HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (13) Citing Articles via Google Scholar Google Scholar Articles by Bures, J. Articles by Fenton, A. A. Search for Related Content PubMed PubMed Citation Articles by Bures, J. Articles by Fenton, A. A.

Neurophysiology of Spatial Cognition

Jan Bures and André A. Fenton

J. Bures and A. A. Fenton are in the Institute of Physiology, Academy of Sciences of the Czech Republic, Videnska 1083, 142 20 Prague 4, Czech Republic.

	Abstract

 
Understanding of the neurophysiology of cognition is advancing through the study of how animals navigate and understand space. Manipulating various classes of spatial information and recording from hippocampal neurons provides a robust model for understanding how the brain stores and constructs the spatial memories that are critical for organizing daily experience.

	Introduction

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
Close your eyes and answer this question: where are you? The answer is not so straightforward, a fact which itself reveals a complexity in how our brains organize, store, and retrieve experience. Let's assume you are at your desk and answer as such. But without changing location, an equally valid answer may be "five paces from the door" or "in the south of Prague." Quite naturally, each answer may have come from memory, since no direct perception of the desk, door, or Prague is necessary. Although each declaration is accurate, the appropriateness depends on the context of the question. This example indicates several important features of cognition that are particularly, but not exclusively, characteristic of spatial cognition. The perception and understanding of space fundamentally involves memory and the implicit construction of a reference or coordinate frame. There are, of course, several reference frames that one can hold in mind, none of which is universally appropriate. The brain clearly stores multiple memories, and to use them it must therefore organize and coordinate them in a way that preserves their integrity.

These commonplace considerations are especially important in light of the progress of research into the mechanisms of degenerative brain diseases (Alzheimer's, Parkinson's) that disturb processes of understanding or cognition. According to one classification, it is the declarative memory of Alzheimer patients that suffers first. Others have classified such memories as being episodic in that they are memories of events identified in time and space. Experimental animal models of cognition have therefore focussed on this class of memory. But is it at all possible to examine declarative or episodic memory in animals? A positive reply to this question requires the availability of nonverbal tests of declarative or episodic memory.

Since episodic memories are fundamentally located within a spatial context, memory research in animals (see Ref. 5 for a review) has focused on the domain of spatial cognition because of several properties. Although space itself cannot be directly perceived, spatial problems and thus spatial memories are of fundamental importance to mobile organisms. Spatial memories form and are necessarily organized within distinct reference frames that themselves must be coordinated. In this review, we first describe the different ways in which common spatial problems appear to be solved. We then summarize our recent animal experiments demonstrating that these different spatial strategies are functionally distinct, that they coexist, and that they depend on autonomous memories that can be studied at the cellular level.

	Modes of navigation

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
All mobile organisms are critically dependent on effective navigation, defined as the purposeful self-directed movement through an environment. Although "way following" is a legitimate spatial behavior, it is not considered a form of navigation since it does not usually allow the subject to take novel or optimal paths from one place to another. Particularly important is place navigation–finding an unmarked place the position of which can only be recognized by its relationship to perceptible distant landmarks or by the memory of the path leading to it from the current position of the animal. Some common examples of navigation behavior include exploring an environment to locate sources of food, dangerous spots, a sexual partner, or simply returning to the safety of home. Experiments using the Morris water maze (9) and other dry maze models of navigation have allowed the following set of navigation behaviors to be classified: 1) Cued navigation to a visible or otherwise directly perceptible target. In this situation the subject moves directly to the target (or the cue marking it) regardless of the target's position in the environment, a behavior that has no or little cognitive demand (Fig. 1A). 2) Idiothesis (8) is navigation to a place that was formerly visited, the position of which is determined by its distance and direction from the subject's current position. The vector to the target place can be computed by "path integration" or "dead reckoning," which is the process of integrating the subject's movements subsequent to visiting the target place (Fig. 1B1). This form of navigation does not necessarily depend on spatial information from the external world. It depends only on interoception, information derived from the subject's movements and hence on a self-centered (egocentric) reference frame. 3) Place navigation to a hidden goal, the position of which is determined relative to intramaze (Fig. 1B2) or extramaze (Fig. 1B3) landmarks that are not directly associated with the goal. In distinction from idiothesis, cued navigation and place navigation are examples of allothesis since they depend on exteroception, the perception of stimuli independent of the subject.

View larger version (40K): [in this window] [in a new window]   FIGURE 1. Different navigation strategies within a circular enclosure in a square room are depicted. The movement trajectory is shown as a vector. Spatial cues are provided by distal room cues outside the enclosure. Three are shown, along with a compass that indicates the inertial coordinate frame of the earth. There are also arena-bound proximal landmarks such as objects, smells, and textures that can be accessed. They are represented by A and B. A: a rat can learn to go directly from a start point in the south (•) to a goal () using distal or proximal allothetic cues or idiothesis. Here the goal is also visible (marked by an X), and thus the animal can also find the target by cued navigation, simply going to the X. B: when the arena is rotated 90°, it is possible to learn which navigation strategy the rat used. If the rat learned the target position by idiothetically computing the vector from the start at the arena periphery (B1), it will look for the goal at different room- and arena-defined places but always according to the vector 60° to the tangent at the start regardless of whether the start is in the room-defined south (black vector) or an arena-defined place 90° away (white vector). This behavior, to navigate using a learned series of movements, is a skill called praxis. Navigation learned according to cues on the arena (B2) will lead the rat to the goal between cues A and B regardless of the start position or the placement of the arena cue configuration. Similarly, navigation according to room cues (B3) will lead the rat to seek the goal in the northeast regardless of the start or the presence of the arena cues.

	Multiple representations of space

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

What is the purpose of such a redundant representation of space? In a stable world, any of them can serve as a backup for the missing one. Moreover, the identity of cued, allothetic, and idiothetic representations indicates the stability of the world. Whenever allothetic and idiothetic representations change their registration it indicates that the ground over which the animal walks moves with respect to more distant features of the environment. This is a situation commonly encountered when we travel on a train and can be illustrated by an animal moving in a fallen tree that floats in a pond.

	Investigating representations of space using navigation tasks

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
Given the existence of multiple simultaneous representations of space, the question arises as to how the brain coordinates these separate representations. Addressing this question in the laboratory requires methods to isolate the separate components of navigation and examine their specific features and interactions. The common analytic tools consist of blockades of sensory inputs and dissociation of the allothetic and idiothetic orientation modes (Fig. 2). Since most of the distant allothetic information is visual, darkness eliminates extramaze allothesis almost completely. On the other hand, darkness does not block allothetic information transmitted by the somatosensory system. The contours of the arena are an important source of allothetic information, which in combination with other tactile and olfactory cues on the walls or on the floor can give the animal full information about its position on the arena and in the room.

View larger version (53K): [in this window] [in a new window]   FIGURE 2. Foraging continuously, as in the place-preference and place-avoidance tasks, requires knowing one's position with respect to cues and a goal. Continuous navigation is depicted here (A1) like in Fig. 1A1. When learned in a stable situation in the light, navigation can be supported by knowing the relations among distal (room-anchored; black dashed lines) allothetic cues and between these cues and the current position and the place of interest as well as the relations among arena-anchored (white dashed lines) proximal allothetic cues and idiothetically derived arena-bound relations. Here and in the dark (A2), the rat walked a 90° arc. Various manipulations can eliminate or restrict the use of the particular navigation modes. Darkness eliminates the room-based (visual) cues, leaving only arena-based cues whose mutual distances and angles can be computed using idiothesis, which in turn must be supported by frequent correction by the arena cues. Rotating the arena in the room with the lights on and leaving the goal fixed in the room (B1) does not eliminate arena-based information but makes it unstable with respect to the goal and thus useless for locating it. Unless the rat can account for the arena movement, even idiothesis cannot help in localization. Alternatively, the room cues can be made irrelevant by keeping the target place fixed from the point of view of the rotating arena (not shown). When the goal rotates with the arena (B2), arena-bound local cues and idiothesis must be used to navigate. As shown here, rotation combined with darkness eliminates the otherwise useless visual room cues. Note that in B1 and B2, as in A, above, the rat walked a 90° arc, but when seen in the reference frame of the room, the rat will appear to move along a larger arc because of the passive movement caused by the arena rotation. Another way to make local arena cues navigationally useless is to "shuffle" the arena surface without passively moving the rat. When this is done in the light and the goal remains fixed in the room (C1), the animal can use room-bound information to locate the goal as well as to correct its idiothesis, which can also be used to find the goal. However, darkness and shuffling removes any means for the animal to update its error-prone idiothesis (C2), which is also the only way to locate the goal.

The rats are pretrained to collect 20-mg food pellets dropped from an overhead feeder to random locations on the surface of an elevated arena 80 cm in diameter. The rat's locomotor activity is recorded with a computerized tracking system. In the place preference task (1), the rat is conditioned to preferentially visit a place since the feeder delivers a pellet only when the animal enters a goal area, e.g., a 20-cm circle in the northeast quadrant of the arena. After the rat hears the click of the feeder and the impact of the pellet, it runs to find it and then returns immediately back to the conditioned place to release another pellet. The rat's behavior alternates between two modes: exploration of the whole arena during pellet retrieval and straight goal-directed place navigation during return to the reinforced area.

In the place avoidance task, the rat forages for pellets continuously but is conditioned to avoid a definite part of the arena (e.g., a northeast sector extending from the center of the arena toward the periphery). Entering this area is punished by mild foot shock (1). In this case, the rat learns to inhibit its search for pellets close to the punishing region.

	Dissociating representations of space using navigation during rotation and darkness

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
In either task, on a stable arena the conditioned place is physically the same from the point of view of the floor or the room (Fig. 2A), but as soon as the arena starts rotating they dissociate into distinct areas. What happens if the preferred or avoided place is defined with respect to the room, i.e., it stays in the northeast quadrant of the arena through which the arena floor is slowly moved (Fig. 2B1). In this case, only the allothetic information that is stable in the room provides information that is useful for defining the conditioned place. All idiothetic information, on the other hand, becomes misleading, and thus the idiothetic solution of the task is rapidly abandoned. This is demonstrated by the fact that after the light is switched off the animal is not able to solve the task, because it has no idiothetic backup that could support it when the room-anchored allothetic information is eliminated by darkness.

Something quite different happens if the conditioned place is defined with respect to a specific part of the arena floor that remains the same irrespective of the rotation (Fig. 2B2). In this case, switching the light off does not deteriorate the performance of the animal that has learned to solve the task without paying attention to remote (misleading) allothetic information. The system that tracks the rat's position must be modified for the rotating arena by providing two displays. One is a "room frame" display corresponding to the tracks recorded by a camera fixed in the room, for example above the center of the arena. This display combines the active locomotion of the animal through the room with its passive movement through the room that is caused by the rotating arena. The other is an "arena frame" display fixed with respect to the arena by using a point on the arena to stabilize the picture of the arena. Arena frame tracks are plotted with respect to this reference point as if viewed from an overhead camera attached to the rotating arena.

In the absence of useful remote allothetic cues, the animal may rely on idiothesis and local (allothetic) cues on the arena surface to navigate (Fig. 2B2). Such local cues are probably not used to directly mark the conditioned place; instead they are used to correct the animal's idiothesis. This is necessary since, unlike the allothetic navigation modes, idiothesis is prone to cumulative errors because without a stable allothetic reference, idiothetic localization must be based on each prior idiothetic judgment of position. Consider the sources of idiothetic information. The proprioceptive signals monitoring the movement of the legs, the vestibular canal receptors recording angular accelerations of the head, and the otolithic receptors sensitive to linear accelerations are continuously active and engaged in path integration computations. What is called "substratal idiothesis" is based on somatosensory and vestibular signals generated during active movement of the animal over the ground. The information received is highly interdependent: the number of steps per second can be directly counted but is at the same time monitored as acceleration by the vestibular accelerometer. Similarly, change of the direction of movement vectors is recorded by joint receptors, asymmetric muscle tonus, and the length of steps made by the left and right limbs, but the overall effect is reflected by angular acceleration perceived with the semicircular canals. What is called "inertial idiothesis" accounts for the perception of passive movement, e.g., when an immobile animal is subjected to linear or angular accelerations during transport. Both substratal and inertial idiothesis allow the animal to assess its position relative to stable surroundings, but an important difference between these two orientation modes is that in the first case vestibular perception is closely supported by tactile contact with the outer world, whereas it is more or less isolated in the second case.

It is difficult to completely remove intramaze cues since the animal may deposit feces and urine and other perceptible landmarks in the course of a session. It is, however, possible to render any arena-based cues navigationally useless by making them unstable with respect to each other and the goal (Fig. 2C). The presence of local olfactory or tactile cues is minimized in the water maze simply because the substrate for navigation, the water, is essentially homogeneous and turbulent. For tasks in which it is advantageous for the rat to locomote for many minutes, the arena surface can be mechanically shuffled. We use an arena constructed from two parts, a central disk and a surrounding belt of equal area (2). The disk and the belt can be rotated around the arena's central axis independently of each other. Thus when the rat is on the disk, the belt can be rotated some amount and when the animal is on the belt the disk can be rotated. This procedure causes any local cues on the disk to intermittently change their relations to cues on the belt. The arena and any surface-bound cues are thus intermittently shuffled without physically disturbing the rat. Thus removing or rendering useless remote allothetic cues by darkness or rotation forces the rat to rely on local cues and idiothesis; when the arena is shuffled in darkness, only idiothesis is left to direct purposeful spatial behavior. Accurate idiothesis under such conditions that prevent it's allothetic updating has been measured to last an average of only 8 m of locomotion, which corresponds to < 2 min for a foraging rat.

Whereas the above examples are extreme cases demonstrating the allothetic or idiothetic solutions of the task, in the interest of learning how the two are learned and interact it is also possible to train the animal to use both solutions, i.e., to avoid at the same time the allothetically and idiothetically defined loci. This "double avoidance" (4) indicates that the two orientation modes are mutually compatible, that they do not lead to an irresolvable conflict, but that under definite conditions they can support an adaptively useful behavior. What is more is that the double avoidance was revealed after the rats learned a place avoidance on a stable arena. Subsequent rotation, without further reinforcement, showed that rats had learned both the allothetic and idiothetic solutions supported by memories acquired and expressed independently.

This fact raises an important question. Does the brain store the two independent spatial memories in different brain systems or does the same neural system encode, coordinate, and manage the memories as separate? The information-processing mechanisms needed to support these two alternatives are quite distinct. We have used electrophysiological recordings of hippocampal "place cells" to try to address this question.

There are good reasons to address this question to the hippocampus. It is an anatomically well-defined structure in the medial temporal lobe that has long been known to be a part of a memory processing system. In particular, an extensive literature has confirmed that the rodent hippocampus is critical for optimal navigation and spatial memories. It is also a site of robust synaptic plasticity characterized by long-term potentiation and depression. Understanding hippocampal electrophysiology also has important clinical relevance in that it is a structure that is damaged in many forms of epilepsy and in Alzheimer's disease.

	Hippocampal neurons

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
A hippocampal place cell is a pyramidal cell in Ammon's horn that fires rapidly only when the rat is in a particular part of the current environment, which is called the cell's "firing field" or "place field" (Fig. 3A). Since each place cell has a unique place field and the collection of place fields cover the accessible space, it is natural to say that the accessible space is mapped onto the place cell population and thus to assume that the hippocampus contains a map of the current environment (Fig. 3B). In fact, it is the very discovery of the place cell phenomenon that prompted the suggestion (12) that the hippocampus is the neural substrate of the cognitive map that animals use to solve complex spatial problems like place navigation (13). In addition to the location specificity of hippocampal output neurons, the hippocampal interneurons also have a reliable spatial firing characteristic (Fig. 3C). They double their firing rate during locomotion and characteristically discharge everywhere but with a cell-specific, environment-specific spatial pattern.

View larger version (50K): [in this window] [in a new window]   FIGURE 3. A: examples of the firing fields of hippocampal place cells. These color-coded firing rate distributions were recorded from the same rat. Darker colors represent higher mean firing rate averaged over a 16-min session in which the rat collected randomly scattered food in a 76-cm diameter cylinder. The color key indicates median firing rates. Notice that different cells have different sized, shaped, and located firing fields. B: cell-specific fields of the four fields in A cover most of the environment's surface, as depicted by overlay. The gray regions of the cylinder were not "mapped" by these 4 cells, but surely some of the nearly 100,000 active hippocampal pyramidal cells would have been active in these areas if they had been recorded. The concept that the environment is mapped by place cells is also based on the observation that firing fields are environment-specific. C: top row shows the same cell recorded in 3 sessions, first while the rat foraged randomly in a cylinder, then on a low-walled circular platform, and then in the cylinder again. The bottom row shows the discharge pattern of a simultaneously recorded hippocampal interneuron (theta cell). The discharge pattern of the place cell in the cylinder is stable across sessions, even though it stopped firing when the rat was on the platform. Other place cells may have fired in different patterns, and cells that were silent in the cylinder may have become active. Hippocampal interneurons also have environment-specific (and cell-specific) firing patterns, although they are not as location specific as place cells. Theta cells are always active, causing powerful inhibition within the hippocampus, which is likely responsible for why pyramidal cells tend only to fire in their firing fields and in most environments not at all.

If place cells reflect the neural processing that underlies spatial cognition, then they should behave in ways that correspond to the rat's spatial behavior. For example, when the rat behaves as if the environment is different, then place cells should remap. However, place cells can also remap the same environment when the animal's task is changed (7). Thus a stronger test of whether place cell activity underlies spatial cognition would be if they did not remap despite changes in the environment as long as the rat continues to navigate successfully.

	Relationship of place cell firing and behavior

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
The use of rotation to dissociate the reference frame of the room and the arena allows us to test this hypothesis. The prediction is that when the room-anchored and arena-anchored spatial reference frames are dissociated by rotation of the arena, place cells involved in spatial cognition should continue to discharge in their firing fields in the same reference frame as the rat's purposeful spatial behavior. Recall that in the double-avoidance experiment rats were first trained to avoid a place on a stable arena, and only later when the arena was rotated was it revealed that they had encoded and were retrieving avoidance memories based in both the room and arena reference frames. Thus we expected that rats trained to forage for randomly scattered food would have stable firing fields in either or both the room or arena reference frames during rotation of the arena. Rats were trained over many days to retrieve randomly scattered food on a stable arena. Then place cells were recorded over 30 min without removing the rat from the arena. The recording was divided into three 10-min phases by the condition of the arena, as follows: 1) stable, meaning that the arena was in the standard condition; 2) rotation, meaning that the arena was rotated continuously at 1 rpm; and 3) stable, meaning that the rotation was stopped. Normal place cells were found in the first stable phase. However, the discharge of most of these cells became spatially disorganized during the rotation (1). There was no remapping; instead, the firing was dissipated across the arena. This was not a permanent change, since just stopping the rotation caused the firing fields observed before the rotation to return. This result can be understood to indicate that the relations between room and arena cues combined to determine firing fields, since only the relations between room and arena cues were disrupted by the rotation. But what does it mean for the relationship between place cell firing and the rat's behavior?

In light of the fact that the same rotation does not disturb spatial behavior, is it possible to conclude that place cell discharge does not contribute to spatial cognition? We must be cautious, since in the random foraging task the rat was not obliged to attend to its spatial surroundings and there was no way to know whether the rat's spatial behavior was changed by the rotation. Perhaps the animals were just as disoriented because their place cell firing was spatially disorganized. This points out an important consideration in evaluating place cell experiments. Just because the activity of place cells is unchanged, or even if it is changed, unless direct parallel evaluation of the animal's behavior is possible, the relationship between the effects on place cells and the subject's cognition or memory cannot be made with any certainty.

The double-avoidance experiment, which showed that rats could navigate on a rotating arena in both the frames of the room and the arena, provides a behavioral framework to directly test whether intact place cell activity is necessary for navigation. Since the arena rotation did not disturb the rat's ability to avoid the arena- and room-defined places, on the hypothesis that place cells are involved in the computations that underlay spatial cognition it is expected that place cell discharge will not be disturbed by the rotation. Furthermore, it is expected that place cell activity will then be defined in both the stable reference frame of the room and the rotating reference frame of the arena.

At first glance this is a difficult requirement to meet. How could a cell's discharge be restricted to a discrete firing field in the room frame and, within the same session, be restricted to a firing field defined by the rotating arena? One possibility is that place cells that are characterized on the stable arena really constitute two classes, one that discharges in the framework of the arena and the other that discharges in the framework of the room. It has already been demonstrated that hippocampal pyramidal cells discharge in distinct reference frames within a complex environment (6). Rats were trained to leave a box and go to a landmark to find food and then return to the box. Both the box and the goal landmark could be moved. The start position of the box was changed once the rat left it so that the rat returned to the box in a different place. While the rat was in the box, the landmark was moved. About half (45%) of the cells with a behavioral correlate discharged like classic place cells; they fired in stable relation to the framework of the room. The remaining cells were either "landmark goal" cells, which discharged in fixed relation to the landmark that cued the goal, "box" cells, which discharged in stable relation to the box, or "box in" or box out" cells, which fired in stable relation to entering or leaving the box, respectively.

	Theories of hippocampal representations

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 
These results beg the question, does it make sense to consider place cell discharge to be fundamentally organized in frameworks? After all, how many reference frames are there in an environment? Eichenbaum and colleagues (14) have argued that in fact hippocampal neurons discharge in firing fields because the firing field is the place where there is a conjunction of the specific stimuli that excite the cell. Since individual place cells are controlled by arbitrary, cell-specific subsets of the spatial stimuli in an environment, the hippocampus therefore encodes a representation of the world that is fundamentally a collection of the combinatorial relationships among stimuli. Thus in environments composed of several salient cues, one place cell will be controlled by the relations between a subset of these stimuli, whereas another place cell will be controlled by the relations among a different subset of these stimuli. The strongest evidence for this view comes from recent cue-controlled experiments, in which place cells were recorded while rats solved an four arm plus maze task requiring them to retrieve food from the end of each arm before revisiting an arm (14). There were four visual cues on a curtain that enclosed the maze, and each maze arm had a unique texture and smell. In the critical "double-rotation" manipulation, the curtain cues were together rotated by 90° and the maze cues were rotated in the opposite direction by 90°. The authors report that 28% of place cells followed the curtain cues, 15% followed the maze cues, 14% stayed fixed to the "background" laboratory frame, and 43% remapped. Additional manipulations support the basic idea that individual place cells were not controlled by the overall spatial organization of the environment; instead, there were arbitrary cell-specific features of the environment that controlled each cell.

This opinion of place cells is in contrast to the "environment mapping" view derived from experiments that show place cells to be controlled by "multiple replaceable stimuli." As originally stated by O'Keefe and Nadel (13), the hippocampus encodes a map fundamentally based on the global organization of spatial relations in the environment. In short, the hippocampus encodes the spatial framework, the geometry of an environment within which the experience of events can presumably be stored. The first such evidence comes from an experiment in which place cells were recorded in a cue-controlled environment and the principal spatial information in a plus maze was provided by four prominent items (11). Place cells throughout the maze continued to discharge in a fixed relation to the constellation of cues when the cues were displaced around the maze by 90° or 180°. The fields were also stable after any one or any pair of these stimuli was removed. In simpler environments, in which the only orienting cue was a card on the wall of a cylinder, angular displacements of the card along the wall caused equal displacements of firing fields, but removing the card (10) or turning out the lights did not disturb the fields if the rat was in the environment before the salient stimulus was made invisible (1). Note, however, that not just any spatial cue can control place cell firing. It was found that three large objects centrally located in a cylindrical arena could not control firing field locations, whereas the same objects did control place cell discharge when they were moved to the periphery of the apparatus (3).

Thus it is not at all clear which of these viewpoints is correct. Does the hippocampus encode arbitrary relationships such as the current position of the subject, or does it organize and encode something more abstract, a framework of spatial (and other) relations that the subject uses to store and organize experience for subsequent retrieval and manipulation?

This question is at the heart of the current investigation of spatial cognition in the rodent. Recall that in humans, the hippocampus seems to be critical for episodic memories, but this has been difficult to test directly in rodents because there are as yet no episodic memory tasks that allow the encoding of an experience to be dissociated from the retrieval of its memory. The cognition of space and the purposeful spatial behavior of animals provide a set of concrete paradigms and tools to lead this investigation.

	References

Top Introduction Modes of navigation Multiple representations of... Investigating representations of... Dissociating representations of... Hippocampal neurons Relationship of place cell... Theories of hippocampal... References

 

1. Bures J, Fenton AA, Kaminsky Y, Rossier J, Sacchetti B, and Zinyuk L. Dissociation of exteroceptive and idiothetic orientation cues: effect on hippocampal place cells and place navigation. Philos Trans R Soc Lond B Biol Sci 352: 1515–1524, 1997.[Abstract/Free Full Text]
2. Bures J, Fenton AA, Kaminsky Y, Wesierska M, and Zahalka A. Rodent navigation after dissociation of the allocentric and idiothetic representations of space. Neuropharmacology 37: 689–699, 1998.[Web of Science][Medline]
3. Cressant A, Muller RU, and Poucet B. Failure of centrally placed objects to control the firing fields of hippocampal place cells. J Neurosci 17: 2531–2542, 1997.[Abstract/Free Full Text]
4. Fenton AA, Wesierska M, Kaminsky Y, and Bures J. Both here and there: simultaneous expression of autonomous spatial memories. Proc Natl Acad Sci USA 95: 11493–11498, 1998.[Abstract/Free Full Text]
5. Gallistel CR. The Organization of Learning. Cambridge, MA: MIT, 1990.
6. Gothard KM, Skaggs WE, Moore KM, and McNaughton BL. Binding of hippocampal CA1 neural activity to multiple reference frames in a landmark-based navigation task. J Neurosci 16: 823–835, 1996.[Abstract/Free Full Text]
7. Markus EJ, Qin Y, Leonard B, Skaggs WE, McNaughton BL, and Barnes CA. Interactions between location and task affect the spatial and directional firing of hippocampal neurons. J Neurosci 15: 7079–7094, 1995.[Abstract]
8. Mittelstaedt ML and Mittelstaedt H. Homing by path integration in the mammal. Naturwissenschaften 67: 566–567, 1980.[Web of Science]
9. Morris RGM. Spatial localisation does not depend on the presence of local cues. Learning Motivation 12: 239–261, 1981.
10. Muller RU and Kubie JL. The effects of changes in the environment on the spatial firing of hippocampal complex-spike cells. J Neurosci 7: 1951–1968, 1987.[Abstract]
11. O'Keefe J and Conway DH. Hippocampal place units in the freely moving rat: why they fire where they fire. Exp Br Res 31: 573–590, 1978.[Web of Science][Medline]
12. O'Keefe J and Dostrovsky J. The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely moving rat. Brain Res 34: 171–175, 1971.[Web of Science][Medline]
13. O'Keefe J and Nadel L. The Hippocampus as a Cognitive Map. Oxford, UK: Clarendon, 1978.
14. Shapiro ML, Tanila H, and Eichenbaum H. Cues that hippocampal place cells encode: dynamic and hierarchical representation of local and distal stimuli. Hippocampus 7: 624–642, 1997.[Web of Science][Medline]
15. Wilson MA and McNaughton BL. Dynamics of the hippocampal ensemble code for space. Science 261: 1055–1058, 1993.[Abstract/Free Full Text]

This article has been cited by other articles:

				J. J. Siegel, J. P. Neunuebel, and J. J. Knierim Dominance of the Proximal Coordinate Frame in Determining the Locations of Hippocampal Place Cell Activity During Navigation J Neurophysiol, January 1, 2008; 99(1): 60 - 76. [Abstract] [Full Text] [PDF]

				T. Bast, B. M. da Silva, and R. G. M. Morris Distinct Contributions of Hippocampal NMDA and AMPA Receptors to Encoding and Retrieval of One-Trial Place Memory J. Neurosci., June 22, 2005; 25(25): 5845 - 5856. [Abstract] [Full Text] [PDF]

				J. M. Cimadevilla, M. Wesierska, A. A. Fenton, and J. Bures Inactivating one hippocampus impairs avoidance of a stable room-defined place during dissociation of arena cues from room cues by rotation of the arena PNAS, March 13, 2001; 98(6): 3531 - 3536. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (13) Citing Articles via Google Scholar Google Scholar Articles by Bures, J. Articles by Fenton, A. A. Search for Related Content PubMed PubMed Citation Articles by Bures, J. Articles by Fenton, A. A.