Open Field Test

Abstract

The open field test, which MeSH classifies under maze learning, is a behavioral assay in which an animal is placed in a novel, enclosed arena and its movement recorded to index locomotor activity and emotionality. Introduced by Calvin Hall in 1934, who scored defecation and ambulation in the rat as measures of individual differences in fearfulness, it is among the oldest and most widely used tests in behavioral neuroscience. Its signature measure is thigmotaxis — a frightened animal hugs the walls and avoids the exposed center — so time spent in the center zone has become the standard inverse index of anxiety-like behavior, sensitive to anxiolytic and anxiogenic drugs. The test is simple to run but hard to interpret, because a single score confounds anxiety with general activity and its outcomes are highly sensitive to apparatus, illumination, and handling.

Keywords: open field test, thigmotaxis, anxiety-like behavior, locomotor activity, emotionality

The open field test is the behavioral neuroscientist's default first look at a rodent. Drop an animal into a brightly lit box it has never seen, record where it goes for a few minutes, and two things fall out at once: how much it moves, and how much it avoids the frightening open middle. That dual yield — activity and emotionality from one short session — is why the paradigm has survived nearly a century and appears in thousands of studies a year, from drug screening to the phenotyping of genetically modified mice (Hall, 1934; Gould et al., 2009). It is also why the test is so often misread: the same simplicity that makes it ubiquitous hides a confound between the two things it measures, and a difference in center time means nothing until general activity has been accounted for (Walsh & Cummins, 1976).

What the Open Field Test Is

The open field test places a single animal in an unfamiliar, walled arena — typically a square or circular box with an open top — and records its behavior for a fixed period, usually five to ten minutes. The arena is conceptually divided into a protected *periphery* adjacent to the walls and an exposed *center*, and the measures taken fall into two families: locomotion (total distance travelled, number of zone crossings, rearing) and emotionality (time in and entries into the center, latency to first enter it, and in the original version defecation) (Hall, 1934; Gould et al., 2009). What makes the apparatus powerful is the conflict it imposes. A rodent in a novel space is driven to explore it, but an open, brightly lit expanse is inherently threatening to a prey animal, so the test pits exploratory drive against fear and reads the balance from where the animal chooses to spend its time.

The test is filed in MeSH beneath maze learning, but this is an artefact of how the indexing vocabulary groups rodent behavioral apparatus rather than a claim that the open field is a maze or that it measures learning. The open field has no goal, no correct path, and no reinforcement; a single undivided arena is almost the definition of the absence of a maze. It sits in that branch of the classification because it shares animals, arena, and automated tracking with the maze tasks, and because it supplies the activity baseline against which maze performance is read — the control that tells a researcher whether a rat that is slow to find a platform has a memory deficit or is simply moving less (Walsh & Cummins, 1976).

What the Test Measures

Calvin Hall built the apparatus to answer a question about individual differences: do rats differ reliably in “emotionality,” and can that difference be scored objectively? His answer was to measure defecation and urination, on the reasoning that autonomic discharge betrays fear, and ambulation, the amount an animal moved (Hall, 1934). Defecation proved an unreliable index and has largely been abandoned, but Hall's deeper move — treating the animal's response to a novel environment as a quantifiable trait — defined the paradigm. The modern descendant of his emotionality score is the animal's distribution in space: a bold animal ventures into the center, a fearful one presses against the walls.

That wall-hugging is *thigmotaxis*, and it is the behavior that carries most of the test's interpretive weight. Thigmotaxis is a robust, evolutionarily sensible anti-predator response — staying near a vertical surface reduces exposure — and its degree tracks the animal's fear state closely enough that the center-avoidance it produces became the field's workhorse measure of anxiety-like behavior (Simon et al., 1994). The crucial inferential step is treating *reduced center time* as *increased anxiety*. That inference is only valid when the animal is actually moving; an animal that freezes in a corner and one that patrols the periphery both show low center time for very different reasons, which is why center measures are always read alongside total locomotion (Walsh & Cummins, 1976; Prut & Belzung, 2003).

Pharmacological Validation

A behavioral measure of anxiety is only as good as its response to drugs of known anxiety-altering action, and it is the pharmacology that earned the open field its status. The decisive evidence is that anxiolytic compounds increase an animal's willingness to enter the center while anxiogenic compounds decrease it. Benzodiazepines such as chlordiazepoxide and diazepam, which reduce anxiety in humans, reliably raise center entries and center time in the open field, and detailed ethological analyses have shown that these drugs specifically release the suppressed exploration of the exposed zone rather than merely stimulating movement (Choleris et al., 2001). The converse manipulations — anxiogenic drugs, acute stress, or bright illumination — push the animal back toward the walls.

This two-way pharmacological control is what licenses the open field as a model of anxiety rather than a mere activity monitor, and it anchored the comprehensive review that consolidated the paradigm as a tool for measuring drug effects on anxiety-like behavior (Prut & Belzung, 2003). The validation also exposes the confound directly: because many anxiolytics and stimulants also change general activity, a clean demonstration requires showing that center occupancy changes *more* than, or independently of, total distance travelled. The ethological approach — scoring the full repertoire of postures and acts rather than a single summary number — was developed precisely to separate a true anxiolytic profile from simple psychomotor stimulation (Choleris et al., 2001).

Measuring Open-Field Behavior

The raw data of a modern open field session is a trajectory: the animal's position sampled many times a second by an overhead camera and tracking software. From that trajectory a standard panel of measures is computed. *Total distance travelled* and *mean velocity* index locomotor activity. *Center time*, *center entries*, and *latency to enter the center* index anxiety-like behavior through thigmotaxis. *Rearing* — standing on the hind limbs — is scored as exploratory and is sensitive to both novelty and stress, making it a useful third dimension that is neither pure locomotion nor pure anxiety (Sturman et al., 2018). A further measure, the *within-session decline* in activity, indexes habituation: a normal animal slows its exploration as the arena becomes familiar, and a failure to habituate can itself be informative.

The persistent methodological problem is that every one of these numbers depends on decisions the experimenter makes. The boundary between center and periphery is arbitrary, and where it is drawn changes the center-time statistic directly; arena size and shape, light level, the colour and odour of the apparatus, the time of day, and above all the way the animal was handled before testing all shift the results, sometimes enough to reverse a drug effect (Sensini et al., 2020). This sensitivity drove repeated calls for standardization and complete reporting, memorably framed as a plea to stop reinventing the wheel with every new laboratory's idiosyncratic protocol (Stanford, 2007; Seibenhener & Wooten, 2015; Kraeuter et al., 2019).

Worked Example

Consider how the center-avoidance measure is computed and why it must be read against activity. Model the arena as a 5×5 grid of 25 equal squares; the inner 3×3 block of 9 squares is the *center* zone and the 16 border squares are the *periphery*. If an animal distributed its time uniformly across the floor, it would spend a fraction 9/25 = 0.36 of the session in the center, so 36% is the no-anxiety baseline set by geometry alone.

Thigmotaxis pulls the animal below that baseline. Let an anxiety index *a* ∈ [0, 1] scale center occupancy as center-time fraction = (1 − *a*) × 0.36. A moderately anxious animal with *a* = 0.5 therefore spends (0.5)(0.36) = 0.18, or 18%, of its time in the center. Suppose an anxiolytic drug lowers the index to *a* = 0.2: center time rises to (0.8)(0.36) = 0.288, or 28.8%. That is an increase of 10.8 percentage points, a 60% relative rise in center occupancy (0.288/0.18 = 1.60).

The interpretive discipline is in the control measure. Imagine total distance travelled is 42.0 m under vehicle and 43.3 m under the drug — essentially unchanged (a 3% difference). Because center time rose 60% while locomotion held flat, the effect cannot be explained by the drug simply making the animal move more, and the anxiolytic interpretation is warranted. Had distance instead risen 55% in step with center time, the “anxiolytic” signal would be indistinguishable from general psychomotor stimulation, and the claim would fail (Prut & Belzung, 2003; Choleris et al., 2001). This is the single most important habit in reading open-field data: no center measure is interpretable without its locomotor control.

Discussion

The open field test endures because it is cheap, fast, requires no training, and yields two of the most frequently asked-for rodent measures — activity and anxiety — from a single brief session. Its history is essentially the story of a measure that outran its own validity and then spent decades being disciplined. Hall's original emotionality score rested on defecation, which did not hold up; the center-avoidance measure that replaced it is far better, but only because the pharmacology supplied an external criterion against which the anxiety interpretation could be checked (Hall, 1934; Prut & Belzung, 2003).

The test's weaknesses are the direct cost of its strengths. One arena, one animal, one short recording cannot cleanly separate the processes it engages, so the open field confounds anxiety with locomotion, novelty-driven exploration with fear-driven inhibition, and trait with state. Walsh and Cummins catalogued these interpretive hazards in 1976, and the field's response has not been to abandon the test but to constrain it — with locomotor controls, with ethological scoring that reads the whole behavioral repertoire, and with standardized protocols that hold the many nuisance variables fixed (Walsh & Cummins, 1976; Choleris et al., 2001; Stanford, 2007). Read with those controls in place, the open field remains a genuinely informative first assay; read naively, it is a generator of artefacts.

Current Directions

Two currents dominate recent work. The first is a reckoning with reproducibility. Open-field results have proven so sensitive to procedural detail that seemingly trivial husbandry choices can swing them: the method and frequency of handling before a test exert sex-specific effects on the very anxiety measures the test is meant to capture, a finding that reframes handling not as a preliminary to the experiment but as part of it (Sensini et al., 2020). This has sharpened the long-standing demand for standardized, fully reported protocols into a concrete research program on which variables must be controlled and disclosed (Stanford, 2007; Kraeuter et al., 2019).

The second is a move beyond summary statistics toward the fine structure of behavior. Rather than collapse a session into a single center-time number, contemporary analyses mine the trajectory for behaviorally meaningful components — for example treating rearing as a distinct, context- and stress-sensitive dimension of exploration rather than noise, and more generally decomposing spontaneous behavior into recurring motifs with machine-vision and unsupervised methods (Sturman et al., 2018). The aim is to recover from the same cheap apparatus a richer and more reliable readout than the classic confounded scores allow, pushing the open field from a blunt two-number instrument toward a high-dimensional description of how an animal copes with a novel world.

Common Misconceptions

Low center time always means high anxiety.

Only when the animal is moving normally. An animal that is sedated, motorically impaired, or frozen also shows little center time, for reasons that have nothing to do with anxiety. Center measures are interpretable only against a measure of total locomotion (Walsh & Cummins, 1976; Prut & Belzung, 2003).

The open field is a kind of maze.

It is filed under maze learning in the MeSH indexing vocabulary, but it has no goal, no correct path, and no reinforcement, and it tests neither learning nor navigation. The classification reflects shared apparatus and its role as an activity baseline for maze studies, not a shared process (Hall, 1934).

A single center-time score is a clean measure of anxiety.

The test confounds anxiety with general activity and is strongly affected by arena size, lighting, handling, and the arbitrary placement of the center boundary. Valid use requires locomotor controls and standardized procedure, which is why the field has repeatedly called for better protocol reporting (Stanford, 2007; Sensini et al., 2020).

Glossary

Ambulation.

The amount an animal moves about the arena, scored by Hall as line crossings and now by total distance travelled; the open field's primary measure of locomotor activity.

Anxiety-like behavior.

A rodent behavioral pattern, such as center avoidance, that resembles human anxiety and responds to anxiolytic and anxiogenic drugs in the expected directions; the construct the open field is used to index.

Anxiogenic.

A drug or manipulation that increases anxiety; in the open field it decreases center entries and center time, pushing the animal toward the walls, and is the pharmacological mirror of an anxiolytic.

Anxiolytic.

A drug that reduces anxiety, such as a benzodiazepine; in the open field it increases center entries and center time, which is the paradigm's key pharmacological validation.

Benzodiazepine.

A class of anxiolytic drug, including chlordiazepoxide and diazepam, whose reliable release of center exploration in the open field is the clearest evidence that the center measure indexes anxiety.

Center zone.

The exposed inner region of the arena, away from the walls; time spent here is the standard inverse index of anxiety. Its boundary is defined by the experimenter and its placement affects the measure.

Defecation score.

The count of fecal boli deposited during a session, Hall's original autonomic index of emotional reactivity; still occasionally reported, though center measures have largely displaced it.

Emotionality.

Hall's term for an animal's reactive fearfulness in a novel environment, which the open field test was created to quantify; the historical ancestor of the modern anxiety-like-behavior construct.

Habituation.

The decline in exploratory activity across a session as the novel arena becomes familiar; a normal within-session feature of open-field behavior and itself an informative measure.

Latency to center.

The time from the start of the session until the animal first enters the center zone; a longer latency reflects stronger initial avoidance of the exposed region.

Locomotor activity.

General movement, indexed by distance travelled and velocity; one of the two things the open field measures and the control against which anxiety measures must be read.

Novel environment.

An unfamiliar arena the animal has not previously explored; novelty is what drives both the exploratory and the fear responses the open field records, and it dissipates as habituation proceeds.

Periphery.

The region of the arena adjacent to the walls, where a fearful animal spends most of its time; the complement of the center zone.

Rearing.

Standing on the hind limbs, with or without wall support; an exploratory behavior sensitive to both novelty and stress, scored as a dimension distinct from horizontal locomotion.

Thigmotaxis.

The tendency to move while maintaining contact with a vertical surface — wall-hugging; the anti-predator response that produces center avoidance and underlies the open field's anxiety measure.

Velocity.

The animal's speed of movement, averaged over the session or sampled moment to moment; a locomotor measure that distinguishes a sedated animal from a merely immobile fearful one.

Key Researchers

Catherine Belzung

(1962-). Université de Tours and the Inserm iBrain unit; her 2003 review with Prut consolidated the open field as a paradigm for measuring the effects of drugs on anxiety-like behavior, and her laboratory works on the neurobiology of depression and anxiety. ORCID

Johannes Bohacek

ETH Zurich; he studies the neurobiology of stress, and his 2018 work dissected exploratory rearing in the open field as a context- and stress-sensitive behavior distinct from horizontal locomotion. ORCID - Google Scholar

Todd D. Gould

University of Maryland School of Medicine; a professor of psychiatry whose edited volume and 2009 methods chapter codified the open field protocol for the behavioral phenotyping of mice. Faculty page

Calvin S. Hall

(1909-1985). American psychologist who originated the open field test in 1934, scoring defecation and ambulation in the rat as measures of individual differences in emotionality; he was later a prominent dream researcher. Wikipedia

Zoltan Sarnyai

James Cook University, where he heads the Laboratory of Psychiatric Neuroscience; his 2019 protocol chapter with Kraeuter standardized the open field test for locomotor and anxiety-like measurement in preclinical models. ORCID - Google Scholar

S. Clare Stanford

University College London, Emeritus Professor of experimental psychopharmacology; her 2007 commentary pressed the field toward standardized open-field methodology and complete reporting. ORCID

Frequently Asked Questions

What is the open field test?

The open field test is a behavioral assay in which an animal is placed in a novel, enclosed arena and its movement recorded to measure locomotor activity and anxiety-like emotionality at once. It was introduced by Calvin Hall in 1934 and is one of the most widely used tests in behavioral neuroscience (Hall, 1934; Gould et al., 2009).

What does the open field test measure?

It yields two families of measure: locomotor activity (total distance, velocity, rearing) and anxiety-like behavior (time in and entries into the exposed center zone, and latency to enter it). The center measures depend on thigmotaxis, the fearful animal's tendency to hug the walls (Simon et al., 1994).

What is thigmotaxis?

Thigmotaxis is the tendency to move while staying in contact with a wall or vertical surface. In a prey animal it is an anti-predator response, and its degree in the open field tracks fear closely enough that the resulting center avoidance became the standard index of anxiety-like behavior (Simon et al., 1994).

How do we know center avoidance reflects anxiety?

Through pharmacology. Anxiolytic drugs such as benzodiazepines reliably increase center entries and center time, while anxiogenic treatments decrease them, and ethological analysis shows the anxiolytics specifically release exploration of the center rather than merely raising activity (Choleris et al., 2001; Prut & Belzung, 2003).

Why must center time be interpreted against locomotion?

Because a single open-field score confounds anxiety with general activity. An animal that moves little — through sedation, impairment, or freezing — shows low center time for reasons unrelated to anxiety, so a center measure means nothing until total distance travelled has been checked (Walsh & Cummins, 1976).

Is the open field test a maze?

No. It is classified under maze learning in MeSH because it shares apparatus and animals with maze tasks and supplies their activity baseline, but it has no goal, path, or reinforcement and tests neither learning nor navigation (Walsh & Cummins, 1976).

Why are open field results so hard to replicate?

Because they are extremely sensitive to procedure. Arena size, illumination, odour, time of day, and especially the way animals are handled before testing all shift the measures, sometimes enough to reverse a drug effect, which is why standardized protocols are repeatedly urged (Sensini et al., 2020; Stanford, 2007).

What is being done to improve the test?

Two things: tighter standardization and reporting of the nuisance variables, and a shift from single summary scores toward fine-grained analysis of the movement trajectory, including treating behaviors such as rearing as distinct, informative dimensions (Kraeuter et al., 2019; Sturman et al., 2018).

References