Spatial maps

location of Spatial maps in the brain

new research indicates that space is represented in several brain systems, each hosting a variety of representations involving functionally specialized cell types. These systems span wide regions of cortex including the hippocampus, entorhinal cortex, pre- and parasubiculum, retrosplenial cortex, parietal cortex, frontal cortex and other areas.

Place cells (hippocampal maps)

Place cells are hippocampal pyramidal cells that fire when the animal is at specific positions in the environment. Different place cells fire at different positions. Place cells are active both in light and dark, suggesting that a single modality such as vision is not responsible for their positional firing.

The brain can read out the activity of a local population of place cells to determine the position of the rat in the box. In experiments where activity is recorded from a large number of cells, the position of the rat can be reconstructed with considerable accuracy [13], indicating that the population of place cells forms a spatial map of the environment. This map could be innate, as is suggested by the fact that place cells are present as early as postnatal day 15 (P15) soon after young rats (pups) open their eyes

Place cells can be shown to encode conjunctions between spatial and olfactory information when this is relevant for the task.

When rats are rewarded in an odor discrimination task, in which they need to recognize odors that do not match a previously presented one, hippocampal neurons can encode not only position but also odor and training rule.

In some situations, the cells encode time intervals rather than positions, such as when the animal is using time to guide its behavior.

The capacity to associate locations with particular experiences could be useful for episodic memory encoding in the hippocampus

Grid cells

grid cells are position-selective neurons in the medial entorhinal cortex (MEC), abundant also in the presubiculum and the parasubiculum. Grid cells are characterized by multiple firing locations that, in an open-field arena, collectively form a hexagonal grid over the entire space available to the animal. Grid cells can differ from each other in their grid spacing, grid phase and grid orientation. It is known that the spacing of grid cells increases along the dorsal–ventral axis of the entorhinal cortex. Similar to place cells, grid cells can be used to reconstruct the position of the rat in the environment and so function as a map of the animal’s position.

Entorhinal maps

The same parahippocampal brain regions that accommodate grid cells also contain two additional types of cells of potential relevance for spatial mapping: head-direction cells and border cells

• Head-direction cells are cells that respond only when the animal is facing a specific azimuth; different head-direction cells are tuned to different allocentric orientations. All directions are represented equally in the cell population. Head-direction cells were discovered in the presubiculum but were later found also in MEC, as well as in several other brain regions

• Border cells respond when the animal is near a boundary of the local environment. Boundary-related cells have been recorded also in the subiculum (Sub), which indirectly links the feedback from CA1 to the MEC, the presubiculum and the parasubiculum

Grid, head-direction, and border cells might have strong innate components, given that rudiments of all three cell types are present when rat pups explore open spaces for the first time between P15 and P20. Together, these cell types could be part of a metric navigation system able to map distances (grid cells), directions (head-direction cells), and vicinity to boundaries (border cells).

Relation between entorhinal and hippocampal maps

Anatomy of the hippocampal formation

The connections between the neocortex, the parahippocampal regions (PHR) and the hippocampal formation (HF).

The neocortex is connected to the hippocampus mainly via two pathways through the parahippocampal cortex.

→ One projects through the perirhinal cortex (PER) and the lateral entorhinal cortex (LEC)

→ the other projects through the postrhinal cortex (POR) and the medial entorhinal cortex (MEC).

Cells that carry information about the position of the animal, such as grid cells, head-direction cells, and border cells, are found in MEC but not in LEC. MEC and LEC project to the same regions in the hippocampus, both via direct projections to each hippocampal subfield and via the indirect trisynaptic circuit through dentate gyrus and CA3. While axons from MEC and LEC to dentate gyrus and CA3 tend to target the same cells, connections to CA1 are split, such that MEC is linked preferentially to the proximal part of CA1, and LEC preferentially to the distal part. This differential connectivity leads to stronger spatial modulation in proximal than distal CA1. The arrow from CA3 to itself stresses the abundance of recurrent connections within area CA3. Signals are routed back from CA1 to the entorhinal cortex either via direct projections, or via the subiculum (Sub), the presubiculum or the parasubiculum.

The hippocampal formation is located deep in the medial temporal lobe of the brain, on both sides of the head. The hippocampal formation includes several tightly connected structures: the dentate gyrus, CA3, CA2, CA1, and the subiculum. Immediately adjacent to it lies the entorhinal cortex, which serves as the main interface between the hippocampus and the rest of the cerebral cortex.

The broader region surrounding this system is called the parahippocampal region. This region includes the perirhinal cortex, the postrhinal cortex, and the entorhinal cortex, which itself is divided into medial (MEC) and lateral (LEC) parts.

The neocortex is the outer layer of the brain and is responsible for processing sensory information, perception, language, reasoning, and complex thought. However, visual features are processed in one region, sounds in another, objects in another, etc.

Information from the neocortex passes through the parahippocampal region via two partially separate pathways.

• One pathway runs through the perirhinal cortex (PER) and then into the lateral entorhinal cortex. The perirhinal cortex is especially involved in object recognition and familiarity judgments. The lateral entorhinal cortex (LEC) then receives this object-based and contextual information. Neurons in the LEC tend to represent information about specific items, events, and aspects of temporal context. This pathway therefore carries primarily non-spatial information.

• The second pathway runs through the postrhinal cortex (POR) and then into the medial entorhinal cortex (MEC). The postrhinal cortex processes spatial layout and environmental context. The medial entorhinal cortex contains specialized neurons such as grid cells, head-direction cells, and border cells. Together, these neurons provide precise spatial information about location and orientation. This pathway therefore carries spatial information.

Episodic memory requires both knowing what happened and knowing where it happened. The lateral pathway emphasizes the “what,” and the medial pathway emphasizes the “where.” The hippocampus integrates these streams so that objects and events can be embedded within spatial and contextual frameworks.

From the entorhinal cortex, information enters the hippocampus through two routes.

• One route is direct, with neurons from the entorhinal cortex projecting straight to CA1.

• The other route is indirect and is known as the trisynaptic circuit. The trisynaptic circuit begins in the entorhinal cortex, projects to the dentate gyrus, then to CA3, and finally to CA1. It is called trisynaptic because there are three major synaptic steps in this pathway.

  • The dentate gyrus consists of densely packed granule cells that are highly selective in their firing. Its primary computational role is pattern separation. When two experiences are similar the cortical input patterns that reach the hippocampus will overlap. The dentate gyrus reduces this overlap by activating only a small and distinct subset of neurons for each input, making memories more distinct from one another.

  • CA3 is composed of pyramidal neurons that are heavily interconnected with one another through recurrent collateral connections. This means that each neuron can influence many others within the same region. Because of this CA3 functions as an autoassociative network. Its primary computational role is pattern completion. If only part of a previously stored memory is activated, CA3 can reactivate the full stored pattern by spreading activity through its recurrent connections. This is particularly important during memory retrieval, when partial cues must trigger complete recollection.

  • CA1 receives input from both CA3 and directly from the entorhinal cortex. This dual input allows CA1 to compare stored memory representations (coming from CA3) with current sensory input (coming directly from entorhinal cortex). It plays a crucial role in novelty detection, updating memories, and refining stored representations.

Importantly, projections from the medial and lateral entorhinal cortex are not evenly distributed in CA1. Inputs from the medial entorhinal cortex preferentially target the proximal part of CA1, while inputs from the lateral entorhinal cortex preferentially target the distal part. This anatomical difference results in stronger spatial coding in proximal CA1 and relatively stronger non-spatial coding in distal CA1. This means that even within CA1, spatial and non-spatial information remain partially organized.

After processing in CA1, information exits the hippocampus. It travels either directly back to the entorhinal cortex or indirectly through the subiculum and related structures such as the presubiculum and parasubiculum. From there, information is sent back out to widespread regions of the neocortex. During memory consolidation, especially during sleep, this loop supports the gradual integration of hippocampal memory traces into long-term cortical storage.

Allocentric coordinates

World-based coordinates. Opposite of egocentric coordinates

Attractor dynamics

Attractor networks are neural networks with one or more stable states. These stable states are determined by the strengths of the recurrent connections between the individual neurons of the network. When the system is started from a location in state space other than the stable state, it will evolve until it arrives at one of the stable states and will then tend to stay there.

Azimuth

Orientation relative to world coordinates.

Fourier-like transform

A classical mathematical Fourier-transform transforms between the description of a signal in position terms to a description of the signal in spatial-frequency terms. The Fourier transform of a sine wave is a narrow pulse. In two dimensions, this resembles the transformation from grid cells (which are oscillatory sine waves) to place cells (which resemble a narrow pulse in space).

Grid orientation

The orientation of the grid pattern relative to an external reference orientation.

Grid phase

The position of the vertices of a grid cell in the x–y plane.

Grid spacing

The distance between adjacent grid vertices, expressed as the average distance from the central peak to any of the vertices of the inner hexagon in the spatial autocorrelogram.

Path integration

Position is the integral of velocity in time. Therefore one way to determine the current position is to sum up all momentary velocities and directions until the current moment. Path integration is a method to determine one’s position from one’s own self movement in this manner, without relying on external landmarks.

Reference frame

An external configuration of landmarks and geometry to which neuronal firing coordinates are associated

Theta rhythm

A dominant regular 8 Hz rhythm recorded in local-field potential signals in many brain areas, such as the hippocampus, the entorhinal cortex and the septum. Neuronal firing of individual cells is phase-modulated Corresponding authors: Derdikman, D. ( by this population rhythm. It occurs mostly when the rat is in movement.

remapping

The same population of place cells can encode or retrieve different maps in different environments or different configurations of the same environment [17–19]. The process of switching to a different map is called remapping

Development of spatial maps

rudiments of the brain’s spatial representation system are present at the time when rat pups make their first navigational movements at approximately 2.5 weeks of age. The pups were implanted with tetrodes on P13, before the eyelids unsealed, and recordings were made a few days later when the animals explored an environment outside the nest for the first few times. Both studies show that a rudimentary brain map of space is present from the first day of outbound movement. The most rapidly developing component of the brain map is the directional representation in the pre- and parasubiculum. Strong directional tuning was apparent already at P15 and P16 when activity was recorded from these regions for the first time. The proportion of direction-tuned cells was similar to that of adult rats, and the degree of directional tuning was not different.

Young animals exploring an open space for the first time also had place cells. The number of place cells in CA1 was only slightly lower at P16–P18 than at older ages, although the spatial tuning and stability of the cells continued to show some development.

Finally, young rats also had rudiments of grid cells. The number of grid cells was lower at P16–P19 than in older rats and the periodic structure of the grid fields was weaker than in the adults. The number of grid cells, and their spatial periodicity, reached adult levels during the first week or two after the onset of navigation. Therefore it seems that a rudimentary map of cells with directional and spatial firing correlates is present when animals navigate the outside world for the first time. Although head-direction cells, place cells and grid cells show slightly different developmental profiles, these cells, or their predecessors, might interact from the outset. The adult-like representation of direction in pre- and parasubiculum in the youngest animals could guide the development of spatial representations in entorhinal cortex and hippocampus, and rudimentary grid cells in entorhinal cortex might provide sufficiently patterned input to the hippocampus to generate place-specific responses in the hippocampal areas. It could also be that in young pups place cells are constructed by a larger proportion of head-direction cells and border cells, and that the contribution of grid cells to the construction of place cells grows with age. The evolution of functional intrinsic connections in MEC during the fourth week could be an essential component for the generation of a combined entorhinal–hippocampal representation of space.

1⃣ What Is Being Studied?

The researchers were studying how the brain’s spatial representation system develops in rat pups.

The spatial representation system is the network of brain regions that allow an animal to know:

• Which direction it is facing

• Where it is located

• How far it has moved

• What the structure of the surrounding environment looks like

This system primarily involves:

• The hippocampus

• The entorhinal cortex (especially the medial entorhinal cortex, MEC)

• The presubiculum and parasubiculum

These areas are all located in the medial temporal lobe.

The researchers implanted tetrodes into rat pups on postnatal day 13 (P13). A tetrode is a recording device made of four closely spaced electrodes that can record the activity of individual neurons. At P13, rat pups’ eyelids are still closed, meaning they have not yet had visual experience of the external world. A few days later, when the pups began exploring outside the nest for the first time (around P15–P16), neuronal recordings were taken.

This allowed researchers to ask a crucial question:

Is the spatial mapping system built from experience, or is it already present before experience?

---

2⃣ What Was Found?

The major finding was that rudiments — meaning early basic forms — of the spatial map were already present at the very first moment of navigation.

This means the brain does not start from zero when an animal begins exploring. Instead, core components are already in place.

Let us now examine each cell type.

---

3⃣ Head-Direction Cells (Pre- and Parasubiculum)What are head-direction cells?

Head-direction cells are neurons that fire when the animal’s head is facing a specific direction in space, regardless of location. For example, one cell may fire whenever the animal faces north, another when facing east.

They function like a neural compass.

Where are they located?

In this study, they were recorded in the presubiculum and parasubiculum, which are regions located between the hippocampus and entorhinal cortex. These areas are strongly connected to the medial entorhinal cortex.

What did the study show?

By P15–P16 — the very first days of navigation — these head-direction cells were already strongly tuned.

The proportion of direction-tuned cells was similar to adult rats.
The precision of their tuning was also similar.

Why is this important?

It means directional coding is essentially mature from the outset.

This suggests that the directional system may be genetically preconfigured and may not require visual experience to develop properly.

What does this tell us?

Direction coding may serve as the foundational scaffold for building the rest of the spatial system.

Because if you know which way you are facing, you can anchor spatial representations around that stable reference.

---

4⃣ Place Cells (CA1)What are place cells?

Place cells are neurons in the hippocampus (especially in CA1 and CA3) that fire when the animal is in a specific location in the environment.

For example, one neuron may fire when the rat is in the upper-left corner of a room.

Together, many place cells form a cognitive map.

Where are they located?

Primarily in CA1 and CA3 of the hippocampus.

What did the study show?

Even at P16–P18, during the first explorations:

• Place cells were already present.

• The number of place cells was only slightly lower than in adults.

• However, their spatial tuning and stability were still developing.

This means:

• The cells fired in location-specific ways.

• But their firing fields were less precise.

• And their maps were less stable over time.

Why does this matter?

It shows that the hippocampus is capable of generating spatially specific representations almost immediately.

However, the refinement of those representations requires further maturation and experience.

---

5⃣ Grid Cells (Medial Entorhinal Cortex)What are grid cells?

Grid cells are neurons in the medial entorhinal cortex that fire in multiple locations arranged in a hexagonal grid pattern across the environment.

Unlike place cells, which fire in one location, grid cells fire in many evenly spaced locations.

They are thought to provide a metric coordinate system for space.

Where are they located?

In the medial entorhinal cortex (MEC), which sits just outside the hippocampus.

What did the study show?

At P16–P19:

• Grid cells were present.

• But there were fewer of them than in adults.

• Their hexagonal structure was weaker.

• Their spatial periodicity was less precise.

Within 1–2 weeks after navigation began, grid cells reached adult-like levels.

Why is this interesting?

Grid cells develop more slowly than head-direction cells.

This suggests that:

• Direction coding may guide grid cell development.

• Intrinsic network maturation within MEC may be required.

---

6⃣ Developmental Order: Why This Sequence?

The developmental sequence appears to be:

1. Head-direction cells mature first.

2. Place cells are present early but refine gradually.

3. Grid cells mature more slowly.

Why might this be?

Directional coding is computationally simpler and may depend on subcortical systems (vestibular and angular velocity inputs) that mature early.

Grid cells, however, depend on:

• Recurrent connectivity within MEC.

• Path integration.

• Oscillatory dynamics.

The passage specifically mentions that intrinsic connections within MEC mature during the fourth week. This suggests grid formation depends on network-level maturation rather than just sensory input.

---

7⃣ How Do These Systems Interact?

The passage proposes several models:

Model 1: Direction Guides Grid and Place

Because head-direction cells are adult-like early on, they may provide stable orientation signals that help organize spatial coding in entorhinal cortex and hippocampus.

Model 2: Early Grid Inputs Generate Place Cells

Even weak grid-like patterns may be sufficient to generate place-specific firing in hippocampus.

Model 3: Place Cells Initially Depend More on Head-Direction and Border Cells

In early development:

• Place cells may rely more heavily on head-direction cells.

• Border cells (cells that fire near boundaries) may help define spatial anchors.

• As grid cells mature, they increasingly contribute.

This suggests the spatial system is interactive from the outset, rather than sequentially assembled.

---

8⃣ Why Is This Real and Important?

This research demonstrates that spatial coding is not purely learned from experience.

Instead:

• Core spatial representations are present before extensive exploration.

• The brain contains preconfigured spatial circuitry.

• Experience refines rather than creates the system.

This supports the idea that spatial navigation is evolutionarily fundamental.

---

9⃣ What Does This Tell Us About Humans?

Humans have homologous structures:

• Hippocampus

• Entorhinal cortex

• Presubiculum

Human neuroimaging shows grid-like signals in entorhinal cortex.

This developmental research suggests:

1. Humans may be born with preconfigured spatial scaffolding.

2. Basic orientation and place coding likely develop very early in infancy.

3. Spatial cognition may form a foundation for episodic memory development.

Since episodic memory depends on spatial context, early maturation of spatial coding may scaffold autobiographical memory later in life.

It also has implications for:

• Developmental disorders

• Early hippocampal damage

• Alzheimer’s disease (which first affects entorhinal cortex)

---

🔬 Big Theoretical Implication

The key theoretical insight is that:

The entorhinal–hippocampal system is not constructed from experience alone. It is a partially innate dynamical system that becomes refined through interaction with the environment.

Directional coding appears foundational.
Grid coding requires network maturation.
Place coding emerges early but stabilizes with development.

Together, they form a progressively refined cognitive map.

remapping

Changes in the environment cause consistent changes in spatial maps in the hippocampus and the entorhinal cortex. These transformations are referred to as remapping [18]. Three main types of transformations can be considered: 

Place cell deformations: Squeezing or stretching the environment can cause a systematic move of the position of the place fields relative to the surrounding boundaries (Figure Ia). Such deformation was reported in 2-D boxes [50] but also along linear tracks [65]. In situations where place cells undergo spatial deformation, grid cells deform too [91]. 

Rate remapping: In some cases, environmental transformations cause a dramatic change in the distribution of firing rates among place cells without an accompanying change in firing positions (Figure Ib). An example of a manipulation of the environment that induces rate remapping is to change the wall color of the box the rat is in from black to white [92]. When place cells undergo rate remapping, simultaneously recorded grid cells do not change their firing in a consistent way [68].

Global (place) remapping: In some cases, environmental changes can change the position of the place fields in an unpredictable way. For example, when the rat is walking in a box in one room (room A) and then in a similar box in another room (room B), the positions and firing rates of the different place cells are apparently unrelated [93,94] (Figure Ic). Global remapping might also occur in the same location when the geometry or other salient properties of the environment change radically [18,68]. When place cells undergo global remapping, the firing vertices of grid cells undergo changes such as shifts in grid phase, grid orientation or grid scale [68]. Place cells do not preserve distance information during global remapping whereas grid cells do; that is, two place cells which had adjacent place fields in one environment might have very distant place fields in a second environment, whereas two grids with similar spacing will shift together such that the spatial phase relationships between the grid fields are conserved [30,68].

1⃣ What Is Remapping?

Remapping refers to systematic changes in spatial firing patterns in the hippocampus and entorhinal cortex when the environment changes.

To understand this, recall:

• Place cells in the hippocampus fire in specific locations.

• Grid cells in the medial entorhinal cortex fire in repeating hexagonal patterns across space.

Together, they form a neural map of the environment.

When the environment changes, the neural map changes too. These changes are not random noise — they are structured transformations. That structured change is called remapping.

Remapping shows that spatial representations are dynamic and context-sensitive, not fixed coordinate systems.

There are three major types.

---

2⃣ Place Cell DeformationsWhat happens?

If you physically stretch or compress the environment, place fields move in a systematic way relative to environmental boundaries.

For example:
If you stretch a square box to make it longer, place fields stretch correspondingly.

This has been observed in:

• Two-dimensional open boxes

• One-dimensional linear tracks

What does this mean?

The place cell map is not rigid. It is anchored to environmental geometry.

Place fields are not coded in absolute coordinates; they are coded relative to environmental boundaries and cues.

What happens to grid cells?

When place fields deform, grid cells also deform.

The hexagonal grid pattern stretches or compresses in alignment with the environmental distortion.

This shows that grid cells and place cells are tightly coupled during geometric transformations.

Why does this happen?

Spatial coding integrates:

• External sensory cues (walls, boundaries)

• Internal path integration signals

When geometry changes, boundary-based anchoring forces the internal spatial metric to rescale.

This suggests that environmental boundaries play a powerful anchoring role in the spatial map.

What does this tell us?

The spatial system preserves relative structure when geometry changes gradually.

It behaves like a flexible coordinate system, not a fixed GPS.

---

3⃣ Rate RemappingWhat happens?

In rate remapping, place cells keep firing in the same locations, but their firing rates change dramatically.

The spatial position of place fields stays constant.
The intensity of firing changes.

For example:
If the wall color changes from black to white, place cells may fire more or less strongly — but in the same locations.

What does this mean?

The hippocampus encodes more than geometry.

It encodes contextual features.

Rate remapping reflects changes in context without changes in spatial layout.

What happens to grid cells?

Grid cells do not show consistent changes during rate remapping.

This suggests that grid cells primarily encode metric spatial structure rather than contextual identity.

Why is this important?

This indicates that:

• Place cells carry both spatial and contextual information.

• Grid cells primarily provide a spatial metric backbone.

Place cells integrate:

• Grid inputs

• Sensory inputs

• Contextual inputs from lateral entorhinal cortex

Rate remapping likely reflects modulation from non-spatial inputs.

What does this tell us about memory?

It suggests that the hippocampus can represent two experiences occurring in the same location but under different contextual conditions.

For humans, this explains how you can remember:

• The same room during different events

• The same place at different times

The spatial layout stays constant, but the memory trace differs.

---

4⃣ Global Remapping

Global remapping is the most dramatic transformation.

What happens?

Place cells change both:

• Their firing locations

• Their firing rates

The new map is essentially unrelated to the old one.

For example:
A rat explores Box A in Room A.
Then it explores an identical box in Room B.

Despite geometric similarity, the place fields rearrange unpredictably.

Two cells that were neighbors in one map may be far apart in another.

What does this mean?

The hippocampus has switched to an entirely new representation.

It treats the two environments as separate contexts.

This is not distortion — it is map replacement.

What happens to grid cells?

Grid cells do not completely randomize.

Instead, they undergo coherent transformations such as:

• Shifts in grid phase (the grid pattern shifts position)

• Changes in orientation (rotates)

• Changes in scale (spacing between grid fields changes)

Importantly, relationships between grid cells are preserved.

If two grid cells had similar spacing before, they maintain that relationship.

Why is this important?

This reveals a key difference:

• Place cells can completely reorganize.

• Grid cells preserve internal metric structure.

Grid cells maintain distance relationships.
Place cells do not.

This suggests that:

Grid cells encode universal metric space.
Place cells encode environment-specific identity.

---

5⃣ Why Does Global Remapping Occur?

Global remapping likely occurs when:

• The context is perceived as fundamentally different.

• Environmental cues signal a new situation.

• The hippocampus detects a major contextual shift.

It is essentially the neural basis of episodic separation.

It prevents interference between memories of similar environments.

Without global remapping, memories from Room A and Room B would overlap.

This is critical for episodic memory.

---

6⃣ Computational Interpretation

Let us compare the three types:

Deformation

The map stretches but preserves structure.

Rate remapping

The map stays spatially stable but changes contextual intensity.

Global remapping

The map reorganizes entirely.

This suggests the hippocampus can flexibly choose among:

• Metric rescaling

• Contextual modulation

• Complete map switching

This flexibility reflects a balance between stability and plasticity.

---

7⃣ What Does This Tell Us About the Entorhinal–Hippocampal System?

The findings suggest:

1. Grid cells provide a stable metric backbone.

2. Place cells integrate spatial and contextual inputs.

3. The hippocampus decides when two experiences belong to:

   • The same map (deformation or rate remapping)

   • Different maps (global remapping)

This implies that the hippocampus performs context classification.

It determines whether an experience is:

• A variation of the same environment

• Or a completely new environment

---

8⃣ What Does This Tell Us About Humans?

Humans have homologous structures:

• Hippocampus

• Entorhinal cortex

Human fMRI studies show remapping-like signals when:

• Context changes

• Task rules change

• Conceptual environments change

This suggests that remapping may not be limited to physical space.

It may extend to abstract spaces such as:

• Conceptual knowledge

• Social hierarchies

• Task structures

Global remapping may underlie how we form separate episodic memories for:

• Two different houses

• The same house before and after renovation

• The same classroom but during different life periods

Rate remapping may explain how emotional tone or context modifies a memory without changing its location.

---

9⃣ Deep Theoretical Implication

Remapping shows that:

The hippocampus is not simply a spatial coordinate calculator.

It is a context-sensitive memory index.

It can:

• Preserve metric structure

• Modify contextual weighting

• Replace entire maps

This dynamic flexibility is what allows episodic memory to be both stable and distinct.

Grid cells preserve universal spatial relationships.
Place cells provide environment-specific indexing.

Together, they enable both continuity and separation.