Spatial Behavior

Abstract

Spatial behavior is the set of processes by which organisms locate themselves, represent the space around them, and move through it to reach goals. This article treats spatial behavior as a family of capacities rather than a single faculty, spanning the cognitive map first proposed by Edward Tolman, the place and grid cells that encode position in the mammalian brain, the neural systems that support human navigation, and the regulation of interpersonal distance studied as proxemics. It sets out how each is measured, what neural machinery underlies it, and how the abilities vary across people and cultures. Three interactive demonstrations let a reader watch a cognitive map produce a novel shortcut, compare the firing fields of place and grid cells, and explore the concentric zones of personal space.

Keywords: spatial behavior, cognitive map, place cells, grid cells, proxemics

Spatial behavior is the range of activities by which an organism perceives, represents, and acts within the space that surrounds it. In the Medical Subject Headings vocabulary the descriptor is filed as a kind of behavior, and it gathers under one heading a set of capacities that psychology has more often studied apart: knowing where one is, building an internal model of a layout, finding a route to a goal, and governing the physical distance kept from others. What unites them is that each is organised around space as the variable the nervous system must estimate and use. The scientific interest of spatial behavior is that it turns an abstract cognitive question — how is knowledge represented? — into one with an unusually direct answer, because the brain encodes space in single neurons whose activity can be recorded as an animal moves. Few domains of cognition connect a subjective construct so tightly to a measurable neural code, and it is this bridge, from Edward Tolman's inferred cognitive map to the place and grid cells that appear to implement it, that makes spatial behavior a model system for cognition at large.

Types of Spatial Behavior

MeSH organises spatial behavior into seven narrower descriptors, listed in Table 1. They are not a single graded scale but a set of partly independent capacities: a person may navigate a city expertly yet keep an unusual interpersonal distance, and an animal may defend a territory without any laboratory sign of spatial learning. The categories are also an indexing classification built to file the biomedical literature, not a theory of how the mind carves space; they cut the domain for retrieval rather than asserting that these seven kinds exhaust or cleanly partition spatial cognition. With that caveat, the tree is a useful map of the field, separating the representational capacities (orientation, learning, processing, navigation) from the social and ecological uses of space (personal space, crowding, territoriality). Two of the seven are developed as their own articles and are linked below; the remainder are named here as MeSH defines them.

The Cognitive Map

The modern study of spatial behavior begins with a challenge to behaviourism. On the dominant stimulus-response view, a rat learned a maze by strengthening a chain of turn responses, each cued by the last, with no internal representation of the maze as a whole. Edward Tolman argued that this could not be the whole story. In Cognitive maps in rats and men he marshalled experiments — latent learning, in which rats that had merely explored a maze unrewarded later learned it far faster than naive rats, and spatial shortcut tests, in which rats took a direct novel path to a goal when the trained route was blocked — to argue that the animal acquires a cognitive map, a map-like internal representation of the environment's layout that supports flexible, inference-like behaviour rather than fixed habit (Tolman, 1948). The claim was that spatial knowledge is representational and relational: the rat knows where the goal is, not merely which sequence of movements once reached it.

The idea lay relatively dormant while behaviourism receded, then returned with a neural interpretation. John O'Keefe and Lynn Nadel proposed that the hippocampus is the seat of Tolman's map — an allocentric, world-centred representation of space, as opposed to an egocentric representation anchored to the body — and that this map underwrites both spatial memory and, more broadly, the storage of experience in its spatial and temporal context (O'Keefe & Nadel, 1978). Figure 1 sets out the brain's spatial representation system that this proposal launched, and the first demonstration lets a reader watch a cognitive map, but not a stimulus-response chain, produce a novel shortcut when a familiar route is blocked.

Place Cells and Grid Cells

Tolman's map became physiology when John O'Keefe and Jonathan Dostrovsky recorded neurons in the rat hippocampus that fired only when the animal occupied a particular location in its environment (O'Keefe & Dostrovsky, 1971). Each such place cell has a place field, a small region of the arena where it is active; together the population tiles the space, so that the animal's location can be read out from which cells are firing. Place cells are the concrete form of the allocentric map: they encode where the animal is in a world-centred frame, remap to form a distinct code for each new environment, and persist as a stable representation the animal can revisit.

Where place cells locate the animal, the metric of the map turned out to lie one synapse upstream. Torkel Hafting, Marianne Fyhn, and Edvard and May-Britt Moser found that neurons in the medial entorhinal cortex fire at the vertices of a regular hexagonal grid tiling the whole environment, so that a single grid cell has many firing fields laid out in a periodic lattice (Hafting et al., 2005). Grid cells supply a distance-and-direction coordinate system — a metric for space that does not depend on external landmarks, and one well suited to path integration, the updating of position from self-motion cues alone — and are organised into modules of increasing scale, a scheme reviewed by Edvard Moser, Emilio Kropff, and May-Britt Moser as the core of the brain's spatial representation system (Moser et al., 2008). The discovery of place and grid cells was recognised with the 2014 Nobel Prize in Physiology or Medicine. Crucially, the same cell types exist in humans: Arne Ekstrom and colleagues, recording from electrodes in patients navigating a virtual town, found human place-responsive and goal-responsive cells in the hippocampus (Ekstrom et al., 2003). Table 2 summarises the principal spatial cell types, and the second demonstration contrasts the single firing field of a place cell with the hexagonal lattice of a grid cell as a marker moves across an arena.

Human Spatial Behavior and the Brain

The animal work predicts that human spatial behavior should depend on the same hippocampal-entorhinal system, and the human evidence bears this out. The most celebrated demonstration came from London taxi drivers, whose licensing requires them to memorise the city's tangled street layout. Eleanor Maguire and colleagues found that these drivers had a larger posterior hippocampus than matched controls, and that the volume increased with years of driving experience — structural evidence that intensive spatial learning reshapes the very region the animal work implicates (Maguire et al., 2000). Neil Burgess, Eleanor Maguire, and John O'Keefe drew the animal and human strands together, arguing that the human hippocampus supports an allocentric map that serves both spatial navigation and the broader function of episodic memory, for which spatial context is a natural scaffold (Burgess et al., 2002).

Human navigation is not the hippocampus alone. Russell Epstein and Nancy Kanwisher identified a region of the parahippocampal place area that responds selectively to scenes and spatial layouts — the visual recognition of where one is, a complement to the hippocampal map of where things are (Epstein & Kanwisher, 1998). And navigation recruits more than one strategy. Giuseppe Iaria, Véronique Bohbot, and colleagues showed that people spontaneously divide into those who navigate by a hippocampus-dependent spatial strategy, learning the relationships among landmarks, and those who navigate by a caudate-dependent response strategy, learning a fixed sequence of turns — a modern echo of Tolman's map-versus-habit distinction, now with distinct neural substrates and shifting with practice (Iaria et al., 2003). This interplay of a flexible hippocampus-based map and a rigid habit system is the human form of the choice between spatial navigation by knowing and by rote.

Personal Space and Proxemics

Not all spatial behavior is about finding one's way; a large part of it is about the space kept from other people. The anthropologist Edward T. Hall founded the study of this in The Hidden Dimension, coining proxemics for the culturally patterned use of interpersonal distance (Hall, 1966). Hall proposed that people carry with them a set of concentric zones — an intimate zone for embrace and whispering, a personal zone for conversation among friends, a social zone for acquaintances and business, and a public zone for formal address — and that the radii of these zones, and the rules for when each may be entered, vary systematically across cultures. Personal space is thus not idiosyncratic but a structured, measurable form of spatial behavior, and its violation reliably produces discomfort and withdrawal.

Proxemics connects the ecological and social branch of spatial behavior to its cognitive core. The zones are egocentric — defined relative to the body rather than the world — which places them opposite the allocentric cognitive map on the same reference-frame dimension that organises the whole domain. And the maintenance of personal space, like territoriality and the stress of crowding, shows that spatial behavior is regulated by emotional and social systems as well as by the navigational machinery of the hippocampus. The third demonstration lets a reader vary an approach distance and read off which of Hall's zones it falls in and the interaction that zone typically supports.

Individual Differences

People differ enormously in spatial ability, and the sources of that variation are a research field in their own right. Thomas Wolbers and Mary Hegarty set out the determinants of navigational ability, distinguishing the contributions of basic spatial abilities, the strategies a person adopts, the quality of their internal representations, and the effects of ageing, and arguing that navigation is best understood as the product of several partly separable component skills rather than a single talent (Wolbers & Hegarty, 2010). Their framework explains why two people can fail at the same route for different reasons — one misjudging distances, another failing to build a survey representation, a third defaulting to a rigid response strategy.

The scale of the variation, and its predictors, became measurable with large online samples. Antoine Coutrot, Hugo Spiers, and colleagues used Sea Hero Quest, a navigation game played by millions, to chart navigational ability across the globe, finding that it varied with age, gender, and — strikingly — with a country's wealth and its cultural practices (Coutrot et al., 2018). The same dataset later revealed an environmental determinant: people who had grown up in cities with more regular, grid-like street layouts navigated the game's environments less well than those raised in more topologically complex, irregular towns, as though a childhood spent solving harder wayfinding problems built a more capable spatial system (Coutrot et al., 2022). Spatial behavior, on this evidence, is shaped by the environment it develops in as much as by any fixed endowment.

Worked Example

Consider the geometry of the grid-cell code, which illustrates how a small set of neural modules can represent a large range of space. Grid cells are organised into modules, each with its own characteristic grid spacing — the distance between neighbouring firing fields — and the spacings increase from one module to the next by an approximately constant ratio. Empirically that ratio is close to r = 1.42. Take the smallest module to have a spacing of s = 30 cm. The spacing of the n-th module is then s_n = 30 × 1.42^(n-1) cm.

Working upward, module 1 spans 30 cm, module 2 spans 30 × 1.42 = 42.6 cm, module 3 spans 42.6 × 1.42 = 60.5 cm, and module 4 spans 60.5 × 1.42 = 85.9 cm. Each module resolves position finely within its own scale but is ambiguous beyond it, because its pattern repeats; combining modules of different scales resolves that ambiguity, much as the hands of a clock jointly specify a time that no single hand could. Now ask how many modules are needed to span from the smallest scale, 30 cm, up to a large environment of 3 m. We require 30 × 1.42^(n-1) ≥ 300, that is 1.42^(n-1) ≥ 10. Taking logarithms, (n - 1) ≥ ln 10 / ln 1.42 = 2.303 / 0.351 = 6.57, so n - 1 = 7 and n = 8. Just eight modules, scaling by a constant ratio, cover a hundredfold range of spatial scale — a compact, efficient code, and the reason a handful of grid modules suffices to locate an animal across environments spanning many metres. The numbers are illustrative of the reported scaling rather than exact measured values, but they reproduce the geometric-progression structure the recordings reveal.

Discussion

Spatial behavior has become one of the best-understood bridges between cognition and its neural basis. Tolman's inference of a cognitive map from behaviour alone (Tolman, 1948) was vindicated in an unusually literal way when place cells (O'Keefe & Dostrovsky, 1971) and grid cells (Hafting et al., 2005) were found to encode location and metric, and the human work extended the same system to people, from the enlarged hippocampi of taxi drivers (Maguire et al., 2000) to scene-selective cortex (Epstein & Kanwisher, 1998) and the strategy split between map and habit (Iaria et al., 2003). Alongside this navigational core sits a social branch — proxemics, personal space, territoriality, and the stress of crowding — that shares the domain's central organising axis, the contrast between egocentric and allocentric frames (Hall, 1966).

Two themes give the field its current momentum. The first is that spatial ability is not one thing but a set of dissociable components that vary across people and cultures (Wolbers & Hegarty, 2010; Coutrot et al., 2018), so that both its measurement and its neuroscience must be componential. The second, more radical, is that the spatial system may not be for space alone: the same hippocampal-entorhinal code that maps physical environments appears to organise abstract, non-spatial knowledge as well, a possibility taken up below and one that would make spatial behavior the model for cognition rather than merely one province of it.

Current Directions

The most striking contemporary claim is that the brain's map is not confined to physical space. Russell Epstein, Hugo Spiers, and colleagues reviewed evidence that the human cognitive map supports not only navigation but the flexible use of spatial knowledge for planning and inference, and that its principles extend beyond the strictly spatial (Epstein et al., 2017). Jacob Bellmund, Christian Doeller, and colleagues pushed the argument further, proposing that grid-like and place-like codes provide a general format for cognitive spaces — that concepts varying along continuous dimensions, from the size of animals to the pitch of sounds, are represented in the same entorhinal-hippocampal geometry that maps rooms and cities (Bellmund et al., 2018). Michael Peer, Russell Epstein, and colleagues extended the geometry itself, arguing that knowledge is structured not only as continuous cognitive maps but also as discrete cognitive graphs of nodes and links, and that the hippocampal system flexibly uses both (Peer et al., 2021).

A second front is the study of spatial behavior at population scale. The Sea Hero Quest programme has shown that digital games can measure navigational ability in millions of people and recover its demographic, cultural, and environmental determinants, including the childhood-environment effect on adult wayfinding (Coutrot et al., 2022). Together these fronts are reframing spatial behavior: from a specialised faculty for getting around, toward a candidate general mechanism by which the brain represents structured knowledge of any kind.

Common Misconceptions

Animals navigate purely by chained stimulus-response habits.

Tolman's shortcut and latent-learning experiments showed that animals form a flexible cognitive map and can take novel routes, not just retrace trained sequences of turns (Tolman, 1948).

Place cells and grid cells are the same thing.

Place cells in the hippocampus have a single firing field marking one location; grid cells in the entorhinal cortex fire at many locations on a periodic hexagonal lattice and supply a metric (O'Keefe & Dostrovsky, 1971; Hafting et al., 2005).

Spatial ability is a single, fixed talent a person either has or lacks.

Navigation is the product of several separable components and strategies, and it varies with age, culture, and the environment a person grew up in (Wolbers & Hegarty, 2010; Coutrot et al., 2022).

The brain's spatial map represents only physical space.

The same hippocampal-entorhinal code appears to map abstract conceptual dimensions too, suggesting spatial representation is a general format for structured knowledge (Bellmund et al., 2018).

Glossary

Allocentric representation.

A world-centred representation of space, encoding the locations of things relative to each other and to the environment rather than to the observer.

Boundary cell.

A neuron that fires as a function of the distance and direction to environmental edges, anchoring the spatial map to the shape of the enclosure.

Cognitive map.

An internal, map-like representation of the spatial layout of an environment that supports flexible navigation and inference, proposed by Tolman.

Egocentric representation.

A body-centred representation of space, encoding locations relative to the observer's own position and heading.

Entorhinal cortex.

A cortical region adjacent to the hippocampus that houses grid cells and supplies the hippocampus with much of its spatial input.

Grid cell.

An entorhinal neuron that fires at the vertices of a regular hexagonal lattice covering the whole environment, providing a metric coordinate system.

Head-direction cell.

A neuron that fires when the head points in a particular direction, functioning as an internal compass independent of location.

Hippocampus.

A medial-temporal-lobe structure that houses place cells and is central to the allocentric cognitive map and to episodic memory.

Landmark.

A distinctive, stable feature of the environment used as a reference point for orientation and route-finding.

Latent learning.

Learning that occurs without reinforcement and is not expressed in behaviour until there is a reason to use it; key evidence for the cognitive map.

Parahippocampal place area.

A scene-selective region of visual cortex that responds to spatial layouts and environments, supporting the recognition of places.

Path integration.

The process of keeping track of one's position by continuously integrating self-motion cues such as speed and heading, independent of landmarks.

Personal space.

The zone around the body that a person maintains against intrusion by others; one of Hall's proxemic zones.

Place cell.

A hippocampal neuron that fires when the animal is in a specific location, its place field; the population encodes the animal's position.

Proxemics.

The study of the culturally patterned use of interpersonal distance, founded by Edward T. Hall.

Spatial navigation.

The planning and execution of movement through space toward a goal, drawing on the cognitive map and on self-motion cues.

Spatial orientation.

Awareness of one's own position and heading relative to the surrounding environment.

Territoriality.

The defence of a bounded area against others, typically of the same species; the ecological branch of spatial behavior.

Wayfinding.

The cognitive process of determining and following a route between an origin and a destination.

Key Researchers

Russell A. Epstein (contemporary). Psychologist at the University of Pennsylvania; he discovered the parahippocampal place area and has led work on the human cognitive map and its extension beyond physical space. ORCID - Google Scholar - Faculty page

Edward T. Hall (1914-2009). American anthropologist at Northwestern University; he founded proxemics, the study of the culturally patterned use of personal space, in The Hidden Dimension. Wikipedia - Wikidata

Eleanor A. Maguire (1970-2025). Neuroscientist at University College London's Wellcome Centre for Human Neuroimaging; her studies of London taxi drivers showed that spatial expertise reshapes the human hippocampus. ORCID - Wikipedia - Wikidata

Edvard I. Moser (contemporary). Neuroscientist at the Norwegian University of Science and Technology and the Kavli Institute for Systems Neuroscience; he co-discovered grid cells in the entorhinal cortex, sharing the 2014 Nobel Prize. ORCID - Google Scholar - Wikipedia

May-Britt Moser (contemporary). Neuroscientist at the Norwegian University of Science and Technology and co-director of the Kavli Institute for Systems Neuroscience; she co-discovered grid cells, sharing the 2014 Nobel Prize. ORCID - Google Scholar - Wikipedia

John O'Keefe (contemporary). Neuroscientist at University College London; he discovered place cells in the hippocampus and co-developed the theory of the hippocampus as a cognitive map, sharing the 2014 Nobel Prize in Physiology or Medicine. ORCID - Wikipedia - Wikidata

Edward C. Tolman (1886-1959). American psychologist at the University of California, Berkeley; his purposive behaviorism and the concept of the cognitive map, set out in Cognitive maps in rats and men, launched the modern study of spatial representation. Wikipedia - Wikidata

Frequently Asked Questions

What is spatial behavior?
Spatial behavior is the family of processes by which organisms locate themselves, build internal representations of space, move through it to reach goals, and regulate the physical distance they keep from others; MeSH classifies it as a kind of behavior (Tolman, 1948).

What is a cognitive map?
A cognitive map is an internal, map-like representation of an environment's layout that supports flexible navigation, such as taking a novel shortcut; Edward Tolman inferred it from experiments in which rats behaved as though they knew where a goal was rather than merely which turns once reached it (Tolman, 1948).

What are place cells and grid cells?
Place cells are hippocampal neurons that fire when an animal is in a particular location; grid cells are entorhinal neurons that fire at the vertices of a hexagonal lattice covering the environment, supplying a metric. Together they give the cognitive map a neural substrate (O'Keefe & Dostrovsky, 1971; Hafting et al., 2005).

Do humans have the same spatial cells as animals?
Yes. Recordings from electrodes in neurosurgical patients navigating a virtual town revealed human place-responsive and goal-responsive cells in the hippocampus, and the same hippocampal-entorhinal system supports human navigation (Ekstrom et al., 2003).

Why are London taxi drivers studied so often?
Their licensing requires memorising the city's complex street layout, and their posterior hippocampus is larger than matched controls' and grows with experience, which is direct evidence that intensive spatial learning reshapes the human brain (Maguire et al., 2000).

Is spatial ability fixed, or does it depend on where a person grows up?
It varies widely and is shaped by environment: using a navigation game played by millions, researchers found that people raised in cities with irregular, complex street layouts navigated better than those from regular grid-like cities (Coutrot et al., 2018; Coutrot et al., 2022).

What is proxemics?
Proxemics is the study of the culturally patterned use of interpersonal distance, founded by the anthropologist Edward T. Hall, who described concentric intimate, personal, social, and public zones whose sizes vary across cultures (Hall, 1966).

Does the brain's spatial map represent anything besides space?
A growing body of work suggests it does: the same entorhinal-hippocampal code that maps physical environments appears to represent abstract conceptual dimensions, making spatial representation a candidate general format for structured knowledge (Bellmund et al., 2018; Epstein et al., 2017).

References