A cognitive map is your brain’s internal model of the world around you, a neural representation that lets you know where you are, remember where things are, and figure out how to get somewhere new without step-by-step instructions. The term was coined in the 1940s by psychologist Edward Tolman, who argued that rats exploring mazes were not just memorizing sequences of turns but building flexible, map-like representations of space. Since then, researchers have discovered specific brain cells that seem to form the hardware behind cognitive maps, and the concept has expanded well beyond physical navigation into how we organize social relationships, abstract knowledge, and even memory itself.
Where the Idea Came From
Tolman’s insight came from a set of experiments known as latent learning studies, in which rats that had previously wandered a maze without any reward were suddenly able to navigate it efficiently once food was introduced. The animals had clearly learned something about the maze’s layout during their unrewarded wandering, even though their behavior at the time gave no sign of it. This was a direct challenge to the prevailing behaviorist view that all learning required reinforcement. Tolman proposed that the rats had formed an internal “cognitive map” of the maze, a representation they could draw upon flexibly when motivation appeared.1PubMed Central. Accounting for sensitivity of latent learning to behavioral statistics with successor representations For decades the idea remained mostly theoretical, because nobody could point to a specific brain mechanism that might store such a map. That changed in the 1970s when neuroscientist John O’Keefe discovered place cells in the rat hippocampus.
The Brain Cells That Build the Map
Flexible navigation depends on a cognitive map of space, and the leading candidate for its physical basis is a network of specialized neurons spread across the hippocampus and nearby entorhinal cortex.2PubMed Central. Hippocampal place cells encode global location but not connectivity in a complex space These cells fall into several categories, each encoding a different piece of spatial information:
- Place cells: Neurons in the hippocampus that fire when you are in a particular location. Each place cell has its own preferred spot, and together the population tiles an environment so that every position has a unique neural signature.
- Grid cells: Found in the entorhinal cortex, these fire at regularly spaced locations that form a hexagonal lattice across the environment. They provide a kind of internal metric, encoding distances and directions between places.3PubMed Central. Context-dependent spatially periodic activity in the human entorhinal cortex
- Head-direction cells: These act like an internal compass, firing when the animal faces a particular direction regardless of where it is.
- Border cells: Neurons that fire near environmental boundaries like walls or edges, anchoring the spatial code to fixed features in the world.4PubMed Central. The Contribution of Spatially Tuned Brain Cells to Human Navigation Performance
It is generally accepted that place cells let you distinguish one location from another, grid cells let you measure distances between locations, head-direction cells let you maintain a sense of heading, and border cells let you reference your position to fixed landmarks like walls.5Current Biology. What Is a Cognitive Map? Together these populations form a kind of GPS-plus-compass system that the brain assembles in real time as you move through a space. The discovery of these cells earned John O’Keefe, May-Britt Moser, and Edvard Moser the 2014 Nobel Prize in Physiology or Medicine.
How the Map Handles Complex Terrain
Most early studies of place cells and grid cells involved rats running across flat, featureless platforms, which left open the question of how the system deals with the irregular, three-dimensional environments that actually exist in nature. Recent work has started filling in this gap. Recordings from rats foraging across ridged terrain show that place cells adjust their firing to reflect the shape of the surface itself. Place fields stretched parallel to the contours of the terrain, consistent with a map that hugs the surface rather than encoding a simple overhead floor plan.6PubMed Central. Hippocampal place cells map terrain geometry independently of behavior Interestingly, the elongation of place fields followed the geometry of the ground rather than the movement patterns of the rats, suggesting the map is genuinely encoding terrain shape and not just reflecting habitual paths.
Animals that live in fully three-dimensional environments push this question further. Recordings from freely flying bats revealed place cells that were active within confined 3D volumes. In more than 90 percent of neurons studied, all three spatial axes were encoded with similar resolution, and the fields collectively blanketed the available room uniformly.7PubMed. Representation of three-dimensional space in the hippocampus of flying bats So while the rat brain appears to use a surface-based map, the bat hippocampus builds something closer to a true volumetric representation. This makes intuitive sense: a flying animal needs to know not just “where on the ground” but “where in the room,” and the brain scales its mapping accordingly.
Replaying the Map While Resting
Building a cognitive map is one thing; storing it for later use is another. One of the more striking discoveries of recent decades is that the hippocampus does not stop working when an animal finishes navigating. During pauses in exploration and during sleep, the hippocampus produces brief bursts of electrical activity called sharp-wave ripples, and during these bursts, place cells fire in rapid sequences that replay the routes the animal recently traveled or is about to take.
Disrupting these ripples during waking pauses causes specific learning and performance deficits that persist throughout training on a spatial task.8PubMed Central. Awake hippocampal sharp-wave ripples support spatial memory The ripples seem to serve as a kind of rehearsal mechanism that strengthens the map. Longer-duration ripples, which replay larger portions of planned routes, improve memory on maze tasks, and experimentally extending ripple duration with optogenetic stimulation boosts performance.9PubMed Central. Long-duration hippocampal sharp wave ripples improve memory The content of these replays is not random. The specific sequences replayed during waking ripples predict what gets replayed again during subsequent sleep, suggesting a tagging mechanism that selects which experiences are preserved for the long term.10PubMed Central. Selection of experience for memory by hippocampal sharp wave ripples In practical terms, the brain appears to practice your routes while you are taking a break, and practice them again while you sleep, and the experiences that get the most waking rehearsal are the ones that become lasting memories.
More Than Just Physical Space
Perhaps the most dramatic expansion of Tolman’s original idea is the growing evidence that the brain’s mapping machinery is not reserved for physical navigation. The same hippocampal and entorhinal circuits appear to organize abstract, non-spatial information into map-like structures.11PubMed Central. Similarities and differences in spatial and non-spatial cognitive maps In one widely discussed experiment, participants watched a cartoon bird whose neck and legs changed in length along two continuous dimensions. Their task was to predict when the bird would match the appearance of certain target birds. Brain imaging revealed a grid-like signal in the entorhinal cortex that tracked movement through this two-dimensional “conceptual space,” even though no physical navigation was involved at all.12Neuron. What Is a Cognitive Map?
Social relationships are another domain that seems to get mapped this way. Researchers have found that the hippocampus tracks an individual’s position in a two-dimensional social space framed by power and affiliation.13Neuron. What Is a Cognitive Map? Organizing Knowledge for Flexible Behavior When people interact with characters who vary in how powerful and how friendly they are, the hippocampal signal correlates with the character’s angle in this social coordinate system. Subsequent work extended this by showing that people reconstruct unseen two-dimensional social hierarchies into a unified cognitive map in the hippocampus and entorhinal cortex, and that they use a grid-like code in the entorhinal cortex when making inferences about relationships between entities in that abstract space.14Nature Neuroscience. Inferences on a multidimensional social hierarchy use a grid-like code The implication is that your brain organizes social knowledge, like who has authority over whom and who gets along with whom, using the same computational architecture it uses to find your way home.
You Do Not Even Need to Move
An especially curious finding is that the brain’s grid-like spatial signals appear even when you are merely imagining moving through a familiar environment. fMRI studies in humans have detected the signature six-fold rotational symmetry of grid cells in the entorhinal cortex during both actual virtual navigation and periods of imagined navigation along the same paths, with similar signal orientation in both conditions.15Current Biology. Grid-like Processing of Imagined Navigation The grid signal was present in real movement compared to stationary periods and showed up in both the left and right entorhinal cortex during imagination. This fits with the everyday experience that you can mentally “walk” through your house or recall the layout of a familiar city without moving at all. Your cognitive map is not switched on only by physical locomotion; it runs in the background whenever you mentally simulate a space.
Grid-like signals in humans were first identified using fMRI while participants explored a virtual environment. The effect was strongest in the right entorhinal cortex, and how coherent the signal was across that region predicted how well participants performed on a spatial memory test.16PubMed Central. Evidence for grid cells in a human memory network So the grid code is not just present in humans; its quality varies between individuals and relates to how well they can navigate.
Senses Beyond Vision
Most people associate cognitive maps with visual information, and for good reason: humans are highly visual animals. But vision is not the only sense that feeds the map. Experiments in virtual environments have shown that odor cues can serve as spatial landmarks, enriching place cell representations and dramatically improving navigation accuracy.17PubMed. Olfactory landmarks and path integration converge to form a cognitive spatial map In these studies, presenting the same odor at different locations generated distinct place cell firing patterns rather than confusing the map. An odor cue at one location even enhanced place cell coverage beyond that spot, extending the map into nearby unknown territory. This suggests the map is built iteratively, with each new landmark anchoring and stretching the representation outward. The human brain can also use olfactory information to build and deploy cognitive maps, which is especially relevant to understanding how people with visual impairments navigate.18PubMed. Cognitive maps: Constructing a route with your snout
Echolocation offers another dramatic example. Egyptian fruit bats can identify their location after being moved and navigate several kilometers using only echolocation, without visual landmarks. A large-scale acoustic model demonstrated how returning echoes from the environment carry enough information to support map-based navigation.19PubMed. Acoustic cognitive map-based navigation in echolocating bats The lesson is that the cognitive map is not tied to any single sense. Whatever sensory channels an animal relies on, the brain can use them to feed the same hippocampal mapping machinery.
When the Map Develops in Children
Children do not arrive in the world with a fully functional cognitive map. The ability to build flexible spatial representations and use them for novel shortcuts develops gradually. Studies tracking navigation performance in children have identified a distinct shift around age nine, when children begin to reliably orient and navigate using cognitive maps rather than relying on memorized routes.20PubMed. The Emergence of Cognitive Maps for Spatial Navigation in 7- to 10-Year-Old Children Seven- and eight-year-olds were considerably less effective than older children at finding shortcuts, suggesting they were still navigating primarily by remembering sequences of turns rather than constructing a map-like overview.
Longitudinal research following the same children over time has found that overall navigation performance reaches adult-level proficiency around age twelve. The ability to integrate information from different routes into a single coherent map, measured by how accurately children can point between routes they have traveled separately, also stabilizes around this age.21PubMed. Longitudinal development of cognitive mapping from childhood to adolescence This timeline overlaps with the protracted development of the hippocampus, which does not reach full structural maturity until adolescence. If you have ever noticed a young child struggling to give directions or find a shortcut through a neighborhood they visit frequently, this developmental lag in cognitive mapping is probably why.
Alzheimer’s Disease and the Breakdown of the Map
Spatial disorientation, getting lost in familiar places, is one of the earliest and most recognizable symptoms of Alzheimer’s disease. This is no coincidence. The entorhinal cortex, where grid cells reside, is among the first brain regions affected by the protein tangles that characterize Alzheimer’s. Research using mouse models of the disease has shown that grid cell firing becomes destabilized, with reduced spatial periodicity, lower stability, and weakened synchrony with other cell types in the navigation network.22PubMed Central. Disruption of the grid cell network in a mouse model of early Alzheimer’s disease These deficits emerge at the earliest stages of protein buildup and coincide with impaired performance on path integration tasks, the ability to keep track of your position based on your own movement.
The disruption appears to stem specifically from reduced integration of self-motion cues rather than problems with processing environmental landmarks.23PubMed. Grid cell disruption in a mouse model of early Alzheimer’s disease reflects reduced integration of self-motion cues In other words, early Alzheimer’s may not make you unable to recognize a landmark, but it impairs your ability to use your own movement to keep your internal position updated between landmarks. Other mouse models of the disease show that tau protein pathology initiated in the entorhinal cortex leads to excitatory neuron loss, degraded grid fields, and spatial memory deficits.24Neuron. Tau Pathology Induces Excitatory Neuron Loss, Grid Cell Dysfunction, and Spatial Memory Deficits Reminiscent of Early Alzheimer’s Disease This line of research has made navigation testing a target for early Alzheimer’s detection, since spatial problems can appear before the memory complaints that typically trigger a clinical visit.
Two Frames of Reference
Your brain maintains at least two different spatial reference frames, and the interplay between them matters for how the cognitive map operates in practice. An allocentric frame represents space in world-centered coordinates: the coffee shop is north of the library, regardless of which direction you are facing. This is the kind of representation that hippocampal place cells are thought to encode. An egocentric frame represents space relative to your own body: the coffee shop is to your left. Parietal cortex regions handle this self-centered perspective, and there is a coordinate transform interface, running through posterior cingulate and retrosplenial cortex, that converts between the two systems.25PubMed. Spatial coordinate transforms linking the allocentric hippocampal and egocentric parietal primate brain systems for memory, action in space, and navigation
You probably shift between these frames without noticing. When you picture the layout of a city you know well, you are likely using an allocentric map. When you reach for a doorknob, you are relying on egocentric coordinates. The cognitive map in the hippocampus is allocentric, a bird’s-eye view, but for it to be useful in guiding your body through space, its output has to be translated back into the egocentric terms your motor system works with. People who are especially good at this translation tend to be better navigators, and damage to the brain regions mediating the transform can leave someone able to describe a route verbally but unable to walk it.
What Happens When You Outsource Navigation to GPS
If the cognitive map is built through active exploration, a natural question is what happens when you stop actively navigating and let a device do the work. A study that assessed lifetime GPS experience in regular drivers alongside several measures of spatial memory found that people with greater GPS use performed worse on spatial memory during self-guided navigation, when they had to find their way without the device.26PubMed Central. Habitual use of GPS negatively impacts spatial memory during self-guided navigation A follow-up three years later, though based on a small sample, found that increased GPS use over that period was associated with a steeper decline in hippocampal-dependent spatial memory. The finding makes sense in light of what we know about how the cognitive map is maintained: if place cells and grid cells are fed by active engagement with the spatial environment, passive turn-by-turn following gives them less to work with. This does not mean GPS is dangerous, but it does suggest that relying on it exclusively may come at a cost to the navigational abilities you fall back on when the battery dies or the signal drops.
Urban Design and Legibility
The concept of cognitive maps also has a life outside neuroscience, in architecture and urban planning. The idea that some environments are easier to build mental maps of than others dates back to Kevin Lynch’s work in the 1960s, and researchers continue to investigate what makes a neighborhood “legible,” meaning easy to understand spatially. A study of historical neighborhoods found that landmarks and navigation signs were the most important factors affecting spatial legibility from the perspective of residents. The same study found that the organic, irregular texture of these neighborhoods produced poor spatial legibility despite their historical charm, because the layouts lacked the connectivity and integration that help people form clear cognitive maps.27Sustainability. The Legibility Efficacy of Historical Neighborhoods in Creating a Cognitive Map for Citizens If you have ever felt confidently oriented in a grid-based city and hopelessly lost in a winding medieval town center, the legibility of the environment’s layout is a big part of why. The cognitive map you build is only as good as the raw material the environment gives you to work with.
Cognitive Maps and Episodic Memory
One of the more recent extensions of cognitive map theory is into episodic memory, your memory for specific personal experiences. Researchers have proposed that the same hippocampal machinery that organizes space might organize memories according to their conceptual similarity, essentially placing experiences on an internal map where similar events sit close together and dissimilar ones sit far apart. A recent preprint found that the Euclidean distances between concepts in such a cognitive map affected how well people remembered images, not only for items whose features were explicitly compared but also for the reference images themselves.28bioRxiv. Task-relevant cognitive maps in episodic memory The implication is that when your brain places a new experience on its cognitive map of concepts, the placement itself influences whether you remember it. Events that are conceptually far from what you expected, outliers on the map, tend to stick in memory more effectively. This is still early-stage research, but it connects the decades-old idea of a spatial map to everyday questions about why some experiences are memorable and others vanish.