept in the new home (e.g., he did not know where the lawnmower
was even if he had used it the day before). He had no familiarity
with magazines he had read before, so would read the same articles
repeatedly. He would eat lunch and a half-hour later could not remember
1.5 Sensory Regions 13

A
8 cm B
C

Hippocampus

Figure 1.7 Depiction of medial temporal lobe resection in patient H. M. Left, inferior view
of the brain illustrating the spatial extent (8 centimeters) of the medial temporal lobe
resection (the shaded region; occipital pole at the bottom). Right, coronal view
corresponding to the dotted line labeled B to the left. The hippocampus is labeled to the
right and the resected medial temporal lobe region, which included the hippocampus and
the surrounding cortex, is illustrated to the left (in black). Both medial temporal lobes
were removed in the patient.

he had eaten. Despite his severe deficit in long-term memory, his working
memory appeared to be intact. He could remember a pair of words or
a three-digit number for several minutes as long as he was not distracted.
These results indicate that hippocampus and the surrounding cortical
regions are critical for long-term memory, which will be supported by
numerous findings in this book.
Long-term memory typically refers to retrieval of previously presented
information. However, the key stages of long-term memory include
encoding, storage, and retrieval. The hippocampus has been associated
with both long-term memory encoding and long-term memory retrieval
(see Chapter 3). Long-term memory storage depends on a process called
memory consolidation, which refers to changes in the brain regions,
including the hippocampus, underlying long-term memory (see
Chapter 3). Thus, all three stages of long-term memory depend on the
hippocampus.

1.5 Sensory Regions

If a person recalls what they had for dinner the night before, they will
almost certainly have a visual experience of what it looked like. This
subjective experience supports the sensory reactivation hypothesis,
14 Types of Memory and Brain Regions of Interest

Figure 1.8 Sensory brain regions of interest. Left, lateral view of the left hemisphere
(occipital pole to the right). Right, inferior view of the left hemisphere (occipital pole at the
bottom). Visual sensory regions (within red ovals) are labeled according to the type of
processing (with the name of each region in parentheses). The arrows (in red) illustrate the
where pathway and the what pathway. Non-visual sensory regions are also illustrated
(within blue ovals) and labeled. (A black and white version of this figure will appear in some
formats. For the color version, please refer to the plate section.)

which is that memory for an event can activate the same brain regions
associated with perception of that event. These sensory memory effects
reflect the contents of memory (e.g., memory for a visual experience
contains visual information). Before considering the evidence from mem-
ory studies that support the sensory reactivation hypothesis, the sensory
regions of the brain associated with visual processing, language/auditory
processing, motor processing, and olfactory processing will be briefly
reviewed. Given that almost all memory studies employ visual stimuli,
visual sensory brain regions will be emphasized.
Figure 1.8 shows brain regions that have been associated with visual
perception (in red). When an object is perceived, the first visual cortical
area called V1 or striate cortex in the back of the brain processes the
features of that object including shape, color, location, and motion.
The name striate cortex comes from the striated appearance of this
region when it is stained with dye. The object continues to be processed
in more anterior brain regions called the extrastriate cortex, so named
because these regions are in addition to the striate cortex. A broad
1.5 Sensory Regions 15

distinction between processing in V1, extrastriate cortex, and more ante-

rior visual cortical regions is that more ventral visual regions (toward the
bottom of the brain) are associated with processing object identity and
more dorsal visual regions (toward the top of the brain) are associated
with processing object location. As such, the visual processing regions
from V1 to ventral extrastriate cortex to ventral temporal cortex are
referred to as the what pathway, and the visual processing regions from
V1 to dorsal extrastriate cortex to parietal cortex are referred to as the
where pathway. These pathways are also hierarchical in nature, with
lower-level processing occurring in early visual regions, such as V1, and
higher-level processing occurring in late visual regions, such as more
anterior ventral temporal cortex (Felleman & Van Essen, 1991).
V1 and extrastriate cortex are left-right reversed such that objects in
the left visual field (i.e., the left side of space) are processed by V1 and
extrastriate cortex in the right hemisphere and objects in the right visual
field (i.e., the right side of space) are processed by V1 and extrastriate
cortex in the left hemisphere. This mapping of the left visual field and the
right visual field onto right early visual areas and left early visual areas,
respectively, is referred to as contralateral visual processing. Extrastriate
cortex includes regions that are specialized for processing different visual
features. A region associated with processing shape is called the lateral
occipital complex (LOC), the eighth visual region called V8 is associated
with processing color, and there is a region associated with processing
motion called MT (this region is named after the middle temporal area in
monkeys and the same label is used even though it is in a different
location in humans).
More complex object processing occurs in more anterior ventral visual
processing regions. These include a face processing region called the
fusiform face area (FFA, which is within the fusiform gyrus) and
a context processing region called the parahippocampal place area
(PPA), which is within the parahippocampal gyrus and is activated for
stimuli that can represent visual context such as places or scenes. It should
be emphasized that even though there are regions that are specialized for
processing certain features or stimulus types, this does not mean these are
the only regions associated with that type of processing. For example,
even though the FFA is a face processing region, there are at least eleven
face processing regions in the brain (Slotnick & White, 2013; see
Chapter 11). This illustrates that objects are represented in the brain by
the pattern of activity across many visual regions (Haxby et al., 2001),
rather than the activity within one visual region. One major problem in
the field of cognitive neuroscience is the popular view that one brain
16 Types of Memory and Brain Regions of Interest

region can be associated with one cognitive process, but this overly
simplistic view is never correct (see Chapter 11). Figure 1.8 also shows
brain regions that have been associated with language processing,
motor processing, and olfactory processing (shown in blue). Language
processing includes auditory/sound processing (in the more posterior
region shown in the figure), word comprehension, and word production
(see Chapter 8).
There is a large body of research that supports the memory sensory
reactivation hypothesis (Slotnick, 2004b). Memory for visual informa-
tion, language information (i.e., sounds or words), motor information
(i.e., actions), and olfactory information (i.e., odors) reactivate the cor-
responding sensory regions of the brain. Within the visual processing
regions, there is also evidence that memory for faces and houses activate
the FFA and PPA, respectively. In the last decade, evidence has also
accumulated that memory for specific features activate the correspond-
ing feature processing brain region. Memory for shape activates LOC
(Karanian & Slotnick, 2015), memory for color activates V8 (Slotnick,
2009a), memory for items in the left visual field or the right visual field
activate the extrastriate cortex in the opposite/contralateral hemisphere
(Slotnick & Schacter, 2006; Slotnick, 2009b), and memory for motion
activates region MT (Slotnick & Thakral, 2011). One functional magnetic
resonance imaging (fMRI) study compared the sensory activity asso-
ciated with the recall of objects and the recall of sounds (Wheeler &
Buckner, 2000). As will be discussed in Chapter 2, fMRI measures the
increases in blood flow that occur in active brain regions. All that needs
to be known at this point is that fMRI can be used to identify the
specific regions of the brain that are associated with a particular cognitive
process. During the study phase, participants saw pictures of objects (e.g.,
a dog) or heard sounds of objects (e.g., the sound of a train) along with
the corresponding word labels (e.g., ‘dog’ or ‘train’). During the test
phase, the previous word labels were presented and participants were
asked to recall whether the corresponding item was previously “seen”
or “heard.” Figure 1.9A shows fMRI activity in the extrastriate cortex
associated with the perception of pictures (in blue/green) and Figure 1.9B
shows fMRI activity in the extrastriate cortex associated with recall of
pictures. Figure 1.9C shows fMRI activity in the auditory processing
cortex associated with the perception of sounds (in red/yellow) and
Figure 1.9D shows fMRI activity in the auditory processing cortex
associated with recall of sounds. These findings illustrate that picture
memory and sound memory reactivate the same regions associated with
picture perception and sound perception, respectively. It is notable that
1.5 Sensory Regions 17

Figure 1.9 Sensory fMRI activity associated with perception and memory. (A) fMRI
activity associated with visual perception (axial view, occipital pole at the bottom). (B) fMRI
activity associated with visual memory (arrow indicates extrastriate cortex). (C) fMRI activity
associated with perception of sounds. (D) fMRI activity associated with memory for sounds
(arrow indicates auditory sensory cortex). (A black and white version of this figure will appear
in some formats. For the color version, please refer to the plate section.)

the spatial extent of activity associated with memory is much smaller

than the spatial extent of activity associated with perception. This is
because the subjective experience associated with memory is not as
detailed as the subjective experience associated with perception.
18 Types of Memory and Brain Regions of Interest

1.6 Control Regions

Control regions guide the construction of explicit memories. Two regions
associated with memory control are the dorsolateral prefrontal cortex
and the parietal cortex. These regions mediate different functions during
memory. For instance, the dorsolateral prefrontal cortex is thought to be
involved in memory selection and the parietal cortex is thought to be
involved in attending to the contents of memory (see Chapters 3 and 8).
The medial temporal lobe, including the hippocampus (which was dis-
cussed in section 1.4 of this chapter), is also considered a control region
(see Chapter 3). When control regions modulate activity in sensory
regions, this is sometimes referred to as a top-down interaction.
One fMRI study aimed to identify the brain regions associated with
memory control (Wheeler & Buckner, 2003). A similar paradigm
described at the end of section 1.5 was employed. During the study
phase, word labels were presented, immediately followed by the
corresponding pictures or sounds and these items were presented either
1 time or 20 times. During the test phase, old word labels from the study
phase or new word labels were presented and participants made
“seen”–“heard”–“new” recognition judgments. It can be assumed that
recognition of items that were presented 1 time during the study phase
required more control than items that were presented 20 times (as
retrieval of items presented numerous times is a relatively automatic
process). A comparison of old word labels that were presented 1 time
versus old word labels that were presented 20 times produced activity in
the dorsolateral prefrontal cortex and the parietal cortex, which indicates
that these regions are associated with memory control.
Another fMRI study assessed whether item memory and context/source
memory produced activity in different regions of the brain (Slotnick, Moo,
Segal & Hart, 2003). This study highlights the comparisons used to isolate
these types of memory and illustrates the corresponding control regions.
As shown in Figure 1.10A, top, during the study phase of both item
memory runs and source memory runs, participants were presented with
abstract shapes in the left visual field or the right visual field and were
instructed to remember each shape and its spatial location. Abstract
shapes were employed to minimize language/verbal processing strategies.
As shown in Figure 1.10A, bottom left, during the test phase of item
memory runs, old and new shapes were presented in the center of the
screen and participants made “old”–“new” recognition judgments.
As shown in Figure 1.10A, bottom right, during the test phase of source
memory runs, old shapes were presented in the center of the screen and
 A B

Item Memory Source Memory

0.1 0.1
Remember Each Shape Remember Each Shape Source Memory
and Side of Screen and Side of Screen

% Signal Change

% Signal Change
Item Memory
Correct Rejection
0 0

Time Time

–0.1 –0.1

Old or New? Right Side or Left Side?

Time Time

Figure 1.10 Item memory and source memory paradigm and fMRI results. (A) Left, illustration of item memory task. Right, illustration of source
memory task. (B) Bottom, fMRI activity associated with source memory (in red) and item memory (in yellow) in the dorsolateral prefrontal cortex and
the parietal cortex (axial view, occipital pole at the bottom). Top, the magnitude of activity (in percent signal change) associated with each event type
extracted from the two circled dorsolateral prefrontal cortex activations (key at the top right). (A black and white version of this figure will appear in
some formats. For the color version, please refer to the plate section.)
20 Types of Memory and Brain Regions of Interest

participants recalled whether each item was previously presented on the

“right” or “left.” To isolate brain activity associated with item memory, the
magnitude of activity associated with new shapes that were correctly
classified as “new” (i.e., correct rejections) was subtracted from the mag-
nitude of activity associated with old shapes that were correctly classified
as “old” (i.e., old-hits) in all regions of the brain. That is, old-hits, which
reflect item memory, were contrasted with new-correct rejections,
a baseline event that did not reflect item memory (as correct rejections
do not involve memory). Brain regions in which the magnitude of activity
associated with old-hits versus new-correct rejections was significantly
greater than zero were assumed to be associated with item memory.
The old-hit versus new-correct rejection comparison is a classic contrast
used to isolate brain activity associated with item memory. To isolate
brain activity associated with source memory, accurate source memory,
which required accurate item memory and spatial location memory, was
contrasted with accurate item memory (i.e., old-hits). As discussed in
Box 1.2, this contrast illustrates subtractive logic, where one process

Box 1.2: Isolating a process with subtractive logic

Subtractive logic has been used for well over a century to measure the speed
of nerve conduction (Helmholtz, 1850) and the speed of mental processing
(Donders, 1868). It is based on the assumption that the two event types differ
only with regard to the process of interest. Although subtractive logic is
widely used in cognitive neuroscience, it produces interpretable results only if
the event types differ by only a single cognitive process. When considering
cognitive neuroscience results, keep in mind that the cognitive processes
associated with an event type can be determined, in part, through introspec-
tion (see Box 1.1). If multiple cognitive processes differ between two event
types, the results are confounded and could be attributed to any of these
processes rather than the cognitive process of interest. To avoid confounded
results, a convincing case needs to be made that the contrast employed
isolates the single cognitive process of interest. Consider the classic contrast
between old-hits and new-correct rejections. These event types differ in that
only old-hits are associated with item memory. However, they also differ in
that only old items have been seen before and thus will produce repetition
priming effects. Therefore, it is uncertain whether brain activity produced by
this contrast is due to item memory or repetition priming. By comparison, the
contrast between old-hits and old-misses (i.e., forgotten old items that are
classified as “new”) does isolate item memory, because this is the only type
of memory that differs between these event types.
1.7 The Organization of This Book 21

(in this case item memory) that is associated with two event types is
subtracted out to isolate the process of interest (in this case source mem-
ory). As shown in Figure 1.10B, bottom, source memory produced activity
in the left dorsolateral prefrontal cortex (in red), while item memory
produced activity in the right dorsolateral prefrontal cortex and the
parietal cortex (in yellow) along with the medial temporal lobe (not
shown). Figure 1.10B, top, shows the magnitude of activity (in percent
signal change, which will be detailed in Chapter 2) associated with accurate
source memory, accurate item memory, and new-correct rejections in the
dorsolateral prefrontal cortex. These activation profiles show that each
dorsolateral prefrontal cortex region is associated with only item memory
or source memory.
The fMRI results in this section and section 1.5 provide a brief intro-
duction to the sensory brain regions and control brain regions associated
with memory. As will be discussed in Chapter 2, fMRI is just one of the
many tools employed in cognitive neuroscience to investigate the brain
mechanisms underlying memory.

1.7 The Organization of This Book

Chapter 2 of this book, “The Tools of Cognitive Neuroscience,” pro-
vides a brief overview of the techniques employed in this field.
Research in cognitive neuroscience is completely dependent on these
methods, which each have their strengths and weaknesses. The next
eight chapters are organized, in part, by memory type. Most chapters
focus on long-term memory, as this is the most widely studied type of
memory. Chapter 3, “Brain Regions Associated with Long-Term
Memory,” details the spatial location of brain regions that have been
associated with long-term memory. Chapter 4, “Brain Timing
Associated with Long-Term Memory,” discusses the timing of the
activity in the brain regions associated with long-term memory.
Chapter 5, “Long-Term Memory Failure,” provides an overview of
the brain mechanisms that underlie long-term memory failure, such
as forgetting and false memory. Chapters 6 and 7, “Working Memory”
and “Implicit Memory,” respectively, discuss the spatial location and
timing of brain regions associated with these kinds of memory.
Chapter 8, “Memory and Other Cognitive Processes,” discusses the
similarities and differences between memory and other cognitive
processes such as attention, imagery, and language. Chapter 9,
“Explicit Memory and Disease,” reviews diseases that affect explicit
memory, such as Alzheimer’s disease. Chapter 10, “Long-Term
22 Types of Memory and Brain Regions of Interest

Memory in Animals,” details the findings from long-term memory

experiments with animals such as rats and monkeys. The final chapter,
“The Future of Memory Research,” provides a synopsis of where the
field of cognitive neuroscience stands now and what the field needs to
do in the future to understand the brain mechanisms underlying mem-
ory. This will require a major shift in how research is conducted in the
field such that we not only investigate which brain regions are asso-
ciated with memory but also investigate when these brain regions are
active and how they interact with one another. This will require embra-
cing more complex techniques, which will be a challenge, but it is also
a very exciting time to investigate the cognitive neuroscience of
memory.

Chapter Summary
• The six pairs of memory types are explicit memory and implicit
memory, long-term memory and working memory, episodic memory
and semantic memory, context memory and item memory, “remem-
bering” and “knowing,” and recollection and familiarity.
• The five brain regions that have been associated with memory are the
occipital cortex, the temporal cortex, the parietal cortex, the dorsolat-
eral prefrontal cortex, and the medial temporal lobe.
• Removal of the medial temporal lobes in patient H. M. caused a
complete loss of long-term memory.
• There are different sensory regions associated with visual processing,
language processing, motor processing, and olfactory processing.
• Within the visual modality, there are different sensory regions asso-
ciated with processing shape (in LOC), color (in V8), spatial location
(in V1 and extrastriate cortex), motion (in MT), faces (in the FFA),
and context (in the PPA).
• The memory control regions in the brain are the dorsolateral prefron-
tal cortex, the parietal cortex, and the medial temporal lobe.

Review Questions
How do explicit memory and implicit memory differ?
How do recollection and familiarity differ?
What are three brain regions that have been associated with memory?
Does region V8 process color or motion?
Is the dorsolateral prefrontal cortex a sensory region or a control
region?
Further Reading 23

Further Reading
Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26, 1–12.
This classic paper introduced “remembering” and “knowing” and
illustrates introspection.
Scoville, W. B. & Milner, B. (1957). Loss of recent memory after bilateral
hippocampal lesions. Journal of Neurology, Neurosurgery, & Psychiatry,
20, 11–21.
This landmark study shows that medial temporal lobe lesions produce
a profound impairment in long-term memory.
Wheeler, M. E., Petersen, S. E. & Buckner, R. L. (2000). Memory’s echo:
Vivid remembering reactivates sensory-specific cortex. Proceedings of
the National Academy of Sciences of the United States of America, 97,
11125–11129.
This fMRI paper illustrates that memory for visual information and
auditory information produce activity in the same brain regions that are
associated with visual perception and auditory perception.
Slotnick, S. D., Moo, L. R., Segal, J. B. & Hart, J., Jr. (2003). Distinct
prefrontal cortex activity associated with item memory and source
memory for visual shapes. Cognitive Brain Research, 17, 75–82.
This fMRI paper shows that item memory and source memory are
associated with activity in the dorsolateral prefrontal cortex, the parietal
cortex, and the medial temporal lobe.
CHAPTER TWO

The Tools of Cognitive Neuroscience

Learning Objectives
• To describe how fMRI measures brain activity and characterize this
method’s spatial resolution and temporal resolution.
• To describe how ERPs measure brain activity and characterize this
method’s spatial resolution and temporal resolution.
• To list one problem with patient lesion evidence.
• To describe how TMS works and characterize this method’s spatial
resolution and temporal resolution.
• To name two methods that could be combined to measure brain activity
with excellent spatial resolution and excellent temporal resolution.

Cognitive neuroscientists employ tools to look inside the brain of partici-

pants while they are actively engaged in a mental process. This is no simple
feat, and the field of cognitive neuroscience has grown with the advent of
techniques that can measure activity in the functioning human brain.
These methods vary in popularity, cost, complexity, spatial resolution,
and temporal resolution. Each technique has advantages and disadvan-
tages and takes years to master. This chapter briefly describes the most
widely used techniques in cognitive neuroscience that will be referred to
throughout this book. Section 2.1 briefly reviews the behavioral measures
that allow for the interpretation of brain activation results. Section 2.2
discusses techniques with high spatial resolution, such as fMRI, which is
the most popular method. fMRI measures the increases in blood flow that
occur in active brain regions. This technique has excellent spatial resolu-
tion but has poor temporal resolution because the blood flow response is
slow. Section 2.3 focuses on techniques with high temporal resolution,
such as event-related potentials (ERPs). ERPs measure voltages (i.e.,
potentials) on the scalp that directly reflect the underlying brain activity.
This technique has excellent temporal resolution and limited spatial
resolution. In section 2.4, techniques with excellent spatial resolution and
excellent temporal resolution are described. These include combined
fMRI and ERPs as well as depth electrode recording from patients who
have electrodes implanted in their brains for clinical reasons. Section 2.5
2.2 High Spatial Resolution Techniques 25

considers evidence from patients with brain lesions and cortical deactiva-
tion methods such as transcranial magnetic stimulation (TMS). Both of
these methods have limited spatial resolution and poor temporal resolu-
tion; however, they can assess whether a brain region is necessary for
a given cognitive process. In section 2.6, the spatial resolution and
temporal resolution of the different techniques are compared. It is con-
cluded that only combined methods, such as fMRI and ERPs, that have
excellent spatial resolution and excellent temporal resolution can be
widely used to track the spatial-temporal dynamics of the functioning
brain. Such combined techniques are the future of cognitive neuroscience
(see Chapter 11).

2.1 Behavioral Measures

As mentioned in Chapter 1, cognitive psychologists use behavioral mea-
sures such as accuracy, reaction time, and subjective experience to isolate
different cognitive processes. Cognitive psychologists typically conduct
research without any consideration of brain activity, as it is not thought to
be necessary to inform the understanding of cognitive processing.
Cognitive neuroscientists use behavioral measures as well, but these are
considered in conjunction with measures of brain activity. For instance,
the comparison of old-hits and old-misses (i.e., accurate versus inaccurate
responses) can be used to isolate brain activity associated with item
memory (see Chapter 1). “Remember” and “know” responses, which
reflect types of subjective experience, are associated with unique patterns
of brain activity (see Chapter 4). Reaction times are faster for old items
than new items, and comparing these event types can isolate brain
activity associated with repetition priming, a type of implicit memory
(see Chapter 7). Although the remaining sections of this chapter focus
on techniques used to measure brain activity, it is important to keep in
mind that brain activity is only meaningful in light of the corresponding
behavioral measures.

2.2 High Spatial Resolution Techniques

fMRI is the most widely used technique in the field of cognitive
neuroscience (see Chapter 11). Using this method, the brain regions
associated with a particular cognitive process can be localized with
excellent spatial resolution. Specifically, the spatial resolution of fMRI
is a few millimeters, which is sufficient to answer numerous questions in
the field of cognitive neuroscience.
26 The Tools of Cognitive Neuroscience

The physics behind fMRI is very complicated and there are entire
textbooks dedicated to this technique (e.g., Huettel, Song & McCarthy,
2014). A brief review is provided here to give a sense of what the fMRI
signal represents. During fMRI, a participant lies on a scanner bed with
their head inside the scanner bore. Figure 2.1A shows an MRI scanner,
which can be used to acquire extremely high resolution anatomic MRI
images (with a typical spatial resolution of 1 millimeter) and fMRI
images (with a typical spatial resolution of 4 millimeters). Note that
both MRI images and fMRI images are acquired on the same MRI
machine using different data acquisition protocols. The participant lies
on their back on the scanning table with their head toward the MRI
machine. Then, the top of the table slides such that their head and body
is inside the scanner bore (i.e., the circular hole). In the figure, the feet
of the participant, which are covered by a white sheet, are shown
protruding from the scanner bore. A huge coil of superconducting
wire surrounds the scanner bore and current flows through this coil.
This produces a strong magnetic field directed along the axis of the bore
(i.e., in the feet-to-head direction). Protons within the brain, which are
hydrogen ions dissociated from water and fat, are usually oriented in
random directions. In the scanner bore, protons act like tiny magnets
and align with the large magnetic field. The protons also rotate/precess,
like a spinning top, at a specific frequency. While a participant is
performing a cognitive task, a stimulating coil applies a smaller mag-
netic field that knocks over the protons in the brain so they precess
perpendicular/orthogonal to the large magnetic field (i.e., in a plane
approximately parallel with the nose and ears). If a brain region is
active because it is involved in the cognitive task, the amount of oxyge-
nated blood in that region will increase. This increase in oxygenated
blood stabilizes the orthogonal precessing protons in that region
(because deoxygenated hemoglobin in the blood destabilizes protons,
such that they return to being aligned with the large magnetic field).
Other magnetic fields are applied so the orthogonal protons precess at
unique frequencies in different spatial locations of the brain. These
orthogonal proton frequency signals are detected by a receiving coil
and then these signals are used to construct an image that identifies the
specific region of the brain associated with that cognitive process.
Of importance, the magnetic fields applied during fMRI are in the low-
energy radio frequency range, so this is a completely safe technique.
The key points are that engaging in a cognitive process (such as memory
retrieval) increases activity in particular brain regions that causes an
increase in blood flow that is detected as fMRI signal.
2.2 High Spatial Resolution Techniques 27

B
379
378
377
376
375
374
0 2 4 6 8 10 12 14
TIME (sec)

C
PERCENT SIGNAL CHANGE

EXTRASTRIATE

.5

L.PREFRONTAL

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
TIME (SEC)

Figure 2.1 MRI scanner and fMRI results. (A) MRI scanner with a participant’s legs (covered
by a sheet) protruding from the bore. (B) Left, one participant’s fMRI activity associated with
28 The Tools of Cognitive Neuroscience

The physics behind fMRI illustrates that this technique measures an

increase in blood flow that is correlated with brain activity rather than
measuring brain activity itself. Almost all fMRI studies today employ an
event-related design, which refers to experimental protocols with
a mixture of different events such that brain activity associated with
individual events types can be identified. In the very first event-related
fMRI study, participants were presented with word stems (e.g., ‘COU’,
‘GRE’) that were shown for 1.5 seconds every 14 to 16 seconds, and the
task was to complete the stems by generating words (e.g., “couple,”
“green”; Buckner et al., 1996). Figure 2.1B, left, shows the brain regions
for one participant in which activity was correlated with the word stem
completion task (more significant activity is shown in yellow). These
regions included the extrastriate cortex (at the bottom), which can be
assumed to reflect visual processing, and the left dorsolateral prefrontal
cortex (at the top), which can be assumed to reflect semantic memory
retrieval (see Chapters 1 and 8). Figure 2.1B, right, shows the word stem
stimulation period (the cyan square) and the corresponding event-
related timecourse of activity within the left dorsolateral prefrontal
cortex (measured by the intensity of fMRI activity over time). There
are a couple of important points to make about the timecourse of fMRI
activity. First, even though the word stem/stimulus was presented in the
0- to 1.5-second period, the magnitude of fMRI activity did not increase
above the baseline level until 4 seconds after stimulus onset. Second, the
fMRI activity did not return to the baseline level until about 10 seconds
after stimulus offset. Figure 2.1C shows the event-related timecourses of
activity within the extrastriate cortex and the left dorsolateral prefrontal
cortex for another participant. The same pattern of activity can be
observed with the magnitude increasing above baseline about 4 seconds
after stimulus onset and not returning to baseline until about 10 seconds

Caption for Figure 2.1 (cont.)

word stem completion (more significant activity is shown in yellow; axial view, occipital pole
at the bottom). Extrastriate cortex activity is shown at the bottom and dorsolateral prefrontal
cortex activity is shown at the top. Right, activation timecourse (intensity as a function of
time after stimulus onset, in seconds) extracted from the left dorsolateral prefrontal cortex
activation. The cyan square represents the stimulus period. (C) Activation timecourses
(percent signal change as a function of time after stimulus onset) extracted from the
extrastriate cortex and the left dorsolateral prefrontal cortex of another participant. (A black
and white version of this figure will appear in some formats. For the color version, please
refer to the plate section.)
2.2 High Spatial Resolution Techniques 29

after stimulus offset. In this case, the magnitude is measured in percent

signal change, which is calculated by taking the difference between
the intensity at each time point and the baseline level of intensity
and then dividing this difference by the baseline intensity. This adjusts
the magnitude such that its value is approximately 0 at stimulus onset
and its maximum value is approximately 1. For example, in Figure 2.1B,
right, the percent signal change for 6 seconds after stimulus onset is
approximately 0.8, which can be calculated from (378 – 375)/375 × 100
(multiplying by 100 converts the value into a percentage).
The previous event-related timecourse results show that fMRI activ-
ity is temporally delayed and extended in time. Even though changes in
brain activity are very rapid (on the scale of milliseconds), blood flow is
very sluggish (on the scale of seconds). This means that fMRI cannot be
used to investigate the timecourse of the functioning brain. For
instance, fMRI could not measure the temporal dynamics of retrieval
from long-term memory, which takes approximately 2 seconds. fMRI
can only provide a static picture of all the brain regions that were active
during a particular cognitive process. Thus, although fMRI has excel-
lent spatial resolution, it has poor temporal resolution. Another major
limitation of fMRI is its cost. MRI machines cost millions of dollars to
purchase and hundreds of thousands of dollars per year to maintain
(see Chapter 11). As discussed in Box 2.1, even though fMRI has poor
temporal resolution, researchers who employ this technique often
believe this method can be used to investigate the temporal dynamics
of brain function.

Box 2.1: fMRI cannot be used to investigate temporal

processing
fMRI has poor temporal resolution, but many scientists who use this technique
believe it can track the temporal dynamics of brain function. For example,
based on the fMRI timecourses from a single participant shown in Figure 2.1C,
it could be argued that the extrastriate cortex was activated earlier in time than
the left prefrontal cortex. However, because fMRI activity reflects blood flow
rather than neural activity, this difference in the timecourse of activation could
just as well have corresponded to a relatively more sluggish blood flow
response in the dorsolateral prefrontal cortex. The major limitation of fMRI is
that this technique provides little if any information about the temporal
dynamics of brain function. Fortunately, there are other tools in the toolbox
that provide excellent temporal resolution.
30 The Tools of Cognitive Neuroscience

Positron emission tomography (PET) is another technique that has

historically been used in the field of cognitive neuroscience. PET has
relatively high spatial resolution and, like fMRI, measures an increase
in blood flow to active brain regions. Before the participant engages
in a task, a low level of radioactive material is injected into their
bloodstream. During the task, there is an increase in blood flow within
active brain regions and this increases radioactive emissions that are
detected and localized to these regions. The temporal resolution of
PET is approximately half a minute. This means all PET studies can be
conducted only using a blocked design, which refers to protocols where
each period (that usually lasts longer than 10 seconds) consists of
a series of the same events. One major problem with blocked designs,
as compared to event-related designs, is that there can be general
processing differences between the types of blocks. For example, if
one type of block is relatively more difficult than another type of block,
this can confound the results because it is unknown whether activity
identified by comparing the blocks is due to differences in the cognitive
process of interest or due to differences in difficulty. As compared to
fMRI, PET has lower spatial resolution, lower temporal resolution,
and is harmful due to the use of radioactive material. As such, PET has
only rarely been used in the field of cognitive neuroscience within the
last decade, since fMRI has become widely available. PET is described
here as it is mentioned in some of the suggested further readings at the
end of each chapter, but results from this technique are almost never
considered in this book.

2.3 High Temporal Resolution Techniques

Event-related potentials (ERPs) can track brain activity in real time.
The term potential is just another word for voltage, and, as its name
implies, ERP studies use event-related designs. During ERP recording,
the participant sits in a comfortable chair and electrodes are placed
on their scalp. Figure 2.2A shows a participant in an ERP experiment
wearing a nylon cap embedded with 128 electrodes along with electrodes
around the eyes to monitor eye movements. This type of system costs
about $100,000 (US) and has no maintenance fee. During a cognitive
task, the underlying brain activity generates electric fields that induce
tiny voltages on the scalp. These fields are generated by adjacent positive
and negative charges (i.e., dipoles) that are created by neuronal activity
that is perpendicular to the cortical surface (Nunez & Srinivasan, 2005).
The voltage at each electrode is amplified approximately 100,000 times
2.3 High Temporal Resolution Techniques 31

B
Rscene – new

anterior

left right

posterior
–0.4 to 2.1 –1.5 to 2.5 0.1 to 2.9 –0.6 to 2.8

500–800 800–1100 1100–1400 1400–1900

Figure 2.2 ERP setup and results. (A) ERP setup that includes a comfortable chair,
a 128-channel electrode cap, and amplifiers (to the right of the chair). (B) ERP topographic
maps (superior views, occipital poles at the bottom; key to the left) associated with
remembering a word was previously paired with a scene versus correctly rejecting new words
as a function of time period (in milliseconds, shown at the bottom below each topographic
map). Electrodes are shown as small black dots (more significant activity is shown in red;
voltage range is shown immediately below each topographic map). (A black and white
version of this figure will appear in some formats. For the color version, please refer to the
plate section.)

and then recorded in a data acquisition computer. Amplifiers are shown

on the small table next to the right arm of the chair in the figure. After the
data are acquired, the voltage responses at each electrode in a time range
of interest (e.g., from –100 to 2000 milliseconds after stimulus onset) are
averaged over all the trials for each event type. The ERP response for
each event is this average voltage response as a function of time.
Although ERPs are typically averaged based on stimulus onset, it is
notable that they can also be averaged based on response onset. ERPs
directly measure neural activity and thus have a temporal resolution in
milliseconds. However, the signal from the brain is blurred in space (due
to the brain, the skull, the cerebral spinal fluid, and the scalp), and activity
from multiple active cortical regions can interact (e.g., the electric fields
32 The Tools of Cognitive Neuroscience

from two opposing active cortical surfaces can cancel each other out).
Owing to such limitations, this technique has a spatial resolution in
centimeters, which is much lower than fMRI. ERP source localization
can be used with the aim of increasing the spatial resolution of this
technique (Slotnick, 2004a). This requires creating a mathematical
model of cortical activity (e.g., a dipole source), the head (e.g., a four-
shell ellipsoid corresponding to the brain, the skull, the cerebral spinal
fluid, and the scalp), and electrode locations. Then, the location and
orientation of cortical activity is adjusted to minimize the difference
between the model ERP activity and the measured ERP activity at all
the electrodes. There are two major limitations to ERP source localiza-
tion. First, there are an infinite number of dipole sources that can give rise
to the same pattern of ERP activity, which is referred to as the inverse
problem. Second, head models are relatively poor. As such, the spatial
resolution of ERP source localization is about 1 centimeter at best. ERP
source localization is not usually employed by cognitive neuroscientists
as it is currently questionable whether it offers a sufficient increase in
spatial resolution.
One ERP memory study illustrates the high temporal resolution and
limited spatial resolution of this technique (Johnson, Minton & Rugg,
2008). During the study phase, words (i.e., names of objects) were
superimposed on a scene or a gray background. For each of these two
conditions, respectively, participants were instructed to either imagine
the corresponding object in the scene or generate a sentence that
incorporated the word. During the test phase, old words from the
study phase and new words were presented and participants made
“old-remember,” “old-know,” or “new” judgments. Figure 2.2B shows
the ERP topographic map (i.e., the magnitude of activity across the scalp)
for old words in the scene condition that were “remembered” (Rscene)
versus new words that were classified as “new” for different time periods.
These topographic maps show a shift from left parietal activity within 500
to 800 milliseconds after stimulus onset to right frontal activity within
1400 to 1900 milliseconds after stimulus onset. These results illustrate the
excellent temporal resolution and limited spatial resolution of ERPs.
First, the magnitude of ERP activity can be observed changing rapidly
over time. If this had been an fMRI study, activity in the parietal cortex
and frontal cortex would be identified, but there would be no information
with regard to the time period in which each these regions were
active. Second, the spatial distribution of ERP activity is relatively
large. Only the general region of the brain that gave rise to ERP activity
can be identified.
2.3 High Temporal Resolution Techniques 33

Figure 2.3 MEG setup. The MEG system (to the left) is housed in a room that is shielded
from electromagnetic waves that could interfere with the signal.

The other methods with high temporal resolution are intimately

related to ERPs. Electroencephalography (EEG) uses the identical
data acquisition methodology as ERPs, but refers to any measure
of brain activity that corresponds to electric fields. This includes ERPs,
but more commonly refers to brain activity that oscillates within a specific
range of frequencies. EEG frequency analysis is a powerful alternative to
the more commonly employed ERP analysis (see Chapter 4). Related to
EEG, magnetoencephalography (MEG) refers to any measure of brain
activity that corresponds to magnetic fields, and also typically refers to
brain activity that oscillates within a specific frequency range. MEG is
measured using superconducting coils that are placed over the scalp.
Figure 2.3 illustrates an MEG machine. Such a large machine is required
because the coils need to be cooled to near absolute zero to maintain
their superconducting properties, which makes this technique many
times more expensive to purchase and maintain than EEG. Like ERPs
that are generated by averaging all the events of a given type from EEG
data during a cognitive task, event-related fields (ERFs) are generated by
averaging all the events of a given type from MEG data. The more
general terms EEG and MEG also refer to ERPs and ERFs. However,
in the field of cognitive neuroscience, the terms ERPs and MEG are more
common and thus will be used for the remainder of this chapter to refer to
ERPs/EEG and MEG/ERFs.
34 The Tools of Cognitive Neuroscience

Box 2.2: Tracking the temporal dynamics of brain function

fMRI is by far the most popular method in the field of cognitive neuroscience
(see Chapter 11). However, brain activity is not a static set of blobs that
represent a cognitive process. Rather, brain activity changes across different
regions in milliseconds. Only techniques with excellent temporal resolution,
such as ERPs, can track the functioning brain. This book highlights the
temporal dimension of brain processing in addition to the spatial dimension
of brain processing. One major advantage of temporal information is that
one can use it to assess whether different brain regions are synchronously
active, which indicates that these regions interact (see Chapters 4 and 6).
This reflects how the brain is actually operating (see Chapter 11).

It is generally believed that MEG has better spatial resolution than

ERPs due to less distortion of the magnetic field from anatomic
structures. These techniques are not identical in terms of their
sensitivity to cognitive and neural processes, but they actually have
similar spatial resolutions (Cohen & Cuffin, 1991; Malmivuo, 2012).
Although the electric field produces ERP activity and the magnetic
field produces MEG activity, the electric field and magnetic field
generated from the same brain region are intrinsically and mathema-
tically linked (Maxwell, 1865; Einstein, 1905). Although there is
a growing body of research in which MEG is employed, MEG and
ERPs have similar spatial resolutions and MEG is much more costly
such that ERPs are more widely used than MEG in the field of
cognitive neuroscience.
ERPs and MEG can measure brain activity only near the scalp and the
activity is blurred or can even be undetectable, which limits the spatial
resolution of these techniques. Critically, as discussed in Box 2.2, only
methods such as ERPs and MEG that directly reflect neural activity have
sufficiently high temporal resolution to track the rapid temporal
dynamics of the functioning brain.

2.4 High Spatial and Temporal Resolution Techniques

The techniques discussed thus far have either excellent spatial resolution
and poor temporal resolution (i.e., fMRI) or excellent temporal resolu-
tion and limited spatial resolution (i.e., ERPs and MEG). One way to get
excellent spatial resolution and excellent temporal resolution is to
2.4 High Spatial and Temporal Resolution Techniques 35

combine methods, such as fMRI and ERPs. Unfortunately, this is rarely

done. One reason is that it takes years to become proficient using even
a single cognitive neuroscience method such that very few laboratories
can employ multiple techniques. Another reason is that the large major-
ity of research in cognitive neuroscience is conducted with fMRI.
There are many reasons for this focus with fMRI that are discussed in
Chapter 11, but this has resulted in relatively few cognitive neuroscience
laboratories using methods with high temporal resolution and even fewer
laboratories that use combined methods with high spatial resolution and
high temporal resolution.
Depth electrode recording is similar to ERP recording, but the elec-
trode is inserted directly into a specific brain region. This technique has
excellent spatial resolution (in the sub-millimeter range) and excellent
temporal resolution (in the millisecond range). However, this technique
is only used in humans under rare circumstances, such as in patients who
have electrodes implanted for clinical reasons. Depth electrode recording
is also referred to as single-cell recording, although this is a misnomer
because each electrode records activity from multiple nearby cells/neu-
rons (in contrast to single-cell recording in non-human animals, where
activity is actually recorded from individual cells; see Chapter 10).
Patients with epilepsy sometimes have electrodes implanted in their
brain in an effort to determine the precise region that gives rise to their
seizures. If such a region is identified, it can be surgically removed in an
effort to treat their condition (see Chapter 9). After the patients have
electrodes implanted in their brain, while the signal is being monitored to
determine the location that gives rise to the seizures, they sometimes
volunteer to participate in memory experiments. Depth electrode record-
ing can be used to measure lower frequency activity that oscillates slower,
which can be used to produce ERPs, or higher frequency activity that
oscillates faster, which reflects spiking/firing of nearby neurons
(Logothetis, Pauls, Augath, Trinath & Oeltermann, 2001). As discussed
in Box 2.3, there is a direct correlation between neural activity, electro-
physiological activity (i.e., electrical activity generated by neuronal firing
that can be measured with ERPs), and fMRI activity.
One memory study recorded from depth electrodes implanted in the
hippocampus and other medial temporal lobe regions of patients that
were being evaluated for epilepsy surgery (Suthana et al., 2015).
Figure 2.4A shows the location of a depth electrode in the left hippo-
campus of one participant. During the study phase, participants were
instructed to learn specific photographs (i.e., targets). During the test
phase, old items/targets, similar items/lures, and new items/foils were
36 The Tools of Cognitive Neuroscience

Box 2.3: Neural activity, electrophysiological activity,

and fMRI activity are correlated
As described earlier in this chapter, fMRI measures an increase in blood flow.
Many skeptics initially questioned whether fMRI activity reflected neural
activity. To answer this question, Logothetis et al. (2001) simultaneously
used depth electrode recording and fMRI to measure V1 activity in monkeys
during visual stimulation. Stimulus duration and contrast were experimen-
tally manipulated and it was found that neuronal activity, electrophysiologi-
cal activity, and fMRI activity were highly correlated (although the fMRI
activity had a delayed onset and extended timecourse). These results indicate
that neuronal spiking in an active brain region produces electrophyiological
activity, which can be measured with ERPs, and increases blood flow to that
region, which can be measured with fMRI. Long-term memory is mediated
by an increase in the neuronal spike rate and amplitude in the hippocampus
(see Chapter 10). This increase in neural activity can be assumed to produce
the increase in hippocampal fMRI activity that is typically observed during
long-term memory (see Chapter 3).

B Target Lure 1 Lure 2 Foil

Firing Rate

42 42 42 42
(Hz)

0 1000 0 1000 0 1000 0 1000

Time (msec)

Figure 2.4 Hippocampal depth electrode placement and results. (A) Depth electrode (black
circle) in the left hippocampus (partial coronal view). (B) For each item type during the test
phase (labeled), hippocampal neuronal spiking (immediately below each label; the line
shows stimulus onset) and firing rate (in Hertz, spikes per second; at the bottom), as
a function of time after stimulus onset (in milliseconds), are shown for a representative
participant.
2.5 Lesions and Temporary Cortical Disruption Techniques 37

presented, and participants responded as to whether or not they had seen

that exact item. Figure 2.4B, from top to bottom, illustrates the different
item types at test, the corresponding hippocampal neuron spike rate, and
the corresponding firing rate. The hippocampal neurons were more
active for targets than for lures or foils, which indicates that this region
is important for remembering specific target faces.
Although there are methods that provide excellent spatial resolution
and excellent temporal resolution, these techniques are currently rarely
used in cognitive neuroscience. The frequency to which depth electrode
recording is employed is limited by clinical utility and invasiveness.
However, combining techniques that are well established, such as fMRI
and ERPs, is feasible for those who aspire to understand the spatial-
temporal brain mechanisms underlying memory (see Chapter 11).

2.5 Lesions and Temporary Cortical Disruption Techniques

The methods discussed thus far can be used to identify brain activity that
is correlated with a particular cognitive process. That is, during the same
time period in which a particular cognitive event occurs, there is an
increase in activity within a certain brain region that is measured using
the preceding methods. It is reasonable to assume that such activity
reflects the brain regions underlying that cognitive process. However, it
is possible that a brain region has nothing to do with that event and is only
co-activated because it has a strong connection with another brain region
underlying the event. This illustrates the problem with correlational
methods such as fMRI, ERPs, and MEG. It is always uncertain whether
brain activity identified using these techniques is necessary for a cognitive
process or is epiphenomenal, like the heat coming off a light bulb.
Patients with naturally occurring brain lesions due to a stroke or
another type of brain trauma can be evaluated to assess whether
a region is necessary for a cognitive process. If a specific region is lesioned
and performance on a cognitive task is selectively impaired, it can be
assumed that the region underlies performance on that task. There are
two major limitations to lesion studies. First, there must be access to
a sufficient number of willing patients with lesions in the brain region of
interest. Second, naturally occurring lesions are almost never restricted
to the single region of interest; therefore, it is not certain which of the
disrupted regions impaired performance on a task.
One study investigated whether patients with hippocampal lesions had
impaired long-term memory (Manns, Hopkins, Reed, Kitchener &
Squire, 2003). Figure 2.5A shows the brain of a control participant
38 The Tools of Cognitive Neuroscience

A B C 1.5
CON1 JS 1-WK CON (n = 7)

Discriminability Score (d`)

H (n = 7)

1.0

0.5

0.0
Remember Know

Figure 2.5 Hippocampal lesion and recognition memory results. (A) The intact hippocampus
of a control participant (CON1) is shown in light gray, as indicated by each arrowhead
(just below the downward slanted black lines; coronal view). (B) The lesioned hippocampus
of one patient (JS) is shown in darker gray as indicated by the arrows (coronal view).
(C) Recognition memory performance (discriminability, d’, between old and new items) for
“remember” and “know” responses in control participants that had a 1-week delay between
the study phase and the test phase (1-WK CON) and in patients with lesions to the
hippocampus (H; key at the top right).

(CON1) with a healthy hippocampus (the light gray area indicated by the
white arrowheads, just below the downward slanted black lines). Six of
the seven patients in the study had hippocampal lesions due to some sort
of loss of oxygen/anoxia (e.g., carbon monoxide poisoning or a heart
attack). Figure 2.5B shows the brain of one patient (JS) with focal lesions
to the hippocampus (indicated by the white arrows), as indicated by their
hippocampus being darker and less uniform in color than the control
participant. Across the patients, the reduction in size of the hippocampus
ranged from 10 to 45 percent, while the size of the surrounding para-
hippocampal gyrus was within normal and ranged from –15 to 15 percent.
One set of experiments compared the “remember” responses and the
“know” responses for control participants and patients. During the study
phase of each experiment, pictures (i.e., faces or abstract line drawings)
or words were presented. During each of the corresponding test
phases, old items and new items were presented and participants
made “old”–“new” recognition judgments and for “old” items made
“remember”–“know” judgments (see Chapter 1). To ensure the control
participants had similar memory performance as the patients, such that
the results would not be confounded by differences in overall memory
performance, the patients were tested immediately after the study phase
and control participants were tested 1 week after the study phase.
As illustrated in Figure 2.5C, memory performance (i.e., d’, which
measures the discriminability between old and new items) for patients
and control participants was similar for items associated with
2.5 Lesions and Temporary Cortical Disruption Techniques 39

“remember” responses and for items associated with “know”

responses. These results suggest that the hippocampus is involved to
a similar degree in both recollection and familiarity, which are two
forms of long-term memory (see Chapter 1). This is a highly debated
topic, as the majority view in the field is that the hippocampus is
preferentially associated with recollection (see Chapter 3). The major
problem with this study is that anoxia is known to cause global cortical
atrophy/shrinking rather than lesions restricted to the hippocampus
(Grubb et al., 2000). Such global atrophy can be seen by comparing
Figure 2.4A and 2.4B, as patient JS has much more dark space
between the brain and the skull. This means that the impaired memory
performance in the patients could have been due to lesions in other
regions of the brain that are associated with long-term memory, such as
the dorsolateral prefrontal cortex and/or the parietal cortex (see
Chapter 3). This illustrates that lesions are almost never restricted to
one region, which is the major limitation of these studies.
Other methods can be used to temporarily disrupt processing in one
region of the brain. The most commonly used technique for this purpose
is transcranial magnetic stimulation (TMS). Figure 2.6A shows a TMS
system. The TMS machine is about the size of two computer towers (the
laptop in the lower right of the figure provides a frame of reference), with
a small screen to control the stimulation parameters (shown at the top
with a blue screen). The TMS participant sits in a chair while
a stimulation coil is held against their scalp manually and/or with
a mechanical arm over a target brain region. During each TMS pulse,
current is passed through the coil, which induces a magnetic field about
the same strength as an MRI machine that disrupts cortical processing
immediately below the coil. One way to identify a target location is to use
landmarks on the head, such as a standard ERP electrode location (see
Chapter 4). However, this option is inherently inaccurate because such
landmarks do not correspond that well to specific brain regions. A much
more accurate method is to use MRI, which can be used to target
a specific anatomic region, or fMRI, which can be used to target
a specific brain activation associated with a cognitive process of interest.
It should be noted that even if these more precise methods of targeting
are employed, the magnetic field generated by the TMS coil disrupts
cortex surrounding the target location to some degree, which limits the
spatial resolution of TMS to about a centimeter. Another limitation of
TMS is that it can disrupt brain regions only near the surface of the scalp.
For example, TMS could be used to target a region of the left dorsolateral
prefrontal cortex but could not target the hippocampus as this region is
40 The Tools of Cognitive Neuroscience

A B

C D 1.0
0.9
Hit rate (moving-stationary judgment)

ns
0.8
0.7
0.6
Target
0.5
0.4
0.3
0.2
0.1
0
No TMS MT TMS No TMS MT TMS
Moving items Stationary items

Figure 2.6 TMS setup and fMRI guided TMS results. (A) TMS system that includes
a stimulation coil (at the top left). (B) TMS coil positioned over motion processing region
MT of a participant. (C) fMRI activity associated with motion perception (in red/yellow) for
one participant (partial lateral view, occipital pole to the left). The bottom half of the head
is shown in a triangular mesh (in brown). The TMS coil is shown by wireframe wheels and
the target point (red sphere) is located within MT, the motion processing region of the
brain. This image is a screenshot of the fMRI guided TMS neuronavigation software that
was used to target MT in real time, with the head and coil identical to the positioning
shown in (B) but zoomed in closer to the coil. (D) TMS results showing a reduced hit rate
(the probability of responding “moving” to previously moving items or “stationary” to
previously stationary items) for moving items following TMS to MT, as compared to no TMS
(the asterisk indicates a significant difference, ns = not significantly different). (A black and
white version of this figure will appear in some formats. For the color version, please refer
to the plate section.)
2.5 Lesions and Temporary Cortical Disruption Techniques 41

too deep inside the brain. One of the most common TMS protocols
stimulates a target region at 1 pulse per second (Hertz) for 10 minutes.
Immediately after the stimulation period, processing in that region is
disrupted for about 8 minutes. During the disruption period, participants
perform a cognitive task. If behavioral performance is impaired on that
task, it can be assumed that the disrupted cortical region is involved in
that cognitive process. The 1 Hertz TMS protocol was illustrated in one
study that targeted motion processing region MT (see Chapter 1) to
assess whether disruption of this region impaired memory for motion
(Slotnick & Thakral, 2011). Figure 2.6B shows the TMS coil held in place
by a mechanical arm and positioned over region MT within the right
hemisphere of one participant. The positioning of the coil was guided by
fMRI activity from the same participant. As shown in Figure 2.6C, for
each participant, the target point (red sphere) within each hemisphere
was identified as fMRI activity associated with perceiving motion (in red/
yellow) within the posterior bank of the ascending limb of the inferior
temporal sulcus, which is the known location of region MT. During the
study phase (without TMS), participants viewed abstract shapes that
were either moving or stationary in the left visual field or the right visual
field. Immediately after the study phase, 1 Hertz TMS was applied to
region MT within one hemisphere for 10 minutes. During TMS, the coil
position (shown as wireframe wheels in Figure 2.6C) was manually
adjusted in real time such that the magnetic field was always focused on
the target point. During the test phase, with activity in region MT dis-
rupted, the shapes from the study phase were presented and participants
classified each item as previously “moving” or “stationary.” Memory
accuracy was measured by the hit rate, the probability of classifying
a previously moving item as “moving” or a stationary item as “station-
ary.” As shown in Figure 2.6D, TMS to region MT impaired memory
accuracy for moving items but did not affect memory accuracy for
stationary items. These TMS results show that region MT is necessary
for remembering motion.
There are other TMS protocols in which processing in the target region
is enhanced/activated rather than disrupted/deactivated (see
Chapter 11), but these are less commonly used. There are also protocols
in which processing is disrupted by stimulating at a specific time point
after stimulus onset, which can increase the temporal resolution of this
technique, but these protocols are not commonly used. Thus, TMS typi-
cally has limited/good spatial resolution, depending on the method of
targeting, and poor temporal resolution.
42 The Tools of Cognitive Neuroscience

Figure 2.7 tDCS setup. The tDCS device is shown on the table (to the left). The electrodes are
positioned on the scalp to stimulate the left dorsolateral prefrontal cortex.

The last technique that will be mentioned is transcranial direct

current stimulation (tDCS). tDCS is similar to TMS in that it tempora-
rily modulates processing in a targeted cortical region but stimulates
with a weak direct current rather than a magnetic field. This current is
subthreshold as it does not induce neural firing but rather alters the
resting potential of the neural membrane (Purpura & McMurtry,
1965). During conventional tDCS, current flows through a target elec-
trode placed on the scalp over the brain region of interest and a return
electrode is placed elsewhere. Figure 2.7 shows a tDCS setup used to
stimulate the left dorsolateral prefrontal cortex, with the return elec-
trode placed over the right eye. Each tDCS electrode, which consists of
a smaller electrode within a saline-soaked sponge, is typically quite
large (e.g., 5 centimeters by 7 centimeters). As a result of the placement
of the electrodes and their large size, the spatial resolution of tDCS is
poor (i.e., worse than TMS). Like the terminals of a battery, the target
tDCS electrode can be either a cathode (where current flows out) or an
anode (where current flows in). Cathodal stimulation typically reduces
cortical activity and anodal stimulation typically enhances cortical
activity (Shin, Foerster & Nitsche, 2015). Stimulation time usually
varies between 15 to 40 minutes, which produces excitatory or
2.6 Method Comparisons 43

inhibitory effects that last beyond the stimulation period. The effects of
tDCS are relatively small such that the studies often need to recruit
a greater number of participants than other methods. Although tDCS
has low spatial resolution and low temporal resolution, it has proven to
be an effective way to assess whether a region is necessary for
a cognitive process of interest. Moreover, a tDCS system costs only
a few thousand dollars, which is an order of magnitude less than a TMS
system, making this technique incredibly cost effective. A relatively
new method called transcranial alternating current stimulation (tACS)
uses the identical setup as tDCS, but the current alternates at a specific
frequency; thus, tACS can stimulate the brain at a desired frequency
(Herrmann, Rach, Neuling & Strüber, 2013).

2.6 Method Comparisons

Figure 2.8 directly compares the techniques discussed in this chapter
based on spatial resolution and temporal resolution. Depth electrode
recording has excellent spatial resolution and excellent temporal
resolution; however, this method can only rarely be used. fMRI has

Figure 2.8 Spatial resolution and temporal resolution for different methods.
44 The Tools of Cognitive Neuroscience

excellent spatial resolution and poor temporal resolution, while ERPs

and MEG have excellent temporal resolution and limited spatial
resolution. TMS, tDCS, and lesions have limited/poor spatial resolu-
tion and poor temporal resolution but can assess whether a region is
necessary for a cognitive process. Lesion evidence has an additional
limitation that lesions are not restricted to one brain region. When
used in isolation, none of the commonly used techniques have
excellent spatial resolution and excellent temporal resolution.
Some cognitive neuroscientists have combined methods to improve
spatial and/or temporal resolution. For instance, fMRI guided TMS has
been used to increase the spatial resolution of TMS, and combining fMRI
and ERPs produces results with excellent spatial resolution and excellent
temporal resolution. By combining techniques with excellent spatial
resolution and excellent temporal resolution, the spatial-temporal
dynamics of brain activity can be measured, which is the future of cogni-
tive neuroscience (see Chapter 11).

Chapter Summary
• fMRI measures blood flow that increases in active brain regions and
has excellent spatial resolution and poor temporal resolution.
• ERPs measure voltages produced by brain activity using scalp electro-
des and have excellent temporal resolution and limited spatial
resolution.
• MEG is similar to ERPs but measures magnetic fields produced by
brain activity using superconducting coils and is much more expensive
than ERPs.
• Depth electrode recording in patients has excellent spatial and tem-
poral resolution but is only rarely done.
• Lesion evidence can assess whether a region is necessary for
a particular cognitive process, has poor spatial and poor temporal
resolution, and is questionable because lesions are not restricted to
one brain region.
• TMS can be used to temporarily disrupt one cortical region using
a magnetic field to assess whether that region is necessary for
a particular cognitive process and typically has limited spatial resolu-
tion and poor temporal resolution.
• tDCS is similar to TMS but uses a weak stimulating current, has poor
spatial resolution, and is much less expensive than TMS.
• Combining techniques such as fMRI and ERPs offers excellent spatial
resolution and excellent temporal resolution.
Further Reading 45

Review Questions
Does fMRI measure blood flow or neural activity?
Do ERPs have excellent spatial resolution, excellent temporal resolution,
or excellent spatial resolution and excellent temporal resolution?
What is one problem with patient lesion evidence?
How does TMS work?
Which two techniques could be combined to produce results with
excellent spatial resolution and excellent temporal resolution?

Further Reading
Buckner, R. L., Bandettini, P. A., O’Craven, K. M., Savoy, R. L.,
Petersen, S. E., Raichle, M. E. & Rosen, B. R. (1996). Detection of
cortical activation during averaged single trials of a cognitive task using
functional magnetic resonance imaging. Proceedings of the National
Academy of Sciences of the United States of America, 93, 14878–14883.
This paper introduced event-related fMRI and illustrates the excellent
spatial resolution and poor temporal resolution of this method.
Johnson, J. D., Minton, B. R. & Rugg, M. D. (2008). Content dependence of the
electrophysiological correlates of recollection. NeuroImage, 39, 406–416.
This paper demonstrates the excellent temporal resolution and limited
spatial resolution of ERPs.
Suthana, N. A., Parikshak, N. N., Ekstrom, A. D., Ison, M. J., Knowlton, B. J.,
Bookheimer, S. Y. & Fried, I. (2015). Specific responses of human
hippocampal neurons are associated with better memory. Proceedings
of the National Academy of Sciences of the United States of America, 112,
10503–10508.
This paper illustrates depth electrode recording, which has excellent spatial
resolution and excellent temporal resolution but is rarely done.
Slotnick, S. D. & Thakral, P. P. (2011). Memory for motion and spatial location
is mediated by contralateral and ipsilateral motion processing cortex.
NeuroImage, 55, 794–800.
This paper uses fMRI guided TMS to assess whether motion processing
region MT is necessary for remembering motion.
CHAPTER THREE

Brain Regions Associated with Long-Term Memory

Learning Objectives
• To identify three brain regions most commonly associated with episodic
memory.
• To compare the brain regions associated with episodic memory and
semantic memory.
• To contrast the two models of long-term memory consolidation.
• To explain what happens during slow wave sleep that promotes
long-term memory consolidation.
• To compare the brain regions associated with memory retrieval and
memory encoding.
• To describe how behavioral performance and hippocampal activity differ
between females and males during long-term memory.
• To explain one way in which the brains of those with superior memory
differ from those with normal memory.

This chapter considers the brain regions associated with long-term

memory, a type of explicit memory (see Chapter 1). Long-term memory
can be broken down into episodic memory and semantic memory.
Episodic memory refers to the detailed retrieval of a previous episode,
such as when someone remembers a happy moment of his or her life.
Semantic memory refers to the retrieval of factual information, such as
the definition of a word or the name of the current president. Semantic
memories are formed through repeated exposure to information
throughout life and lack the details associated with episodic memories.
This information is simply known and there is no memory for the pre-
vious details of the learning experience. Although episodic memory and
semantic memory both refer to conscious forms of retrieval, the degree of
detail and subjective experience associated with these types of memory is
quite different. It follows that the brain regions associated with episodic
memory and semantic memory are also distinct. The first two sections of
the chapter (sections 3.1 and 3.2) consider the brain regions associated
with episodic memory and semantic memory. Section 3.3 will consider
long-term memory consolidation (i.e., the process of creating more
3.1 Episodic Memory 47

permanent memory representations in the brain). In section 3.4, the role

of sleep in long-term memory consolidation is examined. Long-term
memory consolidation requires the interaction between multiple brain
regions in which activity oscillates at specific frequencies. In section 3.5,
the brain regions associated with memory encoding will be reviewed.
Section 3.6 details differences in behavioral performance and brain
activity between females and males (i.e., sex differences) during long-
term memory. In the last section, 3.7, the brains of those with superior
memory are evaluated, including London taxi drivers and those who
compete in World Memory Championships. Although the research on
this topic is sparse, there is convergent evidence that having a superior
memory does not come without a cost.

3.1 Episodic Memory

The term episodic memory can refer to many other related
forms of memory including context memory, source memory,
“remembering,” recollection, and autobiographical memory (see
Chapter 1). Autobiographical memory refers to a specific type of
episodic memory for detailed personal events. Context memory
and source memory refer to accurate retrieval of contextual informa-
tion, such as memory for which side of the street the car is parked,
and “remembering” refers to the subjective experience during
detailed retrieval, such as the visual experience of imagining where
the car is parked. The same brain regions have been associated with
all of these flavors of episodic memory.
Episodic memories are associated with activity in both control regions
and sensory regions of the brain (see Chapter 1). Sensory cortical activity
reflects the contents of memory. For example, visual and auditory pro-
cessing regions of the brain are associated with memory for objects and
sounds, respectively. As the interpretation of activity in sensory regions is
straightforward and not unique to episodic memory, sensory activity will
not be focused on in this chapter. The control regions that mediate
episodic memory include the medial temporal lobe, the dorsolateral
prefrontal cortex, and the parietal cortex (Chapter 1; Wagner,
Shannon, Kahn & Buckner, 2005; Rugg & Vilberg, 2013). Figure 3.1
illustrates these regions, which include the hippocampus and the para-
hippocampal cortex within the medial temporal lobe, the dorsolateral
prefrontal cortex (the unlabeled leftmost activation), and the inferior
parietal cortex (angular gyrus) and the medial parietal cortex (extending
into retrosplenial cortex and posterior cingulate cortex). There are many
48 Brain Regions Associated with Long-Term Memory

Figure 3.1 Regions of the brain associated with episodic memory. fMRI activity (in red/yellow)
in the left hemisphere (left, lateral view; right, medial view; occipital poles toward the center).
(A black and white version of this figure will appear in some formats. For the color version,
please refer to the plate section.)

regions, but keep in mind that the primary regions associated with
episodic memory are the medial temporal lobe, the dorsolateral prefron-
tal cortex, and the parietal cortex.
The roles of the hippocampus and the parahippocampal cortex within
the medial temporal lobe are relatively well understood (Diana,
Yonelinas & Ranganath, 2007). The parahippocampal cortex processes
the context of previously presented information. This could be the spatial
location (e.g., the left or right side of the screen) or the color (e.g., red or
green) of an object that was presented earlier. There is also fMRI evi-
dence that the parahippocampal cortex processes memory for temporal
information, such as the temporal order in which a previous event
unfolded. During the study phase of one experiment, participants learned
the locations of eight different stores and the order they visited the stores
in a virtual environment (Ekstrom, Copara, Isham, Wang & Yonelinas,
2011). During each trial of the test phase, participants were presented
with one target store and were asked which of two other stores was
nearest in distance, which was based on spatial memory, or were asked
which of two other stores was closest in delivery order, which was based
on temporal memory. Both accurate spatial memory and accurate tem-
poral memory activated the parahippocampal cortex. Another medial
temporal lobe region referred to as the perirhinal cortex, which is directly
anterior to the parahippocampal cortex, needs to be mentioned as well
because the perirhinal cortex processes item information (e.g., whether
or not an object has been seen before) and episodes are composed of
individual items. The perirhinal cortex is not usually associated with
detailed episodic memory because item memory can be based on non-
detailed familiarity. For instance, a person might recognize someone they
saw for the first time at a party the week before, because the person looks
3.1 Episodic Memory 49

PRC HC

Item
Cortical input

Binding
PHC

Context

Figure 3.2 Model of medial temporal lobe sub-region function. The perirhinal cortex (PRC)
processes item information, the parahippocampal cortex (PHC) processes context
information, and the hippocampus (HC) binds item information and context information.

familiar, but the previous context may not be remembered (i.e., the
party). The hippocampus is thought to bind item information and context
information together during episodic memory. That is, a memory for
a specific episode contains multiple items, which are processed in the
perirhinal cortex, in a particular context, which is processed in the para-
hippocampal cortex. The hippocampus binds this item information and
context information to create a detailed episodic memory (Slotnick,
2013b). For example, if an individual went on a vacation to Newport
Beach in California and later recalled meeting a friend on the beach, that
individual’s perirhinal cortex would process item information (e.g., the
friend), the parahippocampal cortex would process context information
(e.g., the area of the beach on which they were standing), and the
hippocampus would bind this item information and context information
into a unified memory. Figure 3.2 illustrates this model of medial tem-
poral lobe function. Box 3.1 describes an alternative model of medial
temporal lobe function (Squire, Wixted & Clark, 2007).
The roles of the dorsolateral prefrontal cortex and the parietal
cortex during episodic memory are not well understood. fMRI evi-
dence and lesion evidence indicates that the dorsolateral prefrontal
cortex is important for episodic memory (Mitchell & Johnson, 2009).
There has been speculation that the dorsolateral prefrontal cortex
mediates post-retrieval monitoring (i.e., evaluating the contents of
50 Brain Regions Associated with Long-Term Memory

Box 3.1: Is the hippocampus preferentially associated with

context memory?
There is a large body of evidence indicating that the hippocampus is asso-
ciated with context memory (i.e., recollection) to a greater degree than item
memory (i.e., familiarity; Diana et al., 2007), which supports the standard
model (i.e., the majority view) that this region is important for binding item
information and context information. There is an alternative model (i.e., the
minority view) that the hippocampus is associated with context memory and
item memory to a similar degree (Squire et al., 2007). Proponents of the
alternative model discount the studies reporting greater context memory
than item memory activity in the hippocampus by arguing that differences in
hippocampal activity are due to higher memory strength during context
memory than during item memory (rather than the hippocampus being
preferentially associated with context memory). However, regardless of
whether context memory strength is greater than item memory strength or
item memory strength is greater than context memory strength, the hippo-
campus is associated with context memory (Slotnick, 2013b). Although there
is a growing body of evidence that the hippocampus is preferentially asso-
ciated with context memory, this is still a highly debated topic. A possible
resolution of these opposing viewpoints is that the hippocampus may bind
item information and context information during context memory, and this
region may also bind individual features that comprise an item during item
memory (Slotnick, 2010a).

memory), due to the relatively slow timecourse of activity in this

region (see Chapters 2 and 4), or selecting information that is stored
in other regions (see Chapter 8). The parietal cortex has been hypothe-
sized to mediate multiple functions during episodic memory such as
the accumulation of mnemonic information (i.e., storing the contents
of memory) or attention directed to internal memory representations
(Wagner et al., 2005; see Chapter 8). There is also lesion evidence and
fMRI evidence that suggests the inferior parietal cortex is associated
with accurate autobiographical memory and “remembering” but not
source memory (Cabeza, Ciaramelli, Olson & Moscovitch, 2008).
Determining the functions of the dorsolateral prefrontal cortex and
the parietal cortex during episodic memory is a major topic of
investigation.
It should also be noted that the regions associated with episodic
memory – the dorsolateral prefrontal cortex, the parietal cortex, and
3.2 Semantic Memory 51

the medial temporal lobe – have also been associated with item memory
(Eldridge, Knowlton, Furmanski, Bookheimer & Engel, 2000; Wheeler
& Buckner, 2004; Slotnick & Schacter, 2007). Therefore, these regions
are more generally associated with long-term memory (but not semantic
memory; see Chapter 1 and section 3.2).

3.2 Semantic Memory

Semantic memory refers to knowledge of facts that are learned
through repeated exposure over a long period of time, typically
years. Semantic memory is a type of long-term memory, as retrieval
of such information is conscious, but no details of the learning episodes
are retrieved. Subjectively, semantic memory is associated with
“knowing.” There is an important distinction between semantic
memory, which typically refers to learning through repeated exposure
over years and will be the topic of this section, as compared to item
memory, which typically refers to very recent learning based on one or
a few repetitions (see Chapter 1). It should also be pointed out that
semantic memory refers to definitions and conceptual knowledge, and
thus this cognitive process links the field of memory and the field of
language (see Chapter 8).
Semantic memory has been associated with the left dorsolateral
prefrontal cortex (in a different region associated with episodic
memory), the anterior temporal lobes, and sensory cortical regions
(Gabrieli, Poldrack & Desmond, 1998; Martin & Chao, 2001).
The left dorsolateral prefrontal cortex activity associated with seman-
tic memory may reflect language processing, to some degree (see
Chapter 8). Alternatively, activity in the left dorsolateral prefrontal
cortex may reflect the process of selecting a semantic memory that is
stored in other cortical regions. Semantic memory also activates the
corresponding sensory cortical regions. For example, naming animals
activates more lateral inferior occipital-temporal cortex that has
been associated with perception of living things, while naming tools
activates more medial inferior occipital-temporal cortex that has
been associated with perception of nonliving things (Martin &
Chao, 2001). The region most consistently associated with semantic
memory is the anterior temporal cortex. As shown in Figure 3.3, in
one study of Alzheimer’s patients, impairment in an object naming
task, which depends on intact semantic memory, was most highly
correlated with cortical thinning in the left anterior temporal lobe
(Domoto-Reilly, Sapolsky, Brickhouse & Dickerson, 2012). This
52 Brain Regions Associated with Long-Term Memory

Figure 3.3 Regions of the brain associated with semantic memory. Cortical thinning in
Alzheimer’s patients (in red/yellow) associated with disruption in semantic memory (lateral
views, occipital poles toward the center). (A black and white version of this figure will appear
in some formats. For the color version, please refer to the plate section.)

finding suggests that the left anterior temporal lobe is necessary for
semantic memory.
The role of the anterior temporal lobe during sematic memory
is a current topic of research. One hypothesis is that the anterior temporal
lobe stores semantic information. Alternatively, the anterior temporal
lobe may link information in different cortical regions during semantic
memory in a similar fashion to how the hippocampus links information
during episodic memory. There is some evidence that different regions of
the anterior temporal lobe mediate different types of semantic memory
such as visual semantic memory (e.g., ‘what does a sheep look like?’),
auditory semantic memory (e.g., ‘what does a sheep sound like?’), or
semantic memory for social information (e.g., ‘are sheep friendly?’;
Skipper, Ross & Olson, 2011). In one fMRI study, participants learned
facts about people, buildings, or hammers (Simmons, Reddish, Bellgowan
& Martin, 2010). For example, they learned ‘the brooks hammer is eight
years old’ and ‘patrick was born in little rock’ through repeated exposures
during fMRI. Learning facts about people activated the left and right
anterior temporal lobe to a greater degree than learning facts about
buildings or hammers, while learning facts about buildings or hammers
did not activate this region to a greater degree than the other event types.
These findings suggest that the anterior temporal lobe may be particularly
important during semantic memory for processing social information.
The degree to which the anterior temporal lobe is specialized for proces-
sing semantic memory for social information or is associated with semantic
memory more generally is a topic of future investigation.
3.3 Memory Consolidation 53

3.3 Memory Consolidation

As detailed above, the hippocampus binds information between differ-
ent cortical regions during long-term memory. However, long-term
memories may only depend on the hippocampus for a limited time.
In the standard model of memory consolidation, a long-term memory
representation changes from being based on hippocampal–cortical
interactions to being based on cortical–cortical interactions, which
takes a period of somewhere between 1 to 10 years (Alvarez &
Squire, 1994). The estimated length of time it takes for consolidation
to occur is based on patients with lesions to the hippocampus. For
example, a person with hippocampal damage due to temporary lack
of oxygen might have impaired long-term memories for approximately
1 year before the time of damage, which is called retrograde amnesia
and have intact long-term memories for earlier events. This suggests
that the hippocampus is involved in long-term memory retrieval for
approximately 1 year, as more remote long-term memories no longer
depend on the hippocampus so they are not disrupted.
Although the standard model of consolidation has been very influen-
tial, there is a convincing body of evidence that indicates patients with
hippocampal damage have disrupted long-term memories, particularly
autobiographical memories, for events that occurred over 30 years before
the time of hippocampal damage (Nadel & Moscovitch, 1997). This
suggests that the hippocampus is involved in long-term memory retrieval
for our entire lifetime. Proponents of this model argue that the reason
standard model proponents have not reported disruption of remote long-
term memories is because the measures used to assess memory were not
sufficiently sensitive (Nadel & Bohbot, 2001). To illustrate, a patient with
hippocampal damage may have intact semantic memory for public events
from about 15 years ago (e.g., they could answer the question, ‘which
country did George W. Bush declare war on based on the conjecture that
there were weapons of mass destruction?’) but may have impaired auto-
biographical memory (e.g., they could not remember visiting Disneyland
15 years ago). If only semantic memories were evaluated, the test would
be less sensitive to measuring memory disruption, because semantic
memories are less susceptible to disruption following damage to the
hippocampus (Nadel & Moscovitch, 1997; Winocur & Moscovitch,
2011). This is consistent with the brain regions most commonly associated
with semantic memory (see section 3.2 of this chapter), which do not
include the hippocampus. In one fMRI study, participants answered
questions about news events that had occurred in the past 30 years,
54 Brain Regions Associated with Long-Term Memory

which assessed semantic memory (Smith & Squire, 2009). The magnitude
of hippocampal activity systematically decreased as a function of news
events that were 3 to 6 to 9 years old and then mostly leveled off, which
suggests the hippocampus is most active during memory for recent
events. This evidence was taken to support the standard model of con-
solidation, where the hippocampus is involved during retrieval of only
more recent memories. There are two problems with this interpretation.
First, only semantic memory was tested. As the hippocampus appears to
be primarily involved in detailed long-term memory retrieval such as
episodic memory (see the previous two sections), the finding that the
hippocampus is not as important for remote semantic memories is not
surprising – the investigators used a test that was not sensitive. Second,
the activity in the hippocampus did not drop to zero for older semantic
memories but was well above baseline even for events that were 30 years
old. This indicates that the hippocampus was involved in memory
retrieval for this entire period (and also suggests this region plays some
role during the semantic memory task employed). If the hippocampus
was no longer involved, the magnitude of activity in this region would
have dropped to zero for remote memories. Thus, even evidence from
proponents of the standard model of consolidation appears to support
the alternative hypothesis that the hippocampus is involved during long-
term memory throughout the lifetime.
Although much of the evidence appears to support the involvement of
the hippocampus during long-term memory for events that occurred
decades earlier, such evidence can be difficult to interpret because of
uncertainty about the lesion location(s) and because the memory tasks
employed may rely on the hippocampus (Knowlton & Fanselow, 1998).
Regarding the first point, lesions are rarely restricted to the hippocampus
(see Chapter 2). Traumatic events that damage the hippocampus, such as
lack of oxygen, often damage other regions such as the dorsolateral
prefrontal cortex. If damage did occur to multiple regions, it is uncertain
which damaged region caused the disruption in memory performance.
Even if the standard model of consolidation is correct, it could be the case
that damage to the dorsolateral prefrontal cortex, which is known to be
important during long-term memory retrieval, is the cause for deficits in
long-term memory. This is why lesion evidence is most convincing when
it is restricted to the one region of interest. One study reported results
from patients with focal lesions restricted to the hippocampus following
transient global amnesia (TGA), which is a temporary loss of long-term
memory (i.e., episodic memory and item memory, but not semantic
memory; see Chapter 1) that is often triggered by a highly emotional or
3.3 Memory Consolidation 55

Overall autobiographical memory score

16

14

12
TGA acute
10
Rated score

TGA follow-up
Controls
8

0
Time yrs yrs 0y
rs yrs mt
hs
17 30 >3 t5 2
period 0– 18
– las <1

Figure 3.4 Autobiographical memory disruption for recent and remote events in patients
with hippocampal lesions. Autobiographical memory score plotted as a function of the time
period before the onset of transient global amnesia (TGA). TGA patient performance (in gray)
and control performance, including performance when the TGA patients no longer had
a lesion (follow-up, in black) and performance of participants with no lesions (in white; key to
the right).

physically arousing event (Bartsch, Döhring, Rohr, Jansen & Deuschl,

2011; see Chapter 9). Figure 3.4 shows that TGA patients had autogra-
phical memory disruption in the acute phase – immediately following the
amnesia – for memories that were both recent (less than 12 months old)
and remote (more than 30 years old), as shown by their memory score
being significantly lower than that of control participants (who did not
have a lesion) and their memory score being significantly lower than their
follow-up performance (after the amnesia had resolved).
Regarding the second point on why consolidation evidence is diffi-
cult to interpret, it is important to make sure the task employed does
not rely on the hippocampus. If the task does engage the hippocampus,
retrieval of both recent and remote memories may only appear to rely
on the hippocampus. For example, in one fMRI study, participants
retrieved recent episodic memories and remote episodic memories
and engaged in detailed memory construction for 12 seconds during
both types of retrieval (Bonnici et al., 2012). Memory construction
refers to mental maintenance of an episodic memory and elaboration
on that memory for an extended period of time. A similar magnitude of
56 Brain Regions Associated with Long-Term Memory

Box 3.2: Do both sides of the consolidation debate

believe they are correct?
Those that support the standard model of consolidation and those that
support the alternative model of consolidation strongly believe they are
correct and those with the opposite view are wrong. It is not uncommon
for scientists – even the most intelligent and famous – to be convinced that
their favored hypothesis is true, despite results that contradict it (Platt, 1964).
The truth is uncovered much faster in science when multiple hypotheses are
considered viable and experimental results are used to rule out a particular
hypothesis. The hypothesis that is not contradicted by the results remains the
most viable, until someone comes up with a new hypothesis to test. This is
the scientific method. Fortunately, even though scientific progress is often
hampered by a strong belief in a favored hypothesis, if that hypothesis is
incorrect there will eventually be a sufficient amount of contradictory
evidence that it will be ruled out.

hippocampal activity was reported during both recent and remote

episodic memories; however, it is unclear whether this activity was
due to the process of episodic retrieval or the task of memory construc-
tion, as the latter process is known to activate the hippocampus (Addis,
Wong & Schacter, 2007).
There is a growing body of evidence that the hippocampus is involved
in long-term memories throughout the lifetime. As such, the process
of consolidation does not appear to result in the complete transfer from
hippocampal–cortical memory representations to cortical-cortical mem-
ory representations. As illustrated in Box 3.2, this is still a highly debated
topic and will continue to be hashed out for years to come.

3.4 Consolidation and Sleep

Consolidation refers to changes in the brain regions underlying long-
term memory. Although the process of consolidation takes years, it is
known to begin during the next period of sleep. A primary role of sleep
may be to integrate new memories into our vast memory store with
minimal disruption of old memories. As illustrated in Figure 3.5A,
sleep involves rapid eye movement (REM) periods and non-REM
periods that alternate every ninety minutes, with four stages of progres-
sively deeper non-REM sleep. It is notable that the first half of a night’s
3.4 Consolidation and Sleep 57

A Awake

REM
Sleep stage

NREM-1
NREM-2
NREM-3
NREM-4

12 1 a.m. 2 a.m. 3 a.m. 4 a.m. 5 a.m. 6 a.m. 7 a.m.

midnight Time

Neocortical
Slow Oscillations

Thalamo-Cortical
Spindles

Hippocampal
Ripples
Reactivation

Figure 3.5 Sleep stages and brain oscillations associated with slow wave sleep and long-term
memory consolidation. (A) REM and non-REM (NREM) sleep stages as a function of time for
a typical night of sleep (thick gray bars show REM sleep). (B) Schematic of the brain that
includes the cortex (off white), the thalamus (dark gray structure near the center), and the
hippocampus (light gray structure; occipital pole to the left) in addition to cortical slow waves,
thalamic-cortical sleep spindles, and hippocampal sharp-wave ripples (labels to the right).

sleep is dominated by non-REM sleep, while the amount of REM sleep

increases during the second half of the night. Non-REM stages 3 and 4,
which are referred to as slow wave sleep, are of particular relevance
because these periods are important for consolidation of long-term
memories (Stickgold & Walker, 2005; Ellenbogen, Payne & Stickgold,
2006; Marshall & Born, 2007). By comparison, REM sleep seems to be
particularly important for consolidation of implicit memories.
58 Brain Regions Associated with Long-Term Memory

As its name implies, slow wave sleep is associated with slow (less than
1 Hertz) waves of brain activity that can be measured across the entire
scalp using EEG. These slow waves orchestrate a number of brain
processes that mediate the process of long-term memory consolidation
(Payne, 2010; Born & Wilhelm, 2012). Slow waves alternate between
down-states corresponding to global decreases in brain activity and up-
states corresponding to global increases in brain activity. As shown in
Figure 3.5B, slow waves synchronize other brain waves including tha-
lamic-cortical sleep spindles (that oscillate at frequencies of 11–16
Hertz) and hippocampal sharp-wave ripples (that oscillate at
a frequency of approximately 200 Hertz). Hippocampal sharp-wave
ripples are of particular importance because they are known to coordi-
nate the hippocampal–cortical interactions that reflect the reproduction
of memories from the previous waking period. In short, important long-
term memories from the previous waking period are replayed during
slow wave sleep, which in turn strengthens these memories and results
in consolidation. Although this mechanism for memory consolidation is
based on strengthening of memory representations through repeated
activations, it has been proposed that sleep may also weaken memory
representations of unimportant events to provide a clean slate for the
next day’s events (Axmacher, Draguhn, Elger & Fell, 2009).
Two widely known empirical studies have provided evidence sup-
porting the role of slow wave sleep in long-term memory consolidation.
In one transcranial alternating current stimulation (tACS) study (see
Chapter 2), participants studied pairs of words before going to sleep
(Marshall, Helgadóttir, Mölle & Born, 2006). While the participants
were in slow wave sleep, frontal-lateral electrodes were stimulated at
0.75 Hertz with a very low current to induce slow brain oscillations.
When the participants woke up, they were presented with one of the
words in each pair and were instructed to recall the other word from the
pair, which is a standard long-term memory task. On a different night,
the same participants learned different word pairs before going to sleep
and were not stimulated with current at frontal-lateral electrodes, which
was a measure of baseline performance. In the 0.75 Hertz stimulation
condition, as compared to the non-stimulation condition, the rate of
word recall more than doubled. In comparison, stimulation at 5 Hertz
did not improve performance on the word recall task. The 0.75 Hertz
stimulation condition also improved performance on another long-term
memory task and did not improve performance on two implicit memory
tasks. These results suggest that slow wave sleep is important for
consolidation of long-term memories. In another study, participants
3.5 Memory Encoding 59

learned objects presented at different locations on a computer screen

along with a corresponding sound cue (e.g., the sound ‘meow’ with
a picture of a cat) before a nap (Rudoy, Voss, Westerberg & Paller,
2009). Naps are known to be dominated by slow wave sleep. During the
nap, sound cues for half of the objects were presented. After the nap,
each object was presented and participants recalled the previous spatial
location of that object. Spatial memory performance was better for
objects that were cued during the nap, as compared to objects that
were not cued. When the sound cue was played during sleep this pre-
sumably reactivated the representation of the object in its previous
spatial location via hippocampal–cortical interactions, which in turn
strengthened the memory trace. The findings of these studies provide
evidence that slow wave sleep is important for long-term memory con-
solidation. However, this is a relatively new area of research and much
more work needs to be done to understand the role of sleep in long-term
memory consolidation.

3.5 Memory Encoding

Long-term memory typically refers to the process of retrieval. However,
long-term memory can also refer to encoding, the acquisition of informa-
tion (see Chapter 1). Long-term memory encoding occurs with little if
any effort in everyday life. If a person pays attention to something or it is
meaningful, they will likely remember it later.
In the laboratory, the brain regions associated with long-term
memory encoding are identified using a subsequent memory analysis.
A standard memory paradigm is employed, where a list of items
is presented during the study phase, and then during the test
phase participants are presented with old and new items and make
“old”–“new” recognition judgments. After the experiment is complete,
each item presented during the study phase is classified based on the
subsequent response given to that item during the test phase. For
example, if the words ‘wolf’ and ‘ocean’, along with other words, are
presented during the study phase, and the subsequent responses to
those words during the test phase were “old” and “new,” respectively,
the words at encoding would be classified as subsequently remembered
(a subsequent “old” response to an old word, e.g., responding “old” to
‘wolf’) and subsequently forgotten (a subsequent “new” response to an
old word, e.g., responding “new” to ‘ocean’). Brain activity associated
with successful encoding is identified by contrasting subsequently
remembered items and subsequently forgotten items.
60 Brain Regions Associated with Long-Term Memory

Figure 3.6 Regions of the brain associated with subsequent memory effects. fMRI
activations associated with subsequent memory (in red/yellow; top, lateral views, occipital
poles toward the center; bottom, coronal views, the left image is the most anterior and the
right image is the most posterior). Medial temporal lobe activity, centered on the
hippocampus, is shown near the bottom of each coronal image in both hemispheres. (A black
and white version of this figure will appear in some formats. For the color version, please
refer to the plate section.)

Long-term memory encoding and retrieval may seem like very differ-
ent mental processes, but they are really quite similar. Memory encoding
can be thought of as the initial experience of an event, while memory
retrieval can be thought of as a re-experience of the same event. As such,
it is not surprising that subsequent memory analyses of fMRI studies
have consistently revealed long-term memory encoding activity in the
same regions that have been associated with long-term memory retrieval
(Spaniol et al., 2009; Kim, 2011). Specifically, as shown in Figure 3.6,
successful long-term memory encoding is associated with activity in
sensory regions and control regions, including the dorsolateral prefrontal
cortex, the parietal cortex, and the medial temporal lobe.
Although long-term memory encoding and retrieval are largely
mediated by the same regions of the brain, there are some subtle differ-
ences. For instance, the perirhinal cortex within the medial temporal lobe
is associated with both long-term memory encoding and retrieval of item
information, but the magnitudes of activity in this region during encoding
3.6 Sex Differences 61

and retrieval are of the opposite polarity. During successful item memory
encoding, there is an increase in perirhinal cortex activity (Davachi,
Mitchell & Wagner, 2003), which may reflect magnified processing due
to attention (see Chapter 8). During successful item memory retrieval,
there is a decrease in perirhinal cortex activity (Ross & Slotnick, 2008),
which may reflect more fluent neural processing when items are repeated
at test (see Chapter 7). The hippocampus has also been associated with
long-term memory encoding and retrieval, with an increase in the mag-
nitude of activity during both processes. This region can be thought of as
binding information processed in different cortical regions during both
long-term memory encoding and retrieval. Still, there is evidence that
the patterns of activity within sub-regions of the hippocampus differ
during long-term memory encoding and retrieval (Duncan, Tompary &
Davachi, 2014). Such differences illustrate that the brain regions asso-
ciated with long-term memory encoding and retrieval are not always
identical. Still, for the vast majority of human long-term memory studies
that have investigated encoding and retrieval, the same general pattern of
results has been obtained. That is, successful long-term memory encoding
and retrieval depend on the dorsolateral prefrontal cortex, the parietal
cortex, and the medial temporal lobe.

3.6 Sex Differences

The behavioral performance of females and males can differ on long-
term memory tasks (Andreano & Cahill, 2009). Males often perform
better on spatial memory tasks such as mental rotation, which requires
working memory as the object representation must be kept in mind (see
Chapter 1), and navigation through a previously learned environment.
Females typically perform better on long-term memory tasks that can
depend on verbal memory such as word list recognition and recall,
associative memory, and autobiographical memory. As almost all long-
term memory tasks can be performed using verbal memory strategies,
females generally have better behavioral performance than males. Sex
differences in the brain can account for the relatively superior behavioral
performance of females. Females have larger numbers of estrogen
receptors in the hippocampus and the dorsolateral prefrontal cortex,
two of the three regions associated with long-term memory (see
section 3.1 of this chapter), which can increase the activity of these
regions (Cahill, 2006). The hippocampus and the dorsolateral prefrontal
cortex are also larger in females than males, relative to overall brain size
(Goldstein et al., 2001). Moreover, females have relatively larger
62 Brain Regions Associated with Long-Term Memory

volumes of language processing cortex (see Chapter 1), which likely

contributes to their superior verbal memory.
Females and males often employ different cognitive strategies and
have distinct patterns of brain activity while they perform the same
task. One fMRI study investigated whether there were sex differences
in the hippocampus during memory for object–location associations
(Frings et al., 2006). There were 10 female participants and 10 male
participants. During the study blocks, participants viewed a video as if
they were walking through a virtual environment with five colored
geometric objects. Figure 3.7A illustrates an aerial view of one envir-
onment. During recognition blocks, an aerial view of each object was
shown in an old location or a new location. Participants responded as to
whether each object was in the “old” location or a “new” location.
During baseline blocks, participants were presented with the same
object in two different sizes and responded as to whether the larger
object was on the “left” or “right,” which controlled for visual percep-
tion and motor processing. Each participant also used the following
four-point rating scale to describe the strategy they used to learn the
object locations: (1) completely verbal, (2) more verbal than pictorial,
(3) more pictorial than verbal, and (4) completely pictorial. Although
there was no difference in behavioral performance between female
participants and male participants, the average strategy rating for
female participants was 2.5 and the average strategy rating for male
participants was 4.0. This indicates that female participants employed
more verbal memory strategies and male participants employed purely
spatial/non-verbal memory strategies. The contrast of recognition
blocks and baseline blocks produced hippocampal activity in all but
one participant. Figure 3.7B shows that activity was lateralized to the
left hippocampus in the large majority of female participants (i.e., the
extent of activity was greater in the left hippocampus than the right
hippocampus) and activity was lateralized to the right hippocampus in
the large majority of male participants (i.e., the extent of activity was
greater in the right hippocampus than the left hippocampus). Another
blocked fMRI study investigating sex differences similarly found that
there was greater activity in the left hippocampus of females than males
during memory for verbal items (Banks, Jones-Gotman, Ladowski &
Sziklas, 2012). These findings are consistent with patient studies indi-
cating that lesions in the left medial temporal lobe impair verbal
memory and lesions in the right medial temporal lobe impair visual
memory (see Chapter 9). Outside of the hippocampus, one event-
related fMRI study reported greater activity in the dorsolateral
3.6 Sex Differences 63

B
9 Women
Men
8

7
Number of participants

0
Left Right
Lateralization

Figure 3.7 Object–location virtual environment and hippocampal laterality results.

(A) Example of to-be-remembered object–location associations in a virtual environment
(objects in dark gray; aerial view). (B) Number of female participants (women) and male
participants (men) with hippocampal activity lateralized to the left hippocampus or the right
hippocampus (key at the top right).
64 Brain Regions Associated with Long-Term Memory

prefrontal cortex and the parietal cortex of females than males when
12-second autobiographical memory trials were contrasted with cate-
gory generation trials (e.g., generate seven examples in the category
‘tool’; Young, Bellgowan, Bodurka & Drevets, 2013).
The results of the previous fMRI studies suggest that there is greater
activity in the left hippocampus, the dorsolateral prefrontal cortex, and
the parietal cortex of females than males during long-term memory
retreival. Such amplification of functional activity may reflect increased
recruitment of these regions in females and mediate their generally
superior performance, as compared to males, during long-term memory
tasks. However, all of the previous fMRI studies employed blocked
designs or trials with a long duration such that the results may have
been influenced by differences in difficulty between blocks or trial types
(where a higher magnitude of activity is associated with greater difficulty;
see Chapter 2). Of importance, the sex differences in the hippocampus
were qualitative (i.e., activity was in different regions for females and
males), while a difficulty account predicts quantitative differences (i.e.,
a different extent of activity in the same regions for females and males).
Future fMRI studies on sex differences should employ event-related
designs and conduct the appropriate contrasts to isolate the cognitive
process of interest (e.g., old-hits versus old-misses to isolate item
memory).
Critically, there is a large body of evidence illustrating sex differences
in behavioral performance, neurochemistry, brain anatomy, and brain
function (Cahill, 2006; Andreano & Cahill, 2009). Sex differences have
largely been ignored in the field of cognitive neuroscience due to a lack of
awareness and/or the high cost of fMRI (see Chapter 11), as studies
investigating sex differences require approximately twice as many parti-
cipants. Fortunately, sex differences have recently become a major issue
in the field of behavioral neuroscience. This has resulted in a heightened
awareness of sex differences across disciplines that will motivate an
increase in cognitive neuroscience research on this important topic.

3.7 Superior Memory

Long-term memory abilities naturally vary among people within the
range of normal, but there are rare individuals that have truly superior
memory abilities. Although there are many productive lines of research
investigating the brain basis of memory failure (see Chapter 5), surpris-
ingly little is known about the brain basis of superior memory, which can
be thought of as extreme memory success. This is, in part, because those
3.7 Superior Memory 65

with superior memory are so rare that it is difficult to find participants for
scientific studies. In addition, there are lines of research that are more
popular than others, and superior memory is not currently a major topic
of inquiry in the field of memory. Fortunately, there are a few sparse lines
of research that have begun to shed light on the brain basis of superior
memory.
One line of research on superior memory has focused on London taxi
drivers who must learn the layout of 25,000 city streets and the locations
of thousands of city attractions. In one study, the investigators assessed
whether there were differences in the sizes of brain regions between taxi
drivers and control participants who did not drive taxis (Maguire et al.,
2000). The assumption was that if a brain region was used more (or less)
in the taxi drivers, this region would grow (or shrink) in size. They found
that taxi drivers had changes in the size of only their hippocampus, with
a relative increase in the amount of gray matter within the posterior
hippocampus and a relative decrease in the amount of gray matter within
the anterior hippocampus. Moreover, the types of changes in both types
of hippocampal gray matter size correlated with the length of time they
had been taxi drivers, which ranged from 1.5 to 42 years (with the largest
changes for those who had been taxi drivers the longest). Figure 3.8
shows the positive correlation in the posterior hippocampus.
These posterior hippocampal effects support the alternative model of
consolidation described in section 3.3, where remote long-term memories
continue to rely on this region. In a follow-up study (Maguire, Woollett &

time as taxi driver (months)

0 50 100 150 200 250 300 350 400
6
adjusted VBM responses
posterior hippocampus

-2

-4

-6

Figure 3.8 Change in the size of the posterior hippocampus as a function of time as a
London taxi driver. The size of the posterior hippocampus (shown on the y-axis) was
measured using voxel-based morphology (VBM).
66 Brain Regions Associated with Long-Term Memory

Spiers, 2006), brain region sizes were compared between London taxi
drivers and London bus drivers, who were a better matched control
group in terms of driving experience, stress, and other factors.
The identical pattern of results was obtained, where the taxi drivers
had a relatively larger posterior hippocampus and a relatively smaller
anterior hippocampus than bus drivers, and this correlated with the
length of time they had been driving a taxi. The investigators also
found that taxi drivers were worse at copying a complex drawing from
memory, and a subsequent study from the same group (Woollett, Spiers
& Maguire, 2009) observed the same result and also reported that taxi
drivers were relatively worse than control participants at learning
object–place pairs and word pairs (they also reported that the taxi drivers
had average IQs). These behavioral memory deficits may be due to the
relatively smaller size of the anterior hippocampus in London taxi dri-
vers. Thus, although the taxi drivers had superior memory for navigating
London, it appears to have come at a cost for other forms of memory.
Another group of individuals who have superior memory are those
who have participated in the World Memory Championships and those
who are known for extraordinary memory abilities. One study com-
pared such individuals with control participants to assess whether
there were differences in cognitive abilities, differences in the size of
brain regions, and differences in the magnitude of fMRI activation
during memory tasks (Maguire, Valentine, Wilding & Kapur, 2003).
Those with superior memory did not differ from control participants in
the cognitive abilities tested (e.g., IQ ranges were 95–119 and 98–119,
respectively) or the size of any brain regions. The fMRI tasks required
memory for a sequence of digits (a task where those with superior
memory excelled), memory for a sequence of faces, or memory for
a sequence of snowflakes. Across tasks, those with superior memory
had greater activation in the posterior hippocampus, the retrosplenial
cortex, and the medial superior parietal cortex, which are regions that
have been associated with long-term memory (see section 3.1 of this
chapter). Almost all of the participants with superior memory reported
using a memory strategy called the method of loci. This method entails
first associating each item during the study phase with a specific sequence
of objects as one mentally travels through a familiar setting (e.g., associat-
ing each item in the study phase with the piece of furniture that is “seen”
while mentally traveling through one’s home). To remember the
sequence of items from the study phase, one mentally travels through
the familiar setting again and the study items can be retrieved in the order
they were initially presented. It is particularly notable that those with
3.7 Superior Memory 67

superior memory had activity in the posterior hippocampus, which is

consistent with the London taxi driver findings described above, and
may reflect retrieval of spatial information while the participants
employed the method of loci strategy. A separate case study investigated
another individual with superior memory known as PI who was able to
recall the digits of π to over 65,000 decimal places (Raz et al., 2009).
Like those with superior memory described above, PI had similar
performance as control participants on the large majority of cognitive
tasks. What stood out was his superior ability on working memory
(99.9th percentile), even though his general memory was average (50th
percentile). In addition, PI was impaired at tests of visual memory (3rd
percentile or below). Thus, as with the London taxi drivers, PI’s superior
memory ability appears to have had a cost outside the domain of his
expertise.
A final group of individuals with superior memory that will be con-
sidered have highly superior autobiographical memory. These indivi-
duals have detailed episodic memory for every day of their later
childhood and adult life. If they are given any date, they can recall
the day of the week, the public events that occurred that day, and
detailed autobiographical details from that day. In one study of
participants with highly superior autobiographical memory, their
performance was normal on most standard cognitive tasks (LePort
et al., 2012). Not surprisingly, they scored very high on tests of auto-
biographical memory. A comparison of the different brain regions
between participants with highly superior autobiographical memory
and control participants revealed a number of differences including
greater white matter coherence in the parahippocampal gyrus,
which may reflect greater contextual processing associated with episo-
dic retrieval, and a relatively smaller anterior temporal cortex.
The decrease in size of the anterior temporal cortex, which has
been associated with semantic memory (see section 3.2 of this chapter),
may reflect the disuse of this region because those with highly superior
autobiographical memory rely more on episodic retrieval. Although
tests of semantic memory were not included in this study, the smaller
size of the anterior temporal cortex would suggest that those with highly
superior autobiographical memory may be deficient on such tasks. This
hypothesis should be tested in future studies.
Across the three groups of individuals with superior memory consid-
ered, some commonalities emerge. First, those with superior memory
are often of average intelligence. Second, although they have superior
memory ability in one domain, they often have inferior memory ability in
68 Brain Regions Associated with Long-Term Memory

another domain. This fits with a zero-sum model of cognitive function,

where a gain in one ability or set of abilities is linked to a loss in another
ability or set of abilities. One other issue is whether the brain changes in
those with superior memory preceded their mental ability or whether
practice with their mental ability resulted in the changes to their brain.
In the case of London taxi drivers, it appears to be the latter as their
memory abilities and brain structure returned to normal after they
retired (Woollett et al., 2009). For the other groups, it is likely
a mixture of nature and nurture that drives their modified mental proces-
sing and brain processing. Superior memory research is still in its infancy
and promises to be an exciting new area of research.

Chapter Summary
• In addition to sensory cortical regions, the three control brain regions
most commonly associated with long-term memory (i.e., episodic
memory and item memory) are the dorsolateral prefrontal cortex,
the parietal cortex, and the medial temporal lobe.
• Within the medial temporal lobe, the perirhinal cortex mediates item
memory, the parahippocampal cortex mediates context memory, and
the hippocampus binds item information and context information.
• In addition to sensory cortical regions, semantic memory has been
associated with the left dorsolateral prefrontal cortex and the anterior
temporal lobes.
• One topic of research is whether the anterior temporal lobes store
semantic information directly or whether this region binds information
that is stored in other cortical regions.
• According to the standard model of long-term memory consolidation,
episodic memory changes from being based on hippocampal–cortical
interactions to being based on cortical–cortical interactions, and this
process takes years.
• There is a growing body of evidence that supports an alternative model
of long-term memory consolidation where retrieval always depends on
the hippocampus.
• Long-term memory consolidation occurs primarily during slow wave
sleep.
• Slow wave sleep is associated with widespread cortical modulation at
frequencies of less than 1 Hertz and primarily occurs during stage
3 non-REM sleep, stage 4 non-REM sleep, and naps.
• During sleep, slow waves synchronize other brain waves including
thalamic-cortical sleep spindles (that oscillate at frequencies of 11 to
Further Reading 69

16 Hertz) and hippocampal sharp-wave ripples (that oscillate at

a frequency of approximately 200 Hertz).
• Hippocampal sharp-wave ripples coordinate hippocampal–cortical
interactions that reflect the replay of memories from the previous
waking period, which in turn strengthens these memories and results
in long-term memory consolidation.
• The brain regions associated with long-term memory encoding are nearly
identical to the brain regions associated with long-term memory retrieval.
• Females, as compared to males, typically employ verbal memory strate-
gies to a greater degree and perform better on long-term memory tasks.
• There is some evidence that females, as compared to males, have
a greater magnitude of activity in the dorsolateral prefrontal cortex,
the parietal cortex, and the left hippocampus during long-term
memory.
• There have been some brain changes reported in those with superior
memory, including relatively larger regions of the brain that appear to
support their extraordinary ability.
• Although more research needs to be done on this topic, those with
superior memory are typically of average intelligence, and their extra-
ordinary memory in one domain appears to come at a cost of poor
performance in other domains.

Review Questions
What are the brain regions most commonly associated with episodic
memory?
Is there more evidence supporting the standard model of consolidation or
the alternative model of consolidation?
Which type of sleep is particularly important for long-term memory
consolidation?
Are long-term memory encoding and retrieval relatively similar
cognitive processes or relatively different cognitive processes?
Which strategy do females typically employ to a greater degree than
males during long-term memory tasks?
Do those with superior memory have advanced abilities on most
cognitive tests?

Further Reading
Rugg, M. D. & Vilberg, K. L. (2013). Brain networks underlying episodic
memory retrieval. Current Opinion in Neurobiology, 23, 255–260.
70 Brain Regions Associated with Long-Term Memory

This article reviews evidence indicating that episodic memory is

associated with a core network of brain regions including the dorsolateral
prefrontal cortex, the parietal cortex, and the medial temporal lobe.
Simmons, W. K., Reddish, M., Bellgowan, P. S. & Martin, A. (2010).
The selectivity and functional connectivity of the anterior temporal lobes.
Cerebral Cortex, 20, 813–825.
This fMRI study investigates the nature of semantic memory processing
in the anterior temporal lobes and finds that this region is preferentially
associated with processing social information.
Bartsch, T., Döhring, J., Rohr, A., Jansen, O. & Deuschl, G. (2011). CA1 neurons
in the human hippocampus are critical for autobiographical memory,
mental time travel, and autonoetic consciousness. Proceedings of the
National Academy of Sciences of the United States of America, 108,
17562–17567.
This study illustrates that patients with lesions in the hippocampus have
deficits in both recent and remote episodic memory retrieval, which
supports the alternative model of consolidation.
Born, J. & Wilhelm, I. (2012). System consolidation of memory during sleep.
Psychological Research, 76, 192–203.
This manuscript reviews the role of sleep in long-term memory
consolidation.
Maguire, E. A., Gadian, D. G., Johnsrude, I. S., Good, C. D., Ashburner, J.,
Frackowiak, R. S. & Frith, C. D. (2000). Navigation-related structural
change in the hippocampi of taxi drivers. Proceedings of the National
Academy of Sciences of the United States of America, 97, 4398–4403.
This study shows that taxi drivers who have superior memory for
navigating the streets of London have a relatively larger posterior
hippocampus and a relatively smaller anterior hippocampus.
CHAPTER FOUR

Brain Timing Associated with Long-Term Memory

Learning Objectives
• To understand the timing and location of brain activity associated with
recollection and familiarity.
• To contrast the evidence on both sides of the scientific debate about
activity that has been associated with familiarity.
• To describe what is meant by synchronous activity and how such activity
indicates two brain regions interact.
• To list the three frequency bands of brain activity associated with
long-term memory.

The large majority of human neuroscience research on long-term memory

has focused on identifying the spatial locations of activity associated with
this process (see Chapter 3). Although the temporal dimension of brain
activity is often ignored, this does not mean that brain activity is static
across time. In reality, brain activity changes rapidly across time, and the
temporal dynamics of activity must be tracked to understand the brain
mechanisms underlying memory. This chapter focuses on the timing of
brain activity associated with long-term memory. As discussed previously
(see Chapters 1 and 3), recollection refers to retrieval of detailed informa-
tion, whereas familiarity refers to retrieval of non-detailed information.
The chapter begins by introducing event-related potential (ERP) activa-
tions (see Chapter 2) that have been associated with familiarity and
recollection (section 4.1). Familiarity has been associated with activity in
frontal brain regions that occurs within 300 to 500 milliseconds after
stimulus onset, while recollection has been associated with activity in
parietal brain regions that occurs within 500 to 800 milliseconds
after stimulus onset. In section 4.2, a scientific debate that has focused
on the ERP activity associated with familiarity is discussed. In section 4.3,
it is shown that synchronous activity in two different brain regions (i.e.,
activation timecourses that increase and decrease together) indicates that
these regions interact. Such synchronous activity between regions during
long-term memory typically occurs within specific frequency bands
of activity including the theta frequency band (4 to 8 cycles per second,
72 Brain Timing Associated with Long-Term Memory

i.e., Hertz), the alpha frequency band (8 to 12 Hertz), and the gamma
frequency band (greater than 30 Hertz). Theta activity reflects the inter-
action between the hippocampus and cortical regions during long-term
memory, alpha activity reflects inhibition of cortical regions, and gamma
activity reflects processing of features in different cortical regions that are
combined to create a unified memory.

4.1 Timing of Activity

The large majority of research on the timing of brain activity associated
with long-term memory has been conducted using ERPs. ERP recording
involves placing electrodes on the surface of the scalp that measure
rapidly changing voltages (i.e., potentials) associated with brain activity
(see Chapter 2). Electrodes are named according to the corresponding
lobe of the brain, such as parietal (P) or frontal (F), with odd-numbered
electrodes over the left hemisphere and even-numbered electrodes over
the right hemisphere. For each electrode, the timing of activity associated
with each event type of interest is given by plotting the corresponding
voltage as a function of time. An increase in the voltage magnitude within
a given time period is assumed to reflect greater brain activity in the
underlying region, even though the localization of cortical activity is
inherently uncertain (see Chapter 2).
ERP studies of memory have focused on two brain activity compo-
nents that are thought to reflect familiarity and recollection (Curran,
Schacter, Johnson & Spinks, 2001; Rugg & Curran, 2007), the two types
of long-term memory (see Chapter 1). The first ERP component is
referred to as the mid-frontal old–new effect. The mid-frontal old–new
effect occurs within 300 to 500 milliseconds after trial onset, has
a maximum amplitude over frontal electrodes (mid refers to the middle
of the head, although it is often lateralized to the left hemisphere), and is
greater in amplitude during accurate familiarity-based retrieval of old
items (i.e., responding “old” to an old item) than correct rejection of
new items (i.e., responding “new” to a new item). The second ERP
component is referred to as the left-parietal old–new effect. The left-
parietal old–new effect occurs within 500 to 800 milliseconds after trial
onset, has a maximum amplitude over left parietal electrodes (although it
often occurs over both hemispheres), and is greater in amplitude during
recollection-based retrieval of old items than correct rejection of new
items. In one study that illustrates these ERP components, participants
4.1 Timing of Activity 73

A
Remember
+
Know
5 µV
New

F5 F6

P5 P6

0 800 ms
B
300–500 ms 500–800 ms

Figure 4.1 ERP activity associated with recollection and familiarity. (A) Activation
timecourses (microvolts as a function of milliseconds) at frontal electrodes and parietal
electrodes associated with “remember” responses to old items (recollection), “know”
responses to old items (familiarity), and “new” responses to new items (event, electrode, and
amplitude keys at the top). (B) Topographic maps illustrating the mid-frontal old–new effect
within 300 to 500 milliseconds (left) and the left-parietal old–new effect within 500 to 800
milliseconds (right; superior views, occipital poles at the bottom; more significant activity is
shown in red). (A black and white version of this figure will appear in some formats. For the
color version, please refer to the plate section.)

viewed pictures of object pairs during the study phase and then during the
test phase were presented with pictures of single old objects or new
objects and made “remember”–“know”–“new” judgments (Vilberg,
Moosavi & Rugg, 2006). Figure 4.1A, top left, shows the mid-frontal
old–new effect between 300 to 500 milliseconds at electrode F5, where
activity associated with “know” responses to old items, which can be
assumed to reflect familiarity (see Chapter 1), was more positive in
magnitude than activity associated with “new” responses to new items.
At this same electrode, activity associated with “remember” responses to
old items was also more positive in magnitude than activity associated
74 Brain Timing Associated with Long-Term Memory

with “new” responses to new items (and did not differ in magnitude from
“know” responses to old items), which is expected given that “remem-
bered” items are also familiar. The mid-frontal old–new effect reflects
a change in magnitude of an ERP component that is negative in value,
which is why this component is referred to as the FN400 (i.e., the frontal
component that is negative in magnitude approximately 400 milliseconds
after stimulus onset). Figure 4.1A, bottom left, shows the left-parietal
old–new effect within 500 to 800 milliseconds at electrode P5, where
activity associated with “remember” responses to old items, which can
be assumed to reflect recollection (see Chapter 1), was more positive
in magnitude than activity associated with “know” responses to old
items and “new” responses to new items. Figure 4.1B, left, shows the
topographic map (i.e., the magnitude of activity across the scalp) for the
mid-frontal old–new effect, which in this case is left lateralized, while
Figure 4.1B, right, shows the topographic map for the left-parietal
old–new effect.
The previous results illustrate the generally accepted view that the
mid-frontal old–new effect reflects familiarity and the left-parietal
old–new effect reflects recollection. There is also compelling evidence
that the magnitude of the left-parietal old–new effect corresponds to
the amount of information recollected. In the preceding study (Vilberg
et al. 2006), participants made a graded “remember” response, where
“R2” corresponded to remembering the object that was previously
paired with the old item and “R1” corresponded to remembering
other information (e.g., seeing that old item on the screen during the
study phase but not the object it was paired with). It can be assumed
that “R2” responses were associated with recollection of greater details
than “R1” responses. The magnitude of activity at electrode P5 within
500 to 800 milliseconds was greatest for “R2” responses to old items,
intermediate for “R1” responses to old items, and smallest for “new”
responses to new items (a subsequent ERP study reported the same
pattern of results; Vilberg & Rugg, 2009). These findings suggest that
the magnitude of the left-parietal old–new effect reflects the amount of
information retrieved during recollection.
One long-standing debate in cognitive psychology is whether famil-
iarity and recollection reflect a single process (i.e., they are quantita-
tively different, corresponding to weak memory and strong memory)
or whether they reflect separate processes (i.e., they are qualitatively
different) (Slotnick & Dodson, 2005; Wixted, 2007). The mid-frontal
old–new effect and the left-parietal old–new effect are topographi-
cally separable, temporally separable, and functionally separable.
4.1 Timing of Activity 75

As such, these separate neural measures could be taken to suggest

that familiarity and recollection are separate cognitive processes and
illustrate that neural measures can constrain cognitive accounts (but
see Slotnick, 2013b).
In addition to the mid-frontal old–new effect that occurs within 300
to 500 milliseconds and the left-parietal old–new effect that occurs
within 500 to 800 milliseconds, there is a third ERP component that
occurs within 1000 to 1600 milliseconds. This component has
a maximum amplitude over right frontal electrodes and is greater
during recollection-based or familiarity-based retrieval of old items
than correct rejection of new items (Curran et al., 2001; Vilberg et al.,
2006; Vilberg & Rugg, 2009; Woroch & Gonsalves, 2010). This ERP
component is illustrated in Figure 4.1A, top right, at electrode F6.
Although this ERP component is not typically focused on in memory
studies, it could be referred to as the right-frontal old–new effect.
The late onset of this effect suggests it might reflect post-retrieval
monitoring (i.e., evaluating what was just remembered) or memory
elaboration (i.e., filling in details of the previous experience). Either of
these possibilities is speculative given the paucity of research in this
area. As previous studies have not focused on the right-frontal
old–new effect, future cognitive neuroscience and cognitive psychol-
ogy research will be needed to understand its role during long-term
memory retrieval.
As detailed in Chapter 3, the brain regions associated with long-
term memory include the dorsolateral prefrontal cortex and the
inferior parietal cortex. These regions correspond with the ERP
components at frontal electrodes and parietal electrodes discussed
above. In one study, the same remember–know–new task was con-
ducted using fMRI and ERPs with different sets of participants
(Vilberg & Rugg, 2009). The left-parietal old–new effect in the
ERP experiment and the left inferior parietal cortex in the fMRI
experiment tracked the amount of information retrieved during
recollection. These results suggest that the left inferior parietal cortex
underlies the left-parietal old–new effect. However, the brain regions
underlying the mid-frontal old–new effect and the right-frontal
old–new effect are currently unknown. Additional studies that
combine methods with high spatial resolution and high temporal
resolution, such as fMRI and ERPs (see Chapters 2 and 11), will be
needed to investigate the spatial-temporal dynamics of long-term
memory retrieval.
76 Brain Timing Associated with Long-Term Memory

4.2 The FN400 Debate

As mentioned above, the mid-frontal old–new effect corresponds to
modulation of the ERP FN400 component, which is why it has also
been referred to as the FN400 effect. It has been hypothesized that the
FN400 effect does not reflect familiarity, a conscious process, but rather
reflects repetition priming, a nonconscious process (Paller, Voss &
Boehm, 2007). Repetition priming refers to a change in the magnitude
of brain activity that occurs when an item is repeated and is thought to
reflect more efficient or fluent processing of the item (see Chapters 1
and 7). Amnesic patients with medial temporal lobe damage have
impaired conscious long-term memory (see Chapters 3 and 9) but
relatively normal repetition priming effects (see Chapter 7), which
indicates repetition priming is a nonconscious process. The results
shown in Figure 4.1A, top left, between 300 to 500 milliseconds after
stimulus onset are consistent with the repetition priming hypothesis, even
though they were discussed in section 4.1 as reflecting familiarity. That is,
the magnitude of activity for old items (i.e., “remember” or “know”
responses to old items), which corresponded to the second time items
were presented, is smaller (i.e., closer to zero) than the magnitude of
activity for new items, which corresponded to the first time items were
presented. Thus, based on results such as those shown in Figure 4.1A, top
left, it is unclear whether the reduction in the magnitude of the FN400
associated with old items reflects familiarity or repetition priming.
Paller and colleagues have argued that the FN400 effect corresponds
to a type of repetition priming referred to as conceptual repetition
priming (see Chapter 7). As conceptual or meaning-based processing is
known to occur within the frontal cortex, conceptual repetition prim-
ing effects can manifest as a change in the magnitude of activity within
the frontal cortex for repeated items. In one study that aimed to assess
whether conceptual priming could modulate the FN400, participants
were presented with abstract shapes during the study phase and then
during the test phase were presented with old shapes and new shapes
(Voss, Schendan & Paller, 2010). During both the study phase and the
test phase, participants rated the meaningfulness of each shape on
a four-point scale from “no meaningfulness” to “high meaningfulness,”
where high meaningfulness can be assumed to reflect conceptual proces-
sing (e.g., a shape that looked like a bird activated the concept of “bird”
and should have received a “high meaningfulness” rating). Figure 4.2
shows conceptual priming effects within 300 to 500 milliseconds in
frontal electrodes, particularly over the middle and right hemisphere
4.2 The FN400 Debate 77

+5 µV
High-Meaning Old
Low-Meaning Old
New
0 200 400 600 800 1000 ms

F3i Fza F4i

C3' Cz C4'

P3i Pzi P4i

Figure 4.2 ERP activity associated with conceptual repetition priming. Activation
timecourses at frontal electrodes, central electrodes, and parietal electrodes associated with
high-meaning old items, low-meaning old items, and new items (event key, amplitude by
time key, in microvolts by milliseconds, and electrode location key at the top).

(at electrodes Fza and F4i), as there is a larger magnitude of activity

associated with high-meaning old shapes than low-meaning old
shapes or new shapes. The identical pattern of results was obtained
in a subsequent study that used ancient Chinese characters as stimuli
(Hou, Safron, Paller & Guo, 2013). These changes in the magnitude
of activity could be assumed to reflect conceptual priming. However,
there are two reasons why this conceptual priming activity does
not correspond to the FN400. First, the conceptual priming ERP
component in these studies was positive in magnitude, which
indicates this component was produced by a different brain region
than the region underlying the FN400. Second, the conceptual prim-
ing effect was largest over central-parietal electrodes (e.g., electrodes
Cz and Pzi), which also indicates that the brain region underlying this
effect was different from the region that produces the mid-frontal
old–new effect.
One ERP study directly compared the voltage distributions associated
with the conceptual priming effect and the mid-frontal old–new effect in
the same participants (Bridger, Bader, Kriukova, Unger & Mecklinger,
2012). As shown in Figure 4.3, the conceptual priming effect had
a central-parietal maximum and the mid-frontal old–new (recognition)
78 Brain Timing Associated with Long-Term Memory

+0.0 µV +4.0

Priming Recognition

Figure 4.3 Topographic maps illustrating the conceptual priming effect (left) and the
mid-frontal old–new (recognition) effect (right; time range from 500 to 800 milliseconds;
superior view, occipital pole at the bottom; key at the top, in microvolts).

effect had a frontal maximum. Both the positive magnitude of the con-
ceptual priming effect and its non-frontal maximum contradict the
hypothesis that the FN400 effect reflects conceptual priming.
There is additional evidence indicating that the mid-frontal old–new
effect reflects familiarity rather than conceptual priming. In one ERP
experiment, words were presented during the study phase, and then
during the test phase old words and new words were presented and
participants made “remember”–“know”–“new” judgments (Woollams,
Taylor, Karayanidids & Henson, 2008). A standard mid-frontal old–new
effect was observed, with a more positive magnitude of FN400 activity
associated with “know” responses to old items (and “remember”
responses to old items) than the magnitude of activity associated with
“new” responses to new items. There was also greater FN400 activity
associated with “know” responses to old items than “new” responses to
old items (i.e., forgotten items). This is particularly important because it
can be assumed that conceptual repetition priming effects occur for all
old items, even forgotten items. It follows that if the FN400 reflected
conceptual repetition priming, there would be no difference between
“know” responses to old items and “new” responses to old items.
As there was a difference between these event types, this difference
must reflect some process other than conceptual priming, such as famil-
iarity (the same pattern of results was obtained in a subsequent study;
Woroch & Gonsalves, 2010). These results also contradict the conceptual
repetition priming hypothesis of the FN400. Considered with the results
4.3 Phase and Frequency of Activity 79

Box 4.1: Debates drive science forward

There are seemingly countless experiments that can be conducted. Simply
use a different stimulus or a slightly different task and a new experiment has
been designed. But what keeps people working hard and keeps the field
moving forward is passion. At the beginning of the FN400 debate, Paller and
colleagues made the astute observation that this ERP component could
reflect conceptual repetition priming. At the beginning of the debate, both
sides acknowledged that the FN400 could reflect familiarity, conceptual
repetition priming, or a combination of both of these processes (Paller
et al., 2007; Rugg & Curran, 2007). However, as the debate continued,
proponents of the conceptual priming hypothesis claimed that evidence in
support of the conceptual priming hypothesis was contradictory to
the familiarity hypothesis (Voss et al., 2010), which was not the case. In the
last few years, the FN400 debate has died down, perhaps because the
proponents of the conceptual priming hypothesis realized that their position
was not supported by the evidence. Of primary importance, this debate and
the passion behind it has led to a deeper understanding about the nature of
the FN400 and our understanding of memory more generally.

reviewed above, the current body of evidence indicates that the FN400
reflects familiarity. Box 4.1 discusses how scientific disagreements, such
as the FN400 debate, can drive advances in the field.

4.3 Phase and Frequency of Activity

The previous sections of this chapter focused on individual ERP
components that were each associated with one cognitive process.
There has also been research focusing on two regions that have similar
patterns of activity over time. If the activation timecourses in two
regions are very similar, it can be assumed that these regions interact
during memory. This is like seeing two people doing exactly the same
dance next to each other. It would be reasonable to assume that people
engaged in a synchronized dance are interacting with each other, rather
than being total strangers who happen to be doing exactly the same
dance at exactly the same time.
Synchronized brain activity was illustrated in a spatial memory ERP
investigation where abstract shapes were presented in the left visual field
or the right visual field during the study phase (Slotnick, 2010b). During
the test phase, participants were presented with old and new shapes and
80 Brain Timing Associated with Long-Term Memory

classified each shape as “old and previously on the left,” “old and pre-
viously on the right,” or “new.” To isolate brain activity associated with
memory for items previously presented in the left visual field, accurate
memory for items previously presented on the right (old-right-hits,
a baseline measure that is not associated with memory for items pre-
viously presented in the left visual field) was subtracted from accurate
memory for items previously presented on the left (old-left-hits; i.e.,
subtractive logic was used; see Chapter 1). To isolate brain activity
associated with memory for items previously presented in the right visual
field, old-left-hits were subtracted from old-right-hits. Figure 4.4A shows
that accurate memory for items previously presented on the left pro-
duced activity in the right occipital region, the right temporal region, and
the right frontal region. Figure 4.4B shows that accurate memory for
items previously presented on the right produced activity in the left
occipital region, the left temporal region, and the left frontal region.
These results are consistent with the known organization of perceptual
processing within the brain, where information in the left visual field is
preferentially processed in the right hemisphere and information in the
right visual field is preferentially processed in the left hemisphere (see
Chapter 1). Figure 4.4C illustrates that the activation timecourses in the
left occipital region, the left temporal region, and the left frontal region
(shown in Figure 4.4B) were very similar, increasing and decreasing in
magnitude with the same timing. Such synchronous activity is referred to
as in phase or phase-locked, given the high correspondence between the
peaks and valleys of the voltage magnitudes over time in these regions.
The same pattern of results was observed for the eight time periods when
the frontal and temporal regions were simultaneously active (i.e., the
phase of the activation timecourses did not differ by more than 1 milli-
second) and, to a lesser degree, for the eight time periods when the
frontal regions and the occipital regions were simultaneously active
(i.e., the phase of the activation timecourses did not differ by more than
9 milliseconds). The activity in frontal and temporal regions, in particu-
lar, was almost perfectly synchronous, which suggests these regions inter-
act during spatial memory.
In addition to assessing the phase of activity between two regions, the
frequency of activity can also be evaluated. Frequency refers to the rate
of change in magnitude over time. Frequencies can be low, changing
slowly over time, or high, changing rapidly over time. Brain activation
timecourses, such as the ones illustrated in Figure 4.4C can also be
considered from a frequency perspective, with lower frequencies corre-
sponding to the slower changes in signal over time (e.g., the single large
4.3 Phase and Frequency of Activity 81

Figure 4.4 Topographic maps and activation timecourses illustrating spatial memory
effects. (A) Topographic map corresponding to accurate memory for items previously
presented on the left (old-left-hits) versus accurate memory for items previously presented on
the right (old-right-hits) at 180 milliseconds after stimulus onset (lateral views, occipital
poles toward the center; key to the right, in microvolts). Regions of interest (white ovals)
included left (L) and right (R) frontal (F), parietal (P), temporal (T), and occipital (O) electrodes
(red discs). (B) Topographic map corresponding to old-right-hits versus old-left-hits at
1417 milliseconds after stimulus onset. (C) Activation timecourses corresponding to
old-right-hits versus old-left-hits in the left frontal, the left temporal, and the left occipital
regions of interest from 1377 to 1477 milliseconds after stimulus onset (key at the top right).
(A black and white version of this figure will appear in some formats. For the color version,
please refer to the plate section.)
82 Brain Timing Associated with Long-Term Memory

increase and then decrease in amplitude over the 100-millisecond

period shown), and higher frequencies corresponding to the faster
changes in signal over time (e.g., the multiple increases and decreases
riding along the plateau in the middle of the activation timecourses).
Certain frequencies of brain activity have been associated with memory
and have been linked to particular brain regions. Specifically, memory
has been associated with brain activity that oscillates in the theta
frequency band (4 to 8 Hertz), the alpha frequency band (8 to 12
Hertz), and the gamma frequency band (greater than 30 Hertz). In the
fields of visual perception and visual attention, gamma activity is known
to reflect binding of features that are processed in different cortical
regions (e.g., shape and color; see Chapter 1) and is a mechanism that
underlies perception of unified objects (Engel, Fries & Singer, 2001).
Theta activity reflects the interaction between the hippocampus and
cortical regions during long-term memory, and alpha activity reflects
cortical inhibition.
In addition to modulation of activity within theta, alpha, and gamma
frequency bands during memory, there is evidence that brain regions
with different frequencies of modulation can be in phase with each
other. This is referred to as cross-frequency coupling and indicates
two brain regions interact. In a subsequent long-term memory
electroencephalography (EEG) study, participants viewed pictures
of objects during the study phase and then during the test phase
were presented with old and new pictures of objects and made
“remember”–“know”–“new” judgments (Friese et al., 2013). EEG
uses the same data acquisition techniques as ERP recording, but the
analysis typically focuses on the magnitude of specific frequency bands
of activity (see Chapter 2). As shown in Figure 4.5A, subsequent
“remember” responses to old items (i.e., subsequently remembered
items) as compared to subsequent “new” responses to old items (i.e.,
subsequently forgotten items) were associated with an increase in
theta activity in right frontal regions, a decrease in alpha activity
in anterior and posterior regions, and an increase in gamma activity
in parietal and occipital regions (from 300 to 1300 milliseconds
after stimulus onset). Moreover, as shown in Figure 4.5B, there was
greater cross-frequency coupling for subsequently remembered
than subsequently forgotten items between frontal theta activity and
parietal-occipital gamma activity. The identical pattern of results for
theta activity and gamma activity was observed with the same experi-
mental protocol during memory retrieval (Köster, Friese, Schöne,
Trujillo-Barreto & Gruber, 2014). Based on the known role of
 A Remembered Forgotten

Theta

Alpha

Gamma

+ signs: electrodes with significant difference

Theta +/- 3, Alpha +/- 6, Gamma +/- 0.4 (μV2)

B Modulation
index (MI)

SR > SF

frontal 0.01
Remembered

theta

parietal
gamma

SF > SR
0.005
0.005 0.01 MI
Forgotten

Figure 4.5 EEG frequency band activity associated with subsequently remembered and
forgotten items. (A) Topographic maps illustrating subsequently remembered and subsequently
forgotten theta activity (top), alpha activity (middle), and gamma activity (bottom; superior
views, occipital pole at the bottom of each image; key at the bottom, in microvolts squared).
(B) Left, schematic illustrating frontal theta activity and parietal-occipital gamma activity
cross-frequency coupling. Right, frontal theta modulation of parietal-occipital gamma activity
(as measured by a modulation index, MI) was greater for subsequently remembered (SR)
than subsequently forgotten (SF) items (each dot represents one participant’s remembered MI
versus forgotten MI, with dots above the line showing the SR > SF effect). (A black and white
version of this figure will appear in some formats. For the color version, please refer to the plate
section.)
84 Brain Timing Associated with Long-Term Memory

gamma activity in visual perception and attention, it can be assumed

that the increase in parietal-occipital gamma activity in these studies
reflected an increase in visual object processing associated with
remembered items, and frontal theta activity may have modulated
this gamma activity. Of particular importance, the cross-frequency
coupling evidence suggests that frontal regions and parietal-occipital
regions interacted during long-term memory encoding and retrieval.
There is also evidence that the thalamus plays a key role in modulat-
ing the frequency of brain activity. In an intracranial EEG (iEEG)
subsequent long-term memory study (see Chapter 2), epilepsy patients
were implanted with depth electrodes in the thalamus to record iEEG,
and electrodes were also placed on their scalp to record EEG
(Sweeney-Reed et al., 2014). Participants were presented with photos
of scenes during the study phase and then during the test phase were
presented with old and new scenes and made “old”–“new” recognition
judgments. Subsequent memory effects were associated with an
increase in frontal activity and thalamic activity in the theta frequency
band that was in phase, and there was also cross-frequency coupling
between frontal activity and thalamic activity. In another combined
iEEG and EEG long-term memory study that recorded from depth
electrodes in the thalamus and electrodes on the scalp, recall of visual
information was associated with an increase in phase-locked thalamic-
occipital gamma activity and a decrease in phase-locked alpha activity
in the thalamus and across the entire scalp (Slotnick, Moo, Kraut,
Lesser & Hart, 2002). It was posited that alpha activity may reflect
inhibition of cortical activity, such that the decrease in the magnitude
of alpha activity reflected widespread cortical disinhibition that pre-
ceded memory retrieval. These studies highlight the role of the thala-
mus in regulating theta, alpha, and gamma activity.
The previous findings indicate that theta activity, alpha activity, and
gamma activity play an important role during long-term memory
encoding and retrieval. Theta activity and gamma activity also appear
to be associated with frontal regions and parietal-occipital regions,
respectively. However, there have been relatively few studies investi-
gating frequency modulation, and much more work needs to be done
in this area. In the fields of perception and attention, studies have
used phase lag to shed light on the nature of interactions between
regions. Phase lag is measured by the time (in milliseconds) or angle
(from 0 to 360 degrees, i.e., 0 to 1 cycle) of offset between activation
timecourses in two different brain regions. For instance, in an
attention magnetoencephalography (MEG) study (see Chapter 2),
4.3 Phase and Frequency of Activity 85

Box 4.2: Brain frequency analysis must be more widely

employed
The large majority of electrophysiological research in the field of memory
has focused on ERP components, such as the left-parietal old–new effect,
and has ignored EEG frequency oscillations, such as theta and gamma
activity. Two reasons for the relative popularity of ERP component analysis
is that it is relatively simple to use and that it is employed in the large majority
of ERP laboratories. However, simplicity and popularity should not be the
primary factors that determine the way in which scientists conduct research.
Scientists should learn the techniques that are necessary to answer the
important questions, even if these techniques are complex or unpopular
(see Chapter 11). One of the most fundamental questions in the field of
cognitive neuroscience is how different brain regions interact during mental
processing. EEG frequency analysis can provide insight into how regions
interact by assessing whether there is synchronous activity in multiple regions
at the same frequency or whether there is cross-frequency coupling between
regions. Phase lag between regions can be used to understand the nature of
the interaction between regions, as a positive phase lag suggests the region
with activity that precedes activity in the other region is driving the interac-
tion. EEG frequency analysis could be conducted using the same data that
was acquired to conduct ERP component analysis. Whenever an ERP dataset
is collected and only component analysis is conducted, the valuable fre-
quency and phase information is wasted. The complex aspects of EEG
frequency analysis may be daunting, but it can be learned like any other
technique. If the field of memory is to make major advances in understand-
ing of the mechanisms underlying memory, this technique must be more
widely employed.

there was a 20-millisecond phase lag between frontal gamma activity

and visual sensory gamma activity (Baldauf & Desimone, 2014). This
20-millisecond phase lag indicated that during attention the frontal
cortex produced a top-down signal that modulated processing in
visual sensory regions. Although phase lag has not been focused on
in memory studies, this is a powerful analysis method that should be
employed in future research. As discussed in Box 4.2, complex brain
analysis techniques such as EEG frequency analysis need to be
embraced to understand the brain mechanisms underlying memory
(see Chapter 11).
86 Brain Timing Associated with Long-Term Memory

Chapter Summary
• ERP evidence indicates that the processes of familiarity and recollection
are associated with the mid-frontal old–new effect and the left-parietal
old–new effect, respectively.
• The mid-frontal old–new effect is an ERP component that occurs
within 300 to 500 milliseconds, has a maximum amplitude over frontal
electrodes, and is greater during familiarity-based retrieval of old
items than correct rejection of new items.
• The left-parietal old–new effect is an ERP component that occurs
within 500 to 800 milliseconds, has a maximum amplitude over left
parietal electrodes, and is greater during recollection-based retrieval
of old items than familiarity-based retrieval of old items and correct
rejection of new items.
• The right-frontal old–new effect is an ERP component that occurs
within 1000 to 1600 milliseconds, has a maximum amplitude over right
frontal electrodes, and is greater during recollection-based or familiar-
ity-based retrieval of old items than correct rejection of new items.
• There has been debate with regard to whether the mid-frontal old–new
effect, which corresponds to modulation of the FN400 ERP component,
reflects familiarity or conceptual priming, but the evidence supports the
familiarity account.
• Long-term memory has been associated with activity in the theta
frequency band (4 to 8 Hertz), the alpha frequency band (8 to 12
Hertz), and the gamma frequency band (greater than 30 Hertz).
• Theta activity reflects the interaction between the hippocampus
and cortical regions during long-term memory, alpha activity reflects
cortical inhibition, and gamma activity reflects processing of features
in different cortical regions that are combined to create a unified
memory.

Review Questions
What are the ERP components associated with recollection and
familiarity?
What is one piece of evidence that contradicts the conceptual priming
account of FN400 modulation?
Which brain activity frequency bands have been linked to long-term
memory?
Which frequency band reflects the interaction between the hippocampus
and cortical regions?
Further Reading 87

Further Reading
Vilberg, K. L., Moosavi, R. F. & Rugg, M. D. (2006). The relationship between
electrophysiological correlates of recollection and amount of information
retrieved. Brain Research, 1122, 161–170.
This ERP article provides evidence that the mid-frontal old–new effect
reflects familiarity, the left-parietal old–new effect reflects recollection
and also illustrates the right-frontal old–new effect.
Bridger, E. K., Bader, R., Kriukova, O., Unger, K. & Mecklinger, A. (2012).
The FN400 is functionally distinct from the N400. NeuroImage, 63,
1334–1342.
This ERP study directly compares the topographic maps associated with
familiarity and conceptual priming and finds that conceptual priming
produces a relatively more posterior maximum, which contradicts the
hypothesis that the FN400 reflects conceptual priming.
Slotnick, S. D. (2010b). Synchronous retinotopic frontal-temporal activity
during long-term memory for spatial location. Brain Research, 1330,
89–100.
This ERP study illustrates phase-locked activity between frontal and
temporal regions during spatial memory, which indicates that these
regions interact.
Friese, U., Köster, M., Hassler, U., Martens, U., Trujillo-Barreto, N. & Gruber, T.
(2013). Successful memory encoding is associated with increased
cross-frequency coupling between frontal theta and posterior gamma
oscillations in human scalp-recorded EEG. NeuroImage, 66, 642–647.
This EEG study reports that subsequent memory is associated with
theta activity, alpha activity, and gamma activity, and also shows
evidence of cross-frequency coupling between frontal theta activity
and parietal-occipital gamma activity.
CHAPTER FIVE

Long-Term Memory Failure

Learning Objectives
• To identify the brain regions associated with typical forgetting.
• To understand the experimental paradigms that are used to investigate
retrieval-induced forgetting and motivated forgetting.
• To describe the interaction between the dorsolateral prefrontal cortex
and the hippocampus during retrieval-induced forgetting and
motivated forgetting.
• To compare and contrast the brain regions associated with true
memory, false memory for related information, and false memory for
unrelated information.
• To determine one way in which flashbulb memories exemplify
memory failure.

The previous two chapters focused on the brain mechanisms under-

lying successful long-term memory. The flip side of memory success is
memory failure, and these processes are intimately linked. As will be
discussed in more detail within this chapter, understanding memory
failure furthers our understanding of memory success. Memory failure
can be broadly classified into forgetting and memory distortion.
Everyone is experienced with forgetting and, even though we are
almost never aware of it, memory distortion. This chapter begins by
reviewing the brain regions associated with typical forgetting, which
can be attributed to a lack of attention during encoding (section 5.1).
In section 5.2, the brain mechanisms underlying retrieval-induced
forgetting are considered, which is when retrieval of one item (e.g.,
the word ‘banana’) has an inhibitory effect on related items (e.g., the
word ‘orange’) and increases the rate of forgetting for these items.
The brain regions associated with a related process called motivated
forgetting, an increase in the rate of forgetting for items that one
intentionally tries to forget, is then considered. In the next two
sections of the chapter, 5.3 and 5.4, two types of memory distortion
are considered: false memories (i.e., memories for information that
did not occur) and flashbulb memories (i.e., seemingly picture-like
5.1 Typical Forgetting 89

memories for very surprising and consequential events). It has been

argued that long-term memory failure reflects an adaptive memory
system that works well (Schacter, 1999; Schacter, Guerin &
St. Jacques, 2011). For example, if we remembered everything, our
minds would be too cluttered (e.g., it would be difficult to remember
where one parked today as compared to yesterday). There has been
much less research on the brain basis of memory failure as compared
to memory success. This stems in part from memory failure being
a less popular topic of inquiry. However, investigations of the brain
mechanisms underlying memory failure have provided critical insights
into how our largely successful memory system operates.

5.1 Typical Forgetting

Forgetting in everyday life can usually be attributed to a failure to attend to
information. This could be for numerous reasons such as not being inter-
ested in the material, being distracted by a cell phone, being sleepy, or
thinking about something else. Attention has been shown to be a key
aspect of encoding (see Chapter 8), even when it is not known that memory
will be tested later. For instance, if participants are asked to deeply process
words, such as deciding whether each word in a study list is “pleasant” or
“unpleasant,” their memory performance will be similar whether or not
they know there is a subsequent memory test. This illustrates that success-
fully encoding information requires attention, rather than the knowledge
that the information will be tested at a later time.
As discussed in Chapter 3, brain regions associated with successful
memory encoding are sometimes identified using a subsequent memory
analysis, where subsequently remembered items are compared to subse-
quently forgotten items. This comparison has consistently produced
activity in the dorsolateral prefrontal cortex, the parietal cortex, and
the medial temporal lobe. Otten and Rugg (2001) did something unusual.
They flipped the typical contrast of subsequently remembered items and
subsequently forgotten items and compared subsequently forgotten
items to subsequently remembered items. This was also a subsequent
memory analysis, but was non-standard. They may have conducted this
comparison because they were theoretically interested in the brain
regions associated with subsequent forgetting. Alternatively, they might
have conducted this comparison because it is not uncommon to conduct
statistical contrasts in both directions and they happened to see activity
associated with subsequent forgetting. Box 5.1 highlights that important
findings in science can be accidental.
90 Long-Term Memory Failure

Box 5.1: Scientists should keep their eyes open

for the unexpected
Otten and Rugg (2001) compared subsequent forgetting and subsequent
remembering, which was unusual because all previous fMRI studies had
employed the opposite contrast. The authors probably stumbled upon their
findings because they conducted the comparison in both directions (which is
called a two-tailed statistical test), even though they were interested in the
brain regions associated with only subsequent remembering. These findings
underscore that scientists should always keep their eyes open for the unex-
pected. Unanticipated results often lead to new research directions and
valuable insights into what is actually going on.

Regardless of the motivation for conducting this comparison, they

reported subsequent forgetting activity across two studies within
a number of brain regions including the dorsolateral prefrontal cortex,
the inferior parietal cortex, and the medial parietal cortex. Motivated by
these findings, Wagner and Davachi (2001) reanalyzed their own fMRI
data from two previously published subsequent memory studies and found
subsequent forgetting activity in the same regions reported by Otten
and Rugg in addition to activity within the medial prefrontal cortex.
Subsequent forgetting has been associated with the same pattern of brain
activity in many studies (e.g., Daselaar, Prince & Cabeza, 2004; Shrager,
Kirwan & Squire, 2008). As shown in Figure 5.1A, a meta-analysis of
seventeen subsequent memory studies showed consistent activity in the
dorsolateral prefrontal cortex, the medial prefrontal cortex, the inferior
parietal cortex, and the medial parietal cortex (Kim, 2011).
Subsequent forgetting activity in the dorsolateral prefrontal cortex and
the parietal cortex may seem particularly unusual given that subsequent
remembering activity has been associated with the same regions (see
Chapter 3). However, distinct sub-regions of the frontal cortex and the
parietal cortex have been associated with subsequent remembering and
subsequent forgetting (Kim, 2011). Logically, this must be the case as the
comparison between subsequent remembering and subsequent forget-
ting cannot activate exactly the same region as the opposite comparison
between subsequent forgetting and subsequent remembering. That is, if
brain activity is positive in magnitude for one of these comparisons, it
cannot also be positive in magnitude (and must be negative in magni-
tude) for the other comparison (e.g., 5 − 3 = +2 and 3 − 5 = −2).
5.1 Typical Forgetting 91

Figure 5.1 Subsequent forgetting fMRI activity and default network fMRI activity.
(A) Subsequent forgetting fMRI activity (in red/yellow) in the right hemisphere (top, lateral
view, occipital pole to the left; bottom, medial view, occipital pole to the right). The same
pattern of activity was reported in the left hemisphere. (B) Default network fMRI activity (in
blue/cyan) in the left hemisphere (top, lateral view, occipital pole to the right; bottom, medial
view, occipital pole to the left). The same pattern of activity was reported in the right
hemisphere. (A black and white version of this figure will appear in some formats. For the
color version, please refer to the plate section.)

The pattern of brain activity associated with subsequent forgetting is

the same as the pattern of brain activity that is referred to as the default
network, which is illustrated in Figure 5.1B (compare Figures 5.1A and
5.1B). The default network consists of the regions of the brain that
become active when participants are not engaged in any particular task,
such as when they lay quietly with their eyes closed, passively look at
a fixation point on the screen, or wait between experimental trials. This
network of brain activity has been associated with many cognitive states,
such as day dreaming, mind wandering, lapses of attention, retrieval of
personal information, and planning for the future (Buckner, Andrews-
Hanna & Schacter, 2008). The association between forgetting and the
default network has implications for cognitive psychology research,
where it is typically assumed that forgetting is due only to inattention.
These neural results suggest that forgetting could also be due to other
active cognitive functions such as retrieval of personal information or
planning for the future, which should be topics of future research in
cognitive psychology. Of particular importance here, default network
activity indicates participants are not engaged in the experimental task.
92 Long-Term Memory Failure

As such, it is not surprising that the default network has been associated
with encoding trials that were subsequently forgotten. The participants
were not attending to the stimuli during those trials but, rather, were
engaged in some other mental process. These results indicate that to
minimize forgetting, one should maintain attention to the information
that is being presented. In the real world, this translates into minimizing
distractions and attending to information that is important. For instance,
it is known that people learn better when they are not multitasking (e.g.,
texting or daydreaming while in the classroom). To avoid forgetting, one
needs to focus attention and stay engaged.

5.2 Retrieval-Induced Forgetting

As described in section 5.1, forgetting can be caused by a failure to
engage during encoding, which reflects an inactive mental process (with
regard to the experimental task). In contrast, retrieval-induced forgetting
is an active process where retrieval of an item from memory inhibits
retrieval of related items. For instance, if the word ‘banana’ is recalled,
the memory representation of the related word ‘orange’, which is also
a fruit, will be inhibited to some degree. It is thought that such inhibition
occurs to reduce the likelihood that a similar but incorrect item will be
retrieved (e.g., to avoid mistakenly saying ‘orange’ when one intends to
say ‘banana’).
Retrieval-induced forgetting is investigated using somewhat
complicated paradigms that are required to uncover the effects. These
paradigms include an initial study phase, an intermediate retrieval
practice phase, and a final recall phase. As illustrated in Figure 5.2A,
in one fMRI experiment, participants were presented with word pairs
consisting of a category and an example of that category in the study
phase (Wimber et al., 2008). During the intermediate retrieval practice
phase, participants were presented with a subset of the categories along
with a two-letter word cue and were asked to mentally complete each
word (during this phase, non-presented words from the same categories
were inhibited). In the final recall phase, participants were presented
with all of the categories and word cues corresponding to word pairs
from the study phase. Categories/words that were presented in the
study phase but were not presented in the retrieval practice phase
served as a baseline level of performance (since these were not inhib-
ited). Figure 5.2B, left, shows the retrieval-induced forgetting effect.
There was a lower percentage of recall for words that were
from the same category as words presented during retrieval practice
5.2 Retrieval-Induced Forgetting 93

A STUDY RETRIEVAL FINAL RECALL

PRACTICE
FRUIT FRUIT
P+
Apple K
SPORT FRUIT P–
Tennis FRUIT A
FRUIT Ma SPORT
C
Kiwi FRUIT H
SPORT FRUIT
Ki P–
Hockev O
FRUIT SPORT
C
Mango T
FRUIT FRUIT
Orange M P+

B 80 P–
70 C–
P+
% recalled

60 C+

50

40

30
Item Type

C 10
8
6
4
BA 47

2
0
–2
–4
–6
–8
–20 –10 0 10 20 30 40
Forgetting in %

Figure 5.2 Retrieval-inducted forgetting paradigm, behavioral performance, and fMRI

activity. (A) The paradigm consisted of a study phase, a retrieval practice phase, and a final
recall phase. All of the categories and words/word cues were presented during the study
phase and the final recall phase. Item types during the final recall phase are labeled to the
right. P+ refers to categories/words from the retrieval practice phase, P− refers to words from
the same categories that were practiced but were not presented during the retrieval practice
phase, and C (control) refers to words from categories that were not presented during the
retrieval practice phase. (B) Left, the percentage of words recalled in the P− condition and the
corresponding control condition (C−). Right, the percentage of words recalled in the
P+ condition and the corresponding control condition (C+; key at the top right). (C) Left,
retrieval-induced forgetting activity in the dorsolateral prefrontal cortex (axial view, occipital
pole at the bottom). Right, the magnitude of activity within Brodmann Area (BA) 47
(extracted from the region within the white square on the image to the left) as a function of
the percentage of retrieval-induced forgetting.
94 Long-Term Memory Failure

(P−; e.g., ‘apple’) than the percentage of recall for words that were
from a different category that was not presented during retrieval prac-
tice (C−; e.g., ‘tennis’; i.e., the dark gray bar is lower than the dark
hatched bar). Figure 5.2B, right, shows there was a higher percentage of
recall for words presented during retrieval practice (P+), which were
also presented during the study phase, than for words that were
from a different category that was not presented during retrieval
practice (C+), which were only presented during the study phase (i.e.,
the light gray bar is higher than the light hatched bar). To identify brain
regions associated with retrieval-induced forgetting during the final
recall phase, non-presented words from the same category as those
presented during retrieval practice (P−, which were inhibited) were
compared with practiced words (P+, which were not inhibited).
As shown in Figure 5.2C (left), this contrast produced activity in the
dorsolateral prefrontal cortex. Figure 5.2C (right) shows that, across
participants, the larger the magnitude of activity in the dorsolateral
prefrontal cortex (Brodmann Area, BA, 47; see Chapter 1), the higher the
percentage of retrieval-induced forgetting. This suggests that the dorsolat-
eral prefrontal cortex actively inhibits non-presented words from the same
category as words presented during retrieval practice. In another retrieval-
induced forgetting study, transcranial direct current stimulation (tDCS;
see Chapter 2) was used to disrupt activity in the right dorsolateral pre-
frontal cortex during the practice phase (Penolazzi, Stramaccia, Brago,
Mondini & Galfano, 2014). This completely eliminated the retrieval-
induced forgetting effect, which indicates that the dorsolateral prefrontal
cortex is necessary to produce this type of forgetting.
Another fMRI study used objects as stimuli (e.g., a picture of
Marilyn Monroe or a hat) and also reported that retrieval-induced
forgetting was associated with an increase in activity within the dorso-
lateral prefrontal cortex (Wimber, Alink, Charest, Kriegeskorte &
Anderson, 2015). This study additionally found that retrieval-
induced forgetting of objects was associated with a decrease in activity
within the hippocampus and a decrease in activity within visual sensory
regions. As activity within the hippocampus and visual sensory regions
is known to be associated with successful long-term memory retrieval
of visual information (see Chapters 1 and 3), it appears that the
dorsolateral prefrontal cortex inhibited activity in these regions,
which in turn produced retrieval-induced forgetting. The finding that
visual regions can be inhibited during retrieval-induced forgetting was
complemented by findings from a retrieval-induced forgetting EEG
study that also used visual stimuli and reported an increase in alpha
5.2 Retrieval-Induced Forgetting 95

θ – Amplitude (5–9 Hz)

35 Difference
30 (SR - RE)
SR
25 RE
Signal change (%)

20
15
10
5
0
–8 % 8%
–0.5 0 0.5 1 1.5
Time (sec.)

Figure 5.3 Retrieval-induced forgetting EEG activity. Left, the magnitude of theta activity
(percent signal change) in the selective retrieval (SR) condition was greater than that of the
re-exposure (RE) condition within 0 to 0.5 seconds after stimulus onset (key at the top right).
Right, topographic map illustrating the difference in theta activity between the selective
retrieval condition and the re-exposure condition within 0 to 0.5 seconds after stimulus
onset (superior view, occipital pole at the bottom; key at the bottom, in percent). (A black and
white version of this figure will appear in some formats. For the color version, please refer to
the plate section.)

activity over visual sensory regions (Waldhauser, Johansson &

Hanslmayr, 2012), given that alpha activity reflects inhibitory proces-
sing (see Chapters 4 and 6).
Another retrieval-induced forgetting EEG study used the same stimulus
protocol described above (Figure 5.2A), except that the intermediate
phase consisted of either a selective retrieval phase, where participants
completed word stems from the study phase (e.g., ‘man___’ for ‘mango’),
or a re-exposure phase, where participants studied the same words from the
study phase (Staudigl, Hanslmayr & Bäuml, 2010). Of importance, only
word stem completion produces retrieval-induced forgetting. To investigate
brain activity associated with retrieval-induced forgetting during the
intermediate phase, items in the selective retrieval condition were
compared to items in the re-exposure condition. Figure 5.3 shows that
there was greater theta activity over the lateral frontal cortex in the selec-
tive retrieval (SR) condition as compared to the re-exposure (RE) condi-
tion. An increase in lateral frontal theta activity during the selective
retrieval condition was also observed in another EEG retrieval-induced
forgetting study (Waldhauser et al., 2012). As theta activity has been
96 Long-Term Memory Failure

associated with interactions between the frontal cortex and the hippocam-
pus (see Chapter 4), these EEG results provide complementary evidence to
the fMRI results above (Wimber et al., 2015) that indicate the dorsolateral
prefrontal cortex and the hippocampus interact to produce retrieval-
induced forgetting.

5.3 Motivated Forgetting

Like retrieval-induced forgetting, motivated forgetting refers to an active
process where retrieval of an item from memory is suppressed. However,
unlike retrieval-induced forgetting, which is an automatic process, moti-
vated forgetting is an intentional process. In everyday life, it could be
psychologically beneficial to forget unpleasant or traumatic memories.
The experimental paradigms used to study motivated forgetting are
similar to those used to study retrieval-induced forgetting. There is an
initial study phase, an intermediate phase that leads to forgetting of some
items, and a final recall phase. The intermediate think/no-think phase is
unique to motivated forgetting paradigms. For each item in this phase,
participants are instructed to think about/rehearse that item or they are
instructed not to think about/rehearse that item.
The first study that investigated the brain regions associated with moti-
vated forgetting employed fMRI (Anderson et al., 2004). During the study
phase, pairs of words were presented (e.g., ‘ordeal–roach’, ‘steam–train’,
and ‘jaw–gum’). During the think/no-think phase, the initial words of
some pairs were shown in red (e.g., ‘ordeal’), which meant the associated
word (e.g., ‘roach’) should not be thought about, the initial words of some
pairs were shown in green (e.g., ‘steam’), which meant the associated word
(e.g., ‘train’) should be rehearsed, and the initial words of some pairs were
not shown, which served as a baseline measure of memory performance.
During the final recall phase, all of the initial words of the pairs were
shown (e.g., ‘ordeal’, ‘steam’, and ‘jaw’) and participants were instructed
to retrieve the associated words (e.g., ‘roach’, ‘train’, and ‘gum’).
The percentage of associated words recalled in the no-think condition
was lower than the percentage of associated words recalled in the baseline
condition, which reflected motivated forgetting. The percentage of asso-
ciated words recalled in the think condition was higher than baseline
performance, which was expected due to the additional rehearsal. Brain
activity associated with motivated forgetting was identified by contrasting
no-think trials (which were associated with subsequent forgetting) and
think trials (which were not associated with subsequent forgetting).
Motivated forgetting was associated with an increase in activity within
5.4 False Memories 97

the dorsolateral prefrontal cortex and a decrease in activity within the

hippocampus. Moreover, across participants, an increase in the
degree of motivated forgetting was associated with an increase in
the magnitude of activity in the dorsolateral prefrontal cortex (similar
to Figure 5.2C, right). In a more recent fMRI study of motivated
forgetting, words were paired with objects during the study phase
(Gagnepain, Henson & Anderson, 2014). For example, the word
‘duty’ was paired with a picture of binoculars. The contrast of no-
think versus think trials in the intermediate phase produced an
increase in activity within the dorsolateral prefrontal cortex and
a decrease in activity within the hippocampus (as in the previous
study) in addition to a decrease in activity within visual sensory
regions.
It has been proposed that motivated forgetting may be due to retrieval
of distracting information rather than inhibition of the no-think informa-
tion. That is, participants may be thinking of something else during the
no-think instruction, rather than suppressing memory for the no-think
item. However, retrieval of distracting information would also engage the
hippocampus, which would predict an increase in the magnitude of
activity within this region. As motivated forgetting has been associated
with a decrease in activity within the hippocampus, this argues against the
possibility that participants retrieve distracting information during moti-
vated forgetting (Benoit & Anderson, 2012; Depue, 2012).
A review of the literature has shown that motivated forgetting
consistently produces an increase in activity within the dorsolateral
prefrontal cortex and a decrease in activity within the hippocampus
(Anderson & Hanslmayr, 2014). In addition, as reported above, moti-
vated forgetting of visual information produces a decrease in activity
within visual sensory regions (Gagnepain et al., 2014). This overall
pattern of brain activity during motivated forgetting is identical to
that of retrieval-induced forgetting described in section 5.2. These
findings provide convergent evidence that active forgetting, whether
retrieval-based or motivated, is caused by a top-down signal within
the dorsolateral prefrontal cortex that inhibits the hippocampus and
sensory cortical regions.

5.4 False Memories

False memory refers to remembering something that never happened.
False memories often stem from memory for the general theme of previous
events, which is referred to as gist. The Deese-Roediger-McDermott
98 Long-Term Memory Failure

(DRM) paradigm is commonly used to study false memory (Deese,

1959; Roediger & McDermott, 1995). In the DRM paradigm,
lists of associated words are presented during the study phase (e.g.,
‘web’, ‘insect’, ‘bug’, ‘fly’), and then during the test phase old words,
new related words (e.g., ‘spider’), and new unrelated words are
presented and participants make “old”–“new” recognition judgments.
Participants have very high levels of false memories for new related
words in these paradigms (they usually respond “old” to ‘spider’ in
the example above). One study even found that the rate of “remem-
ber” responses, which correspond to retrieval of specific details (see
Chapter 1), did not differ between true memories and false memories
(Roediger & McDermott, 1995). It is thought that when the associated
words are presented during the study phase in such paradigms,
participants learn the gist of the list, and this leads to a false memory
for the related item. It has been argued that remembering gist is an
important feature of our memory system (Schacter et al., 2011).
Typically, memory for gist is useful as it allows us to remember general
information without getting bogged down by useless details. For exam-
ple, when a person sees a friend (or an enemy) it makes more sense
for them to remember the gist of that person rather than retrieve all of
their previous interactions. As illustrated below, the brain regions
associated with true memory and gist-based false memory are very
similar.
One fMRI study used a DRM paradigm with abstract shapes as stimuli
(Slotnick & Schacter, 2004). During the study phase, multiple sets of
similar abstract shapes were shown. During the test phase, participants
were presented with old shapes from the study phase, new related shapes
that were similar to the previously studied shapes, and new unrelated
shapes and made “old”–“new” recognition judgments. True memory
activity was isolated by comparing “old” responses to old items (i.e.,
old-hits) and “new” responses to new unrelated items (i.e., correct-
rejections), and false memory activity was isolated by comparing “old”
responses to new related items (i.e., related-false alarms) and correct
rejections. Figure 5.4A, left, shows that both true memory and false
memory activated the dorsolateral prefrontal cortex and the parietal
cortex. True memory and false memory were also associated with activity
in the hippocampus (not shown). Figure 5.4A, right, shows that true
memory and false memory were also associated with activity in later
visual processing regions (one region is illustrated within the black
circle). Thus, both true memory and false memory produced activity in
the core regions associated with long-term memory retrieval: the
5.4 False Memories 99

Old-hit
Related-false alarm
Signal (% change)

0.3
0.2
0.1
0
–0.1
–0.2

0 4 8 12 16
Time (s)

Figure 5.4 Regions of the brain commonly and differentially associated with true memory
and related false memory. (A) fMRI activity (in orange) associated with both true memory
and false memory (left, superior view; right, inferior view; occipital poles at the bottom).
Activity in a later visual region is shown within the black circle. (B) Right, the contrast of
true memory and false memory (old-hits > related-false alarms) produced activity in early
visual regions (inferior view, occipital pole at the bottom), as shown within the black
circle. Left, activation timecourses (percent signal change as a function of time after
stimulus onset, in seconds) associated with true memory (old-hits) and false memory
(related-false alarms; key at the top). (A black and white version of this figure will appear
in some formats. For the color version, please refer to the plate section.)
100 Long-Term Memory Failure

dorsolateral prefrontal cortex, the parietal cortex, the hippocampus, and

sensory processing regions (see Chapters 1 and 3). A review of the
literature showed that these regions are commonly associated with true
memory and false memory (Schacter & Slotnick, 2004). This large degree
of overlapping brain activity for true memory and false memory explains
why participants responded “old” to both old items and new related
items.
However, there were differences in brain activity between true memory
and false memory. Most notably, as shown in Figure 5.4B, there was greater
activity for true memory (i.e., old-hits) than false memory (i.e., related-false
alarms) in more posterior early visual processing regions, including V1 (see
Chapter 1). These findings indicate that activity in early sensory regions can
distinguish between true memory and false memory. The same pattern of
visual area activity was reported in a subsequent study that used words as
stimuli (i.e., true memory and false memory produced activity in later visual
processing regions and true memory produced greater activity in early
visual regions; Kim & Cabeza, 2007). If early visual regions can distinguish
between true memory and false memory, why don’t participants use this
information to respond “new” to related items? Slotnick and Schacter
(2004) reasoned that if participants had conscious access to this information
they would have used it to correctly reject new related items and, therefore,
activity in early visual processing regions may reflect nonconscious proces-
sing. In a follow-up study using exactly the same experimental procedure
(Slotnick & Schacter, 2006), we found that activity in early visual processing
regions reflected repetition priming, a form of nonconscious memory (see
Chapters 1 and 7). Thus, although the brain can distinguish between true
memories and false memories in these paradigms, the mind does not have
access to this information. In a recent fMRI study, we employed stimulus
and task conditions that evoked conscious processing in early visual regions
(Karanian & Slotnick, 2016). True memory produced greater activity than
false memory in early visual regions, which is consistent with the previous
results; however, false memory also produced activity in these regions,
including V1. These findings indicate false memory can activate early visual
regions under certain conditions, which is a topic of future research. More
broadly, these most recent findings question whether fMRI findings should
be implemented in the courtroom to assess whether eyewitness testimony
reflects false memory (Schacter & Loftus, 2013).
True memory for old items and false memory for new related items
have been associated with activation of the dorsolateral prefrontal cortex
and the hippocampus, which could be taken to suggest that these regions
play similar roles during these mental processes. However, patient
5.4 False Memories 101

lesion evidence indicates that the dorsolateral prefrontal cortex and the
hippocampus play different roles during true memory and false memory
(Schacter & Slotnick, 2004). Amnesic patients with lesions that included
the hippocampus have a lower rate of true memories for old items and
a lower rate of false memories for new related items, which indicates that
the hippocampus plays a similar role during true memory and false
memory. This is consistent with the known role of the hippocampus
in binding information that is processed in different cortical regions
(see Chapter 3). True memory involves binding specific information
such as how it looked, its meaning, and where it was presented on the
computer screen, while false memory involves binding non-specific/gist
information such as how the related items looked on the screen, the
theme of the related items, and where the related items were presented
on the computer screen. Box 5.2 discusses how similarities in the
brain regions associated with memory success and memory failure can

Box 5.2: Memory failure is interesting and provides insight

into memory success
Memory researchers are typically interested in how successful memory
operates. So why do they study memory failure? One reason is that it is
inherently interesting. False memories are fascinating because they correspond
to retrieval of events that never occurred, just like visual illusions are interesting
because something is seen that is not there. Understanding memory failure is
also interesting because it could help us figure out ways to improve our
memory (e.g., by attending to relevant information and limiting distraction).
A second, more academic, reason to investigate the brain regions associated
with memory failure is that this line of inquiry provides insight into the role of
these regions during memory success. For example, the hippocampus has been
associated with both true memory and false memory, which suggests that this
region is generally involved in binding information, regardless of whether
or not the memory is real. In addition, there is evidence that the dorsolateral
prefrontal cortex inhibits the hippocampus during memory inhibition.
Although memory inhibition does result in forgetting the inhibited information,
which can be considered memory failure, it is also an important feature of
memory success. Inhibition of related information makes one less likely to
retrieve non-target information, and it is sometimes psychologically beneficial
for us to forget traumatic events. Understanding that binding and inhibition are
a normal part of remembering gives a more complete and accurate picture of
how our largely successful memory system operates.
102 Long-Term Memory Failure

provide insight into the mechanisms underlying memory. In contrast to

hippocampal lesions that impair true memory and false memory, lesions
to the dorsolateral prefrontal cortex impair false memory to a greater
extent than true memory. This suggests that the dorsolateral prefrontal
cortex may play a more important role during false memory than true
memory, which is consistent with the fMRI evidence indicating that there
is greater activity in this region during false memory than true memory
(Slotnick & Schacter, 2007).
Although the large majority of research has focused on false memory
for new related items, as described above, one can also have false memory
for new unrelated items. One fMRI study examined the brain regions
associated with true memory for abstract shapes, false memory for new
related shapes, and false memory for new unrelated shapes (Garoff-
Eaton, Slotnick & Schacter, 2006). Consistent with the findings above,
true memory for old shapes and false memory for new related shapes
produced activity in the dorsolateral prefrontal cortex, the parietal
cortex, the hippocampus, and later visual regions. Figure 5.5 shows that
false memory for new unrelated shapes produced greater activity within
the superior posterior temporal cortex, a region associated with language
processing (see Chapters 1 and 8). These results suggest that subsequent
false memory for new unrelated shapes were mediated by verbal labels
that were shared between these shapes and old shapes. For example,
a participant might have used the verbal label “butterfly” to help them
remember a shape during the study phase. Then, during the test phase,

Figure 5.5 Brain activity associated with unrelated false memory. Left, fMRI activity (in
red/yellow) associated with false memory for new unrelated items (lateral view, occipital
pole to the right). Right, activation timecourses (percent signal change as a function of
time after stimulus onset, in seconds) extracted from activity in language processing
cortex (within the white circle to the left; key to the right). (A black and white version of
this figure will appear in some formats. For the color version, please refer to the plate
section.)
5.5 Flashbulb Memories 103

a new unrelated shape might have also looked like a butterfly, which led
to a false memory. Another fMRI study used abstract shapes that were
either moving or stationary at study, and then at test old shapes were
presented and participants made “moving”–“stationary” judgments
(Karanian & Slotnick, 2014a). False memory for motion (i.e., responding
“moving” to a previously stationary shape), which can be considered
a false memory of an unrelated feature, produced activity in language
processing cortex. This finding provides additional evidence that false
memory for unrelated information is due to language processing (i.e.,
verbal labels). It should also be noted that true memory for motion
produced greater activity than false memory for motion in the hippo-
campus (Karanian & Slotnick, 2014b), which suggests that false memory
for unrelated information does not engage the hippocampus to the same
degree as false memory for related information.
The previous evidence indicates there are two different types of false
memory. False memory for related information is based on gist proces-
sing and involves many of the same brain regions associated with true
memory, including the dorsolateral prefrontal cortex, the parietal cortex,
and the hippocampus. False memory for unrelated information is based
on incorrect assignment of verbal labels and has been associated with
activity in language processing cortex. As only a few studies have inves-
tigated the brain regions associated with false memory for unrelated
information, this is a topic of future research in cognitive neuroscience
and cognitive psychology.

5.5 Flashbulb Memories

At unexpected moments in our lives, very surprising and important
events occur that can lead to a flashbulb memory (Brown & Kulik,
1977). The flashbulb was a tiny one-time use bulb that provided
illumination for cameras in the 1960s and 1970s that were replaced
by electronic flashes that are used on cameras/smartphones today.
The flashbulb metaphor was used to suggest that a very surprising
and consequential event could create a picture-like/extremely accu-
rate memory of the circumstances in which the information was first
learned. Examples of such events are the assassination of John
F. Kennedy, the Space Shuttle Challenger explosion, the New York
City terrorist attack on September 11, 2001 (which will be subse-
quently referred to as 911), and the death of a loved one. A flashbulb
memory refers to retrieval of contextual details such as where someone
was when they first learned of the event, what they were doing, who
104 Long-Term Memory Failure

told them, what happened immediately afterward, and how they felt
(Brown & Kulik, 1977).
Flashbulb memories were initially proposed to be extremely vivid and
accurate; however, subsequent research showed that these memories
are much less like a photograph than their name implies. One study
measured the accuracy of flashbulb memories for 911 by comparing
participant responses to contextual questions (e.g., ‘how did you first
learn about it?’, ‘where were you?’, ‘what were you doing?’) at 1 week,
11 months, and 35 months after the event (Hirst et al., 2009). It can be
assumed that memory accuracy was best 1 week after the event.
Therefore, any inconsistencies between responses at 11 months and 1
week or between 35 months and 1 week could be attributed to memory
failure (e.g., forgetting or false memory). Memory accuracy was approxi-
mately 60 percent correct 11 months after the event and remained
approximately the same after 35 months. A follow-up study approxi-
mately 10 years after 911 showed about the same level of memory
accuracy (Hirst et al., 2015). A similar rate of memory failure was
reported for flashbulb memories that were over two decades old for
a 1980 train explosion in Bologna, Italy (Cubelli & Della Sala, 2008).
These studies illustrate that the accuracy of flashbulb memories is well
within the range of normal episodic memories. One unique feature of
flashbulb memories is that they are associated with very high confidence
ratings (greater than 4 out of 5) that barely diminish over time (Hirst
et al., 2015). Such unfounded confidence is a different type of memory
failure from forgetting and underlies the seemingly vivid nature of flash-
bulb memories.
Surprisingly few studies have investigated the brain basis of flashbulb
memories. It would not be possible to study flashbulb memory at encod-
ing because the events that lead to a flashbulb memory are by definition
unexpected. One study assessed whether patients with lesions to the
frontal lobe or patients with lesions to the medial temporal lobe, includ-
ing the hippocampus, had impaired memory for 911 (Davidson, Cook,
Glisky, Verfaellie & Rapcsak, 2005). The investigators tested memory
for the context in which participants first learned of the event, which
corresponds to flashbulb memory details as described above, along with
memory for facts (e.g., ‘where did the event occur?’, who were the
‘people involved?’), which corresponds to item memory details.
Participant responses within 1 month and about 6 months later were
compared to assess memory inconsistencies. As compared to participants
without brain lesions, patients with medial temporal lobe lesions were
impaired at both context memory and item memory (a similar finding was
5.5 Flashbulb Memories 105

reported by Metternich, Wagner, Schulze-Bonhage, Buschmann &

McCarthy, 2013). Davidson et al. also found that patients with frontal
lesions were only impaired at context memory. These results suggest that
the frontal cortex is particularly important for flashbulb memory retrie-
val. This is consistent with the known role of the dorsolateral prefrontal
cortex in memory for contextual information (Mitchell & Johnson, 2009).
To date, there has been only one fMRI study on flashbulb memory
for 911 (Sharot, Martorella, Delgado & Phelps, 2007). Three years
after 911, participants saw the word ‘September’ or ‘summer’, which
cued them to retrieve an episodic memory of an event that occurred on
911 or in the preceding summer. Some of the participants were close
to the site of the attack (approximately 2 miles), while other partici-
pants were far from the site of the attack (approximately 5 miles).
The authors of the study focused on the role of the amygdala,
a medial temporal lobe region that is associated with emotional
memory (see Chapter 8). For the close group of participants, but not
the far group of participants, there was greater amygdala activity dur-
ing memory for 911 than during memory for the preceding summer.
The authors concluded that the amygdala is important during flashbulb
memories. However, there is a major problem with this interpretation.
All of the participants could be assumed to have had flashbulb
memories. Therefore, if activity in a region differs between the
participant groups, as it did in the amygdala, it should have been
concluded that this region was not critical for flashbulb memories.
The amygdala activation can be attributed to a greater magnitude
of emotional experience in the close group, rather than having anything
to do with flashbulb memory. Across both participant groups,
a comparison between memories of 911 and memories of the preceding
summer produced activity in the dorsolateral prefrontal cortex,
which is consistent with the patient lesion evidence discussed above,
and the parietal cortex. These findings indicate that flashbulb
memories are associated with a greater magnitude of activity in the
same regions that have been associated with normal episodic memories
(see Chapter 3).
The behavioral evidence indicates that flashbulb memories have
similar rates of memory failure as normal episodic memories and the
brain evidence indicates that flashbulb memories depend on the same
regions as normal episodic memories. As such, flashbulb memories can
be considered a normal type of episodic memory, except they are asso-
ciated with an abnormally high degree of confidence. Thus, flashbulb
memories are not as picture-like as their name implies.
106 Long-Term Memory Failure

Chapter Summary
• Typical forgetting can be attributed to a failure to focus on information
during encoding and is associated with activity in the dorsolateral
prefrontal cortex, the medial prefrontal cortex, the inferior parietal
cortex, and the medial parietal cortex.
• The same regions of the brain are associated with typical forgetting
and the default network.
• Retrieval-induced forgetting is associated with an increase in activity
within the dorsolateral prefrontal cortex and a decrease in activity
within the hippocampus (in addition to a decrease in activity within
visual sensory regions for visual items).
• Motivated forgetting is associated with the same regions of the brain as
retrieval-induced forgetting.
• True memory for old items and false memory for new related items are
both associated with the dorsolateral prefrontal cortex, the parietal
cortex, and the hippocampus.
• There is usually greater activity in early sensory cortical regions during
true memory than false memory for new related items.
• False memory for new unrelated items is associated with activity in
language processing regions.
• Flashbulb memories have been associated with activity in the same
brain regions as regular episodic memories.
• Flashbulb memories have a similar rate of forgetting and distortion as
regular episodic memories but are associated with a very high level of
confidence.

Review Questions
Which brain regions have been associated with typical forgetting?
What are the three stages of the paradigm used to investigate retrieval-
induced forgetting?
How do the two types of false memories differ in terms of mental
processing and brain processing?
Are flashbulb memories extremely accurate?

Further Reading
Kim, H. (2011). Neural activity that predicts subsequent memory and forgetting:
A meta-analysis of 74 fMRI studies. NeuroImage, 54, 2446–2461.
Further Reading 107

This review article conducts a meta-analysis of seventeen fMRI studies

and shows that subsequent forgetting is associated with activity in the
dorsolateral prefrontal cortex, the medial prefrontal cortex, the lateral
parietal cortex, and the medial parietal cortex.
Anderson, M. C., Ochsner, K. N., Kuhl, B., Cooper, J., Robertson, E.,
Gabrieli, S. W., Glover, G. H. & Gabrieli, J. D. (2004). Neural systems
underlying the suppression of unwanted memories. Science, 303, 232–235.
This fMRI study shows that directed forgetting is associated with an
increase in activity within the dorsolateral prefrontal cortex and
a decrease in activity within the hippocampus.
Slotnick, S. D. & Schacter, D. L. (2004). A sensory signature that distinguishes
true from false memories. Nature Neuroscience, 7, 664–672.
This fMRI study illustrates that true memories and related false memories
are associated with activity in the dorsolateral prefrontal cortex, the
parietal cortex, and the hippocampus, while true memories produce
greater activity than false memories in early visual cortical regions.
Sharot, T., Martorella, E. A., Delgado, M. R. & Phelps, E. A. (2007). How
personal experience modulates the neural circuitry of memories
of September 11. Proceedings of the National Academy of Sciences of the
United States of America, 104, 389–394.
This fMRI study shows that flashbulb memories, like other episodic
memories, are associated with activity in the dorsolateral prefrontal cortex
and the parietal cortex.
CHAPTER SIX

Working Memory

Learning Objectives
• To identify the brain regions that are thought to store the contents of
working memory.
• To describe how information is coded in early sensory regions during
visual working memory.
• To list three shortcomings of the evidence or analysis techniques that
have been used to associate working memory and the hippocampus.
• To compare and contrast the brain activity frequency bands associated
with working memory and long-term memory.
• To understand what types of changes take place in the brain after
extensive training on working memory tasks.

Working memory refers to actively holding information in mind during

a relatively short period of time, typically seconds (see Chapter 1).
Like most long-term memory paradigms, working memory paradigms
consist of a study phase, a delay period, and a test phase. During working
memory paradigms, information is actively kept in mind during the delay
period. Working memory is an explicit process as its contents dominate
conscious experience. Working memory has been associated with activity
in the dorsolateral prefrontal cortex, the parietal cortex, and sensory
processing regions. Thus, the regions associated with working memory
are similar to those associated with long-term memory (see Chapter 3),
with the notable absence of medial temporal lobe regions such as the
hippocampus. Section 6.1 of this chapter details the brain regions that
store the contents of working memory during the delay period. It has long
been thought that the contents of working memory are stored in the
dorsolateral prefrontal cortex, but more recent evidence indicates that
storage also takes place in early sensory cortical regions such as V1.
In section 6.2, the evidence is evaluated that claims to link working
memory with the hippocampus. In section 6.3, brain activity associated
with working memory that oscillates at particular frequencies is consid-
ered, which includes alpha activity and gamma activity. This also mirrors
the findings of long-term memory (see Chapter 4), except for the lack of
6.1 The Contents of Working Memory 109

working memory theta activity. Finally, in section 6.4, changes in brain

activity are highlighted that have been linked to training-related
increases in working memory capacity. These findings suggest that
extensive training (e.g., multiple times a week for many weeks) on work-
ing memory tasks can produce long-term improvements in behavioral
performance, change the way the brain functions for a period well
beyond the time of training, and perhaps even increase intelligence.

6.1 The Contents of Working Memory

Working memory contents refer to any type of information that one
actively maintains in mind, such as a clip of a song that someone can’t
get out of their head, the address of a party someone mentally repeats
before typing it in their phone, or the code written on the dry erase board
in the laboratory that is visualized as someone walks to the copy machine
down the hall. As with long-term memory, the large majority of research
on working memory has used visual items as stimuli.
Maintaining information in working memory has long been shown to
activate later sensory cortical regions (e.g., the fusiform face area, but not
V1; Slotnick, 2004b). One fMRI study investigated the brain regions
associated with working memory for faces, houses, or spatial locations
(Sala, Rämä & Courtney, 2003). Figure 6.1A illustrates the paradigm. For
each trial, an initial instruction indicated which type of information
should be maintained during the working memory delay period (i.e.,
house identity, face identity, or spatial location). This was followed by
the sample presentation/study phase with items or spatial locations to be
maintained in working memory, the delay period that lasted 9 seconds,
and the test phase where participants responded as to whether or not
a stimulus or spatial location was from the study phase. It should be
underscored that either identity or spatial location was held in working
memory, rather than both identity and spatial location. As shown in
Figure 6.1B, the contrast between the face working memory delay period
and the house working memory delay period produced activity in the
lateral fusiform cortex, which has been associated with face perception,
while the opposite contrast produced activity in the medial fusiform/
parahippocampal cortex, which has been associated with house/scene
perception (i.e., face delay period activity and house delay period activity
was observed in the fusiform face area and the parahippocampal place
area, respectively; see Chapter 1). In addition, the contents of working
memory during the delay period produced activity in different regions
of the dorsolateral prefrontal cortex. Figure 6.1C shows that the
110 Working Memory

HOUSE IDEN +

Instruction
3 Sec Instruction +
Delay
3 Sec
+
Sample
Presentation Memory
3 sec Delay Test
9 sec 3 sec ITI
3 sec

B C
(Location > Control only)

MedFus 0.35% 0.30%

Superior Frontal Areas

Inferior Frontal Areas

(Face > Control only)
Z 0.30% 0.25%
Face MedFus 0.25%
0.20%
identity 0.20%
0.15% 0.15%
2.34 0.10% 0.10%
0.05%
House 0.05%
0.00%
identity –0.05% 0.00%
LatFus LatFus –0.10% –0.05%

Figure 6.1 Object or location working memory paradigm and fMRI results. (A) On each
trial, a cue instructed participants whether to maintain object (face or house) information
or spatial location information during the working memory delay period. Items were
presented during the sample/study phase, followed by the delay period, the test phase,
and an inter-trial-interval (ITI) before the onset of the next trial (the time of each period, in
seconds, is shown under each panel). (B) Maintenance of faces during the delay period
produced activity (in red/yellow) in the lateral fusiform cortex (i.e., the fusiform face area)
and maintenance of houses during the delay period produced activity (in cyan/purple) in
the medial fusiform/parahippocampal cortex (i.e., the parahippocampal place area; axial
view, occipital pole at the bottom). (C) Left, activity (percent signal change) in the superior
dorsolateral prefrontal cortex (identified by contrasting working memory for spatial
locations and control trials) was associated with maintenance of spatial locations (in
green) to a greater degree than maintenance of faces (in red) and houses (in blue). Delay
period activity corresponds to time points 2 to 4 (paradigm timing key at the top). Right,
activity in the inferior dorsolateral prefrontal cortex (identified by contrasting working
memory for faces and control trials) was associated with maintenance of faces (in red)
and houses (in blue) to a greater degree than maintenance of spatial locations (in green).
(A black and white version of this figure will appear in some formats. For the color version,
please refer to the plate section.)

superior dorsolateral prefrontal cortex was associated with working

memory for spatial location to a greater degree than working memory
for faces or houses, and the inferior dorsolateral prefrontal cortex was
associated with working memory for faces and houses to a greater
degree than working memory for spatial location. This processing
6.1 The Contents of Working Memory 111

Figure 6.2 Sustained working memory fMRI activity in the dorsolateral prefrontal cortex.
Left, sustained activity (in gray/white) during the working memory delay period in the
dorsolateral prefrontal cortex (the rightmost activation) and the parietal cortex (the leftmost
activation; lateral view, occipital pole to the left). Right, working memory delay period
activation timecourse (percent signal change as a function of time from study phase onset)
extracted from the dorsolateral prefrontal cortex region within the white circle to the left.
The delay period is illustrated by the gray bar.

distinction in the prefrontal cortex is consistent with the ventral what/

identity and dorsal where/spatial location processing pathways in the
brain (see Chapter 1).
The previously reported distinction in the dorsolateral prefrontal
cortex is intriguing because these were sensory effects, given that they
were associated with item identity and spatial location, but this region is
associated with memory control (see Chapter 1). Many studies have
reported the ventral-dorsal what–where working memory processing
distinction in the prefrontal cortex (Slotnick, 2004b), and a decade ago
the dominant view was that the dorsolateral prefrontal cortex was the
primary storage site for the contents of working memory. As illustrated in
Figure 6.2, this was because sustained activity was consistently observed
in the prefrontal cortex during the working memory delay period and
also because sustained activity was not observed in early visual sensory
regions during the working memory delay period (e.g., V1). However,
Curtis and D’Esposito (2003) proposed that the dorsolateral prefrontal
cortex activity may reflect memory control processes, such as directing
attention to internal representations that are stored in sensory cortical
regions, which would also be sustained during the delay period (see
Chapter 8). Evidence for working memory delay period activity in early
visual sensory regions, which is detailed below, took years to emerge and
provides some support for the view that the dorsolateral prefrontal
112 Working Memory

cortex is associated with memory control rather than memory sensory

effects.
The first fMRI study to report sustained working memory activity in
V1 employed a new analysis technique (Harrison & Tong, 2009). During
the delay period, participants maintained an orientation grating (i.e.,
alternating parallel light and dark bars) at ~25 or ~115 degrees from
horizontal for 11 seconds. This stimulus is known to produce a robust
response in V1 because this region responds to line orientation.
Moreover, rather than conducting a contrast to identify regions with
sustained activity during the delay period, multi-voxel pattern analysis
and a pattern classification algorithm were employed to evaluate the
pattern of activity across V1 that was associated with each orientation.
Such a pattern of activity can be complex, with some voxels being positive
in magnitude, some voxels being negative in magnitude, and some voxels
having a magnitude of zero. A subset of the trials was used to identify the
unique pattern of activity in V1 associated with holding the ~25 degree
grating in working memory and the unique pattern of activity in V1
associated with holding the ~115 degree grating in working memory.
For each of the remaining trials, the pattern classification algorithm
used these two unique patterns to predict which of the two orientation
gratings the participant was holding in working memory. That is, on
a given trial, if the pattern of activity in V1 matched the ~25 degree
grating pattern better than the ~115 degree grating pattern, the pattern
classification algorithm predicted the participant was holding the ~25
degree grating in working memory and vice versa. As there were two
gratings, chance performance was 50 percent correct. This procedure
produced a classification accuracy of over 70 percent correct, which
shows that there was sustained activity in V1 during the working memory
delay period that reflected the maintenance of stimulus orientation.
Similar results were obtained in visual areas V2, V3, and V4 within the
extrastriate cortex. Another fMRI study also used multi-voxel pattern
analysis and a pattern classification algorithm to identify sustained activ-
ity in early visual areas during the working memory delay period
(Serences, Ester, Vogel & Awh, 2009). In this study, either the orienta-
tion of a grating (at ~45 degrees or ~135 degrees) or a color (red or green)
were maintained in working memory during a 10-second delay period.
As in the previous study, classification accuracy was greater than chance
performance, which indicated that sustained activity in V1 during
the working memory period reflected the maintenance of orientation
information and color information. A recent working memory fMRI
study also used multi-voxel pattern analysis and a pattern classification
6.1 The Contents of Working Memory 113

algorithm and found evidence that activity in V1 and V2 reflected

maintenance of spatial location information (Pratte & Tong, 2014).
Specifically, classification accuracy was higher than chance for the orien-
tation grating held in working memory based on activity in contralateral
early visual regions but not ipsilateral early visual regions, (e.g., for right
visual field stimuli, in V1 within the left hemisphere but not the right
hemisphere). These findings are consistent with the known contralateral
spatial organization of these regions (see Chapter 1; similar results were
reported by Sprague, Ester & Serences, 2014). In a TMS study, tempor-
ary disruption of V1 in either the left hemisphere or the right hemisphere
impaired performance on a working memory task for items in the con-
tralateral visual field to a greater degree than items in the ipsilateral
visual field (van de Ven, Jacobs & Sack, 2012). This finding shows that
V1 activity is necessary to maintain an accurate stimulus representation
during working memory. In summary, multi-voxel pattern analyses and
pattern classification algorithms have revealed sustained activity reflect-
ing the maintenance of orientation, color, and spatial location in early
sensory cortical regions during working memory.
The previous results support the view that the contents of working
memory may be mediated by the sensory cortex rather than the
dorsolateral prefrontal cortex. One fMRI study aimed to distinguish
between these possibilities by employing multi-voxel pattern analysis
and a pattern classification algorithm based on activity in both the visual
sensory cortex and the dorsolateral prefrontal cortex (Sreenivasan,
Vytlacil & D’Esposito, 2014). During the study phase, two faces and
two houses were presented and participants were instructed to remember
the faces, the houses, or both faces and houses during the 9- second delay
period. Relatively large regions of interest were analyzed. Visual sensory
regions included extrastriate cortical regions, the parahippocampal
gyrus, and the fusiform gyrus (in both hemispheres). Dorsolateral pre-
frontal regions included the middle frontal gyrus and the inferior frontal
gyrus in both hemispheres. The average magnitude of activity was greater
than zero during the delay period for all three trial types in the dorso-
lateral prefrontal cortex but not in the visual sensory cortex. This is
consistent with previous findings that observed sustained increases in
activity within only the dorsolateral prefrontal cortex using conventional
methods. However, in line with the findings detailed above, multi-voxel
pattern analysis and a pattern classification algorithm revealed sustained
activity for all three trial types in both the visual sensory cortex and the
dorsolateral prefrontal cortex. The patterns of activity associated with
holding faces, houses, and faces/houses must be distinct, otherwise the
114 Working Memory

classification accuracy would have been at chance levels. The authors

made the insightful assumption that a region that stored the contents of
working memory should have meaningful patterns of activity. It was
assumed that the pattern associated with faces should be more similar
to the pattern associated with faces/houses (because they have the face
representation in common) than the pattern associated with houses and
that the pattern associated with houses should be more similar to the
pattern associated with faces/houses (because they have the house repre-
sentation in common) than the pattern associated with faces. That is, for
each individual item type (i.e., faces or houses), the pattern should be
more similar for faces/houses than for the opposite category. This
assumption was tested by evaluating the rate of classification errors
based on activity in the visual sensory cortex and based on activity in
the dorsolateral prefrontal cortex. In the visual sensory cortex, for faces
and houses, the rate of classification errors for faces/houses was greater
than the rate of classification errors for the opposite category, as expected
from a region that reflects the contents of working memory. In the
dorsolateral prefrontal cortex, for faces and houses, there was no differ-
ence in the rate of classification errors for faces/houses and the rate of
classification errors for the opposite category. These findings suggest that
the contents of working memory are primarily stored in the sensory
cortex rather than the dorsolateral prefrontal cortex. However, it is
important to keep in mind that these are the results from a single study
and they do not rule out the hypothesis that the contents of working
memory are stored in the dorsolateral prefrontal cortex.
The evidence reviewed above suggests that the contents of working
memory are stored in the sensory cortex. Future work will be needed to
further evaluate the role of the visual sensory cortex and the dorsolateral
prefrontal cortex during working memory. As illustrated in Box 6.1, this
will be a topic of research for many years to come.

6.2 Working Memory and the Hippocampus

The hippocampus is known to be associated with episodic memory and
item memory, two kinds of long-term memory (see Chapter 3).
If a long-term memory fMRI study is conducted, it is expected that
there will be activation in the hippocampus, and patients with a lesion
in the hippocampus are expected to have impaired long-term memory.
In contrast, until recently, working memory fMRI studies have not
reported activity in the hippocampus, and patients with lesions
restricted to the hippocampus have not had impaired working memory.
6.2 Working Memory and the Hippocampus 115

Box 6.1: Are the contents of working memory stored

in the dorsolateral prefrontal cortex?
For decades, sustained activity in the dorsolateral prefrontal cortex has been
interpreted as reflecting the active storage of information during the
working memory delay period. The recent evidence reviewed in this section
indicates that the patterns of activity in visual sensory regions reflect the
contents of working memory. In light of this recent evidence, does that mean
that we should abandon the hypothesis that information is stored in the
dorsolateral prefrontal cortex during working memory? Not at all. These are
not exclusive hypotheses – information during working memory could be
stored in both the dorsolateral prefrontal cortex and the visual sensory
cortex. Much more work needs to be done to evaluate whether the dorso-
lateral prefrontal cortex, the visual sensory cortex, or both of these regions
store the contents of working memory.

The classic example is patient H. M., who had both medial temporal
lobes removed, including the hippocampus, which caused a severe
deficit in long-term memory but did not cause a deficit in working
memory (see Chapter 1).
There have been recent claims that working memory, like long-term
memory, may be associated with the hippocampus. One fMRI study used
a novel working memory paradigm in an effort to uncover activity in the
hippocampus (Hannula & Ranganath, 2008). During the study phase,
four objects were presented at random locations along the perimeter of
a three by three grid. During the 11-second delay phase, participants
were instructed to mentally rotate the objects on the grid 90 degrees.
During the test phase, participants responded as to whether or not
the test objects/locations matched the mentally rotated object set.
The contrast of subsequently correct and subsequently incorrect
responses during the study phase produced activity in the hippocampus.
The contrast of correct and incorrect responses during the test phase also
produced activity in the hippocampus. However, the contrast of subse-
quently correct and subsequently incorrect responses during the delay
period did not activate the hippocampus. There was also no sustained
activity in the hippocampus during the delay period after collapsing over
accuracy. The authors took the hippocampal activations during the study
phase and during the test phase as evidence that the hippocampus is
associated with working memory. However, there are serious problems
with this interpretation. First, there was no hippocampal activity during
116 Working Memory

the delay period, which is the only phase that actually reflects working
memory. Second, during the study phase and test phase, novel stimuli
were presented on the screen, and it is known that the hippocampus is
activated by such stimuli. Third, the stimuli and task depended heavily on
spatial processing, and it is known that the hippocampus is associated
with spatial processing. Fourth, even though this was a working memory
paradigm, during the study phase and during the test phase, long-term
memory encoding processes were presumably operating (as anything
that is attended is encoded; see Chapter 5). Therefore, it is not possible
to determine whether the hippocampus was active due to working
memory processing or due to long-term memory processing during the
study phase and the test phase. Another fMRI study used an insightful
paradigm and analysis to tease apart working memory processing and
long-term memory processing (Bergmann, Rijpkema, Fernández &
Kessels, 2012). During the working memory study phase, four
face–house pairs were sequentially presented, followed by a 10-second
delay. During the working memory test phase, three face–house pairs
were presented and participants responded as to whether each was the
same or rearranged. After all the working memory trials were complete,
there was a surprise recognition memory test. Only brain activity during
the study phase was evaluated. Activity associated with accurate working
memory encoding was isolated by contrasting subsequently correct and
subsequently incorrect working memory responses for trials in which
long-term memory was incorrect (i.e., long-term memory was constant
and subtracted out in the contrast; see Chapter 1). Activity associated
with accurate long-term memory encoding was isolated by contrasting
subsequently correct and subsequently incorrect long-term memory
responses for trials in which working memory was correct (i.e., working
memory was constant and subtracted out in the contrast). Working
memory encoding did not produce activity in the hippocampus; however,
long-term memory encoding did produce activity in the hippocampus.
These results indicate that the working memory encoding findings
reported by Hannula and Ranganath (2008) could be attributed to long-
term memory encoding. A recent reanalysis of the Hannula and
Ranganath (2008) fMRI data employed multi-voxel pattern analysis
(Libby, Hannula & Ranganath, 2014). Hippocampal activity was again
reported using non-standard comparisons and analyses of the study
phase data, and no hippocampal activity was associated with the delay
period, the standard measure of working memory. Thus, to date, there is
no compelling fMRI evidence that the hippocampus is associated with
working memory.
6.2 Working Memory and the Hippocampus 117

There has also been recent brain lesion evidence that has attempted to
link the hippocampus to working memory. One study investigated work-
ing memory performance in three epilepsy patients who had their right
medial temporal lobe structures removed, including the hippocampus
(Finke et al., 2008). Figure 6.3A shows the paradigms. During the sam-
ple/study phase of color working memory trials, squares of different
colors were presented followed by a 900- or 5000-millisecond delay.
During the probe/test phase, participants responded as to whether or
not the stimulus matched the color of one of the items from the study
phase. A similar protocol was used for spatial location trials. Association
trials required participants to maintain both color and location informa-
tion during the delay period. Figure 6.3B shows that the patients with
medial temporal lobe lesions performed normally on all conditions
except for the association task at the 5000-millisecond delay, where
they were impaired. The authors interpreted this as showing evidence
that the hippocampus plays a significant role during working memory for
associations. This fits with the view that the hippocampus mediates
binding of information (see Chapter 3), given that there was only impair-
ment in the association condition. However, there are multiple problems
with this interpretation. First, the impairment was only observed at the
5000-millisecond delay. Working memory processes would be expected
to operate at both 900- and 5000-millisecond delays, and thus if the
hippocampus was associated with working memory there should have
been impaired performance at both delays. Long-term memory pro-
cesses, by comparison, would be more dominant at the longer delay,
and thus if the hippocampus was associated with long-term memory
there should have been a larger impairment at the longer delay, as was
observed. Second, the hippocampus is known to be associated with
spatial processing, which was required in color–location working memory
trials. Thus, the impaired performance might have been due to a problem
in spatial processing that was only observable in the association condition
that was more difficult. Third, the lesions included multiple medial tem-
poral lobe structures including the right amygdala, the hippocampus, the
entorhinal cortex, and the perirhinal cortex. As such, it is unclear whether
the impaired performance was caused by a lesion to the hippocampus or
one of these other regions.
In a direct response to the last two shortcomings, another study
investigated working memory performance on non-spatial working
memory tasks with a patient who had a lesion that was restricted to the
hippocampus (Baddeley, Allen & Vargha-Khadem, 2010). This patient
had an over 50 percent reduction in the volume of the hippocampus
118 Working Memory

A SAMPLE DELAY PROBE

200 ms 900 / 5000 ms match non-match

COLOR

LOCATION

ASSOCIATION

B COLOR LOCATION ASSOCIATION

100 100 100
correct responses [%]

90 90 90

80 80 80

70 patients 70 70
controls

900 5000 900 5000 900 5000

delay [ms] delay [ms] delay [ms]

Figure 6.3 Color and/or location working memory paradigms and medial temporal lobe
lesion results. (A) During each color working memory trial, illustrated at the top, colored
squares were presented during the sample/study phase, there was a 900- or 5000-
millisecond delay period, and then there was a probe/test phase in which participants made
“match”–“non-match” judgments. The same paradigm was used for location and
association (i.e., color and location) trials, illustrated at the middle and bottom, respectively.
(B) Performance (percent correct) on the color, location, and association working memory
tasks as a function of delay period duration (in milliseconds) for patients with medial
temporal lobe damage and control participants that did not have a brain lesion (asterisks
indicate significantly impaired performance in the patients as compared to control
participants). (A black and white version of this figure will appear in some formats. For the
color version, please refer to the plate section.)
6.3 Working Memory and Brain Frequencies 119

within both hemispheres. One task required maintenance of color–shape

associations and the other task required word associations. The patient
was not impaired on either of these working memory tasks. In a follow-up
study, the same patient was tested on working memory tasks requiring
maintenance of color, location, color–location, or object–location
(Allen, Vargha-Khadem & Baddeley, 2014). The object–location work-
ing memory trials were also followed by a long-term memory recall task,
where each object was presented and the participant selected its previous
location. The patient was not impaired on any of the working memory
tasks, but performance was at chance on the long-term memory task
(control participants performed well on both tasks). These results show
that the hippocampus is associated with long-term memory rather than
working memory. Moreover, they indicate that the deficit in working
memory performance reported by Finke et al. (2008) was due to lesions in
regions other than the hippocampus. These results would be bolstered by
future studies of patients with lesions restricted to the hippocampus,
which will presumably also show no deficits in working memory
performance.
Considering the findings above, there does not appear to be any
compelling evidence that the hippocampus is associated with work-
ing memory. This is not surprising given that there have been hun-
dreds of fMRI studies and hippocampal lesion studies on working
memory. If working memory was associated with the hippocampus,
this association would have become evident by now. This does not
mean that scientists should stop looking for such an association, but
until such evidence is uncovered and survives scrutiny, it is sensible
to conclude that the hippocampus is not associated with working
memory.

6.3 Working Memory and Brain Frequencies

Like long-term memory, working memory has been reported to be asso-
ciated with brain activity in the theta frequency band (4 to 8 Hertz), the
alpha frequency band (8 to 12 Hertz), and the gamma frequency band
(greater than 30 Hertz; see Chapter 4). As with long-term memory, alpha
activity reflects inhibition, while gamma activity reflects binding of infor-
mation in different cortical regions. However, as discussed below, the role
of theta activity during working memory is questionable at best.
One study employed EEG to investigate theta activity, alpha activity,
and gamma activity during working memory (Sauseng et al., 2009).
The paradigm is illustrated in Figure 6.4A. On each trial, during the
120 Working Memory

Cue

Memory Array

200 ms

Retention Interval
100 ms

Probe
900 ms

2000 ms

B
0.01
3.0
Theta -locked Gamma Phase

Alpha Activity [µV/m2]

1.5
Synchronization

contralateral
0.00 ipsilteral 0.0

–1.5

–3.0
Load 2 Load 3 Load 4 Load 6
–0.01
Load 2 Load 3 Load 4 Load 6

Figure 6.4 Color working memory paradigm and EEG results. (A) During each trial, an arrow
cued one hemifield. The memory array/study phase consisted of two to six colored squares in
each hemifield, followed by a retention interval/delay period where the stimuli in the cued
hemifield were maintained, and then during the probe/test phase participants indicated
whether or not any of the colors in the cued hemifield had changed. (B) Left, theta-gamma
synchronization as a function of the number of items in working memory (i.e., working
memory load) at contralateral and ipsilateral occipital-parietal recording sites (key to the
right). Right, alpha activity as a function of working memory load at contralateral and
ipsilateral occipital-parietal recording sites. (A black and white version of this figure will
appear in some formats. For the color version, please refer to the plate section.)

study phase, participants were cued to maintain two to six colored

squares in one visual field and ignore the colored squares in the other
visual field. Participants maintained the stimuli in working memory
during the 900-millisecond delay period and then decided whether the
color of any of the stimuli in the attended visual field had changed.
6.3 Working Memory and Brain Frequencies 121

Figure 6.4B, left, shows that theta-gamma cross-frequency coupling (see

Chapter 4) increased during the delay period over contralateral (but not
ipsilateral) occipital and parietal electrodes (i.e., maintenance of items in
the left visual field produced theta-gamma synchronization over right
visual regions, and vice versa). This was particularly apparent as working
memory load increased from two to four items, and may have decreased
at load six because this number of stimuli was greater than the capacity of
working memory. These findings suggest that theta-gamma activity over
contralateral visual regions reflected the contents of working memory,
which is consistent with the findings detailed in the first section of this
chapter. It is important to mention that the occipital-parietal theta activ-
ity observed should not be interpreted as reflecting frontal-hippocampal
interactions that occur during long-term memory. Figure 6.4B, right,
shows that alpha activity increased during the delay period over ipsilat-
eral (but not contralateral) posterior electrodes. This supports the view
that alpha activity reflects suppression of visual activity, as it would be
beneficial during this task to suppress the to-be-ignored stimulus repre-
sentations in ipsilateral visual regions. A subsequent MEG study that
required working memory for colors at particular spatial locations also
reported an increase in gamma and alpha activity as a function of working
memory load; however, there was no increase in theta activity (Roux,
Wibral, Mohr, Singer & Uhlhaas, 2012).
Another EEG study investigated the frequency of brain activity during
working memory for specific items as compared to working memory for
the order in which items were presented (Hsieh, Ekstrom & Ranganath,
2011). During each trial of the study phase, four kaleidoscope images,
which looked like multi-colored complex snowflakes, were presented one
at a time. Participants received a cue before the trial instructing them to
remember the four stimuli (on item memory trials) or to remember the
temporal order of the four stimuli (on order memory trials). Participants
maintained these items during the 4-second delay period. During the test
phase on item memory trials, an old item and a similar item were pre-
sented and participants selected the one they thought was “old.”
On order memory trials, two old items were presented and participants
selected the one they thought was presented “first.” The authors
analyzed brain frequencies during the delay period, which is notably
different from the methods employed by the same group in section 6.2
of this chapter (Hannula & Ranganath, 2008; Libby et al., 2014). Hsieh
et al. (2011) reported greater posterior alpha activity during working
memory for item information than during working memory for temporal
order information and greater frontal theta activity during working
122 Working Memory

memory for temporal order information than item information.

The analysis did not consider gamma activity. The posterior alpha activ-
ity during working memory for items is similar to the studies described
above and may have reflected inhibition of visual regions that were not
actively maintaining the stimulus representations. The frontal theta activ-
ity during working memory for temporal order may have reflected
frontal–hippocampal interactions. In support of this possibility, a recent
study recorded from depth electrodes in the hippocampus and reported
theta-gamma cross-frequency coupling during working memory for faces
that were presented sequentially during the study phase (Chaieb et al.,
2015). One commonality between the two studies that reported frontal
theta activity (Hsieh et al., 2011) and hippocampal theta activity (Chaieb
et al., 2015) is that they both presented stimuli sequentially. As such,
participants may have been processing the temporal order of the stimuli
during the working memory delay period, which was required in the first
study and may have been done incidentally in the second study. As it is
known that the hippocampus is involved in temporal order processing
(see Chapter 10), the theta/hippocampal activity reported in these studies
can be attributed to temporal order processing rather than working
memory.
The previous findings indicate that posterior gamma activity reflects
the contents of visual working memory and posterior alpha activity
reflects suppression of irrelevant information. Although working
memory studies in cognitive psychology have typically focused on
the contents of working memory, the robust alpha activity described
above indicates that future cognitive psychology studies should also
investigate the effects of working memory on irrelevant distractors.
Theta activity has only sometimes been observed in working memory
studies (e.g., Roux et al., 2012; for a review, see Roux & Uhlhaas,
2014), which indicates that such activity does not reflect working
memory and rather reflects processes that sometimes occur during
certain working memory tasks. As discussed in Box 6.2, hippocampal
activity does not appear to be associated with working memory.

6.4 Brain Plasticity and Working Memory Training

One line of research has investigated whether training on a working
memory task produces changes in brain activity, which is referred to as
brain plasticity. There is some evidence that extensive training on
a working memory task not only improves performance on that task
6.4 Brain Plasticity and Working Memory Training 123

Box 6.2: Working memory does not depend on the

hippocampus
A hypothesis can be generated about anything. One popular hypothesis is
that working memory depends on the hippocampus. There is currently no
convincing support for this hypothesis as working memory has very rarely
activated the hippocampus and hippocampal lesions have very rarely
disrupted working memory. As the hippocampus is known to be involved
in long-term memory, spatial processing, temporal processing, and novel
stimulus processing, all the observed hippocampal activations during
working memory can be discounted due to these confounds or unjustified
analyses. The lack of an association between working memory and the
hippocampus is called a null finding (i.e., a result that is not statistically
significant). One can never be absolutely certain that a null finding is correct,
which is referred to as accepting the null hypothesis. It could be argued that
the analysis techniques used have not been sensitive enough to uncover
hippocampal activity during working memory or that working memory
deficits following hippocampal lesions require subtle tests that have not yet
been employed. Although those are theoretical possibilities, there have been
thousands of studies on working memory and the brain, and none of them
has convincingly linked working memory to the hippocampus. Thus, it can
currently be concluded that such a link does not exist.

but can also improve intelligence (Jaeggi, Buschkuehl, Jonides & Perrig,
2008).
One fMRI study assessed the changes in brain activity following
training on a working memory task (Jolles, Grol, Van Buchem,
Rombouts & Crone, 2010). During each trial of the study phase,
three to five objects were sequentially presented and participants
were asked to verbally encode these objects in the order presented.
Participants were then cued to either maintain the objects in the pre-
viously presented order (in the maintenance condition) or reverse the
order of the objects (in the manipulate condition) during the delay
period. During the test phase, one of the objects was presented and
participants pressed a button to indicate the position of that item in the
sequence. Participants practiced this task for about 25 minutes, three
times per week, for 6 weeks. Working memory delay period activity was
measured before practice (time point 1) and after 6 weeks of practice
(time point 2), and a subset of participants completed a behavioral
test 6 months after that (time point 3, to assess whether there were
124 Working Memory

A B
100
Time point 2 > Time point 1
90
Accuracy (% correct)

80

70

60
load 3
50 load 4
load 5

Time point 1 Time point 2 Time point 3

Figure 6.5 Behavioral effects and brain effects of working memory training. (A) Working
memory accuracy (percent correct) as a function of time (time point 1 = pre-training, time
point 2 = 6 weeks of training, time point 3 = 6 months after time point 2) and load (key at the
bottom right). (B) fMRI activity (in dark gray) at time point 2 versus time point 1 (axial view,
occipital pole at the bottom).

long-term effects of training). Figure 6.5A shows that working memory

accuracy did improve with training, particularly for higher working
memory loads of four or five items, and these improvements were
sustained 6 months after training. Figure 6.5B illustrates that training
produced an increase in activity within the anterior prefrontal cortex
and the parietal cortex during the working memory delay period.
An earlier working memory fMRI study that employed a spatial loca-
tion paradigm reported that 5 weeks of training similarly produced an
increase in activity within the dorsolateral prefrontal cortex and the
parietal cortex and also produced a decrease in activity within another
region of the dorsolateral prefrontal cortex (Olesen, Westerberg &
Klingberg, 2004). Although it may seem paradoxical that working
memory training would produce both an increase in activity and
a decrease in activity within the dorsolateral prefrontal cortex, this is
a huge brain region that is associated with many cognitive functions.
The decreases in activity likely reflected stimulus or response fluency
due to training, which corresponds to repetition priming (a type of
implicit memory; see Chapter 7). The increases in activity likely
reflected strategies that were employed to make a difficult task more
manageable, such as chunking (where multiple items are associated
with one another) or increasing attention to the items held in working
memory (see Chapter 8).
6.4 Brain Plasticity and Working Memory Training 125

Working memory training has been associated with both increases

and decreases in the dorsolateral prefrontal cortex and the parietal
cortex in many studies, particularly when the training was for many
hours across multiple weeks (Klingberg, 2010; Li et al., 2015). Training
for less than an hour has only been associated with decreases in activity
within these regions and in visual sensory regions, which can be attrib-
uted to repetition priming. These findings indicate that ramping up
activity in the dorsolateral prefrontal cortex and the parietal cortex in
these tasks requires a lot of training. When participants received
extensive training on working memory tasks, an increase in behavioral
performance on other related and non-related tasks has also been
reported (Jaeggi et al., 2008; Klingberg, 2010). This is likely because
the dorsolateral prefrontal cortex and the parietal cortex are involved
in many cognitive functions including working memory, long-term
memory, imagery, and attention (see Chapter 8). As discussed in
Box 6.3, working memory may simply be another label for imagery.
Although a number of studies have not reported an increase in perfor-
mance on non-trained tasks, these were null findings (which are always
questionable). Future studies should ensure there is extensive working
memory training and employ a broad range of tasks in an effort to
better understand this process.

Box 6.3: Does working memory exist?

The field of working memory is much larger than the field of imagery.
To illustrate, a PubMed.gov (article database) search for the terms ‘working
memory’ and ‘fMRI’ identified over three times the number of articles than
the search for the terms ‘imagery’ and ‘fMRI’. However, the cognitive pro-
cesses and brain regions associated with working memory and imagery
appear to be identical (see Chapter 8). Any cognitive process should be
broken down into its most fundamental operations. Perception is the most
basic cognitive operation and is associated with activity in only sensory
processing regions. Imagery can be described as a weak form of perception
that also activates sensory processing regions but depends on dorsolateral
prefrontal cortex and parietal cortex control regions. Working memory gets
its name because it refers to the active (working) maintenance of previously
presented (remembered) information. However, this description seems
overly complex, as remembering information that was just presented is not
really memory (at least not the way the term is commonly used), and it can be
126 Working Memory

Box 6.3: (cont.)

argued that the process of working memory is nothing more than imagery.
By comparison, long-term memory requires retrieval of previously learned
information and depends on the hippocampus, and thus is distinct from
imagery. It is arguable that since working memory is not associated with any
mental processes or brain regions beyond those associated with imagery,
working memory does not exist as a separate cognitive function. This has no
functional relevance. These are just labels, and those who investigate work-
ing memory will continue to call it working memory. However, scientists who
are primarily interested in the brain mechanisms underlying memory should
be less interested in imagery/working memory than in definitive types of
memory (i.e., long-term memory and implicit memory).

Chapter Summary
• Sustained working memory activity has long been observed in the
dorsolateral prefrontal cortex, which was thought to reflect the
contents of working memory.
• Recently, multi-voxel pattern analyses and pattern classification
algorithms have been used to uncover working memory activity in
early sensory cortical regions such as V1.
• Although there have been claims that the hippocampus is associated
with working memory, these findings are questionable based on the
methods employed (e.g., the analysis was not restricted to the delay
period) and/or confounding factors.
• Brain activity in the alpha frequency band and the gamma frequency
band have been consistently associated with visual working memory.
• Brain activity in the theta frequency band during working memory can
be attributed to confounding factors.
• Extensive training can produce increases and decreases in activity
within the dorsolateral prefrontal cortex and the parietal cortex,
which can be attributed to training-related plasticity of the brain and
repetition priming, respectively.
• There is some evidence that extensive training on working memory
tasks can improve performance on non-related tasks and even increase
intelligence.
Further Reading 127

Review Questions
How do the regions thought to store the contents of working memory
today differ from the regions thought to store the contents of working
memory 10 years ago?
What are the analysis procedures that have been used to uncover
working memory activity in early sensory regions?
Why is the evidence linking working memory to the hippocampus
questionable?
Which brain activity frequency bands has working memory been
consistently associated with?
Does training for any duration on a working memory task increase
activity in the dorsolateral prefrontal cortex?

Further Reading
Sala, J. B., Rämä, P. & Courtney, S. M. (2003). Functional topography of
a distributed neural system for spatial and nonspatial information
maintenance in working memory. Neuropsychologia, 41, 341–356.
This fMRI study illustrates the view that the contents of working memory
are stored in the dorsolateral prefrontal cortex but not in early sensory
regions.
Harrison, S. A. & Tong, F. (2009). Decoding reveals the contents of visual
working memory in early visual areas. Nature, 458, 632–635.
This fMRI study is the first to show that multi-voxel pattern analysis and
a pattern classification algorithm can reveal sustained patterns of activity in
early sensory regions, including V1.
Hannula, D. E. & Ranganath, C. (2008). Medial temporal lobe activity predicts
successful relational memory binding. The Journal of Neuroscience, 28,
116–124.
This fMRI study aimed to link working memory to the hippocampus, but
employed a paradigm that depended heavily on spatial processing and an
analysis that was not restricted to the delay period.
Sauseng, P., Klimesch, W., Heise, K. F., Gruber, W. R., Holz, E., Karim, A. A.,
Glennon, M., Gerloff, C., Birbaumer, N. & Hummel, F. C. (2009). Brain
oscillatory substrates of visual short-term memory capacity. Current
Biology, 19, 1846–1852.
This study shows that visual working memory is associated with EEG
activity in the gamma frequency band and the alpha frequency band
within occipital regions and parietal regions and includes an elegant
TMS experiment that provides a causal link between alpha activity and
inhibition of distracting items.
128 Working Memory

Olesen, P. J., Westerberg, H. & Klingberg, T. (2004). Increased prefrontal and

parietal activity after training of working memory. Nature Neuroscience, 7,
75–79.
This fMRI study is one of the first to show that extensive training on
a working memory task can produce increases in activity within the
dorsolateral prefrontal cortex and the parietal cortex.
CHAPTER SEVEN

Implicit Memory

Learning Objectives
• To describe the behavioral effects and brain effects that typically occur
during implicit memory.
• To identify the brain regions associated with implicit memory.
• To characterize the brain activity frequency bands associated with
implicit memory.
• To detail the different neural models of implicit memory.
• To determine whether there is convincing evidence that implicit memory
is associated with the hippocampus.
• To describe two different patterns of brain activity that occur during
skill learning.

In everyday life, the term memory is used to refer to the conscious

experience of a previous event. However, when an event is repeated,
there can also be behavioral effects and brain effects that occur outside of
conscious experience. Implicit memory refers to a lack of conscious
experience or awareness of previously learned information. This includes
more efficient or fluent processing of an item when it is repeated (i.e.,
repetition priming) and skill learning (see Chapter 1). Section 7.1 of this
chapter considers the brain regions that have been associated with impli-
cit memory, which include the dorsolateral prefrontal cortex and sensory
processing regions (a subset of the regions associated with long-term
memory; see Chapters 1 and 3). In section 7.2, the frequency bands of
activity associated with implicit memory are discussed, which include
gamma activity and alpha activity (a subset of the frequency bands of
activity associated with long-term memory; see Chapter 4). Although
there is some overlap between implicit memory and long-term memory
with regard to the associated regions and frequency bands of activity,
there are many notable differences that will be highlighted in this
chapter. For instance, in direct opposition to the increases in cortical
activity associated with long-term memory, implicit memory is typically
associated with decreases in cortical activity. Section 7.3 details theore-
tical models of neural activity that underlie implicit memory effects and
130 Implicit Memory

discusses ways in which these models can be distinguished from one

another. In the fourth section, 7.4, evidence is considered that has
claimed to link the hippocampus to implicit memory, which if true
would contradict the evidence that this region is associated with only
long-term memory. The last section, 7.5, focuses on skill learning by
evaluating how brain activity changes over time, from the initial stage
of learning that depends on long-term memory to a later stage of learning
that depends on implicit memory. As skill learning has been investigated
using tasks that are too simplistic and training durations that are too
short, more realistic paradigms will need to be employed to study this
important topic in the future.

7.1 Brain Regions Associated with Implicit Memory

When an event is first experienced, many brain regions are activated. For
example, if someone views a picture of a bison from the Badlands in
South Dakota, activity would occur in many visual regions and in the left
dorsolateral prefrontal cortex, which is involved in language/conceptual
processing (see Chapters 1 and 8). If they experience the same event at
a later time (e.g., they viewed the same picture), they would both process
the information faster and there would be a relative decrease in the
magnitude of activity in the same cortical regions, which is referred to
as repetition priming. The classic view is that the reduction in the magni-
tude of brain activity reflects more fluent or efficient processing for
repeated events. Such reductions in brain activity occur whether or not
the event is processed consciously (e.g., whether or not the person
remembers previously seeing the picture of the bison).
The large majority of studies that have investigated the brain regions
associated with implicit memory have employed repetition priming
paradigms. During the study phase of these paradigms, items such as
objects are presented and participants make a perceptual or conceptual
decision as quickly as possible such as “symmetrical”–“asymmetrical”
judgments, “larger”–“smaller” than a reference size judgments, or
“animate”–“inanimate” judgments. During the test phase, old and
new items are presented and participants make the same perceptual
or conceptual decision, again as quickly as possible. There are two
aspects of the paradigm that are very important. First, in contrast to
a direct task such as old–new recognition that encourages long-term
memory retrieval (which is an explicit/conscious process), an indirect
task is employed that asks about perceptual or conceptual properties
of items and does not require long-term memory retrieval. Second,
7.1 Brain Regions Associated with Implicit Memory 131

participants are asked to respond as quickly as possible, which aims

to minimize reliance on long-term memory. Repetition priming
paradigms appear to largely isolate nonconscious processing, as
response times are much shorter (typically less than 1 second) than
those associated with long-term memory (which are typically greater
than 2 seconds). Moreover, unlike long-term memory tasks (see
Chapter 3), repetition priming and other implicit memory tasks do
not depend on the medial temporal lobe (Squire, 1992; Schacter,
Dobbins & Schnyer, 2004), which will be discussed in section 7.4 of
this chapter.
The reduction in cortical activity associated with repetition priming,
which is also referred to as repetition suppression or adaptation, is one
of the most robust effects in the field of cognitive neuroscience. One
repetition priming fMRI investigation employed pictures of objects
(Koutstaal et al., 2001). Figure 7.1A illustrates the paradigm. During
the study phase, a list of objects was presented four times and partici-
pants quickly responded as to whether each object was larger or
smaller than a 13-inch square box. During the test phase, same/old,
different (i.e., perceptually different from old items with the same
name), or novel/new items were presented and participants made the
same size judgments. Behavioral repetition priming effects were
observed as participants responded faster to same/old items than to
novel/new items (and responded at an intermediate speed to different
items). Brain activity reductions associated with repetition priming
during the test phase were isolated by contrasting novel/new items
and same/old items. This contrast may seem unusual, as long-term
memory activity is often isolated by contrasting correct responses to
old items and new items (see Chapters 1 and 3). However, the opposite
contrast is required to isolate implicit memory activity because new
(unprimed) items are associated with a greater magnitude of activity
than old (primed) items. Figure 7.1B shows that repetition priming was
associated with activity in the left and right dorsolateral prefrontal
cortex (left image, the two large activations at the top) and activity in
the left and right ventral occipital-temporal cortex, within the fusiform
gyrus (right image, the two large activations at the bottom). There was
also repetition priming activity in the left and right posterior lateral
temporal cortex (not shown). The dorsolateral prefrontal cortex and
posterior lateral temporal cortex reductions in activity were thought to
reflect more efficient conceptual/language processing for repeated
objects, particularly in the left hemisphere (see Chapters 1 and 8),
and the occipital cortex reductions in activity were interpreted as
132 Implicit Memory

A B

NOVEL > REPEATED SAME

TEST

STUDY

C L Fusiform R Fusiform
0.5 0.5
Novel Novel
0.4 Same 0.4 Same
Different Different
% SIGNAL CHANGE

% SIGNAL CHANGE
0.3 0.3

0.2 0.2

0.1 0.1

0 0

–0.1 –0.1

–0.2 –0.2
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
TIME (sec) TIME (sec)

Figure 7.1 Repetition priming paradigm and fMRI results. (A) Left, during the study phase,
objects were presented. Right, during the test phase, the same/old objects, different objects
with the same name, and novel/new objects were presented. (B) Decreases in fMRI activity
for repeated same/old items as compared to novel/new items. Left, dorsolateral prefrontal
cortex activity is shown at the top left and the top right. Right, ventral occipital cortex activity
is shown at the bottom left and the bottom right (axial views, occipital poles at the bottom).
(C) Event-related activation timecourses (percent signal change as a function of time after
stimulus onset) extracted from the left fusiform cortex and the right fusiform cortex for same/
old, different, and novel/new items (key at the top right of each image). (A black and white
version of this figure will appear in some formats. For the color version, please refer to the
plate section.)

reflecting more efficient visual processing for repeated objects.

Figure 7.1C illustrates the event-related activation timecourses that
were extracted from the left fusiform cortex and the right fusiform
cortex. In the left fusiform cortex, repetition priming was observed for
same/old items and there was also some degree of repetition priming
for different items (i.e., the magnitudes of activity for both of these
event types were lower than the magnitude of activity for novel/new
items). This suggests that the left fusiform cortex shows some degree of
7.1 Brain Regions Associated with Implicit Memory 133

repetition priming for both old items and items that are perceptually
different but share the same name. By comparison, in the right fusi-
form cortex, repetition priming was only observed for same/old items,
as there was no difference between the activation profiles for novel/
new and different items. This suggests that the right fusiform cortex
shows repetition priming effects for only exactly the same items.
A similar pattern of results was observed in another repetition priming
fMRI study that used objects (Vuilleumier, Henson, Driver & Dolan,
2002). Repeated objects and perceptually different objects with the
same name (which correspond to the same and different conditions in
the previous study) both produced reductions in activity within the left
inferior dorsolateral prefrontal cortex, which can be assumed to reflect
conceptual repetition priming. For half of the objects during the test
phase, the viewpoint (i.e., the angle at which the object was presented)
was changed. In the left posterior fusiform cortex, reductions in activ-
ity, relative to new items, were reported for objects that were presented
from the same viewpoint and for objects that were presented from
a different viewpoint. In the right posterior fusiform cortex, reductions
in activity relative to new items were reported for objects that were
presented from the same viewpoint but not for objects that were
presented from a different viewpoint. This again shows that the same
items and similar items produce repetition priming effects in the left
fusiform cortex, but only the same items produce repetition priming
effects in the right fusiform cortex.
An fMRI guided TMS study (see Chapter 11) investigated whether
the left inferior dorsolateral prefrontal cortex is necessary for object
repetition priming during a living–non-living task (Wig, Grafton, Demos
& Kelly, 2005). For each participant, during an initial fMRI session, the left
inferior dorsolateral prefrontal cortex region associated with object repeti-
tion priming was identified, and repetition priming effects were also
observed in the occipital cortex. Then, in a second session, TMS was
applied to the left inferior dorsolateral prefrontal cortex region or
a motor cortex control region during the study phase using a set of new
objects. Finally, in a subsequent fMRI session, old objects from the second
session and new objects were presented to evaluate behavioral repetition
priming effects and brain repetition priming effects. TMS to the left
inferior dorsolateral prefrontal cortex (but not the control region) elimi-
nated both behavioral repetition priming effects (i.e., participants were no
longer faster at classifying repeated objects from the second session) and
brain repetition priming effects in the left dorsolateral prefrontal cortex.
Repetition priming effects were intact within the occipital cortex,
134 Implicit Memory

prefrontal cortex lateral temporal cortex

visual cortices

Least Stimulus Most

specificity

Figure 7.2 Review of cortical repetition priming effects. Repetition priming effects have
been consistently observed in the dorsolateral prefrontal cortex (in green), the lateral
temporal cortex (in red), and in the visual cortices (in blue) within the posterior occipital
cortex and the ventral occipital-temporal processing stream. Within visual cortical regions,
more posterior regions are the most stimulus specific and more anterior regions are the least
stimulus specific (lateral view, occipital pole to the left; key at the bottom). (A black and
white version of this figure will appear in some formats. For the color version, please refer to
the plate section.)

presumably because these effects depended on perceptual processing

rather than conceptual processing. These findings show that the left
inferior dorsolateral prefrontal cortex is necessary for intact behavioral
repetition priming and intact brain repetition priming effects in this
region.
The results of the preceding studies indicate that the dorsolateral
prefrontal cortex and posterior lateral temporal cortex, particularly in
the left hemisphere, reflect repetition priming of conceptual/language
information (see Chapter 8), while ventral occipital-temporal cortex
reflects repetition priming of perceptual information. These findings are
consistent with a review of numerous priming studies illustrated in
Figure 7.2 (Schacter, Wig & Stevens, 2007). This review concluded that
conceptual repetition priming effects, which are less dependent on the
7.2 Brain Timing Associated with Implicit Memory 135

perceptual features of the items, occur in the dorsolateral prefrontal

cortex and in the lateral temporal cortex (particularly in the left
hemisphere) . In addition, perceptual repetition priming effects occur in
visual cortex. More specific perceptual overlap is required for repetition
priming between the study stimulus and the test stimulus in more poster-
ior visual regions, and less specific perceptual overlap is required for
repetition priming between the study stimulus and the test stimulus in
more anterior visual regions (and more specific overlap is required for
repetition priming in the right hemisphere).
Repetition priming has almost always been associated with decreases
in cortical activity, which supports the view that this type of implicit
memory reflects more fluent or efficient processing. It should also be
noted that there have been reports of repetition priming-related increases
in visual cortical activity, particularly for items such as abstract shapes or
unfamiliar objects (Henson, Shallice & Dolan, 2000; Slotnick & Schacter,
2006). Such increases in activity have only rarely been observed because
almost all studies use familiar items as stimuli. However, the fact that
such repetition priming-related increases in activity can occur indicates
that the term repetition suppression is too restrictive and does not apply
to all forms of repetition priming in the brain. There is recent behavioral
evidence that such increases in cortical activity during repetition priming
may be due to the increased allocation of attention to repeated unfami-
liar items (Thakral, Jacobs & Slotnick, forthcoming). This is an exciting
topic for future research.

7.2 Brain Timing Associated with Implicit Memory

Implicit memory studies that have used techniques with high temporal
resolution have largely employed repetition priming paradigms. It is
important to keep in mind that repetition priming (for familiar items) is
associated with a decrease in the magnitude of fMRI activity. In all
previous studies that have been considered in this book (see Chapters 3
and 4), memory has produced an increase in fMRI activity and an
increase in the magnitude of electrophysiological activity (i.e., ERPs
and ERFs; see Chapter 2), which indicates that these measures are
correlated. Therefore, one might expect that repetition priming should
produce a decrease in the magnitude of ERP or ERF activity, but this has
not been consistently observed. Such ERP/ERF effects have been mixed
across studies, with some studies showing no changes in magnitude,
others showing a decrease in magnitude, and others showing an increase
in magnitude.
136 Implicit Memory

Although repetition priming ERP/ERF findings have been unclear,

repetition priming has consistently affected EEG/MEG gamma activ-
ity and alpha activity. In an EEG repetition priming study, words (e.g.,
‘hug’) and pseudowords (e.g., ‘wug’) were presented once or twice and
participants made “word”–“nonword” judgments (Fiebach, Gruber &
Supp, 2005). For the first presentation of both words and nonwords,
there was an increase in activity within the gamma frequency band
(25–80 Hertz) at occipital electrodes and parietal electrodes from
200 to 350 milliseconds after stimulus onset. Figure 7.3A shows
that this posterior gamma activity was higher in magnitude and more
phase-locked for the first presentation of words than for the second
presentation of words (i.e., there was a decrease in the magnitude of
gamma activity and phase-locking for repeated items). The identical
pattern of results was obtained in another EEG repetition priming
study that employed objects as stimuli (Gruber & Müller, 2005).
In an MEG repetition priming study, objects were presented once or
repeated, and participants mentally named each item and pressed
a button as soon as they could identify it (Gilbert, Gotts, Carver &
Martin, 2010). Repeated objects, as compared to novel objects, were
associated with an increase in activity within the alpha frequency
band (centered at 12 Hertz). Figure 7.3B shows that this increase in
alpha activity occurred within the right fusiform gyrus (which
was identified using source localization; see Chapter 2) starting 200
milliseconds after stimulus onset. The same pattern of alpha activity
was observed in the right dorsolateral prefrontal cortex. Another repe-
tition priming MEG study similarly reported an increase in alpha
activity in the dorsolateral prefrontal cortex and the inferior temporal
cortex from 190 to 270 milliseconds after stimulus onset (Ghuman, Bar,
Dobbins & Schnyer, 2008). Furthermore, there was a phase lag of
approximately 30 milliseconds between the alpha activity in the dorso-
lateral prefrontal cortex and the alpha activity in the inferior temporal
cortex (see Chapter 4). This suggests these regions were interacting in
a top-down manner, with the dorsolateral prefrontal cortex driving the
alpha activity in the inferior temporal cortex. A repetition priming
intracranial EEG study in patients with intractable epilepsy comple-
mented the previous results (Engell & McCarthy, 2014). Repeated
versus novel faces were associated with a decrease in gamma activity
and an increase in alpha activity within the fusiform cortex starting
from 100 to 300 milliseconds after stimulus onset.
The preceding repetition priming EEG/MEG findings provide
a consistent pattern of results. Starting at about 200 milliseconds,
7.2 Brain Timing Associated with Implicit Memory 137

A 1st 2nd

0.4
0.2

Cz
0

P3 Pz P4
μV 2

O1 O2

B
1.6
log10 (Powerstim/Powerprestim)

P < .001
1.2

0.8

P < .05 0.4 Repeated

Novel
n.s. 0

–0.4
0 100 200 300 400 500
time (ms)

Figure 7.3 Repetition priming EEG and MEG results. (A) Magnitude of posterior EEG gamma
activity (in microvolts squared; key to the left) and phase-locking (indicated by the lines) for
the 1st presentation of words (i.e., new words) and the 2nd presentation of words (i.e., old
words; circles show electrode locations; superior views, occipital poles at the bottom). (B)
Left, repetition priming increases in the magnitude of MEG alpha activity within the right
fusiform gyrus for repeated/old items as compared to novel/new items (coronal view;
statistical significance key to the left, n.s. = not significant). Right, event-related activity
extracted from one region to the left (within the white square) illustrating repetition priming
effects (log of power/magnitude for each stimulus divided by the baseline power/magnitude
before each stimulus as a function of time, in milliseconds, after stimulus onset; key to the
right; asterisks indicate significant differences).

repetition priming is associated with a decrease in gamma activity within

visual processing regions and an increase in alpha activity within visual
processing regions and the dorsolateral prefrontal cortex. As gamma
activity reflects processing in visual cortical regions (see Chapters 4
and 6), repetition priming-related decreases in gamma activity can be
assumed to reflect decreases in visual cortical activity. As alpha activity
reflects inhibitory processing, repetition priming-related increases in
138 Implicit Memory

alpha activity may reflect the mechanism by which visual cortical activity
is reduced (possibly via a top-down signal from the dorsolateral prefron-
tal cortex). As these repetition priming EEG/MEG gamma/alpha effects
both reflect a decrease in cortical activity, they complement the fMRI
results discussed in the first section of this chapter.
The pattern of EEG/MEG gamma activity and alpha activity may
provide a clue as to why ERP/ERF priming effects have not been
consistently observed across studies. ERPs/ERFs are averages of
activity across all frequency ranges. If repetition priming decreases
the magnitude of gamma activity and increases the magnitude of alpha
activity, they might cancel each other out when averaged. There may
be stimuli or tasks (or participants) with more dominant gamma activ-
ity, and repetition priming ERP/ERF effects may occur under only
these conditions. Providing some support for this possibility, one
repetition priming study that reported changes in gamma activity but
not alpha activity also reported a decrease in the magnitude of ERP
activity for old items relative to new items (Gruber & Müller, 2005),
whereas another repetition priming study that reported changes in
both gamma activity and alpha activity did not observe ERP priming
effects (Engell & McCarthy, 2014). Further research will be required
to solve the mystery behind the inconsistent repetition priming ERP/
ERF effects.

7.3 Models of Implicit Memory

The evidence that has been reviewed thus far has shown that repetition
priming of familiar items produces a decrease in brain activity, as
measured by fMRI activity and gamma activity. The classic explanation
is that repetition priming produces more fluent or efficient processing,
but this explanation is vague. There are three theoretical models of
neural activity that describe what actually may be happening during
repetition priming (Grill-Spector, Henson & Martin, 2006). These
models apply to other forms of implicit memory as well.
Figure 7.4A depicts the pattern of activity within individual neurons
when an item is first presented. These neurons are interconnected (as
shown by the lines) and could reflect activity across the cortex (e.g.,
multiple centimeters away from one another), such as between the
face and house processing regions of the ventral occipital-temporal
cortex (see Chapter 1), or could reflect activity of neurons within one
cortical region. For the first presentation, many neurons are respon-
sive to that item, as illustrated by intermediate to high response rates
7.3 Models of Implicit Memory 139

Figure 7.4 Models of repetition priming. (A) Activity in individual neurons (circles, with lines
illustrating interneuron connections), as measured by mean firing rate, to the first
presentation of an item (key to the right). The activation timecourse (number of spikes as
a function of time after stimulus onset) is shown for the two neurons on the right. (B) Activity
in the same neurons to the second presentation of an item, as dictated by the fatigue model
(to the left), the sharpening model (in the middle), and the facilitation model (to the right).

(i.e., light gray and white circles, respectively) and the activation
profiles for the two neurons on the right. Figure 7.4B, left, shows
the fatigue model of repetition priming, where a repeated item is
associated with a decrease in the magnitude of activity for all of the
neurons (i.e., the previously light gray and white circles are dark gray
and light gray, respectively). This is called the fatigue model because
it describes what the neurons would do if they were less responsive/
fatigued after being active. Figure 7.4B, middle, shows the sharpening
model of repetition priming, where a repeated item is associated with
a decrease in the magnitude of activity for neurons that were not
maximally active, with the same magnitude of activity for neurons
that were maximally active (i.e., the previously light gray circles are
now black and the white circles are unchanged). This is called the
sharpening model because only the previously most active neurons
are responsive, which means the representation of the previous item
is more spatially restricted/sharper. Figure 7.4B, right, shows the
facilitation model of repetition priming, where a repeated item is
associated with the same magnitude of activity for all neurons, but
all of the activations occur at a faster rate (i.e., the activation time-
courses are compressed in time). An additional model of repetition
140 Implicit Memory

priming has been proposed based on an increase in synchronous

activity (Gotts, Chow & Martin, 2012). However, the increase in the
magnitude of alpha activity described in section 7.2 likely reflects
cortical inhibition, which is a mechanism of reduced cortical activity
rather than a separate model.
There is no evidence that supports the facilitation model of repetition
priming. For instance, ERPs/ERFs do not have more rapid timecourses
for repeated items as compared to new items. This leaves the fatigue
model and the sharpening model as the two viable models of repetition
priming. Fortunately, these models can be distinguished by evaluating
repetition priming effects for neurons that initially produced the highest
magnitude of activity as compared to neurons that initially produced
a lower magnitude of activity. The fatigue model stipulates that repeti-
tion will reduce the magnitude of activity in all cortical neurons, which
predicts the reduction of activity during repetition priming will be highest
in neurons that were initially the most active. For example, if the maxi-
mally responsive neurons (the white circles in Figure 7.4A) have
a magnitude of 10 and less responsive neurons (the light gray circles in
Figure 7.4A) have a magnitude of 6, the fatigue model (Figure 7.4B, left)
might predict a reduction in magnitude by half for all neurons such that
the previous maximally responsive neurons would have a magnitude of 5
and the less responsive neurons would have a magnitude of 3 (i.e., the
maximally responsive neurons would have a relatively larger reduction of
magnitude than less responsive neurons). In contrast, the sharpening
model stipulates that repetition will only reduce the magnitude of activity
in less responsive neurons, which predicts the reduction of activity during
repetition priming will be the highest in neurons that were initially less
active. Using the numbers from the previous example, if the maximally
responsive neurons have a magnitude of 10 and less responsive neurons
have a magnitude of 6, the sharpening model might predict a reduction
by half for only less responsive neurons such that the previous
maximally responsive neurons would still have a magnitude of 10 and
the less responsive neurons would have a magnitude of 3 (i.e., the maxi-
mally responsive neurons would have a relatively smaller reduction in
magnitude than less responsive neurons). One fMRI study evaluated
the fatigue model and the sharpening model of repetition priming by
comparing the relative magnitudes of activity in the ventral occipital-
temporal cortex for different categories of items such as faces and houses
(Weiner, Sayres, Vinberg & Grill-Spector, 2010). In the lateral ventral
occipital-temporal cortex, there was a similar proportional reduction in
the magnitude of activity for categories that initially produced the highest
7.4 Implicit Memory and the Hippocampus 141

magnitude of response and for categories that initially produced a lower

magnitude of response, which supports the fatigue model. In the medial
ventral occipital-temporal cortex, there was a greater proportional reduc-
tion in the magnitude of activity for categories that initially produced
a lower magnitude of response, which supports the sharpening model.
These results suggest that the nature of repetition priming may be distinct
in different cortical regions. As this is the result of only one study, work
along the same lines will be needed in the future to further investigate
these models of repetition priming.

7.4 Implicit Memory and the Hippocampus

There is an abundance of evidence that the hippocampus is associated
with long-term memory (see Chapter 3) and that this region is not
associated with implicit memory. This has been shown in numerous
patient lesion studies, where damage to the medial temporal lobe that
includes the hippocampus typically produces a dramatic impairment in
long-term memory performance and little or no impairment in implicit
memory performance. This has also been shown in numerous fMRI
studies, as long-term memory consistently activates the hippocampus
(Slotnick, 2013b) but implicit memory does not activate this region.
A few studies with unusual tasks and analyses have reported that the
hippocampus might be associated with implicit memory. If this were
true, it would question the evidence-based view that the hippocampus
is associated with only long-term memory. During the associative prim-
ing task, pairs of unrelated words are presented during the study phase
(e.g., ‘cloud–flower’, ‘cave–reason’, ‘table–plane’). During the test phase,
participants are shown intact word pairs with the second word as a stem
(e.g., ‘cloud–flo___’) or rearranged word pairs with the second word as
a stem (e.g., ‘cave–pla___’), and participants complete the word stem as
quickly as possible with the first word that comes to mind. Behavioral
associative priming is reflected by a greater proportion of word stems
completed from the study phase for intact word pairs than for rearranged
word pairs. As this is an indirect and speeded task, such associative
priming task effects could be interpreted as relying on implicit memory.
However, as discussed in Box 7.1, the employment of an indirect task
does not necessarily mean that participants base their responses on only
implicit memory. For the associative memory task, participants could
also complete the word fragment based on long-term memory (i.e., the
first word and stem could have cued recollection of the second word from
the study phase). Such explicit memory contamination during this
142 Implicit Memory

Box 7.1: That task doesn’t map onto that process

Many scientists assume that a task is based on one cognitive process.
In memory research, indirect tasks, which do not require thinking back to
the study phase, are generally assumed to reflect implicit memory, while
direct tasks, which require thinking back to the study phase, are generally
assumed to reflect long-term memory. However, just because a particular
type of task is employed doesn’t necessarily mean that participants are going
to behave in the desired or expected way. Even though participants are not
asked to remember previously presented items during an indirect task, they
might still automatically or intentionally retrieve this information.
Participants have free will and they often do unexpected things. Along the
same lines, even though participants are asked to remember previously
presented items during a direct task, there are still going to be implicit
memory effects. That is, no task is process-pure. To assess the cognitive
processes involved in a task, all possible cognitive strategies that participants
might use to perform a task should be considered. Behavioral analysis and
post-experiment questionnaires should be used to assess the strategies
participants employed during a task. For instance, tasks that are based
largely on implicit memory should have much more rapid response times,
and participants should have no explicit knowledge items were repeated.

indirect task appears likely given that performance on this task has been
correlated with other measures of long-term memory and that patients
with severe amnesia were impaired at this task (Schacter et al., 2004).
Although the contrast of intact word pairs and rearranged word pairs
during an associative priming task produced an increase of activity
within the medial temporal lobe, this was interpreted as reflecting long-
term memory (Badgaiyan, Schacter & Alpert, 2003). Thus, although
there has been an association between the medial temporal lobe (includ-
ing the hippocampus) and associative priming, this can be attributed to
the use of long-term memory during this indirect task.
Unlike the associative priming task, which is thought to be based, in
part, on long-term memory, a strong claim has been made that the
contextual cueing task is based on only implicit memory (Chun &
Jiang, 1998). Figure 7.5 shows a stimulus display for one version of this
task, where participants were instructed to quickly detect the direction
(“left” or “right”) of a rotated T embedded in many rotated Ls with
different orientations. Twelve unique contexts (i.e., the configuration of
Ls) were repeated thirty times during the experiment and cued the target
7.4 Implicit Memory and the Hippocampus 143

Figure 7.5 Contextual cuing stimulus display. Each stimulus display consists of a target
(i.e., a rotated T) within a context (i.e., many rotated Ls). Participants indicate whether the
target is pointing “left” or “right.”

location, while the other half of the contexts were new. Participants were
faster at detecting the direction of the target for repeated contexts than
for new contexts. In a follow-up study, patients with damage to the
medial temporal lobe that included the hippocampus were found to be
impaired on the contextual cuing task, which was interpreted as a link
between the hippocampus and implicit memory (Chun & Phelps, 1999).
However, an alternative reason that these medial temporal lobe patients
were impaired on this task is that the task was associated with long-term
memory. There is evidence for explicit contamination in the original
study that introduced the contextual cuing task (Chun & Jiang, 1998),
as nearly half of the participants in one experiment said they were aware
that the contexts were repeated. An fMRI study of contextual cuing
conducted by a different research group also provided compelling
evidence for explicit contamination during this task (Preston &
Gabrieli, 2008). The same stimulus paradigm depicted in Figure 7.5 was
used, where twelve contexts (i.e., the rotated Ls) were each repeated
twenty times along with new contexts and participants quickly detected
the direction of the target (i.e., the rotated T). After fMRI was complete,
participants were given a surprise test to measure whether they had
long-term memory for the repeated contexts. They were presented
with the twelve repeated contexts and the twelve new contexts and
made “old”–“new” recognition judgements. Recognition memory
144 Implicit Memory

performance was at 58 percent correct, which is above the chance/

guessing rate of 50 percent and indicates that many of the participants
used long-term memory during this task. Contextual cuing perfor-
mance was associated with activity in the perirhinal cortex (i.e., novel
contexts produced greater activity than repeated contexts), which
likely reflected repetition priming or familiarity. This suggests that
the previously described impairment in contextual cuing performance
in patients with medial temporal lobe damage (Chun & Phelps, 1999)
may have been due to damage to the perirhinal cortex rather than
damage to the hippocampus. These findings indicate that the contex-
tual cuing task can reflect both implicit memory and long-term memory
and that there is no convincing evidence that this task is associated with
the hippocampus.
Another fMRI study also claimed to provide evidence of hippocampal
involvement during implicit memory, as measured by eye movements for
faces and scenes (Hannula & Ranganath, 2009). During the study phase,
participants viewed face–scene pairs and assessed whether or not the face
belonged in the place depicted by the scene (e.g., a particular face in
a kitchen scene). During each trial of the test phase, one of the previous
scenes was presented for 1 second followed by a 7-second delay and
participants were instructed to use the scene as a cue to retrieve the
associated face. Then, three of the faces from the study phase (including
the one that matched that scene) were presented and participants
selected the face that they thought matched the scene. In addition
to making an explicit face recognition response (i.e., the button press
indicating the previous face that matched the scene), eye movements
were monitored to probe the viewing time for each face. Viewing time
could be maximal for the face that matched that scene or maximal for
a face that did not match that scene. The key trials were those in which
explicit face recognition was incorrect (i.e., they picked the wrong face
with the button press) and there was higher viewing time for the correct
face than for the incorrect face. It is reasonable to assume that such an
increase in viewing time to the correct face reflected implicit memory, as
long-term memory (as measured by the button press) failed on these
trials. There was hippocampal activity during these trials, but it did not
occur when the three faces were presented and implicit processing could
have occurred. Rather, the hippocampal activity occurred when the scene
was presented alone, before the faces were presented. The major pro-
blem is that there is no basis to argue that implicit memory was operating
when the scene was presented alone. During this period, participants
were instructed to retrieve the face that had been paired with
7.4 Implicit Memory and the Hippocampus 145

that scene, and they presumably retrieved one of the faces from the study
phase. It can be assumed that they retrieved the incorrect face during this
period, since they subsequently made an incorrect button response,
which reflects the process of false memory. As it is known that false
memories can produce activity in the hippocampus (see Chapter 5),
activity during the scene cue can be attributed to false memory rather
than implicit memory. This is supported by the subsequent incorrect face
choice, where participants may have looked at the matching face longer
(because they had a weak memory trace of the face–scene pair) but then
picked a different face that better matched their false memory from the
scene period. Therefore, this study does not provide any compelling
evidence that implicit memory is associated with the hippocampus.
This section has evaluated and discounted cases in which implicit
memory has been associated with the hippocampus. The association
between long-term memory and the hippocampus is based on a massive
amount of evidence, and there is no convincing evidence that implicit
memory is associated with the hippocampus. As discussed in Box 7.2, the
widely held evidence-based view that implicit memory is not associated
with the hippocampus may have been challenged in an effort to achieve
scientific success. If the aim is to find an association between implicit
memory and the hippocampus, future studies will need to make a much
stronger case that their task reflects only implicit memory. However,
based on the wealth of evidence showing that the hippocampus is asso-
ciated with long-term memory but not implicit memory, this is unlikely to
happen.

Box 7.2: One path to success in science

Becoming well known in science is typically a long process. It can take years
or decades to make significant progress on a scientific question. Those that
do become well known receive many benefits including better jobs, more
grant funding, exemplary students, and publications in prestigious journals.
One way in which scientists can increase their chance of success is to work on
topics that are controversial. In this way, they become a central player in
a heated scientific debate. A related way to increase their chance of success is
to try and find evidence that contradicts a commonly held view. However,
the quality of the science that challenges a widely held view is sometimes
questionable. Fortunately, such research will be critically evaluated by a huge
community of scientists and the truth will be revealed.
146 Implicit Memory

7.5 Skill Learning

Proficiency in skills such as playing a musical instrument, martial arts, or
chess takes years of training. Skill learning involves multiple stages
including an early stage that depends largely on long-term memory and
a late stage that depends largely on implicit memory.
Scientific studies of skill learning typically require participants to
repeat a relatively simple task and track the changes in brain activity
over time. In an fMRI skill learning study, participants learned
a sequence of five taps between the tip of the thumb and the tip of the
fingers of their left hand (Ma et al., 2010). For example, one sequence was
5, 2, 4, 3, 5 (fingers were numbered in order from the index finger, 2, to the
little finger, 5). Participants practiced this sequence for 15 minutes
per day for 4 weeks. fMRI was conducted on the first day (before train-
ing), after 2 weeks of practice, and after 4 weeks of practice while
participants alternated between performing the learned sequence or
resting (the control condition) for 2-minute periods. Figure 7.6A shows
that the rate of movement doubled from day 1 to day 14, and there was
little increase in the rate from day 14 to day 28. The relatively flat increase
in performance from week 2 to week 4 indicates that the finger tapping
sequence had been learned well and was presumably based largely on

A B
90 2
Pre-training
80 Week 2
Week 4
Percent signal change

1.5
70
Movement rate

60
1
50

40
0.5
30

20 0
0 7 14 21 28 M1 SMA BG
Day

Figure 7.6 Skill learning behavioral results and fMRI results. (A) Rate of finger tapping
(number of sequences per minute) as a function of training day. (B) The magnitude of fMRI
activity (percent signal change) in the primary motor cortex (M1), the supplementary motor
area/cortex (SMA), and the basal ganglia (BG) before training (pre-training), after 2 weeks of
training, and after 4 weeks of training (key at the top right).
7.5 Skill Learning 147

implicit memory. Figure 7.6B illustrates that in motor processing

regions – the primary motor cortex (M1), the supplementary motor
area/cortex (SMA), and a sub-cortical region called the basal ganglia
(BG) – there was an increase in activity 2 weeks after training followed
by a decrease in activity 4 weeks after training. This study also reported
a progressive decrease in activity over time within the dorsolateral pre-
frontal cortex and in the cerebellum, a region of the brain that has been
associated with motor coordination. Previous studies have observed
similar decreases in activity within the dorsolateral prefrontal cortex
(Floyer-Lea & Matthews, 2005) and the cerebellum (Ungerleider,
Doyon & Karni, 2002) after weeks of training. Of additional relevance,
the hippocampus has been associated with sequence learning on the first
or second day of training, but not after more extensive training (Penhune
& Doyon, 2002; Steele & Penhune, 2010).
Although there are multiple regions associated with sequence learn-
ing, the findings can be interpreted in a straightforward manner.
The decrease in activity within the hippocampus and the dorsolateral
prefrontal cortex with more training likely reflects less dependence on
long-term memory for the sequence (see Chapter 3). Similarly, the
decrease in activity within the cerebellum can be attributed to a lower
degree of motor coordination required with increased practice.
The initial increase in activity within motor processing regions (after 2
weeks) may reflect a shift from explicit control in the dorsolateral
prefrontal cortex to increased processing in lower-level motor processing
regions (Diedrichsen & Kornysheva, 2015). The subsequent decrease in
activity within motor processing regions (from week 2 to week 4) might
reflect more efficient or fluent processing, as occurs with repetition
priming.
Although studies of skill learning have begun to shed light on the
brain basis of this process, this line of research has multiple short-
comings. First, the tasks employed to date, such as finger tapping, are
much more simplistic than skills that are learned in everyday life, such
as martial arts. Second, the amount of practice on these tasks (i.e., no
more than a few weeks) is far less than the amount of practice on tasks
that are learned in everyday life. On this issue, many skill learning
studies investigate brain activity during one session or during a couple
of days of training, but learning during this early period can be assumed
to be contaminated by long-term memory (these studies were not
considered as they are not relevant to this chapter). Third, previous
work has focused on the changes in brain activity associated with motor
skill learning and have ignored cognitive skill learning, such as gaining
148 Implicit Memory

proficiency in chess or video games. For instance, one behavioral study

showed that chess experts can process chess board configurations
unconsciously (Kiesel, Kunde, Pohl, Berner & Hoffman, 2009), which
corresponds to implicit memory for the identity and movement possi-
bilities of the chess pieces. Future studies should employ more realistic
tasks and extensive training to better understand the brain mechanisms
associated with skill learning.

Chapter Summary
• Implicit memory is commonly investigated using repetition priming para-
digms, where participants make speeded responses during indirect tasks.
• Behavioral repetition priming is evidenced by faster reaction times to
old items than new items.
• Repetition priming for familiar items has been associated with
a decrease in fMRI activity within the dorsolateral prefrontal cortex,
the posterior lateral temporal cortex (primarily in the left hemisphere),
and perceptual processing regions.
• Repetition priming for familiar items has been associated with
a decrease in gamma activity and an increase in alpha activity, which
both correspond to a decrease in cortical activity.
• The two viable neural models of repetition priming are the fatigue
model and the sharpening model.
• There is no convincing evidence that implicit memory is associated
with the hippocampus.
• As a skill is learned over time, there is a decrease in activity within the
dorsolateral prefrontal cortex in addition to an initial increase in
activity (within the first 2 weeks) and a subsequent decrease in activity
(from 2 to 4 weeks) within motor processing regions.

Review Questions
How does brain activity differ during repetition priming of familiar items
and repetition priming of unfamiliar items?
Which brain regions have been associated with implicit memory?
How do implicit memory fMRI effects and frequency band effects relate
to one another?
What are the two viable neural models of repetition priming?
Is there convincing evidence that implicit memory is associated with the
hippocampus?
In what way could skill learning paradigms be improved?
Further Reading 149

Further Reading
Koutstaal, W., Wagner, A. D., Rotte, M., Maril, A., Buckner, R. L. &
Schacter, D. L. (2001). Perceptual specificity in visual object priming:
Functional magnetic resonance imaging evidence for a laterality
difference in fusiform cortex. Neuropsychologia, 39, 184–199.
This fMRI study illustrates the reductions in the magnitude of cortical
activity associated with repetition priming for familiar items.
Engell, A. D. & McCarthy, G. (2014). Repetition suppression of face-selective
evoked and induced EEG recorded from human cortex. Human Brain
Mapping, 35, 4155–4162.
This intracranial EEG study illustrates the two temporal effects of
repetition priming for familiar items, a decrease in gamma activity and an
increase in alpha activity.
Grill-Spector, K., Henson, R. & Martin, A. (2006). Repetition and the brain:
Neural models of stimulus-specific effects. Trends in Cognitive Sciences, 10,
14–23.
This paper reviews multiple neural models of repetition priming.
Hannula, D. E. & Ranganath, C. (2009). The eyes have it: Hippocampal activity
predicts expression of memory in eye movements. Neuron, 63, 592–599.
This fMRI paper claims to provide evidence that the hippocampus is
associated with implicit memory, but the effects can be attributed to false
memory.
Ma, L., Wang, B., Narayana, S., Hazeltine, E., Chen, X., Robin, D. A., Fox, P. T. &
Xiong, J. (2010). Changes in regional activity are accompanied with changes
in inter-regional connectivity during 4 weeks motor learning. Brain
Research, 1318, 64–76.
This fMRI study shows how brain activity changes over 4 weeks of
training during skill learning of a finger tapping sequence.
CHAPTER EIGHT

Memory and Other Cognitive Processes

Learning Objectives
• To describe the cognitive processes and brain regions associated with
visual attention.
• To compare the brain regions associated with visual attention to working
memory and long-term memory.
• To describe the cognitive processes and brain regions associated with
visual imagery.
• To compare the brain regions associated with visual imagery to working
memory and long-term memory.
• To list the two primary brain regions associated with language processing
and name two ways in which language processing is relevant to memory.
• To identify the two regions that interact to enhance memory for
emotional information.

Attention is focused on the contents of all explicit memories.

The experience of detailed recollection seems similar to the experience
of vivid imagery. This chapter compares the cognitive processes and
brain regions associated with memory to the cognitive processes and
brain regions associated with attention, imagery, language, and emotion.
Section 8.1 reviews the brain regions that have been associated
with attention, which include sensory processing regions in addition to
dorsolateral prefrontal cortex and parietal cortex control regions. These
regions are similar to the regions that have been associated with working
memory and long-term memory (except for the additional dependence of
long-term memory on the medial temporal lobe; see Chapters 3 and 6).
In section 8.2 of this chapter, the brain regions associated with imagery
are reviewed, which also include sensory processing regions, the dorso-
lateral prefrontal cortex, and the parietal cortex. The cognitive processes
and brain processes associated with visual imagery are compared to the
cognitive processes and brain processes associated with working memory
and long-term memory. Section 8.3 details the regions of the brain
associated with language processing, which include the left inferior
dorsolateral prefrontal cortex and the left posterior lateral temporal
8.1 Attention and Memory 151

cortex. These regions are of relevance to memory studies, which often

use words and meaningful objects as stimuli that have language/concep-
tual representations. The final section, 8.4, considers the brain regions
that have been associated with emotion, which include the amygdala (a
region just anterior to the hippocampus) and the dorsolateral prefrontal
cortex. During memory for emotional information, it appears that the
amygdala interacts with the hippocampus to amplify memory encoding
and memory consolidation. The role of the dorsolateral prefrontal cortex
is also contemplated, as this region has been associated with attention,
memory, language, and emotion. This suggests that the dorsolateral
prefrontal cortex is not exclusively linked to any particular cognitive
process.

8.1 Attention and Memory

Imagine an instructor actively lecturing in the front of a class with a clock
immediately above her head. To avoid insulting her, one of the students
keeps looking at her face and shifts his attention to the clock, checks the
time, and then shifts his attention back to her face. This illustrates the
process of shifting attention in space. One can also pay attention to
different visual features of an item such as its motion (e.g., when someone
is catching a ball) or its color (e.g., when someone is shopping for
clothes). Behavioral research has shown that attention enhances the
processing of items, which is reflected by more accurate and faster
responses. For example, in a widely used paradigm in the field of atten-
tion, participants are presented with a cue at the central fixation point
(e.g., an arrowhead) that directs their attention to either the left visual
field or the right visual field (Posner, 1980). Participants are faster at
detecting a stimulus when it is attended (i.e., when they are cued to attend
to the same visual field as the target) than when the identical stimulus is
unattended (i.e., when they are cued to attend to the opposite visual field
as the target). It is as if attended items are more salient or brighter than
unattended items.
The cognitive neuroscience of attention constitutes a completely sepa-
rate field from the cognitive neuroscience of memory. Attention effects
in the brain are investigated by comparing activity associated with
attended items/locations with the same items/locations when they are
unattended. Attention has been associated with both sensory regions and
control regions of the brain (see Chapter 1). The sensory effects are well
described by the gain model of attention, which stipulates that attention
amplifies the magnitude of brain activity in sensory processing regions.
152 Memory and Other Cognitive Processes

These sensory attention effects in the brain can be assumed to give rise to
the enhanced behavioral attention effects described above. The sensory
effects of attention associated with different locations in the visual field
are typically studied by simultaneously presenting stimuli in both visual
fields. Participants are cued to attend to stimuli in either the right visual
field or the left visual field while always looking at the central fixation
point. Figure 8.1A illustrates a representative stimulus display. In this
example, when one of the two overlapping arrowheads at the fixation
point briefly flashes red (shown in the top panel of the figure) this cues
participants to shift attention to the corresponding visual field.
Participants then maintain attention to the flashing checkerboard stimu-
lus in that visual field (which is illustrated by the dotted circle), ignore the
stimulus in the opposite visual field, and press a button when they detect
a small red square that infrequently occurs within the attended stimulus
(shown in the second panel of the figure). When the other arrowhead at
the fixation point flashes red and points to the opposite hemifield (shown
in the third panel of the figure), it cues participants to shift attention to
the opposite visual field. In this way, participants shift attention between
the stimulus in the right visual field and the stimulus in the left visual field,
and the stimulus in the opposite visual field serves as the unattended
control stimulus. It should be highlighted that the stimulus display is
identical the entire time (except for the infrequent targets), which
ensures that perceptual processing is constant and that changes in brain
activity can be attributed to attention rather than perception.
The sensory effects of attention were illustrated in an fMRI study
where stimuli were presented in the left visual field and the right visual
field (Hopfinger, Woldorff, Fletcher & Mangun, 2001). On each trial,
participants were cued to attend to the left visual field or the right visual
field while always looking at the central fixation point. As shown in
Figure 8.1B, attention to the right visual field versus attention to the
left visual field produced activity in the left extrastriate cortex, while
attention to the left visual field versus attention to the right visual field
produced activity in the right extrastriate cortex. A subsequent fMRI
study using a similar paradigm also reported contralateral attention
effects in early visual regions, including V1, V2, and V3 (Slotnick,
Schwarzbach & Yantis, 2003). Such contralateral attention effects in
visual processing regions (see Chapter 1) are typically observed in
studies of spatial attention. Attention to other features also increases
the magnitude of activity in the corresponding feature processing
regions. For instance, attention to color produced activity in the
color processing region within the ventral visual processing stream
8.1 Attention and Memory 153

B C

Right > Left Left > Right

Figure 8.1 Spatial attention paradigm and fMRI results. (A) Attention stimulus display with
two overlapping arrowheads at the central fixation point and a flashing checkerboard stimulus
within each visual field. When one arrowhead briefly turns red, participants shift attention to
the corresponding visual field/stimulus (illustrated by the dotted circle). Participants press
a button when they detect a small red square within the attended location/stimulus and ignore
the unattended location/stimulus. (B) Contralateral attention activity in early visual regions
(axial view, occipital pole at the bottom). The contrast of attention to the right visual field and
attention to the left visual field (Right > Left) produced activity in the left extrastriate cortex (in
purple/cyan), while the contrast of attention to the left visual field and attention to the right
visual field (Left > Right) produced activity in the right extrastriate cortex (in red/yellow).
(C) Attention control activity in the dorsolateral prefrontal cortex (the rightmost activation)
and the parietal cortex (the leftmost activation) of the right hemisphere (in purple/yellow;
lateral-posterior view, occipital pole to the left). (A black and white version of this figure will
appear in some formats. For the color version, please refer to the plate section.)
154 Memory and Other Cognitive Processes

(see Chapter 1; Liu, Slotnick, Serences & Yantis, 2003). Attention

effects were illustrated in another fMRI study that used moving dots as
stimuli (Thakral & Slotnick, 2009). While always looking at the central
fixation point, participants viewed a field of dots that moved toward
the fixation point for 14 seconds and either attended to the moving dots
and detected when the dots briefly slowed down (in the attention condi-
tion) or perceived the moving dots without any task (in the perception
condition). A comparison between the attention periods and the percep-
tion periods revealed sensory activity in the motion processing region
(see Chapter 1). As shown in Figure 8.1C, this contrast also produced
activity within dorsolateral prefrontal cortex and parietal cortex
control regions. The control regions of attention consistently include
the dorsolateral prefrontal cortex and the parietal cortex (Corbetta &
Shulman, 2002).
Working memory paradigms include a study phase when to-be-
remembered stimuli are presented, a delay period when the stimuli are
actively maintained, and a test phase when a stimulus is presented and
participants decide whether or not it was from the study phase (see
Chapter 6). The process of working memory is reflected during the
delay period. Working memory has been associated with activity within
sensory processing regions and activity in dorsolateral prefrontal cortex
and parietal cortex control regions, which are the same regions that have
been associated with attention.
Working memory for information in either the left visual field or the
right visual field has been associated with activity in contralateral early
visual regions (see Chapter 6). These working memory contralateral
visual sensory effects mirror the contralateral visual sensory effects that
have been associated with spatial attention.
The overlap between dorsolateral prefrontal cortex and parietal cortex
control regions during working memory and attention has been used to
make the case that these cognitive processes are linked (Awh, Vogel &
Oh, 2006; Gazzaley & Nobre, 2012). One study compared the pattern
of fMRI activity during similar spatial working memory and spatial
attention paradigms (Ikkai & Curtis, 2011). During each trial of the
working memory paradigm, a single spatial location in either the left
visual field or the right visual field was maintained for 7.5 to 13.5 seconds
during the delay period. During each trial of the attention paradigm,
a cue directed attention to the left visual field or the right visual field for
7.5 to 13.5 seconds before the onset of the target stimulus. Both spatial
working memory and spatial attention produced similar patterns of
activity in the dorsolateral prefrontal cortex and the parietal cortex
8.1 Attention and Memory 155

during the delay/sustained attention period. This is not surprising given

that the working memory and attention paradigms were designed to be so
similar. However, working memory paradigms typically involve mainte-
nance of multiple items or spatial locations and use more complex stimuli
(e.g., faces and houses) during the delay period (see Chapter 6), while
attention paradigms typically have a short delay between the cue and the
target and use relatively simple stimuli (e.g., a checkerboard pattern).
Still, that these paradigms could be easily manipulated to map onto one
another and a similar pattern of control region activity was observed
suggest that working memory and attention are similar cognitive
processes.
Future studies that aim to investigate the link between working
memory and attention should employ paradigms that are representa-
tive of both processes (without distorting the paradigms too much)
in the same participants (rather than comparing the brain regions
associated with these cognitive processes in different groups of parti-
cipants). Of importance, one can characterize any working memory
paradigm, even one with multiple spatial locations or complex stimuli,
as requiring sustained attention to the contents of working memory.
As such, working memory is intimately related to the process of
sustained attention.
Long-term memory paradigms include a study phase, where a list of
to-be-remembered stimuli are presented, and a test phase, where old
and (usually) new stimuli are presented and participants make
“old”–“new” recognition judgments and/or context memory judg-
ments (see Chapters 1 and 3). Episodic memory and item memory,
two types of long-term memory, have been associated with activity
within sensory processing regions, the dorsolateral prefrontal cortex,
the parietal cortex, and the medial temporal lobe. Except for the
medial temporal lobe, these are the same regions that have been
associated with attention. It has been proposed that attention may
operate on the internal representation of an item during retrieval
(Wagner, Shannon, Kahn & Buckner, 2005; Cabeza, Ciaramelli,
Olson & Moscovitch, 2008). For example, when someone recalls
where they put their keys before leaving the house, they selectively
attend to the location of the keys in this memory representation. This
example illustrates that the brain regions associated with long-term
memory may, to some degree, reflect the process of attention.
There is evidence that long-term memory and attention are associated
with the same sensory effects. One long-term memory fMRI-ERP study
used abstract shapes as stimuli (Slotnick, 2009b). During the study phase,
156 Memory and Other Cognitive Processes

Figure 8.2 Spatial memory fMRI and ERP results. (A) Contralateral memory fMRI activity in
early visual regions (posterior view). The contrast of accurate memory for items previously on the
left and accurate memory for items previously on the right (old-left-hit > old-right-hit) produced
activity in the right extrastriate cortex (in red), while the contrast of accurate memory for items
previously on the right and accurate memory for items previously on the left (old-right-hit > old-
left-hit) produced activity in the left extrastriate cortex (in blue; key at the bottom). (B)
Contralateral memory ERP activity, with occipital regions of interest (ROIs) and temporal ROIs
demarcated (electrode locations are shown by small red discs; key at the center, in microvolts).
Left, topographic map associated with accurate memory for items previously on the right versus
accurate memory for items previously on the left (old-right-hit – old-left-hit) at 154 milliseconds
after stimulus onset (posterior view) and the corresponding occipital cortex dipole source in the
left hemisphere (in blue; coronal view). Right, topographic map associated with accurate
memory for items previously on the left versus accurate memory for items previously on the right
(old-left-hit – old-right-hit) at 180 milliseconds after stimulus onset (posterior view) and the
corresponding occipital cortex dipole source in the right hemisphere (in red; coronal view).
(A black and white version of this figure will appear in some formats. For the color version,
please refer to the plate section.)

shapes were presented to the left or right of the central fixation point.
During the test phase, old and new shapes were presented at the fixation
point, and participants classified each item as “old and previously on the
left,” “old and previously on the right,” or “new.” Participants were
encouraged to use visual strategies rather than verbal strategies (e.g., to
visualize the item as it was previously presented, rather than remember-
ing the verbal label “left” or “right”). As shown in Figure 8.2A, the fMRI
contrast of accurate memory for items previously presented in the left
8.1 Attention and Memory 157

visual field (old-left-hits) and accurate memory for items previously

presented in the right visual field (old-right-hits) produced activity in
the right extrastriate cortex, while the contrast of accurate memory for
items previously presented in the right visual field (old-right-hits) and
accurate memory for items presented in the left visual field (old-left-
hits) produced activity in the left extrastriate cortex. Figure 8.2B shows
that the same pattern of contralateral visual activity was observed using
ERPs between 100 to 200 milliseconds after stimulus onset, as illu-
strated by the topographic maps and corresponding dipole source loca-
tions (see Chapter 2). This memory ERP effect corresponds to the
contralateral P1 effect that has consistently been reported in ERP
studies of spatial attention (i.e., an increase in the magnitude of con-
tralateral activity that occurs 100 to 200 milliseconds after stimulus
onset; Hopfinger et al., 2001). A subsequent fMRI study that used the
same paradigm reported the identical pattern of contralateral early
visual region activity during long-term memory encoding (Thakral &
Slotnick, 2013). These long-term memory contralateral visual sensory
effects are similar to the contralateral visual sensory effects that have
been associated with spatial attention.
As illustrated in Figure 8.3, a meta-analysis of thirty-six studies
found that the same dorsolateral prefrontal cortex and parietal cortex
control regions were associated with attention, working memory
(which is relevant to the preceding discussion), and episodic/long-
term memory retrieval (Naghavi & Nyberg, 2005). In a meta-analysis
of ninety-three fMRI studies, the same parietal regions that have been
associated with attention were also associated with long-term memory
encoding (Uncapher & Wagner, 2009). A recent fMRI study conducted
a detailed analysis of the degree of overlap between attention activity
and long-term memory activity within the parietal lobe (Hutchingson
et al., 2014). They conducted an analysis on individual participants (rather
than averaging activity across participants or looking for consistent activity
across studies) and found that long-term memory produced activity
in exactly the same regions that were associated with attention. These
findings provide compelling evidence that the process of attention is
engaged during long-term memory. Such individual participant results
are particularly compelling because they are not limited by the variability
that is introduced when the analysis is based on separate groups of parti-
cipants. That is, individual participant results assess whether the exact
same brain regions are associated with two different cognitive processes,
while analyzing separate groups of participants blurs activations such that
they might appear similar when they are actually distinct.
158 Memory and Other Cognitive Processes

A Attention

B Working memory

C Episodic retrieval

Figure 8.3 Meta-analysis of control region activity associated with attention, working
memory, and episodic memory retrieval. (A) Dorsolateral prefrontal cortex and parietal cortex
activations associated with attention (lateral views, occipital poles toward the center).
(B) Dorsolateral prefrontal cortex and parietal cortex activations associated with working
memory. (C) Dorsolateral prefrontal cortex and parietal cortex activations associated with
episodic memory retrieval.

Although attention appears to operate during long-term memory,

long-term memory is also associated with activity in the medial temporal
lobe. As such, long-term memory is a distinct cognitive process, with
attention being a separate process that operates to enhance processing of
8.2 Imagery and Memory 159

the internal memory representation. Future work is needed to shed light

on the nature of the relationship between attention and long-term mem-
ory by directly comparing the activity associated with these processes
across the entire brain on an individual participant basis.

8.2 Imagery and Memory

Stephen Kosslyn, a brilliant cognitive neuroscientist who studies ima-
gery, has long been interested in how people answer the question, “What
shape are a German Shephard’s ears?” Most people report creating
a visual mental image of a German Shephard and then “looking” at the
shape of its ears. Visual imagery has been shown to activate the same
sensory regions of the brain as visual perception, including V1 and
extrastriate cortex regions but also depends on dorsolateral prefrontal
cortex and parietal cortex control regions (Kosslyn, Ganis & Thompson,
2001; Pearson, Naselaris, Holms & Kosslyn, 2015). One fMRI study
aimed to compare the sensory regions and control regions associated
with visual perception, visual imagery, and visual attention (Slotnick,
Thompson & Kosslyn, 2005). As shown in Figure 8.4A, left, during
the visual perception condition, flashing checkerboard wedges (which
are known to activate early visual regions) rotated around the central
fixation point and participants identified whether a small red square
that infrequently flashed was “inside” or “outside” of the wedges.
As shown in Figure 8.4A, right, during the visual imagery condition,
only the outer edges of the wedges were shown and participants were
instructed to vividly imagine the entire flashing wedges (i.e., imagine the
wedges in as much visual detail as possible) while making the same
“inside”–“outside” judgments. The visual attention control condition
used the same stimulus as the imagery condition, but participants were
instructed not to imagine the flashing wedges and identified whether the
red square was in the “left” visual field or the “right” visual field.
Figure 8.4B, left, shows the activity (i.e., retinotopic maps) in early visual
regions associated with perception for one representative participant.
The term retinotopic map refers to activations in early visual regions
where adjacent locations in the visual field are mapped onto adjacent
locations on the cortex (which is the way that the visual field maps
onto the retina of each eye). Different locations in the visual field are
illustrated by different colors, and early visual regions are labeled (each
black line indicates the border between adjacent visual regions).
Figure 8.4B, middle and right, shows activity associated with imagery
and attention, respectively. Critically, the retinotopic map associated
160 Memory and Other Cognitive Processes

Time

B Perception Imagery Attention

C
Perception + Imagery
Imagery + Attentiom

Figure 8.4 Visual perception, imagery, and attention paradigms and fMRI results. (A) Left,
perception stimulus display, with flashing checkerboards rotating around the central fixation
point. Right, imagery and attention stimulus display with only the outer arcs of the flashing
checkerboards rotating around the central fixation point. During the perception and imagery
condition, participants determined whether a briefly flashed small red square was “inside”
or “outside” of the stimuli. During the attention condition, participants determined whether
the small red square was in the “left” visual field or the “right” visual field. (B) Perception,
imagery, and attention retinotopic maps for a representative participant (posterior view;
colors correspond to different spatial locations in the visual field as shown by the semi-circle
8.2 Imagery and Memory 161

with imagery, as compared to the retinotopic map associated with atten-

tion, was much more similar to the retinotopic map associated with
perception. The regions in the imagery map that were not observed in
the attention map are demarcated by the cyan ovals. These findings were
consistent across the participants in the study and illustrate that imagery
effects in early visual regions are similar to, albeit weaker than, percep-
tion effects. With regard to sensory activation, visual imagery has been
characterized as a weak form of perception (Pearson et al., 2015).
Figure 8.4C shows the regions of the brain that were associated with
sustained periods of perception and imagery, which included visual pro-
cessing regions (in green), and regions of the brain that were associated
with sustained periods of imagery and attention (in orange), which
included the dorsolateral prefrontal cortex and the parietal cortex.
These findings indicate that visual perception and visual imagery are
associated with overlapping activity in sensory regions, which is consis-
tent with the retinotopic map findings, and that visual imagery and visual
attention are associated with the same control regions.
Visual imagery and visual working memory are inseparable cognitive
processes. During visual working memory paradigms, participants are
typically presented with stimuli and then maintain a mental represen-
tation of the stimuli during a delay period (see Chapter 6). During
visual imagery paradigms, participants are typically presented with
stimuli and then imagine the stimuli during a delay period. Both
cognitive processes activate visual sensory processing regions,
including V1, both cognitive processes are associated with dorsolateral
prefrontal cortex and parietal cortex control regions, and neither
cognitive process depends on the medial temporal lobe. Despite the
similarity between these cognitive processes, the body of literature on
working memory is separate from the body of literature on imagery.

Caption for Figure 8.4 (cont.)

key between the perception and imagery retinotopic maps). Early visual regions are labeled
(in black) and cyan ovals show the regions where imagery produced greater retinotopic
activity than attention. The repeating patters of colors (e.g., yellow to red to yellow to red in
the upper left hemisphere) correspond to repeated visual field representations in early visual
areas (e.g., the lower right quadrant in the visual field has a unique representation in dorsal
V1, V2, and V3 of the left hemisphere). (C) Activity associated with both perception and
imagery (in green) and activity associated with both imagery and attention (in orange; key to
the right). (A black and white version of this figure will appear in some formats. For the color
version, please refer to the plate section.)
162 Memory and Other Cognitive Processes

The only difference between these cognitive processes seems to be in

how the stimulus representation is described during the delay period.
In the working memory literature, the stimulus is described as being
maintained, while in the imagery literature, the stimulus is described
as being imagined. Maintaining visual information seems like another
way of saying imagining visual information, and these cognitive func-
tions have been associated with the same sensory regions and control
regions. These striking similarities in paradigms, cognitive processes,
and brain regions suggest that working memory is simply another label
for imagery (see Chapter 6). If a convincing case is to be made that
working memory is distinct from imagery, the same stimulus paradigm
with either working memory instructions or imagery instructions would
need to produce activity in distinct brain regions. This seems unlikely,
but it is a topic of future research.
Long-term memory requires retrieval of information that was not
kept in mind since the initial encoding event, whereas visual imagery
requires keeping a stimulus in mind that was just presented. These
cognitive processes are not identical, and the brain regions associated
with these processes are not the same. Only long-term memory has
been associated with activity in the medial temporal lobe. However,
visual long-term memory and visual imagery can both reflect detailed
mental representations. Furthermore, both long-term memory and
imagery have been associated with activity in sensory processing
regions, the dorsolateral prefrontal cortex, and the parietal cortex
(see Chapter 3).
One fMRI study aimed to investigate the common and distinct brain
regions associated with visual recollection and visual imagery (Slotnick,
Thompson & Kosslyn, 2012). During the familiarization phase, line
drawings of objects (e.g., a zebra and a feather) were presented. For
each object, participants were instructed to memorize the object in detail,
press a button such that the object disappeared and then imagine the
object exactly as it appeared, and then press a button such that the object
reappeared and correct their mental image. The list of objects was
repeated in this way three times. Such a familiarization phase is common
in imagery studies and aims to ensure mental images are vivid/visually
detailed. Participants then completed study and test phases, which
are common in memory studies. Previously familiarized objects were
presented during each study phase, and participants were instructed
to remember each item. During the memory test phase, word labels
were presented corresponding to old objects, new objects, or
control responses (‘left’, ‘center’, or ‘right’). Participants made
8.2 Imagery and Memory 163

“remember”–“know”–“new” judgments for old and new words or

pressed the corresponding (left, center, or right) button for control
words. During the imagery test phase, the same type of word labels were
presented, but participants were instructed to imagine the corresponding
object as vividly as possible and made “high vividness”–“medium
vividness” –“low vividness” judgments for old words and new words or
pressed the corresponding button for control words. Both memory-old
-“remember” responses and imagery-old-“high vividness” responses, as
compared to control responses (which required word processing and
motor processing), produced activity in the same regions of the dorso-
lateral prefrontal cortex, the parietal cortex, and visual sensory regions,
including V1. The comparisons between memory-old-“remember”
responses and imagery-old-“high vividness” responses produced greater
activity in visual sensory regions, which suggests that the memory task
was associated with more detailed visual representations. Taken
together, the results of this study suggest that long-term memory and
imagery share many of the same brain processes, but that these cognitive
processes are not identical.
One relatively new line of research in the field of memory has
focused on the brain regions associated with autobiographical mem-
ory for past events (see Chapter 3) and imagined autobiographical
memory for future events. For instance, the following description was
given when a participant was asked to imagine an autobiographical
memory 5 years in the future after seeing the cue word ‘dress’:
“My sister will be finishing . . . her undergraduate education,
I imagine some neat place, Ivy league private school . . . it would be
a very nice spring day and my mom and dad will be there . . .” (for the
full description, see Addis, Wong & Schacter, 2007, p. 1375). In this
fMRI study, both autobiographical memory for past events and ima-
gined autobiographical memory for future events produced activity in
visual regions, the dorsolateral prefrontal cortex, the parietal cortex,
and the medial temporal lobe (including the hippocampus). Although
these results could be taken to suggest that imagery is associated with
activity in the medial temporal lobe, this is not a standard imagery
task. Imagining future autobiographical events involves retrieval of past
information (e.g., parents, in the example above), as acknowledged by
the authors, in addition to memory encoding of the constructed event,
both of which would be expected to produce activity in the medial
temporal lobe (see Chapter 3). Thus, such medial temporal lobe activa-
tions can be attributed to the process of long-term memory encoding
and retrieval rather than imagery.
164 Memory and Other Cognitive Processes

Motor Angular
Cortex Gyrus

Visual
Cortex

Broca’s Wernicke’s
Area Area

Figure 8.5 Language processing regions (lateral view, occipital pole to the right).
Regions (in different shades of gray) are labeled and arrows indicate the direction of
information flow between regions.

8.3 Language and Memory

In the late nineteenth century, a patient with a lesion in the left inferior
dorsolateral prefrontal cortex was reported to have an isolated word
production deficit (i.e., the patient could not speak but could understand
speech), and another patient with a lesion in the left posterior superior
temporal cortex was reported to have an isolated comprehension deficit
(i.e., the patient could not understand speech but could speak). These
areas were named Broca’s area and Wernicke’s area, respectively, after
the scientists who reported the findings. Figure 8.5 illustrates the location
of these regions, along with other regions associated with language
processing (Price, 2000). Word production has been associated with
Broca’s area, which is just inferior and anterior to the motor cortex,
while visual word comprehension has been associated with the visual
cortex, the angular gyrus (within the inferior parietal cortex), and
Wernicke’s area. The association of Broca’s area and Wernicke’s area
with language production and language comprehension, respectively, is
the classic model of language processing. However, more recent evidence
has indicated that language production and language comprehension are
both associated with Broca’s area and Wernicke’s area. Of particular
relevance to memory research, processing a word’s meaning, which is
referred to in the field of language as semantic processing, can activate
8.3 Language and Memory 165

Broca’s area, Wernicke’s area, the angular gyrus, and more anterior
superior temporal cortex (Price, 2000; Vigneau et al., 2006; Friederici &
Gierhan, 2013). The key point is that semantic/conceptual processing has
been associated with activity in the left inferior dorsolateral prefrontal
cortex and the left posterior superior temporal cortex.
Language processing, and more specifically word processing, is an
important aspect of memory studies because words are often used as
stimuli and meaningful objects are associated with semantic proces-
sing. For example, when a person sees a picture of a sheep, they not
only process the visual aspects of the animal but know what it sounds
like (“bah”), where it can be found (a farm), and how it can be useful
to humans (to make wool). This type of semantic or conceptual
representation is associated with activity in language processing
regions including the left inferior dorsolateral prefrontal cortex and
the left posterior superior temporal cortex (i.e., Broca’s area and
Wernicke’s area, respectively). Multiple examples of language proces-
sing during memory have already been touched on in this book.
Semantic memory, which refers to knowledge of facts that are learned
through repeated exposure over a long period of time, has been
associated with activity in the left dorsolateral prefrontal cortex (see
Chapter 3). False memory often occurs due to a verbal label that is
shared with true memories and has been associated with activity in
the left dorsolateral prefrontal cortex and the left posterior superior
temporal cortex (see Chapter 5). Conceptual priming effects have also
been associated with the left dorsolateral prefrontal cortex and the
left posterior superior temporal cortex (see Chapter 7). When stimuli
can be processed semantically/conceptually, activity in language
processing regions is often observed.
Although some dorsolateral prefrontal cortex activations that
have been observed during memory tasks can be attributed to language
processing, this is not always the case. Retrieval-induced forgetting has
been associated with the right dorsolateral prefrontal cortex, and this
region is thought to mediate inhibition rather than language processing
(see Chapter 5). Semantic memory might activate a region of the left
dorsolateral prefrontal cortex that is more anterior than Broca’s area
(Gabrieli, Poldrack & Desmond, 1998). These findings underscore that
dorsolateral prefrontal cortex activity does not necessarily reflect
language processing. Still, the large majority of stimuli that are
employed in memory studies do have semantic/conceptual representa-
tions, which is why memory often activates regions associated with
language processing. As discussed in Box 8.1, such detailed
166 Memory and Other Cognitive Processes

Box 8.1: The benefits of understanding other cognitive

processes
Cognitive neuroscientists who study memory benefit from having a detailed
understanding of other cognitive processes. Attention is often equated to
enhanced processing during memory encoding, imagery is often equated to
detailed sensory processing during memory construction, and language is
often equated to verbal encoding or retrieval. However, attention, imagery,
and language are all separate and rich topics within cognitive psychology
and cognitive neuroscience. As illustrated in this chapter, understanding
these cognitive processes and their associated brain regions can provide
novel insight into the mechanisms underlying memory. For instance, when
the same brain region is associated with different cognitive processes, this
suggests there may be a common process. One important line of future work
will be to identify the common and disparate brain mechanisms underlying
memory and other cognitive processes.

comparisons between activity associated with memory and activity

associated with other cognitive processes provide insight into the
brain mechanisms underlying memory.

8.4 Emotion and Memory

The field of affective neuroscience focuses on the brain regions associated
with emotional processing, and this field is largely distinct from the field of
cognitive neuroscience. However, these fields overlap when stimuli are
employed that evoke emotions (e.g., fear, disgust, or happiness) in studies
that employ cognitive neuroscience techniques. Processing of emotional
stimuli such as a picture of a spider, a skull, or a gun, as compared to
processing neutral stimuli, activate many regions of the brain including the
amygdala, the orbitofrontal cortex (the part of the frontal cortex just above
the eyes/orbits), and the dorsolateral prefrontal cortex (Lindquist, Wager,
Kober, Bliss-Moreau & Barrett, 2012). Figure 8.6 shows the location of the
amygdala, which is a small region just anterior to the hippocampus.
The amygdala is the core region of the brain associated with processing
emotional stimuli and is thought to serve as a hub that is broadly connected
to many regions of the brain (Pessoa & Adolphs, 2010; Lindquist et al.,
2012). Box 8.2 discusses the role of the dorsolateral prefrontal cortex
during emotional processing, as compared to other types of cognitive
processing.
8.4 Emotion and Memory 167

Figure 8.6 The amygdala and the hippocampus. The amygdala (in dark gray) and the
hippocampus (in light gray) in each hemisphere are shown within a semi-transparent brain
(lateral-anterior view, occipital pole to the right).

Box 8.2: The dorsolateral prefrontal cortex is associated

with many cognitive processes
The dorsolateral prefrontal cortex has been associated with memory,
attention, imagery, language, and emotion. Scientists who conduct research
on one of these topics often assume the dorsolateral prefrontal cortex is
primarily associated with the cognitive process they investigate. Although
one region of the left inferior dorsolateral prefrontal cortex (i.e., Broca’s area)
appears to be specialized for language processing, the dorsolateral prefron-
tal cortex is huge and activity associated with these different cognitive
processes appears to be largely overlapping. What does the dorsolateral
prefrontal cortex actually do? One possibility is that this region mediates
a common cognitive process or processes. For instance, activity in this region
might reflect selecting information that is processed in other regions of the
brain, such as the remembered item, the attended stimulus, the imagined
stimulus, the to-be-spoken word, or the type of emotion. The dorsolateral
prefrontal cortex might also reflect inhibition of information that is irrelevant
in other regions of the brain. Selection of relevant material and inhibition of
irrelevant material are related, as they both allow performance of a particular
goal (i.e., focusing on information of relevance). Moreover, both selection
and inhibition may reflect the more fundamental process of attention, which
168 Memory and Other Cognitive Processes

Box 8.2: (cont.)

may be shared by all of these cognitive processes. Another possibility is that
the dorsolateral prefrontal cortex is a flexible region that reorganizes its
function to reflect the rules required to perform each task (Miller,
Freedman & Wallis, 2002). If that were the case, the function of this region
could differ for each type of cognitive process. These broad functions of the
dorsolateral prefrontal cortex are not exclusive. The dorsolateral prefrontal
cortex may be involved in selection/inhibition that is common to all of these
cognitive functions and may also be involved in learning the task rules that
are unique to each of these cognitive functions. The key point is that activity
in the dorsolateral prefrontal cortex should not be assumed to reflect only
one cognitive process.

Since emotional information is associated with increased mental

processing and brain processing, it should not be surprising that
memory for emotional stimuli is typically superior to memory for
neutral stimuli (except for situations where emotional information is
overwhelming such as during a traumatic event, which can impair
memory). Processing of emotional stimuli is thought to enhance long-
term memory, in part, through the interaction of the amygdala and the
hippocampus (Phelps, 2004). The anatomic proximity of the amygdala
and the hippocampus supports the view that these regions interact, as
the amygdala seamlessly extends from the hippocampus (it is often
difficult to distinguish these regions from one another, even on a high
resolution MRI of an individual participant). The amygdala appears
to enhance processing in the hippocampus during both encoding of
emotional stimuli and consolidation of emotional stimuli. In one
fMRI study, participants were presented with photographs of emo-
tional stimuli (half positive, such as a cute kitten, and half negative,
such as a forest fire) mixed with neutral stimuli during the study phase
and then during the test phase they made “old”–“new” recognition
judgments (Mickley Steinmetz, Schmidt, Zucker & Kensinger, 2012).
Subsequently remembered items (old-hits) versus subsequently forgot-
ten items (old-misses) produced greater activity for emotional items
than neutral items in the hippocampus, the amygdala, the orbitofrontal
cortex, and the dorsolateral prefrontal cortex. These results support the
view that emotional memories, as compared to neutral memories, are
associated with enhanced processing in multiple regions of the brain.
Further Reading 169

Chapter Summary
• Visual attention increases activity in visual sensory regions and is also
associated with activity in dorsolateral prefrontal cortex and parietal
cortex control regions.
• Visual working memory is associated with the same sensory regions
and control regions associated with attention, which likely reflects
attention to the contents of working memory.
• Visual long-term memory is associated with the same regions asso-
ciated with visual attention in addition to the medial temporal lobe,
which indicates this cognitive process is distinct from attention.
• Imagery and working memory share the same cognitive operations
and are associated with the same brain regions (i.e., the sensory cortex,
the dorsolateral prefrontal cortex, and the parietal cortex), which
suggests these are the same cognitive process.
• Semantic memory, false memory, and conceptual repetition priming
have been associated with activity in language processing regions
within the left inferior dorsolateral prefrontal cortex (i.e., Broca’s
area) and the left posterior superior temporal cortex (i.e., Wernicke’s
area).
• Memory for emotional information is thought to be enhanced through
the interaction of the amygdala and the hippocampus.

Review Questions
Which brain regions have been associated with visual attention and visual
working memory?
How do the brain regions associated with visual attention and visual long-
term memory differ?
Are imagery and working memory different cognitive processes?
What are the two primary brain regions associated with language
processing?
Which brain region interacts with the hippocampus during memory for
emotional information?

Further Reading
Ikkai, A. & Curtis, C. E. (2011). Common neural mechanisms supporting spatial
working memory, attention and motor intention. Neuropsychologia, 49,
1428–1434.
170 Memory and Other Cognitive Processes

This fMRI investigation shows a similar pattern of activity in the

dorsolateral prefrontal cortex and the parietal cortex during spatial
working memory and spatial attention.
Slotnick, S. D., Thompson, W. L. & Kosslyn, S. M. (2012). Visual memory and
visual mental imagery recruit common control and sensory regions of the
brain. Cognitive Neuroscience, 3, 14–20.
This fMRI study shows activity in the same regions of the visual sensory
cortex, the dorsolateral prefrontal cortex, and the parietal cortex during
visual long-term memory and visual imagery.
Friederici, A. D. & Gierhan, S. M. (2013). The language network. Current
Opinion in Neurobiology, 23, 250–254.
This review paper highlights the regions of the brain associated with
language processing, which include the same regions that have been
associated with semantic/conceptual processing in memory studies.
Mickley Steinmetz, K. R., Schmidt, K., Zucker, H. R. & Kensinger, E. A. (2012).
The effect of emotional arousal and retention delay on
subsequent-memory effects. Cognitive Neuroscience, 3, 150–159.
The findings of this fMRI paper support the hypothesis that the
amygdala and the hippocampus interact during memory for emotional
stimuli.
Figure 1.4 Brain regions associated with memory. Each region is shown within red ovals
and labeled. (A) Lateral view of the right hemisphere oriented with the occipital pole to the
left. Cortical surface gyri and sulci in this figure and all subsequent figures are shown in light
and dark gray. (B) Coronal view corresponding to the position in the lateral view indicated by
the dashed vertical line. (C) Axial view corresponding to the position in the lateral view
indicated by the dashed horizontal line.
Figure 1.8 Sensory brain regions of interest. Left, lateral view of the left hemisphere
(occipital pole to the right). Right, inferior view of the left hemisphere (occipital pole at the
bottom). Visual sensory regions (within red ovals) are labeled according to the type of
processing (with the name of each region in parentheses). The arrows (in red) illustrate the
where pathway and the what pathway. Non-visual sensory regions are also illustrated
(within blue ovals) and labeled.
Figure 1.9 Sensory fMRI activity associated with perception and memory. (A) fMRI activity
associated with visual perception (axial view, occipital pole at the bottom). (B) fMRI activity
associated with visual memory (arrow indicates extrastriate cortex). (C) fMRI activity
associated with perception of sounds. (D) fMRI activity associated with memory for sounds
(arrow indicates auditory sensory cortex).
 A B

Item Memory Source Memory

0.1 0.1
Remember Each Shape Remember Each Shape Source Memory
and Side of Screen and Side of Screen

% Signal Change

% Signal Change
Item Memory
Correct Rejection
0 0

Time Time

–0.1 –0.1

Old or New? Right Side or Left Side?

Time Time

Figure 1.10 Item memory and source memory paradigm and fMRI results. (A) Left, illustration of item memory task. Right, illustration of source
memory task. (B) Bottom, fMRI activity associated with source memory (in red) and item memory (in yellow) in the dorsolateral prefrontal cortex and
the parietal cortex (axial view, occipital pole at the bottom). Top, the magnitude of activity (in percent signal change) associated with each event type
extracted from the two circled dorsolateral prefrontal cortex activations (key at the top right).
 A

B
379
378
377
376
375
374
0 2 4 6 8 10 12 14
TIME (sec)

C
PERCENT SIGNAL CHANGE

EXTRASTRIATE
.5

L.PREFRONTAL

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
TIME (SEC)

Figure 2.1 MRI scanner and fMRI results. (A) MRI scanner with a participant’s legs (covered
by a sheet) protruding from the bore. (B) Left, one participant’s fMRI activity associated with
word stem completion (more significant activity is shown in yellow; axial view, occipital pole
at the bottom). Extrastriate cortex activity is shown at the bottom and dorsolateral prefrontal
cortex activity is shown at the top. Right, activation timecourse (intensity as a function of
time after stimulus onset, in seconds) extracted from the left dorsolateral prefrontal cortex
activation. The cyan square represents the stimulus period. (C) Activation timecourses
(percent signal change as a function of time after stimulus onset) extracted from the
extrastriate cortex and the left dorsolateral prefrontal cortex of another participant.
 A

B
Rscene – new

anterior

left right

posterior
–0.4 to 2.1 –1.5 to 2.5 0.1 to 2.9 –0.6 to 2.8

500–800 800–1100 1100–1400 1400–1900

Figure 2.2 ERP setup and results. (A) ERP setup that includes a comfortable chair, a 128-
channel electrode cap, and amplifiers (to the right of the chair). (B) ERP topographic maps
(superior views, occipital poles at the bottom; key to the left) associated with remembering
a word was previously paired with a scene versus correctly rejecting new words as a function
of time period (in milliseconds, shown at the bottom below each topographic map).
Electrodes are shown as small black dots (more significant activity is shown in red; voltage
range is shown immediately below each topographic map).
 A B

C D 1.0
0.9
Hit rate (moving-stationary judgment)

ns
0.8
0.7
0.6
Target
0.5
0.4
0.3
0.2
0.1
0
No TMS MT TMS No TMS MT TMS
Moving items Stationary items

Figure 2.6 TMS setup and fMRI guided TMS results. (A) TMS system that includes
a stimulation coil (at the top left). (B) TMS coil positioned over motion processing region MT
of a participant. (C) fMRI activity associated with motion perception (in red/yellow) for one
participant (partial lateral view, occipital pole to the left). The bottom half of the head is
shown in a triangular mesh (in brown). The TMS coil is shown by wireframe wheels and the
target point (red sphere) is located within MT, the motion processing region of the brain. This
image is a screenshot of the fMRI guided TMS neuronavigation software that was used to
target MT in real time, with the head and coil identical to the positioning shown in (B) but
zoomed in closer to the coil. (D) TMS results showing a reduced hit rate (the probability of
responding “moving” to previously moving items or “stationary” to previously stationary
items) for moving items following TMS to MT, as compared to no TMS (the asterisk indicates
a significant difference, ns = not significantly different).
Figure 3.1 Regions of the brain associated with episodic memory. fMRI activity (in red/
yellow) in the left hemisphere (left, lateral view; right, medial view; occipital poles toward
the center).

Figure 3.3 Regions of the brain associated with semantic memory. Cortical thinning in
Alzheimer’s patients (in red/yellow) associated with disruption in semantic memory (lateral
views, occipital poles toward the center).
Figure 3.6 Regions of the brain associated with subsequent memory effects. fMRI
activations associated with subsequent memory (in red/yellow; top, lateral views, occipital
poles toward the center; bottom, coronal views, the left image is the most anterior and the
right image is the most posterior). Medial temporal lobe activity, centered on the
hippocampus, is shown near the bottom of each coronal image in both hemispheres.

A
Remember
+
Know
5 µV
New

F5 F6

P5 P6

0 800 ms
B
300–500 ms 500–800 ms

Figure 4.1 ERP activity associated with recollection and familiarity. (A) Activation
timecourses (microvolts as a function of milliseconds) at frontal electrodes and parietal
electrodes associated with “remember” responses to old items (recollection), “know”
responses to old items (familiarity), and “new” responses to new items (event, electrode,
and amplitude keys at the top). (B) Topographic maps illustrating the mid-frontal old–new
effect within 300 to 500 milliseconds (left) and the left-parietal old–new effect within 500 to
800 milliseconds (right; superior views, occipital poles at the bottom; more significant
activity is shown in red).
Figure 4.4 Topographic maps and activation timecourses illustrating spatial memory
effects. (A) Topographic map corresponding to accurate memory for items previously
presented on the left (old-left-hits) versus accurate memory for items previously presented on
the right (old-right-hits) at 180 milliseconds after stimulus onset (lateral views, occipital
poles toward the center; key to the right, in microvolts). Regions of interest (white ovals)
included left (L) and right (R) frontal (F), parietal (P), temporal (T), and occipital (O) electrodes
(red discs). (B) Topographic map corresponding to old-right-hits versus old-left-hits at 1417
milliseconds after stimulus onset. (C) Activation timecourses corresponding to old-right-hits
versus old-left-hits in the left frontal, the left temporal, and the left occipital regions of
interest from 1377 to 1477 milliseconds after stimulus onset (key at the top right).
 A Remembered Forgotten

Theta

Alpha

Gamma

+ signs: electrodes with significant difference

Theta +/- 3, Alpha +/- 6, Gamma +/- 0.4 (μV2)

B Modulation
index (MI)

SR > SF

frontal 0.01
Remembered

theta

parietal
gamma

SF > SR
0.005
0.005 0.01 MI
Forgotten

Figure 4.5 EEG frequency band activity associated with subsequently remembered and
forgotten items. (A) Topographic maps illustrating subsequently remembered and
subsequently forgotten theta activity (top), alpha activity (middle), and gamma activity
(bottom; superior views, occipital pole at the bottom of each image; key at the bottom, in
microvolts squared). (B) Left, schematic illustrating frontal theta activity and parietal-
occipital gamma activity cross-frequency coupling. Right, frontal theta modulation of
parietal-occipital gamma activity (as measured by a modulation index, MI) was greater for
subsequently remembered (SR) than subsequently forgotten (SF) items (each dot represents
one participant’s remembered MI versus forgotten MI, with dots above the line showing the
SR > SF effect).
Figure 5.1 Subsequent forgetting fMRI activity and default network fMRI activity. (A)
Subsequent forgetting fMRI activity (in red/yellow) in the right hemisphere (top, lateral view,
occipital pole to the left; bottom, medial view, occipital pole to the right). The same pattern of
activity was reported in the left hemisphere. (B) Default network fMRI activity (in blue/cyan)
in the left hemisphere (top, lateral view, occipital pole to the right; bottom, medial view,
occipital pole to the left). The same pattern of activity was reported in the right hemisphere.

θ – Amplitude (5–9 Hz)

35 Difference
30 (SR - RE)
SR
25 RE
Signal change (%)

20
15
10
5
0
–8 % 8%
–0.5 0 0.5 1 1.5
Time (sec.)

Figure 5.3 Retrieval-induced forgetting EEG activity. Left, the magnitude of theta activity
(percent signal change) in the selective retrieval (SR) condition was greater than that of the
re-exposure (RE) condition within 0 to 0.5 seconds after stimulus onset (key at the top right).
Right, topographic map illustrating the difference in theta activity between the selective
retrieval condition and the re-exposure condition within 0 to 0.5 seconds after stimulus onset
(superior view, occipital pole at the bottom; key at the bottom, in percent).
 A

Old-hit
Related-false alarm

0.3
Signal (% change)

0.2
0.1
0
–0.1
–0.2

0 4 8 12 16
Time (s)

Figure 5.4 Regions of the brain commonly and differentially associated with true memory and
related false memory. (A) fMRI activity (in orange) associated with both true memory and
false memory (left, superior view; right, inferior view; occipital poles at the bottom). Activity in
a later visual region is shown within the black circle. (B) Right, the contrast of true memory and
false memory (old-hits > related-false alarms) produced activity in early visual regions (inferior
view, occipital pole at the bottom), as shown within the black circle. Left, activation timecourses
(percent signal change as a function of time after stimulus onset, in seconds) associated with
true memory (old-hits) and false memory (related-false alarms; key at the top).

Figure 5.5 Brain activity associated with unrelated false memory. Left, fMRI activity (in red/
yellow) associated with false memory for new unrelated items (lateral view, occipital pole to
the right). Right, activation timecourses (percent signal change as a function of time after
stimulus onset, in seconds) extracted from activity in language processing cortex (within the
white circle to the left; key to the right).
 A

HOUSE IDEN +

Instruction
3 Sec Instruction +
Delay
3 Sec
+
Sample
Presentation Memory
3 sec Delay Test
9 sec 3 sec ITI
3 sec

B C
(Location > Control only)

MedFus 0.35% 0.30%

Superior Frontal Areas

Inferior Frontal Areas

(Face > Control only)
Z 0.30% 0.25%
Face MedFus 0.25%
0.20%
identity 0.20%
0.15% 0.15%
2.34 0.10% 0.10%
0.05%
House 0.05%
0.00%
identity –0.05% 0.00%
LatFus LatFus –0.10% –0.05%

Figure 6.1 Object or location working memory paradigm and fMRI results. (A) On each trial,
a cue instructed participants whether to maintain object (face or house) information or
spatial location information during the working memory delay period. Items were presented
during the sample/study phase, followed by the delay period, the test phase, and an inter-
trial-interval (ITI) before the onset of the next trial (the time of each period, in seconds, is
shown under each panel). (B) Maintenance of faces during the delay period produced activity
(in red/yellow) in the lateral fusiform cortex (i.e., the fusiform face area) and maintenance of
houses during the delay period produced activity (in cyan/purple) in the medial fusiform/
parahippocampal cortex (i.e., the parahippocampal place area; axial view, occipital pole at
the bottom). (C) Left, activity (percent signal change) in the superior dorsolateral prefrontal
cortex (identified by contrasting working memory for spatial locations and control trials) was
associated with maintenance of spatial locations (in green) to a greater degree than
maintenance of faces (in red) and houses (in blue). Delay period activity corresponds to time
points 2 to 4 (paradigm timing key at the top). Right, activity in the inferior dorsolateral
prefrontal cortex (identified by contrasting working memory for faces and control trials) was
associated with maintenance of faces (in red) and houses (in blue) to a greater degree than
maintenance of spatial locations (in green).
 A SAMPLE DELAY PROBE
200 ms 900 / 5000 ms match non-match

COLOR

LOCATION

ASSOCIATION

B COLOR LOCATION ASSOCIATION

100 100 100
correct responses [%]

90 90 90

80 80 80

70 patients 70 70
controls
900 5000 900 5000 900 5000
delay [ms] delay [ms] delay [ms]

Figure 6.3 Color and/or location working memory paradigms and medial temporal lobe
lesion results. (A) During each color working memory trial, illustrated at the top, colored
squares were presented during the sample/study phase, there was a 900- or 5000-
millisecond delay period, and then there was a probe/test phase in which participants made
“match”–“non-match” judgments. The same paradigm was used for location and
association (i.e., color and location) trials, illustrated at the middle and bottom, respectively.
(B) Performance (percent correct) on the color, location, and association working memory
tasks as a function of delay period duration (in milliseconds) for patients with medial
temporal lobe damage and control participants that did not have a brain lesion (asterisks
indicate significantly impaired performance in the patients as compared to control
participants).
 A

Cue

Memory Array

200 ms

Retention Interval
100 ms

Probe
900 ms

2000 ms

B
0.01
Theta -locked Gamma Phase

3.0
Alpha Activity [µV/m2]
Synchronization

1.5
contralateral
0.00 ipsilteral 0.0

–1.5

–3.0
–0.01 Load 2 Load 3 Load 4 Load 6
Load 2 Load 3 Load 4 Load 6

Figure 6.4 Color working memory paradigm and EEG results. (A) During each trial, an arrow
cued one hemifield. The memory array/study phase consisted of two to six colored squares in
each hemifield, followed by a retention interval/delay period where the stimuli in the cued
hemifield were maintained, and then during the probe/test phase participants indicated
whether or not any of the colors in the cued hemifield had changed. (B) Left, theta-gamma
synchronization as a function of the number of items in working memory (i.e., working
memory load) at contralateral and ipsilateral occipital-parietal recording sites (key to the
right). Right, alpha activity as a function of working memory load at contralateral and
ipsilateral occipital-parietal recording sites.
 A B

NOVEL > REPEATED SAME

TEST

STUDY

C L Fusiform R Fusiform
0.5 0.5
Novel Novel
0.4 Same 0.4 Same
Different Different
% SIGNAL CHANGE

% SIGNAL CHANGE

0.3 0.3

0.2 0.2

0.1 0.1

0 0

–0.1 –0.1

–0.2 –0.2
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
TIME (sec) TIME (sec)

Figure 7.1 Repetition priming paradigm and fMRI results. (A) Left, during the study phase,
objects were presented. Right, during the test phase, the same/old objects, different objects
with the same name, and novel/new objects were presented. (B) Decreases in fMRI activity
for repeated same/old items as compared to novel/new items. Left, dorsolateral prefrontal
cortex activity is shown at the top left and the top right. Right, ventral occipital cortex activity
is shown at the bottom left and the bottom right (axial views, occipital poles at the bottom).
(C) Event-related activation timecourses (percent signal change as a function of time after
stimulus onset) extracted from the left fusiform cortex and the right fusiform cortex for same/
old, different, and novel/new items (key at the top right of each image).
 prefrontal cortex lateral temporal cortex

visual cortices

Least Stimulus Most

specificity

Figure 7.2 Review of cortical repetition priming effects. Repetition priming effects have
been consistently observed in the dorsolateral prefrontal cortex (in green), the lateral
temporal cortex (in red), and in the visual cortices (in blue) within the posterior occipital
cortex and the ventral occipital-temporal processing stream. Within visual cortical regions,
more posterior regions are the most stimulus specific and more anterior regions are the least
stimulus specific (lateral view, occipital pole to the left; key at the bottom).
 A

B C

Right > Left Left > Right

Figure 8.1 Spatial attention paradigm and fMRI results. (A) Attention stimulus display with
two overlapping arrowheads at the central fixation point and a flashing checkerboard
stimulus within each visual field. When one arrowhead briefly turns red, participants shift
attention to the corresponding visual field/stimulus (illustrated by the dotted circle).
Participants press a button when they detect a small red square within the attended location/
stimulus and ignore the unattended location/stimulus. (B) Contralateral attention activity in
early visual regions (axial view, occipital pole at the bottom). The contrast of attention to the
right visual field and attention to the left visual field (Right > Left) produced activity in the left
extrastriate cortex (in purple/cyan), while the contrast of attention to the left visual field and
attention to the right visual field (Left > Right) produced activity in the right extrastriate
cortex (in red/yellow). (C) Attention control activity in the dorsolateral prefrontal cortex (the
rightmost activation) and the parietal cortex (the leftmost activation) of the right hemisphere
(in purple/yellow; lateral-posterior view, occipital pole to the left).
Figure 8.2 Spatial memory fMRI and ERP results. (A) Contralateral memory fMRI activity in
early visual regions (posterior view). The contrast of accurate memory for items previously on
the left and accurate memory for items previously on the right (old-left-hit > old-right-hit)
produced activity in the right extrastriate cortex (in red), while the contrast of accurate
memory for items previously on the right and accurate memory for items previously on the
left (old-right-hit > old-left-hit) produced activity in the left extrastriate cortex (in blue; key at
the bottom). (B) Contralateral memory ERP activity, with occipital regions of interest (ROIs)
and temporal ROIs demarcated (electrode locations are shown by small red discs; key at the
center, in microvolts). Left, topographic map associated with accurate memory for items
previously on the right versus accurate memory for items previously on the left (old-right-hit –
old-left-hit) at 154 milliseconds after stimulus onset (posterior view) and the corresponding
occipital cortex dipole source in the left hemisphere (in blue; coronal view). Right,
topographic map associated with accurate memory for items previously on the left versus
accurate memory for items previously on the right (old-left-hit – old-right-hit) at 180
milliseconds after stimulus onset (posterior view) and the corresponding occipital cortex
dipole source in the right hemisphere (in red; coronal view).
 A

Time

B Perception Imagery Attention

C
Perception + Imagery
Imagery + Attentiom

Figure 8.4 Visual perception, imagery, and attention paradigms and fMRI results. (A) Left,
perception stimulus display, with flashing checkerboards rotating around the central fixation
point. Right, imagery and attention stimulus display with only the outer arcs of the flashing
checkerboards rotating around the central fixation point. During the perception and imagery
condition, participants determined whether a briefly flashed small red square was “inside” or
“outside” of the stimuli. During the attention condition, participants determined whether the
small red square was in the “left” visual field or the “right” visual field. (B) Perception,
imagery, and attention retinotopic maps for a representative participant (posterior view;
colors correspond to different spatial locations in the visual field as shown by the semi-circle
key between the perception and imagery retinotopic maps). Early visual regions are labeled
(in black) and cyan ovals show the regions where imagery produced greater retinotopic
activity than attention. The repeating patters of colors (e.g., yellow to red to yellow to red in
the upper left hemisphere) correspond to repeated visual field representations in early visual
areas (e.g., the lower right quadrant in the visual field has a unique representation in dorsal
V1, V2, and V3 of the left hemisphere). (C) Activity associated with both perception and
imagery (in green) and activity associated with both imagery and attention (in orange; key to
the right).
A B
1.0 Old
Similar

Proportion of Response
0.8
New
0.6

0.4

0.2

0.0
Controls aMCI
Lures

C E

CA1 LS minus LO
CA3/DG 1.5
Mean activity: Separation Contrast

SUB Controls
ERC aMCI
1
PRC
0.5

D LCA3/DG 0

–0.5

–1

–1.5
LCA3/DG LCA1 LSUB LERC

LERC

Figure 9.2 Pattern separation paradigm, behavioral results, and fMRI results for control
participants and aMCI patients. (A) Illustration of the stimulus paradigm that included old
objects, similar objects/lures, and new objects. (B) Proportion of “old,” “similar,” and “new”
responses to lures for control participants (controls) and aMCI patients (key at the top left).
(C) Medial temporal lobe regions of interest included the hippocampal sub-regions CA1,
CA3/dentate gyrus (DG), and the subiculum (SUB) in addition to the entorhinal cortex (ERC)
and the perirhinal cortex (PRC; left hemisphere partial coronal view; key to the right). (D) For
aMCI patients, as compared to control participants, the contrast of pattern separation (lure-
“similar” responses, LS) and pattern completion (lure-“old” responses, LO) produced a larger
magnitude of fMRI activity within the left hemisphere hippocampal CA3/DG (LCA3/DG) sub-
region (in orange/yellow) and a smaller magnitude of fMRI activity within the left hemisphere
entorhinal cortex (LERC; in blue/cyan). (E) Magnitude of fMRI activity associated with the
pattern separation versus pattern completion contrast (LS minus LO) in different medial
temporal lobe regions for control participants and aMCI patients (key at the top right).
Figure 10.4 Memory replay in the rat. Left, during exploration the rat walks down a path
from the starting point (left circle) to the end point (right circle with dots/food). Hippocampal
theta activity is shown below the path (in green). Place cells in the hippocampus (numbered 1
through 4) fire when the rat is in a specific region of the path. Middle, when the rat reaches
the food, a hippocampal sharp-wave ripple (in blue) coordinates the same place cells to
rapidly fire in reverse order. Right, during slow wave sleep, a hippocampal sharp-wave ripple
coordinates the same place cells to rapidly fire in forward order.

A B C

water port
1
treadmill

neuron

21
0 time (s) 16

water port water port

Figure 10.5 Time cell behavioral apparatus and neural activity. (A) The figure-eight maze
apparatus (in red) with the treadmill at the center (in gray; superior view). (B) Schematic of
the apparatus and task. Each rat entered the treadmill, ran for more than 16 seconds, and
then alternated between going through the right arm of the maze (red arrow) and going
through the left arm of the maze (cyan arrow; water ports labeled). (C) Response of 21
hippocampal neurons during a 16-second period on the treadmill for one rat sorted from the
top to the bottom from the neuron that responded earliest in time to the neuron that
responded latest in time (red indicates a higher firing rate).
 A B

Places Other people’s thoughts

Faces
Bodies

Figure 11.1 Past phrenology map and present brain map. (A) Spurzheim’s phrenology map
from 1827 (lateral view, occipital pole to the right). (B) Kanwisher’s brain map from 2010
(lateral view, occipital pole to the right).

Figure 11.2 Face processing and shape processing fMRI activity. (A) Face versus object activity
is shown in blue, with regions of interest labeled, and object versus face activity is shown in
purple (key at the top left; left, inferior view, occipital pole at the bottom; top right, lateral view,
occipital pole to the left; bottom right, lateral view, occipital pole to the right; L = left, R = right,
FFA = fusiform face area, ATFP = anterior temporal face patch, Amy = amygdala, OFC =
orbitofrontal cortex, OFA = occipital face area, fSTS = face-selective region in the superior
temporal sulcus, and IFS = inferior frontal sulcus). (B) Magnitude of activity (percent signal
change) associated with faces, shapes in the left visual field (shape-LVF), shapes in the central
visual field (shape-CVF), and shapes in the right visual field (shape-RVF) in the right FFA (RFFA)
and the left FFA (LFFA). Brackets illustrate statistical comparisons between faces and the other
event types (asterisks indicate significant
CHAPTER NINE

Explicit Memory and Disease

Learning Objectives
• To describe the changes in brain anatomy and fMRI activity in patients
with amnestic mild cognitive impairment.
• To identify the regions of the brain that atrophy in patients with early
Alzheimer’s disease and learn the proteins that are accumulated in these
regions.
• To compare the behavioral performance and fMRI activations of mild
traumatic brain injury patients and healthy control participants during
working memory tasks.
• To understand how surgery on medial temporal lobe epilepsy patients
has revealed associations between the left medial temporal lobe and
the right medial temporal lobe and verbal long-term memory and visual
long-term memory.
• To specify the location of the hippocampal lesion that causes transient
global amnesia.

The previous chapters of this book have focused on the neural basis of
memory in healthy adults. This chapter discusses five neurological
diseases that affect the brain regions associated with explicit memory.
Section 9.1 discusses patients with amnestic mild cognitive impairment.
These patients have long-term memory deficits due to atrophy of medial
temporal lobe regions including the hippocampus. Within a few years
of being diagnosed with amnestic mild cognitive impairment, about half
of these individuals are diagnosed with Alzheimer’s disease, the topic of
section 9.2. Patients with early Alzheimer’s disease have more severe
impairment of long-term memory and atrophy of the medial temporal
lobe and the parietal lobe, two regions that have been associated with
long-term memory (see Chapter 3). Alzheimer’s disease patients also
have abnormally high levels of proteins in the medial temporal lobe and
the parietal lobe, which is thought to further disrupt processing in these
regions. Section 9.3 focuses on patients with mild traumatic brain injury,
who typically perform normally on working memory tasks but have
increased fMRI activity within the dorsolateral prefrontal cortex and
172 Explicit Memory and Disease

the parietal cortex, relative to healthy control participants. It is

generally believed that such increases in fMRI activity reflect compen-
sation, where these regions are recruited to perform normally on the
task. In section 9.4, patients with medial temporal lobe epilepsy are
considered. These patients can elect to have a region in their medial
temporal lobe removed in an effort to reduce the frequency of their
seizures. Removal of regions in the left temporal lobe produces deficits
in verbal long-term memory, while removal of regions in the right
temporal lobe produces deficits in visual long-term memory. In the
final section, 9.5, patients with transient global amnesia are discussed
(such patients were briefly discussed in Chapter 3). These patients have
a sudden onset of amnesia that lasts for less than 24 hours and is caused
by a small temporary lesion to a specific sub-region of the hippocam-
pus. Although transient global amnesia is almost always triggered by
emotional or physical stress, its underlying mechanisms have remained
elusive for over half a century.

9.1 Amnestic Mild Cognitive Impairment

Amnestic mild cognitive impairment (aMCI) occurs in a small but
significant percentage of adults who are older than 60 years of age, with
incidence increasing as a function of age. Individuals with aMCI have
a selective impairment in long-term memory, as compared to healthy
aged-matched control participants, and are unimpaired in other cognitive
domains such as attention and language. Although aMCI patients are
often described as having a selective impairment in episodic memory,
they have impaired item memory as well (e.g., they might forget about
a recently made appointment). There is a convincing body of evidence
indicating that the long-term memory impairment in aMCI patients is
due to atrophy of medial temporal lobe sub-regions that is accompanied
by a paradoxical increase in fMRI activity within the medial temporal
lobe during long-term memory (Dickerson & Sperling, 2008; Leal &
Yassa, 2013).
In one study, structural MRI was used to compare the size of the
hippocampus and the entorhinal cortex in aMCI patients and control
participants (Stoub et al., 2006). The entorhinal cortex is a sub-region of
the medial temporal lobe that connects the perirhinal cortex and the
hippocampus (see the following paragraph and Chapter 10). It is notable
that in most cognitive neuroscience fMRI studies, the entorhinal cortex
and the perirhinal cortex are both referred to as the perirhinal
cortex, a region that has been associated with item memory/familiarity
9.1 Amnestic Mild Cognitive Impairment 173

B
2.5 0.8
LH
Hippocampal Volume

RH
Entorhinal Volume

2.0
Mean Normalized
Mean Normalized

0.6
1.5
0.4
1.0
0.2
0.5

0.0 0.0
Aged Controls MCI Aged Controls MCI

Figure 9.1 Hippocampus and entorhinal cortex segmentation and volumes of these
regions in control participants and amnestic mild cognitive impairment (aMCI) patients.
(A) Segmentations (white outlines) of the hippocampus (in the left hemisphere) and the
entorhinal cortex (in the right hemisphere) of a representative participant (coronal view).
(B) Mean hippocampal volume and entorhinal cortex volume within the left hemisphere (LH)
and the right hemisphere (RH) of healthy aged-matched control participants and aMCI
patients (key at the top right).

(see Chapters 3 and 10). Figure 9.1A outlines the hippocampus, in the left
hemisphere, and the entorhinal cortex, in the right hemisphere, of
a representative participant. Figure 9.1B shows that aMCI patients had
a smaller hippocampal volume and a smaller entorhinal cortex volume in
both hemispheres, as compared to age-matched control participants,
which indicates there was atrophy of these regions. In addition, the
white matter pathway between the entorhinal cortex and the hippocam-
pus, which is referred to as the perforant path, also had a smaller volume
in aMCI patients than control participants, and this was the only white
174 Explicit Memory and Disease

matter region in the entire brain that differed in volume. These results
indicate that the long-term memory impairments in aMCI patients are
due to isolated atrophy in the entorhinal cortex and the hippocampus.
Medial temporal lobe atrophy can be assumed to disrupt processing in
that region and thus would be expected to produce a decrease in the
magnitude of fMRI activity. However, aMCI patients typically show an
increase in the magnitude of fMRI activity within the medial temporal
lobe. One long-term memory fMRI study evaluated the magnitude of
activity within different medial temporal lobe sub-regions in aMCI
patients and control participants (Yassa et al., 2010). Figure 9.2A illus-
trates the stimulus paradigm. Each run consisted of a series of objects that
were new items (e.g., the clover in the top panel), old items (e.g., the duck
in the fifth panel), or similar items/lures (e.g., the clover in the bottom
panel). Participants classified each item as “old,” “similar,” or “new.”
The two critical event types are “similar” responses to lures, which
reflect the process of pattern separation (where participants distinguish/
separate old items and lures), and “old” responses to lures, which reflect
the process of pattern completion (where participants respond based on/
complete the common patterns between old items and lures). Pattern
separation reflects correct responses and pattern completion reflects
incorrect responses. Pattern completion is another way of referring to
false memories for new related items (see Chapter 5). The behavioral
performance of aMCI patients and control participants was similar for
old items and new items, but differed for lures. Figure 9.2B shows that
aMCI patients responded “old” to lures at a higher rate than control
participants and responded “similar” to lures at a lower rate than control
participants (the rate of “new” responses to lures did not differ between
groups). These behavioral results indicate that aMCI patients shift from
pattern separation to pattern completion. In the real world, this might
translate into aMCI patients having a high rate of false memories to new
related items (e.g., incorrectly recognizing a person who looks similar to
someone they actually know). The medial temporal lobe sub-regions
that were evaluated during pattern completion and pattern separation
for aMCI patients and control participants are shown in Figure 9.2C.
These regions included the hippocampal sub-regions CA1, CA3/dentate
gyrus (DG), and the subiculum (SUB) in addition to the entorhinal
cortex (ERC) and the perirhinal cortex (PRC). Each medial temporal
lobe sub-region is associated with different types of processing (see
Chapter 10). As illustrated in Figures 9.2D and 9.2E, the contrast of
“similar” responses to lures (pattern separation) and “old” responses to
lures (pattern completion) produced a higher magnitude of activity in the
A B
1.0 Old
Similar

Proportion of Response
0.8
New
0.6

0.4

0.2

0.0
Controls aMCI
Lures

C E

CA1 LS minus LO
CA3/DG 1.5
Mean activity: Separation Contrast

SUB Controls
ERC aMCI
1
PRC
0.5

D LCA3/DG 0

–0.5

–1

–1.5
LCA3/DG LCA1 LSUB LERC

LERC

Figure 9.2 Pattern separation paradigm, behavioral results, and fMRI results for control
participants and aMCI patients. (A) Illustration of the stimulus paradigm that included old
objects, similar objects/lures, and new objects. (B) Proportion of “old,” “similar,” and “new”
responses to lures for control participants (controls) and aMCI patients (key at the top left).
(C) Medial temporal lobe regions of interest included the hippocampal sub-regions CA1,
CA3/dentate gyrus (DG), and the subiculum (SUB) in addition to the entorhinal cortex (ERC)
and the perirhinal cortex (PRC; left hemisphere partial coronal view; key to the right).
(D) For aMCI patients, as compared to control participants, the contrast of pattern separation
(lure-“similar” responses, LS) and pattern completion (lure-“old” responses, LO) produced
a larger magnitude of fMRI activity within the left hemisphere hippocampal CA3/DG (LCA3/
DG) sub-region (in orange/yellow) and a smaller magnitude of fMRI activity within the left
hemisphere entorhinal cortex (LERC; in blue/cyan). (E) Magnitude of fMRI activity associated
with the pattern separation versus pattern completion contrast (LS minus LO) in different
medial temporal lobe regions for control participants and aMCI patients (key at the top right).
(A black and white version of this figure will appear in some formats. For the color version,
please refer to the plate section.)
176 Explicit Memory and Disease

left CA3/DG sub-region and a lower magnitude of activity in the left

entorhinal cortex of aMCI patients than control participants. The relative
decrease in entorhinal cortex activity of aMCI patients can be attributed
to atrophy of this region, which is described above. However, as the
hippocampus also shows atrophy in aMCI patients, the relative increase
in activity within the CA3/DG sub-region is unexpected.
There are two different hypotheses that could explain the increase in
fMRI activity within the CA3/DG sub-region of the hippocampus in
aMCI patients. The first hypothesis is that the increase in activity reflects
compensation for disrupted neural processing (i.e., the hyperactivity is
functional and enhances behavioral performance). Using the results
above to illustrate, if this hypothesis is correct, the increase in CA3/DG
activity would reflect increased processing in this region to successfully
complete the task. The second hypothesis is that the increase in activity
reflects non-compensatory disruption of normal processing (i.e., the
hyperactivity is non-functional and does not enhance and might even
impair behavioral performance). If this hypothesis is correct, the increase
in CA3/DG activity might reflect disrupted inhibitory processes (i.e.,
the increase in activity could be due to disinhibition rather than
compensation; see Gallagher & Koh, 2011). To distinguish between
these hypotheses, a recent fMRI study employed a similar paradigm to
the one used in the previous study and reduced the magnitude of activity
within the CA3/DG sub-region of the hippocampus in aMCI patients
with the anti-epileptic drug levetiracetam (Bakker, Albert, Krauss, Speck &
Gallagher, 2015). If the increase in activity within the CA3/DG
sub-region reflects compensation, reducing this activity should impair
behavioral performance; however, if the increase in activity reflects
disrupted neural processing, reducing this activity might enhance beha-
vioral performance. As in the previous study, without medication, the
behavioral performance of aMCI patients showed a shift from pattern
separation to pattern completion, as compared to control participants,
and pattern separation produced a higher magnitude of activity in aMCI
patients than control participants in the hippocampal CA3/DG sub-
region. After treatment of the aMCI patients with levetiracetam,
behavioral performance improved and was similar to control participants
and the magnitude of activity in the hippocampal CA3/DG sub-region
was reduced such that it no longer differed from control participants.
These findings suggest that the relatively higher magnitude of fMRI
activity within the CA3/DG sub-region during pattern separation reflects
a non-compensatory change in processing related to neural disruption in
aMCI patients.
9.2 Alzheimer’s Disease 177

9.2 Alzheimer’s Disease

Alzheimer’s disease (AD) is the most common cause of cognitive
deficits in older adults. The first cognitive problem in early AD patients
is impaired long-term memory. About half of aMCI patients, who have
atrophy in medial temporal lobe sub-regions (including the hippocam-
pus; see section 9.1), are diagnosed with AD within a few years (Tromp,
Dufour, Lithfous, Pebayle & Després, 2015). As AD progresses from
earlier to later stages, atrophy starts in the medial temporal lobe,
extends to the parietal lobe, and finally includes the frontal lobe
(Reiman & Jagust, 2012; Tromp et al., 2015). The long-term memory
impairment in early AD patients can be attributed to the disrupted
processing in the hippocampus and parietal cortex, two regions that
have been associated with this cognitive process (see Chapter 3).
As the disease progresses, other cognitive functions are disrupted
such as attention and language, which both depend on the dorsolateral
prefrontal cortex (see Chapter 8).
As described in section 9.1, aMCI has been associated with increases in
fMRI activity within medial temporal lobe sub-regions during long-term
memory tasks, as compared to control participants. In aMCI patients who
progress to AD, as atrophy increases in the medial temporal lobe, there is
a relative decrease in fMRI activity in this region (Dickerson & Sperling,
2008; Leal & Yassa, 2013). In early AD patients, as atrophy begins in the
parietal cortex and the frontal cortex, there have also been reports of
increases in fMRI activity within cortical regions. It is uncertain whether
these increases in cortical fMRI activity reflect a compensatory mechan-
ism, which is often assumed to be the case, or reflect non-compensatory
hyperactivity due to neural disruption.
In addition to brain atrophy, AD patients have abnormally high levels
of proteins in different brain regions. In the medial temporal lobe, the
accumulation of tau protein leads to neurofibrillary tangles. In cortical
regions, such as the parietal cortex in early AD, the accumulation of
amyloid-β protein leads to amyloid plaques. The neurofibrillary tangles
in the medial temporal lobe and amyloid plaques in cortical regions can
be assumed to disrupt neural processing in these regions. There is an
influential hypothesis that there is a causal relationship between default
network activity that leads to deposition of amyloid that results in atro-
phy and disrupted metabolic activity, which impairs long-term memory
in AD patients (Buckner et al., 2005). As detailed in Chapter 5, regions in
the default network are active when participants are not engaged in a task
and include the dorsolateral prefrontal cortex, the medial prefrontal
178 Explicit Memory and Disease

cortex, the inferior parietal cortex, and the medial parietal cortex. In AD
patients, amyloid deposition occurs in the same regions, which suggests
that default network activity may lead to amyloid deposition. However,
the link between amyloid deposition and atrophy is tenuous, as AD
patients initially have atrophy in the medial temporal lobe and the
parietal cortex. Thus, in early AD patients, there is no correlation
between amyloid deposition and atrophy in either the medial temporal
lobe (where there is low amyloid deposition but significant atrophy) or
the frontal cortex (where there is high amyloid deposition but little
atrophy). This lack of correlation questions the hypothesis that a high
level of amyloid deposition causes atrophy in AD patients. However, it is
still possible that a high level of amyloid deposition causes atrophy in
susceptible brain regions, such as the parietal cortex. Perhaps a higher
level of amyloid deposition, which occurs in late AD patients, is neces-
sary to produce atrophy in the frontal cortex. This is a topic for future
investigation.
The high level of amyloid deposition in the parietal cortex and the
frontal cortex of AD patients suggests that accumulation of this protein
disrupts neural processing in these cortical regions and produces long-
term memory deficits. Interestingly, there is considerable variation in the
level of amyloid deposition in the brains of healthy older adults. If high
amyloid deposition is a causal factor in developing AD, older adults with
low levels of amyloid should be at decreased risk for developing this
disease. There is some evidence that cognitive engagement and exercise
engagement throughout life may reduce the amyloid level in the brains of
healthy older adults. In one study, cortical amyloid level was measured
in older adults as a function of cognitive engagement, and this was
compared to the cortical amyloid levels in AD patients and young adults
(Landau et al., 2012). Amyloid level was measured using PET (see
Chapter 2) with a radioactive substance that binds to this protein called
Pittsburgh Compound B (PiB). Participants rated the frequency in which
they engaged in cognitively demanding tasks such as reading, writing,
going to the library, or playing games at five different ages (6, 12, 18, 40,
and their current age). Healthy older adults with greater cognitive
engagement throughout their lifetime, as measured by the average
cognitive activity at the five ages, had lower levels of PiB uptake/amyloid
in default network regions. Moreover, the healthy older adults in the
lowest one-third of lifetime cognitive engagement had PiB/amyloid levels
that were equivalent to AD patients, and the healthy older adults in the
highest one-third of lifetime cognitive engagement had PiB/amyloid
levels that were equivalent to young adults. Another study measured
9.3 Mild Traumatic Brain Injury 179

the level of AD biomarkers in healthy older adults as a function of

exercise (Liang et al., 2010). Cortical amyloid level was measured with
PiB using PET, and tau protein level was measured in the cerebrospinal
fluid using a spinal tap. Participants rated the frequency and duration that
they engaged in walking, jogging, and running for the previous 10 years.
Exercise engagement was the average metabolic equivalent hours per
week during that period. As a reference value, the American Heart
Association recommends 7.5 metabolic equivalent hours per week for
older adults, which is about 30 minutes of moderate exercise 5 days per
week. Figures 9.3A and 9.3B show that older adults with higher levels of
exercise engagement had lower levels of PiB/amyloid and lower levels of
tau. It is particularly striking that none of the older adults who exercised
more than the recommended 7.5 metabolic equivalent hours per week
(demarcated by the horizontal dashed lines) had abnormal levels of PiB
or tau (normal levels are below the vertical dashed lines). The results of
the previous two studies suggest that cognitive engagement and exercise
engagement throughout life reduce the levels of amyloid protein and tau
protein in the brain. As these are the primary biomarkers in AD, mental
processing and physical activity may reduce the risk of contracting this
disease.

9.3 Mild Traumatic Brain Injury

Traumatic brain injury is relatively common in the general population
and the large majority of these injuries are mild (McDonald, Saykin &
McAllister, 2012; Mayer, Bellgowan & Hanlon, 2015). There are many
causes for mild traumatic brain injury (mTBI) such as motor vehicle
accidents, sports-related injuries, and blasts during military combat.
Patients with mTBI do not have any brain abnormalities, as measured
using structural neuroimaging methods such as anatomic MRI.
The diagnosis of mTBI includes loss of consciousness for less than 30
minutes and post-traumatic amnesia for less than 24 hours. Patients with
mTBI can have attention and memory deficits, but these typically resolve
within a few weeks. In the last decade, there have been an increasing
number of fMRI studies that have reported differences between the
pattern of brain activity in mTBI patients and control participants during
working memory tasks. As will be discussed below, there are many
factors that can impact the fMRI findings such as the severity of the
head injury, a history of previous head injuries, the delay between the
trauma and the time of testing, and the persistence of symptoms (e.g.,
headache, dizziness, nausea, and insomnia).
180 Explicit Memory and Disease

Exercise Engagement (MET-hours/week)

18.0
16.5
15.0
13.5
12.0
10.5
9.0
7.5
6.0
4.5
3.0
1.5
0.0

–0.4 –0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
PiB (Mean Cortical Binding Potential)
B
Exercise Engagement (MET-hours/week)

18.0
16.5
15.0
13.5
12.0
10.5
9.0
7.5
6.0
4.5
3.0
1.5
0.0

0 85 170 255 340 425 510 595 680 765 850

CSF Tau (pg/mL)

Figure 9.3 Relationship between exercise engagement and Alzeimer’s disease biomarkers
in older adults. (A) Exercise engagement, as measured by metabolic equivalent hours per
week (MET-hours/week), as a function of cortical Pittsburgh compound-B (PiB).
The horizontal dashed line shows the recommended amount of exercise by the American
Heart Association and the vertical dashed line shows the upper boundary of the normal
range. (B) Exercise engagement as a function of cerebral spinal fluid tau protein (CSF Tau,
picograms per milliliter).
9.3 Mild Traumatic Brain Injury 181

Working memory has been associated with activity in the dorsolateral

prefrontal cortex and the parietal cortex (see Chapter 6). Many studies
have shown a greater extent and magnitude of fMRI activity in mTBI
patients than control participants in these regions during working mem-
ory tasks. One fMRI study used a working memory task to investigate
whether there were any differences in brain activity between mTBI
patients and control participants (McAllister et al., 2001). Injuries were
due to car crashes, falls, or sports and recreation. Patients were tested
within 1 month of their injury and were excluded if they had a previous
TBI with loss of consciousness. Figure 9.4A, top, illustrates the 1-back
working memory task, where participants heard a string of consonants
every 3 seconds and they responded when a letter was repeated. This task
requires maintenance of the previous letter in working memory so that it
can be compared to the current letter. Figure 9.4A, bottom, illustrates the
2-back working memory task, where participants again heard a string of
consonants but in this case responded when the current letter matched
the letter that was presented before the previous letter. The 2-back task
requires the maintenance of the previous two letters in working memory;
therefore, the contents of working memory can be assumed to be greater
in the 2-back condition than the 1-back condition. More generally, this
working memory task is referred to as the n-back task, as the number of
items maintained in working memory can vary. Figure 9.4B shows that
the performance of mTBI patients and control participants did not differ
on the 1-back task or the 2-back task. To investigate the brain regions
associated with working memory, 2-back blocks were contrasted with
1-back blocks. As illustrated in Figure 9.4C, mTBI patients had a greater
extent and magnitude of fMRI activity in the dorsolateral prefrontal
cortex and the parietal cortex than control participants. A more
recent fMRI study also employed the n-back working memory task to
investigate the brain activation differences between mTBI patients and
control participants (Dettwiler et al., 2014). All of the participants
had a concussion due to sports-related injuries and were tested 2 days,
2 weeks, and 2 months after their injury. Only 1 participant out of 15 had
symptoms 2 months after the injury. Consistent with the previous study,
the behavioral performance of mTBI patients and control participants
did not differ, and the 2-back versus 1-back contrast produced greater
fMRI activity in mTBI patients than control participants within the
dorsolateral prefrontal cortex at all three time points and within the
parietal cortex at the first two time points. The greater activity in mTBI
patients within the dorsolateral prefrontal cortex 2 months after injury
are particularly concerning because they indicate there are differences in
182 Explicit Memory and Disease

A
1 back

B X X A G G L F G N N T C P P H

2 back

K S B S Q M M G M R C J C F Q F

B C
100 MTBI 2bk > 1bk
Controls

80
Percent Correct

Controls
60

40

20

MTBI
0
1-back 2-back
Task Condition

Figure 9.4 N-back paradigm, behavioral results, and fMRI results for mild traumatic brain
injury (mTBI) patients and control participants. (A) Illustration of the 1-back task (top) and
the 2-back task (bottom). Arrowheads indicate correct responses. (B) Accuracy (percent
correct) on the 1-back and 2-back tasks for mTBI patients and control participants (key at the
top right). (C) fMRI activity (in gray/white) produced by the contrast between 2-back and
1-back blocks for control participants (top) and mTBI patients (bottom; superior view,
occipital poles to the left).

brain processing even after behavioral symptoms have resolved. This

indicates there can be persistent brain disruption in these individuals,
as indicated by fMRI, even though there are no behavioral symptoms or
brain abnormalities observable with anatomic neuroimaging methods.
As mTBI patients may be more sensitive to repeated head trauma, it is
arguable that they should not be allowed to continue participating in
impact sports until their fMRI activity returns to normal. Future studies
should evaluate later time points, such as 6 months and 1 year after brain
injury, to determine when there is no longer abnormal fMRI activity in
mTBI patients. Another fMRI study employed a virtual reality spatial
navigation task and evaluated mTBI patients within 30 days of their
9.3 Mild Traumatic Brain Injury 183

head injury in an effort to investigate the activity associated with spatial

long-term memory (Slobounov et al., 2010). All cases were sports related
and relatively mild: there was no loss of consciousness, post-traumatic
amnesia lasted less than 30 minutes, there were no clinical symptoms 10
days after the injury, and no participants had a history of mTBI. As in the
previous studies, there was no difference in behavioral performance
between mTBI patients and healthy control participants, and mTBI
patients had greater fMRI activity within the dorsolateral prefrontal
cortex and the parietal cortex. The contrast used to identify this activity
did not isolate spatial long-term memory retrieval from other cognitive
processes associated with spatial navigation (e.g., differences in percep-
tual processing). Still, these findings provide convergent evidence that
there is increased fMRI activity within the dorsolateral prefrontal cortex
and the parietal cortex in mTBI patients, even after relatively mild head
injury.
The previous results indicate there are increases in the magnitude of
fMRI activity within the dorsolateral prefrontal cortex and parietal
cortex during working memory in mTBI patients, as compared to
control participants. However, there is also evidence that the magni-
tude of fMRI activity decreases in mTBI patients with more severe or
repeated head injuries. One working memory fMRI study included
mTBI patients with more severe sports-related head injuries (Chen
et al., 2004). The not-so-mild mTBI patients were tested 1 to 14 months
after the most recent head injury, the large majority of participants
had multiple previous concussions, and 15 of the 16 participants had
persistent symptoms. Each working memory trial consisted of four to-
be-remembered items presented sequentially, a 1-second delay period,
and an old item from the previous set or a new item, and participants
classified this item as “old” or “new” (see Chapter 6). The baseline task
used the identical paradigm, but the first four items were the same and
the final item indicated which button to press (such that the task did not
require working memory). Items on each trial were either abstract
drawings or abstract words, and multiple trials of the same type were
blocked in 1-minute periods. Behavioral performance on the visual
working memory task and the verbal working memory task did not
differ between the mTBI patients and control participants. For both
tasks, the contrast of working memory blocks versus control blocks
produced greater activity in the dorsolateral prefrontal cortex in con-
trol participants than in mTBI patients, which is in direct opposition to
the previous findings for less severe mTBI patients. Moreover, partici-
pants with greater post-concussive symptoms had a smaller magnitude
184 Explicit Memory and Disease

and extent of fMRI activity within the dorsolateral prefrontal cortex

during visual working memory blocks. The same pattern of fMRI
results was obtained in a subsequent study that employed the identical
visual working memory task and a similar group of not-so-mild mTBI
participants (Gosselin et al., 2011). Of additional relevance, repeated
mTBI and sub-concussive head injuries (e.g., due to boxing or profes-
sional football) can lead to chronic traumatic encephalopathy (CTE;
Mez, Stern & McKee, 2013). CTE patients have brain atrophy
that includes the frontal lobe and the medial temporal lobe, among
other regions, which can lead to deficits in behavior (e.g., aggression),
mood (e.g., depression), and cognitive functions (e.g., attention and
memory).
Considering all of the mTBI findings, it appears that patients with
less severe head injury have an increase in fMRI activity within the
dorsolateral prefrontal cortex and the parietal cortex, at least for 1 to
2 months after the injury, and patients with more severe head injury
have a decrease in fMRI activity within the dorsolateral prefrontal
cortex, relative to control participants. This increase and decrease in
fMRI activity within the dorsolateral prefrontal cortex for mTBI
patients with less severe head injury and more severe head injury,
respectively, is reminiscent of the increase and decrease in fMRI
activity within the hippocampus for aMCI and AD patients, respec-
tively. As discussed in Box 9.1, future research will be needed to

Box 9.1: The nature of increased fMRI activity following

mTBI
Behavioral performance on working memory tasks is typically similar for
mTBI patients and control participants, but mTBI patients can have greater
fMRI activity within the dorsolateral prefrontal cortex and the parietal
cortex. Such increases in fMRI activity are generally thought to reflect
a compensatory mechanism, given that these regions are known to be
associated with working memory and thus may be recruited to a greater
degree to successfully perform the task. An alternative hypothesis is that
the increase in fMRI activity reflects a non-compensatory hyperactivation
that reflects disrupted neural processing, similar to what appears to occur
in the hippocampal CA3/DG sub-region of aMCI patients (see section 9.1
of this chapter). One major difference between these groups is that mTBI
patients have normal behavioral performance, whereas aMCI patients have
impaired behavioral performance. The normal behavioral performance
9.3 Mild Traumatic Brain Injury 185

Box 9.1: (cont.)

in mTBI patients would seem to favor the hypothesis that increases in fMRI
activity within the dorsolateral prefrontal cortex and the parietal cortex
reflect a compensatory mechanism. To distinguish between the compen-
satory hypothesis and the non-compensatory hypothesis in mTBI patients,
future studies could experimentally reduce activity in cortical regions
and determine how this affects behavioral performance. For instance,
following Bakker et al. (2015), mTBI patients could be given levetiracetam
to reduce the magnitude of cortical activity, or perhaps 1 Hz TMS could be
used to reduce cortical activity (see Chapter 2). If reduced cortical activity
in mTBI patients impairs behavioral performance, that would favor the
compensatory hypothesis; however, if it does not impair behavioral per-
formance, that would favor the non-compensatory hypothesis.

determine whether the fMRI activation increases in patients with

mTBI reflect a compensatory mechanism or non-compensatory
hyperactivation.
Research on mTBI has two major limitations that should be
mentioned. The first limitation is that many fMRI studies have used
non-uniform groups of mTBI patients with regard to the type of head
injury, the severity of head injury, the region of impact, the time period
between the concussion and testing, the degree of post-concussive
symptoms, and the number and severity of previous concussions.
As fMRI analysis identifies activity that is consistent across partici-
pants and mTBI patients with varying severity of head injury have
different patterns of activity, conducting fMRI analysis on non-
uniform groups of mTBI patients will generally yield null findings.
As such, the low magnitude of fMRI activity in some mTBI studies
may be due to the participants having a wide range of head injuries.
Although this is not an issue for studies that have focused on mTBI
patients with less severe head injury and no history of concussions,
future mTBI studies that focus on patients with more severe head
injury should use more uniform groups of participants. As discussed
in Box 9.2, a second limitation of mTBI studies is that they have not
employed tasks and analysis methods that isolate the cognitive
processes of working memory or long-term memory. Future studies
should use up-to-date cognitive neuroscience tasks and techniques to
186 Explicit Memory and Disease

Box 9.2: Selecting a commonly used task versus

selecting the best task
Many mTBI fMRI studies have employed the n-back task to investigate the
brain regions associated with working memory. The n-back task does involve
maintenance of information in working memory but also involves shifting
attention from the previous item that was maintained to the next item that
needs to be maintained. As such, the n-back task confounds working mem-
ory and attentional shifting (see Chapter 8), which is one reason why this task
is currently rarely used in the field of cognitive neuroscience to investigate
working memory. In other mTBI working memory fMRI studies, the trials
were blocked such that the working memory delay period was not isolated
from the encoding period and the retrieval period and the blocked design
also introduced a difficulty confound (see Chapter 2). One study employed
a spatial long-term memory task, but the contrasts employed did not isolate
this process. As such, none of the working memory or long-term memory
studies that have investigated the changes in fMRI activity following mTBI
have isolated the cognitive process of interest. In the clinical setting, the
repeated use of the same task across studies is often done to maintain
consistency with previous work. However, this should be done only if the
task and analysis methods employed isolate the cognitive process of interest.
Future mTBI studies should use up-to-date cognitive neuroscience tasks,
such as event-related working memory paradigms that can isolate activity
associated with the delay period (see Chapter 6) and event-related long-term
memory paradigms that can isolate activity associated with retrieval (see
Chapter 3). In this way, mTBI studies will be able to connect to the fMRI
literature in the field of cognitive neuroscience that offers a wealth of
findings that can be used as a basis of comparison.

investigate the differences in brain activity between mTBI patients and

control participants.

9.4 Medial Temporal Lobe Epilepsy

Patients with epilepsy have recurrent seizures that are often caused by
abnormal activity within the medial temporal lobe, including the hippo-
campus and the surrounding cortical regions (Willment & Golby, 2013).
Patients with medial temporal lobe epilepsy (mTLE), which is also
referred to as anterior temporal lobe epilepsy, sometimes continue to
9.4 Medial Temporal Lobe Epilepsy 187

have disabling seizures despite taking anti-seizure medication. Some of

these patients elect to have surgical removal of their seizure focus, the
region of the brain from which seizures originate, with the aim of redu-
cing the frequency of their seizures. In the 1940s and 1950s, many mTLE
patients with medically intractable seizures had removal of the medial
temporal lobe within both hemispheres (i.e., bilateral removal), as was
the case with patient H. M. (see Chapter 1), but this caused complete
and devastating anterograde amnesia. Based on such devastating
outcomes, mTLE patient surgeries after that time have attempted to
remove only the specific seizure focus within one hemisphere (i.e.,
unilateral removal).
For half a century, there has been evidence that unilateral temporal
lobe surgery in mTLE patients can produce material-specific long-term
memory deficits (Milner, 1968). Removal of left medial temporal lobe
regions can impair long-term memory for verbal information, while
removal of right medial temporal lobe regions can impair long-term
memory for visual information. One study illustrates the verbal long-
term memory impairments following left temporal lobe surgery in right-
handed participants (Blakemore & Falconer, 1967). There were 54
patients who had removal of a left temporal lobe region and 32 patients
who had removal of a right temporal lobe region. Patients were tested on
a verbal long-term memory task 1 year after surgery. The task was to
learn eight pairs of auditorily presented words through repetition and
testing until the list was correctly remembered three times. Patients with
left temporal lobe removal had three times as many errors in learning the
word pair lists than the number of errors before the surgery, while the
number of errors did not differ before and after surgery for patients with
right temporal lobe removal. Another study illustrates visual long-term
memory impairments following removal of right temporal lobe regions
(Jones-Gotman, 1986). There were 33 patients who had removal of
a left temporal lobe region and 34 patients who had removal of a right
temporal lobe region. All but 2 participants were right handed, and they
had language processing lateralized to the left hemisphere. Patients were
classified into two groups based on removal of smaller areas of the
hippocampus (h) and the parahippocampal gyrus or removal of larger
areas of the hippocampus (H) and the parahippocampal gyrus. The task
was to learn thirteen abstract designs by copying them on a blank piece of
paper (while they were shown) and then drawing them on a blank piece
of paper based on memory (while they were not shown) until at least
twelve of them were successfully recalled two times. There was
a maximum of ten copy-recall trials. Figure 9.5A illustrates one set of
188 Explicit Memory and Disease

B
16

14

12
Mean Percent Forgetting

10

0
Normal Left Temporal Right Temporal
Control h H h H

Figure 9.5 Stimuli and behavioral results for control participants and medial temporal
lobe epilepsy patients following removal of left or right medial temporal lobe regions.
(A) Illustration of one set of abstract designs. (B) Percent forgetting for control participants
and mTLE patients who had removal of the parahippocampal gyrus and smaller (h) or larger
(H) regions of the hippocampus from the left or right medial temporal lobe.
9.4 Medial Temporal Lobe Epilepsy 189

designs. Recall of the shapes was assessed by determining the number

of errors 24 hours after the learning phase, as compared to the number of
errors on the last learning trial. As shown in Figure 9.5B, patients with
right temporal lobe removal, particularly those with larger hippocampal
lesions, had a greater number of recall errors than control participants,
while the number of recall errors did not differ between patients with left
temporal lobe removal and control participants. Another mTLE patient
study used a similar learning procedure to the previous study for both
words and designs (Glosser, Deutsch, Cole, Corwin & Saykin, 1998).
Consistent with the findings from the previous two studies, recognition
memory accuracy was lower for words than designs following removal of
left temporal lobe regions and recognition memory accuracy was lower
for designs than words following removal of right temporal lobe regions.
It should be mentioned that verbal long-term memory deficits are con-
sistently observed following removal of left medial temporal lobe
regions, while visual long-term memory deficits are less consistently
observed following removal of right medial temporal lobe regions
(Willment & Golby, 2013). This may be due to verbal learning strategies
that are employed during certain visual memory tasks (e.g., the stimulus
in the lower right corner of Figure 9.5A could be encoded as “wind
blowing northeast”). If verbal strategies are employed, long-term mem-
ory for visual information would reflect activity in the left medial tem-
poral lobe rather than the right medial temporal lobe. This may explain
why removal of right medial temporal lobe regions do not disrupt visual
long-term memory performance in all studies. Of importance, many
mTLE patient studies have reported verbal long-term memory and
visual long-term memory impairments following removal of left and
right medial temporal lobe regions, respectively. This indicates that
verbal long-term memory is preferentially associated with the left medial
temporal lobe and visual long-term memory is preferentially associated
with the right medial temporal lobe.
The material-specific processing distinction in the left medial temporal
lobe and the right medial temporal lobe described above has been used to
guide surgical planning for mTLE patients. For decades, patients with
mTLE have often taken part in the intracarotid amobarbital test before
surgery to assess whether removal of the seizure focus is likely to disrupt
long-term memory or language comprehension and production. This test
involves injection of sodium amobarbital, a sedative, into one of the
internal carotid arteries, which temporarily disrupts processing in the
anterior two-thirds of the corresponding hemisphere (Glosser et al.,
1998). The patient is then tested to assess the degree to which this impairs
190 Explicit Memory and Disease

behavioral performance. For example, in an mTLE patient with a seizure

focus in the right hemisphere, if the right hemisphere intracarotid
amobarbital test does not reveal impairment in language or long-term
memory, a larger region of the medial temporal lobe could be removed to
increase the likelihood of relieving seizures with a low risk of subsequent
cognitive deficits. In another mTLE patient with a seizure focus in the
right hemisphere, if the right hemisphere intracarotid amobarbital test
does reveal impairment in language or long-term memory, a smaller
region of the medial temporal lobe might be removed to decrease the
likelihood of subsequent cognitive impairments with a lower likelihood
of seizure reduction (such patients may elect not to have surgery to
minimize the risk of cognitive impairments). Although the intracarotid
amobarbital test has widely been used to inform surgical planning, there
are two major problems with this method. First, it is an invasive proce-
dure and thus poses a small but serious risk to the patient. Second, it
anaesthetizes a large part of a hemisphere and thus cannot pinpoint the
specific regions that are associated with language and memory. Some
scientists have investigated whether fMRI could replace the intracarotid
amobarbital test, as this method is safe and has excellent spatial resolu-
tion (see Chapter 2). To evaluate this possibility, mTLE patients have
undergone the intracarotid amobarbital test and, at a different time,
fMRI during language and long-term memory tasks. Lateralization to
one hemisphere for a particular fMRI contrast is measured by comparing
the number of voxels activated in one hemisphere with the number of
voxels activated in the other hemisphere. For example, a contrast that
isolates verbal long-term memory processing would produce more active
voxels in the left medial temporal lobe than in the right medial temporal
lobe. In one review paper, there was an 80 percent concordance between
hemispheric laterality using the intracarotid amobarbital test and hemi-
spheric laterality using fMRI during language and verbal long-term
memory tasks, and hemispheric laterality using fMRI was not improved
by also considering the intracarotid amobarbital test results (Binder,
2011). In the future, fMRI may completely replace the intracarotid
amobarbital test to more safely and accurately guide surgical planning
for patients with unilateral mTLE.

9.5 Transient Global Amnesia

Although the phenomenon of transient global amnesia (TGA) was
named over half a century ago, its etiology is still unknown.
The following criteria are used to diagnose a patient with TGA: (1)
9.5 Transient Global Amnesia 191

there is clear anterograde amnesia, (2) the attack must last no longer than
24 hours, (3) the individual must not have clouding of consciousness (e.g.,
drowsiness) and they must know their personal identity, (4) the attack
must be witnessed by another person, (5) there should be no other
neurological symptoms during or after the attack (e.g., problems speak-
ing or partial paralysis), and (6) there should be no recent history of head
injury or epilepsy (Hodges & Warlow, 1990; Quinette et al., 2006). TGA
patients often have retrograde amnesia for hours before the attack and
have anterograde amnesia for 1 to 10 hours. They usually repeat the same
questions, such as “where am I?” and “why am I here?” because they
forget that they had already asked a question and received an answer.
The most common events that precipitate an attack are emotional stress,
physical effort, contact with hot or cold water, or sexual intercourse.
TGA patients are usually middle-aged or elderly adults, and accompany-
ing symptoms can include headache, nausea, and dizziness. After some-
one is diagnosed with TGA, the course of treatment is simply to wait for
the amnesia to resolve on its own.
Until about the last decade, neuroimaging techniques did not reveal
any abnormalities in the brains of TGA patients. More recently,
a growing body of evidence has indicated that TGA is caused by lesions
in the CA1 region of the hippocampus. One study used diffusion-
weighted imaging (DWI), an MRI technique that is sensitive to the
diffusion of water that tracks white matter pathways, to assess whether
there were any brain abnormalities in 20 TGA patients (Yang, Kim &
Kim, 2008). DWI was conducted 5 to 23 hours after the onset of attack.
Figure 9.6 shows brain images of 6 representative TGA patients that
show small lesions in the hippocampus (indicated by white arrows). All
20 patients had at least one 1 to 3 millimeter lesion in the lateral aspect of
the hippocampus, which corresponds to the CA1 sub-region (see
Figure 9.2C; the lateral CA1 sub-region is in yellow to the left).
Another study employed standard anatomic MRI and found cavities
that were greater than 3 millimeters within the lateral hippocampus,
which includes the CA1 sub-region, in 14 of 15 TGA patients, as
compared to smaller (less than 2 millimeters) cavities found in control
participants (Nakada, Kwee, Jujii & Knight, 2005). A recent study used
standard anatomic MRI and tested 108 TGA patients 24 to 72 hours after
the onset of attack and found that the large majority of patients had
lesions in the CA1 sub-region of the hippocampus (Döhring, Schmuck &
Bartsch, 2014). These results provide compelling evidence that a lesion to
the CA1 sub-region of the hippocampus causes TGA.
192 Explicit Memory and Disease

Figure 9.6 Brain images of transient global amnesia patients. Representative patients/
cases are shown with arrows indicating lateral hippocampal lesions (for each case, an
axial view is shown to the left with the occipital pole at the bottom and a coronal view is
shown to the right).

TGA patients have also been shown to have impaired performance on

long-term memory tasks. One study employed DWI to look for brain
abnormalities in 14 TGA patients 48 to 72 hours after attack onset
(Bartsch et al., 2010). All of the patients had 1 to 7 millimeter lesions in
the lateral hippocampus within the CA1 sub-region, and there were no
lesions outside of the hippocampus. A few hours after the onset of attack,
the TGA patients took part in a virtual reality maze task where they
solved the maze from different starting points during the learning phase
(i.e., they navigated to the end point). Then, during the test phase, they
solved the maze again three times from different starting points. During
the test phase, TGA patients navigated the maze in a largely random
fashion and took almost three times longer than control participants,
which shows they were impaired on this long-term memory task.
Another study by the same research group showed that 16 TGA patients
with lesions in the CA1 region of the hippocampus had impaired auto-
biographical memory, another type of long-term memory (Bartsch,
Döhring, Rohr, Jansen & Deuschl, 2011; this study was discussed in
Chapter 3). It is also notable that the hippocampal lesions that are
Chapter Summary 193

apparent within a few days of the TGA attack are no longer visible 4 to 6
months later (Bartsch et al., 2010).
The previous findings provide compelling evidence that TGA is caused
by a temporary lesion in the CA1 region of the hippocampus. This is
consistent with the critical role of the hippocampus during long-term
memory (see Chapter 3). Intriguingly, the mechanism underlying hippo-
campal lesions in TGA patients is still unknown. One hypothesis is that
TGA patients have blood flow problems due to vascular blockage.
However, TGA patients do not have greater vascular risk factors, such
as high blood pressure, high cholesterol, or diabetes, than healthy control
participants (Quinette et al., 2006). The only risk factor that has been
associated with TGA is a history of migraine headaches. As emotional or
physical stress almost always triggers TGA attacks and stress can produce
changes in blood flow, it may be that hippocampal CA1 lesions are due
to stress-induced decreases in blood flow to this sub-region.
The hippocampal CA1 sub-region may be particularly susceptible to
reductions in blood flow because it is supplied by one large artery,
while the other hippocampal sub-regions are supplied by one large artery
and many small arteries (Yang et al., 2008). Although the mechanisms
underlying TGA still remain a mystery, the temporary focal lesions in the
hippocampal CA1 sub-region of TGA patients provide a unique oppor-
tunity for future collaborations between cognitive neuroscientists and
neurologists to investigate the specific role of this region in long-term
memory.

Chapter Summary
• Patients with aMCI have impaired long-term memory that is caused by
atrophy of the hippocampus and the entorhinal cortex within the
medial temporal lobe.
• aMCI patients, relative to control participants, have an increase in the
magnitude of fMRI activity within the hippocampus during long-term
memory.
• As compared to aMCI patients, early AD patients have a greater
impairment in long-term memory, more severe atrophy of the medial
temporal lobe, and also have atrophy of the parietal lobe.
• AD patients have abnormally high levels of tau protein in the medial
temporal lobe, which produces neurofibrillary tangles, and amyloid-β
protein in cortical regions, which produces amyloid plaques.
• The behavioral performance of mTBI patients and control partici-
pants is similar during working memory tasks, but mTBI patients
194 Explicit Memory and Disease

have a greater extent and magnitude of fMRI activity within the

dorsolateral prefrontal cortex and the parietal cortex.
• In mTLE patients, removal of left medial temporal lobe regions
produces deficits in verbal memory and removal of right medial
temporal lobe regions produces deficits in visual memory.
• TGA patients have anterograde amnesia for less than 24 hours, and it
is almost always triggered by emotional or physical stress.
• TGA is caused by a temporary lesion to the CA1 sub-region of the
hippocampus, although the mechanisms underlying such lesions are
unknown.

Review Questions
What are the differences in brain anatomy and fMRI activity between
aMCI patients and control participants?
Which brain regions atrophy in early AD patients and what protein
accumulates in each region?
Do mTBI patients with no symptoms and no history of concussion have
a greater or lesser extent of fMRI activity than control participants
during working memory?
Based on mTLE surgical outcomes, is the left or right medial temporal
lobe associated with verbal long-term memory?
Which hippocampal sub-region is lesioned in TGA patients?

Further Reading
Yassa, M. A., Stark, S. M., Bakker, A., Albert, M. S., Gallagher, M. & Stark, C. E.
(2010). High-resolution structural and functional MRI of hippocampal
CA3 and dentate gyrus in patients with amnestic Mild Cognitive
Impairment. NeuroImage, 51, 1242–1252.
This fMRI study shows that aMCI patients, as compared to control
participants, have impaired behavioral performance and increased activity
within the hippocampal CA3/dentate gyrus sub-region during pattern
separation, a type of long-term memory.
Buckner, R. L., Snyder, A. Z., Shannon, B. J., LaRossa, G., Sachs, R.,
Fotenos, A. F., Sheline, Y. I., Klunk, W. E., Mathis, C. A., Morris, J. C. &
Mintun, M. A. (2005). Molecular, structural, and functional
characterization of Alzheimer’s disease: Evidence for a relationship
between default activity, amyloid, and memory. The Journal of
Neuroscience, 25, 7709–7717.
Further Reading 195

This highly influential paper hypothesizes that default network activity

causes amyloid deposition that leads to cortical atrophy and long-term
memory dysfunction in AD patients.
McAllister, T. W., Sparling, M. B., Flashman, L. A., Guerin, S. J.,
Mamourian, A. C. & Saykin, A. J. (2001). Differential working memory
load effects after mild traumatic brain injury. NeuroImage, 14, 1004–1012.
This working memory fMRI study shows that mTBI patients, as compared
to control participants, have similar behavioral performance and increased
activity within the dorsolateral prefrontal cortex and the parietal cortex.
Jones-Gotman, M. (1986). Right hippocampal excision impairs learning and
recall of a list of abstract designs. Neuropsychologia, 24, 659–670.
This mTLE patient study shows that long-term memory for visual designs is
impaired following removal of right medial temporal lobe regions,
particularly if larger hippocampal regions are resected.
Bartsch, T., Schönfeld, R., Müller, F. J., Alfke, K., Leplow, B., Aldenhoff, J.,
Deuschl, G. & Koch, J. M. (2010). Focal lesions of human hippocampal
CA1 neurons in transient global amnesia impair place memory. Science,
328, 1412–1415.
This study shows that TGA patients, as compared to control participants,
have a lesion in the CA1 sub-region of the hippocampus and are severely
impaired on a virtual reality maze long-term memory task.
CHAPTER TEN

Long-Term Memory in Animals

Learning Objectives
• To identify the regions of the medial temporal lobe that are associated
with item memory, context memory, and binding item information
and context information in rats, cats, and monkeys.
• To understand how long-term potentiation links cortical regions to the
hippocampus.
• To compare the brain regions that have been associated with memory
replay in rats and the brain regions associated with episodic memory in
humans.
• To detail the paradigms that have been used to uncover time cells in the
hippocampus of rats and monkeys.
• To describe one type of behavioral evidence and one type of brain
evidence that indicates mammals have episodic memory.

This book is on the cognitive neuroscience of memory, so why is there

a chapter on animal memory? One reason is that the same brain
processes associated with memory in animals are often associated with
memory in humans. These can be considered core brain processes that
mediate memory across species. A second reason is that certain techni-
ques can be used only on animals, such as targeted single-cell recording
and brain lesions. The results of such techniques offer a detailed view
into the brain mechanisms underlying memory that is not available in
humans. This chapter focuses on long-term memory in animals, which
relates to the large majority of research conducted with humans.
Section 10.1 shows that rats, cats, and monkeys have a medial temporal
lobe organization that is the same as humans. The perirhinal cortex is
associated with item memory, the parahippocampal cortex is associated
with context memory, and the hippocampus is associated with binding
item information and context information. In section 10.2, long-term
potentiation in the hippocampus is discussed, which is the mechanism
by which cortical regions link to the hippocampus. Section 10.3 reviews
evidence for memory replay in rats, which refers to reactivation of the
same brain regions in the same or the reverse temporal sequence that
10.1 The Medial Temporal Lobe 197

were activated during a previous event. Such replay activity has been
reported in the hippocampus, the prefrontal cortex, the parietal cortex,
and the visual sensory cortex, which are the same regions that have been
associated with episodic memory in humans (see Chapter 3). In section
10.4, time cells in the rat hippocampus are discussed. Time cells are active
at specific moments after the beginning of an event. The last section, 10.5,
considers the behavioral evidence and the brain evidence that indicates
animals have episodic memory. This has proven to be a controversial
topic because animals cannot tell us whether or not they “remember.”
However, a cumulating body of evidence indicates that animals, and in
particular mammals, have episodic memory.

10.1 The Medial Temporal Lobe

Evaluating whether an item is “old” or “new” is one of the most basic
forms of memory and is referred to as item memory (see Chapter 1).
In human item memory paradigms, items are presented during the
study phase and then during the test phase old and new items are
presented and participants make “old”–“new” recognition judgments.
Because humans cannot communicate with animals, we cannot simply
ask them to make “old”–“new” recognition judgments. Therefore,
unique paradigms have been developed to investigate item memory
in animals.
The spontaneous object recognition task was relatively recently
developed to investigate item memory in animals (Winters, Saksida
& Bussey, 2008). As shown in Figure 10.1, during the sample/study
phase a rat explores two identical objects (e.g., soccer balls) for
a limited amount of time. The rat is then separated from the study
objects during the retention delay, which lasts minutes to ensure sub-
sequent performance is based on long-term memory rather than work-
ing memory. In this illustration, the rat is placed in the left part of the
chamber and the sliding door is lowered before the retention delay.
One of the two objects is replaced by a new object and then the sliding
door is lifted and the rat is allowed to explore the objects. Rats have
a preference for exploring the new item (e.g., the inverted cup), which
indicates it recognizes – and is less interested in – the old item (e.g., the
soccer ball). This behavior reflects a type of item memory. Rats with
lesions restricted to the perirhinal cortex, within the medial temporal
lobe, are impaired on the spontaneous object recognition task,
while rats with lesions restricted to the hippocampus are not usually
impaired on this task (Winters et al., 2008; Eichenbaum, Sauvage,
198 Long-Term Memory in Animals

Sample Phase Choice Phase

Retention
Delay
(variable)

Figure 10.1 Spontaneous object recognition task. Left, during the sample/study phase the
rat explores two identical objects. Middle, the rat is separated from the study objects by
a sliding door during the retention delay (i.e., the rat moves to the left and the door is
lowered). Right, during the choice/test phase, the door is lifted and the rat is allowed to
explore the objects.

Fortin, Komorowski & Lipton, 2012). This indicates the perirhinal

cortex is associated with item memory.
The delayed nonmatching-to-sample task is similar to, but more
complicated than, the spontaneous object recognition task and was
also developed to test item memory in animals. During the sample/
study phase, one item is presented (e.g., a card with a plus sign on it)
followed by a delay period that lasts minutes. During the test phase,
the old item and a new item (e.g., a card with a square on it) are
each presented on top of a bowl. Only one of the bowls has a food
reward in it. The task is to select the nonmatching/new item (i.e., the
square in this example), which always covers the food, and the
animal can eat if it makes the correct selection. The animals are
hungry during these trials and thus motivated to select the new item,
which depends on recognition of the old item. This task requires
selection of the nonmatching/new item, rather than the matching/old
item, because the old item will be associated with repetition priming,
a type of implicit memory (see Chapters 1 and 7), and the aim is to
require the animal to respond based on long-term memory. Note
that in the field of behavioral neuroscience (see Chapter 1), explicit
memory (including long-term memory) and implicit memory are
referred to as declarative memory and nondeclarative memory,
respectively. One drawback of the delayed nonmatching-to-sample
task is that the animal requires extensive training to perform the
10.1 The Medial Temporal Lobe 199

task, which has been a basis for criticism that performance on this
task is based on only long-term memory. The concern is that a task
practiced hundreds or thousands of times during the training period
will reflect performance of a skill, which is a largely nonconscious
process (see Chapter 7). Putting this potential concern aside, rats
and monkeys with lesions restricted to the perirhinal cortex are
impaired on the delayed nonmatching-to-sample task, while rats
with lesions restricted to the hippocampus or parahippocampal
cortex are not impaired on this task (Eichenbaum, Yonelinas &
Ranganath, 2007; Winters et al., 2008; Eichenbaum et al., 2012).
This is the same pattern of results observed for the spontaneous
object recognition task and provides convergent evidence that the
perirhinal cortex is associated with item memory.
The lesion evidence described above that links the perirhinal cortex to
item memory in animals is consistent with the fMRI findings in humans
that the perirhinal cortex is associated with item memory, the parahip-
pocampal cortex is associated with context memory, and the hippocam-
pus is associate with binding item information and context information
(see Chapter 3). Anatomic studies of the medial temporal lobe in rats,
cats, and monkeys indicate that this medial temporal lobe organization is
conserved across other mammals as well (Manns & Eichenbaum, 2006).
Figure 10.2A illustrates the organization of the medial temporal lobe for
rats, cats, and monkeys. Cortical regions associated with nonspatial/
object processing input to the perirhinal cortex (PER), the lateral entorh-
inal area/cortex (LEA), and the CA1 and subiculum (SUB) sub-regions
of the hippocampus. Cortical regions associated with spatial/context
processing input to the postrhinal/parahippocampal cortex (POR), the
medial entorhinal area/cortex (MEA), and the CA3 and dentate gyrus
(DG) sub-regions of the hippocampus. These separate medial temporal
lobe pathways are extensions of the what pathway and the where
pathway, respectively (see Chapter 1), that converge onto the hippocam-
pus. Figure 10.2B shows the phylogenic/evolutionary tree for selected
mammals. It is notable that rats and cats split off from a common
ancestor with humans about 100 million years ago, yet the organization
of the medial temporal lobe has been highly conserved across these three
species. Note that the findings in monkeys (such as the macaque
monkey, which is shown in the figure near humans on the tree) apply to
humans, since they are our close evolutionary relatives, and thus there is
a very similar organization to all primate brains (including monkeys,
apes, and humans). The findings in this section provide convergent
evidence that the perirhinal cortex is associated with item memory, the
200 Long-Term Memory in Animals

A B Kangaroo

nonspatial (object) cortical input spatial (context) cortical input Hedgehog

Bat
Cat
Pig

Tree Shrew

PER POR
Macaque
Human

LEA MEA

Mouse
CA1/SUB CA3/DG Rat
Tenrec
n-1 Echidna
n-2

200 100 0
Years Prior to Present Day (in millions)

Figure 10.2 Medial temporal lobe organization and phylogenic tree of mammals.
(A) Schematic of the medial temporal lobe based on findings from rats, cats, and monkeys.
Left, nonspatial/object pathway from cortical input to the perirhinal cortex (PER) to the
lateral entorhinal area/cortex (LEA) to the CA1 and subiculum (SUB) sub-regions of the
hippocampus. Right, spatial/context pathway from cortical input to the postrhinal/
parahippocampal cortex (POR) to the medial entorhinal area/cortex (MEA) to the CA3 and
dentate gyrus (DG) sub-regions of the hippocampus. Arrows illustrate connections between
regions and direction of information flow. (B) Phylogenic/evolutionary tree of selected
mammals.

parahippocampal cortex is associated with context memory, and the

hippocampus is associated with binding item information and context
information in all mammals.

10.2 Long-Term Potentiation

As described in section 10.1, the hippocampus binds item information
and context information. For instance, if someone was with a group of
friends celebrating at a restaurant, the sights, sounds, and smells would
produce activity in multiple cortical regions, and this activity would be
linked through the hippocampus. If the same person visited the restau-
rant a few months later, the smell of the restaurant upon entering might
produce the same pattern of activity in the olfactory cortex of that person
and, via the previously established links through the hippocampus, trig-
ger reactivation of visual cortex and auditory cortex that reflect episodic
10.2 Long-Term Potentiation 201

memory of the previous celebration. One line of behavioral neuroscience

research has focused on understanding the mechanisms by which cortical
regions link to the hippocampus, which is referred to as long-term
potentiation.
Before detailing the mechanisms of long-term potentiation, the basics
of communication between neurons will be briefly reviewed. Neurons
usually consist of short dendrites, which receive input from other neu-
rons, a cell body, and a long axon, which transmits information to another
neuron. If the sum of the inputs to a neuron produce a sufficiently
positive voltage, the axon fires (i.e., there is an action potential) starting
near the cell body and this action potential travels to the axon terminal
far from the cell body (neuronal firing can be measured using depth
electrode recording; see Chapter 2). There is a synaptic cleft between
the axon terminal of the pre-synaptic neuron and the dendrite of the post-
synaptic neuron. The action potential at the axon terminal causes
the release of a neurotransmitter (a chemical substance that allows for
communication between neurons) from the pre-synaptic neuron.
The neurotransmitter traverses the synaptic cleft and binds to specific
receptors – proteins embedded in the cell wall that act as gateways for
positive or negative ions – on the dendrite of the post-synaptic neuron.
This changes the structure of the receptors so that positive or negative
ions flow into the dendrite of the post-synaptic neuron, which increases or
decreases the amplitude of the voltage/potential of the post-synaptic
neuron. Glutamate is the primary neurotransmitter that produces the
flow of positive ions into the dendrite of the post-synaptic neuron, which
give rise to an excitatory post-synaptic potential (i.e., an increase in
voltage) that can cause the axon of the post-synaptic neuron to fire if
the sum of all the inputs produces a sufficiently positive voltage.
As described below, long-term potentiation produces an increase in the
amplitude and rate of neuronal firing.
Long-term potentiation was discovered over 40 years ago using depth
electrode stimulation and recording in the hippocampus of rabbits
(Bliss & Lømo, 1973). Figure 10.3A illustrates the experimental setup,
which consisted of a stimulating electrode placed in the perforant path
(PP), which connects the entorhinal cortex to the hippocampus (see
Chapter 9), and a recording electrode placed in the dentate gyrus (D)
sub-region of the hippocampus. In one experiment, the perforant path
electrode was stimulated at 10 to 15 Hertz for 10 seconds (i.e., 100 to 150
pulses) every 30 minutes to 1 hour in a 3-hour period (for a total of 4
trains of stimulation). The neuronal response in the dentate gyrus was
also tested at regular intervals by stimulating the perforant path electrode
202 Long-Term Memory in Animals

A Rec B

CA 1 Stim

Sub
CA 3 D 2 mV
PP
10 msec

Figure 10.3 Long-term potentiation experimental setup and results. (A) Depiction of the
experimental setup in the rabbit hippocampus with the stimulating electrode (Stim) in the
perforant path (PP) and the recording electrode (Rec) in the dentate gyrus (D) sub-region.
The arrow indicates the direction of information flow from the entorhinal cortex (not shown)
to the hippocampus. Other hippocampal sub-regions are also shown including CA3, CA1, and
the subiculum (Sub). (B) Neural responses after conditioning (solid lines), which illustrate
long-term potentiation, and neural response before conditioning (dotted line; millivolt by
millisecond key at the bottom right).

and measuring the action potential with the recording electrode.

As shown in Figure 10.3B, the dentate gyrus action potentials approxi-
mately 3 hours after the last train of conditioning stimulation (solid lines)
had higher amplitudes and more rapid responses than the action poten-
tials before conditioning (dotted line). This increase in amplitude and
response rate following conditioning reflects long-term potentiation in
the hippocampus. In another experiment in the same study, long-term
potentiation was produced by stimulating with 1 train at 100 Hertz for 3
to 4 seconds, which showed that higher frequency stimulation for
a shorter duration induced long-term potentiation. Another study stimu-
lated and recorded from rat hippocampal slices to evaluate whether the
length of time between stimulation trains would affect long-term poten-
tiation (Larson, Wong & Lynch, 1986). Each of the stimulation trains
consisted of 4 pulses at 100 Hertz and there were 5 to 20 stimulation trains
separated by intervals of 0.1, 0.2, 1.0, or 2.0 seconds (i.e., 10.0, 5.0, 1.0, or
0.5 Hertz). Long-term potentiation was measured by the amplitude of the
neuronal response following conditioning compared to the amplitude of
the response before conditioning. Of primary importance, an inter-burst
frequency of 5.0 Hertz produced an approximately 25 percent increase in
neuronal response amplitude, which was nearly double the increase at
other inter-burst frequencies. This finding shows that conditioning with
10.3 Memory Replay 203

theta frequency bursts maximizes long-term potentiation and supports

the evidence indicating that activity in this frequency band reflects the
interaction between the hippocampus and cortical regions during long-
term memory (see Chapter 4). In the previous studies, high-frequency
stimulation during conditioning is a model for how active cortical regions
connect to the hippocampus. The long-term potentiation that develops
can be assumed to reflect the link between a cortical region and the
hippocampus.
The molecular mechanisms underlying long-term potentiation in the
hippocampus are a major topic of investigation within the field of
neuroscience. Long-term potentiation is caused by many cellular cas-
cades involving interactions between ions, cellular machinery, and
receptors. An excitatory post-synaptic potential can lead to the mod-
ification of current receptors, the addition of new receptors, and an
increase in the surface area of dendrites (Bliss & Collingridge, 1993;
Baudry et al., 2015). All of these changes make the post-synaptic
neuron more sensitive to subsequent neurotransmitter release. This is
another way of describing a stronger link between the pre-synaptic
neuron and the post-synaptic neuron that gives rise to long-term
potentiation. It should also be mentioned that there is hippocampal
long-term depression in the hippocampus, which refers to a decrease in
neuronal response magnitude following activation (Bear & Abraham,
1996; Kemp & Manahan-Vaughan, 2007). As cognitive neuroscience
techniques can measure only neural firing rate (see Chapter 2), the
molecular mechanisms of long-term potentiation and long-term
depression are beyond the scope of this book.

10.3 Memory Replay

Slow wave sleep is reflected by EEG modulation of less than 1 Hertz and
is important for memory consolidation in humans (see Chapter 3).
Such slow waves synchronize other brain waves such as hippocampal
sharp-wave ripples that oscillate at a frequency of approximately
200 Hertz. These sharp-wave ripples coordinate hippocampal–cortical
interactions that reflect the replay of memories, which strengthens these
memories and results in consolidation.
In rats, hippocampal sharp-wave ripples similarly orchestrate consoli-
dation during slow wave sleep and have also been shown to coordinate
memory replay during non-exploratory waking states such as eating,
drinking, grooming, or quiet wakefulness (O’Neill, Pleydell-Bouverie,
Dupret & Csicsvari, 2010; Girardeau & Zugaro, 2011). Figure 10.4
204 Long-Term Memory in Animals

Figure 10.4 Memory replay in the rat. Left, during exploration the rat walks down a path
from the starting point (left circle) to the end point (right circle with dots/food). Hippocampal
theta activity is shown below the path (in green). Place cells in the hippocampus (numbered 1
through 4) fire when the rat is in a specific region of the path. Middle, when the rat reaches
the food, a hippocampal sharp-wave ripple (in blue) coordinates the same place cells to
rapidly fire in reverse order. Right, during slow wave sleep, a hippocampal sharp-wave ripple
coordinates the same place cells to rapidly fire in forward order. (A black and white version of
this figure will appear in some formats. For the color version, please refer to the plate
section.)

illustrates memory replay in a rat. During exploration (Figure 10.4, left),

the rat walks down a path from the starting point (left, the empty circle)
to the end point (right, the circle with dots, which represent food).
Hippocampal theta activity occurs during this exploration (in green
below the path), which likely reflects the interaction between the
hippocampus and cortical regions during episodic memory encoding
(see section 10.2 and Chapter 4). Place cells in the hippocampus
(numbered 1 through 4) fire (illustrated by the vertical bars below the
cells) when the rat is in a specific region of the path and correspond to the
location of the rat as it moves from the starting point to the end point.
Such place cells in the rat hippocampus have long been known to code for
particular locations in space and represent a spatial map of the environ-
ment (O’Keefe & Dostrovsky, 1971). Hippocampal place cells have also
been reported in monkeys (Matsumura et al., 1999) and humans
(Ekstrom et al., 2003). When the rat reaches the food (Figure 10.4,
middle), a hippocampal sharp-wave ripple (in blue below the path)
coordinates the same place cells to fire. There are two interesting aspects
of the firing pattern. First, the cells fire in reverse order (i.e., cell 4 then
cell 3 then cell 2 then cell 1). Second, the firing is compressed in time. This
rapid reverse replay of the memory may reflect the rat imagining the path
it just took from the food back to the starting point, perhaps to
strengthen the memory of this route so the rat can return to the starting
10.4 Time Cells 205

location after it eats. During slow wave sleep (Figure 10.4, right),
a hippocampal sharp-wave ripple again coordinates the rapid firing of
the same place cells, but this time in forward order. This forward replay
may reflect strengthening the memory from the starting point to the end
point, so the rat can return to the food if it finds itself at the starting point
again in the future.
Memory replay in rats has also been shown in other cortical regions
beyond the hippocampus. Synchronous activity between the hippocam-
pus and the prefrontal cortex has been shown to occur during spatial
memory replay (Preston & Eichenbaum, 2013), which suggests that these
regions interact during episodic memory consolidation. Replay of activity
that occurred during spatial exploration has also been reported in the
visual sensory cortex (Ji & Wilson, 2007) and the parietal cortex (Qin,
McNaughton, Skaggs & Barnes, 1997), which can be assumed to
reflect the reinstatement of the detailed visual experience that occurred
previously. These findings show that spatial memory replay produces
activity in the same regions that have been associated with episodic
memory in humans: the hippocampus, the prefrontal cortex, the parietal
cortex, and the sensory cortex (see Chapter 3).
Hippocampal sharp-wave ripples appear to be critical for episodic
memory consolidation in rats and humans. They have also been observed
in every other mammalian species that has been tested including mice,
bats, rabbits, cats, and monkeys (Buzsáki, 2015). In contrast, hippocam-
pal sharp-wave ripples and slow wave sleep have not been reported in
birds (Rattenborg, Martinez-Gonzalez, Roth & Pravosudov, 2011).
These findings suggest that hippocampal sharp-wave ripple coordinated
memory replay is a mechanism of episodic memory consolidation that
has been conserved across many species (the furry ones), but not all
species (the feathery ones).
Although there is some evidence that slow wave sleep is important for
episodic memory consolidation in humans (see Chapter 3), there is no
evidence at this time in humans for forward memory replay or backward
memory replay during quiet waking states or slow wave sleep. This is
a future area of research in humans that stems directly from the recent
findings in rats.

10.4 Time Cells

Damage to the hippocampus has been shown to selectively impair mem-
ory for the temporal order of previously presented stimuli in both rodents
and humans (Eichenbaum, 2014). For example, in a study with humans, if
206 Long-Term Memory in Animals

four objects (e.g., a shoe, a bowl, a hammer, and an apple) were sequen-
tially presented during the study phase, “old”–“new” recognition mem-
ory would be intact but memory for the order of the items (e.g., ‘did the
shoe come before the hammer?’) would be impaired. This pattern of
results is consistent with the known roles of the medial temporal lobe
regions discussed above, where item memory is associated with the
perirhinal cortex but item-to-context binding is associated with the hip-
pocampus. Here, the temporal position of an item can be considered its
temporal context.
As mentioned in section 10.3, hippocampal place cells in rats have
been investigated for over 40 years. Within the last decade, there has
been an exciting discovery of rat hippocampal time cells. Similar to
a place cell that is active at a specific place (e.g., the middle of a path),
a time cell is active at a specific time (e.g., 5 seconds after the beginning
of an event). There are only a handful of recent studies on time cells.
A PubMed.gov search for the terms ‘hippocampus’ and ‘time cells’
identified only 16 articles at the time this chapter was written. Half of
the articles were reviews and all of the articles were published within
the last 5 years. By comparison, the terms ‘hippocampus’ and ‘place
cells’ identified over 600 articles. One study investigated hippocampal
time cells while each rat was running on a treadmill (Kraus, Robinson,
White, Eichenbaum & Hasselmo, 2013). Figure 10.5A shows the appa-
ratus (the treadmill is at the center in gray), which was a figure-eight
maze. Figure 10.5B shows a schematic of the apparatus and the task.
While recording from 96 cells in the hippocampus (using an array of
single-cell electrodes; see Chapter 2), the rat entered the treadmill, ran
for more than 16 seconds, and then alternated between going through
the right arm of the maze (illustrated by the red arrow) and going
through the left arm of the maze (illustrated by the cyan arrow).
The rats were thirsty and received water at water ports, which moti-
vated them to perform the task. Figure 10.5C shows the response of 21
hippocampal neurons during a 16-second period on the treadmill for
one rat (red corresponds to a higher firing rate). These neurons were
sorted from the top to the bottom from the neuron that responded
earliest in time to the neuron that responded latest in time. The pattern
of activity clearly shows that the neurons were active as a function of
the time the rat was running on the treadmill. A similar pattern was
observed for other treadmill running periods and for other rats. These
findings provide compelling evidence that there are time cells in the
hippocampus. The progressive firing rate in time for time cells is
analogous to the progressive firing rate in space for place cells
10.4 Time Cells 207

A B C

water port
1

treadmill

neuron
21
0 time (s) 16

water port water port

Figure 10.5 Time cell behavioral apparatus and neural activity. (A) The figure-eight maze
apparatus (in red) with the treadmill at the center (in gray; superior view). (B) Schematic
of the apparatus and task. Each rat entered the treadmill, ran for more than 16 seconds,
and then alternated between going through the right arm of the maze (red arrow) and
going through the left arm of the maze (cyan arrow; water ports labeled). (C) Response of
21 hippocampal neurons during a 16-second period on the treadmill for one rat sorted
from the top to the bottom from the neuron that responded earliest in time to the neuron
that responded latest in time (red indicates a higher firing rate). (A black and white
version of this figure will appear in some formats. For the color version, please refer to the
plate section.)

(compare Figure 10.5C to Figure 10.4, left). It is notable that there is

more dispersion for cells that were activated later in time (i.e., there
is a larger spread of activity over time in higher numbered cells). This is
typically observed in hippocampal time cell studies and may reflect the
increase in uncertainty in the amount of time that has passed since the
beginning of the period. One potential problem with these findings is
that the rat was running a longer distance as time passed; therefore,
these could be distance cells rather than time cells. To address this
potential problem, the investigators varied the speed of the treadmill in
different periods that the rats were on the treadmill. In this way, they
could look for hippocampal cells that varied as a function of time but
not distance, and, of importance, these types of cells were observed.
A previous study similarly measured activity in hippocampal cells
while rats ran on a wheel for 10 or 20 seconds and alternated between
the left arm and the right arm of a figure-eight maze (Pastalkova,
Itskov, Amarasingham & Buzsáki, 2008). In that study, cells also fired
as a function of time on the wheel. The activity of these time cells was
related to the phase of theta activity, as has been observed for place
208 Long-Term Memory in Animals

cells. These investigators also measured activity in hippocampal cells

while the rats ran on the wheel for water or ran on the wheel for fun
without the subsequent maze running task, and there was no evidence
of time cell activity. Moreover, the configuration of time cells differed
depending on whether the rat was subsequently going to run through
the left arm of the maze or the right arm of the maze. These findings
suggest that hippocampal cells do not simply measure time, but reflect
cognitive processing such as imagery of a memory-guided future
action. For instance, while a rat is running on a wheel, it may be
visualizing the path to the left that it will need to take to receive
a water reward. The link between time cell activity and subsequent
memory guided action is an intriguing topic of future research.
The previous studies provided evidence for hippocampal time cells
while the rats were running in place. Although the rats were techni-
cally in one place in space, there are confounds with movement such
as changes in trajectory and head motion. In addition, one can
imagine the treadmill or wheel as a linear path that is bent into
a circle and then connected at each end. As it is known that place
cells are activated as rats move down a path, time cells could actually
be place cells that are activated at different locations along the tread-
mill/wheel. One study addressed these potential issues by keeping the
head of each rat motionless during a delayed matching-to-sample task
that used odors as stimuli (MacDonald, Carrow, Place & Eichenbaum,
2013). This task was similar to a delayed non-matching-to-sample object
task (discussed in section 10.1 of this chapter), except the task was to
select the odor at test that matched the sample. There was a 2- to
5-second delay period between the sample odor and the test odor.
Even with completely restricted head motion, hippocampal time cells
were observed during the delay period and fired in coordination with
theta activity. Like the findings above, cells that fired at later time
points were active over a more dispersed time period and a different
configuration of time cells was associated with each of the sample
odors. These results indicate that the hippocampal time cell activity
observed in other studies was not due to confounds associated with
movement.
One study investigated whether there were time cells in the hippocam-
pus of monkeys (Naya & Suzuki, 2011). During the study phase, two
objects were sequentially presented (e.g., a sunflower and then a bowtie).
During the test phase, two old objects and one new object were presented
along the corners of an imaginary triangle (e.g., a sunflower, a pumpkin,
and a bowtie), and the monkey had been trained to select the first object
10.4 Time Cells 209

Box 10.1: Animal research often guides human research

Research with animals is conducted using techniques that cannot be used
with humans, such as targeted single-cell recording. In humans, single-cell
studies are extremely rare and electrode placement is restricted to locations
that will potentially benefit the patient (e.g., the probable location of
a seizure focus; see Chapters 2 and 9). Moreover, the brains of patients
with implanted electrodes are necessarily different from the general popu-
lation, or they wouldn’t have electrodes implanted in their brain.
By comparison, single-cell electrodes can be placed in very specific regions
of interest in animals with normal brain function. This allows for the
detailed investigation of animal brain activity that often leads to new
discoveries, such as time cells in the hippocampus. Now that time cells
have been observed in rats, scientists will look for time cells in humans.
Along the same lines, evidence for memory replay in the hippocampus and
other cortical regions in rats (discussed in section 10.3) will motivate
scientists to look for memory replay evidence in humans. These examples
illustrate that research with animals often guides research with humans.
As will be clear in section 10.5, research with humans can also guide
research with animals.

from the study phase (e.g., the sunflower). Then the selected object
disappeared and the monkey had been trained to select the second
object from the study phase (e.g., the bowtie). Time cells were defined
as having activity that differentiated between the period when the first
object was presented and the period when the second object was pre-
sented during the study phase (i.e., cells that were active during one of
those study periods, but not both periods). By comparison, item cells
were defined as having activity that differentiated between the objects
presented during the study phase rather than the time period. They found
that hippocampal cells were predominantly time cells and perirhinal
cortex cells were predominantly item cells.
The findings of this section indicate that there are time cells in the
hippocampus of rats and monkeys. This is an exciting new line of research
in animals that did not exist a decade ago. Much more work needs to be
done on this topic in animals, such as determining why the configuration
of time cells depends on the subsequent task. As discussed in Box 10.1,
the discovery of hippocampal time cells in animals opens up a new line of
research in human memory.
210 Long-Term Memory in Animals

10.5 Episodic Memory

Episodic memory involves retrieval of what items comprised the event,
where the event took place, and when the event occurred. Retrieval of
such detailed what-where-when information requires mentally traveling
back in time to the previously experienced event. Such mental time travel
is a key component of episodic memories and is associated with the
subjective experience of “remembering” rather than “knowing”
(Tulving, 1985; see Chapter 1).
Humans can report whether they “remember” or “know,” and
“remember” responses can be assumed to reflect episodic memory.
One problem with interpreting the mental states of animals is that
they cannot tell us about their subjective experience (Tulving, 2005;
Suddendorf & Corballis, 2007). Just over a decade ago, Endel Tulving,
who introduced the term episodic memory and linked this process
to mental time travel, concluded that the evidence – at that time –
indicated that animals do not have episodic memory (Tulving, 2005).
It is useful to go through the type of evidence Tulving considered
important in order to highlight what would make a convincing case
that animals have episodic memory. First, Tulving stated that the tasks
employed did not differentiate between item memory and context
memory, and of particular importance did not assess memory for con-
text/temporal memory (this was generally true in 2005). Second,
Tulving stated there was no evidence that different regions of the
medial temporal lobe were associated with item memory and context
memory. The dominant view at the time was that long-term memory,
which includes both item memory and context memory, was associated
with the medial temporal lobe as a unified system (this is now the
alternative model of medial temporal lobe function; see Chapter 3).
In other words, Tulving’s criteria for episodic memory is behavioral
evidence that animals have memory for temporal information and
brain evidence that a distinct region of the medial temporal lobe is
associated with context/temporal memory.
In the last decade, behavioral evidence has accumulated that indicates
mammals have memory for temporal information (i.e., Tulving’s first
requirement for episodic memory in animals). In one study, rats were
evaluated to assess whether they could remember what type of food they
had eaten, where they had eaten it, and when they had eaten it (Babb &
Crystal, 2006). Figure 10.6A illustrates the first phase of one trial in which
rats were allowed to retrieve grape-, raspberry-, and chow-flavored food
pellets from four arms of an eight-arm radial maze. After the first phase,
10.5 Episodic Memory 211

A B C C C D Raspberry
1.0
G C G Grape
0.8

p(Revisits)
C C C
0.6
R C R C 0.4
C C
G Grape
0.2
R Raspberry
C Chow 0.0
– Closed door
Short Long
Retention Interval

Figure 10.6 Time delay memory task and behavioral results. (A) In the first phase of
one trial rats were allowed to retrieve grape-, raspberry-, and chow-flavored food
pellets from four arms of an eight-arm radial maze (key at the bottom). (B) In short/
1-hour delay trials, chow-flavored pellets were placed in the locations that were not
available in the first phase. (C) In long/6-hour delay trials, chow pellets were placed in
the locations that were not available in the first phase and the grape- and the
raspberry-flavored pellets were placed in the same locations as in the first phase.
(D) Probability that the rats revisited the raspberry arm and the grape arm as a function
of retention interval (key at the top).

there was either a 1-hour delay period or a 6-hour delay period.

Figure 10.6B illustrates the second phase of the trial that occurred after
a short/1-hour delay in which chow-flavored pellets were placed in the
locations that were not available in the first phase. Figure 10.6C illus-
trates the second phase of the trial that occurred after a long/6-hour delay
in which chow pellets were placed in the locations that were not available
in the first phase and the grape- and the raspberry-flavored pellets were
placed in the same locations as in the first phase. Figure 10.6D shows that
at the short delay rats rarely visited the grape- or the raspberry-flavored
pellet arms of the maze, while at the long delay the rats frequently visited
those arms. This suggests that the rats remembered the time delay (i.e.,
1 hour ago or 6 hours ago) between the first phase and the second phase,
which influenced their behavior. In a follow-up experiment, one of the
non-chow-flavored pellets was devalued by pairing that flavor with
a substance that caused taste aversion during the delay period. This
caused the rats to almost completely avoid visiting the corresponding
arm of the radial maze during the second phase, but did not cause them to
reduce visits to the other non-chow-flavored pellet. This indicates that
the rats remembered what type of food pellet was in each arm of the
radial maze during the first phase. These results provide evidence that
rats can remember what-where-when information.
212 Long-Term Memory in Animals

An earlier study in scrub jays tested for temporal memory in birds

(Clayton & Dickinson, 1998). The experiment capitalized on the fact that
these birds store/cache their excess food for future use. During the first
phase, they stored either worms or peanuts in a specific area. During
the second phase 120 hours later, they stored the opposite type of food in
a different area (e.g., if they stored peanuts the first time, they stored
worms the second time). During the third phase 4 hours later, they were
allowed to recover their food of choice. Of importance, the birds prefer to
eat the worms, but these decayed after too long and so were only edible
after the 4-hour delay between the second and third phases (i.e., if the
worms were buried in the second phase). Peanuts were edible regardless
of when they were cached. On trials in which worms were buried in
the second phase (and thus were edible), the birds preferentially recov-
ered the worms, but on trials in which worms were buried in the first
phase (and thus were inedible), the birds recovered only the peanuts.
These results indicate that the birds remembered where and when each
type of food was cached. Tulving (2005) noted that this evidence in scrub
jays can be attributed to memory for temporal information and appeared
to reflect episodic memory, but that such evidence had not been reported
for other animals.
The previous behavioral results suggest that rats and scrub jays have
memory for temporal information, which could reflect episodic memory.
However, one criticism of these findings is that the behavior could alter-
natively be based on familiarity, which is why it has been referred to as
episodic-like memory. That is, the rats could have used the following rule:
if the non-chow food arm is more familiar (at the 1-hour delay) do not
seek food there, but if the non-chow food arm is less familiar (at the
6-hour delay) seek food there. The scrub jays could have used the
following rule: if the worm cache area is more familiar (at the 4-hour
delay) recover food there, but if the worm cache area is less familiar (at
the 124-hour delay) recover food in the other (peanut cache) area. Such
rules seem complicated, but they show there is a possibility that this type
of behavioral evidence may reflect familiarity rather than episodic
memory.
There is additional behavioral evidence for temporal memory in ani-
mals that cannot be attributed to familiarity. In one study, pigs entered
feeding crates that were painted with blue stripes or red leaves (Špinka,
Duncan & Widowski, 1998). After they entered the crates and ate, they
were confined to the crates with blue stripes for 30 minutes and confined
to the crates with red leaves for 240 minutes. For 16 days, they were
forced to alternate between the two types of crates each morning and
10.5 Episodic Memory 213

were allowed to choose the type of crate to enter in the afternoon. On the
first day, half of the pigs chose the crates with blue stripes and half of the
pigs chose the crates with red leaves (as expected by random selection).
However, on the last day, the majority (75 percent) of the pigs chose the
crates with blue stripes. These findings indicate that pigs prefer short
confinement and also suggest that they associated the crates with blue
stripes and red leaves with a shorter confinement time and a longer
confinement time, respectively. That is, the pigs appear to have remem-
bered the duration of time that was associated with each type of crate,
which can be assumed to reflect episodic memory.
The impressive memory ability of dolphins is routinely displayed dur-
ing performances, such as those at SeaWorld. During these shows,
a dolphin performs a complex series of actions based on a gesture made
by a human. Memory for these actions can be assumed to be largely
implicit, as they can be considered highly practiced skills (see Chapter 7).
However, in a clever study, bottlenosed dolphins have been shown to
think back in time to a previous experience (Mercado, Murray,
Uyeyama, Pack & Herman, 1998). The dolphins tested could perform
over sixty different behaviors, such as ‘swim in a circle, belly up, fins
waving’, ‘toss an object with tail’, or ‘leap belly up, mouth open, fins
waving’. The dolphins generally performed the specific commands indi-
cated by the gestures, which is not surprising given that this is what
a hungry show dolphin will do for a fish reward. One of the commands
that these dolphins were trained to perform was referred to as the creative
command, which is of particular interest because it signaled the dolphin
to perform a behavior that had not been done recently. Accurate perfor-
mance of the creative command did not reflect familiarity because more
recent behaviors were more familiar. Repeating a non-recent behavior
required the dolphin to think back in time to the previous behaviors it
had performed and recall one that was not recent. Moreover, the process
of recall, rather than recognition, can be assumed to reflect episodic
memory.
All of the preceding behavioral results used tasks that could be
described as somewhat complicated. The most straightforward task that
depends on memory for temporal information is the temporal order task.
Rats and monkeys have proven to be capable of accurately reporting the
temporal order of previously presented odors or objects (see section 10.4
of this chapter). Future studies in other animals that appear to be capable
of memory for temporal information, such as pigs and dolphins (and
elephants, see below), should also be tested on these relatively simple
temporal order tasks.
214 Long-Term Memory in Animals

A Kangaroo Fruit Bat Echidna B

Elephant

1 cm Chinchilla 5 mm Pig
1 cm
Hedgehog
hippocampus

5 cm

5 mm
5 mm 1 cm

Figure 10.7 Hippocampal anatomy in mammals. (A) Brains of mammals (labeled at the top
of each image) with darkly stained hippocampal dentate gyrus and CA sub-regions (coronal
view; scale bar at the bottom left).

Even more compelling than the behavioral findings is a large body

of evidence that has accumulated in the last decade that one region of
the medial temporal lobe – the hippocampus – is associated with
context/temporal memory in mammals (i.e., Tulving’s second require-
ment for episodic memory). Research with rats and monkeys has
shown that item memory is associated with the perirhinal cortex and
context memory is associated with the hippocampus (see section 10.1
and Eichenbaum et al., 2007). Moreover, the organization of the
hippocampus is highly conserved across rats, cats, and monkeys.
As shown in Figure 10.7A, other mammalian species have a similar
hippocampal organization, including dentate gyrus and CA sub-
regions (darkly stained in the coronal slice of each animal).
Figure 10.7B shows an anatomic image of a male African elephant
brain with the hippocampus labeled. The elephant hippocampus is of
similar complexity to humans and more complex than other mam-
mals, including more layers in the dentate gyrus and more connec-
tions between neurons (Patzke et al., 2014). Elephants are also known
to have incredible spatial memory (Hart, Hart & Pinter-Wollman,
2008). For instance, during times of drought, the matriarch will lead
the herd for hundreds of miles toward water holes. Moreover, mem-
ory for the location of a specific water hole (e.g., a potential source of
water during a severe drought) can last for decades. The similarity
between the elephant hippocampus and the human hippocampus
coupled with the remarkable spatial memory of elephants provide
compelling evidence that they have episodic memory.
10.5 Episodic Memory 215

Hippocampal time cells, which have been observed in rats and mon-
keys, are active at specific periods following the onset of an event, such as
when a rat is running on a treadmill, but are only active if there is
a subsequent to-be-remembered task (e.g., if the rat is supposed to turn
left at the fork in the maze the next time after getting off the wheel; see
section 10.4 of this chapter). This dependence in the configuration of time
cell activity on a to-be-remembered task suggests these cells are asso-
ciated with imagery of a future action, which requires mental time travel,
the key characteristic of episodic memory. Such time cells provide a brain
mechanism that is likely utilized during episodic memory in rats and
other mammals.
Some of the most compelling evidence that animals can have episodic
memory stems from the discovery of memory replay in the hippocampus,
as discussed in section 10.3 of this chapter. Memory replay refers to the
reactivation of brain activity associated with a previous experience in the
correct temporal order. Memory replay, which has predominantly been
observed in rats, has been shown to be coordinated by hippocampal
sharp-wave ripples during periods of slow wave sleep and quiet wakeful-
ness. A study of bottlenose dolphins also showed evidence for memory
replay during periods of sleep or rest (Kremers, Jaramillo, Böye,
Lemasson & Hausberger, 2011). The dolphins heard recorded humpback
whale sounds – a 14-second sequence of five calls repeated eight times –
that were broadcast at the beginning of shows approximately two to three
times per day for many days. The whale sounds are very different from
the whistles and burst-pulsed vocalizations typically made by dolphins.
Sounds from the dolphins were recorded during subsequent days and
nights. It was found that the dolphins made whale-like sound produc-
tions, mostly at night but also during quiet restfulness while swimming
slowly or floating. Such sounds were never observed before the dolphins
heard the whale sounds. Furthermore, these whale-like productions
appeared to be speeded up in time, which is reminiscent of the memory
replay in the brain that is compressed in time that has been observed in
rats (as discussed in section 10.3 of this chapter). A set of human obser-
vers classified actual humpback whale sounds, dolphin whistles, the
dolphin whale-like productions played at normal speed, and the dolphin
whale-like productions played at half speed. As expected, the dolphin
whistles were classified as produced by a dolphin. Of most importance,
a similar percentage of dolphin whale-like productions played at half
speed and humpback whale sounds were classified as produced by
a whale. These findings suggest that dolphins, like rats, have memory
replay.
216 Long-Term Memory in Animals

Based on the evidence of memory replay during sharp-wave ripples in

the hippocampus in rats, one of the champions of the view that animals
do not have episodic memory (Suddendorf & Corballis, 2007) has made
a surprising reversal of position and has stated “it seems highly likely
from an evolutionary perspective that this activity is homologous to that
involved in mental time travel in humans” (Corballis, 2013, p. 5). This
underscores the strength of the memory replay evidence in support of the
view that mammals have the capacity for mental time travel and episodic
memory.
Do all animals have episodic memory? Given that hippocampal
sharp-wave ripples coordinate memory replay and have been observed
in all mammals that have been tested, it can be concluded that all
mammals have episodic memory. In contrast to mammals, there is
no evidence for sharp-wave ripples in the hippocampus of birds. It is
possible that birds use another mechanism to coordinate memory
replay, such as a different frequency of hippocampal activity.
Alternatively, it is possible that birds do not have memory replay and
do not have episodic memory. Future work will be needed to assess
whether there is memory replay evidence or other evidence for
episodic memory in birds.
The current chapter provides behavioral evidence and brain
evidence that satisfies both of Tulving’s criteria for episodic memory
in mammals. As discussed in Box 10.2, given that episodic memory is
one of the highest forms of cognitive processing, the conclusion that
mammals have episodic memory has implications for the way that
humans treat them.

Box 10.2: Implications of episodic memory in mammals

Episodic memory is one of the highest forms of cognitive processing in
humans. It involves mental time travel to a previous event and reflects
detailed conscious experience of that event. As reviewed in this chapter,
a growing body of evidence indicates that mammals, like humans, have
episodic memory. This supports a larger body of research indicating that
mammals can be highly intelligent and share many of the same cognitive
abilities as humans (de Waal, 2016). Although invasive experiments in
mammals have provided novel insight into the mechanisms of memory,
their advanced cognitive abilities must be weighed against the potential
benefits of animal research.
Review Questions 217

Chapter Summary
• The medial temporal lobe in rats, cats, and monkeys has the same
organization as humans, where the perirhinal cortex is associated with
item memory, the parahippocampal cortex is associated with context
memory, and the hippocampus is associated with binding item
information and context information.
• Long-term potentiation produces an increase in the amplitude and
rate of neural activity within the hippocampus that reflects the link
between this region and a cortical region.
• Memory replay in rats is associated with activity in the hippocampus,
the prefrontal cortex, the parietal cortex, and the visual sensory
cortex, which are the same regions associated with episodic memory
in humans.
• Hippocampal time cells in rats have been shown to be active
during non-task periods (such as running on a wheel) and require
a to-be-remembered task, and hippocampal time cells in monkeys
have been shown to be active during temporal order memory
tasks.
• Hippocampal time cells in rats fire as a function of theta activity.
• There is a growing body of behavioral evidence and brain evidence,
such as performance on temporal order tasks and memory replay
activity in the brain, that indicates mammals, like humans, have episo-
dic memory.

Review Questions
Which regions of the medial temporal lobe are conserved between
humans and other mammals?
How does long-term potentiation produce a link between a cortical
region and the hippocampus?
What are the brain regions that have been associated with memory replay
in rats and episodic memory in humans?
How are theta activity and hippocampal sharp-wave ripples related to
memory replay?
What are the two types of paradigms that have been used to uncover time
cells in the hippocampus?
Does the evidence that has accumulated today indicate that mammals
have episodic memory?
218 Long-Term Memory in Animals

Further Reading
Manns, J. R. & Eichenbaum, H. (2006). Evolution of declarative memory.
Hippocampus, 16, 795–808.
This paper shows that the organization of the medial temporal lobe in
rats, cats, and monkeys is the same as the mediate temporal lobe
organization in humans.
Girardeau, G. & Zugaro, M. (2011). Hippocampal ripples and memory
consolidation. Current Opinion in Neurobiology, 21, 452–459.
This paper reviews evidence for memory replay in rats, including memory
replay coordination by hippocampal sharp-wave ripples and the
importance of memory replay for memory consolidation.
Naya, Y. & Suzuki, W. A. (2011). Integrating what and when across the primate
medial temporal lobe. Science, 333, 773–776.
This study investigates temporal memory and item memory using
single-cell recording in different regions of the monkey medial temporal
lobe.
Kremers, D., Jaramillo, M. B., Böye, M., Lemasson, A. & Hausberger, M. (2011).
Do dolphins rehearse show-stimuli when at rest? Delayed matching of
auditory memory. Frontiers in Psychology, 2, 386.
This study shows evidence for speeded memory replay in vocalizations
made by dolphins during periods of quiet wakefulness or sleep, which are
the same characteristics of memory replay activity that has been
observed in the hippocampus of rats.
CHAPTER ELEVEN

The Future of Memory Research

Learning Objectives
• To understand the similarities between phrenology and fMRI.
• To list two advantages of ERPs over fMRI.
• To describe how brain region interaction studies are conducted.
• To characterize how the field of cognitive neuroscience will change in the
future.
• To specify whether research on temporal processing in the brain will
increase in the future.

Research on human memory is completely dependent on the methods

that are employed in the field of cognitive neuroscience, and thus
the future of memory research will follow the future of cognitive
neuroscience. This final chapter focuses on the cognitive neuroscience
techniques that have been employed in the past and the techniques that
will be employed in the future. Section 11.1 describes the similarities
between fMRI, which identifies brain regions associated with a cognitive
process, and phrenology, a pseudoscience from two centuries ago in
which each protrusion of the skull was associated with a particular beha-
vioral characteristic. In section 11.2, fMRI is directly compared to ERPs.
As fMRI has poor temporal resolution, only ERPs can measure the
temporal dynamics of the functioning brain. A cost–benefit analysis
favors ERPs, and government agencies are starting to increase funding
for research that employs ERPs. Section 11.3 discusses research investi-
gating brain region interactions, which will also receive increased govern-
ment funding. Brain region interaction research has only recently started
to be conducted and involves brain activity frequency analysis or mod-
ulating one brain region and measuring how that changes activity in
another brain region. Section 11.4 provides an overview of the field of
cognitive neuroscience in the future. A distinction is made between
human brain mapping, which refers to identifying the brain regions
associated with a cognitive process using fMRI, and research that
investigates brain region interactions using EEG frequency analysis
and combined techniques. It is predicted that human brain mapping
220 The Future of Memory Research

research will be assimilated by the field of cognitive psychology and that

the field of cognitive neuroscience will consist of human brain region
interaction research and will be an area within the field of behavioral
neuroscience. The final section, 11.5, shines a spotlight on the dimension
of time. To date, temporal processing in the brain has received less
attention than spatial localization. However, time is the future of the
cognitive neuroscience of memory.

11.1 Phrenology and fMRI

One of the primary aims of this book was to highlight findings from
studies that employed techniques with excellent temporal resolution.
Despite this aim, the large majority of the findings reviewed were based
on fMRI, as this is the most widely used technique in the field of cognitive
neuroscience. One major problem with fMRI is that it provides little
information with regard to the temporal dynamics of brain function
(see Chapter 2). A related problem is that many cognitive neuroscientists
believe that one brain region can be associated with one cognitive
process. This one-region-to-one-process mapping is reminiscent of the
pseudoscience called phrenology.
About two hundred years ago, Franz Joseph Gall and his collaborator
Johann Gaspar Spurzheim pioneered the system of phrenology. Based on
Gall’s observations of hundreds of human skulls, he proposed there were
twenty-seven skull protrusions on the scalp and that each protrusion was
associated with a particular behavior such as the propensity of combative-
ness, the sentiment of hope, or the sense of color. Figure 11.1A illustrates
a phrenology map from Spurzheim (1827) with each number representing
a different behavioral characteristic. Phrenology was based on the follow-
ing assumptions (van Wyhe, 2004, 16–17):
1 “Aptitudes and tendencies [that is, faculties] are inborn in humans
and animals.”
2 These have their “seat, their basis, in the brain.”
3 & 4 “Not only are the aptitudes and tendencies varied and independent,
but in addition they are essentially separate and independent of one
other, therefore they must have their seat in various and independent
parts of the brain.”
5 “From the various divisions of the various organs, and the varying
development of these, arises the varying shapes of the brain.”
6 “From the composition and development of particular organs arises the
particular shape of particular parts of the brain or regions of the same.”
11.1 Phrenology and fMRI 221

A B

Places Other people’s thoughts

Faces
Bodies

Figure 11.1 Past phrenology map and present brain map. (A) Spurzheim’s phrenology
map from 1827 (lateral view, occipital pole to the right). (B) Kanwisher’s brain map from
2010 (lateral view, occipital pole to the right). (A black and white version of this figure will
appear in some formats. For the color version, please refer to the plate section.)

7 “From the genesis of the bones of the skull from infancy to the
greatest age, the shape of the exterior surface of the skull is deter-
mined by the shape of the brain; therefore so far as the outer surface
of the skull and the inner coincide, and no exception is made for the
usual contours, particular aptitudes and tendencies can be con-
cluded.” This was to be determined by examining the shape and
contours of a head with the hands.
If cognitive processes are substituted for aptitudes and tendencies and
fMRI activity is substituted for the shape of the skull, fMRI can be
considered a kind of phrenology (Uttal, 2003). Phrenologists have assumed
a one-to-one mapping between a skull contour and a behavioral character-
istic, and many fMRI users have assumed a one-to-one mapping between
a brain activation and a cognitive process.
The fusiform face area (FFA) illustrates a brain region that is widely
believed to selectively process faces (see Chapter 1). The FFA, within the
right fusiform gyrus on the ventral surface of the brain, was first identified
in a visual perception fMRI study (Kanwisher, McDermott & Chun,
1997). In that study, many different types of stimuli were shown including
faces, objects, hands, and houses, and the FFA was found to be more
responsive to faces than non-face stimuli. These findings were taken to
indicate that the FFA selectively processes faces, and the FFA was even
222 The Future of Memory Research

referred to as a face processing “module” in the title of the paper. Since

that time, hundreds of articles have been published under the assumption
that the FFA is the face processing region. Like the FFA, other regions of
the brain have been identified that are thought to selectively process
certain types of stimuli or information including places (the parahippo-
campal place area, the PPA; see Chapter 1), bodies (in the extrastriate
body area, the EBA), and other people’s thoughts (in the right temporal
parietal junction, the rTPJ). Nancy Kanwisher (2010) has taken a strong
position and claimed that each of these regions may be “primarily, if not
exclusively, engaged in processing its preferred stimulus class” (p. 11164).
This statement indicates that these brain regions operate independently,
which is exactly what was assumed by phrenologists (see points 3 and 4
above). Figure 11.1B shows Kanwisher’s brain map from 2010, which is
similar to Spurzheim’s phrenology map from 1827 (compare
Figures 11.1A and 11.1B).
There are multiple lines of evidence showing that the FFA, the primary
example of specialized processing in the brain, is not specialized for
processing faces. Figure 11.1B shows face processing activity in the super-
ior temporal sulcus, which indicates that other regions of the brain
process faces in addition to the FFA. As shown in Figure 11.2A, one
fMRI study reported face versus object processing activity in eleven
different brain regions (Slotnick & White, 2013). This study tested the
hypothesis that the FFA is associated with shape processing rather than
face processing. As shown in Figure 11.2B, within the right hemisphere
FFA (the classic face processing region) and the left hemisphere FFA
(the left hemisphere homologue of the right hemisphere FFA), the
magnitude of activity associated with face perception did not differ
from the magnitude of activity associated with shape perception in the
contralateral visual field (e.g., the magnitude of activity in the right FFA
did not differ between faces and shapes in the left visual field). This
suggests that the FFA is actually associated with processing shape infor-
mation in the contralateral visual field rather than faces. Since faces are
typically composed of a greater number of internal shapes (e.g., the eyes
and the mouth) than objects, this might explain why the FFA has
appeared to be specialized for processing faces. Moreover, there is an
abundance of evidence that the FFA processes stimulus types other than
faces. In one fMRI study, participants viewed faces, houses, cats, sham-
poo bottles, scissors, shoes, and chairs (Haxby et al., 2001). Each stimulus
type was associated with a unique pattern of activations and deactivations
distributed across the ventral temporal cortex, rather than being
restricted to one or a few regions. A multi-voxel pattern analysis and
11.1 Phrenology and fMRI 223

Figure 11.2 Face processing and shape processing fMRI activity. (A) Face versus
object activity is shown in blue, with regions of interest labeled, and object versus face
activity is shown in purple (key at the top left; left, inferior view, occipital pole at the
bottom; top right, lateral view, occipital pole to the left; bottom right, lateral view,
occipital pole to the right; L = left, R = right, FFA = fusiform face area, ATFP = anterior
temporal face patch, Amy = amygdala, OFC = orbitofrontal cortex, OFA = occipital face
area, fSTS = face-selective region in the superior temporal sulcus, and IFS = inferior
frontal sulcus). (B) Magnitude of activity (percent signal change) associated with faces,
shapes in the left visual field (shape-LVF), shapes in the central visual field (shape-CVF),
and shapes in the right visual field (shape-RVF) in the right FFA (RFFA) and the left FFA
(LFFA). Brackets illustrate statistical comparisons between faces and the other event
types (asterisks indicate significant differences, ns = not significantly different). (A black
and white version of this figure will appear in some formats. For the color version, please
refer to the plate section.)
224 The Future of Memory Research

a pattern classification algorithm based on half of the trials were used to

identify which category the participants viewed on the other half of the
trials (see Chapter 6). That is, using half the data, the pattern of activity
associated with each stimulus type was identified. Then, for each trial in
the other half of the data, the pattern of activity associated with that item
was matched to each of the previously identified patterns and the best
match corresponded to the predicted category for that item. Even after
excluding the maximally responsive brain regions to faces, which
included the FFA, face trials could be identified on 100 percent of the
trials. This shows that face processing occurs outside the FFA. Moreover,
the pattern of activity in regions maximally responsive to faces, which
included the FFA, was used to identify all the other stimulus categories
with over 70 percent accuracy. This indicates that activity in the FFA
reflects processing of all the other stimulus types. In further support of
the last point, a meta-analysis of fMRI studies showed that the average
magnitude of activity in the FFA during face perception was 2.4 percent
signal change, while the average magnitude of activity in the FFA during
non-face perception (e.g., objects, cars, and bodies) was 1.0 percent signal
change (Slotnick, 2013a). The fact that the magnitude of non-face activity
in the FFA was greater than zero shows that this region is associated with
processing non-face stimuli. The preceding findings provide compelling
evidence that the FFA does not selective process faces and that face
processing is not restricted to this area. Thus, there is no basis to refer
to this region as the fusiform face area.
The previous example illustrates that fMRI results can be interpreted
in overly simplistic ways. Such simplicity is one of the reasons for the
attraction to fMRI by the news media (Beck, 2010). fMRI is also alluring
because it provides explanations of behavior based on the brain, such as
the brain basis for romantic love (Bartels & Zeki, 2000), and there is
increased confidence in biological results. fMRI is appealing to many
cognitive neuroscientists and cognitive psychologists for the same
reasons.
This section emphasized what can go wrong if only fMRI is employed
and it is assumed that one process is associated with one brain region or
a few brain regions. As illustrated in numerous scientific findings
reviewed in this book, each cognitive process is mediated by many
brain regions, and these brain regions are activated at different times
and interact with one another. If cognitive neuroscientists are to under-
stand the brain mechanisms underlying memory, we need to increase our
focus on brain timing. This can be achieved only by using methods with
excellent temporal resolution such as ERPs.
11.2 fMRI versus ERPs 225

11.2 fMRI versus ERPs

Figure 11.3 shows the number of fMRI papers and ERP papers published
from 1995 to 2015 in Nature Neuroscience, Neuron, and The Journal of
Neuroscience, the three journals with the highest-impact factors in the
field of cognitive neuroscience. The number of fMRI papers in these
journals is consistently over ten times the number of ERP papers. This
illustrates that fMRI is the gold standard in the field. Although fMRI has
excellent spatial resolution, it has poor temporal resolution (see
Chapter 2). This method takes a picture of all the brain regions that
were active in a 2-second time period. As brain activity changes on the
millisecond time scale, fMRI is approximately one thousand times too
slow to measure temporal processing in the brain.
Only ERPs can track the rapid temporal dynamics of the functioning
brain, but this technique is employed much less often in the field of
cognitive neuroscience. One reason is that it is not considered the gold
standard, so fewer scientists are drawn to this technique. A related reason
is that there are far fewer laboratories that employ ERPs, which results in
a lower number of scientists trained in the use of this technique. A final
reason is that ERP data acquisition and analysis is complex and there is

180
Number of articles in high-impact journals

160

140

120

100
fMRI
ERPs
80

60

40

20

0
1995 2000 2005 2010 2015
Year

Figure 11.3 Number of fMRI and ERP articles in the highest-impact cognitive neuroscience
journals (from 1995 to 2015; key to the right).
226 The Future of Memory Research

not a large research community to generate widely accessible analysis

procedures. The limiting factor in the widespread use of ERPs is the
relatively low number of laboratories that currently employ this
technique.
Fortunately, there are multiple factors that are expected to drive
the increased use of ERPs in the relatively near future. The first factor
that should increase the use of ERPs is the recent shift in the aims of
government agencies to fund brain research that employs techniques
with high temporal resolution. For instance, the description of
the recent National Institutes of Health BRAIN (Brain Research
through Advancing Innovative Neurotechnologies) Initiative states,
“By accelerating the development and application of innovative tech-
nologies, researchers will be able to produce a revolutionary new
dynamic picture of the brain that, for the first time, shows how
individual cells and complex neural circuits interact in both time
and space.” Along the same lines, the National Science Foundation
program guidelines for grant submissions in the cognitive neu-
roscience area states, “New frontiers in cognitive neuroscience
research have emerged from investigations that integrate data at
different spatial and temporal scales from a variety of techniques.”
Given that only ERPs can track brain activity in time, in the future
more funds will be allocated to research using this technique.
The second factor that should increase the use of ERPs is
that the cost of an ERP system is much less than an fMRI system.
A 128-channel ERP setup costs about $100,000 (US) and there are
almost no maintenance costs, while a 3 Tesla fMRI system costs about
$6 million (including the initial cost and maintenance costs for 10
years). An ERP system is also typically situated inside the scientist’s
laboratory, rather than being housed in a separate building or
at another university, as is almost always the case with fMRI.
The third factor that should increase the use of ERPs is an increase
in research using EEG frequency analysis within the last decade (see
Chapters 4 and 6). Recall that EEG uses the same acquisition
methodology as ERPs (see Chapter 2). EEG frequency analysis can
be used to test whether the activity in two brain regions is in phase,
which indicates that the regions interact. Brain region interaction
research will also receive additional funding in the future (see the
BRAIN Initiative statement above). As discussed in Box 11.1,
government agencies will increase funding for ERP research and
combined fMRI–ERP research.
11.3 Brain Region Interactions 227

Box 11.1: Funding will increase for ERP research

fMRI research is extremely expensive. Since MRI machines are so expensive
to purchase, install, and maintain, it typically costs at least $500 per hour
to rent MRI time. Assuming each session lasts 2 hours and each participant
is paid $100, an fMRI experiment with 20 participants would cost $22,000.
By comparison, the only costs associated with an ERP experiment are
participant payments, so the same experiment would cost $2,000.
In terms of funding needed per experiment, fMRI costs more than ten
times as much as ERPs. Said another way, over ten ERP experiments could
be funded for the same cost as one fMRI experiment. For two decades, the
government has primarily been investing money to pay for fMRI research,
which has supported the explosive growth of this technique in cognitive
neuroscience. However, government agencies are becoming more reluc-
tant to fund experiments that employ only fMRI. ERP studies can be funded
for a fraction of the cost, and combined fMRI–ERP experiments can be
conducted for about the same cost as fMRI experiments alone. As such,
funding for ERP research will increase in the future.

11.3 Brain Region Interactions

Many brain regions are activated during any given cognitive process.
As discussed in section 11.1 of this chapter, even the seemingly simple
process of face perception produces activity in at least eleven distinct
brain regions. More broadly, it is known that the visual cortical proces-
sing system is composed of over thirty brain regions, and there is massive
connectivity between regions with approximately 40 percent of all pairs
of regions directly connected (Felleman & Van Essen, 1991). This indi-
cates that even visual perceptual processing is mediated by numerous
brain regions, and even more brain regions are involved during memory.
For instance, long-term memory is associated with activity in sensory
processing regions and is also associated with activity in the dorsolateral
prefrontal cortex, the parietal cortex, and the medial temporal lobe (see
Chapter 3).
One future direction in the field of cognitive neuroscience will be to
determine which brain regions interact, the nature of these interactions,
and the timing of these interactions. One way to assess brain region
interactions is to conduct EEG frequency analysis, as illustrated by
many examples in this book (see Chapters 4 and 6). Brain region inter-
actions can also be evaluated by modulating activity in one brain region
228 The Future of Memory Research

and measuring how this changes activity in another brain region. TMS is
typically used to modulate activity in one brain region and then the
resultant changes in activity within another brain region are measured
using fMRI or ERPs. Such studies have only rarely been conducted
because they require knowledge of how to use multiple cognitive
neuroscience techniques and there are technical challenges to combining
multiple techniques. For instance, in addition to the physical constraints
of getting a TMS coil into the proper position inside an already tight-
fitting MRI scanner bore (see Chapter 2), the TMS coil will also distort
the MRI magnetic field and the results unless the coil is magnetically
shielded. Combining TMS and ERP is also challenging because both
systems have to be housed in a small space and each TMS pulse induces
a high amplitude current in the electrodes.
One study combined TMS and fMRI to assess how the right dorsolat-
eral prefrontal cortex interacts with early visual regions (Ruff et al.,
2006). Figure 11.4A shows the two targets of TMS stimulation for one
participant. These included the posterior middle frontal gyrus in a region
referred to as the frontal eye field (FEF), which has been associated with
eye movements and shifts in spatial attention, and the vertex, a common
control site for TMS stimulation. The vertex is at the top of the head at
the standard electrode location Cz, which is defined by the intersection
between the line connecting the preauricular point on each ear (i.e., the
most anterior point of the small curve in the outer ear that is just above
the tragus) and the line connecting the nasion (i.e., the depression at the
top of the nose) and the inion (i.e., the protrusion at the back of the skull).
During fMRI, five pulses of 9 Hertz TMS were applied to one of the

A B
Frontal vs. Vertex TMS
Mean TMS Effect

0 Center
Periphery

–2
Frontal TMS Mean V1 V2 V3 V4
Vertex TMS Region

Figure 11.4 Brain region interaction TMS target sites and fMRI visual sensory effects during
perception. (A) TMS was used to activate the posterior middle frontal gyrus (light gray
asterisk) or the vertex, the control site (dark gray asterisk; lateral view, occipital pole to the
left; key at the bottom right). (B) Change in fMRI activity (mean TMS effect) in central visual
field representations and peripheral visual field representations within visual regions V1, V2,
V3, and V4 (key to the right; asterisks indicate significant differences).
11.3 Brain Region Interactions 229

target sites, while the participants viewed a complex visual stimulus that
was moving and changing color or viewed a blank visual field.
Of importance, this TMS sequence activated the targeted cortical region
(as compared to 1 Hertz TMS that inhibits the targeted cortical region;
see Chapter 2). Of additional relevance, central visual field stimuli (i.e.,
corresponding to where someone is looking/fixating) are mapped more
posterior in the brain than peripheral visual field stimuli (i.e., the visual
field area surrounding where someone is looking/fixating). As shown in
Figure 11.4B, TMS to the right FEF, as compared to TMS to the vertex,
produced an increase in activity within early visual regions (V1–V4)
associated with peripheral visual field locations and a decrease in activity
within early visual regions associated with central visual field locations.
The same effects were observed regardless of whether participants
viewed the complex stimulus or the blank visual field, which indicates
the right FEF and early visual regions interact regardless of whether
there is visual stimulation. Of primary importance, these results suggest
that the right FEF activates early visual regions associated with the
peripheral visual field and deactivates early visual regions associated
with the central visual field.
Another study used TMS and fMRI to investigate the interaction
between the dorsolateral prefrontal cortex and visual sensory regions
during working memory (Feredoes, Heinen, Weiskopf, Ruff & Driver,
2011). During each study phase, three target faces or three target houses
were presented. During the delay period, TMS was applied to the right
dorsolateral prefrontal cortex, as depicted in Figure 11.5A. Three pulses
of TMS were applied at either a high intensity, which activated the
region, or a low intensity, which served as a baseline level of stimulation.
During each test phase, an old item or a new item was presented that
matched the category of the study phase (e.g., faces were presented
during the study phase and the test phase) and participants made an
“old”–“new” recognition judgment. On half of the trials, three distracter
stimuli from the opposite category were presented during the delay
period (e.g., faces during the study and test phases and houses during
the delay period). During the delay period, the magnitude of fMRI
activity was measured in the FFA and the PPA, which are illustrated in
Figure 11.5B. This study aimed to determine whether the dorsolateral
prefrontal cortex interacted with visual sensory regions to facilitate the
maintenance of targets from the study phase or to inhibit distracters
during the delay period. Figure 11.5C shows that high-intensity versus
low-intensity TMS to the right dorsolateral prefrontal cortex increased
activity in the FFA during the delay period when faces were targets and
 A B

C .15

.125

.1
% Signal Change

.075

.05

.025

–.025

–.05 face target face target house target house target

distracter absent house distracter distracter absent face distracter
FFA

.15

.125

.1
% Signal Change

.075

.05

.025

–.025

–.05 face target face target house target house target

distracter absent house distracter distracter absent face distracter

PPA

Figure 11.5 Brain region interaction TMS target site, visual sensory regions of interest, and
fMRI effects during working memory. (A) TMS was used to activate the right middle frontal
11.3 Brain Region Interactions 231

houses were distractors and increased activity in the PPA during the
delay period when houses were targets and faces were distractors.
These results indicate that the dorsolateral prefrontal cortex activates
visual sensory regions that maintain information from the study phase
during the delay period, but only in the presence of distracters.
The previous two studies not only identified the brain regions that were
interacting but also revealed the nature of these interactions. These
studies go far beyond simply identifying the brain regions that are asso-
ciated with a given cognitive process. Such brain interaction studies will
become more common as government funding for this type of research is
increased and more laboratories employ multiple techniques.
It should be noted that scientists who employ fMRI sometimes use
an analysis technique called structural equation modeling (SEM) or
dynamic causal modeling (DCM) with the aim of measuring how differ-
ent brain regions interact with one another. Briefly, these modeling/
mathematical techniques compare the fMRI activation timecourses in
a few brain regions of interest. If the activation timecourses between two
regions are correlated, the regions are assumed to be linked. If there is
a phase shift in the timecourse of fMRI activity between two regions, it is
assumed that the region with the activity that occurs earlier in time
modulates the activity in the other region. These would be reasonable
assumptions if fMRI had sufficient temporal resolution, but this is not the
case (see Chapter 2). There are a number of serious problems with SEM
and DCM including: (1) it is known that the fMRI activation timecourses
can differ between regions for physiological regions, such as the relatively
slow response in the prefrontal cortex, which will produce errors in the
directions of the interactions, (2) the model results change if additional
regions are included in the analysis, which means the results are not
stable and cannot be trusted, and (3) the models almost never adequately
fit the data, which is not acceptable from a statistical perspective.
Although employing SEM or DCM is admirable because it suggests
fMRI investigators want to study brain region interactions, the above
limitations are serious enough to cast doubt on the findings. Fortunately,

Caption for Figure 11.5 (cont.)

gyrus within the dorsolateral prefrontal cortex (lightning bolt; lateral view, occipital pole to
the left). (B) FFA (in black) and PPA (in dark gray; axial view, occipital pole to the left).
(C) Delay period activity (percent signal change) as a function of face or house target and
distracter conditions in the FFA (top) and the PPA (bottom; asterisks indicate significant
differences).
232 The Future of Memory Research

EEG frequency analysis and combined techniques can be used to reliably

measure brain regions interactions.

11.4 The Future of Cognitive Neuroscience

In the first chapter of this book, the field of cognitive neuroscience was
described as the intersection between the field of cognitive psychology
and the field of behavioral neuroscience. According to Michael
Gazzaniga, the first author of an excellent textbook on cognitive
neuroscience (Gazzaniga, Ivry & Mangun, 2014), he and the famous
cognitive psychologist George Miller came up with the name cognitive
neuroscience in the back seat of a New York City taxi in the late 1970s.
Michael Gazzaniga and some other world-renowned scientists
founded the Cognitive Neuroscience Society in 1994; therefore, the
field of cognitive neuroscience has been growing for about two
decades.
Based on its name, cognitive neuroscience seems to fit nicely at
the intersection between cognitive psychology and behavioral
neuroscience. However, as illustrated in Figure 11.6, top, this view of
cognitive neuroscience is a vision from the past. The fields of cognitive
psychology and cognitive neuroscience have changed and are continu-
ing to change. Cognitive psychology, the study of human mental
processing, has historically only employed behavioral measures such
as accuracy and reaction time (see Chapter 2). With the emergence of
cognitive neuroscience, cognitive psychologists have increasingly
begun to consider brain results. For example, any cognitive psychology
task can be employed during fMRI to identify the associated brain
regions. This is one way that many cognitive psychologists and cogni-
tive neuroscientists conduct experiments. However, identifying the
brain region(s) associated with a cognitive process (i.e., human brain
mapping) in no way measures brain region interactions and is reminis-
cent of phrenology (see section 11.1 of this chapter). Cognitive psy-
chologists currently have sessions at their conferences and sections in
their textbooks that focus on human brain mapping results.
This illustrates that the field of cognitive psychology is beginning to
assimilate human brain mapping results.
The field of cognitive neuroscience has been changing as well.
Cognitive neuroscientists are increasingly measuring brain region
interactions with EEG frequency analysis or by combining multiple
techniques such as TMS and fMRI or TMS and ERPs. These brain region
interaction studies go well beyond simple brain mapping and are
11.4 The Future of Cognitive Neuroscience 233

The Past

Cognitive Psychology Behavioral Neuroscience

Cognitive Neuroscience
Human mental processing Animal brain mechanisms

The Future

Cognitive Psychology Behavioral Neuroscience

Human mental processing and Animal brain mechanisms
human brain mapping

Cognitive Neuroscience
Human brain mechanisms
underlying mental procesing

Figure 11.6 The relationships between the fields of cognitive psychology, cognitive
neuroscience, and behavioral neuroscience in the past (at the top) and in the future (at
the bottom).

investigating the actual mechanisms of the human brain that underlie

a cognitive process. Such brain region interaction studies fit perfectly
within the field of behavioral neuroscience. Specifically, in systems
neuroscience, a subfield of behavioral neuroscience, scientists determine
which brain regions interact, when they interact, and how they interact
(e.g., it is assessed whether one brain region activates or deactivates
another brain region). Brain region interaction studies can be described
as systems neuroscience conducted with humans rather than non-human
animals.
234 The Future of Memory Research

The field of cognitive neuroscience is in transition. Currently, the large

majority of cognitive neuroscientists conduct human brain mapping stu-
dies, but an increasing number of cognitive neuroscientists will conduct
brain interaction studies. As illustrated in Figure 11.6, bottom, it is
predicted that, in the future, human brain mapping will be completely
assimilated by cognitive psychology (i.e., there will be no meaningful
distinction between cognitive psychology and human brain mapping)
and that human brain region interaction research will be a unique sub-
field of behavioral neuroscience that will be the new form of cognitive
neuroscience. It needs to be underscored that there is nothing wrong with
cognitive psychology. The best cognitive neuroscientists have almost all
trained in cognitive psychology and continue to conduct cognitive psy-
chology research. The point is that cognitive psychology (including
human brain mapping) is not neuroscience, as it does not investigate
the mechanisms of the functioning brain.

11.5 A Spotlight on the Fourth Dimension

Whenever one actually remembers having seen or heard, or learned, something, he includes
in this act . . . the consciousness of “formerly”; and the distinction of “former” and “latter”
is a distinction in time.
(Aristotle, [350 BCE] 1941, p. 608)

Albert Einstein wrote, the “system of values x, y, z, t . . . completely

defines the place and time of an event” (1905, p. 43). In the current
book, the dimension of time has been emphasized. This applies to the
time in which brain activity occurs and the frequency of brain activity
(i.e., oscillations in time; see Chapter 4). This also applies to memory for
the time at which a previous event occurred, such as mental time travel
during episodic memories (see Chapter 10).
As discussed earlier in this chapter, cognitive neuroscientists have
largely focused on identifying the location of brain activity associated
with a given cognitive process. Of course, it is important to identify
the brain regions involved in a cognitive process, but it is just as
important to identify the time these brain regions are active and
how these brain regions interact. The past and present focus on
brain localization has been largely driven by the popularity of fMRI
research.
In his paper on strong inference, Platt (1964) stresses that to make
rapid scientific progress in any field, multiple hypotheses must be gener-
ated and crucial experiments must be conducted to eliminate incorrect
11.5 A Spotlight on the Fourth Dimension 235

hypotheses. Strong inference is just another name for the scientific

method that was developed by Francis Bacon.
Bacon (1620) highlighted that the scientific method relies on exclusion
of hypotheses:
Only when the rejection and exclusion has been performed in proper fashion
will there remain (at the bottom of the flask, so to speak) an affirmative form,
solid, true and well-defined (the volatile opinions having now vanished into
smoke). (p. 127)
The cognitive neuroscience of memory has many important questions
that can be addressed only by applying the scientific method. For
instance, it is known that multiple regions of the brain are associated
with long-term memory including the dorsolateral prefrontal cortex, the
parietal cortex, the medial temporal lobe, and sensory processing regions
(see Chapters 1 and 3). How and when does the dorsolateral prefrontal
cortex modulate sensory processing regions? How and when does the
parietal cortex modulate sensory processing regions? Do the dorsolateral
prefrontal cortex and the parietal cortex interact? Do the dorsolateral
prefrontal cortex and the hippocampus interact? Regarding modulation,
one hypothesis is that the interaction is positive (i.e., produces activation)
and another hypothesis is that the interaction is negative (i.e., produces
deactivation). Regarding timing, one hypothesis is that the interaction
occurs early in time and another hypothesis is that the interaction occurs
late in time. A particular hypothesis can be ruled out by conducting the
crucial experiment using the proper techniques.
If we are to understand the mechanisms of human brain function,
cognitive neuroscientists must turn to more complex methods such as
EEG frequency analysis and combined techniques. Although this may
seem daunting, real progress requires investigators to learn the appro-
priate methods to answer important questions, rather than letting the
method they know dictate the questions they can answer.
This was stated eloquently by Platt (1964, p. 351):
Beware of the man of one method or one instrument, either experimental or
theoretical. He tends to become method-oriented rather than problem-
oriented. The method-oriented man is shackled; the problem-oriented man
is at least reaching freely toward what is most important. Strong inference
redirects a man to problem-orientation, but it requires him to be willing
repeatedly to put aside his last methods and teach himself new ones.
As discussed in Box 11.2, future cognitive neuroscientists should learn
a technique that can measure the time at which brain regions are active.
236 The Future of Memory Research

Box 11.2: Future cognitive neuroscientists should learn

methods that measure brain timing
The major strength of fMRI is its excellent spatial resolution, but this method
has poor temporal resolution. As discussed in this chapter, although ERPs
offer excellent temporal resolution, this method is currently underutilized in
the field of cognitive neuroscience. Moreover, EEG frequency analysis can be
used to investigate brain region interactions. The next generation of cogni-
tive neuroscientists should increasingly employ methods that can track the
temporal dynamics of brain activity if we are to understand the mechanisms
underlying human memory.

The dimension of time in the brain is the future of the cognitive

neuroscience of memory.

Chapter Summary
• Phrenology and fMRI are similar in that both assume a skull protru-
sion or brain activation is associated with one behavioral characteristic
or cognitive process.
• ERPs have much better temporal resolution than fMRI and ERP
research costs much less than fMRI.
• Government agencies will increase funding for research that employs
techniques with high temporal resolution and research that investi-
gates brain region interactions.
• Brain region interaction methods either measure synchronous activity
in different brain regions or modulate activity in one brain region and
then measure the resultant change in activity in another brain region.
• In the future, it is predicted that human brain mapping research will
be assimilated by the field of cognitive psychology, and cognitive neu-
roscience research will focus on investigations of brain region interactions
and will be a subfield of behavioral neuroscience.
• In the future, there will be an increase in research on temporal proces-
sing in the brain.

Review Questions
How are phrenology and fMRI similar?
What are two advantages of ERPs over fMRI?
Further Reading 237

Do brain region interaction studies always involve disrupting one brain

region and measuring activity in another brain region?
Will the field of cognitive neuroscience be completely assimilated by the
field of cognitive psychology in the future?
How will research on temporal processing in the brain change in the
future?

Further Reading
Kanwisher, N., McDermott, J. & Chun, M. M. (1997). The fusiform face area:
A module in human extrastriate cortex specialized for face perception.
The Journal of Neuroscience, 17, 4302–4311.
This fMRI paper introduced the fusiform face area, a region of the brain
that is still widely believed to be specialized for processing faces.
Slotnick, S. D. & White, R. C. (2013). The fusiform face area responds
equivalently to faces and abstract shapes in the left and central visual
fields. NeuroImage, 83, 408–417.
This fMRI paper shows that face perception produces activity in eleven
different brain regions and that the FFA is similarly associated with face
processing and shape processing.
Feredoes, E., Heinen, K., Weiskopf, N., Ruff, C. & Driver, J. (2011). Causal
evidence for frontal involvement in memory target maintenance by
posterior brain areas during distracter interference of visual working
memory. Proceedings of the National Academy of Sciences of the United
States of America, 108, 17510–17515.
This TMS-fMRI paper illustrates how multiple techniques can be used to
investigate brain region interactions.
Platt, J. R. (1964). Strong inference. Science, 146, 347–353.
This paper stresses the benefits of designing crucial experiments to rule
out hypotheses and employing the methods required to rapidly advance
scientific progress.
Glossary

action potential transient increase in voltage that travels down the axon of
a neuron.
AD see Alzheimer’s disease.
affective neuroscience field that focuses on the brain regions associated with
emotional processing, which is largely distinct from the field of cognitive
neuroscience.
alpha frequency band brain activity that oscillates between 8 and 12 Hertz.
Alzheimer’s disease a disease that, in the early stages, is primarily associated
with a selective impairment in long-term memory due to atrophy and protein
deposition in the medial temporal lobe and the parietal lobe.
aMCI see amnestic mild cognitive impairment.
amnesia impaired long-term memory.
amnestic mild cognitive impairment a disease that often progresses to
Alzheimer’s disease and is associated with a selective impairment in long-
term memory in older adults due to atrophy of the hippocampus and the
entorhinal cortex.
amyloid-β protein protein that accumulates in cortical regions of patients with
Alzheimer’s disease.
amyloid plaques agglomeration of Amyloid-β.
anterior temporal lobe epilepsy see medial temporal lobe epilepsy.
anterograde amnesia impaired long-term memory after the time of brain
damage.
associative priming task memory paradigm in which pairs of unrelated words
are presented during the study phase and then during the test phase
participants are shown intact or rearranged word pairs with the second word
as a stem and complete the word stem as quickly as possible with the first
word that comes to mind.
associative memory memory for an association between two items.
autobiographical memory a type of episodic memory for detailed personal
events.
axial view a slice of the brain that is approximately parallel to the nose and ears.
axon region of a neuron that transmits information to another neuron.
axon terminal the end of the axon that is far from the cell body.
BA Brodmann area, see Brodmann areas.
baseline event an event that is not associated with the cognitive process of interest.
behavioral experiments studies that measure only behavior, such as accuracy
and reaction time.
Glossary 239

behavioral neuroscience the study of the brain mechanisms underlying

behavior in animals.
bilateral relating to both hemispheres.
blocked design experimental protocols where each period has a relatively long
duration and consists of a series of the same events.
brain plasticity changes in the brain.
Brodmann areas distinct regions of the brain that were numbered by Korbinian
Brodmann in 1909.
central visual field part of the visual field where you are looking/fixating.
chunking associating multiple items with one another such that they can be
encoded as a single item.
chronic traumatic encephalopathy a disease caused by repeated mTBI and sub-
concussive head injuries that is associated with cognitive impairments in
attention and long-term memory due to atrophy that includes the frontal
lobes and the medial temporal lobes.
cognitive neuroscience the study of the brain mechanisms underlying human
mental processing.
cognitive psychology the study of human mental processes.
conceptual repetition priming change in the magnitude of activity within the
dorsolateral prefrontal frontal cortex that reflects decreased meaning-based
processing for repeated items. See also repetition priming.
contents of memory sensory memory effects.
context memory memory for the context of a previously presented item.
contextual cueing task memory paradigm in which participants quickly detect
a single target embedded in many similar distractors (i.e., the context), where
half of the contexts are repeated and the other half of the contexts are new.
contralateral P1 effect a marker of spatial attention in ERP studies that is
manifested by an increase in the magnitude of contralateral occipital/visual
activity that occurs 100 to 200 milliseconds after stimulus onset.
contralateral visual processing mapping of the left visual field and the right
visual field onto the right early visual areas and the left early visual areas,
respectively.
coronal view a slice of the brain that is approximately parallel to the face.
correct rejections new items correctly classified as “new.”
cross-frequency coupling activity that oscillates at different frequencies in two
brain regions that is in phase.
CTE see chronic traumatic encephalopathy.
Cz electrode location on the scalp that is defined by the intersection between
the line connecting the Preauricular point on each ear and the line
connecting the Nasion and the Inion.
declarative memory term used in animals that is equivalent to explicit memory
in humans.
240 Glossary

default network regions of the brain that become active when participants are
not engaged in the experimental task.
delayed matching-to-sample task paradigm to test old–new recognition
memory in animals where performance is based on selection of an old item
rather than a new item.
delayed nonmatching-to-sample task paradigm to test old–new recognition
memory in animals where performance is based on selection of a new item
rather than an old item.
dendrite region of a neuron that receives input from another neuron.
depth electrode recording method with excellent spatial resolution and
excellent temporal resolution in which electrodes measure neural activity.
diffusion-weighted imaging MRI technique that is sensitive to the diffusion of
water that tracks white matter pathways.
dipole adjacent positive and negative charge that is a model of cortical activity.
direct task paradigm in which participants are asked to recall old items, make
“old”–“new” recognition judgments, or make some other type of explicit
memory judgments.
dorsal toward the top of the brain.
dorsolateral prefrontal cortex frontal cortex that consists of the dorsal and
lateral surface that is anterior to the motor processing regions.
DRM paradigm experimental protocol where lists of associated words are
presented during the study phase and then during the test phase old words,
new related words, and new unrelated words are presented and participants
make “old”–“new” recognition judgments.
DWI see diffusion-weighted imaging.
dynamic causal modeling see structural equation modeling.
EBA see extrastriate body area.
EEG see electroencephalography.
electroencephalography method that uses the identical data acquisition
methodology as event-related potentials, but refers to any measure of brain
activity that corresponds to electric fields and typically refers to brain activity
that oscillates within a specific frequency range.
electrophysiological activity electrical activity generated by neuronal firing that
can be measured with event-related potentials.
episodic memory the detailed retrieval of a previous episode.
ERFs see event-related fields.
ERP component a single peak or valley in the event-related potential activation
timecourse.
ERPs see event-related potentials.
event-related design experimental protocols with a mixture of different events.
event-related fields method with excellent temporal resolution and limited
spatial resolution that measures magnetic fields using superconducting coils
Glossary 241

immediately above the scalp that directly reflect the underlying brain
activity.
event-related potentials method with excellent temporal resolution and limited
spatial resolution that measures voltages using electrodes on the scalp that
directly reflect the underlying brain activity.
excitatory post-synaptic potential an increase in voltage within the post-
synaptic neuron.
explicit memory conscious memory.
explicit memory contamination when participants use explicit memory during
an indirect task that has been assumed to rely on only implicit memory.
extrastriate body area region of the extrastriate cortex that is thought to be
specialized for processing bodies or body parts.
extrastriate cortex early visual processing regions that are anterior to V1.
facilitation model neural model of repetition priming where a repeated item is
associated with the same magnitude of activity for all neurons but the
activations occur at a faster rate.
false memory memory for information that did not occur.
familiarity all the types of non-detailed long-term memory.
fatigue model neural model of repetition priming where a repeated item is
associated with a similar proportional decrease in the magnitude of activity
for all of the neurons.
FEF see frontal eye field.
FFA see fusiform face area.
flashbulb memory seemingly picture-like memory for a very surprising and
consequential event.
fMRI see functional magnetic resonance imaging.
FN400 frontal ERP component that is negative in magnitude and peaks at
approximately 400 milliseconds after stimulus onset. See also, mid-frontal
old–new effect.
frontal eye field region of the frontal cortex at the intersection of the superior
frontal sulcus and the precentral sulcus that is associated with eye
movements and shifts in spatial attention.
functional magnetic resonance imaging method with excellent spatial
resolution and poor temporal resolution that measures increases in blood
flow that occur in active brain regions.
fusiform face area a visual region that preferentially processes faces.
gain model of attention view that attention amplifies the magnitude of brain
activity in sensory processing regions.
gamma frequency band brain activity that oscillates at a rate greater than 30
Hertz.
gist memory for the general theme of previous events.
glutamate the primary excitatory neurotransmitter.
242 Glossary

highly superior autobiographical memory rare memory ability of individuals

who, for any date of their life, can accurately and vividly remember the day
of the week, public events, and personal events.
hippocampal sharp-wave ripples hippocampal–cortical activity at a frequency
of approximately 200 Hertz that reflects replay of long-term memories.
human brain mapping identifying the brain regions associated with a cognitive
process using fMRI.
iEEG see intracranial EEG.
implicit memory nonconscious memory.
in phase activation timecourses in two brain regions that are very similar,
increasing and decreasing in magnitude with the same timing.
indirect task paradigm in which participants are asked about perceptual or
conceptual properties of items that is usually assumed to rely on only implicit
memory.
inferior view view of the brain from directly below.
inion the protrusion at the back of the skull.
intracarotid amobarbital test test to assess language and long-term memory
function in mTLE patients following injection of sodium amobarbital into
one of the internal carotid arteries to disrupt processing in the corresponding
hemisphere.
intracranial EEG EEG recording using depth electrodes implanted in the brain.
introspection the examination of your own mental processes.
inverse problem there are an infinite number of dipole sources that can give rise
to the same pattern of activity measured on the scalp using ERPs or MEG.
item memory memory for whether an item is “old” or “new.”
“knowing” subjective experience corresponding to the lack of detailed
retrieval during long-term memory.
lateral occipital complex a visual sensory region that preferentially processes
shape.
lateral view view of the brain from the side.
left-parietal old–new effect ERP component that occurs within 500 to 800
milliseconds, has a maximum amplitude over left parietal electrodes, and is
greater during recollection-based retrieval of old items than familiarity-
based retrieval of old items or correct rejection of new items.
left visual field the left half of space.
LOC see lateral occipital complex.
long-term depression a decrease in neuronal response magnitude following
activation.
long-term memory type of explicit memory in which information is not actively
maintained during the delay period.
long-term potentiation an increase in neuronal response magnitude following
activation.
Glossary 243

magnetoencephalography method that uses the identical data acquisition

methodology as event-related fields, but refers to any measure of brain
activity that corresponds to magnetic fields and typically refers to brain
activity that oscillates within a specific frequency range.
medial temporal lobe the upper medial aspect of the temporal lobe that
includes the perirhinal cortex, the parahippocampal cortex, and the
hippocampus.
medial temporal lobe epilepsy a disease associated with seizures caused by
abnormal brain functioning within the medial temporal lobe.
medial view viewing a hemisphere from the opposite direction as a lateral view.
MEG see magnetoencephalography.
memory consolidation changes in the brain regions underlying long-term
memory that takes years.
memory construction mental maintenance and elaboration on an episodic
memory for an extended period of time.
memory contents see contents of memory.
memory replay the reactivation of the same brain regions in the same or the
reverse temporal sequence that were activated during a previous event.
method of loci memory strategy where to-be-remembered items are associated
with a sequence of previously stored objects as one mentally travels through
a familiar setting.
mid-frontal old–new effect ERP component that occurs within 300 to 500
milliseconds, has a maximum amplitude over frontal electrodes, and is
greater during familiarity-based retrieval of old items than correct rejection
of new items. See also, FN400.
mild traumatic brain injury an injury caused by an impact to the head that
causes loss of consciousness for less than 30 minutes and post-traumatic
amnesia for less than 24 hours.
motivated forgetting an increase in the rate of forgetting for items that
a participant intentionally tries to forget.
MT a visual sensory region that preferentially processes motion.
mTBI see mild traumatic brain injury.
mTLE see medial temporal lobe epilepsy.
multi-voxel pattern analysis analysis of the pattern of fMRI activity across
many voxels.
n-back task a task in which items are sequentially presented and participants
are instructed to respond when the current item matches the item that was
presented n items previously.
nasion the depression at the top of the nose.
neurofibrillary tangles agglomeration of tau protein.
neurotransmitter a chemical substance that allows for communication between
neurons.
244 Glossary

nondeclarative memory term used in animals that is equivalent to implicit

memory in humans.
null finding a result that is not statistically significant.
occipital pole the most posterior part of the occipital lobe.
old-hits old items correctly classified as “old.”
old-misses old items incorrectly classified as “new”.
old–new recognition when old items and new items are presented and
participants make an “old” or “new” judgment for each item.
orientation grating stimulus with alternating parallel light and dark bars.
parahippocampal place area a visual region that preferentially processes visual
context such as places or scenes.
pattern classification algorithm computer program that learns the pattern of
brain activity associated with each trial type from a subset of trials and then
classifies the remaining trials based on how well the pattern of activity for
each trial matches the previously learned patterns.
pattern completion inaccurately responding “old” to new similar items.
pattern separation accurately responding “similar” to new similar items.
perforant path the white matter pathway between the entorhinal cortex and the
hippocampus.
peripheral visual field part of the visual field surrounding where you are
looking/fixating.
PET see positron emission tomography.
phase lag time (in milliseconds) or angle (from 0 to 360 degrees, i.e., 0 to 1 cycle)
of offset between activation timecourses in two brain regions.
phase-locked see in phase.
phrenology a pseudoscience from two centuries ago in which each protrusion of
the skull was associated with a particular behavioral characteristic.
PiB see Pittsburgh Compound B.
Pittsburgh Compound B radioactive substance that binds to amyloid-β protein
and can be measured using positron emission tomography to measure the
amount of this protein in different brain regions.
place cells neurons in the hippocampus that are active when an animal is in
a specific location.
positron emission tomography method with good spatial resolution and poor
temporal resolution that measures radioactive emissions associated with
increases in blood flow that occur in active brain regions.
PPA see parahippocampal place area.
preauricular point the most anterior point of the small curve in the outer ear
that is just above the tragus.
process-pure a hypothetical task that is associated with either implicit memory
or explicit memory.
recall retrieval of information based on an associated memory cue.
Glossary 245

receptor protein embedded in the cell wall that acts as a gateway for positive or
negative ions.
recollection all the types of detailed long-term memory.
REM rapid eye movement.
“remembering” subjective experience corresponding to detailed retrieval
during long-term memory.
repetition priming more efficient or fluent processing of an item when it is
repeated.
repetition suppression See repetition priming.
retinotopic map activations in early visual regions where adjacent locations in
the visual field are mapped onto adjacent locations on the cortex.
retrieval-induced forgetting when retrieval of one item has an inhibitory effect
on a related item, which increases the rate of forgetting for related items.
retrograde amnesia impaired long-term memory before the time of brain damage.
right-frontal old–new effect ERP component that occurs within 1000 to 1600
milliseconds, has a maximum amplitude over right frontal electrodes, and is
greater during recollection-based or familiarity-based retrieval of old items
than correct rejection of new items.
right temporal parietal junction brain region at the intersection of the temporal
lobe, the parietal lobe, and the occipital lobe that is widely believed to be
specialized for processing the thoughts of other people.
right visual field the right half of space.
rTPJ see right temporal parietal junction.
seizure focus the region of the brain from which seizures originate.
semantic memory retrieval of factual information that is learned over a long
period of time.
semantic processing extracting the meaning or conceptual representation of
a word or object.
sensory reactivation hypothesis hypothesis that memory for an event activates
the same brain regions associated with perception of that event.
sex differences differences between females and males.
sharpening model neural model of repetition priming where a repeated item is
associated with a decrease in the magnitude of activity for neurons that were
not maximally active and is associated with the same magnitude of activity
for neurons that were maximally active.
short-term memory see working memory.
single-cell recording method used in animals with excellent spatial resolution
and excellent temporal resolution in which an electrode measures activity
from a single neuron.
slow wave sleep non-REM sleep stages 3 and 4 associated with slow (less than
1 Hertz) waves of brain activity that can be measured across the entire
scalp using EEG.
246 Glossary

source memory see context memory.

spontaneous object recognition task paradigm to test old–new recognition
memory in animals where performance is based on preference for a new item
over an old item.
striate cortex see V1.
strong inference another name for the scientific method.
structural equation modeling analysis technique that compares the activation
timecourses in different brain regions in an effort to measure how they
interact with one another.
subsequent memory analysis sorting items in the study phase based on
responses during the test phase into subsequently remembered items and
subsequently forgotten items.
superior memory extraordinary memory ability of certain individuals in one
domain.
superior view view of the brain from directly above.
synaptic cleft the gap between the axon terminal of the pre-synaptic neuron and
the dendrite of the post-synaptic neuron.
Systems neuroscience subfield of behavioral neuroscience in which scientists
determine which brain regions interact, when they interact, and how they
interact.
tACS see transcranial alternating current stimulation.
tau protein protein that accumulates in the medial temporal lobe of patients
with Alzheimer’s disease.
tDCS see transcranial direct current stimulation.
TGA see transient global amnesia.
thalamic-cortical sleep spindles thalamic-cortical activity at frequencies of 11 to
16 Hertz that corresponds to the interaction between the thalamus and the
cortex during slow wave sleep.
theta frequency band brain activity that oscillates between 4 and 8 Hertz.
time cells neurons that are active at specific moments after the beginning of an
event.
TMS see transcranial magnetic stimulation.
top-down interaction when control regions modulate activity in sensory regions.
topographic map the magnitude of activity across the scalp.
transcranial alternating current stimulation method with poor spatial resolution
and poor temporal resolution in which a cortical region is temporarily
disrupted using a weak alternating current between two electrodes.
transcranial direct current stimulation method with poor spatial resolution and
poor temporal resolution in which a cortical region is temporarily disrupted
using a weak direct current between two electrodes.
transcranial magnetic stimulation method with limited spatial resolution and
poor temporal resolution in which a cortical region is temporarily
Glossary 247

deactivated or activated using a local magnetic field generated by

a stimulation coil.
transient global amnesia a temporary loss of memory that is usually triggered by
a highly emotional or physically arousing event.
unilateral relating to one hemisphere.
V1 the first visual sensory processing region.
V8 a visual sensory region that preferentially processes color.
ventral toward the bottom of the brain.
vertex point at the top of the head that is defined by the intersection between
the line connecting the preauricular points on the ears and the line
connecting the nasion and the inion.
what pathway visual regions from V1 to ventral extrastriate cortex to ventral
temporal cortex that process item identity.
where pathway visual regions from V1 to dorsal extrastriate cortex to parietal
cortex that process item spatial location.
working memory type of memory in which information is actively maintained
during the delay period.
References

Addis, D. R., Wong, A. T. & Schacter, D. L. (2007). Remembering the past

and imagining the future: Common and distinct neural substrates
during event construction and elaboration. Neuropsychologia, 45,
1363–1377.
Allen, R. J., Vargha-Khadem, F. & Baddeley, A. D. (2014). Item-location
binding in working memory: Is it hippocampus-dependent?
Neuropsychologia, 59, 74–84.
Alvarez, P. & Squire, L. R. (1994). Memory consolidation and the medial
temporal lobe: A simple network model. Proceedings of the National
Academy of Sciences of the United States of America, 91, 7041–7045.
Anderson, M. C. & Hanslmayr, S. (2014). Neural mechanisms of motivated
forgetting. Trends in Cognitive Sciences, 18, 279–292.
Anderson, M. C., Ochsner, K. N., Kuhl, B., Cooper, J., Robertson, E.,
Gabrieli, S. W., Glover, G. H. & Gabrieli, J. D. (2004). Neural systems
underlying the suppression of unwanted memories. Science, 303,
232–235.
Andreano, J. M. & Cahill, L. (2009). Sex influences on the neurobiology of
learning and memory. Learning & Memory, 16, 248–266.
Aristotle [350 BCE] (1941). On memory and reminiscence. In R. McKean
(Ed.), J. I. Beare (Trans.), The Basic Works of Aristotle (pp. 607–617).
New York: Random House.
Awh, E., Vogel, E. K. & Oh, S. H. (2006). Interactions between attention and
working memory. Neuroscience, 139, 201–208.
Axmacher, N., Draguhn, A., Elger, C. E. & Fell, J. (2009). Memory processes
during sleep: Beyond the standard consolidation theory. Cellular and
Molecular Life Sciences, 66, 2285–2297.
Babb, S. J. & Crystal, J. D. (2006). Episodic-like memory in the rat. Current
Biology, 16, 1317–1321.
Bacon, F. [1620] (2000). The New Organon. L. Jardine & M. Silverthorne
(Eds.). Cambridge University Press.
Baddeley, A., Allen, R. & Vargha-Khadem, F. (2010). Is the hippocampus
necessary for visual and verbal binding in working memory?
Neuropsychologia, 48, 1089–1095.
Badgaiyan, R. D., Schacter, D. L. & Alpert, N. M. (2003). Priming of new
associations: A PET study. Neuroreport, 14, 2475–2479.
Bakker, A., Albert, M. S., Krauss, G., Speck, C. L. & Gallagher, M. (2015).
Response of the medial temporal lobe network in amnestic mild
References 249

cognitive impairment to therapeutic intervention assessed by fMRI

and memory task performance. NeuroImage: Clinical, 7, 688–698.
Baldauf, D. & Desimone, R. (2014). Neural mechanisms of object-based
attention. Science, 344, 424–427.
Banks, S. J., Jones-Gotman, M., Ladowski, D. & Sziklas, V. (2012). Sex
differences in the medial temporal lobe during encoding and
recognition of pseudowords and abstract designs. NeuroImage, 59,
1888–1895.
Bartels, A. & Zeki, S. (2000). The neural basis of romantic love.
NeuroReport, 11, 3829–3834.
Bartsch, T., Döhring, J., Rohr, A., Jansen, O. & Deuschl, G. (2011). CA1
neurons in the human hippocampus are critical for autobiographical
memory, mental time travel, and autonoetic consciousness.
Proceedings of the National Academy of Sciences of the United States of
America, 108, 17562–17567.
Bartsch, T., Schönfeld, R., Müller, F. J., Alfke, K., Leplow, B., Aldenhoff, J.,
Deuschl, G. & Koch, J. M. (2010). Focal lesions of human hippocampal
CA1 neurons in transient global amnesia impair place memory.
Science, 328, 1412–1415.
Baudry, M., Zhu, G., Liu, Y., Wang, Y., Briz, V. & Bi, X. (2015). Multiple
cellular cascades participate in long-term potentiation and in
hippocampus-dependent learning. Brain Research, 1621, 73–81.
Bear, M. F. & Abraham, W. C. (1996). Long-term depression in
hippocampus. Annual Review of Neuroscience, 19, 437–462.
Beck, D. M. (2010). The appeal of the brain in the popular press. Perspectives
on Psychological Science, 5, 762–766.
Benoit, R. G. & Anderson, M. C. (2012). Opposing mechanisms support the
voluntary forgetting of unwanted memories. Neuron, 76, 450–460.
Bergmann, H. C., Rijpkema, M., Fernández, G. & Kessels, R. P. (2012). Distinct
neural correlates of associative working memory and long-term memory
encoding in the medial temporal lobe. NeuroImage, 63, 989–997.
Binder, J. R. (2011). Functional MRI is a valid noninvasive alternative to
Wada testing. Epilepsy & Behavior, 20, 214–222.
Blakemore, C. B. & Falconer, M. A. (1967). Long-term effects of anterior
temporal lobectomy on certain cognitive functions. Journal of
Neurology, Neurosurgery, & Psychiatry, 30, 364–367.
Bliss, T. V. & Collingridge, G. L. (1993). A synaptic model of memory:
Long-term potentiation in the hippocampus. Nature, 361, 31–39.
Bliss, T. V. & Lømo, T. (1973). Long-lasting potentiation of synaptic
transmission in the dentate area of the anaesthetized rabbit following
stimulation of the perforant path. The Journal of Physiology, 232,
331–356.
250 References

Bonnici, H. M., Chadwick, M. J., Lutti, A., Hassabis, D., Weiskopf, N. &
Maguire, E. A. (2012). Detecting representations of recent and remote
autobiographical memories in vmPFC and hippocampus. The Journal
of Neuroscience, 32, 16982–16991.
Born, J. & Wilhelm, I. (2012). System consolidation of memory during sleep.
Psychological Research, 76, 192–203.
Bridger, E. K., Bader, R., Kriukova, O., Unger, K. & Mecklinger, A. (2012).
The FN400 is functionally distinct from the N400. NeuroImage, 63,
1334–1342.
Brodmann, K. (1909). Vergleichende Lokalisationslehre der Grosshirnrinde
in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Leipzig:
Johann Ambrosius Barth Verlag.
Brown, R. & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99.
Buckner, R. L., Andrews-Hanna, J. R. & Schacter, D. L. (2008). The brain’s
default network: Anatomy, function, and relevance to disease. Annals
of the New York Academy of Sciences, 1124, 1–38.
Buckner, R. L., Bandettini, P. A., O’Craven, K. M., Savoy, R. L.,
Petersen, S. E., Raichle, M. E. & Rosen, B. R. (1996). Detection of
cortical activation during averaged single trials of a cognitive task
using functional magnetic resonance imaging. Proceedings of the
National Academy of Sciences of the United States of America, 93,
14878–14883.
Buckner, R. L., Snyder, A. Z., Shannon, B. J., LaRossa, G., Sachs, R.,
Fotenos, A. F., Sheline, Y. I., Klunk, W. E., Mathis, C. A., Morris, J. C.
& Mintun, M. A. (2005). Molecular, structural, and functional
characterization of Alzheimer’s disease: Evidence for a relationship
between default activity, amyloid, and memory. The Journal of
Neuroscience, 25, 7709–7717.
Buzsáki, G. (2015). Hippocampal sharp wave-ripple: A cognitive biomarker
for episodic memory and planning. Hippocampus, 25, 1073–1188.
Cabeza, R., Ciaramelli, E., Olson, I. R. & Moscovitch, M. (2008). The parietal
cortex and episodic memory: An attentional account. Nature Reviews
Neuroscience, 9, 613–625.
Cahill, L. (2006). Why sex matters for neuroscience. Nature Reviews
Neuroscience, 7, 477–484.
Chaieb, L., Leszczynski, M., Axmacher, N., Höhne, M., Elger, C. E. & Fell, J.
(2015). Theta-gamma phase-phase coupling during working memory
maintenance in the human hippocampus. Cognitive Neuroscience, 6,
149–157.
Chen, J. K., Johnston, K. M., Frey, S., Petrides, M., Worsley, K & Ptito, A.
(2004). Functional abnormalities in symptomatic concussed athletes:
An fMRI study. NeuroImage, 22, 68–82.
References 251

Chun, M. M. & Jiang, Y. (1998). Contextual cueing: Implicit learning and

memory of visual context guides spatial attention. Cognitive
Psychology, 36, 28–71.
Chun, M. M. & Phelps, E. A. (1999). Memory deficits for implicit contextual
information in amnesic subjects with hippocampal damage. Nature
Neuroscience, 2, 844–847.
Clayton, N. S. & Dickinson, A. (1998). Episodic-like memory during cache
recovery by scrub jays. Nature, 395, 272–274.
Cohen, D. & Cuffin, B. N. (1991). EEG versus MEG localization accuracy:
Theory and experiment. Brain Topography, 4, 95–103.
Corballis, M. C. (2013). Mental time travel: A case for evolutionary
continuity. Trends in Cognitive Science, 17, 5–6.
Corbetta, M. & Shulman, G. L. (2002). Control of goal-directed and
stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3,
201–215.
Cubelli, R. & Della Sala, S. (2008). Flashbulb memories: Special but not
iconic. Cortex, 44, 908–909.
Curran, T., Schacter, D. L., Johnson, M. K. & Spinks, R. (2001). Brain
potentials reflect behavioral differences in true and false recognition.
Journal of Cognitive Neuroscience, 13, 201–216.
Curtis, C. E. & D’Esposito, M. (2003). Persistent activity in the prefrontal
cortex during working memory. Trends in Cognitive Sciences, 7,
415–423.
Daselaar, S. M., Prince, S. E. & Cabeza, R. (2004). When less means more:
Deactivations during encoding that predict subsequent memory.
NeuroImage, 23, 921–927.
Davachi, L., Mitchell, J. P. & Wagner, A. D. (2003). Multiple routes to
memory: Distinct medial temporal lobe processes build item and
source memories. Proceedings of the National Academy of Sciences of
the United States of America, 100, 2157–2162.
Davidson, P. S., Cook, S. P., Glisky, E. L., Verfaellie, M. & Rapcsak, S. Z.
(2005). Source memory in the real world: A neuropsychological study
of flashbulb memory. Journal of Clinical and Experimental
Neuropsychology, 27, 915–929.
De Waal, F. (2016). Are We Smart Enough to Know How Smart Animals
Are? New York: W. W. Norton & Company.
Deese, J. (1959). On the prediction of occurrence of particular verbal
intrusions in immediate recall. Journal of Experimental Psychology, 58,
17–22.
Depue, B. E. (2012). A neuroanatomical model of prefrontal inhibitory
modulation of memory retrieval. Neuroscience & Biobehavioral
Reviews, 36, 1382–1399.
252 References

Dettwiler, A., Murugavel, M., Putukian, M., Cubon, V., Furtado, J. &
Osherson, D. (2014). Persistent differences in patterns of brain
activation after sports-related concussion: A longitudinal functional
magnetic resonance imaging study. Journal of Neurotrauma, 31,
180–188.
Diana, R. A., Yonelinas, A. P. & Ranganath, C. (2007). Imaging recollection
and familiarity in the medial temporal lobe: A three-component model.
Trends in Cognitive Sciences, 11, 379–386.
Dickerson, B. C. & Sperling, R. A. (2008). Functional abnormalities of the
medial temporal lobe memory system in mild cognitive impairment
and Alzheimer’s disease: Insights from functional MRI studies.
Neuropsychologia, 46, 1624–1635.
Diedrichsen, J. & Kornysheva, K. (2015). Motor skill learning between
selection and execution. Trends in Cognitive Sciences, 19, 227–233.
Döhring, J., Schmuck, A. & Bartsch, T. (2014). Stress-related factors in the
emergence of transient global amnesia with hippocampal lesions.
Frontiers in Behavioral Neuroscience, 8, 287.
Domoto-Reilly, K., Sapolsky, D., Brickhouse, M. & Dickerson, B. C. (2012).
Naming impairment in Alzheimer’s disease is associated with left
anterior temporal lobe atrophy. NeuroImage, 63, 348–355.
Donders, F. C. (1868). Over de snelheid van psychische processen.
Onderzoekingen gedaan in het Physiologisch Laboratorium der
Utrechtsche Hoogeschool, 1868–1869, Tweede reeks, II, 92–120.
Duncan, K., Tompary, A. & Davachi, L. (2014). Associative encoding and
retrieval are predicted by functional connectivity in distinct hippocampal
area CA1 pathways. The Journal of Neuroscience, 34, 11188–11198.
Eichenbaum, H. (2014). Time cells in the hippocampus: A new dimension for
mapping memories. Nature Reviews Neuroscience, 15, 732–744.
Eichenbaum, H., Sauvage, M., Fortin, N., Komorowski, R. & Lipton, P.
(2012). Towards a functional organization of episodic memory in the
medial temporal lobe. Neuroscience & Biobehavioral Reviews, 36,
1597–1608.
Eichenbaum, H., Yonelinas, A. P. & Ranganath, C. (2007). The medial
temporal lobe and recognition memory. Annual Review of
Neuroscience, 30, 123–152.
Einstein, A. [1905] (1952). On the electrodynamics of moving bodies.
In W. Perrett & G. B. Jeffery (Trans.), The Principle of Relativity:
A Collection of Original Memoirs on the Special and General Theory of
Relativity (pp. 37–65). New York: Dover Publications.
Ekstrom, A. D., Copara, M. S., Isham, E. A., Wang, W. C. & Yonelinas, A. P.
(2011). Dissociable networks involved in spatial and temporal order
source retrieval. NeuroImage, 56, 1803–1813.
References 253

Ekstrom, A. D., Kahana, M. J., Caplan, J. B., Fields, T. A., Isham, E. A.,
Newman, E. L. & Fried, I. (2003). Cellular networks underlying human
spatial navigation. Nature, 425, 184–188.
Eldridge, L. L., Knowlton, B. J., Furmanski, C. S., Bookheimer, S. Y. &
Engel, S. A. (2000). Remembering episodes: A selective role for the
hippocampus during retrieval. Nature Neuroscience, 3, 1149–1152.
Ellenbogen, J. M., Payne, J. D. & Stickgold, R. (2006). The role of sleep in
declarative memory consolidation: Passive, permissive, active or none?
Current Opinion in Neurobiology, 16, 716–722.
Engel, A. K., Fries, P. & Singer, W. (2001). Dynamic predictions: Oscillations
and synchrony in top-down processing. Nature Reviews Neuroscience,
2, 704–716.
Engell, A. D. & McCarthy, G. (2014). Repetition suppression of
face-selective evoked and induced EEG recorded from human cortex.
Human Brain Mapping, 35, 4155–4162.
Felleman, D. J. & Van Essen, D. C. (1991). Distributed hierarchical
processing in the primate cerebral cortex. Cerebral Cortex, 1, 1–47.
Feredoes, E., Heinen, K., Weiskopf, N., Ruff, C. & Driver, J. (2011). Causal
evidence for frontal involvement in memory target maintenance by
posterior brain areas during distracter interference of visual working
memory. Proceedings of the National Academy of Sciences of the
United States of America, 108, 17510–17515.
Fiebach, C. J., Gruber, T. & Supp, G. G. (2005). Neuronal mechanisms of
repetition priming in occipitotemporal cortex: Spatiotemporal
evidence from functional magnetic resonance imaging and
electroencephalography. The Journal of Neuroscience, 25, 3414–3422.
Finke, C., Braun, M., Ostendorf, F., Lehmann, T. N., Hoffmann, K. T.,
Kopp, U. & Ploner, C. J. (2008). The human hippocampal formation
mediates short-term memory of colour-location associations.
Neuropsychologia, 46, 614–623.
Floyer-Lea, A. & Matthews, P. M. (2005). Distinguishable brain activation
networks for short- and long-term motor skill learning. Journal of
Neurophysiology, 94, 512–518.
Friederici, A. D. & Gierhan, S. M. (2013). The language network. Current
Opinion in Neurobiology, 23, 250–254.
Friese, U., Köster, M., Hassler, U., Martens, U., Trujillo-Barreto, N. &
Gruber, T. (2013). Successful memory encoding is associated with
increased cross-frequency coupling between frontal theta and
posterior gamma oscillations in human scalp-recorded EEG.
NeuroImage, 66, 642–647.
Frings, L., Wagner, K., Unterrainer, J., Spreer, J., Halsband, U. &
Schulze-Bonhage, A. (2006). Gender-related differences in
254 References

lateralization of hippocampal activation and cognitive strategy.

Frontiers in Human Neuroscience, 7, 279.
Gabrieli, J. D., Poldrack, R. A. & Desmond, J. E. (1998). The role of left
prefrontal cortex in language and memory. Proceedings of the National
Academy of Sciences of the United States of America, 95, 906–913.
Gagnepain, P., Henson, R. N. & Anderson, M. C. (2014). Suppressing
unwanted memories reduces their unconscious influence via targeted
cortical inhibition. Proceedings of the National Academy of Sciences of
the United States of America, 111, E1310–1319.
Gallagher, M. & Koh, M. T. (2011). Episodic memory on the path to
Alzheimer’s disease. Current Opinion in Neurobiology, 21, 929–934.
Garoff-Eaton, R. J., Slotnick, S. D. & Schacter, D. L. (2006). Not all false
memories are created equal: The neural basis of false recognition.
Cerebral Cortex, 16, 1645–1652.
Gazzaley, A., & Nobre, A. C. (2012). Top-down modulation: Bridging selective
attention and working memory. Trends in Cognitive Sciences, 16, 129–135.
Gazzaniga, M. S., Ivry, R. B. & Mangun, G. R. (2014). Cognitive
Neuroscience: The Biology of the Mind (4th edn.). New York:
W. W. Norton & Company.
Ghuman, A. S., Bar, M., Dobbins, I. G. & Schnyer, D. M. (2008). The effects
of priming on frontal-temporal communication. Proceedings of the
National Academy of Sciences of the United States of America, 105,
8405–8409.
Gilbert, J. R., Gotts, S. J., Carver, F. W. & Martin, A. (2010). Object
repetition leads to local increases in the temporal coordination of
neural responses. Frontiers in Human Neuroscience, 4, 30.
Girardeau, G. & Zugaro, M. (2011). Hippocampal ripples and memory
consolidation. Current Opinion in Neurobiology, 21, 452–459.
Glosser, G., Deutsch, G. K., Cole, L. C., Corwin, J. & Saykin, A. J. (1998).
Differential lateralization of memory discrimination and response bias
in temporal lobe epilepsy patients. Journal of the International
Neuropsychological Society, 4, 502–511.
Goldstein, J. M., Seidman, L. J., Horton, N. J., Makris, N., Kennedy, D. N.,
Caviness, V. S. Jr., Faraone, S. V. & Tsuang, M. T. (2001). Normal
sexual dimorphism of the adult human brain assessed by in vivo
magnetic resonance imaging. Cerebral Cortex, 11, 490–497.
Gosselin, N., Bottari, C., Chen, J. K., Petrides, M., Tinawi, S., de Guise, E. &
Ptito, A. (2011). Electrophysiology and functional MRI in post-acute
mild traumatic brain injury. Journal of Neurotrauma, 28, 329–341.
Gotts, S. J., Chow, C. C. & Martin, A. (2012). Repetition priming and
repetition suppression: A case for enhanced efficiency through neural
synchronization. Cognitive Neuroscience, 3, 227–237.
References 255

Grill-Spector, K., Henson, R. & Martin, A. (2006). Repetition and the brain:
Neural models of stimulus-specific effects. Trends in Cognitive
Sciences, 10, 14–23.
Grubb, N. R., Fox, K. A., Smith, K., Best, J., Blane, A., Ebmeier, K. P.,
Glabus, M. F. & O’Carroll, R. E. (2000). Memory impairment in
out-of-hospital cardiac arrest survivors is associated with global
reduction in brain volume, not focal hippocampal injury. Stroke, 31,
1509–1514.
Gruber T. & Müller, M. M. (2005). Oscillatory brain activity dissociates
between associative stimulus content in a repetition priming task in the
human EEG. Cerebral Cortex, 15, 109–116.
Hannula, D. E. & Ranganath, C. (2008). Medial temporal lobe activity
predicts successful relational memory binding. The Journal of
Neuroscience, 28, 116–124.
Hannula, D. E. & Ranganath, C. (2009). The eyes have it: Hippocampal activity
predicts expression of memory in eye movements. Neuron, 63, 592–599.
Harrison, S. A. & Tong, F. (2009). Decoding reveals the contents of visual
working memory in early visual areas. Nature, 458, 632–635.
Hart, B. L., Hart, L. A. & Pinter-Wollman, N. (2008). Large brains and
cognition: Where do elephants fit in? Neuroscience & Biobehavioral
Reviews, 32, 86–98.
Haxby, J. V., Gobbini, M. I., Furey, M. L., Ishai, A., Schouten, J. L. &
Pietrini, P. (2001). Distributed and overlapping representations of
faces and objects in ventral temporal cortex. Science, 293, 2425–2430.
Helmholtz, H. von (1850). Vorläufiger Bericht über die
Fortpflanzungsgeschwindigkeit der Nervenreizung. Archiv für
Anatomie, Physiologie und wissenschaftliche Medicin, 71–73.
Henson, R., Shallice, T. & Dolan, R. (2000). Neuroimaging evidence for
dissociable forms of repetition priming. Science, 287, 1269–1272.
Herrmann, C. S., Rach, S., Neuling, T. & Strüber, D. (2013). Transcranial
alternating current stimulation: A review of the underlying
mechanisms and modulation of cognitive processes. Frontiers in
Human Neuroscience, 7, 279.
Hirst, W., Phelps, E. A., Buckner, R. L., Budson, A. E., Cuc, A.,
Gabrieli, J. D., Johnson, M. K., Lustig, C., Lyle, K. B., Mather, M.,
Meksin, R., Mitchell, K. J., Ochsner, K. N., Schacter, D. L., Simons, J. S.
& Vaidya, C. J. (2009). Long-term memory for the terrorist attack
of September 11: Flashbulb memories, event memories, and the factors
that influence their retention. Journal of Experimental Psychology:
General, 138, 161–176.
Hirst, W., Phelps, E. A., Meksin, R., Vaidya, C. J., Johnson, M. K.,
Mitchell, K. J., Buckner, R. L., Budson, A. E., Gabrieli, J. D., Lustig, C.,
256 References

Mather, M., Ochsner, K. N., Schacter, D., Simons, J. S., Lyle, K. B,

Cuc, A. F. & Olsson, A. (2015). A ten-year follow-up of a study of
memory for the attack of September 11, 2001: Flashbulb memories and
memories for flashbulb events. Journal of Experimental Psychology:
General, 144, 603–623.
Hodges, J. R. & Warlow, C. P. (1990). Syndromes of transient amnesia:
Towards a classification. A study of 153 cases. Journal of Neurology,
Neurosurgery, and Psychiatry, 53, 834–843.
Hopfinger, J. B., Woldorff, M. G., Fletcher, E. M. & Mangun, G. R. (2001).
Dissociating top-down attentional control from selective perception
and action. Neuropsychologia, 39, 1277–1291.
Hou, M., Safron, A., Paller, K. A. & Guo, C. (2013). Neural correlates of
familiarity and conceptual fluency in a recognition test with ancient
pictographic characters. Brain Research, 1518, 48–60.
Hsieh, L. T., Ekstrom, A. D. & Ranganath, C. (2011). Neural oscillations
associated with item and temporal order maintenance in working
memory. The Journal of Neuroscience, 31, 10803–10810.
Huettel, S. A., Song, A. W. & McCarthy, G. (2014). Functional Magnetic
Resonance Imaging (3rd edn.). Sunderland, MA: Sinauer Associates.
Hutchinson, J. B., Uncapher, M. R., Weiner, K. S., Bressler, D. W.,
Silver, M. A., Preston, A. R. & Wagner, A. D. (2014). Functional
heterogeneity in posterior parietal cortex across attention and episodic
memory retrieval. Cerebral Cortex, 24, 49–66.
Ikkai, A. & Curtis, C. E. (2011). Common neural mechanisms supporting
spatial working memory, attention and motor intention.
Neuropsychologia, 49, 1428–1434.
Jaeggi, S. M., Buschkuehl, M., Jonides, J. & Perrig, W. J. (2008). Improving
fluid intelligence with training on working memory. Proceedings of the
National Academy of Sciences of the United States of America, 105,
6829–6833.
James, W. (1890). The principles of psychology (Vol. I). New York: Dover
Publications.
Ji, D. & Wilson, M. A. (2007). Coordinated memory replay in the visual
cortex and hippocampus during sleep. Nature Neuroscience, 10,
100–107.
Johnson, J. D., Minton, B. R. & Rugg, M. D. (2008). Content dependence of
the electrophysiological correlates of recollection. NeuroImage, 39,
406–416.
Jolles, D. D., Grol, M. J., Van Buchem, M. A., Rombouts, S. A. &
Crone, E. A. (2010). Practice effects in the brain: Changes in cerebral
activation after working memory practice depend on task demands.
NeuroImage, 52, 658–668.
References 257

Jones-Gotman, M. (1986). Right hippocampal excision impairs learning and

recall of a list of abstract designs. Neuropsychologia, 24, 659–670.
Kanwisher, N. (2010). Functional specificity in the human brain: A window
into the functional architecture of the mind. Proceedings of the
National Academy of Sciences of the United States of America, 107,
11163–11170.
Kanwisher, N., McDermott, J. & Chun, M. M. (1997). The fusiform face area:
A module in human extrastriate cortex specialized for face perception.
The Journal of Neuroscience, 17, 4302–4311.
Karanian, J. M. & Slotnick, S. D. (2014a). The cortical basis of true memory
and false memory for motion. Neuropsychologia, 54, 53–58.
(2014b). False memory for context activates the parahippocampal cortex.
Cognitive Neuroscience, 5, 186–92.
(2015). Memory for shape reactivates the lateral occipital complex. Brain
Research, 1603, 124–132.
(2016). False memories activate early sensory cortex. Manuscript
submitted for publication.
Kemp, A. & Manahan-Vaughan, D. (2007). Hippocampal long-term
depression: Master or minion in declarative memory processes? Trends
in Neurosciences, 30, 111–118.
Kiesel, A., Kunde, W., Pohl, C., Berner, M. P. & Hoffmann, J. (2009). Playing
chess unconsciously. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 35, 292–298.
Kim, H. (2011). Neural activity that predicts subsequent memory and
forgetting: A meta-analysis of 74 fMRI studies. NeuroImage, 54,
2446–2461.
Kim, H. & Cabeza, R. (2007). Differential contributions of prefrontal, medial
temporal, and sensory-perceptual regions to true and false memory
formation. Cerebral Cortex, 17, 2143–2150.
Klingberg, T. (2010). Training and plasticity of working memory. Trends in
Cognitive Sciences, 14, 317–324.
Knowlton, B. J. & Fanselow, M. S. (1998). The hippocampus, consolidation
and on-line memory. Current Opinion in Neurobiology, 8, 293–296.
Kosslyn, S. M., Ganis, G. & Thompson, W. L. (2001). Neural foundations of
imagery. Nature Reviews Neuroscience, 2, 635–642.
Köster, M., Friese, U., Schöne, B., Trujillo-Barreto, N. & Gruber, T. (2014).
Theta-gamma coupling during episodic retrieval in the human EEG.
Brain Research, 1577, 57–68.
Koutstaal, W., Wagner, A. D., Rotte, M., Maril, A., Buckner, R. L. &
Schacter, D. L. (2001). Perceptual specificity in visual object priming:
Functional magnetic resonance imaging evidence for a laterality
difference in fusiform cortex. Neuropsychologia, 39, 184–199.
258 References

Kraus, B. J., Robinson, R. J., White, J. A., Eichenbaum, H. &

Hasselmo, M. E. (2013). Hippocampal “time cells”: Time versus path
integration. Neuron, 78, 1090–1101.
Kremers, D., Jaramillo, M. B., Böye, M., Lemasson, A. & Hausberger, M.
(2011). Do dolphins rehearse show-stimuli when at rest? Delayed
matching of auditory memory. Frontiers in Psychology, 2, 386.
Landau, S. M., Marks, S. M., Mormino, E. C., Rabinovici, G. D., Oh, H.,
O’Neil, J. P., Wilson, R. S. & Jagust, W. J. (2012). Association of
lifetime cognitive engagement and low β-amyloid deposition. Archives
of Neurology, 69, 623–629.
Larson, J., Wong, D. & Lynch, G. (1986). Patterned stimulation at the theta
frequency is optimal for the induction of hippocampal long-term
potentiation. Brain Research, 368, 347–350.
Leal, S. L. & Yassa, M. A. (2013). Perturbations of neural circuitry in aging,
mild cognitive impairment, and Alzheimer’s disease. Ageing Research
Reviews, 12, 823–831.
LePort, A. K., Mattfeld, A. T., Dickinson-Anson, H., Fallon, J. H.,
Stark, C. E., Kruggel, F., Cahill, L. & McGaugh, J. L. (2012).
Behavioral and neuroanatomical investigation of Highly Superior
Autobiographical Memory (HSAM). Neurobiology of Learning and
Memory, 98, 78–92.
Li, X., Xiao, Y. H., Zhao, Q., Leung, A. W., Cheung, E. F. & Chan, R. C.
(2015). The neuroplastic effect of working memory training in healthy
volunteers and patients with schizophrenia: Implications for cognitive
rehabilitation. Neuropsychologia, 75, 149–162.
Liang, K. Y., Mintun, M. A., Fagan, A. M., Goate, A. M., Bugg, J. M.,
Holtzman, D. M., Morris, J. C. & Head, D. (2010). Exercise and
Alzheimer’s disease biomarkers in cognitively normal older adults.
Annals of Neurology, 68, 311–318.
Libby, L. A., Hannula, D. E. & Ranganath, C. (2014). Medial temporal lobe
coding of item and spatial information during relational binding in
working memory. The Journal of Neuroscience, 34, 14233–14242.
Lindquist, K. A., Wager, T. D., Kober, H., Bliss-Moreau, E. & Barrett, L. F.
(2012). The brain basis of emotion: A meta-analytic review. Behavioral
and Brain Sciences, 35, 121–143.
Liu, T., Slotnick, S. D., Serences, J. T. & Yantis, S. (2003). Cortical mechanisms
of feature-based attentional control. Cerebral Cortex, 13, 1334–1343.
Logothetis, N. K., Pauls, J., Augath, M., Trinath, T. & Oeltermann, A. (2001).
Neurophysiological investigation of the basis of the fMRI signal.
Nature, 412, 150–157.
Ma, L., Wang, B., Narayana, S., Hazeltine, E., Chen, X., Robin, D. A.,
Fox, P. T. & Xiong, J. (2010). Changes in regional activity are
References 259

accompanied with changes in inter-regional connectivity during 4

weeks motor learning. Brain Research, 1318, 64–76.
MacDonald, C. J., Carrow, S., Place, R. & Eichenbaum, H. (2013). Distinct
hippocampal time cell sequences represent odor memories in
immobilized rats. The Journal of Neuroscience, 33, 14607–14616.
Maguire, E. A., Gadian, D. G., Johnsrude, I. S., Good, C. D., Ashburner, J.,
Frackowiak, R. S. & Frith, C. D. (2000). Navigation-related structural
change in the hippocampi of taxi drivers. Proceedings of the National
Academy of Sciences of the United States of America, 97, 4398–4403.
Maguire, E. A., Valentine, E. R., Wilding, J. M. & Kapur, N. (2003). Routes
to remembering: The brains behind superior memory. Nature
Neuroscience, 6, 90–95.
Maguire, E. A., Woollett, K. & Spiers, H. J. (2006). London taxi drivers and
bus drivers: A structural MRI and neuropsychological analysis.
Hippocampus, 16, 1091–1101.
Malmivuo, J. (2012). Comparison of the properties of EEG and MEG in
detecting the electric activity of the brain. Brain Topography, 25, 1–19.
Manns, J. R. & Eichenbaum, H. (2006). Evolution of declarative memory.
Hippocampus, 16, 795–808.
Manns, J. R., Hopkins, R. O., Reed, J. M., Kitchener, E. G. & Squire, L. R.
(2003). Recognition memory and the human hippocampus. Neuron, 37,
171–180.
Marshall, L. & Born, J. (2007). The contribution of sleep to
hippocampus-dependent memory consolidation. Trends in Cognitive
Sciences, 11, 442–450.
Marshall, L., Helgadóttir, H., Mölle, M. & Born, J. (2006). Boosting slow
oscillations during sleep potentiates memory. Nature, 444, 610–613.
Martin, A. & Chao, L. L. (2001). Semantic memory and the brain: Structure
and processes. Current Opinion in Neurobiology, 11, 194–201.
Matsumura, N., Nishijo, H., Tamura, R., Eifuku, S., Endo, S. & Ono, T.
(1999). Spatial- and task-dependent neuronal responses during real
and virtual translocation in the monkey hippocampal formation.
The Journal of Neuroscience, 19, 2381–2393.
Maxwell, J. C. (1865). A dynamical theory of the electromagnetic field.
Philosophical Transactions of the Royal Society of London, 155, 459–512.
Mayer, A. R., Bellgowan, P. S. & Hanlon, F. M. (2015). Functional magnetic
resonance imaging of mild traumatic brain injury. Neuroscience &
Biobehavioral Reviews, 49, 8–18.
McAllister, T. W., Sparling, M. B., Flashman, L. A., Guerin, S. J.,
Mamourian, A. C. & Saykin, A. J. (2001). Differential working
memory load effects after mild traumatic brain injury. NeuroImage, 14,
1004–1012.
260 References

McDonald, B. C., Saykin, A. J. & McAllister, T. W. (2012). Functional MRI

of mild traumatic brain injury (mTBI): Progress and perspectives from
the first decade of studies. Brain Imaging and Behavior, 6, 193–207.
Mercado, E., Murray, S. O., Uyeyama, R. K., Pack, A. A. & Herman, L. M.
(1998). Memory for recent actions in the bottlenosed dolphin (Tursiops
truncates): Repetition of arbitrary behaviors using an abstract rule.
Animal Learning & Behavior, 26, 210–218.
Metternich, B., Wagner, K., Schulze-Bonhage, A., Buschmann, F. &
McCarthy, R. A. (2013). Flashbulb memories in patients with temporal
lobe epilepsy. Epilepsy & Behavior, 28, 71–77.
Mez, J., Stern, R. A. & McKee, A. C. (2013). Chronic traumatic
encephalopathy: Where are we and where are we going? Current
Neurology and Neuroscience Reports, 13, 407.
Mickley Steinmetz, K. R., Schmidt, K., Zucker, H. R. & Kensinger, E. A.
(2012). The effect of emotional arousal and retention delay on
subsequent-memory effects. Cognitive Neuroscience, 3, 150–159.
Miller, E. K., Freedman, D. J. & Wallis, J. D. (2002). The prefrontal cortex:
Categories, concepts and cognition. Philosophical Transactions of the
Royal Society of London B: Biological Sciences, 357, 1123–1136.
Milner, B. (1968). Disorders of memory after brain lesions in man.
Neuropsychologia, 6, 175–179.
Mitchell, K. J. & Johnson, M. K. (2009). Source monitoring 15 years later:
What have we learned from fMRI about the neural mechanisms of
source memory? Psychological Bulletin, 135, 638–677.
Nadel, L. & Bohbot, V. (2001). Consolidation of memory. Hippocampus, 11,
56–60.
Nadel, L. & Moscovitch, M. (1997). Memory consolidation, retrograde
amnesia and the hippocampal complex. Current Opinion in
Neurobiology, 7, 217–227.
Naghavi, H. R. & Nyberg, L. (2005). Common fronto-parietal activity in
attention, memory, and consciousness: Shared demands on
integration? Consciousness and Cognition, 14, 390–425.
Nakada, T., Kwee, I. L., Fujii, Y. & Knight, R. T. (2005). High-field, T2
reversed MRI of the hippocampus in transient global amnesia.
Neurology, 64, 1170–1174.
Naya, Y. & Suzuki, W. A. (2011). Integrating what and when across the
primate medial temporal lobe. Science, 333, 773–776.
Nunez, P. L. & Srinivasan, R. (2005). Electric Fields of the Brain:
The Neurophysis of EEG (2nd edn.). New York: Oxford University Press.
O’Keefe, J. & Dostrovsky, J. (1971). The hippocampus as a spatial map:
Preliminary evidence from unit activity in the freely-moving rat. Brain
Research, 34, 171–175.
References 261

Olesen, P. J., Westerberg, H. & Klingberg, T. (2004). Increased prefrontal

and parietal activity after training of working memory. Nature
Neuroscience, 7, 75–79.
O’Neill, J., Pleydell-Bouverie, B., Dupret, D. & Csicsvari, J. (2010). Play it
again: Reactivation of waking experience and memory. Trends in
Neurosciences, 33, 220–229.
Otten, L. J. & Rugg, M. D. (2001). When more means less: Neural activity
related to unsuccessful memory encoding. Current Biology, 11,
1528–1530.
Paller, K. A., Voss, J. L. & Boehm, S. G. (2007). Validating neural correlates
of familiarity. Trends in Cognitive Sciences, 11, 243–250.
Pastalkova, E., Itskov, V., Amarasingham, A. & Buzsáki, G. (2008).
Internally generated cell assembly sequences in the rat hippocampus.
Science, 321, 1322–1327.
Patzke, N., Olaleye, O., Haagensen, M., Hof, P. R., Ihunwo, A. O. &
Manger, P. R. (2014). Organization and chemical neuroanatomy of the
African elephant (Loxodonta africana) hippocampus. Brain Structure
and Function, 219, 1587–1601.
Payne, J. D. (2010). Memory consolidation, the diurnal rhythm of cortisol,
and the nature of dreams: A new hypothesis. International Review of
Neurobiology, 92, 101–134.
Pearson, J., Naselaris, T., Holmes, E. A. & Kosslyn, S. M. (2015). Mental
imagery: Functional mechanisms and clinical applications. Trends in
Cognitive Sciences, 19, 590–602.
Penhune, V. B. & Doyon, J. (2002). Dynamic cortical and subcortical
networks in learning and delayed recall of timed motor sequences.
The Journal of Neuroscience, 22, 1397–1406.
Penolazzi, B., Stramaccia, D. F., Braga, M., Mondini, S. & Galfano, G.
(2014). Human memory retrieval and inhibitory control in the brain:
Beyond correlational evidence. The Journal of Neuroscience, 34,
6606–6610.
Pessoa, L. & Adolphs, R. (2010). Emotion processing and the amygdala:
From a “low road” to “many roads” of evaluating biological
significance. Nature Reviews Neuroscience, 11, 773–783.
Phelps, E. A. (2004). Human emotion and memory: Interactions of the
amygdala and hippocampal complex. Current Opinion in
Neurobiology, 14, 198–202.
Platt, J. R. (1964). Strong inference. Science, 146, 347–353.
Posner, M. I. (1980). Orienting of attention. The Quarterly Journal of
Experimental Psychology, 32, 3–25.
Pratte, M. S. & Tong, F. (2014). Spatial specificity of working memory
representations in the early visual cortex. Journal of Vision, 14(3):22, 1–12.
262 References

Preston, A. R. & Eichenbaum, H. (2013). Interplay of hippocampus

and prefrontal cortex in memory. Current Biology, 23,
R764–773.
Preston, A. R. & Gabrieli, J. D. (2008). Dissociation between explicit
memory and configural memory in the human medial temporal lobe.
Cerebral Cortex, 18, 2192–2207.
Price, C. J. (2000). The anatomy of language: Contributions from functional
neuroimaging. Journal of Anatomy, 197, 335–359.
Purpura, D. P. & McMurtry, J. G. (1965). Intracellular activities and evoked
potential changes during polarization of motor cortex. Journal of
Neurophysiology, 28, 166–185.
Qin, Y. L., McNaughton, B. L., Skaggs, W. E. & Barnes, C. A. (1997).
Memory reprocessing in corticocortical and hippocampocortical
neuronal ensembles. Philosophical Transactions of the Royal Society of
London B: Biological Sciences, 352, 1525–1533.
Quinette, P., Guillery-Girard, B., Dayan, J., de la Sayette, V., Marquis, S.,
Viader, F., Desgranges, B. & Eustache, F. (2006). What does transient
global amnesia really mean? Review of the literature and thorough
study of 142 cases. Brain, 129, 1640–1658.
Rattenborg, N. C., Martinez-Gonzalez, D., Roth, T. C., & Pravosudov, V. V.
(2011). Hippocampal memory consolidation during sleep:
A comparison of mammals and birds. Biological Reviews of the
Cambridge Philosophical Society, 86, 658–691.
Raz, A., Packard, M. G., Alexander, G. M., Buhle, J. T., Zhu, H., Yu, S. &
Peterson, B. S. (2009). A slice of pi: An exploratory neuroimaging
study of digit encoding and retrieval in a superior memorist. Neurocase,
15, 361–372.
Reiman, E. M. & Jagust, W. J. (2012). Brain imaging in the study of
Alzheimer’s disease. NeuroImage, 61, 505–516.
Roediger, H. L. & McDermott, K. B. (1995). Creating false memories:
Remembering words not presented in lists. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 21, 803–814.
Ross, R. S. & Slotnick, S. D. (2008). The hippocampus is preferentially
associated with memory for spatial context. Journal of Cognitive
Neuroscience, 20, 432–446.
Roux, F. & Uhlhaas, P. J. (2014). Working memory and neural oscillations:
Alpha-gamma versus theta-gamma codes for distinct WM
information? Trends in Cognitive Science, 18, 16–25.
Roux, F., Wibral, M., Mohr, H. M., Singer, W. & Uhlhaas, P. J. (2012).
Gamma-band activity in human prefrontal cortex codes for the number
of relevant items maintained in working memory. The Journal of
Neuroscience, 32, 12411–12420.
References 263

Rudoy, J. D., Voss, J. L., Westerberg, C. E. & Paller, K. A. (2009).

Strengthening individual memories by reactivating them during sleep.
Science, 326, 1079.
Ruff, C. C., Blankenburg, F., Bjoertomt, O., Bestmann, S., Freeman, E.,
Haynes, J. D., Rees, G., Josephs, O., Deichmann, R. & Driver, J.
(2006). Concurrent TMS-fMRI and psychophysics reveal frontal
influences on human retinotopic visual cortex. Current Biology, 16,
1479–1488.
Rugg, M. D. & Curran, T. (2007). Event-related potentials and recognition
memory. Trends in Cognitive Sciences, 11, 251–257.
Rugg, M. D. & Vilberg, K. L. (2013). Brain networks underlying episodic
memory retrieval. Current Opinion in Neurobiology, 23, 255–260.
Sala, J. B., Rämä, P. & Courtney, S. M. (2003). Functional topography of
a distributed neural system for spatial and nonspatial information
maintenance in working memory. Neuropsychologia, 41, 341–356.
Sauseng, P., Klimesch, W., Heise, K. F., Gruber, W. R., Holz, E.,
Karim, A. A., Glennon, M., Gerloff, C., Birbaumer, N. &
Hummel, F. C. (2009). Brain oscillatory substrates of visual short-term
memory capacity. Current Biology, 19, 1846–1852.
Schacter, D. L. (1999). The seven sins of memory. Insights from psychology
and cognitive neuroscience. American Psychologist, 54, 182–203.
Schacter, D. L. & Loftus, E. F. (2013). Memory and law: What can cognitive
neuroscience contribute? Nature Neuroscience, 16, 119–123.
Schacter, D. L. & Slotnick, S. D. (2004). The cognitive neuroscience of
memory distortion. Neuron, 44, 149–160.
Schacter, D. L., Dobbins, I. G. & Schnyer, D. M. (2004). Specificity of
priming: A cognitive neuroscience perspective. Nature Reviews
Neuroscience, 5, 853–862.
Schacter, D. L., Guerin, S. A. & St Jacques, P. L. (2011). Memory distortion:
An adaptive perspective. Trends in Cognitive Sciences, 15, 467–474.
Schacter, D. L., Wig, G. S. & Stevens, W. D. (2007). Reductions in cortical
activity during priming. Current Opinion in Neurobiology, 17, 171–176.
Scoville, W. B. & Milner, B. (1957). Loss of recent memory after bilateral
hippocampal lesions. Journal of Neurology, Neurosurgery, &
Psychiatry, 20, 11–21.
Serences, J. T., Ester, E. F., Vogel, E. K. & Awh, E. (2009). Stimulus-specific
delay activity in human primary visual cortex. Psychological Science,
20, 207–214.
Sharot, T., Martorella, E. A., Delgado, M. R. & Phelps, E. A. (2007). How
personal experience modulates the neural circuitry of memories
of September 11. Proceedings of the National Academy of Sciences of
the United States of America, 104, 389–394.
264 References

Shin, Y. I., Foerster, Á. & Nitsche, M. A. (2015). Reprint of: Transcranial

direct current stimulation (tDCS) – Application in neuropsychology.
Neuropsychologia, 74, 74–95.
Shrager, Y., Kirwan, C. B. & Squire, L. R. (2008). Activity in both
hippocampus and perirhinal cortex predicts the memory strength of
subsequently remembered information. Neuron, 59, 547–553.
Simmons, W. K., Reddish, M., Bellgowan, P. S. & Martin, A. (2010).
The selectivity and functional connectivity of the anterior temporal
lobes. Cerebral Cortex, 20, 813–825.
Skipper, L. M., Ross, L. A. & Olson, I. R. (2011). Sensory and semantic
category subdivisions within the anterior temporal lobes.
Neuropsychologia, 49, 3419–3429.
Slobounov, S. M., Zhang, K., Pennell, D., Ray, W., Johnson, B. &
Sebastianelli, W. (2010). Functional abnormalities in normally
appearing athletes following mild traumatic brain injury: A functional
MRI study. Experimental Brain Research, 202, 341–354.
Slotnick, S. D. (2004a). Source localization of ERP generators.
In T. C. Handy (ed.), Event-Related Potentials: A Methods Handbook
(pp. 149–166). Cambridge, MA: The MIT Press.
(2004b). Visual memory and visual perception recruit common
neural substrates. Behavioral and Cognitive Neuroscience Reviews, 3,
207–221.
(2009a). Memory for color reactivates color processing region.
NeuroReport, 20, 1568–1571.
(2009b). Rapid retinotopic reactivation during spatial memory. Brain
Research, 1268, 97–111.
(2010a). Does the hippocampus mediate objective binding or subjective
remembering? NeuroImage, 49, 1769–1776.
(2010b). Synchronous retinotopic frontal-temporal activity during
long-term memory for spatial location. Brain Research, 1330, 89–100.
(2013a). Controversies in Cognitive Neuroscience. Basingstoke, UK:
Palgrave Macmillan.
(2013b). The nature of recollection in behavior and the brain.
NeuroReport, 24, 663–670.
Slotnick, S. D. & Dodson, C. S. (2005). Support for a continuous
(single-process) model of recognition memory and source memory.
Memory & Cognition, 33, 151–170.
Slotnick, S. D. & Schacter, D. L. (2004). A sensory signature that
distinguishes true from false memories. Nature Neuroscience, 7,
664–672.
(2006). The nature of memory related activity in early visual areas.
Neuropsychologia, 44, 2874–2886.
References 265

Slotnick, S. D. & Schacter, D. L. (2007). The cognitive neuroscience of

memory and consciousness. In P. D. Zelazo, M. Moscovitch &
E. Thompson (Eds.), Cambridge Handbook of Consciousness
(pp. 809–827). New York: Cambridge University Press.
Slotnick, S. D. & Thakral, P. P. (2011). Memory for motion and spatial
location is mediated by contralateral and ipsilateral motion processing
cortex. NeuroImage, 55, 794–800.
Slotnick, S. D. & White, R. C. (2013). The fusiform face area responds
equivalently to faces and abstract shapes in the left and central visual
fields. NeuroImage, 83, 408–17.
Slotnick, S. D., Schwarzbach, J. & Yantis, S. (2003). Attentional inhibition of
visual processing in human striate and extrastriate cortex. NeuroImage,
19, 1602–1611.
Slotnick, S. D., Moo, L. R., Kraut, M. A., Lesser, R. P. & Hart, J. Jr. (2002).
Interactions between thalamic and cortical rhythms during semantic
memory recall in human. Proceedings of the National Academy of
Sciences of the United States of America, 99, 6440–6443.
Slotnick, S. D., Moo, L. R., Segal, J. B. & Hart, J., Jr. (2003). Distinct
prefrontal cortex activity associated with item memory and source
memory for visual shapes. Cognitive Brain Research, 17, 75–82.
Slotnick, S. D., Thompson, W. L. & Kosslyn, S. M. (2005). Visual mental
imagery induces retinotopically organized activation of early visual
areas. Cerebral Cortex, 15, 1570–1583.
(2012). Visual memory and visual mental imagery recruit common
control and sensory regions of the brain. Cognitive Neuroscience, 3,
14–20.
Smith, C. N. & Squire, L. R. (2009). Medial temporal lobe activity during
retrieval of semantic memory is related to the age of the memory.
The Journal of Neuroscience, 29, 930–938.
Spaniol, J., Davidson, P. S., Kim, A. S., Han, H., Moscovitch, M. &
Grady, C. L. (2009). Event-related fMRI studies of episodic encoding
and retrieval: Meta-analyses using activation likelihood estimation.
Neuropsychologia, 47, 1765–1779.
Špinka, M., Duncan, I. J. H. & Widowski, T. M. (1998). Do domestic pigs
prefer short-term to medium-term confinement? Applied Animal
Behavior Science, 58, 221–232.
Sprague, T. C., Ester, E. F. & Serences, J. T. (2014). Reconstructions of
information in visual spatial working memory degrade with memory
load. Current Biology, 24, 2174–2180.
Spurzheim, J. G. (1827). Outlines of Phrenology; Being also a Manual of
Reference for the Marked Busts. London, England: Treuttel, Wurtz, &
Richter.
266 References

Squire, L. R. (1992). Memory and the hippocampus: A synthesis from

findings with rats, monkeys, and humans. Psychological Review, 99,
195–231.
Squire, L. R., Wixted, J. T. & Clark, R. E. (2007). Recognition memory and
the medial temporal lobe: A new perspective. Nature Reviews
Neuroscience, 8, 872–883.
Sreenivasan, K. K., Vytlacil, J. & D’Esposito, M. (2014). Distributed and
dynamic storage of working memory stimulus information in
extrastriate cortex. Journal of Cognitive Neuroscience, 26,
1141–1153.
Staudigl, T., Hanslmayr, S. & Bäuml, K. H. (2010). Theta oscillations reflect
the dynamics of interference in episodic memory retrieval. The Journal
of Neuroscience, 30, 11356–11362.
Steele, C. J. & Penhune, V. B. (2010). Specific increases within global
decreases: A functional magnetic resonance imaging investigation of
five days of motor sequence learning. The Journal of Neuroscience, 30,
8332–8341.
Stickgold, R. & Walker, M. P. (2005). Memory consolidation and
reconsolidation: What is the role of sleep? Trends in Neurosciences, 28,
408–415.
Stoub, T. R., deToledo-Morrell, L., Stebbins, G. T., Leurgans, S.,
Bennett, D. A. & Shah, R. C. (2006). Hippocampal disconnection
contributes to memory dysfunction in individuals at risk for
Alzheimer’s disease. Proceedings of the National Academy of Sciences
of the United States of America, 103, 10041–10045.
Suddendorf, T. & Corballis, M. C. (2007). The evolution of foresight: What is
mental time travel, and is it unique to humans? Behavioral and Brain
Sciences, 30, 299–313.
Suthana, N. A., Parikshak, N. N., Ekstrom, A. D., Ison, M. J., Knowlton, B. J.,
Bookheimer, S. Y. & Fried, I. (2015). Specific responses of human
hippocampal neurons are associated with better memory. Proceedings
of the National Academy of Sciences of the United States of America,
112, 10503–10508.
Sweeney-Reed, C. M., Zaehle, T., Voges, J., Schmitt, F. C., Buentjen, L.,
Kopitzki, K., Esslinger, C., Hinrichs, H., Heinze, H. J., Knight, R. T. &
Richardson-Klavehn, A. (2014). Corticothalamic phase synchrony and
cross-frequency coupling predict human memory formation. Elife, 3,
e05352.
Thakral, P. P. & Slotnick, S. D. (2009). The role of parietal cortex during
sustained visual spatial attention. Brain Research, 1302, 157–166.
(2013). The role of spatial attention during spatial encoding. Cognitive
Neuroscience, 4, 73–80.
References 267

Thakral, P. P., Jacobs, C. M. & Slotnick, S. D. (forthcoming). An attention

account of neural priming. Memory.
Tromp, D., Dufour, A., Lithfous, S., Pebayle, T. & Després, O. (2015).
Episodic memory in normal aging and Alzheimer disease: Insights
from imaging and behavioral studies. Ageing Research Reviews, 24,
232–262.
Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26, 1–12.
(2005). Episodic memory and autonoesis: Uniquely human? In H. S. Terrace
& J. Metcalfe (Eds.), The Missing Link in Cognition: Origins of
Self-Reflective Consciousness (pp. 3–56). New York: Oxford
University Press.
Uncapher, M. R. & Wagner, A. D. (2009). Posterior parietal cortex and
episodic encoding: Insights from fMRI subsequent memory effects and
dual-attention theory. Neurobiology of Learning and Memory, 91,
139–154.
Ungerleider, L. G., Doyon, J. & Karni, A. (2002). Imaging brain plasticity
during motor skill learning. Neurobiology of Learning and Memory, 78,
553–564.
Uttal, W. R. (2003). The New Phrenology. Cambridge, MA: The MIT Press.
van de Ven, V., Jacobs, C. M. & Sack, A. T. (2012). Topographic contribution
of early visual cortex to short-term memory consolidation:
A transcranial magnetic stimulation study. The Journal of
Neuroscience, 32, 4–11.
van Wyhe, J. (2004). Phrenology and the Origins of Victorian Scientific
Naturalism. Aldershot, England: Ashgate Publishing Limited.
Vigneau, M., Beaucousin, V., Hervé, P. Y., Duffau, H., Crivello, F.,
Houdé, O., Mazoyer, B. & Tzourio-Mazoyer, N. (2006). Meta-
analyzing left hemisphere language areas: Phonology, semantics, and
sentence processing. NeuroImage, 30, 1414–1432.
Vilberg, K. L. & Rugg, M. D. (2009). Functional significance of
retrieval-related activity in lateral parietal cortex: Evidence from fMRI
and ERPs. Human Brain Mapping, 30, 1490–1501.
Vilberg, K. L., Moosavi, R. F. & Rugg, M. D. (2006). The relationship
between electrophysiological correlates of recollection and amount of
information retrieved. Brain Research, 1122, 161–170.
Voss, J. L., Schendan, H. E. & Paller, K. A. (2010). Finding meaning in novel
geometric shapes influences electrophysiological correlates of
repetition and dissociates perceptual and conceptual priming.
NeuroImage, 49, 2879–2889.
Vuilleumier, P., Henson, R. N., Driver, J. & Dolan, R. J. (2002). Multiple
levels of visual object constancy revealed by event-related fMRI of
repetition priming. Nature Neuroscience, 5, 491–499.
268 References

Wagner, A. D. & Davachi, L. (2001). Cognitive neuroscience: Forgetting of

things past. Current Biology, 11, R964–967.
Wagner, A. D., Shannon, B. J., Kahn, I. & Buckner, R. L. (2005). Parietal
lobe contributions to episodic memory retrieval. Trends in Cognitive
Sciences, 9, 445–453.
Waldhauser, G. T., Johansson, M. & Hanslmayr, S. (2012). Alpha/beta
oscillations indicate inhibition of interfering visual memories.
The Journal of Neuroscience, 32, 1953–1961.
Weiner, K. S., Sayres, R., Vinberg, J. & Grill-Spector, K. (2010).
fMRI-adaptation and category selectivity in human ventral temporal
cortex: regional differences across time scales. Journal of
Neurophysiology, 103, 3349–3365.
Wheeler, M. E. & Buckner, R. L. (2003). Functional dissociation among
components of remembering: Control, perceived oldness, and content.
The Journal of Neuroscience, 23, 3869–3880.
(2004). Functional-anatomic correlates of remembering and knowing.
NeuroImage, 21, 1337–1349.
Wheeler, M. E., Petersen, S. E. & Buckner, R. L. (2000). Memory’s echo:
Vivid remembering reactivates sensory-specific cortex. Proceedings of
the National Academy of Sciences of the United States of America, 97,
11125–11129.
Wig, G. S., Grafton, S. T., Demos, K. E. & Kelley, W. M. (2005). Reductions
in neural activity underlie behavioral components of repetition
priming. Nature Neuroscience, 8, 1228–1233.
Willment, K. C. & Golby, A. (2013). Hemispheric lateralization interrupted:
Material-specific memory deficits in temporal lobe epilepsy. Frontiers
in Human Neuroscience, 7, 546.
Wimber, M., Alink, A., Charest, I., Kriegeskorte, N. & Anderson, M. C.
(2015). Retrieval induces adaptive forgetting of competing
memories via cortical pattern suppression. Nature Neuroscience, 18,
582–589.
Wimber, M., Bäuml, K. H., Bergström, Z., Markopoulos, G., Heinze, H. J. &
Richardson-Klavehn, A. (2008). Neural markers of inhibition in human
memory retrieval. The Journal of Neuroscience, 28, 13419–13427.
Winocur, G. & Moscovitch, M. (2011). Memory transformation and systems
consolidation. Journal of the International Neuropsychological Society,
17, 766–780.
Winters, B. D., Saksida, L. M. & Bussey, T. J. (2008). Object recognition
memory: Neurobiological mechanisms of encoding, consolidation and
retrieval. Neuroscience & Biobehavioral Reviews, 32, 1055–1070.
Wixted, J. T. (2007). Dual-process theory and signal-detection theory of
recognition memory. Psychological Review, 114, 152–176.
References 269

Woollams, A. M., Taylor, J. R., Karayanidis, F. & Henson, R. N. (2008).

Event-related potentials associated with masked priming of test cues
reveal multiple potential contributions to recognition memory. Journal
of Cognitive Neuroscience, 20, 1114–1129.
Woollett, K., Spiers, H. J. & Maguire, E. A. (2009). Talent in the taxi:
A model system for exploring expertise. Philosophical Transactions of
the Royal Society of London. Series B, Biological Sciences, 364,
1407–1416.
Woroch, B. & Gonsalves, B. D. (2010). Event-related potential correlates of
item and source memory strength. Brain Research, 1317, 180–191.
Yang, Y., Kim, S. & Kim, J. H. (2008). Ischemic evidence of transient global
amnesia: Location of the lesion in the hippocampus. Journal of Clinical
Neurology, 4, 59–66.
Yassa, M. A., Stark, S. M., Bakker, A., Albert, M. S., Gallagher, M. &
Stark, C. E. (2010). High-resolution structural and functional MRI of
hippocampal CA3 and dentate gyrus in patients with amnestic Mild
Cognitive Impairment. NeuroImage, 51, 1242–52.
Young, K. D., Bellgowan, P. S., Bodurka, J. & Drevets, W. C. (2013).
Functional neuroimaging of sex differences in autobiographical
memory recall. Human Brain Mapping, 34, 3320–3332.
Author Index

Abraham, W. C. 203 Bohbot, V. 53

Addis, D. R. 56, 163 Bonnici, H. M. 55
Adolphs, R. 166 Bookheimer, S. Y. 51
Albert, M. S. 176, 194 Born, J. 57, 58
Alink, A. 94 Böye, M. 215
Allen, R. 117, 119 Brickhouse, M. 51
Alpert, N. M. 142 Bridger, E. K. 77
Alvarez, P. 53 Brown, R. 103
Amarasingham, A. 207 Buckner, R. L. 16, 18, 28, 47, 51, 91, 155, 177
Anderson, M. C. 94, 96, 97 Buschkuehl, M. 123
Andreano, J. M. 61, 64 Buschmann, F. 105
Aristotle 234 Bussey, T. J. 197
Augath, M. 35 Buzsáki, G. 205, 207
Awh, E. 112, 154
Axmacher, N. 58 Cabeza, R. 50, 90, 100, 155
Cahill, L. 61, 64
Babb, S. J. 210 Carrow, S. 208
Baddeley, A. 117, 119 Carver, F. W. 136
Bader, R. 77 Chaieb, L. 122
Badgaiyan, R. D. 142 Chao, L. L. 51
Bakker, A. 176, 185, 195 Charest, I. 94
Baldauf, D. 85 Chen, J. K. 183
Banks, S. J. 62 Chow, C. C. 140
Bar, M. 136 Chun, M. M. 142, 143, 144, 221
Barnes, C. A. 205 Ciaramelli, E. 50, 155
Barrett, L. F. 166 Clark, R. E. 49
Bartels, A. 224 Clayton, N. S. 212
Bartsch, T. 55, 191, 192, 193 Cohen, D. 34
Baudry, M. 203 Cole, L. C. 189
Bäuml, K. H. 95 Collingridge, G. L. 203
Bear, M. F. 203 Copara, M. S. 48
Beck, D. M. 224 Corballis, M. C. 210, 216
Bellgowan, P. S. 52, 179 Corbetta, M. 154
Bergmann, H. C. 116 Corwin, J. 189
Berner, M. P. 148 Courtney, S. M. 109
Binder, J. R. 190 Crone, E. A. 123
Blakemore, C. B. 187 Crystal, J. D. 210
Bliss, T. V. 201, 203 Csicsvari, J. 203
Bliss-Moreau, E. 166 Cubelli, R. 104
Boehm, S. G. 76 Cuffin, B. N. 34
Author Index 271

Curran, T. 72, 75, 79 Falconer, M. A. 187

Curtis, C. E. 111, 154 Fanselow, M. S. 54
Fell, J. 58
Davachi, L. 61, 90 Felleman, D. J. 15, 227
Davidson, P. S. 104, 105 Feredoes, E. 229
Deese, J. 98 Fernández, G. 116
Delgado, M. R. 105 Fiebach, C. J. 136
Della Sala, S. 104 Finke, C. 117
Demos, K. E. 133 Flashman, L. A. 195
Depue, B. E. 97 Fletcher, E. M. 152
Desimone, R. 85 Floyer-Lea, A. 147
Desmond, J. E. 51, 165 Foerster, Á. 42
D’Esposito, M. 111, 113 Fortin, N. 197–198
Després, O. 177 Freedman, D. J. 168
Dettwiler, A. 181 Friederici, A. D. 165
Deuschl, G. 55, 192 Friese, U. 82
Deutsch, G. K. 189 Frings, L. 62
Diana, R. A. 48, 50 Fujii, Y. 191
Dickerson, B. C. 51, 172, 177
Dickinson, A. 212 Gabrieli, J. D. 51, 143, 165
Diedrichsen, J. 147 Gagnepain, P. 97
Dobbins, I. G. 131, 136 Galfano, G. 94
Dodson, C. S. 6, 74 Gallagher, M. 174, 176
Döhring, J. 55, 191, 192 Ganis, G. 159
Dolan, R. 133, 135 Garoff-Eaton, R. J. 102
Domoto-Reilly, K. 51 Gazzaley, A. 154
Donders, F. C. 20 Gazzaniga, M. S. 232
Dostrovsky, J. 204 Ghuman, A. S. 136
Doyon, J. 147 Gierhan, S. M. 165
Draguhn, A. 58 Gilbert, J. R. 136
Driver, J. 133, 229 Girardeau, G. 203
Dufour, A. 177 Glisky, E. L. 104
Duncan, I. J. H. 212 Glosser, G. 189
Duncan, K. 61 Golby, A. 186, 189
Dupret, D. 203 Goldstein, J. M. 61
Gonsalves, B. D. 75, 78
Eichenbaum, H. 197–198, 199, 205, Gosselin, N. 184
206, 208 Gotts, S. J. 136, 140
Einstein, A. 34, 234 Grafton, S. T. 133
Ekstrom, A. D. 48, 121, 204 Grill-Spector, K. 138, 140
Elger, C. E. 58 Grol, M. J. 123
Ellenbogen, J. M. 57 Grubb, N. R. 39
Engel, A. K. 82 Gruber, T. 82, 136, 138
Engel, S. A. 51 Guerin, S. A. 89
Engell, A. D. 136, 138 Guerin, S. J. 195
Ester, E. F. 112, 113 Guo, C. 77
272 Author Index

Hanlon, F. M. 179 Kahn, I. 47, 155

Hannula, D. E. 115, 116, 121, 144 Kanwisher, N. 221, 222
Hanslmayr, S. 95, 97 Kapur, N. 66
Harrison, S. A. 112 Karanian, J. M. 16, 100, 103
Hart, B. L. 214 Karayanidi, F. 78
Hart, J. Jr 18, 84 Karni, A. 147
Hart, L. A. 214 Kemp, A. 203
Hasselmo, M. E. 206 Kensinger, E. A. 168
Hausberger, M. 215 Kessels, R. P. 116
Haxby, J. V. 15, 222 Kiesel, A. 148
Helgadóttir, H. 58 Kim, A. S. 60
Helmholtz, H. 20 Kim, H. 90, 100
Henson, R. 78, 97, 133, Kim, J. H. 191
135, 138 Kim, S. 191
Herman, L. M. 213 Kirwan, C. B. 90
Herrmann, C. S. 43 Kitchener, E. G. 37
Hirst, W. 104 Klingberg, T. 124, 125
Hodges, J. R. 191 Knight, R. T. 191
Hoffmann, J. 148 Knowlton, B. J. 51, 54
Holmes, E. A. 159 Kober, H. 166
Hopfinger, J. B. 152, 157 Koh, M. T. 176
Hopkins, R. O. 37 Komorowski, R. 197–198
Hou, M. 77 Kornysheva, K. 147
Hsieh, L. T. 121, 122 Kosslyn, S. M. 159, 162
Huettel, S. A. 26 Köster, M. 82
Hutchinson, J. B. 157 Koutstaal, W. 131
Kraus, B. J. 206
Ikkai, A. 154 Krauss, G. 176
Isham, E. A. 48 Kraut, M. A. 84
Itskov, V. 207 Kremers, D. 215
Ivry, R. B. 232 Kriegeskorte, N. 94
Kriukova, O. 77
Jacobs, C. M. 113, 135 Kulik, J. 103
Jaeggi, S. M. 123, 125 Kunde, W. 148
Jagust, W. J. 177 Kwee, I. L. 191
James, W. 7
Jansen, O. 55, 192 Ladowski, D. 62
Jaramillo, M. B. 215 Landau, S. M. 178
Ji, D. 205 Larson, J. 202
Jiang, Y. 142, 143 Leal, S. L. 172, 177
Johansson, M. 95 Lemasson, A. 215
Johnson, J. D. 32 LePort, A. K. 67
Johnson, M. K. 49, 72, 105 Lesser, R. P. 84
Jolles, D. D. 123 Li, X. 125
Jones-Gotman, M. 62, 187 Liang, K. Y. 179
Jonides, J. 123 Libby, L. A. 116, 121
Author Index 273

Lindquist, K. A. 166 Moo, L. R. 18, 84

Lipton, P. 197–198 Moosavi, R. F. 73
Lithfous, S. 177 Moscovitch, M. 50, 53, 155
Liu, T. 154 Müller, M. M. 136
Loftus, E. F. 100 Murray, S. O. 213
Logothetis, N. K. 35
Lømo, T. 201 Nadel, L. 53
Lynch, G. 202 Naghavi, H. R. 157
Nakada, T. 191
Ma, L. 146 Naselaris, T. 159
MacDonald, C. J. 208 Naya, Y. 208
Maguire, E. A. 65–66 Neuling, T. 43
Malmivuo, J. 34 Nitsche, M. A. 42
Mamourian, A. C. 195 Nobre, A. C. 154
Manahan-Vaughan, D. 203 Nunez, P. L. 30
Mangun, G. R. 152, 232 Nyberg, L. 157
Manns, J. R. 37, 199
Marshall, L. 57, 58 Oeltermann, A. 35
Martin, A. 51, 52, 136, 138, 140 Oh, S.H. 154
Martinez-Gonzalez, D. 205 O’Keefe, J. 204
Martorella, E. A. 105 Olesen, P. J. 124
Matsumura, N. 204 Olson, I. R. 50, 52, 155
Matthews, P. M. 147 O’Neill, J. 203
Maxwell, J. C. 34 Otten, L. J. 89
Mayer, A. R. 179
McAllister, T. W. 179, 181 Pack, A. A. 213
McCarthy, G. 26, 136, 138 Paller, K. A. 59, 76, 77, 79
McCarthy, R. A. 105 Pastalkova, E. 207
McDermott, J. 221 Patzke, N. 214
McDermott, K. B. 98 Pauls, J. 35
McDonald, B. C. 179 Payne, J. D. 57, 58
McKee, A. C. 184 Pearson, J. 159, 161
McMurtry, J. G. 42 Pebayle, T. 177
McNaughton, B. L. 205 Penhune, V. B. 147
Mecklinger, A. 77 Penolazzi, B. 94
Mercado, E. 213 Perrig, W. J. 123
Metternich, B. 105 Pessoa, L. 166
Mez, J. 184 Phelps, E. A. 105, 143, 144, 168
Mickley Steinmetz, K. R. 168 Pinter-Wollman, N. 214
Miller, E. K. 168 Place, R. 208
Milner, B. 12, 187 Platt, J. R. 56, 234, 235
Minton, B. R. 32 Pleydell-Bouverie, B. 203
Mitchell, J. P. 49, 61, 105 Pohl, C. 148
Mohr, R. M. 121 Poldrack, R. A. 51, 165
Mölle, M. 58 Posner, M. I. 151
Mondini, S. 94 Pratte, M. S. 113
274 Author Index

Pravosudov, V. V. 205 Schulze-Bonhage, A. 105

Preston, A. R. 143, 205 Schwarzbach, J. 152
Price, C. J. 164, 165 Scoville, W. B. 12
Purpura, D. P. 42 Segal, J. B. 18
Serences, J. T. 112, 113, 154
Qin, Y. L. 205 Shallice, T. 135
Quinette, P. 191, 193 Shannon, B. J. 47, 155
Sharot, T. 105
Rach, S. 43 Shin, Y. I. 42
Rämä, P. 109 Shrager, Y. 90
Ranganath, C. 48, 115, 116, 121, 144, 199 Shulman, G. L. 154
Rapcsak, S. Z. 104 Simmons, W. K. 52
Rattenborg, N. C. 205 Singer, W. 82, 121
Raz, A. 67 Skaggs, W. E. 205
Reddish, M. 52 Slobounov, S. M. 183
Reed, J. M. 37 Slotnick, S. D. 6, 15, 16, 18, 32, 41, 49, 50, 51,
Reiman, E. M. 177 61, 74, 75, 79, 84, 98, 100, 101, 102, 103,
Rijpkema, M. 116 111, 135, 141, 152, 154, 155, 157, 159, 162,
Robinson, R. J. 206 222, 224
Roediger, H. L. 98 Smith, C. N. 54
Rohr, A. 55, 192 Song, A. W. 26
Rombouts, S. A. 123 Spaniol, J. 60
Ross, L. A. 52 Sparling, M. B. 195
Ross, R. S. 61 Speck, C. L. 176
Roth, T. C. 205 Sperling, R. A. 172, 177
Roux, F. 121, 122 Spiers, H. J. 65–66
Rudoy, J. D. 59 Špinka, M. 212
Ruff, C. C. 228, 229 Spinks, R. 72
Rugg, M. D. 32, 47, 72, 73, 74, 75, 79, 89, 90 Sprague, T. C. 113
Spurzheim, J. G. 220
Sack, A. T. 113 Squire, L. R. 37, 49, 50, 53, 54, 90, 131
Safron, A. 77 Sreenivasan, K. K. 113
Saksida, L. M. 197 Srinivasan, R. 30
Sala, J. B. 109 Stark, C. E. 194
Sapolsky, D. 51 Stark, S. M. 194
Sauseng, P. 119 Staudigl, T. 95
Sauvage, M. 197–198 Steele, C. J. 147
Saykin, A. J. 179, 189, 195 Stern, R. A. 184
Sayres, R. 140 Stevens, W. D. 134
Schacter, D. L. 16, 51, 56, 72, 89, 91, 98, 100, Stickgold, R. 57
101, 102, 131, 134, 135, 142, 163 St. Jacques, P. L. 89
Schendan, H. E. 76 Stoub, T. R. 172
Schmidt, K. 168 Stramaccia, D. F. 94
Schmuck, A. 191 Strüber, D. 43
Schnyer, D. M. 131, 136 Suddendorf, T. 210, 216
Schöne, B. 82 Supp, G. G. 136
Author Index 275

Suthana, N. A. 35 Wagner, K. 105

Suzuki, W. A. 208 Waldhauser, G. T. 95
Sweeney-Reed, C. M. 84 Walker, M. P. 57
Szikla, V. 62 Wallis, J. D. 168
Wang, W. C. 48
Taylor, J. R. 78 Warlow, C. P. 191
Thakral, P. P. 16, 41, 135, 154, 157 Weiner, K. S. 140
Thompson, W. L. 159, 162 Weiskopf, N. 229
Tompary, A. 61 Westerberg, C. E. 59
Tong, F. 112, 113 Westerberg, H. 124
Trinath, T. 35 Wheeler, M. E. 16, 18, 51
Tromp, D. 177 White, J. A. 206
Trujillo-Barreto, N. 82 White, R. C. 15, 222
Tulving, E. 6–7, 210, 212 Wibral, M. 121
Widowski, T. M. 212
Uhlhaas, P. J. 121, 122 Wig, G. S. 133, 134
Uncapher, M. R. 157 Wilding, J. M. 66
Unger, K. 77 Wilhelm, I. 58
Ungerleider, L. G. 147 Willment, K. C. 186, 189
Uttal, W. R. 221 Wilson, M. A. 205
Uyeyama, R. K. 213 Wimber, M. 92, 94, 96
Winocur, G. 53
Valentine, E. R. 66 Winters, B. D. 197–198, 199
Van Buchem, M. A. 123 Wixted, J. T. 5, 49, 74
van de Ven, V. 113 Woldorff, M. G. 152
Van Essen, D. C. 15, 227 Wollett, K. 65–66, 68
van Wyhe, J. 220 Wong, A. T. 56, 163
Vargha-Khadem, F. 117, 119 Wong, D. 202
Verfaellie, M. 104 Woollams, A. M. 78
Vigneau, M. 165 Woollett, K. 66
Vilberg, K. L. 47, 73, 74, 75 Woroch, B. 75, 78
Vinberg, J. 140
Vogel, E. K. 112, 154 Yang, Y. 191, 193
Voss, J. L. 59, 76, 79 Yantis, S. 152, 154
Vuilleumier, P. 133 Yassa, M. A. 172, 174, 177
Vytlacil, J. 113 Yonelinas, A. P. 48, 199

Wager, T. D. 166 Zeki, S. 224

Wagner, A. D. 47, 50, 61, Zucker, H. R. 168
90, 157 Zugaro, M. 203
Subject Index

action potential 201 brain anatomy 8–12, 13–16, 166

affective neuroscience 166–168 medial temporal lobe in animals 199,
alpha frequency band 84, 119–122, 136–138 214
Alzheimer’s disease (AD) 51, 171, 177–179 BRAIN initative (NIH) 226
amnesia brain mapping 232, 234
anterograde 12, 191 brain plasticity (brain training) 122–125
hippocampal lesions 12–13, 101, 191–193 Broca’s area 164–165
retrograde 12, 53, 191 Brodmann areas 11–12
transient global 54–55, 190–193
amnestic mild cognitive impairment central sulcus 11
(aMCI) 172–176 cerebellum 147
amygdala 105, 166–168 chronic traumatic encephalopathy 184
amyloid 177–179 chunking 124
anatomy of the brain see brain anatomy cognitive impairment
animal studies 2, 196–197, 209 Alzheimer’s disease 51, 171, 177–179
episodic memory 210–216 mild (amnestic) 172–176
ethics 216 cognitive neuroscience, as a field of study 2,
long-term potentiation 201–203 232–234
medial temporal lobe and long-term cognitive psychology 2, 25, 232, 234
memory 197–200 color processing area (V8) 15, 152
memory replay 203–205, 215–216 conceptual repetition priming 76–79, 165
time cells 205–209, 215 concussion (traumatic brain injury) 179–186
anterior temporal lobe consolidation of memory 53–56, 65
semantic memory 51–52 memory replay in animals 203–205,
superior autobiographical memory 67 215–216
anterograde amnesia 12, 191 and sleep 56–59, 203, 204
associative memory 6 context memory (source memory) 5–6, 47
associative priming task 141–142 location of brain activity 15, 18–21,
attention 151–159 49, 50
and not forgetting 89, 92 contextual cueing task 142–144
autobiographical memory 47 contralateral visual processing 15, 80, 113,
and imagery for future events 163 152, 154
superior 67 contralateral P1 effect 155–157
transient global amnesia 54–55 control regions of the brain 18–21
attention 154
behavioral measures 25 episodic memory 47–51
behavioral neuroscience 2, 233 semantic memory 51
see also animal studies see also dorsolateral prefrontal cortex;
birds 212, 216 hippocampus; parietal cortex
blocked designs 30 cross-frequency coupling 82–84
Subject Index 277

declarative memory 198 epilepsy, medial temporal lobe 186–190

default network 90–92, 177–178 episodic memory 5
delayed matching-to-sample task 208 in animals 210–216
delayed non-matching-to-sample task location of brain activity 18–21, 47–51,
198–199 214–216
dementia (Alzheimer’s disease) 51, 171, see also autobiographical memory;
177–179 context memory; recollection;
depth electrode recording 35–37, 43–44 “remembering”
direct tasks 142 ERFs (event-related fields) 33
distortion of memory ERPs see event-related potentials
false memories 97–103, 165, 174 ethics of animal studies 216
flashbulb memories 103–105 event-related design 28
dolphins 213, 215 event-related fields (ERFs) 33
dorsolateral prefrontal cortex event-related potentials (ERPs) 30–32
anatomy 10 compared with other techniques 34,
attention 154 43–44
control of memory 18–21, 167–168 FN400 effect 74, 76–79
episodic memory 49–50 implicit memory/repetition priming 135,
false memories 98–100, 102, 165 138
flashbulb memories 105 increased future use 85, 225–226
forgetting recollection vs familiarity 72–75
motivated 96–97 excitatory post-synaptic potential 201, 203
retrieval-induced 94, 165 exercise, and Alzheimer’s disease 178–179
typical 89–90 explicit memory 3–7
implicit memory/repetition priming see also individual types
131–132, 133–135 explicit memory contamination 141
interaction with visual cortex 228–231 extrastriate cortex 14–15, 16, 152
language processing 165
semantic memory 51, 165 face processing area (FFA) 15, 109, 221–224
skill learning 147 facilitation model of repetition priming 139,
working memory 109–112, 113–114 140
interaction with visual cortex 229–231 false memories 97–103, 165
after mild TBI 181–183, 184–185 in aMCI 174
training 124–125 familiarity 6
DRM (Deese-Roediger-McDermott) location of brain activity 51–52
paradigm 97–98 a separate process from recollection?
dynamic causal modeling (DCM) 231–232 74–75, 212
timing of brain activity 72–75
electroencephalography (EEG) 33, 82–85, FN400 effect 74, 76–79
136–138 see also item memory; “knowing”;
slow wave sleep 58, 203 semantic memory
see also event-related potentials (ERPs) fatigue model of repetition priming 139,
elephants 214 140–141
emotional processing 166–168 females, long-term memory 61–64
encoding of memory 59–61, 89, 92 FFA (fusiform face area) 15, 109, 221–224
entorhinal cortex 172–174 flashbulb memories 103–105
278 Subject Index

fMRI see functional magnetic resonance hippocampal sharp-wave ripples 58,

imaging 203–205
FN400 effect 74, 76–79 hippocampus
forgetting 88–89 amnestic mild cognitive impairment
motivated 96–97 172–176
retrieval-induced 92–96, 165 anatomy 10
typical 89–92 in animals see animal studies
frequency of brain activity 71–72, 80–82 binding of item and context 49, 200–201
implicit memory/repetition priming context memory 50
136–138 depth electrode experiments 35–37
long-term memory 82–85 emotional processing 168
long-term potentiation 202–203 episodic memory 49, 214–216
retrieval-induced forgetting 95–96 false memories 98–101
slow waves 58, 203 forgetting
working memory 119–122 motivated 96–97
frontal eye field (FEF) 228 retrieval-induced 94–96
frontal lobe see dorsolateral prefrontal implicit memory 141–145
cortex interactions with cortex 58, 202–203,
functional magnetic resonance imaging 204
(fMRI) 25–29, 43–44, 225 in London taxi drivers 65–66, 68
correlation with electrophysiological long-term memory 12–13, 37–39, 61–64
activity 36 long-term potentiation 201–203
disadvantages/problems memory consolidation 53–56, 65, 203
expense 227 memory replay in animals 203–205,
investigation of interactions using 215–216
SEM or DCM 231–232 memory encoding 61
oversimplistic interpretation of results sex differences 61–64
221–224 skill learning 147
poor temporal resolution 28–29, 225 time cells 205–209, 215
surgical planning in medial temporal lobe transient global amnesia 191–193
epilepsy 190 working memory 114–119, 123
funding of research 227 H. M. (case study) 12–13, 114–115
fusiform cortex, implicit memory/repetition
priming 132–133 imagery 159–163
fusiform face area (FFA) 15, 109, 221–224 and working memory 125–126, 161–162
implicit memory 3–4
gain model of attention 151 consolidation during REM sleep 57
gamma frequency band 82–85, 119–122, locations of brain activity 130–135,
136–138 147
gender differences in long-term memory neural models 138–141
61–64 repetition priming 4, 76–79
gist 97–98, 103 role of the hippocampus 141–145
gyri 10–11 skill learning 146–148
timing of brain activity 135–138
head injuries (traumatic brain injury) inattention (typical forgetting) 89–92
179–186 indirect tasks 142
Subject Index 279

interactions between different brain regions long-term memory 4–5

227–233 in animals see animal studies
dorsolateral prefrontal cortex and visual and attention 155–159
cortex 228–231 consolidation 53–56, 65
hippocampus and cortex 58, 202–203, 204 memory replay in animals 203–205,
intracarotid amobarbital test 189–190 215–216
introspection 6–7 and sleep 56–59, 203, 205
inverse problem 32 encoding 59–61, 89, 92
item memory 5, 7, 18–21, 48–49 episodic memory 5, 18–21, 47–51
in animals 197–200 in animals 210–216
failure 171–172
“knowing” 6–7, 37–39 Alzheimer’s disease 51, 171, 177–179
amnestic mild cognitive impairment
language processing 16, 164–166 172–176
and false memories 102–103, 165 after epilepsy surgery 186–190
verbal memory 61–62, 187, 189 false memories 97–103, 165, 174
lateral occipital complex (LOC) 15 flashbulb memories 103–105
left-parietal old–new effect 72–75 forgetting 89–97
lesional studies 37–39, 43–44 after mild TBI 182–183
hippocampal lesions transient global amnesia 54–55, 192–193
long-term memory 12–13, 37–39 and the hippocampus 12–13, 37–39,
working memory 117–119 61–64, 214–216
perirhinal complex (in animals) 197–198, and imagery 162–163
199 semantic memory 5, 51–52, 53–54, 165
surgical planning in medial temporal lobe sex differences 61–64
epilepsy 186–190 superior abilities 64–68
location of brain activity timing
attention 154, 157 ERP results of recollection vs
emotional processing 166–168 familiarity 72–75
episodic memory 18–21, 47–51, 214–216 FN400 effect 74, 76–79
false memories 98–103, 165 phase and frequency 71–72, 79–85
forgetting long-term potentiation 201–203
motivated 96–97
retrieval-induced 94–96, 165 magnetoencephalography (MEG) 33–34,
typical 89–90 43–44
imagery 161 males, long-term memory 61–64
implicit memory 130–135, 147 medial temporal lobe 10
language processing 16, 164–166 Alzheimer’s disease 177
semantic memory 51–52, 53–54, 165 atrophy in amnestic mild cognitive
skill learning 147 impairment 172–176
visual perception 14–16 epilepsy surgery 186–190
working memory 109–114, 154, 229–231 episodic memory 48–49
see also interactions between different hippocampus see hippocampus
brain regions long-term memory in animals 197–200
London taxi drivers 65–66, 68 parahippocampal cortex 48, 67
long-term depression, hippocampal 203 PPA 15, 109
280 Subject Index

medial temporal lobe (cont.) working memory

perirhinal cortex after mild TBI 181–183, 184–185
contextual cueing task 144 training 124–125
item memory 48–49, 60–61, 197–198, pattern classification algorithm technique
199 112–114
memory consolidation 53–56, 65 pattern separation/pattern completion in
and sleep 56–59, 203, 205 aMCI 174
memory construction 55–56 perforant path 173–174
memory encoding 59–61, 89, 92 perirhinal cortex
memory replay in animals 203–205, 215–216 contextual cueing task 144
memory types 3–7 item memory 48–49, 60–61
method of loci 66 in animals 197–198, 199
mid-frontal old–new effect 72–75 PET (positron emission tomography) 30
FN400 effect 74, 76–79 phase of brain activity 79–80
mild traumatic brain injury (mTBI) 179–186 phase lag 84–85
motion processing area (MT) 15, 41, 154 phrenology 220–221
motivated forgetting 96–97 physical exercise, and Alzheimer’s disease
motor processing regions 11 178–179
skill learning 147 π, ability to remember digits of 67
MRI see functional magnetic resonance place cells 204–205
imaging (fMRI) positron emission tomography (PET) 30
multi-voxel pattern analysis technique PPA (parahippocampal place area) 15,
112–114 109
multitasking 92 prefrontal cortex see dorsolateral prefrontal
cortex
n-back test 181, 186 primate studies 199, 208–209
neurofibrillary tangles 177
neuronal communication 201 recall 5
neurotransmitters 201 recollection 6
nondeclarative memory see implicit memory compared with imagery 162–163
location of brain activity 47–51
occipital lobe see visual cortex a separate process from familiarity?
old–new recognition 4 74–75, 212
timing of brain activity 72–75 timing of brain activity 72–75
orientation gratings 112–113 see also context memory; episodic
memory; “remembering”
parahippocampal cortex 48, 67 REM (rapid eye movement) sleep 56–57
parahippocampal place area (PPA) “remembering” 6–7, 37–39, 47, 72–74
15, 109 false memories 98
parietal cortex 18 repetition priming 4
Alzheimer’s disease 177–178 FN400 effect 76–79
attention 154 location of brain activity 130–135
episodic memory 50 neural models 138–141
false memories 98–100 timing of brain activity 135–138
forgetting 89–90 retinotopic maps 159–161
left-parietal old–new effect 75 retrieval-induced forgetting 92–96, 165
Subject Index 281

retrograde amnesia 12, 53, 191 spontaneous object recognition task

right-frontal old–new effect 75 197–198
striate cortex (V1) 11, 14–15, 112–113
scientific method 90, 234–236 strong inference 234–235
animal studies 209, 216 structural equation modeling (SEM)
competing hypotheses 56, 79, 115, 231–232
145 subsequent memory analysis 59
semantic memory 5 subtractive logic 20
location of brain activity 51–52, 53–54, sulci 10–11
165 synaptic cleft 201
semantic processing 164–165 synchronicity in brain activity 71–72, 79–80
sensory cortex 13–16 systems neuroscience 233
and semantic memory 51
and working memory 109–114 tau protein 177
see also visual cortex TBI (traumatic brain injury) 179–186
sensory reactivation hypothesis 16–17 temporal lobe
sex differences in long-term memory 61–64 anterior 51–52, 67
shape processing lateral 131–132, 134–135
in the fusiform face area 222 medial see hippocampus; medial
in the lateral occipital complex 15 temporal lobe
sharpening model of repetition priming 139, superior posterior 102–103, 165
140–141 temporal memory
short-term memory see working memory in animals 210–213
single-cell recording (depth electrode parahippocampal cortex 48
recording) 35–37, 43–44 time cells 205–209, 215
skill learning 4, 146–148 temporal resolution/timing of brain activity
sleep, memory consolidation 56–59, 203, implicit memory 135–138
205 long-term memory
source memory (context memory) 5–6, 47 ERP results of recollection vs
location of brain activity 18–21, 49, 50 familiarity 72–75
spatial attention 152, 154–157 FN400 effect 74, 76–79
spatial location processing pathway (where phase and frequency 71–72, 79–85
pathway) 15, 109–111 techniques 24–25, 236
spatial memory compared 43–44
in elephants 214 EEG 33
after mild TBI 182–183 ERPs 30–32, 34, 225–226
place cells 204–205 fMRI 28–29, 225
sex differences 61 with high spatial and temporal
spatial resolution of brain activity 24–25, resolution 34–37, 44
43–44 MEG 33–34
ERPs 31–32 TGA (transient global amnesia) 54–55,
fMRI 25–29 190–193
PET 30 thalamic-cortical sleep spindles 58
techniques with both high spatial and thalamus 84
temporal resolution 34–37, 44 theta frequency band 82–85, 95–96,
see also location of brain activity 119–122, 202–203, 204
282 Subject Index

time cells 205–209, 215 imagery 125–126, 159–163

timing of brain activity see temporal interactions with the dorsolateral
resolution/timing of brain activity prefrontal cortex 228–231
transcranial alternating current stimulation repetition priming 135
(tACS) 43 retrieval-induced forgetting 94–96
transcranial direct current stimulation working memory 109–114, 154, 161–162,
(tDCS) 42–44 229–231
transcranial magnetic stimulation (TMS) visual memory 16–17, 187–189
39–41, 43–44
combining with other experimental Wernicke’s area 164–165
techniques 228 what pathway 15, 109–111
transient global amnesia (TGA) 54–55, where pathway 15, 109–111
190–193 working memory (short-term memory) 4–5,
traumatic brain injury (TBI) 179–186 108–109
and attention 154–155
verbal memory 61–62, 187, 189 frequency of brain activity 119–122
verbal processing see language processing and imagery 125–126, 161–162
vertex 228 location of brain activity 109–114, 154
visual cortex 8, 11, 14–16 hippocampus 114–119, 123
attention 151–154 interactions between dorsolateral
central field vs peripheral field 229 prefrontal cortex and visual cortex
contralateral visual field 15, 80, 113, 152, 229–231
154 after mild TBI 179–182, 183–184
contralateral P1 effect 155–157 training and brain plasticity 122–125
false memories 98–103 World Memory Championship contestants
gamma activity 82 66–67