612

Cognitive neuroscience
Thomas D Albright*, Eric R Kandel† and Michael I Posner‡
The last decade of the 20th century has seen the development disparate subfields of biology into one coherent disci-
of cognitive neuroscience as an effort to understand how the pline. Almost all of the other unifications have been
brain represents mental events. We review the areas of spearheaded by the synthetic power of molecular biology.
emotional and motor memory, vision, and higher mental Cognitive neuroscience is distinctive in that the impor-
processes as examples of this new understanding. Progress tant impetus has come from other sources; in particular, a
in all of these areas has been swift and impressive, but much large part of the impetus has come from psychology and
needs to be done to reveal the mechanisms of cognition at the from systems neuroscience.
local circuit and molecular levels. This work will require new
methods for controlling gene expression in higher animals and That psychology should be a driving force for such a syn-
in studying the interactions between neurons at multiple levels. thesis is perhaps not surprising. Psychology provides the
agenda for the brain sciences; it poses the questions about
Addresses mental activity that we ultimately want to address. In
*Howard Hughes Medical Institute and Salk Institute, 10010 North addition, within the first half of the 20th century, psychol-
Torrey Pines Road, La Jolla, California 92037, USA; ogy underwent a remarkable increase in explanatory
e-mail: [email protected]
† Howard Hughes Medical Institute and Center for Neurobiology and power, evolving from a philosophical tradition based on
Behavior, College of Physicians and Surgeons of Columbia University, introspection to an independent scientific discipline cen-
722 West 168th Street, New York, New York 10032, USA; tered on psychophysics and behaviorism. As psychology
e-mail: [email protected] matured in the first half of the 20th century, it increasing-
‡ Sackler Institute, Department of Psychiatry, Box 140, Weill Medical
ly advocated an empirical behaviorist approach that ended
College of Cornell University, 1300 York Avenue, New York,
New York 10021, USA; e-mail: [email protected] up reducing the focus of interest in psychology to observ-
able aspects of behavior. Psychologists argued that
Current Opinion in Neurobiology 2000, 10:612–624 attempts to quantify non-observable mental events, such
0959-4388/00/$ — see front matter as perception, imagery, thinking, retention, problem solv-
© 2000 Elsevier Science Ltd. All rights reserved. ing, or consciousness were at best speculative and
Abbreviations
unstable. This emphasis on observable indices of behav-
5-HT 5-hydroxytryptamine ior resulted in making psychology a rigorously
CaMKII Ca2+/calmodulin-dependent protein kinase II experimental discipline but one that was so narrowly
CRE cAMP-responsive element focused that it excluded from the study of behavior most
CREB CRE binding protein
of the really fascinating features of mental life. The re-
CS conditioned stimulus
E-LTP early-stage LTP awakening of interest in internal events led to the
ERP event-related potential emergence of modern cognitive psychology by making us
fMRI functional magnetic resonance imaging realize that our knowledge of the world is based on per-
I-LTP intermediate form of LTP ception, and that perception is an act of construction that
IT inferior temporal
L-LTP late-phase LTP depends not only on the information inherent in the stim-
LTD long-term depression ulus but also on the mental structure of the perceiver.
LTP long-term potentiation Ulric Neisser laid out the task of cognitive psychology. In
MST medial superior temporal his classic monograph of 1967 [1], he wrote:
MT middle temporal
NMDA N-methyl-D-aspartate
NR1 NMDA receptor type 1 subunit “…the world of experience is produced by the man
PET positron emission tomography who experiences it…There certainly is a real world
PKA cAMP-dependent protein kinase of trees and people and cars and even books, and it
PP1 protein phosphatase 1
rtTA reverse tTA
has a great deal to do with our experience of these
tetO tet operator objects. However, we have no direct immediate
tTA tetracycline-responsive transcription factor access to the world, nor to any of its properties…
US unconditioned stimulus Whatever we know about reality has been mediated
not only by the organs of sense but by complex sys-
Introduction tems which interpret and reinterpret sensory
The last decade of the 20th century, the Decade of the information.”
Brain, has also been the Decade of Cognitive “. . . the term ‘cognition’ refers to all the processes
Neuroscience. It has been the decade in which the merg- by which the sensory input is transformed, reduced,
er of cognitive psychology and neural science has begun elaborated, stored, recovered and used . . .”
to realize its promise. The joining of neural science and
cognitive psychology is the most recent in a series of sci- Neisser, and the other earlier pioneers in cognitive psy-
entific unifications that have brought together the chology, pointed out that to study mental operations,
 Cognitive neuroscience Albright, Kandel and Posner 613

cognitive psychologists needed to focus on the flow of sen- different sites within the brain. We briefly consider the
sory information, from its transduction by appropriate systems problem before focusing on the molecular prob-
sensory receptors to its eventual use in memory and action. lems of storage (for a current review, see [5]).
This implied that each perceptual or motor act has an
internal representation in the brain, evidenced by a pattern The systems problem of memory
of activity in a specific set of interconnected cells. The pat- Procedural (implicit) memory: there are multiple forms
tern of connections also stores information, in memory, It is now generally accepted that declarative (explicit)
about the perception and the motor act. memory requires for storage the medial temporal lobe and
the hippocampus, whereas procedural (implicit) memory
In the 1970s and 1980s, the early years of cognitive psychol- does not. One of the advances of the past decade has been
ogy, most types of internal representations were not really the further documentation that procedural memory has
accessible for a meaningful experimental analysis (for inter- many subcomponents, and that essentially all cognitive
esting exceptions, see [2–4]). Fortunately, the emergence of systems — be they perceptual or motor — involving a
cognitive psychology was soon followed by, and indeed large variety of different neural systems may be capable of
helped stimulate, significant progress in systems neural sci- storing some types of procedural memory. For example,
ence and in brain imaging. By the 1990s, these approaches various types of motor learning involve the cerebellum;
had come together, making it possible, for the first time, to learned fear the amygdala; operant conditioning and habit
study directly internal sensory and motor representations in learning the basal ganglia; priming the neocortex; and per-
both nonhuman primates and human beings. A major goal of ceptual learning, including habituation and sensitization,
this new direction was to map elementary cognitive func- the primary sensory pathways. This distribution of proce-
tions onto specific neuronal systems. dural memory across neural systems was already well
appreciated in the 1980s, and, in some cases, such as motor
We cannot, in this brief review, document the substantial learning involving the cerebellum, the circuitry involved in
progress that has been made in mapping internal represen- learning was already well studied (see reviews in [6–10]).
tations of cognitive functions. We therefore will restrict our Here, we illustrate how our knowledge of implicit memory
review to three areas in which we have direct experience: mechanisms has expanded in the 1990s by focusing briefly
memory storage, perception, and higher mental functions. on just two examples, fear conditioning in the amygdala
Our purpose in considering memory storage is to illustrate and motor learning in the cerebellum.
that some elementary aspects of cognitive processes, in
particular the switch from short- to long-term memory, can The amygdala and emotional memory
now be studied at the molecular level. In discussing per- Considerable evidence from both humans and experimen-
ception, we illustrate the influence that psychophysics, tal animals now indicates that the amygdala is critical for
especially psychophysical studies in intact, awake behav- the expression of emotion and intervenes between the
ing primates, has exerted on the study of the neuronal hypothalamus and brainstem nuclei — the regions con-
organization and function of the visual system. Finally, we cerned with the somatic expression of emotion — and the
illustrate the remarkable power that we have gained in the cingulate, parahippocampal, and prefrontal cortices — the
study of complex cognitive processes in humans from the neocortical areas concerned with conscious feeling. For
ability to image the living, behaving human brain. example, electrical stimulation of the amygdala in humans
produces feelings of fear and apprehension, whereas dam-
Memory storage age to the amygdala in experimental animals produces
During the past decade, the study of memory storage has tameness. Consistent with this idea, studies using positron
been characterized by the attempt to map the molecular emission tomography (PET) and functional magnetic res-
biology of synaptic plasticity onto ideas about memory onance imaging (fMRI) clearly show that recognition of
systems that emerged from cognitive psychology in the emotional expression in faces involves the amygdala [11].
1980s. The earlier, cognitive psychological studies of
patients with brain lesions made it apparent that memory As early as 1956, Lawrence Weiskrantz [12] had demon-
was not a unitary faculty of mind but had at least two strated that lesions of the amygdala in experimental
major forms: a declarative (explicit) form concerned with animals also impair learned fear. In subsequent work,
the knowledge of what something was about — a knowl- Michael Davies [13], Michael Fanselow [14], Joseph
edge about facts and events — and a procedural (implicit) LeDoux [15] and their colleagues have delineated the out-
form concerned with the knowledge of how to do some- lines of a neural circuit for learned fear. They found that
thing — a knowledge about perceptual and motor one subnucleus of the amygdala, the lateral nucleus, is
procedures. In studying these two independent memory required for procedural memory of fear conditioning to a
processes, it has proven convenient to divide the study of neutral tone [15]. Information about the CS, the tone, is
memory storage into two parts: the systems problem of carried to the lateral nucleus via two pathways: the thalamo-
memory, which is concerned with where in the brain amygdala projection from the auditory thalamus and the
memories are stored, and the molecular problem of mem- cortico-amygdala projection from the auditory cortex.
ory, which is concerned with how memories are stored in Long-term potentiation (LTP) has been observed in vivo
614 A decade of Current neurobiology

in the projection from the thalamus to the amygdala after the cerebellar cortex and the deep nuclei, with each playing
tetanization of the thalamic input [16]. Importantly, synap- a different role. The cerebellar cortex has the special func-
tic change resembling LTP has been observed in this tion of learning the timing of movement. In so doing, the
pathway after naturally occurring fear conditioning [17]. cerebellar cortex guides the learning in the deep nuclei,
which may allow learning to be transferred from the cortex
Because of the anatomical complexity of the amygdala to the deep nuclei. Since one of the issues in declarative
(especially when compared with more ordered structures learning is the transfer of information from the hippocampus
such as the hippocampus and the cerebellum), precise and the medial temporal lobe to other areas of neocortex,
molecular characterization of the plasticity at these defined this insight may prove to be of general significance (see dis-
synapses has not yet been achieved. However, recent find- cussion of Declarative memory storage below).
ings have begun to make some progress. Thus, blockade of
noradrenaline receptors has been found to interfere with Recent studies indicate that the cerebellum probably does
formation of emotional memory in humans, suggesting even more than participate in motor learning; it also par-
that the cAMP pathway in the amygdala may be required ticipates in associative word learning [29], learning of
[18]. Disruption of cAMP-dependent kinase (PKA) in fact tactile mazes [30], and perception of time intervals
disrupts fear conditioning [19,20]. More recently, LTP has between successive stimuli [31].
been described in slices of the amygdala at both the
synapse from the thalamus and the cortex to the lateral Declarative (explicit) memory storage: the medial temporal
amygdala, and, like LTP in the hippocampus, it has been system appears to have only a temporary role
found to have an early phase and a protein-synthesis- A key feature of declarative memory is that the medial tem-
dependent late phase [21]. The induction of the early poral lobe is involved in memory for a limited period of
phase is postsynaptic, but the expression is presynaptic time. The initial evidence came from studies of the patient
and also requires PKA [22,23]. H.M., as well as the two other patients described by
Penfield and Milner [32] who had good memory for remote
Memory for motor acts and the cerebellum events that occurred years before their surgery. Recently,
Well into the 1970s and 1980s, many neural scientists still studies in experimental animals have tested this idea rigor-
thought that the function of the cerebellum was restricted ously and obtained clear evidence for temporally graded
to coordinating voluntary movements: gait, posture, retrograde amnesia, which covered a period ranging from a
speech, and other skilled movements. This view began to few days to about a month before surgery (reviewed in [5]).
change in the early 1980s, when a series of studies by
Richard Thompson, Masao Ito, Mitchell Glickstein, Tom These data have given rise to the idea that the medial tem-
Thach and their colleagues provided a variety of indepen- poral lobe structures direct a gradual process of
dent evidence that the cerebellum is also critical for the reorganization and stabilization by changing the organiza-
formation of learned motor responses ([6–9]; for early tion of cortical representations, perhaps by binding together
theoretical discussions, see also [24,25]). For example, the separate cortical regions that store memory for a whole
Thompson and co-workers [6,26,27] found that lesions of event [33,34]. After sufficient time has passed, the hip-
the cerebellar cortex produced deficits in conditioned eye- pocampal formation may not be needed to support storage
blink responses while sparing the unconditioned or retrieval of declarative memory, and long-term memory
responses. The site for this plasticity appears to reside in is fully dependent on the neocortex (reviewed in [35]).
the mossy-fiber–parallel-fiber Purkinje cell synapse [27].
Indeed, conditioned responses can be obtained by simply The molecular mechanisms of memory storage
substituting for the CS direct electrical stimulation of A role for CREB in procedural (implicit) and declarative
brainstem (pontine) nuclei and pairing that with electrical (explicit) memory
stimulation of fibers from the inferior olive as the US [27]. Although declarative and nondeclarative memory use dif-
ferent neural systems and different logic, the two
The work on eyeblink conditioning is part of a larger tra- memory systems have in common distinct temporal phas-
dition of work on the role of the cerebellum in learning, a es. There is a short-term memory lasting minutes to
tradition that includes the studies of Masao Ito [9] on the hours, and a long-term memory lasting days or even
modification of the vestibular–ocular reflex and the studies longer. In both memory systems, long-term memory dif-
of Thomas Thach [28] on the role of mossy and climbing fers from short-term memory not only in time course but
fibers in monkeys learning to adjust their wrist movements also in molecular mechanisms. Long-term memory, but
to unexpected changes in load. not short-term memory, requires the synthesis of new
protein (reviewed in [36,37]). Studies during the past
A comparison of learning in the vestibular–ocular reflex to decade in Drosophila, Aplysia, and mice suggest that
learning of classical eyeblink conditioning by Raymond et al. cAMP, PKA, and the cAMP-responsive transcription fac-
[10] suggests that these two quite different behaviors show tor CREB are critically involved in the conversion of
a surprisingly conserved role for the cerebellum in motor short-term to long-term memory for both declarative and
learning. In each case, the plasticity is distributed between nondeclarative memory.
 Cognitive neuroscience Albright, Kandel and Posner 615

Indeed, in both Drosophila and Aplysia, CREB learning- protein synthesis and the kinase PKA [46]. Moreover,
related transcriptional activation appears to be the there is increasing evidence for rapid receptor insertion
rate-limiting step in the conversion of short-term to long- and remodeling during E-LTP and for actual structural
term synaptic plasticity and behavioral memory. This was changes during L-LTP (reviewed in [47]).
first shown for synaptic plasticity by injecting into the
nucleus of the sensory neuron of Aplysia an oligonucleotide Development of gene targeting by homologous recombina-
with a cAMP-responsive element (CRE) [38]. This tion in embryonic stem cells has made it possible to test
oligonucleotide selectively blocked long-term facilitation some of these ideas genetically. The initial studies exam-
without affecting short-term facilitation. A similar result ined mice with targeted knockout of CaMKII and the
was later obtained by injecting antibodies specific for tyrosine kinase fyn, kinases that had previously been impli-
CREB1 [39]. Conversely, injecting a phosphorylated form cated in LTP in pharmacological studies [44,48]. The
of the recombinant CREB1a protein, a key regulator of genetic studies revealed that mice lacking CaMKIIα dis-
transcription, was sufficient to induce long-term facilita- played a partial loss of E-LTP in the CA1 neurons of the
tion [40]. Injection of antibodies against Aplysia CREB2 hippocampus and severe impairment on spatial memory
(ApCREB2), a negative regulator that inhibits CREB1- tasks [49–51]. Similarly, analysis of mice with targeted dele-
mediated transcription, led to long-term facilitation tions of the tyrosine kinase fyn also showed deficits in
accompanied by a growth of new synaptic connections [40] E-LTP as well as in spatial memory. Mice with deletions of
when paired with a single pulse of 5-HT (which normally the non-receptor tyrosine kinases src and yes were normal
produces short-term facilitation). Thus, ApCREB2 acts as [52]. By contrast, expression of a dominant-negative
a repressor of long-term facilitation by functionally com- inhibitor of PKA in neurons of the forebrain using the
peting with the CREB1a activator. In parallel experiments CaMKIIα promoter [53] resulted in mice that had a normal
in Drosophila, the behavioral switch from short- to long- E−LTP but a dramatically attenuated L-LTP. These genet-
term memory was found to be regulated by changing the ically modified mice learned a contextual task as well as
activity ratio between CRE binding activator dCREB2a wild-type animals, had a perfectly good short-term memory
and repressor dCREB2b proteins, the fly homologues of when tested 1 hour after training, but were impaired in
mammalian CREB and Aplysia CREB1 [36,41]. For exam- selective long-term memory when tested 24 hours later.
ple, when the expression of the CREB activator is induced Thus, as in Aplysia and Drosophila, the PKA signaling path-
before training, a single odor–shock pairing produces a way in mice seems to be important for maintaining both
long-term memory for the odor. LTP and memory for prolonged periods of time.

These results in Drosophila suggest that this CREB-medi- Previously, Roussoudan Bourtchouladze et al. [54] had
ated induction of transcription is necessary to produce the examined the role of CREB in the mouse. They found that
long-lasting changes in synaptic strength required for the a deletion that eliminated only the alpha and delta isoforms
long-term storage of memories [36,37]. The results in led to impairment in LTP and in long-term memory storage
Aplysia suggest that the CREB switch functions at the level [54]. The memory deficit was similar to that seen in mice
of the individual synapse to convert a short-lasting increase expressing the PKA inhibitor; initial learning and short-
in synaptic strength produced by covalent modifications of term memory were intact, whereas long-term memory was
existing proteins to one that is long-lasting and produced by impaired. Moreover, work by Daniel Storm and colleagues
the synthesis of new proteins (see also [42,43]). [55,56] has provided further strong evidence that CREB is
indeed involved in mouse synaptic plasticity. They pro-
Reverse genetics in the mouse duced a transgenic mouse in which a lacZ reporter gene is
What about the switch to long-term declarative memory? activated by a CREB-responsive promoter, and they found
In the past decade, methods for modifying individual that this reporter is activated both by L-LTP in vitro [55]
genes in mice have become available and have helped and by certain forms of hippocampus-dependent learning
define the mouse as the mammalian model system par in vivo [56]. This demonstrates that CREB or CREB-like
excellence for the genetic study of declarative memory stor- transcription factors are in fact activated under circum-
age. Mice exhibit a memory for space and objects that stances that lead to plasticity and suggests a causal role
corresponds to human declarative memory, requiring the similar to that seen in Aplysia and Drosophila.
hippocampus and the medial temporal lobe. Moreover,
the hippocampus has a form of synaptic plasticity, called Regionally and temporally restricted gene expression
LTP, thought to be a candidate mechanism for this sort of The traditional genetically modified mouse lines men-
memory storage. Pharmacological experiments in the tioned above have identified some genes that are necessary
1980s and early 1990s first indicated that LTP itself has to develop a normal learning and memory phenotype in the
stages, much like long-term facilitation in Aplysia. There adult, and suggest genes that may vary in their allelic form
is an early-stage LTP (E-LTP) that requires covalent in a normal population to give rise to the normal variation in
modification mediated by Ca2+/calmodulin-dependent cognitive ability. These lines of mice, however, suffer from
protein kinase II α (CaMKIIα) and the tyrosine kinase the limitation that many molecules likely to be important
fyn [44,45], and a late-phase LTP (L-LTP) that requires for adult learning are also likely to be important for normal
616 A decade of Current neurobiology

development. In addition, these gene manipulations often Overexpression of calcineurin leads to an impairment in an
affect a variety of different brain regions. One example of intermediate form of LTP (I−LTP) in hippocampus and to
how to circumvent this problem is provided by work on a defect in spatial memory in the Morris water maze
regional and regulated gene expression in the brain [57–59]. [63,64,66]. Since the transgene can readily be switched on
and off by giving or removing doxycycline, it has the great
Mice with a targeted deletion in the NR1 subunit of the advantage that it can be used to study not only memory
NMDA receptor die shortly after birth [60]. Previously, Joe storage but also memory retrieval. Mutant mice that
Tsien and colleagues [57,58] succeeded in obtaining condi- express the calcineurin transgene transiently, after learning
tional deletion restricted to forebrain neurons by using the has been acquired and after spatial memory has already
CRE-loxP systems. By combining the forebrain-specific been stored, have an apparent defect in the retrieval of the
CaMKIIα promoter with the bacterial CRE-loxP recombi- spatial information. This retrieval defect is not attributable
nase system, they were able to knock out the NR1 gene to a disruption in memory storage because it could be
specifically in the CA1 region of the postnatal hippocampus, reversed when the transgene expression was turned off by
without affecting other structures. This restricted knockout stopping doxycycline administration. Thus, with the use of
avoided the perinatal lethality of a complete NR1 knockout. regulated genetic modification one can not only control for
As predicted, the mice showed a deficit in CA1 LTP and a potential developmental abnormalities associated with a
severe deficit in spatial learning, strongly supporting a role genetic change but also begin to explore the various
for NMDA-dependent LTP in hippocampus-dependent phases of memory acquisition, storage and retrieval [64].
learning. More recently, Tsien and colleagues [61] have car-
ried out the reverse experiment. They expressed a mutant Future directions
form of NR1 that allows greater Ca2+ influx, and found that The use of mouse genetics to investigate complex behav-
it enhances LTP and memory storage. ioral traits such as learning and memory is at an early stage,
but it promises to extend cognitive neuroscience into a
A further advance has been the ability to regulate gene new, molecular genetic direction. Moreover, the work in
expression not just regionally but also temporally, using both invertebrates and mice suggests that many of the
the tetracycline system (see e.g. [62]). In a group of exper- basic molecular mechanisms for memory may be con-
iments, a cell-type-specific promoter, such as the served across species, allowing insights from invertebrates
CaMKIIα promoter, is used to drive expression of the to be applied to the mammalian brain. In the fly, for exam-
tetracycline-responsive transcription factor tTA in fore- ple, the characterization of a new learning and memory
brain neurons of one line of mice. In a second line, the mutants will hopefully provide a fuller understanding of
transgene to be regulated is linked to a promoter consist- the critical genes involved in memory storage. However,
ing of multiple repeats of the tet operator (tetO) linked to since the developmental integrity of many brain regions is
a minimal eukaryotic promoter element. When these two necessary for the proper performance of even the simplest
lines are mated so that both transgenes are introduced into memory task, and since memory storage requires some of
a single mouse, the tetO-linked gene is activated specifi- the most basic cellular signaling mechanisms, there will be
cally in those cells that express tTA. The expression of the many developmentally important genes that affect learn-
tetO-linked transgene can then be suppressed by oral ing and memory. The difficult task facing the field,
administration of the tetracycline analogue doxycycline therefore, will be to distinguish those mutations that affect
[59]. In another class of experiments, a mutant form of the the core cellular signaling mechanisms that are used to
tetracycline repressor is used that induces transcription encode memories from those mechanisms that modulate
only in the presence of doxycycline [62]. This reverse tTA, these mechanisms or affect the development of the basic
or rtTA, has been used to obtain inducible and reversible circuits important for performing the learning task or for
expression of a Ca2+-activated protein phosphatase cal- storing the learned information. The use of progressively
cineurin in forebrain neurons [63,64]. Calcineurin opposes better anatomically restricted and temporally regulated
the action of PKA and other protein kinases. Calcineurin genetic modification in the mouse will be critical for dis-
has a high affinity for Ca2+, even higher than that of tinguishing those genes that directly affect memory
CaMKIIα. At low-frequency stimulation, the amount of encoding from those that affect memory indirectly through
Ca2+ coming into the cell through the NMDA receptor is developmental, motivational, or perceptual mechanisms.
small and activates calcineurin but not CaMKIIα.
Calcineurin, in turn, can dephosphorylate protein phos- Vision
phatase Inhibitor-1, which activates protein phosphatase 1 In his pioneering text, which first appeared 50 years ago,
(PP1) and leads to long-term depression (LTD) of synaptic Donald Hebb [67] observed that “we know virtually noth-
transmission, which is the mirror opposite of LTP [65]. In ing about what goes on between the arrival of an excitation
contrast, higher frequencies of stimulation lead to greater at a sensory projection area and its later departure from the
Ca2+ influx. This, in turn, activates kinases, including motor area of the cortex…” “Something like thinking inter-
PKA, that phosphorylate and block Inhibitor-1, thereby venes,” and although it would be hard to disagree with that
shutting off the phosphatase cascade. In fact, PKA and cal- proposition, the goal of cognitive neuroscience has been to
cineurin target the same residue on Inhibitor-1. flesh out that ‘something’ in a form that is more satisfying
 Cognitive neuroscience Albright, Kandel and Posner 617

to both psychologists and neurobiologists alike. In part In a set of experiments conducted by Thomas Albright,
because its operations span the chasm that Hebb lament- Gene Stoner, and colleagues [71–73], contextual cues
ed, the visual system has served as a proving ground for unrelated to visual motion (e.g. luminance or stereoscopic
this goal. By tracing the flow of visual information from cues that elicit a percept of surface depth ordering) were
retina to motor control circuits we can, in principle, deter- found to markedly alter perceived motion, even though
mine how its representation by the brain contributes to the retinal stimulus motion remained unchanged. In addition,
various cognitive processes that constitute thinking, such the responses of many motion-sensitive neurons in cortical
as perception, recognition, imagery, decision making, and visual area MT were found to co-vary with perceived
motor planning. motion rather than with retinal motion, demonstrating that
the formation of perceptual or ‘scene-based’ neuronal rep-
The 1990s will long be remembered as a turning point in resentations is accomplished at early stages in the visual
this effort. As often recorded in these pages, recent processing hierarchy [73–75].
advances in cognitive neuroscience are many. This state of
affairs owes much to the fact that neurobiologists have Related evidence for scene-based representations has
with increasing frequency turned to experimental psychol- come from experiments in which perceived features are
ogy (and vice versa) for guidance, inspiration, and tools. not physically present in the retinal stimulus, but their
Products of this new-found interdisciplinary success presence in the visual scene is implied by contextual cues.
include discoveries regarding the correspondence between In one such study, Rüdiger von der Heydt and colleagues
neuronal and perceptual events, the role of context in per- [76] studied the neuronal basis of a perceptual phenome-
ceptual processing, the neuronal substrates of attention non known as ‘illusory contours’, in which contextual cues
and decision making, the plasticity of adult sensory repre- imply the presence of an occluding surface and the edges
sentations, and the role of such in perceptual learning and of the surface are seen, even though they are not actually
sensory–motor spatial coordinate transformations. present in the stimulus. These investigators discovered
that many neurons fire in a way that matches the percept,
Linking neuronal and perceptual events as though a real contour had been placed in the receptive
We entered the past decade secure in the knowledge that field. In another study, John Assad and John Maunsell [77]
the stimulus selectivities of visual neurons resemble the exploited the fact that observers generally infer the con-
basic elements of perceptual experience, such as simple tinuous motion of an object when it moves behind an
forms, motions, and colors. Lacking, however, was evidence occluder. By using remote contextual cues to place the
for a specific causal relationship between neuronal and per- path of occluded motion within the receptive fields of
ceptual events. The problem was remedied by William motion-sensitive MT neurons, these investigators found
Newsome and colleagues, who united classic methods from that many neurons respond in a way that matches the per-
experimental psychology with modern neurobiological ceptual inference of motion, in the absence of any real
techniques (for a review, see [68]). These investigators dis- motion in the receptive field. The significance of these
covered a close relationship between perceptual motion findings lies, of course, in the fact that the neuronal rep-
sensitivity and the sensitivity of neurons in cortical visual resentations reflect — as does perception — the enduring
area MT [69], which strongly suggested that neuronal activ- structural and relational qualities of the observer’s
ity within area MT is a constituent of the perceptual external environment [78].
experience of motion. Icing came in the form of another
study in which Newsome and colleagues modified the per- Visual attention
ceptual experience of motion by artificially activating small The primate visual system has a limited information pro-
collections of MT neurons [70]. The results from these cessing capacity. An exciting area of research in the 1990s
experiments have offered sound reassurance that the stim- has been that addressing the means and conditions under
ulus selectivities of visual neurons account for perceptual which this limited capacity — visual attention — is
experience, and they have paved the way for an under- dynamically allocated. Work in this area has revealed two
standing of the underlying mechanisms. basic types of attentional phenomena, which may have
distinct neuronal substrates. One effect, known as ‘atten-
Sensation versus perception tional facilitation’, is the improved processing of a
The Newsome experiments solved one puzzle, but as stimulus when it appears at an attended location. Early
many neurobiologists of the 1990s gained sophistication in investigations of the effects of focal brain lesions in
perceptual psychology, they were forced to confront another: humans implicated the parietal lobe in attentional facili-
as understood up to that point, the stimulus selectivities of tation. In subsequent physiological studies of parietal
visual neurons encoded properties of the retinal stimulus. cortex in non-human primates, Michael Goldberg and
But things perceived reflect the ‘meaning’ of the stimulus, colleagues [79] found that for many neurons an attended
as defined by the content of the visual scene that leads to visual stimulus elicited a much larger sensory response
its appearance. Several studies carried out during the past than did an identical unattended stimulus. Similar facili-
decade have sought to distinguish neuronal representa- tatory effects have since been reported for other cortical
tions of sensory and perceptual events. visual areas [80,81].
618 A decade of Current neurobiology

The other basic attentional effect that has been studied search for neuronal responses that ‘predict’ an impending
extensively is known as ‘attentional selection’. This effect motor response to a visual cue (see e.g. [89]). In a recent
refers to the phenomenon in which a target stimulus (i.e. study, Michael Shadlen and colleagues [90] exploited the
the thing you’re looking for) is selected from among other fact that difficult decisions generally require time to accu-
stimuli that are competing for attention. In the mid-1980s, mulate relevant information, hence predictive neurons
Robert Desimone and colleagues [82] found that receptive should exhibit responses that increase in magnitude in par-
field profiles of individual neurons in cortical areas V4 and allel with the observer’s decision confidence. These
IT contract around the attended stimulus, excluding unat- investigators found that neurons in prefrontal cortex do
tended stimuli. These findings of selection at the neuronal exactly that. Collectively, these novel studies of the deci-
level imply that information about an attended stimulus is sion process have identified relevant neuronal substrates
carried to higher processing stages, at the expense of infor- and have led to promising theories regarding ways in which
mation about unattended stimuli. Selective effects have visual information is flexibly mapped to action.
now been reported for many visual areas, including areas
V1, V2, V4, MT, MST, and IT (see e.g. [83–86]), indicat- Perceptual learning
ing that selective mechanisms operate simultaneously on It is a central tenet of neurobiology that the sensory neo-
multiple feature maps. cortex reaches its mature state of organization following a
brief period of postnatal plasticity known as the critical
Physiological studies are beginning to target the underly- period. One of the most important discoveries of the past
ing mechanisms of selective attention. Much of the work decade, however, is the large extent to which this plastici-
in this area has been inspired by the ‘biased competition’ ty continues throughout life. Adult plasticity enables forms
model advocated by Desimone and Duncan [87]. of cognitive flexibility such as perceptual learning, which
According to this view, pieces of incoming sensory infor- is the improvement with practice in the ability to discrim-
mation compete for neuronal representation (and, inate sensory attributes. Early hints of this plastic potential
ultimately, control of the observer’s actions), and the com- came from studies demonstrating that the adult cortex
petition is biased such that behaviorally relevant inputs are undergoes a local functional reorganization to compensate
facilitated. Support for this model comes from a recent for damage to the sensory periphery [91–93]. This com-
study by Desimone and colleagues [86], who found that pensation may be mediated by intrinsic cortical
the neuronal response to two unattended stimuli placed connections, which appear to undergo rapid changes in
together in the receptive field was approximately the aver- synaptic efficacy, as well as a slower process of sprouting
age of the responses elicited by the two stimuli presented and synaptogenesis [94,95]. The belief that this form of
independently. By contrast, when either stimulus was plasticity also underlies adult perceptual learning is sup-
attended — thus giving it a competitive advantage — the ported by evidence that training on perceptual tasks leads
neuronal response approximated that elicited by the to reorganization of cortical sensory maps [96] and
attended stimulus alone. Although these findings beg improvements in the sensitivities of cortical neurons [97].
important and difficult questions regarding the source of
the biasing signal, as well as the local circuit and synaptic Transforming signals from visual space to motor space
interactions that lead to changes in the receptive field pro- Another success story from cognitive neuroscience in the
file, they provide a striking example of the gains afforded 1990s centers on the problem of converting visual inputs to
by the convergence of psychology and neurobiology. signals that can guide actions. One long-standing view
holds that retinal signals, which represent visual space in a
Perceptual decisions coordinate frame that shifts with every movement of the
While intelligent behavior depends upon knowledge of eyes, are re-mapped into a more generic and stable coordi-
one’s external environment (i.e. perception), it also nate frame based, for example, on the positions of objects
requires deciding which actions are appropriate given that relative to the observer’s head or body. Consistent with this
knowledge. This ‘decision process’ has been the focus of view, Richard Andersen and colleagues [98] discovered that
several revealing experiments over the past decade. One the magnitude of response to a visual stimulus varies with
goal has been to identify neuronal activity that is correlat- the angle of gaze. Because they take eye position into
ed with the decision to execute a particular action in account, these ‘gain field’ responses yield a head-centered
response to a particular sensory stimulus, rather than sim- map of visual space that is distributed across a population of
ply correlated with either stimulus or action alone. In a parietal neurons [99]. More recent studies have obtained
series of studies, Earl Miller and colleagues [88] found that intriguing evidence for explicit cellular representations of
the responses of neurons in prefrontal cortex — an area space in pre-motor cortex. Carl Olson and Sonya Gettner
long believed to play a role in the organization of complex [100], for example, found that some premotor neurons rep-
behavior — change as new sensory–motor relationships are resent spatial location relative to the parts of a visible object
learned, such that individual neurons come to represent (i.e. ‘object-based’ coordinates), independent of the posi-
new behaviorally relevant conjunctions of stimulus and tion of the object’s image on the retina. Perhaps even more
action. Other groups have adopted a different approach to remarkable are the findings of Michael Graziano, Charles
the decision process, the principal feature of which is a Gross and colleagues [101], who reported the existence of
 Cognitive neuroscience Albright, Kandel and Posner 619

premotor neurons that represent the position of a visual their organization into circuits, their pathologies and
stimulus relative to the position of the observer’s forearm. change with experience.
Body-part-centered representations of this sort appear well
suited for orchestrating specific limb movements to stimuli Imaging has strengthened the correspondences
that are near those body parts. between the brain anatomy of humans and that of
experimental animals
The binding problem A major advance in making cross-species comparisons has
In reviewing major themes of cognitive neuroscience been the development of flat maps that provide a two-
research in the 1990s, we would be remiss to exclude a dimensional surface for mapping the complex folds of the
topic that has captured enormous attention in disciplines human brain [106]. There has been progress in efforts to
ranging from visual physiology to philosophy, while relate retinotopic human visual areas to the maps
nonetheless remaining one of the most unsettled. The obtained from cellular recording in primates [107].
topic in question is the use of temporal binding codes to Evidence that attention enhances activity in V1 (for a
represent complex conjunctions of information carried by review, see [108]), has made it possible to explore atten-
individual neurons. The potential utility of such codes has tional influences in a brain area where the detailed
long been recognized. Empirical support came in the form cellular structure is better understood. Cellular studies in
of physiological data from Charles Gray, Wolf Singer and monkeys indicate the importance of attention in integrat-
colleagues [102], which suggested that visual features (e.g. ing visual effects that occur outside the classic receptive
edges) perceived as parts of the same object are represent- field of V1 neurons, for example, in perceiving contours
ed by neurons that fire synchronously. Others, however, [109]. The interaction of attention with V1 circuitry may
have failed to find such support (or have raised objections also be important in understanding the early visual sys-
on theoretical grounds), and the unprecedented polariza- tem plasticity described in the previous section.
tion of opinions on the subject remains palpable as the
decade draws to a close (e.g. see reviews in the October Integration of human and animal studies is not limited to
1999 issue of Neuron). In light of the importance of the the early visual system. Studies of parietal neurons, locat-
binding problem for both a functional and mechanistic ed within brain areas shown to be active in imaging studies
understanding of cognition, and the attention and of spatial attention [110,111], have provided evidence of a
resources that have been directed at the problem in recent map of locations organized by their current importance
years, this persistent lack of consensus is both surprising [112] that could serve as the basis for human working
and greatly disappointing. One can only hope that resolu- memory for location [113]. Using new fMRI methods, it
tion will come from the application of new concepts and has been possible to separate operations performed by the
techniques in the next decade. superior and inferior areas of the parietal lobe. These
results [111] suggest that the temporal–parietal junction is
Imaging higher cognitive functions critical for shifts of attention toward unexpected visual
Imaging has been critical for the localization of mental stimuli, a finding that supports the importance of this area
processes in the neglect of space opposite the lesioned hemisphere
As one reads journals devoted to cognitive neuroscience or found in patients suffering from strokes that affect the
human brain mapping, it is hard to imagine that there temporal–parietal junction.
could ever have been doubts that there was a specific
anatomy related to higher mental processes. Only in the Timing and the functional connectivity of neural circuits
past decade have neuroimaging studies using PET and Of equal importance to precision in space is the ability to
fMRI demonstrated a pervasive form of localization in a say exactly when and for how long an anatomical area is
wide variety of cognitive and emotional tasks [103,104]. active and when information is being exchanged between
The localization of mental operations, as described in the areas. Because mental operations occur in the range of tens
first section of this review, has made psychology a full to hundreds of milliseconds, it has been useful to relate
partner in efforts to understand human brain mechanisms. areas of activity found in imaging studies to the distribu-
tion of electrical activity recorded from the scalp [114] or
The distributed nature of the activations in any real cog- by depth electrodes [115]. For example, using combined
nitive task helps explain why Karl Lashley [105] and ERP and fMRI methods, it has been shown that although
others could have thought that the brain operated as a attention influences activity in primary visual cortex, this
whole. However, in tasks involving language, mental takes place only after the information has been processed
imagery, spatial navigation and working memory, where in prestriate areas [116].
we have been able to dissect them into plausible compu-
tations, it is these components not the task themselves A different approach to the circuitry of high-level cognition
that are localized. Of course localization is only a start is to examine functional connectivity by studying the cor-
toward the achievement of cognitive neuroscience. As relation between brain areas on the basis of hemodynamic
described below, imaging has provided an important [117] or electro-magnetic [118] measurements. These
impetus for exploring the evolution of mental operations, efforts provide an approach to the transfer of information
620 A decade of Current neurobiology

between brain areas by specifying within a task exactly the Neuroimaging studies have provided us with some mech-
time when their activity is correlated. anisms by which experience at different time scales might
change anatomy or circuitry on a temporary or permanent
Recently, an adaptation of magnetic imaging called diffu- basis. One way in which brain circuitry can be altered is
sion tensor imaging has been used to produce images of called priming, which refers to changes in the efficiency of
the white matter connecting brain areas [119,120]. The processing a target when part or all of the pathway involved
time course of myelination of neuronal pathways can be has been previously activated. Combined neuroimaging
used to test theories of when in development particular and cellular studies show that priming works by reducing
behaviors emerge. Perhaps this method, when combined or tuning the number of neurons required to process the
with measuring correlations in electrical activity, will allow target [131]. Studies in human subjects have shown that
us to predict when a given brain circuit is sufficiently priming can take place within a second and that it may
developed to support the learning of complex skills. In a take place even when the person is unaware of the identi-
recent study [121], for example, differences in white mat- ty of the prime [132]. Priming may help account for
ter tracts in the temporal parietal area of the left moment-to-moment differences in the thoughts generated
hemisphere were related to reading skill within both a nor- in response to a given environmental challenge or strate-
mal and a reading-impaired population. gies used to solve a problem.

Pathology: the search for subtle functional changes in A few minutes of practice has been shown to be sufficient to
disease change the circuit that processes information from one that
Brain damage due to strokes and tumors can be observed involves a high level of complex computation to one that has
readily in structural images, but other abnormalities may an already compiled answer waiting for output [123,133].
involve more subtle functional changes. The use of struc- Changes of pathways can help explain the shifts that take
tural images to map brain lesions has allowed investigators place when extensive practice renders a skill automatic.
to combine the data from sets of patients with large cere-
bral lesions showing by the overlap in lesion location the Somewhat slower are the task-related increases in cerebral
brain areas that appear to cause the deficit. For example, a tissue found in sensory [134] and motor [135] systems with
set of patients with large lesions of the left hemisphere all extended practice. These changes in visual areas were dis-
showed expressive aphasia provided that they had a loss of cussed in the previous section. The effects of extensive
neurons in the anterior insula [122]. This finding support- practice have also been documented in studies of adults
ed results obtained with PET showing that the insula is an who learned languages either as children or later in life
important pathway for highly automated verbal output, [136] and musicians who had extensive practice on musical
such as reading words aloud [123]. instruments [137]. Although there are strong hints that the
relative plasticity of these skills may differ between chil-
Tasks involving monitoring of emotion and cognition acti- dren and adults, full documentation of this form of critical
vate separate areas of the frontal midline [124–126]. In period is still to be developed.
some studies, the cognitive and emotional areas appear to
be mutually inhibitory, with cognitive tasks reducing blood The next decade?
flow in areas related to emotion, as well as the reverse Donald Hebb and his fellow pioneers in cognitive neural
[126]. These areas appear to be important in some forms of science would surely be pleased at the promiscuous bed-
psychopathology. For example, PET scans of patients with fellows that psychologists and neurobiologists have now
schizophrenia who have never been on medication indicat- become, and with their offspring, which is a fuller demon-
ed an abnormality in the left globus pallidus [127], which stration that specific cognitive information is represented
is the outflow of a major dopamine circuit that modulates in the activities of specific neuronal populations. These
cellular activity in midline frontal areas such as the anteri- recent successes notwithstanding, it seems likely that
or cingulate gyrus [128]. This abnormal function might Hebb would also recognize the weakness in our current
help to explain both the curious neglect of the right side of view: lack of information about how these neuronal repre-
space shown in early schizophrenia and the changes in cir- sentations are achieved mechanistically. How, for example,
cuitry within the anterior cingulate observed in the brains are contextual cues assessed by cortical neurons to form
of schizophrenic patients postmortem [129]. visual representations that coincide with perceptual expe-
rience? And how are such representations altered as a
Plasticity and shaping of neural circuits by experience function of experience? Answers to these and other mech-
A major achievement of infancy research in the past anistic questions require, at the very least, detailed
decade has been to show that human infants enter the information about the patterns of anatomical connections
world with some mechanisms related to the processing of in the cerebral cortex, and the functional properties con-
language, objects, faces and numbers, as well as with the ferred by specific circuit components.
ability to imitate motor routines (for a review, see [130]).
How does experience shape these initial mechanisms into Determining the local circuit organization of the cerebral
the complex skills present in the adult? cortex and how that organization relates to the processing
 Cognitive neuroscience Albright, Kandel and Posner 621

of region-specific information is, however, dauntingly powerful molecular and genetic tools find their calling in
complex, and its elucidation — important though it may the service of cognitive neuroscience, and that the field will
be — is among the most formidable challenges facing cog- continue to advance through a global circuit-based
nitive neuroscience in the next decade. Hope lies in some approach to cognitive representation by the brain. These
remarkable new experimental approaches, which promise are indeed heady times for the young field of cognitive
both fine-scale assessment of functional circuitry in the neuroscience. Although, as noted by Hebb 50 years ago,
cortex and pictures of the global patterns of neuronal activ- there still is “a long way to go before we can speak of under-
ity associated with specific cognitive states. standing the principles of behavior to the degree that we
understand the principles of chemical reaction”, the time
One of the most exciting prospects for fine-scale analysis for that understanding is now — at least — in full view.
of functional circuitry can be found in methods for region-
ally restricted and temporally regulated control of gene Acknowledgements
expression. These methods took the stage in the 1990s in We thank Marc Tessier-Lavigne for his helpful editing of this paper.
TD Albright and ER Kandel are Investigators of the Howard Hughes
the form of mouse germ line transgenic manipulations, and Medical Institute. The work of MI Posner was supported by a grant from
they have become extremely powerful tools for analysis of the James S McDonnell Foundation to the Sackler Institute.
the cellular and molecular bases of learning and memory.
Can we reasonably expect to use these new molecular tools References
to study neural systems in animals such as non-human pri- 1. Neisser U: Cognitive Psychology. New York: Appleton-Century
mates, animals that possess a rich and human-like Crofts; 1967.
repertoire of cognitive skills, but for which germ line gene 2. Mountcastle VB: Central nervous mechanisms in
transfer is all but impossible? Much cause for optimism mechanoreceptive sensibility. In Handbook of Physiology, Section 1:
The Nervous System, Vol. III: Sensory Processes, Part 2. Edited by
stems from newly developed methods that employ viruses Darian-Smith I. Bethesda, Maryland: American Physiological Society;
as vectors for gene transfer, which promise to eliminate 1984:789-878.
dependence upon germ line transgenics and thus allow 3. Wurtz RH, Goldberg ME, Robinson DL: Brain mechanisms of visual
technology for temporal and cell-type specific gene regula- attention. Sci Am 1982, 246:124-135.
tion to be imported to primates. Imagine, for example, the 4. O’Keefe J, Dostrovsky J: The hippocampus as a spatial map:
incredible possibilities for analysis of functional circuit preliminary evidence from unit activity in the freely moving rat.
Brain Res 1971, 34:171-175.
components afforded by the ability to switch on and off a
5. Milner B, Squire LR, Kandel ER: Cognitive neuroscience and the
specific class of cells in a specific cortical area while a mon- study of memory. Neuron 1998, 3:3445-3468.
key is engaged in a cognitive task! 6. Thompson RF: Neural mechanisms of classical conditioning in
mammals. Philos Trans R Soc Lond B Biol Sci 1990, 329:161-170.
At the other end of the continuum of mechanistic issues, 7. Thach WT, Goodkin HP, Keating JG: Cerebellum and the adaptive
we face the problem of measuring and interpreting the coordination of movement. Annu Rev Neurosci 1992, 15:403-442.
interactions between individual neurons and between 8. Glickstein M, Yeo C: The cerebellum and motor learning. J Cogn
groups of neurons. Work of the past decade attests that the Neurosci 1990, 2:69-80.
single neuron approach to perception and cognition has 9. Ito M: Long-term depression. Annu Rev Neurosci 1989, 12:85-102.
been profoundly successful. As we move beyond the rep- 10. Raymond JL, Lisberger STG, Mauk M: The cerebellum: a neuronal
resentational phenomenology of single neurons to confront learning machine. Science 1996, 272:1126-1131.
mechanistic questions, however, we find that it is impossi- 11. Morris JS, Frith CD, Perrett DI, Rowland D, Young AN, Calder AJ,
ble to proceed without a more global approach to neural Dolan RJ: A differential neural response in the human amygdala to
fearful and happy facial expressions. Nature 1996, 383:812-815.
coding. If, for example, we expect to understand the
12. Weiskrantz L: Behavioral changes associated with ablation of the
mechanisms that underlie the interactive contributions of amygdala complex in monkeys. J Comp Physiol Psychol 1956,
a stimulus in the visual field and of working memory to 49:381-391.
selective visual attention, then it is essential that we be 13. Davis M: The role of the amygdala in fear and anxiety. Annu Rev
able to monitor simultaneously neuronal events in all the Neurosci 1992, 15:353-375.
relevant interconnected brain regions. This is not an easy 14. Fanselow M: Neural organization of the defensive behavior system
task, nor is the associated problem of interpreting the mul- responsible for fear. Psychonomic Bull Rev 1994, 1:429-438.

titude of potential neuronal interactions. New techniques 15. LeDoux JE: Fear in the brain, where have we been and where are
we going? Biol Psychiatry 1998, 44:1229-1238.
based on multielectrode recording, optical imaging of neu-
ronal activity, and fMRI — each used in conjunction with 16. Clugnet M, LeDoux JD: Synaptic plasticity in fear conditioning
circuits: induction of LTP in the lateral nucleus of the amygdala by
the behavioral methods that have proved so valuable over stimulation of the medial geniculate body. J Neurosci 1990,
the past decade — are beginning to make important 10:2818-2824.
inroads in this area. 17. Rogan MT, Staubli UV, LeDoux JE: Fear conditioning induces
associative long-term potentiation in the amygdala. Nature 1997,
390:604-607.
In the decade of the 1990s, cognitive neuroscience thrived
18. Cahill L, Prins B, Weber M, McGaugh JL: Beta-adrenergic activation
by bringing together psychology and neurobiology. We now and memory for emotional events. Nature 1994, 371:702-704.
have every reason to expect that the next decade will yield 19. Bourtchouladze R, Abel T, Berman N, Gordon R, Lapidus K,
a similarly mature molecular biology of cognition, in which Kandel ER: Different training procedures recruit either one or two
622 A decade of Current neurobiology

critical periods for contextual memory consolidation, each of 41. Yin JC, Wallach JS, Del Vecchio M, Wilder EL, Zhou H, Quinn WG,
which requires protein synthesis and PKA. Learn Mem 1998, Tully T: Induction of a dominant negative CREB transgene
5:365-374. specifically blocks long-term memory in Drosophila. Cell 1994,
79:49-58.
20. Schafe GE, Nadel NV, Sullivan GM, Harris A, LeDoux JE: Memory
consolidation for contextual and auditory fear conditioning is 42. Martin KC, Casadio A, Zhu H, Yaping E, Rose J, Bailey CH, Chen M,
dependent on protein synthesis, PKA, and MAP kinase. Learn Kandel ER: Synapse-specific transcription-dependent long-term
Mem 1999, 6:97-110. facilitation of the sensory to motor neuron connection in Aplysia:
a function for local protein synthesis in memory storage. Cell
21. Huang Y-Y, Martin KC, Kandel ER: Both PKA and MAP kinase are 1998, 91:927-938.
required for the macromolecular synthesis-dependent late phase
of LTP in the amygdala. J Neurosci 2000, in press. 43. Casadio A, Martin KC, Giustetto M, Zhu H, Chen M, Bartsch D,
Bailey CH, Kandel ER: A transient neuron-wide form of CREB-
22. Huang Y-Y, Kandel ER: Postsynaptic induction and PKA-dependent mediated long-term facilitation can be stabilized at specific
expression of LTP in the lateral amygdala. Neuron 1998, synapses by local protein synthesis. Cell 1999, 99:221-237.
21:169-178.
44. Nicoll RA, Kauer JHA, Malenka RC: The current excitement in long-
23. Weisskopf MG, Bauer EB, LeDoux JE: L-type voltage-gated calcium term potentiation. Neuron 1988, 1:97-103.
channels mediate NMDA independent associative long-term
potentiation at thalamic input synapses to the amygdala. 45. Bliss TV, Collingridge GL: A synaptic model of memory: long-term
J Neurosci 1999, 19:10512-10519. potentiation in the hippocampus. Nature 1993, 361:31-39.
24. Marr D: A theory of the cerebellar cortex. J Physiol 1969, 202:437-470. 46. Frey U, Huang Y-Y, Kandel ER: Effects of cAMP simulate a late
stage of LTP in hippocampal CA1 neurons. Science 1993,
25. Albus JS: A theory of cerebellar function. Mol Biosci 1971, 10:25-61.
260:1661-1664.
26. McCormick DA, Lavond DG, Clark GA, Kettner RE, Rising CE,
47. Lüscher C, Nicoll R, Malenka RC, Muller D: Synaptic plasticity and
Thompson RF: The engram found? Role of the cerebellum in
dynamic modulation of the postsynaptic membrane. Nat Neurosci
classical conditioning of nictitating membrane and eyelid
2000, 3:545-550.
responses. Bull Psychol Soc 1981, 18:103-105.
48. O’Dell TJ, Kandel ER, Grant SN: Long-term potentiation in the
27. Thompson RF, Thompson JK, Kim JJ, Krupa DJ, Shinkman PG: The
nature of reinforcement in cerebellar learning. Neurobiol Learn hippocampus is blocked by tyrosine kinase inhibitors. Nature
Mem 1998, 70:150-176. 1991, 353:558-560.

28. Thach WT: Correlation of neural discharge with pattern and force 49. Silva AJ, Paylor R, Wehner JM, Tonegawa S: Impaired spatial
of muscular activity, joint position, and direction of intended next learning in alpha-calcium-calmodulin kinase II mutant mice.
movement in motor cortex and cerebellum. J Neurophysiol 1978, Science 1992, 257:206-211.
41:654-676. 50. Silva AJ, Stevens CF, Tonegawa S, Wang Y: Deficient hippocampal
29. Fiez JA, Petersen SE, Cheney MK, Raichle ME: Impaired non-motor long-term potentiation in alpha-calcium-calmodulin kinase II
learning and error detection associated with cerebellar damage. mutant mice. Science 1992, 257:201-206.
Brain 1992, 115:155-178. 51. Malinow R: Silencing the controversy in LTP? Neuron 1998,
30. van Mier H, Tempel L, Perlmutter JS, Raichle ME, Petersen SE: 21:1226-1227.
Generalization of practice-related effects in motor learning using 52. Grant SG, O’Dell TJ, Karl K, Stein PL, Soriano P, Kandel ER:
the dominant and nondominant hand measured by PET. Soc Impaired long-term potentiation, spatial learning, and
Neurosci Abstr 1995, 21:1441. hippocampal development in Fyn mutant mice. Science 1992,
31. Ivry RB, Keele SW, Diener HC: Dissociation of the lateral and 258:1903-1910.
medial cerebellum in movement timing and movement execution. 53. Abel T, Nguyen PV, Barad M, Deuel TAS, Kandel ER,
Exp Brain Res 1998, 73:167-180. Bourtchouladze R: Genetic demonstration of a role for PKA in the
32. Penfield W, Milner B: Memory deficits induced by bilateral lesions late phase of LTP and in hippocampus-based long-term memory.
in the hippocampal zone. AMA Arch Neurol Psychiatry 1958, Cell 1997, 88:615-626.
79:475-497. 54. Bourtchouladze R, Frenguelli B, Blendy J, Cioffi D, Schutz G, Silva AL:
33. Alvarez P, Squire LR: Memory consolidation and the medial Deficient long-term memory in mice with a targeted mutation of
temporal lobe: a simple network model. Proc Natl Acad Sci USA the cAMP-responsive element-binding protein. Cell 1994,
1994, 91:7041-7045. 79:59-68.

34. McClelland JL, McNaughton BL, O’Reilly RC: Why there are 55. Impey S, Mark M, Villacres EC, Poser S, Chavkin C, Storm DR:
complementary learning systems in the hippocampus and Induction of CRE-mediated gene expression by stimuli that
neocortex: insights from the successes and failures of generate long-lasting LTP in area CA1 of the hippocampus.
connectionist models of learning and memory. Psychol Rev 1995, Neuron 1996, 16:973-982.
102:419-457.
56. Impey S, Smith DM, Obrietan K, Donahue R, Wade C, Storm DR:
35. Squire LR, Alvarez P: Retrograde amnesia and memory Stimulation of cAMP response element (CRE)-mediated
consolidation: a neurobiological perspective. Curr Opin Neurobiol transcription during contextual learning. Nat Neurosci 1998,
1995, 5:169-177. 7:595-601.
36. Yin JCP, Tully T: CREB and the formation of long-term memory. 57. Tsien JZ, Chen DF, Gerber D, Tom C, Mercer EH, Anderson DJ,
Curr Opin Neurobiol 1996, 6:264-268. Mayford M, Kandel ER, Tonegawa S: Subregion- and cell type-
restricted gene knockout in mouse brain. Cell 1996,
37. Kandel ER, Pittenger C: The past, the future and the biology of 87:1317-1326.
memory storage. Philos Trans R Soc Lond B Biol Sci 1999,
354:2027-2052. 58. Tsien JZ, Huerta PT, Tonegawa S: The essential role of
hippocampal CA1 NMDA receptor-dependent synaptic plasticity
38. Dash PK, Hochner B, Kandel ER: Injection of the cAMP-responsive in spatial memory. Cell 1996, 87:1327-1338.
element into the nucleus of Aplysia sensory neurons blocks long-
term facilitation. Nature 1990, 345:718-721. 59. Mayford M, Bach ME, Huang Y-Y, Wang L, Hawkins RD, Kandel ER:
Control of memory formation through regulated expression of a
39. Bartsch D, Casadio A, Karl KA, Serodio P, Kandel ER: CREB1 CaMKII transgene. Science 1996, 274:1678-1683.
encodes a nuclear activator, a repressor, and a cytoplasmic
modulator that form a regulatory unit critical for long-term 60. Li Y, Erzurumi RS, Chen C, Jhaveri S, Tonegawa S: Whisker-related
facilitation. Cell 1998, 95:211-223. neuronal patterns fail to develop in the trigeminal brainstem of
NMDAR1 knockout mice. Cell 1994, 76:427-437.
40. Bartsch D, Ghirardi M, Skehel PA, Karl KA, Herder SP, Chen M,
Bailey CH, Kandel ER: Aplysia CREB2 represses long-term 61. Tang YP, Shimizu E, Dube GR, Rampon C, Kerchner GA, Zhuo M,
facilitation: relief of repression converts transient facilitation into Liu-G, Tsien JZ: Genetic enhancement of learning and memory in
long-term functional and structural change. Cell 1995, 83:979-992. mice. Nature 1999, 401:63-69.
 Cognitive neuroscience Albright, Kandel and Posner 623

62. Gossen M, Freundlieb S, Bender G, Muller G, Hillen W, Bujard H: 86. Reynolds JH, Chelazi L, Desimone R: Competitive mechanisms
Transcriptional activation by tetracyclines in mammalian cells. subserve attention in macaque areas V2 and V4. J Neurosci 1999,
Science 1995, 268:1766-1769. 19:1736-1753.
63. Mansuy IM, Mayford M, Jacob B, Kandel ER, Bach ME: Restricted 87. Desimone R, Duncan J: Neural mechanisms of selective visual
and regulated overexpression reveals calcineurin as a key attention. Annu Rev Neurosci 1995, 18:193-222.
component in the transition from short- to long-term memory.
Cell 1998, 92:39-49. 88. Assad WF, Rainer G, Miller EK: Neural activity in the primate
prefrontal cortex during associative learning. Neuron 1998,
64. Mansuy IM, Winder DG, Moallem TM, Osman M, Mayford M, 21:1399-1407.
Hawkins RD, Kandel ER: Inducible and reversible gene expression
with the rtTA system for the study of memory. Neuron 1998, 89. Glimcher PW, Sparks DL: Movement selection in advance of action
21:257-265. in the superior colliculus. Nature 1992, 355:542-545.

65. Mulkey RM, Endo S, Shenolikar S, Malenka RC: Involvement of a 90. Kim JN, Shadlen MN: Neural correlates of a decision in the
calcineurin/inhibitor-1 phosphatase cascade in hippocampal dorsolateral prefrontal cortex of the macaque. Nat Neurosci 1999,
long-term depression. Nature 1994, 369:486-488. 2:176-185.

66. Winder DG, Mansuy IM, Osman M, Moallem TM, Kandel ER: Genetic 91. Merzenich MM, Recanzone GH, Jenkins WM, Grajski KA: Adaptive
and pharmacological evidence for a novel, intermediate phase of mechanisms in cortical networks underlying cortical contributions
−LTP) suppressed by calcineurin. Cell
long-term potentiation (I− to learning and nondeclarative memory. Cold Spring Harb Symp
1998, 92:25-37. Quant Biol 1990, 55:873-887.

67. Hebb DO: The Organization of Behavior: A Neuropsychological 92. Gilbert CD, Wiesel TN: Receptive field dynamics in adult primary
Theory. New York: Wiley; 1949. visual cortex. Nature 1992, 356:150-152.

68. Albright TD: Motion perception and the mind-body problem. Curr 93. Chino YM, Kaas JH, Smith EL 3rd, Langston AL, Cheng H: Rapid
Biol 1991, 1:391-393. reorganization of cortical maps in adult cats following restricted
deafferentation in retina. Vision Res 1992, 32:789-796.
69. Newsome WT, Britten KH, Movshon JA: Neuronal correlates of a
perceptual decision. Nature 1989, 341:52-54. 94. Darian-Smith C, Gilbert CD: Axonal sprouting accompanies
functional reorganization in adult cat striate cortex. Nature 1994,
70. Salzman CD, Britten KH, Newsome WT: Cortical microstimulation 368:737-740.
influences perceptual judgements of motion direction. Nature
1990, 346:174-177. 95. Das A, Gilbert CD: Receptive field expansion in adult visual cortex
is linked to dynamic changes in strength of cortical connections.
71. Stoner GR, Albright TD, Ramachandran VS: Transparency and J Neurophysiol 1995, 74:779-792.
coherence in human motion perception. Nature 1990,
344:153-155. 96. Recanzone CH, Schreiner CE, Merzenich MM: Plasticity in the
frequency representation of primary auditory cortex following
72. Stoner GR, Albright TD: The interpretation of visual motion: discrimination training in adult owl monkeys. J Neurosci 1993,
evidence for surface segmentation mechanisms. Vision Res 1996, 13:87-103.
36:1291-1310.
97. Zohary E, Celebrini S, Britten KH, Newsome WT: Neuronal plasticity
73. Duncan RO, Albright TD, Stoner GR: Occlusion and the that underlies improvement in perceptual performance. Science
interpretation of visual motion: perceptual and neuronal effects of 1994, 263:1289-1292.
context. J Neurosci 2000, 20:5885-5897.
98. Andersen RA, Mountcastle VB: The influence of the angle of gaze
74. Stoner GR, Albright TD: Neural correlates of perceptual motion upon the excitability of the light-sensitive neurons of the
coherence. Nature 1992, 358:412-414. posterior parietal cortex. J Neurosci 1983, 3:532-548.
75. Stoner GR, Albright TD: Image segmentation cues in motion 99. Andersen RA, Essick GK, Siegel RM: Encoding of spatial location
processing: implications for modularity in vision. J Cogn Neurosci by posterior parietal neurons. Science 1985, 230:456-458.
1993, 5:129-149.
100. Olson CR, Gettner SN: Object-centered direction selectivity in the
76. Peterhans E, von Der Heydt R: Subjective contours — bridging the macaque supplementary eye field. Science 1995, 269:985-988.
gap between psychophysics and physiology. Trends Neurosci
1991, 14:112-119. 101. Graziano MS, Yap GS, Gross CG: Coding of visual space by
premotor neurons. Science 1994, 266:1054-1057.
77. Assad JA, Maunsell JH: Neuronal correlates of inferred motion in
primate posterior parietal cortex. Nature 1995, 373:518-521. 102. Gray CM, Konig P, Engel AK, Singer W: Oscillatory responses in cat
visual cortex exhibit inter-columnar synchronization which reflects
78. Albright TD: My most true mind thus makes mine eye untrue. global stimulus properties. Nature 1989, 338:334-337.
Trends Neurosci 1995, 18:331-333.
103. Cabeza R, Nyberg L: Imaging cognition. II: an empirical analysis of
79. Bushnell MC, Goldberg ME, Robinson DL: Behavioral enhancement 275 PET and fMRI studies. J Cog Neurosci 2000, 12:1-47.
of visual responses in monkey cerebral cortex. I. Modulation in
posterior parietal cortex related to selective visual attention. 104. Fox PT, Parsons LM, Lancaster JL: Beyond the single study:
J Neurophysiol 1981, 46:755-772. function/location metanalysis in cognitive neuroimaging. Curr
Opin Neurobiol 1998, 8:178-187.
80. Spitzer H, Desimone R, Moran J: Increased attention enhances both
behavioral and neuronal performance. Science 1988, 240:338-340. 105. Lashley KS: Brain Mechanisms and Intelligence: A Quantitative Study
of Injuries to the Brain. Chicago: University of Chicago Press; 1929.
81. McAdams CJ, Maunsell JHR: Effects of attention on orientation-
tuning functions of single neurons in macaque cortical area V4. 106. Van Essen DC, Drury HA, Joshi S, Miller MI: Functional and
J Neurosci 1999, 19:431-441. structural mapping of human cerebral cortex: solutions are in the
surfaces. Proc Natl Acad Sci USA 1998, 95:788-795.
82. Moran J, Desimone R: Selective attention gates visual processing
in the extrastriate cortex. Science 1985, 229:782-784. 107. Tootell RB, Mendola JD, Hadjikhani NK, Ledden PJ, Liu AK, Reppas JB,
Sereno MI, Dale AM: Functional analysis of V3A and related areas
83. Treue S, Maunsell JH: Attentional modulation of visual motion in human visual cortex. J Neurosci 1997, 17:7060-7078.
processing in cortical areas MT and MST. Nature 1996,
382:539-541. 108. Posner MI, Gilbert CD: Attention and primary visual cortex. Proc
Natl Acad Sci USA 1999, 96:2585-2587.
84. Connor CE, Preddie DC, Gallant JL, Van Essen DC: Spatial attention
effects in macaque area V4. J Neurosci 1997, 17:3201-3214. 109. Ito M, Gilbert CD: Attention modulates contextual influences in the
primary visual cortex of alert monkeys. Neuron 1999, 22:593-604.
85. Croner LJ, Albright TD: Segmentation by color influences
responses of motion-sensitive neurons in the cortical middle 110. Corbetta M, Miezin FM, Dobmeyer S, Shulman GL, Petersen SE:
temporal visual area. J Neurosci 1999, 19:3935-3951. Selective and divided attention during visual discrimination of
624 A decade of Current neurobiology

shape, color, and speed: functional anatomy by positron emission 124. Bush G, Whalen PJ, Rose BR, Jenike MA, McInerney SC, Rauch SL:
tomography. J Neurosci 1991, 11:2388-2402. The counting Stroop: an interference task specialized for
functional neuroimaging-validation study with functional MRI.
111. Corbetta M, Kincade MJ, Ollinger JM, McAvoy MP, Shulman GL: Hum Brain Mapp 1998, 6:270-280.
Voluntary orienting is dissociated from target detection in human
posterior parietal cortex. Nat Neurosci 2000, 3:292-297. 125. Bush G, Luu P, Posner MI: Cognitive and emotional influences in
the anterior cingulate cortex. Trends Cogn Sci 2000, 4:215-222.
112. Gottlieb JP, Kusunoki M, Goldberg ME: The representation of visual
salience in monkey parietal cortex. Nature 1998, 391:481-484. 126. Drevets WC, Raichle ME: Reciprocal suppression of regional blood
flow during emotional versus higher cognitive processes:
113. Awh E, Jonides J, Smith EE, Schumacher EH, Koeppe RA, Katz S: implications for interactions between emotion and cognition. Cog
Dissociation of storage and rehearsal in verbal working memory: Emotion 1998, 12:353-385.
evidence from positron emission tomography. Psychol Sci 1996,
7:25-31. 127. Early TS, Reiman EM, Raichle ME, Spitznagel EL: Left globus
pallidus abnormality in never-medicated patients with
114. Hillyard SA, Anllo-Vento L: Event-related brain potentials in the schizophrenia. Proc Natl Acad Sci USA 1987, 84:561-563.
study of visual selective attention. Proc Natl Acad Sci USA 1998,
95:781-787. 128. Dolan RJ, Fletcher P, Frith CD, Friston KJ, Frackowiak RS, Grasby PM:
Dopaminergic modulation of impaired cognitive activation in the
115. Abdullaev YG, Bechtereva MP: Neuronal correlate of the higher- anterior cingulate cortex in schizophrenia. Nature 1995,
order semantic code in human prefrontal cortex. Int J Psychophysiol 378:180-182.
1993, 14:167-177.
129. Benes F: Alterations of neural circuitry within layer II of anterior
116. Martinez A, Anllo-Vento L, Sereno MI, Frank LR, Buxton RB, cingulate cortex in schizophrenia. J Psychiatr Res 1999, 33:511-512.
Dubowitz DJ, Wong EC, Hinrichs H, Heinze HJ, Hillyard SA:
Involvement of striate and extrastriate visual cortical areas in 130. Gazzaniga MS: The New Cognitive Neurosciences, edn 2.
Cambridge, Massachusetts: MIT Press; 2000.
spatial attention. Nat Neurosci 1999, 2:364-369.
131. Jiang Y, Haxby JV, Martin A, Ungerleider LG, Parasuraman R:
117. Worsley KJ, Cao T, Paus M, Petrides M, Evans AC: Application of
Complementary neural mechanisms for tracking items in human
random field theory to functional connectivity. Hum Brain Mapp working memory. Science 2000, 287:643-646.
1998, 6:364-367.
132. Dehaene S, Naccache L, Le Clec HG, Koechlin E, Muller M,
118. Gevins A: Electrophysiological imaging of brain function. In Brain Dehaene-Lamertz G, Vad DE, Mortele PF, Le Bihan D: Imaging
Mapping: The Methods. Edited by Toga AW, Mazziotta JC. San Diego: unconscious semantic priming. Nature 1998, 395:597-600.
Academic Press; 1996:259-273.
133. Petersen SE, Van Mier H, Fiez JA, Raichle ME: The effects of
119. Le Bihan D: Diffusion and Perfusion Magnetic Resonance Imaging. practice on the functional anatomy of task performance. Proc Natl
New York: Raven Press; 1995. Acad Sci USA 1998, 95:853-860.
120. Conturo TE, Lori NF, Cull TS, Akbudak E, Snyder AZ, Shimony JS, 134. Merzenich MM, Jenkins WM: Cortical representation of learned
McKinstry RC, Burton H, Raichle ME: Tracking neuronal fiber behaviors. In Memory Concepts. Edited by Anderson P, Hvalby B,
pathways in the living human brain. Proc Natl Acad Sci USA 2000, Paulsen B, Hokfelt B. Amsterdam: Elsevier; 1995:437-455.
96:10422-10427.
135. Karni A, Meyer G, Rey-Hipolito C, Jezzard P, Adams MM, Turner R,
121. Klingberg T, Hedehus M, Temple E, Salz T, Gabrieli JD, Moseley ME, Ungerleider LG: The acquisition of skilled motor performance: fast
Poldrack RA: Microstructure of temporo-parietal white matter as a and slow experience-driven changes in primary motor cortex.
basis for reading ability: evidence from diffusion tensor magnetic Proc Natl Acad Sci USA 1998, 95:861-868.
resonance imaging. Neuron 2000, 25:493-500.
136. Kim KH, Relkin NR, Lee KM, Hisch J: Distinct cortical areas
122. Dronkers NF: A new brain region for coordinating speech associated with native and second languages. Nature 1997,
articulation. Nature 1996, 384:159–161. 388:171-174.
123. Raichle ME, Fiez JA, Videen TO, MacLeod A-MK, Pardo JV, Fox PT, 137. Pantev ET, Wienbruch C, Rockstroh B, Taub E: Increased cortical
Petersen SE: Practice-related changes in human brain functional representation of the fingers of the left hand in string players.
anatomy during non-motor learning. Cereb Cortex 1994, 4:8-26. Science 1995, 270:305-307.