Spatial Navigation and Hippocampal Place Cell Firing:
The Problem of Goal Encoding
P Lenck-Santini, V Hok, E Save, P Banquet, P Gaussier, R Müller, Bruno
Poucet
To cite this version:
P Lenck-Santini, V Hok, E Save, P Banquet, P Gaussier, et al.. Spatial Navigation and
Hippocampal Place Cell Firing: The Problem of Goal Encoding. Reviews in the Neurosciences,
Walter de Gruyter, 2004, 15 (2), pp.89 - 107. <10.1515/REVNEURO.2004.15.2.89>. <hal01384901>
HAL Id: hal-01384901
https://hal.archives-ouvertes.fr/hal-01384901
Submitted on 7 Dec 2016
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
© Freund & Pettman, U.K.
Reviews in the Neurosciences, !5, 89-107 (2004)
Spatial Navigation and Hippocampal Place Cell Firing:
The Problem of Goal Encoding
Β. Poucet1, P.P. Lenck-Santini2, V. Hok1, E. Save1, J.P. Banquet3, P. Gaussier4 and R.U. Müller2·5
2
' Laboratory of Neurobiology and Cognition, CNRS-Universite de Provence, Marseille, France,
Department of Physiology and Pharmacology, SUNY Downstate Medical Center, Brooklyn, NY, USA,
3
Institut des Neurosciences et Modelisation, Paris, France,
4
Groupe Neurocybernetique, Universite de Cergy-Pontoise, France and
5
MRC Center for Synaptic Plasticity, Department of Anatomy, University of Bristol, Bristol, UK
SYNOPSIS
Place cells are hippocampal neurons whose
discharge is strongly related to a rat's location
in the environment. The existence of such cells,
combined with the reliable impairments seen in
spatial tasks after hippocampal damage, has led
to the proposal that place cells form part of an
integrated neural system dedicated to spatial
navigation. This hypothesis is supported by the
strong relationships between place cell activity
and spatial problem solving, which indicate that
the place cell representation must be both
functional and in register with the surroundings
for the animal to perform correctly in spatial
tasks. The place cell system nevertheless requires
other essentia) elements to be competent, such as
a component that specifies the overall goal of the
animal and computes the path required to take
the rat from its current location to the goal.
Here, we propose a model of the neural network
responsible for spatial navigation that includes
goal coding and path selection. In this model,
the hippocampal formation allows for place
recognition, and stores the set of places that can
be accessed from each position in the environment. The prefrontal cortex is responsible for
encoding goal location and for route planning.
Accepted: 21 January, 2 0 0 4
Reprint address:
B r u n o Poucet
Laboratory of N e u r o b i o l o g y and Cognition
Centre National de la R e c h e r c h e Scientifique
L N C / C N R S , 31 c h e m i n J o s e p h - A i g u i e r
13402 Marseille c e d e x 20, F r a n c e
e-mail: p o u c e t @ l n f . c n r s - m r s . f r
VOLUME 15, NO. 2, 2004
The nucleus accumbens translates paths in
neural space into appropriate locomotor activity
that moves the animal towards the goal in real
space. The complete model assumes that the
hippocampal output to nucleus accumbens and
prefrontal cortex provides information for
generating solutions to spatial problems. In
support of this model, we finally present
preliminary evidence that the goal representation necessary for path planning might be
encoded in the prelimbic/infralimbic region of
the medial prefrontal cortex.
KEY W O R D S
hippocampus, frontal cortex, unit recordings, place
cells, spatial navigation, rat
1. I N T R O D U C T I O N
Twenty-five years ago, O'Keefe and Nadel /76/
put together critical evidence that the hippocampus
carries out fundamental computations involved in
spatial navigation. The core of this theory was
provided by a discovery made a few years earlier.
While O'Keefe and Dostrovsky Π 51 were looking
for the behavioral discharge correlates of neurons
recorded from the hippocampus of freely moving
rats, they were struck by the strong positional
selectivity of some of the cells; their firing was so
strongly correlated with the rat's location in the
environment that they were named 'place cells'.
The discovery of place cells was an important
step in understanding the neural basis of spatial
processing since such cells provide ideal building
blocks for implementing the capability to navigate.
89
90
Β. POUCET ET AL.
It is obvious that for a rat to solve a complex spatial
problem, a problem that requires a representation of
the environmental layout, an initial requirement is
for the rat to locate itself. It is only once selflocalization is achieved that planning a path toward
a potential goal is possible. Performance in the
water maze navigation task, in which the rat is
demonstrably able to find rapidly a hidden platform
using the array of available visual cues, is currently
the prototypical illustrative example /55,63,1 13/.
Even though the spatial function of place cells is
hardly disputable, their direct involvement in the
generation of paths for navigation has remained
controversial. At first glance, the connection
between place cells and spatial memory is supported by the extensive lesion literature which
reveals that damage to the hippocampus results in
severe impairments in a wide variety of spatial
tasks /35,64,71,83,85/. This conclusion is corroborated by more recent work in genetically
modified mice with impaired long-term potentiation (LTP) and in rats after pharmacological
blockade of LTP; in either case place cells are
abnormal and there are corresponding performance
deficits in navigational tasks /11,42,57,93,94/.
Unfortunately, there are other interpretations of
the lesion data that render the support provided for
the cognitive mapping theory less than convincing.
For instance, the spatial memory deficits may result
not from a loss of the representation of the current
environment but instead could be due to a loss of
the ability to store episodic memory that is crucial
for performance /16,20,118/. The inherent difficulty
of inferring function from gross or subtle lesion
methods prevents us from gaining unambiguous
answers via this method 1 .
In this review we therefore take a different,
more direct approach that attempts to relate spatial
behavior to place cell discharge. The evidence we
present confirms that there are close ties between
the activity of single cells and the overall spatial
1
B e s i d e the u n a v o i d a b l e p r o b l e m of synaptic reorganizations
that follow p e r m a n e n t lesions, interpretation of lesion-induced
deficits is difficult b e c a u s e c o m p l e x behavior is often the
result of several, p e r h a p s sequentially arranged processes.
T h e r e f o r e , the o b s e r v a b l e deficit m a y result f r o m the loss of
only o n e of these processes, with the others being undisturbed
behavior of the rat and therefore lends strong
support for the direct involvement of place cells in
navigational processing. Whether such processing
is the main role of place cells, or just a particular
case of a more general function is left for future
work, but the spatial nature of place cell signals
suggests at least that space is an important aspect of
hippocampal function.
The strong spatial signals carried by hippocampal neurons do not mean, however, that the
hippocampus is sufficient to accomplish all
navigational functions. To the contrary, our review
emphasizes a key aspect of navigation, the coding
of the goal location, which appears to be more or
less independent of the hippocampus, at least to the
extent that no representation of goal within the
hippocampus has yet been found. We intend to
propose a model of the neural network responsible
for spatial navigation that includes goal selection.
After summarizing some of the main properties of
place cells, we go on to present preliminary
evidence that goal coding might depend on the
interplay between several brain structures, and
suggest that among these the prelimbic area of the
prefrontal cortex is crucial.
2. F U N D A M E N T A L P R O P E R T I E S O F P L A C E C E L L S
The place cells phenomenon is seen when one
monitors the activity of single pyramidal neurons in
the CA1 and CA3 areas of the hippocampus and
simultaneously follows the location of the rat as it
moves freely in the experimental environment.
Each place cell is intensely active only when the
rat's head is in a cell-specific part of the environment (called the 'firing field'), and is usually silent
elsewhere in the environment /64/. When rats are
exposed for the first time to a new environment,
place cells become active after a few minutes
of exploration /7,29,117/. Once firing fields are
established, their locations may be stationary for
weeks or months /68,108/. In addition, place cell
properties are very similar whether the rat is
engaged in an explicit spatial learning task /78/ or
during simple foraging in an open field /68/. Even
though place cell discharge may be modulated by
non-spatial variables /2,114,116/ or by the direction
when on a linear track /59,66/, spatial location is by
REVIEWS IN THE NEUROSCIENCES
N E U R A L BASES OF NAVIGATION
far the best and most consistent correlate of their
activity.
The location-specific nature of place cell
activity is possible only if the animal extracts
information about its environment. O'Keefe and
Conway /73/ were the first to address experimentally the question of the sensory control of
place cell firing. Using a T-maze inside a set of
curtains, they first demonstrated that firing fields
could be controlled by a set of four cues (a light, a
card, a fan, and a buzzer) placed within a curtained
'controlled-cue' environment some distance off the
maze arms. When the set of cues was rotated as a
whole, firing fields rotated equally. Thus, distal
cues manipulated by the experimenter were used by
the place cell system to anchor location-selective
activity. Control by explicit cues is also seen when
only a single landmark is present. For example,
Muller and Kubie 1611 recorded place cells in a cuecontrolled environment in which the only intended
orienting landmark was a salient cue card attached
to the wall of the recording cylinder. To measure
the control exerted by the card over firing fields,
the card was rotated around the center of the
cylinder in successive sessions. Under these circumstances, card rotation consistently resulted in
equal rotation of firing fields. The implication is
that the spatial firing of the cells is anchored to a
reference frame provided by the wall card. Similar
control was observed when the 2-dimensional card
was replaced with three 3-dimensional objects set
against the cylinder wall to form a triangular
configuration, but not when the same objects were
near the center of the cylinder /12,13/.
Although most experimental work has focused
on the control exerted over place cell firing by
visual information, other sensory cues may also be
important. For example, firing fields remain
relatively stable in the absence of visual information, provided olfactory cues on the floor are not
scrambled 1911. Recordings from blind rats also
suggest that tactile information is useful to anchor
location-selective activity in a reference frame
provided by slender objects /96/. Finally, clear
evidence indicates that auditory cues can exert
control over firing fields when visual information is
removed (Kubie and Muller, unpublished observations).
VOLUME 15, NO. 2, 2004
91
Self-motion information also may be important
in the absence of external cues /58,72,89,100/ or
when such cues are unreliable /38,95/. To use selfmotion information, the rat would update its
position by tracking changes in position using
signals derived from a variety of sources including
the vestibular and proprioceptive systems and
motor efference copy. A major problem with the
use of such relative information is that it leads to
unavoidable error accumulation. In the absence of
recalibration, the total error may become so large
that any self-localization becomes impossible.
Difficulties with pure or nearly pure self-motion
based positioning are confirmed by the finding that
place cell discharge becomes spatially unreliable
when external cues are severely restricted 1911.
To summarize, place cell firing is controlled by
both external and self-motion (idiothetic) cues that
interact with each other /38/ such that either type
can be dominant in different circumstances. Nevertheless, distal visual cues seem to be of primary
importance in the spatial control of firing fields
/67,73/; motion-related cues might allow ongoing
calculation of the rat's position without external
reference, thereby reducing attention to external
cues. Nevertheless, the rat still occasionally has to
recalibrate its position to correct for errors, and the
most usual basis for such recalibration is visual
information /60,87/.
3. PLACE CELL FIRING AND THE CODING
OF DISTINCT ENVIRONMENTS
Although place cell firing is characteristically
stable in constant surroundings, it can be markedly
different while the rat explores a different environment. Muller and Kubie 1611 recorded place cells in
two differently-shaped apparatuses, a cylinder and a
rectangular box, and found that the spatial firing
patterns of most cells were greatly changed between the two environments. Generally, cells that
fired in the two environments had fields that were
very different in location, size, and shape and
discharge rate. Other cells had fields in one
apparatus but not in the other. This 'remapping'
phenomenon suggests that each unique environment is represented by a distinct subset of
hippocampal pyramidal cells; cells common to both
92
Β. P O U C E T E T AL.
subsets have different firing fields. Although the
development of a new set of firing fields may be
very rapid, occurring in a matter of minutes /29,42,
117/, a slower time course is also possible 11,521.
Ultimately, however, place cells collectively
provide information about both the rat's current
environment and the rat's location within the
environment.
The extent to which each environment is
represented independently from other environments
is an important issue. We recently explored an
interesting case showing that two representations
can be independent even when the represented
environments are physically connected with a Ushaped runway that allows the rat to go from one
environment to the other /81 /. Place cells were first
recorded while rats explored two boxes that
differed in many ways, such as shape, color and
floor texture. As expected, each place cell had a
strongly different firing pattern in the two boxes.
The crucial manipulation was to change one of the
two boxes, causing a remapping of cell activity in
the changed box. Despite this clear result, the
spatial firing patterns of the same cells were in
general unaltered in the other, unchanged, box
(Fig. 1). We infer from this outcome that the
hippocampal representation is local to the immediate surroundings; an independent map is
formed for each apparatus and is activated only
when the rat enters the represented apparatus. Thus,
at least under the conditions used in this study, the
hippocampal map appeared to be mainly activated
and updated according to sensory data from the
rat's current environment. These results suggest
that place cell-based navigation is possible only
within the currently perceived environment and that
a switch of the map must take place for navigation
to be correct in distant, undetectable portions of
space.
4. PLACE CELL FIRING A N D NAVIGATION
4.1. P r e c i s i o n a n d reliability o f s p a t i a l c o d i n g b y
place cells
Beyond location-specific firing itself, what other
evidence is there for a role of place cells in spatial
learning and memory? One first issue is the
accuracy of spatial encoding by place cells. Since
every region of the environment accessible to the
rat is represented with about equal probability, how
precise is the information provided by the place cell
No remap
Phase 1
Fig. 1:
Remap
Phase 2
After place cells are recorded while the rat explores two distinctly shaped connected boxes, one box is substituted. This
change induces a local remapping of fields (shown as grey-shaded ellipses) in the changed box but discharge is mostly
unaffected in the unchanged box even though the rat can freely commute between the two boxes. From /81/.
r i - v i e w s in ι πι; ni:uros(.'Ii;nct:s
NEURAL BASES OF NAVIGATION
population about the rat's location within the
environment? Current estimates of the number of
place cells active for a specific apparatus is about
40% of the number of pyramidal cells. Since the
number of pyramidal cells is roughly 500,000 per
hippocampus, one imagines that many cells fire at
each location traversed by the rat. Very basic interpolations of instantaneous firing for small numbers
of simultaneously recorded cells allowed Wilson
and McNaughton /117/ to reconstruct the rat's path
with fair precision. With more sophisticated rules
that take phase-coded firing into account /77/,
Brown et al. /8/ were able to superimpose on the
rat's actual path the calculated path knowing the
instantaneous firing of just 16 simultaneously
recorded place cells (see also Jensen and Lisman
/39/). Thus, the ensemble activity of place cells is a
good predictor of the rat's location. This prediction
holds, however, better at the population level than
the single cell level /27/, since the activity of
individual place cells can be extremely variable
from one run through the firing field to another, a
phenomenon referred to as 'overdispersion' /19/.
4.2. Firing properties of place cells during navigation
If the firing of place cells relates in one way or
another to oriented behavior, then one should
expect differences in activity according to whether
the rat is unambiguously engaged in navigation or
simply moving in an apparently random fashion in
space. Although there is some evidence at a
structural level that hippocampal metabolic activity
is greater during a spatial task than during a control
task 161, there are only hints that place cell firing is
modulated by the rat's behavior. Thus, preliminary
results seem to indicate subtle modifications of
place cell firing during navigational behavior as
opposed to foraging, in which the rat engages in
random searching behavior for food (Fenton,
personal communication; see also ΙΊ9Γ). Comparisons of firing during navigation and foraging were
done during performance of the 'place preference
task' 1921 which was designed to allow place cell
recordings during both behaviors. In this task, the
rat has to enter a circumscribed, unmarked goal
zone in a circular environment and stay there for a
VOLUME 15. NO. 2. 2004
93
short period of time. A white cue card attached to
the cylinder wall is the only spatial landmark the rat
can use as a reference to find the goal zone. When
the rat goes to the goal zone and dwells for a long
enough time, an overhead dispenser is triggered to
release a single food pellet. Since the released
pellet can land anywhere in the cylinder, the rat has
to leave the goal zone to find it. To receive another
reward, the rat has to return to the goal zone.
Because the goal is spatially dissociated from the
wall card, the rat has to use a place strategy, in
which the goal location is computed from its spatial
relationship with the card (Fig. 2). Thus, each
successful 'trial' consists of two steps, one in which
the rat follows a path oriented towards the goal
zone (goal-directed navigation) and another in
which the rat follows a seemingly random trajectory while looking for the pellet (undirected
foraging).
Comparing passes through the field during
foraging and goal approach, subtle differences were
seen in several characteristics of CA1 place cell
firing. The clearest change during navigation was
an increase of firing reliability or, in other words,
decreased overdispersion /19/. The discharge rate
was only marginally increased during navigation
although cells fired more bursts than during
foraging. The inference is that the positional signal
of place cells is stronger during navigation than
during foraging, as if the rat attends more to the
current surroundings.
4.3. Relationship between the rat's spatial
performance and place cell spatial firing fields
How intimately is place cell discharge related to
behavior? In our view, probing this relationship
requires concomitant analysis at the cellular and
behavioral levels in normal animals after changes
are made in the environment. The idea is simply
that if the spatial information coded by place cells
is important for spatial problem solving, then
changes in the environment that result in changes of
place cell firing should be reflected in the animal's
choices. The next sections provide a brief review of
studies based on this principle.
94
Ii. POUCETETAL.
sj
1) Move to the goal
2) Wait 2 sec
\
Food is delivered
pa. '
3) Find the food
Fig. 2:
Place preference task. The rat is required to move to a specific location relative to a cue card (goal-directed navigation)
and wait there for 2 seconds. This triggers the delivery of a food pellet whose final location may be anywhere in the
cylinder, thus requiring the rat to search for it (undirected foraging). From /92/.
4.3.1. Navigation data supporting the spatial mapping
theory
O ' K e e f e and Speakman /78/ were the first to ask
how well changes in place cell activity corresponded to changes in behavior. Rats were trained to
go to one arm of a plus-shaped maze based on its
location relative to a fixed set of extra-maze cues.
On 'memory trials', these cues were removed
before the rat was allowed to make its choice.
When rats selected an arm in the absence of
adequate information, the firing fields of place cells
stayed in register with their behavioral choice, even
when the rat chose the wrong arm /78/.
A second indication of a link between place
cells and behavior is shown by progressive removal
of available sensory information, which often
resulted in a misregistration by ± 120° on a symmetric Y-maze of otherwise unchanged firing fields
/51/. Rats were trained to perform a spatial alternation task on the Y-maze using for landmark
information a prominent white card positioned
midway between two of the arms. Several different
cue manipulations including rotation or removal of
the card could cause the angular position of firing
fields to shift out of register by ± 120° from the
position seen with the card in its standard position.
In parallel with inconsistent field positions, alternation performance underwent a strong deterioration.
The nature of errors, classified according to the
required alternation sequence, indicated that the
rats were disoriented. In addition, the direction of
misregistration of firing fields often predicted which
arms the rat would visit erroneously. Overall, the
results suggest that inducing a mismatch between
the environment and the hippocampal representation of the environment disrupts the rat's behavior,
in clear agreement with the predictions of the
spatial mapping theory.
The functional role of place cells was further
investigated by looking at how the requirements of
a navigational task determined the effects of
disrupting their positional firing patterns /49/. The
cognitive map theory predicts that integrity of
hippocampal activity is critical for place navigation, but less important for cued (beacon)
navigation behavior /80/. Thus, we looked at
performance by normal animals in place and
beacon navigation tasks after environmental manipulations that disturbed the relationship between
the place cell representation and the cues used to
solve the problems. The theory predicts that such
disturbances will disrupt performance of place
navigation but not of beacon navigation.
REVIEWS IN THE NEIJROSCIENCES
N E U R A L BASES OF NAVIGATION
Place cells were recorded while rats performed
the place preference task (see Section 4.2). To
modify relationships between visible stimuli and
place cell activity, we made 'hidden' or 'visible'
90° rotations of the card on the cylinder wall and,
when present, independent rotations of the disk on
the cylinder floor. Hidden rotations were made with
the rat out of the cylinder and generally caused
equal firing field rotations. Visible rotations were
made with the rat inside the cylinder and very often
did not cause fields to rotate /95/. Thus, hidden card
rotations generally induced firing fields to stay in
register with the goal location whereas visible card
rotations generally induced firing fields to shift out
of register with the goal location. When a card
rotation did not induce field rotation, performance
95
was greatly impaired if the task required place
navigation to an arbitrary location relative to the
cue (Fig. 3). In contrast, performance was altered to
only a small extent when the animal had to go
directly towards the cue card, and not at all when
the cue was irrelevant to success because the goal
was a disk on the floor that was moved independently of the cue card /49/. Thus, reorganizing the
relationship between place cells and the environment affected performance in three highly matched
tasks to an extent directly predicted by the
navigational requirements of the task and, therefore,
according to the cognitive mapping theory, to an
extent predicted by the importance of place cells for
solving the task.
Field location is in register with the cue
Field location is out of register with the cue
Fig. 3:
W h e n rotation of the c u e is f o l l o w e d by c o r r e s p o n d i n g rotation of the firing field (ellipse), the h i p p o c a m p a l r e p r e s e n t a t i o n
is in register with the c u e card ( s h o w n as a thick line o u t s i d e the c y l i n d e r ) and t h u s with the n a v i g a t i o n a l task. T h e a n i m a l
usually p e r f o r m s a c o r r e c t search for the goal (location indicated by ' X ' ) . W h e n the f i r i n g field d o e s not f o l l o w c u e
rotation, the h i p p o c a m p a l representation is m i s a l i g n e d with t h e cue. T h e rat p e r f o r m s t h e task p o o r l y ; it usually s e a r c h e s
f o r the goal at a location consistent with the e r r o n e o u s position of the firing field ( b l a c k X ) , but inconsistent with the c u e
(grey X). F r o m /49/.
VOI.IIMI· 15. NO. 2. 2004
Β. POUCET ET AL.
96
4.3.2. Contradictory
findings
The recent literature provides an interesting
counter-example of the relationship between place
cell firing and spatial performance /37/ (see also
/33/). Place cells were recorded while rats had to
solve a navigational task whose acquisition had
been previously shown to depend on hippocampal
integrity. T h e task required the rat to go to a
specific corner in a square box in the center of a
room richly provided with diverse cues. On each
trial, the rat was first allowed to forage freely for
rice grains scattered on the floor of the box,
allowing place cell activity to be sampled everywhere in the box. Next, a tone was sounded to
signal the availability of a food reward at the goal
corner. Since the location of the goal corner was
constant from one trial to another, a well-trained rat
would immediately g o to the correct corner to get a
chocolate reward. At the outset, the arena walls and
lloor were black as they were during the extensive
training period necessary to get the animals to solve
the task reliably. O n c e the fields and behavioral
performance was characterized in the black box,
trials were conducted after the black box was
replaced with a white box. In both boxes, the rat
had to reach the same corner relative to the room
cues. Although the change from a black box to a
white box induced remapping for most place cells,
navigational performance remained essentially
intact /37/. Choice performance dropped from
- 9 0 % to - 7 0 % correct, but this statistically reliable
drop left performance in the white box well above
the 2 5 % chance level. Thus, rats could still perform
correctly even though their hippocampal m a p was
considerably disrupted by the change in the box
color.
At first glance, this o u t c o m e is hard to reconcile
with the idea that place cells drive navigation, but
we would like to propose an alternative explanation
that focuses on details of the rat's behavior. In this
view, the initial solution of this task relies on a
hippocampal representation of the box + laboratory.
Once such a representation is built, however, the rat
can switch to other strategies to perform correctly
/80/. Therefore, once the rat knew how to go
reliably to the goal corner in the study by Jeflery et
id. /37/. it could begin to use a beacon strategy
instead of a spatial strategy. In this view, the rat
identifies a room cue that is directly ' b e h i n d ' the
goal corner and simply heads there. A solution
method of this sort can be performed in the absence
of a functional hippocampus /46,74,82/. In this
study by Jeffery et al. /37/, reliance on a beacon
strategy is encouraged by the variety of cues in the
environment, by the simple geometric shape of the
box, and finally by the extensive period necessary
to train the rats to perform the task reliably.
We predict, in other words, that the same
hippocampal damage that prevents the rat from
learning the task /37/ would hardly affect performance after the 200 or more trials necessary to get to
criterion /46/. If the h i p p o c a m p u s is no longer
necessary for solution, remapping would not be
expected to degrade performance in the task, much
as misregistration of fields is not associated with
poor performance when a beacon directly signals
the goal /49/. Thus, we believe that the contradictory data reported by Jeffery et cd. /3 7/ can be
accounted for by the use of a beacon strategy by the
rat.
4.3.3. Place cell firing during behavior in response to
a spatial change
Preliminary exploration is necessary for rats to
solve spatial problems, presumably because key
information is gathered as the animal visits each
part of a new environment /18,90/. The idea that
information gathering is a fundamental function of
exploration is strengthened by the finding that
animals systematically re-explore modified portions
of a familiar environment /86/. Re-exploration occurs
regardless of whether the modification involves
changes in the relative positions of objects inside
an arena (spatial changes) or whether the modification involves substitution of a novel object for a
familiar object (non-spatial changes). The occurrence of change-induced re-exploration implies the
existence of a stored representation of the surroundings since behavior could not reliably be altered
unless it were possible to compare the current
situation to a previous situation. A key feature of
re-exploration after spatial changes is that it is
selectively reduced or eliminated by hippocampal
lesions, whereas re-exploration after non-spatial
changes is hardly affected /98/ but would probably
be affected by lesions of perirhinal cortex 753,70,
REVIEWS IN Till·; NEUROSCIENt'ES
NEURAL BASIIS OF NAVIGATION
111/. The implication is that the ability to recognize
the arrangement (but not the identity) of objects in
the arena d e p e n d s on the stored h i p p o c a m p a l
representation.
At the place cell level, r e m a p p i n g is a clear
c o n s e q u e n c e of an a n i m a l ' s e x p o s u r e to a strongly
modified e n v i r o n m e n t . It is, h o w e v e r , unclear h o w
more subtle m a n i p u l a t i o n s , such as object substitutions or r e a r r a n g e m e n t s , m i g h t alter place cell
activity. A c c o r d i n g to the e f f e c t s of h i p p o c a m p a l
lesions, spatial c h a n g e s
should
modify
the
positional firing characteristics o f place cells,
whereas non-spatial c h a n g e s should leave place cell
activity unaffected.
T o address this issue, we simultaneously
monitored the r a t ' s exploratory behavior while
recording place cell activity after positional rearrangements of objects or after object substitution
in a familiar e n v i r o n m e n t /50/. Overall, the results
c o n f i r m e d our basic prediction since field m o d i fication w a s observed a f t e r spatial changes, but not
after non-spatial c h a n g e s . T h e modification, h o w ever, was limited in a very interesting way since it
affected only the fields near the displaced objects;
fields further f r o m the o b j e c t s near the apparatus
wall were generally unaltered. W e c o n c l u d e that the
layout of the apparatus and the objects t h e m s e l v e s
are coded separately (see also /24,91/). Interestingly. place cells failed to respond to object
substitution, regardless o f field location. T h i s
suggests that object identity is not the primary
information signaled by place cell firing. Instead,
objects appear to act as m a r k e r s of specific
locations, and their spatial a r r a n g e m e n t is usually
m o r e important than their identities /13/.
In s u m m a r y , there is a striking parallel between
lesion and single cell data. M u c h as h i p p o c a m p a l
lesions disrupt behavioral reactions to spatial
changes, but hardly a f f e c t reactions to non-spatial
changes, pyramidal cell activity is m o d i f i e d by a
spatial c h a n g e but is hardly affected by a n o n spatial change. W e take this overall c o r r e s p o n d e n c e
as strong evidence that place cells are actively
involved in re-exploration and thus in spatial
processing.
VOLUMI . 15. NO. 2. 200-4
97
4.4. Relationships between hippocampal cell firing
and rats' performance in non-spatial tasks
M a n y unit recording studies indicate that hippoc a m p a l cells may d o m o r e t h a n j u s t e n c o d e the
a n i m a l ' s position. For e x a m p l e , h i p p o c a m p a l cells
recorded f r o m rats p e r f o r m i n g a non-spatial version
o f the radial a r m - m a z e task s h o w activity c h a n g e s
associated with specific visual-tactile c u e s on the
m a z e a r m s /120/. Similarly, in olfactory d i s c r i m i n a tion tasks, s o m e h i p p o c a m p a l c o m p l e x - s p i k e cells
s h o w time-locked activity to v a r i o u s task-relevant
behaviors, such as o d o r s a m p l i n g and a p p r o a c h
m o v e m e n t s /17.116/. S u c h e x a m p l e s o f non-spatial
correlates of h i p p o c a m p a l p y r a m i d a l cell
firing
d e m o n s t r a t e the diversity o f t u n i n g of h i p p o c a m p a l
p y r a m i d a l cells and suggest the c o m p l e x i t y of the
c o m p u t a t i o n s that m i g h t be p e r f o r m e d by the
hippocampus.
T h e likelihood that the h i p p o c a m p u s is involved
in processing non-spatial ' r e l a t i o n a l ' i n f o r m a t i o n
d o e s not, h o w e v e r , vitiate the idea that hippoc a m p a l neurons are c o m p o n e n t s of an environmental m a p and are ultimately essential for the
generation of navigation paths. T h e structure and
anatomical c o n n e c t i o n s of the h i p p o c a m p u s suggest
that it is particularly well suited for p r o c e s s i n g
i n f o r m a t i o n in a special w a y , and spatial p r o c e s s i n g
m a y be a prototypical illustration of this m o d e , a
specific instance of a w i d e r set o f f u n c t i o n s . T h u s ,
e v e n if spatial processing d o e s not c a p t u r e the full
f u n c t i o n of the h i p p o c a m p u s , the strong relationships between place cell activity and spatial
behavior underscore the idea that any r e a s o n a b l e
account of h i p p o c a m p a l f u n c t i o n has to include the
p r o c e s s i n g of spatial i n f o r m a t i o n . T h i s notion is
further supported by the existence, in areas connected to the h i p p o c a m p u s , of cells also carrying
p o w e r f u l and c o m p l e m e n t a r y spatial signals (e.g..
head direction cells / 105-107/). Head direction cells
share m a n y properties with place cells including the
w a y s in which they respond to external (mainly
visual) and self-motion c u e s /44,107.1 15/. Interestingly, the activity o f head direction cells correlates
well with spatial behavior so that spatial choices
are usually in register with the preferred firing
direction of head direction cells /15/. Finally, the
strong coupling of head direction cells and place
Β. POUCET ET AL.
98
cells /44/ is highly suggestive of an integrated
system dedicated at least partly to navigation.
5. N E U R A L I M P L E M E N T A T I O N
OF PLACE NAVIGATION:
THE PROBLEM OF THE GOAL LOCATION
We have presented evidence that place cells
contribute in an important way to the ability of rats
to solve spatial problems. Accordingly, the neural
machinery that allows an animal to select efficient
and perhaps optimal paths through the environment
should include the hippocampal place cell system
as a key component. Nevertheless, the main signal
carried by place cells, the rat's moment-to-moment
location in its current environment, is not sufficient
for the rat to compute paths. At least two other
crucial kinds of information are necessary for it to
be possible to compute complete paths, namely, a
representation of the goal and of the sequence of
movements to get from the current location to the
goal. We now briefly describe two classes of
solutions to how the goal and progress to the goal
are represented.
In one class of solutions, the goal is somehow
signaled by place cell activity and the hippocampus
itself computes the optimal path2. In the second
class of solutions, a larger network of structures
solves the whole navigational problem such that
each module in this system is dedicated to a
specific aspect of spatial processing. After considering these possibilities we will present
preliminary evidence in favor of the idea that
spatial navigation requires a larger network and
will emphasize the importance of medial frontal
cortex and striatum.
5.1.Hippocampus-only
models
Following the discovery of place cells, O'Keefe
and Nadel /76/ proposed a model system, centered
on the hippocampus, whose main role was to
"
This
does
not
mean
that
place
cells
contain
all
the
i n f o r m a t i o n n e c e s s a r y to d o the c o m p u t a t i o n . In fact, m a n y
m o d e l s a s s u m e that the c o m p l e m e n t a r y p r o p e r t i e s o f place
cells, h e a d direction cells and o t h e r p o p u l a t i o n s o f cells in t h e
h i p p o c a m p u s and p a r a - h i p p o c a m p a l a r e a s c o n t r i b u t e t o t h e
implement spatial navigation functions. Since then,
several investigators have proposed schemes that
rely on the fact that place cells and the hippocampal
machinery can potentially provide the animal with
all information required to compute paths through
the environment. It is not in the scope of this short
review to describe these models fully /110/ but it is
useful to state their basics if only to say how they
are lacking.
Hippocampus-only models assume that inputs to
hippocampal place cells provide the information
necessary to build place representations. They
differ, however, in how such place information is
used during goal-oriented navigation. Some models
depend on topological relationships in which the rat
gets from its current location to the goal by
traversing a series of places along a previously
experienced route /5,54,99,110/. In such models,
the path is formed by using the properties of LTP.
As the rat runs from its starting position to the goal
during learning, place cells along the route fire in
sequence so that synapses connecting place cells
with adjacent fields are strengthened. When the rat
is again at the starting point of such a sequence, the
whole series of place cells can be replayed, thereby
allowing the rat to follow the path from cell to cell
so that it eventually gets to the goal. Models of this
sort encounter serious difficulties when the rat (or
robot) learns routes to two (or more) goals that
cross since near the crossing the direction of
further progress is ambiguous. Elaborations of
path-following models can solve this difficulty 1221.
In a related model, exploration by the rat allows
place cells to be linked by synapses whose strength
represent in a general way the distance between
their firing fields 1691. Thus, LTP-modifiable
connections between place cells with fields that are
close to each other would be strengthened because
the cells fire in close temporal contiguity. In
contrast, connections between place cells with
widely separated fields would be weak because
such cells cannot fire together. This encoding of
distance is a consequence of connecting place cells
by Hebbian-like synapses and could occur in the
recurrent pyramidal cell to pyramidal cell connections in CA3. This model has the advantage that it
does not rely on previously taken paths; it permits
the animal to find an optimal path from any starting
g e n e r a t i o n of o p t i m a l p a t h s / 1 0 , 1 1 5 / .
REVIEWS IN T H E NEUROSCIKNCES
N E U R A L BASES OF NAVIGATION
point to any goal location in a familiar environment.
Metric models also do not depend on following
previously taken paths /9,10,112,119/. In such
models, the rat is able to compute the vector that
points from its current location to the goal by a
change of activity in place cells whose firing fields
are at the goal. It is possible to find a vector for
each initial location so that the direction and
distance to the goal can be computed from any
starting point /110/.
In common to topological and metric models is
that the representation of the goal location is
directly assigned to place cell activity. This is a
questionable assumption since there is little evidence that goal location in any way influences place
cell activity. In particular, spatial firing patterns do
not seem to undergo systematic changes when the
goal is moved /51,102,109/. For example, in the
absence of visual information, firing fields do not
stay in register with the single arm of a symmetric
Y-maze at whose end food is available even though
this cue is sufficient to orient the animal /51/.
Similarly, hippocampal place cell firing fields do
not tend to occur in higher numbers near the goal
zone in any of the three versions of the place
preference task used to study how place cell activity
is related to behavior /49/.
In contrast to these negative findings, there have
been several reports that firing fields occur in
excess numbers at goal locations /31,32,45/. Does
such accumulation reflect a true representation of
the goal or is it possibly due to methodological
practices? In the cited studies, increased firing at
the goal could be caused by reward-induced
modulation of neuronal activity rather than by
signaling of the goal location by place cells. In
addition, in most experimental designs, the rat is
required to spend more time in the vicinity of the
goal than elsewhere in the environment. Since place
cell firing is not zero outside firing fields, excess
time in a certain location that happens to be the
goal may cause an apparent excess of local activity.
I or example, in the place preference task, rats
indeed spend more time in the zone designated to
trigger food pellet release. Under these circumstances. the number of action potentials fired in the
trigger zone appears elevated compared to other
VOLUME 15. NO. 2. 2004
99
regions outside the firing field, but this is mainly
because the time spent per unit area is lower
elsewhere. Since each report of accumulation of
fields near the goal involved preferred sampling of
the environment near the goal, the issue of inhomogeneous sampling is crucial.
At present, our position has the following form.
We cannot assert that goal location or anticipated
arrival at a goal has no effect on place cell activity.
Nevertheless, in the absence of more convincing
evidence for such signals we assume that goals are
not represented in this way. In part, our willingness
to take this position depends on recent, positive
evidence that goal location in fact influences the
activity of cells in medial prefrontal cortex, a topic
considered near the end of this paper.
5.2. Distributed models
The lack of goal representation in place cell
firing suggests that such a representation exists in
one or more other brain structures and therefore
implies that the navigational system includes more
areas than just the hippocampus. Distributed models
differ according to the functions presumably served
by each area but share the idea that navigation is
separable into three processes /3,21,62/. In the first
process, the map associated with the current
environment is activated, thereby allowing selflocalization. In the second process, the goal
location is identified on the map and a decision is
made about how to get to the goal. In the third
process, locomotor commands that take the rat
towards the goal are generated using information
from the goal-location machinery.
We focus here on the model of Gaussier et al.
/3,21 / since we use it as the framework for
explaining some recent empirical findings. In common with virtually all models, its first component is
presumed to include the hippocampal formation
which is responsible for identification of the current
environment and determination of the animal's
position within this environment; together these are
called 'place recognition'. The hippocampal formation also learns and stores (possibly in the recurrent
connections of CA3) the set of places that can be
accessed from each position in the environment
that we may refer to as "transition learning". Such
100
Β. POUCET ET AL.
learning relies on associations between current and
adjacent locations, possibly via the strengthening of
synapses between place cells with fields close to
each other /69/ (Section 5.1). The outcome is a
purely topological representation of space such that
each place acts as an attractor that triggers the
representation of nearby locations. The Gaussier
model avoids assigning to place cells information
about the motivational valence of the animal's
current location or of neighboring locations. As a
result, the search for optimal paths is shifted f r o m
the hippocampal module to the other components
of the proposed navigational network.
The complete model assumes that the hippocampal output to the nucleus accumbens and prefrontal cortex provides information for generating
solutions to spatial problems. Using elementary
information about spatial transitions provided by
the hippocampus, the ventral striatum can build
simple sensori-motor sequences that select motor
responses depending on an a n i m a l ' s immediate
needs ('drives'). By itself, the hippocampal/striatal
system allows only automatic performance of
rewarded paths /80/ but cannot produce the
complex, flexible action sequences required to
achieve novel goals as in spatial navigation.
The second anatomical component in the model
is the prefrontal cortex which is thought responsible
for encoding goal location and for route planning.
Specifically, the prelimbic/infralimbic portion of
the prefrontal cortex receives input directly from
the ventral hippocampus /36/, some of whose
pyramidal cells are believed to function as place
cells although they occur with lower probability
/88/ or have lower resolution /40/ than place cells in
dorsal hippocampus. The hippocampal input may
therefore provide positional information to the key
portions of prefrontal cortex. Prelimbic/infralimbic
cortex also receives input from the amygdala that
may be the source of information about the reward
value of different locations in the environment.
In contrast to the simple sequences possible
with only the hippocampal plus striatal modules,
the complex planning necessary for true navigation
requires participation of the prefrontal cortex. By
hypothesis, hippocampally
supplied
transition
information from the entire environment is linked
to each other in the prefrontal cortex so that it form
continuous paths. In the language of graph theory,
each atom of transition information from the
hippocampus is an ' e d g e ' that connects a pair of
' n o d e s ' , and the complete set of transition information forms an 'adjacency matrix'. The j o b of the
prefrontal cortex is to build a path from the current
location to the goal using the transition information
from the hippocampus. Thus, a path is a set of
linked transitions that represents a complex set of
actions independent of specific goals as during
latent learning. Because motivational signals are
also available to prefrontal cortex from amygdalar
inputs /34/, paths can be related to the animal's
goal.
The third anatomical component of the Gaussier
model is the nucleus accumbens (ventral striatum)
that translates paths in neural space into appropriate
locomotor activity that m o v e s the animal towards
the goal in real space. The nucleus accumbens gets
hippocampal input via the subiculum and fornix
/26,41/ but also receives afferente f r o m prefrontal
cortex, the amygdala, the hypothalamus and ventral
tegmental area, the latter implying strong dopaminergic innervation. Importantly, cells in the nucleus
accumbens respond both to spatial positional
information and reward expectancy /48.101/. In
contrast, cell discharge in the prefrontal cortex is
mostly associated with temporary storage of
information in m e m o r y /4,28,61/. Moreover, both
prefrontal cortex and nucleus accumbens are tightly
related to motor systems through connections with
(lie ventral pallidum and dorsal striatum and the
premotor/motor cortex, respectively.
The model thus assumes that goals are represented by motivational nodes in the prefrontal
cortex. Activation of these nodes progressively
diffuses through the prefrontal graph of learned
transitions leading to the goal allowing identification of the optimal/shortest path to the goal.
Finally, the ventral striatum merges this prefrontal
top-down information with hippocampal information about possible transitions from the rat's current
location. If several transitions are possible, the
prefrontal activity biases for the selection of the
most relevant transition and the corresponding
motor output and generates local progression
towards the goal. By iteration, this process allows
REVIEWS IN T H E NEUROSCIENCES
101
N E U R A L BASES OF NAVIGATION
for the execution of the shortest path to the goal.
With regard to spatial navigation, the behavioral
output of the three-module system is the production
of smooth and direct paths oriented towards the
goal location 73,21/(Fig. 4).
5.3. Preliminary evidence for goal cells:
A unit recording study of the prefrontal cortex
In the model proposed by Gaussier et al. /21/,
spatial planning relies on the activity of a prefrontal
network which associates places with their motivational valence and is therefore able to specify goals.
If the model is correct, the firing of at least some
neurons in the prefrontal cortex should reflect a
combination of spatial and motivational correlates.
This dual encoding is required to begin the hypothesized backpropagation of a motivational signal
through the network to permit path building.
Recordings made during the place preference
task /92/ (see Section 4.2.) strongly suggest the
existence of goal cells in prefrontal cortex /30/. In
the particular version of the place preference task
that was used, the rat had to go to the 'trigger zone"
inside a cylinder to cause release of a pellet. Rats
were then observed to go reliably to the 'landing
zone", the place underneath the feeder where
released pellets first touched the floor. To actually
eat the pellet, however, the rat had to forage around
the cylinder area since the pellet scattered widely
after dropping. Thus, there are two potential fixed
goals in the cylinder, the trigger zone and the
landing zone. Reward itself was, however, broadly
distributed over the whole floor. This version of the
place preference task makes it possible to disentangle the goal value of places from their reward
value; firing in the trigger zone or the landing zone
cannot be caused by the food reward itself since
eating could occur anywhere.
Action selection
Motivation
-G2- - BD
Prefrontal Module
Striatal Module
Hippocampal Module
d/dt
Fig. 4:
A sketch of the model. In this architecture, CA3-CA1 learn transitions between places. Accumbens (ACC) and prefrontal
cortex (PF) learn, store, and perform temporospatia! sequences. Prefrontal activity produces a bias in accumbens activity,
leading to the selection of the response most appropriate to the organism's current goal. Reprinted from Banquet et al. ßl
with permission.
VOLUME 15. NO. 2. 2004
102
Β. POUCET ET AL.
The key finding of this study was that a substantial proportion ( - 2 0 % ) of cells in the prelimbic
and infralimbic parts of medial frontal cortex had
clear spatial correlates in the cylinder. This is
remarkable since no evidence for delimited firing
fields or even reliable regions of elevated firing was
found when the rat had simply to forage for food in
the cylinder /84/. Thus, prefrontal neurons are
much more likely to display location-specific firing
when the rat is engaged in planning than when it
simply wanders about in the environment. A second
important finding is that neurons with spatial
correlates were more likely to be found in the
prelimbic/infralimbic areas than in the more dorsal
anterior cingulate area. This functional result is
consistent with the distribution of anatomical
connections from the ventral hippocampus which
project to the more ventral aspects of the medial
frontal cortex /36/. A third finding was that most
( - 7 5 % ) of the spatial firing fields of prefrontal
neurons were found in the immediate vicinity of
either the trigger zone or the landing zone (Fig. 5),
the two fixed goal regions. Spatial correlates of this
type seem strongly tied to the motivational valence
of specific places. In short, neurons that discharge
in this fashion have precisely the properties expected of cells encoding spatial goals, cells necessary
for planning optimal paths in the environment.
It would not be surprising if neurons with
similar characteristics can be found in other brain
areas (e.g., the ventral striatum as suggested by
Gaussier's model or the parietal cortex as suggested
by other models /14/). Nevertheless we stress that
they are not found in the hippocampus, but rather in
Cell 1
Fig. 5:
Cell 2
Firing rate maps of two prefrontal cells showing
high spatial selectivity for the trigger zone (left) and
the landing zone (right). See text for details. From
/30/.
a structure that is separately implicated in planning
/25/ and has strong connections to the hippocampus. The finding of goal-related cells in prefrontal cortex provides a strong basis from which to
conclude that spatial navigation relies on a
dispersed, dedicated network that includes not only
the hippocampus but also the head direction
system, prefrontal cortex and at least parts of the
striatum.
6. C O N C L U S I O N
In line with the original theory of O ' K e e f e and
Nadel /76/, we believe that the basic properties of
hippocampal place cells imply that they form part
of an integrated neural system for spatial
navigation; a key additional part is contributed by
head direction cells /104/. The strong relationships
between place cell activity and spatial problem
solving indicate that the place cell representation
must be both functional and in register with the
surroundings for the animal to perform correctly in
spatial tasks. The system composed of 'spatially
tuned cells' (place cells and head direction cells)
nevertheless requires other essential elements to be
competent. In particular, we have focused on the
need for a component that could specify the overall
goal of the animal and could compute the path
required to take the rat from its current location to
the goal; we devoted much less attention to the
need for an output mechanism that converts
planning decisions into specific actions. The
essential message in this paper is that preliminary
evidence exists that the goal representation
necessary for path planning might be encoded in
the prelimbic/infralimbic region of the medial
prefrontal cortex.
Even though our strong belief is that the
hippocampal formation and prefrontal cortex are
parts of a distributed neural network supporting
complex spatial behavior, there are several unsolved issues that we raise briefly. One key
question involves the possibility that the hippocampal formation does more than implement a
representation of the rat's current environment and
its location in that environment. For example, it is
reasonable to ask whether .signals from the prefrontal cortex and striatum are sent back to the
REVIEWS IN ΊΊ IE NEUROSCIHNCES
NEURAL BASI-S OF NAVIGATION
h i p p o c a m p u s . S u c h f e e d b a c k c o u l d a l l o w t h e rat to
k e e p track o f its w h e r e a b o u t s a l o n g t h e p l a n n e d
path. A l t h o u g h recent e v i d e n c e s h o w s that d a m a g e
to the p r e f r o n t a l c o r t e x alters t h e p r o p e r t i e s o f
h i p p o c a m p a l p l a c e cell f i r i n g /47/, it is u n c l e a r
h o w i n f o r m a t i o n f r o m t h e f r o n t a l c o r t e x r e a c h e s the
h i p p o c a m p a l s y s t e m . S i m i l a r l y , recent w o r k ind i c a t e s that p r e l i m b i c frontal c o r t e x is r e q u i r e d for
learning o b j e c t - p l a c e a s s o c i a t i o n s / 4 3 / w h e r e a s t h e
h i p p o c a m p u s is n e c e s s a r y f o r l o n g - t e r m retention
o f such a s s o c i a t i o n s 1231. T h e s e f i n d i n g s are c o n sistent with t h e role o f t h e p r e f r o n t a l cortex
postulated a b o v e . T h e y also s u g g e s t that in t h e long
t e r m the h i p p o c a m p u s e v e n t u a l l y has a c c e s s to
i n f o r m a t i o n a b o u t t h e v a l u e of p l a c e s , t h u s r a i s i n g
the interesting possibility that t h e n a t u r e o f t h e
p l a c e cell signal m i g h t c h a n g e with t i m e and
experience.
In a d i f f e r e n t vein, t h e e n h a n c e d
spatial
o r g a n i z a t i o n of t h e d i s c h a r g e o f b o t h h i p p o c a m p a l
place cells a n d frontal n e u r o n s s e e n w h e n the rat
e n g a g e s in n a v i g a t i o n c o m p a r e d to s i m p l e undirected e x p l o r a t i o n s u g g e s t s that there are (at
least) t w o m o d e s of activity in t h e s e regions.
A l t h o u g h o u r strong h y p o t h e s i s is that the spatial
signal seen in s o m e frontal cortical cells r e p r e s e n t s
t h e e n c o d i n g o f goals, it is m o r e d i f f i c u l t to
u n d e r s t a n d w h y place cell f i r i n g a p p e a r s m o r e
organized
during
active
spatial
navigation.
A l t h o u g h the t w o m o d e s c o u l d reflect a d i f f e r e n t
sort of attention to the spatial layout d u r i n g
f o r a g i n g as o p p o s e d to m o r e directed b e h a v i o r , it
could also be that t h e m o d e s reflect f u n d a m e n t a l l y
different forms of processing.
Lastly, e v e n if p l a c e cells are truly involved in
spatial n a v i g a t i o n , they m i g h t c o n t r i b u t e to o t h e r
a s p e c t s of c o m p l e x b e h a v i o r , s u c h as the storage of
e p i s o d i c m e m o r y as p r o p o s e d by E i c h e n b a u m et eil.
/16/, rapid l e a r n i n g /56/, or c o n s o l i d a t i o n o f recent
m e m o r i e s /103/. In such m o d e l s , it is usually
a s s u m e d that the h i p p o c a m p u s is not n e c e s s a r y for
l o n g - t e r m m e m o r y . S o w h y is place cell activity
persistent and reliable for w e e k s or m o n t h s / l 0 8 / ?
W e believe that a d d r e s s i n g t h e s e i s s u e s is necessary
for a c h i e v i n g a full u n d e r s t a n d i n g o f the f u n c t i o n of
the h i p p o c a m p u s a n d its p l a c e cells. It is o u r f u r t h e r
intuition, h o w e v e r , that rapid p r o g r e s s in unders t a n d i n g the neural basis o f n a v i g a t i o n d e p e n d s on
VOI.UMI·: 15. NO. 2. 2004
103
s o l v i n g t h e m y s t e r y of goal r e p r e s e n t a t i o n a n d the
w a y s in w h i c h m o t i v a t i o n is c o m b i n e d w i t h w h a t
m a y be a s y s t e m that is o t h e r w i s e p u r e l y g e o m e t r i c .
ACKNOWLEDGEMENTS
S u p p o r t f o r this w o r k w a s p r o v i d e d by the
C N R S . t h e M E N R T (AC1 N e u r o s c i e n c e s integratives et c o m p u t a t i o n n e l l e s ) , U S P H S - N I H g r a n t s
N S 2 0 6 8 6 a n d N S 3 7 1 5 0 . a n d an M R C ( U K )
O v e r s e a s Initiative G r a n t .
REFERENCES
1. A m b r o g i Lorenzini C G , Baldi E, B u c c h e r e l l i C,
Sacclietti B, T a s s o n i G. N e u r a l t o p o g r a p h y and c h r o n o logy o f m e m o r y c o n s o l i d a t i o n : a r e v i e w of f u n c t i o n a l
inactivatioh findings. N e u r o b i o l L e a r n M e m 1999; 71:
1-18.
2. A n d e r s o n Ml, J e f f e r y KJ. H e t e r o g e n e o u s m o d u l a t i o n of
place cell firing by c h a n g e s in c o n t e x t . J N e u r o s c i 2 0 0 3 ;
23: 8 8 2 7 - 8 8 3 5 .
3. Banquet JP. G a u s s i e r P, Q u o y M, Revel A, B u r n o d Y.
C o r t i c o - h i p p o c a m p a l m a p s and n a v i g a t i o n strategies in
robots and rodents. In: Hallan B, F l o r e a n o D, Hallam J,
Hallam B, Hages G, M e y e r JA, eds. F r o m A n i m a l s to
A n i m a l s 7. Boston, M A : M I T Press, 2 0 0 2 ; 7: 141-150.
4. Batuev A S . Kursina N P , S h u t o v A P . Unit activity of the
medial wall of the frontal cortex d u r i n g d e l a y e d p e r f o r m ance in rats. B e h a v Brain Res 1990; 4 1 : 9 5 - 1 0 2 .
5. Blum Kl, Abbott LF. A model of spatial m a p formation in
the h i p p o c a m p u s of the rat. Neur C o m p u t 1996; 8: 85-93.
6. B o n t e m p i B, L a u r e n t - D e m i r C, D e s t r a d e C. J a f f a r d R.
T i m e - d e p e n d e n t reorganization of brain circuitry u n d e r lying m e m o r y storage: e v i d e n c e u s i n g brain i m a g i n g .
N a t u r e 1999; 4 0 0 : 6 7 1 - 6 7 5 .
7. Bostock E, Muller RU, K u b i e JL. E x p e r i e n c e - d e p e n d e n t
m o d i f i c a t i o n s of h i p p o c a m p a l place cell firing. H i p p o c a m p u s 1991; 1: 193-206
8. B r o w n EN, Frank L M . T a n g D, Q u i r k M C , Wilson M A .
A statistical p a r a d i g m for neural s p i k e train d e c o d i n g
applied to position prediction f r o m e n s e m b l e firing
patterns of h i p p o c a m p a l place cells. J N e u r o s c i 1998;
18: 741 1-7425.
9. B u r g e s s N, B e c k e r S, King JA, O ' K e e f e J. M e m o r y for
e v e n t s and their spatial context: m o d e l s and e x p e r i m e n t s .
Phil T r a n s Roy Soc Lond Β 2 0 0 1 ; 3 5 6 : 1493-1503.
10. B u r g e s s N. Recce M, O ' K e e f e J. A m o d e l o f hippoc a m p a l f u n c t i o n . Neural N e t w o r k s 1994; 7: 1065-1081.
11. Clio Y H , G i e s e KP, Tanila HT, Silva A.I. E i c h e n b a u m
II. A b n o r m a l h i p p o c a m p a l spatial r e p r e s e n t a t i o n s in
a C a M K I I l : s " A and C R E B " ' - m i c e . S c i e n c e 1998: 279:
867-869.
104
Β. POUCET ET AL.
12. Cressant A, Muller RU, Poucet B. Failure of centrally
placed objects to control the firing fields of hippocampal
place cells. J Neurosci 1997; 1 7 : 2 5 3 1 - 2 5 4 2 .
13. Cressant A, Muller RU, Poucet B. A further study of the
control of place cell firing by intra-apparatus objects.
Hippocampus 1999; 9: 422-431.
14. Crowe DA, Chafee MV, Averbeck BB, Georgopoulos
AP. Neural activity in primate parietal area 7a related to
spatial analysis of visual mazes. Cereb Cortex 2003; 14:
23-34.
15. Dudchenko P, Taube JS. Correlation between head
direction cell activity and spatial behavior on a radial
maze. Behav Neurosci 1997; 1 1 1 : 3 - 1 9 .
16. Eichenbaum Η, Dudchenko Ρ, Wood Ε, Shapiro ML,
Tanila H. The hippocampus, m e m o r y , and place cells: Is
it spatial memory or a memory space? Neuron 1999; 23:
209-226.
17. Eichenbaum Η, Kuperstein Μ, Fagan Α, Nagode J. Cuesampling and goal approach correlates of hippocampal
unit activity in rats performing an odor discrimination
task. J Neurosci 1987; 7: 716-732.
18. Ellen P, Parko EM, Wages C, Doherty D, Herrmann T.
Spatial problem solving by rats: exploration and cognitive maps. Learn Motiv 1982; 13: 81-94.
19. Fenton AA, Muller RU. Place cell discharge is
extremely variable during individual passes of the rat
through the firing field. Proc Natl Acad Sei USA 1998;
95: 3182-3187.
20. Frank LM, Brown EN, Wilson M. Trajectory encoding
in the hippocampus and entorhinal cortex. Neuron 2000;
27: 169-178.
21. Gaussier P, Revel A, Banquet JP, Babeau V. From view
cells and place cells to cognitive map learning: processing stages of the hippocampal system. Biol Cybern
2002; 86: 15-28.
22. Gerstner W, Abbott LF. Learning navigational maps
through potentiation and modulation of hippocampal
place cells. J Comput Neurosci 1997; 4: 79-94.
23. Gilbert PE, Kesner RP. Memory for objects and their
location: the role of the hippocampus in retention of
object-place associations. Neurobiol Learn Mem 2003;
81: 39-45.
24. Gothard KM, Skaggs WE, Moore KM, McNaughton
BL. Binding of hippocampal CA I neural activity to
multiple reference frames in a landmark-based navigation task. J Neurosci 1995; 16: 823-835.
25. Granon S, Poucet B. Medial prefrontal lesions in the rat
and spatial navigation: evidence for impaired planning.
Behav Neurosci 1995; 109: 474-484.
26. Groenewegen HJ, Vermeulen-Van Der Zee P, Te
Kortschot A, Witter MP. Organization of the projections
from the subiculum to the ventral striatum in the rat. A
study using anterograde transport of Phaseolus
vtilgarin
leucogglutinin. Neuroscience 1987; 23: 103-120.
27. Harris KD, Csicsvari J, Hirase H, Dragoi G, Buzsaki G.
Organisation of cell assemblies in the hippocampus.
Nature 2003; 424: 552-556.
28. Hasegawa RP, Blitz A M , Geller N L , Goldberg ME.
Neurons in monkey prefrontal cortex that track past or
predict future performance. Science 2000; 290: 17861789.
29. Hill AJ. First occurrence of hippocampal spatial firing in
a new environment. Exp Neurol 1978; 62: 282-297.
30. Hok V, Save E, Poucet B. Unit recordings in the rat
medial frontal cortex during performance of a spatial
navigation task. Soc Neurosci Abst 2003; 29: 519.15.
31. Hollup SA, Molden S, Donnett JG, Moser MB, Moser
EI. Accumulation of hippocampal place fields at the goal
location in an annular watermaze task. J Neurosci 2001;
21: 1635-1644.
32. Holscher C, Jacob W, Mallot HA. Reward modulates
neuronal activity in the hippocampus of the rat. Behav
Brain Res 2003; 142: 181-191.
33. Huxter JR, Thorpe C M , Martin G M , Harley CW. Spatial
problem solving and hippocampal place cell firing in
rats: control by an internal sense of direction carried
across environments. Behav Brain Res 2001; 123: 3748.
34. Ishikawa A, Nakamura S. Convergence and interaction
of hippocampal and amygdalar projections within the
prefrontal cortex in the rat. J Neurosci 2003; 23: 99879995.
35. Jarrard LE. Selective hippocampal lesions and behavior:
effects of kainic acid lesions on performance of place
and cue tasks. Behav Neurosci 1983; 97: 873-889.
36. Jay T M , Glowinski J, Thierry A M . Selectivity of the
hippocampal projection to the prelimbic area of the
prefrontal cortex in the rat. Brain Res 1989; 505: 337340.
37. Jeffery KJ, Gilbert A, Burton S, Strudwick A. Preserved
performance in a hippocampal-dependent spatial task
despite complete place cell remapping. Hippocampus
2003; 13: 175-189.
38. Jeffery KJ, O ' K e e f e J. Learned interaction of visual and
idiothetic cues in the control of place field orientation.
Exp Brain Res 1999; 127: 151-161.
39. Jensen O, Lisman JE. Position reconstruction from an
ensemble of hippocampal place cells: contribution of
theta phase coding. J Neurophysiol 2000; 83: 26022609.
40. Jung MW, Wiener SI, McNaughton BL. Comparison of
spatial firing characteristics of units in dorsal and ventral
hippocampus of the rat. J Neurosci 1994; 14: 73477356.
41. Kelley AE, Domesick VB. The distribution of the projection from the hippocampal formation to the nucleus
accumbens in the rat: an anterograde- and retrogradehorseradish peroxidase study. Neuroscience 1982; 7:
2321-2335.
42. Kentros C, Hargreaves E, Hawkins RD, Kandel ER,
Shapiro Μ, Muller RU. Abolition of long-term stability
of new hippocampal place cell maps by N M D A receptor
blockade. Science 1998; 280: 2121-2126.
REVIEWS IN ΊΊ IC NEUROSC1KNCES
NEURAL BASES OF NAVIGATION
43. Kesner RP, R a g o z z i n o M E . T h e role of the prefrontal
cortex in o b j e c t - p l a c e learning: a test of the attribute
specificity m o d e l . B e h a v Brain Res 2 0 0 3 ; 146: 159-165.
44. Knierim JJ, K u d r i m o t i HS, M c N a u g h t o n BL. Place cells,
105
Suster M , W e a v e r K L . D e c i p h e r i n g t h e h i p p o c a m p a l
polyglot: the h i p p o c a m p u s as a path integration system.
J E x p Biol 1996; 119: 173-185.
landmark
59. M c N a u g h t o n BL, Barnes C A , O ' K e e f e J. T h e contributions of position, direction, and velocity to single unit
45. Kobayashi T, Tran A H , N i s h i j o Η, O n o T, M a t s u m o t o
G. Contribution of h i p p o c a m p a l place cell activity to
activity in the h i p p o c a m p u s of f r e e l y - m o v i n g rats. E x p
Brain Res 1 9 8 3 ; 5 2 : 4 1 - 4 9 .
60. M c N a u g h t o n BL, Knierim JJ, W i l s o n M A . V e c t o r
learning and f o r m a t i o n of goal-directed navigation in
rats. N e u r o s c i e n c e 2 0 0 3 ; 117: 1025-1035.
e n c o d i n g and the vestibular f o u n d a t i o n s of spatial
cognition: n e u r o p h y s i o l o g i c a l and c o m p u t a t i o n a l m e c h a -
head direction cells, and the learning of
stability. J N e u r o s c i 1995; 15: 1648-1659.
46. Kubie
JL,
Sutherland
R,
Muller
RU.
Hippocampal
lesions p r o d u c e a t e m p o r a l l y - g r a d e d retrograde a m n e s i a
on a dry version of the Morris s w i m m i n g task.
P s y c h o b i o l o g y 1999; 27: 3 1 3 - 3 3 0 .
47. Kyd J, Bilkey D K . P r e f r o n t a l cortex lesions m o d i f y the
spatial properties of h i p p o c a m p a l place cells. C e r e b
Cortex 2 0 0 3 ; 1 3 : 4 4 4 - 4 5 1 .
48. Lavoie A M , M i z u m o r i SJY. Spatial, m o v e m e n t - and
r e w a r d - s e n s i t i v e d i s c h a r g e by medial ventral striatum
n e u r o n s of rats. Brain Res 1994; 6 3 8 : 157-168.
49. Lenck-Santini PP. Muller RU, S a v e E, Poucet B.
Relationships b e t w e e n p l a c e cell firing fields and
navigational d e c i s i o n s by rats. J N e u r o s c i 2 0 0 2 ; 2 2 :
9035-9047.
50. Lenck-Santini PP, Poucet B. A study of place cell
activity and e x p l o r a t o r y b e h a v i o r f o l l o w i n g spatial and
non-spatial c h a n g e s in the e n v i r o n m e n t . Soc N e u r o s c i
Abst 2 0 0 3 ; 198.4.
51. Lenck-Santini PP, S a v e E, Poucet B. E v i d e n c e for a
relationship between the firing patterns of h i p p o c a m p a l
place cells and r a t ' s p e r f o r m a n c e in a spatial m e m o r y
task. H i p p o c a m p u s 2 0 0 1 ; 11: 3 7 7 - 3 9 0 .
52. Lever C, Willis T, C a c u c c i F, B u r g e s s N , O ' K e e f e J.
Long-term plasticity in h i p p o c a m p a l place-cell representation by e n v i r o n m e n t a l g e o m e t r y . N a t u r e 2 0 0 2 ; 4 1 6 :
90-94.
53. Liu P, Bilkey D K . The e f f e c t of excitotoxic lesions
centered on the h i p p o c a m p u s or perirhinal cortex in
object recognition and spatial m e m o r y tasks. Behav
N e u r o s c i 2 0 0 1 ; 115: 9 4 - 1 1 1 .
54. Mataric MJ. A distributed m o d e l for m o b i l e robot
e n v i r o n m e n t learning and navigation. Tech Rep M I T Al
L a b 1990; T R I 2 2 8 .
55. M a t t h e w s DB, Best PJ. E v i d e n c e for the flexible use of
nisms. In: G a z z a n i g a M , ed. T h e C o g n i t i v e N e u r o sciences. Boston, M A : M I T Press, 1994; 5 8 5 - 5 9 5 .
61. Miller EK, Erickson C A , D e s i m o n e R. N e u r a l m e c h a nisms of visual w o r k i n g m e m o r y in p r e f r o n t a l cortex of
the m a c a q u e . J N e u r o s c i 1996; 16: 5 1 5 4 - 5 1 6 7 .
62. M i z u m o r i SJY, C o o p e r B G , L e u t g e b S, Pratt W E . A
neural system analysis o f a d a p t i v e navigation. Mol
N e u r o b i o l 2 0 0 1 ; 21: 5 7 - 8 2 .
63. Morris R G M . Spatial localization d o e s not require the
p r e s e n c e o f local cues. Learn Motiv 1981; 12: 2 3 9 - 2 6 0 .
64. Morris R G M , G a r r u d P, R a w l i n s J N P , O ' K e e f e J. Place
navigation impaired in rats with h i p p o c a m p a l lesions.
Nature 1 9 8 2 ; 2 9 7 : 6 8 1 - 6 8 3 .
65. Muller RU. A quarter o f c e n t u r y of p l a c e cells. N e u r o n
1996; 17: 8 1 3 - 8 2 2 .
66. Muller RU, Bostock E M , T a u b e JS, K u b i e JL. O n the
directional firing properties of h i p p o c a m p a l place cells.
J N e u r o s c i 1994; 1 4 : 7 2 3 5 - 7 2 5 1 .
67. Muller RU, K u b i e JL. T h e e f f e c t s o f c h a n g e s in the
e n v i r o n m e n t on the spatial f i r i n g of h i p p o c a m p a l
c o m p l e x - s p i k e cells. J N e u r o s c i 1987; 7: 1951-1968.
68. Muller R U , K u b i e JL, R a n c k JB. Spatial firing patterns
of h i p p o c a m p a l c o m p l e x - s p i k e cells in a fixed e n v i r o n ment. J N e u r o s c i 1987; 7: 1935-1950.
69. Muller RU, Stead M, Pach J. T h e h i p p o c a m p u s as a
cognitive g r a p h . J G e n Physiol 1996; 107: 6 6 3 - 6 9 4 .
70. M u m b y D G , G l e n n MJ, N e s b i t t C, Kyriazis D A .
Dissociation in r e t r o g r a d e m e m o r y f o r object disc r i m i n a t i o n s and object recognition in rats with perirhinal cortex d a m a g e . B e h a v Brain Res 2 0 0 2 ; 132: 2 1 5 226.
71. N a d e l L. F o r u m : Is the h i p p o c a m p a l f o r m a t i o n preferentially involved in spatial b e h a v i o r ?
1991; 1: 2 2 1 - 2 9 2 .
Hippocampus
spatial k n o w l e d g e in the rat. P s y c h o b i o l o g y 1997; 2 5 :
294-302.
56. McClelland JL, M c N a u g h t o n BL, O ' R e i l l y RC. W h y
72. O ' K e e f e J. Place units in the h i p p o c a m p u s of the freely
m o v i n g rat. E x p N e u r o l 1976; 51: 7 8 - 1 0 9 .
73. O ' K e e f e J, C o n w a y D H . H i p p o c a m p a l place units in the
there are c o m p l e m e n t a r y learning s y s t e m s in the h i p p o c a m p u s and n e o c o r t e x : insights f r o m the successes and
freely m o v i n g rat: w h y they fire w h e r e they fire. E x p
Brain Res 1 9 7 8 ; 3 1 : 5 7 3 - 5 9 0 .
failures of c o n n e c t i o n i s t m o d e l s of learning and m e m o r y .
Psychol Rev 1995; 102: 4 1 9 - 4 5 7 .
74. O ' K e e f e J, C o n w a y D H . O n the trail o f the h i p p o c a m p a l
e n g r a m . Physiol Psychol 1980; 8: 2 2 9 - 2 3 8 .
57. M c l l u g h TJ, Blum Kl, Tsien JZ, T o n e g a w a S, Wilson
75. O ' K e e f e J, D o s t r o v s k y J. T h e h i p p o c a m p u s as a spatial
M A . Impaired h i p p o c a m p a l representation of space in
C A I-specific N M D A R 1 k n o c k o u t mice. Cell 1996; 87:
1339-1349.
58. M c N a u g h t o n BL, B a r n e s C A , Gerrard JL, G o t h a r d K,
J u n g M W , Knierim JJ, K u d r i m o t i H, Qin Y, S k a g g s W E ,
VOLUME 15. NO. 2. 2004
m a p . Preliminary e v i d e n c e f r o m unit activity
freely m o v i n g rat. Brain Res 1971; 34: 171-175.
in the
76. O ' K e e f e J, N a d e l L. H i p p o c a m p u s as a C o g n i t i v e M a p .
O x f o r d : C l a r e n d o n , 1978.
106
Β. POUCET ET AL.
77. O ' K e e f e
J,
Recce
hippocampal
ML.
Phase
relationship
p l a c e units and t h e E E G
Hippocampus
between
theta
rhythm.
C A I r e g i o n o f t h e h i p p o c a m p u s . Cell 1996; 87: 13511361.
95. R o t e n b e r g A , M u l l e r R U . V a r i a b l e p l a c e - c e l l c o u p l i n g to
1993;3:317-330.
78. O ' K e e f e J, S p e a k m a n A. S i n g l e unit activity in t h e rat
a c o n t i n u o u s v i e w e d s t i m u l u s : e v i d e n c e that the h i p p o -
h i p p o c a m p u s d u r i n g a spatial m e m o r y t a s k . E x p Brain
c a m p u s acts as a p e r c e p t u a l s y s t e m . Phil T r a n s R o y S o c
L o n d B 1997; 3 5 2 : 1 5 0 5 - 1 5 1 3 .
R e s 1987; 68: 1-27.
79. O l y p h e r A V , L a n s k y P, F e n t o n A A . P r o p e r t i e s o f the
extra-positional
signal
in
hippocampal
place
cell
d i s c h a r g e d e r i v e d f r o m t h e o v e r d i s p e r s i o n in location-
firing
o f h i p p o c a m p a l p l a c e cells in blind rats. J N e u r o -
sci 1998; 18: 1 8 1 8 - 1 8 2 6 .
97. S a v e E, N e r a d
s p e c i f i c firing. N e u r o s c i e n c e 2 0 0 2 ; 111: 5 5 3 - 5 6 6 .
80. P a c k a r d M G , M c G a u g h JL. Inactivation o f h i p p o c a m p u s
and c a u d a t e n u c l e u s w i t h l i d o c a i n e d i f f e r e n t i a l l y a f f e c t s
e x p r e s s i o n o f p l a c e a n d r e s p o n s e learning.
96. S a v e E, C r e s s a n t A , T h i n u s - B l a n c C, Poucet B. Spatial
Neurobiol
81. P a z - V i l l a g r ä n V , S a v e E, R i e h l e A , P o u c e t B. I n d e p e n -
hippo-
c a m p a l p l a c e cells. H i p p o c a m p u s 2 0 0 0 ; 10: 6 4 - 7 6 .
98. S a v e
E,
Poucet
exploration
Learn M e m 1996; 6 5 : 6 5 - 7 2 .
L, P o u c e t Β C o n t r i b u t i o n of m u l t i p l e
s e n s o r y i n f o r m a t i o n t o p l a c e field stability in
and
B,
Foreman
N,
Buhot
r e a c t i o n s to spatial
MC.
Object
non
spatial
and
c h a n g e s in h o o d e d rats f o l l o w i n g d a m a g e to parietal
dent c o d i n g o f c o n n e c t e d e n v i r o n m e n t s b y p l a c e cells.
c o r t e x or d o r s a l h i p p o c a m p u s . B e h a v N e u r o s c i
Soc Neurosci Abst 2003; 2 9 : 9 1 . 1 9 .
106: 4 4 7 - 4 5 6 .
82. P e a r c e
JM,
Roberts
ADL,
Good
M.
Hippocampal
lesions d i s r u p t n a v i g a t i o n b a s e d o n c o g n i t i v e m a p s but
99. S c h m a j u k N A , T h i e m e
of
R e v 1993; 100: 1 6 3 - 1 8 2 .
spatial
84. P o u c e t B. S e a r c h i n g f o r the spatial c o r r e l a t e s o f unit
in t h e p r e l i m b i c area of t h e rat m e d i a l f r o n t a l
c o r t e x . B e h a v Brain R e s 1997; 84: 151-159.
cognition in the rat. Crit Rev Neurobiol 1997; 11: 101-120.
86. P o u c e t B, C h a p u i s N , D u r u p M, T h i n u s - B l a n c C . A
s t u d y o f e x p l o r a t o r y b e h a v i o r a s an index o f spatial
1986;
14:
93-100.
Save
E,
Lenck-Santini
PP.
Sensory
and
2 0 0 0 ; 11: 9 5 - 1 I I .
visual
patterns
motion
of
information
hippocampal
on
place
the
cells.
J N e u r o s c i 1995; 15: 1 7 3 - 1 8 9 .
101. S h i b a t a R, M u l d e r A B , T r u l l i e r O , W i e n e r SI. Position
neurons
of
rats
complementary
alternating
types
of
between
spatial
tasks
cues.
requiring
Neuroscience
2001; 108:391-411.
102. S p e a k m a n A, O ' K e e f e J. H i p p o c a m p a l c o m p l e x
spike
cells d o not c h a n g e their p l a c e fields if the goal is m o v e d
hippocampus
of
rats. N e u r o R e p o r t
1994;
5:
2045-2048.
findings
firing
of hippo-
c a m p a l p l a c e cells in the d a r k d e p e n d s o n the r a t ' s r e c e n t
e x p e r i e n c e , J N e u r o s c i 1990; 10: 2 0 0 8 - 2 0 1 7 .
Neglected
aspects
of
with rats, m o n k e y s , and h u m a n s . P s y c h o l
Rev 1992; 9 9 : 1 9 5 - 2 3 1 .
104. T a b u c h i
invariant
89. Q u i r k G J , M u l l e r R U , K u b i e J L . T h e
MJ.
1990;2:544-555.
103. S q u i r e LR. M e m o r y a n d t h e h i p p o c a m p u s : a s y n t h e s i s
from
88. Poucet B, T h i n u s - B l a n c C , M u l l e r R U . Place cells in the
90. R e n n e r
and
firing
within a c u e c o n t r o l l e d e n v i r o n m e n t . Eur J N e u r o s c i
B,
m e m o r y p r o p e r t i e s o f p l a c e cells firing. Rev N e u r o s c i
ventral
vestibular
sensitivity in p h a s i c a l l y d i s c h a r g i n g n u c l e u s a c c u m b e n s
85. P o u c e t B, B e n h a m o u S. T h e n e u r o p s y c h o l o g y o f spatial
87. P o u c e t
and
100. S h a r p PE, Blair H T , Etkin D, T z a n e t o s D B . I n f l u e n c e s
t h e s e s o n their s t r u c t u r e and n e u r a l m e c h a n i s m s . P s y c h o l
k n o w l e d g e in h a m s t e r s . A n i m L e a r n B e h a v
Purposive behavior
c o g n i t i v e m a p p i n g : a n e u r a l n e t w o r k m o d e l . Biol C y b e r n
1992; 67: 1 6 5 - 1 7 4 .
not h e a d i n g v e c t o r s . N a t u r e 1998; 3 9 6 : 7 5 - 7 7 .
83. P o u c e t B. Spatial c o g n i t i v e m a p s in a n i m a l s : n e w h y p o -
firing
AD.
1992;
ET,
Mulder AB,
W i e n e r SI. R e w a r d
place responses and
reward
site
activity in h i p p o c a m p a l n e u r o n s o f b e h a v i n g rats. H i p p o c a m p u s 2 0 0 3 ; 13: 117-132.
105. T a u b e JS. H e a d direction cells a n d t h e
exploratory
and
i n v e s t i g a t o r y b e h a v i o r . P s y c h o b i o l o g y 1990; 18: 16-22.
value
associated
neurophysio-
logical basis f o r a s e n s e o f d i r e c t i o n . Prog N e u r o b i o l
1998; 55: 2 2 5 - 2 5 6 .
9 1 . Rivard B, Li Y , L e n c k - S a n t i n i PP, P o u c e t B, M u l l e r R U .
106. T a u b e JS, M u l l e r R U , R a n c k J B Jr. H e a d - d i r e c t i o n cells
R e p r e s e n t a t i o n of o b j e c t s in s p a c e by t w o c l a s s e s of
r e c o r d e d f r o m t h e p o s t s u b i c u l u m in f r e e l y m o v i n g rats.
h i p p o c a m p a l p y r a m i d a l cells. J G e n Physiol; in press.
I. D e s c r i p t i o n
92. R o s s i e r J, S c h e n k F, K a m i n s k y Y. B u r e s J. T h e p l a c e
p r e f e r e n c e task: a n e w tool f o r s t u d y i n g t h e
and
quantitative
analysis.
J
Neurosci
1990; 10: 4 2 0 - 4 3 5 .
relation
107. T a u b e JS, M u l l e r R U , R a n c k J B Jr. H e a d - d i r e c t i o n cells
b e t w e e n b e h a v i o r and p l a c e cell activity in rats. B e h a v
r e c o r d e d f r o m t h e p o s t s u b i c u l u m in f r e e l y m o v i n g rats.
Neurosci 2000; 114:273-284.
II. E f f e c t s of e n v i r o n m e n t a l m a n i p u l a t i o n s . J N e u r o s c i
93. R o t e n b e r g A , A b e l T, H a w k i n s R D , K a n d e l E R , M u l l e r
R U . Parallel instabilities o f l o n g - t e r m p o t e n t i a t i o n , p l a c e
1990; 10: 4 3 6 - 4 4 7 .
108. T h o m p s o n LT, Best PJ. L o n g - t e r m stability of the p l a c e -
cells, a n d l e a r n i n g c a u s e d b y d e c r e a s e d p r o t e i n k i n a s e A
field
activity. J N e u r o s c i 2 0 0 0 ; 2 0 : 8 0 9 6 - 8 1 0 2 .
h i p p o c a m p u s o f f r e e l y b e h a v i n g rats. Brain Res
94. R o t e n b e r g A , M a y f o r d M , H a w k i n s R D , K a n d e l
ER,
M u l l e r R U . M i c e e x p r e s s i n g a c t i v a t e d C a M K l l lack low
f r e q u e n c y L T P and d o not f o r m s t a b l e p l a c e cells in t h e
activity o f s i n g l e units r e c o r d e d f r o m the dorsal
1990;
509: 2 9 9 - 3 0 8 .
109. T r u l l i e r O , S h i b a t a R, M u l d e r A B , W i e n e r SI. H i p p o c a m p a l n e u r o n a l position s e l e c t i v e l y r e m a i n s
fixed
to
REVIEWS IN Ί Ί IE NEUROSCI I iNCI-S
NEURAL BASES OF NAVIGATION
room c u e s only in rats alternating b e t w e e n place
navigation and b e a c o n a p p r o a c h tasks. E u r J N e u r o s c i
1999; 1 1 : 4 3 8 1 - 4 3 8 8 .
110. Trullier O , W i e n e r SI, B e r t h o z A, M e y e r JA. Biologically based artificial n a v i g a t i o n systems: r e v i e w and
prospects. Prog N e u r o b i o l 1997; 51: 4 8 3 - 5 4 4 .
111. Wan H, A g g l e t o n JP, B r o w n M W . D i f f e r e n t contributions of t h e h i p p o c a m p u s and perirhinal cortex to
recognition m e m o r y . J N e u r o s c i 1999; 19: 1142-1148.
112. Wan HS, T o u r e t z k y DS, Redish A D . T o w a r d s a
c o m p u t a t i o n a l theory of rat navigation. In: M o z e r M,
S m o l e n s k y P, T o u r e t z k y DS, Elmau JL, W e i g e n d A, eds.
P r o c e e d i n g s of the 1993 C o n n e c t i o n i s t M o d e l s S u m m e r
School. N e w Y o r k : L a w r e n c e Erlbauin, 1994; 11-19.
I 13. W h i s h a w 1Q. Formation of a place learning set in the rat:
a n e w p r o c e d u r e for n e u r o b e h a v i o r a l studies. Physiol
Behav 1985; 26: 8 4 5 - 8 5 1 .
114. W i e n e r SI. Spatial, behavioral and sensory correlates
of h i p p o c a m p a l CA I c o m p l e x spike cell activity:
implications f o r i n f o r m a t i o n p r o c e s s i n g f u n c t i o n s . Prog
N e u r o b i o l 1996; 4 9 : 3 5 5 - 3 6 1 .
VOLUME 15. NO. 2. 2004
107
115. W i e n e r SI, B e r t h o z A, Z u g a r o M B . M u l t i s e n s o r y p r o c e s s i n g in the e l a b o r a t i o n o f p l a c e and head direction
responses by limbic s y s t e m n e u r o n s . C o g n Brain Res
2 0 0 2 ; 14: 75-90.
116. W i e n e r SI, Paul C A , E i c h e n b a u m Η. Spatial a n d
behavioral correlates of h i p p o c a m p a l n e u r o n a l activity.
J Neurosci 1 9 8 9 ; 9 : 2 7 3 7 - 2 7 6 3 .
117. Wilson M A , M c N a u g h t o n BL. D y n a m i c s of the h i p p o c a m p a l e n s e m b l e c o d e for space. S c i e n c e 1993; 2 6 1 :
1055-1058.
118. W o o d ER, D u d c h e n k o PA.
Η. H i p p o c a m p a l n e u r o n s
d i f f e r e n t types of m e m o r y
s a m e location. N e u r o n 2 0 0 0 ;
Robitseck RJ, E i c h e n b a u m
encode information about
e p i s o d e s o c c u r r i n g in the
27: 623-633.
119. W o r d e n R. N a v i g a t i o n by f r a g m e n t fitting: a theory o f
h i p p o c a m p a l f u n c t i o n . H i p p o c a m p u s 1992; 2: 165-188.
120. Y o u n g BJ, Fox G D , E i c h e n b a u m H. C o r r e l a t e s of
h i p p o c a m p a l c o m p l e x - s p i k e cell activity in rats p e r f o r m ing a non spatial radial m a z e task. J N e u r o s c i 1994; 14:
6553-6563.