Fear

Abstract

Fear is a form of emotion: a rapid, organized response to a recognized, present threat that mobilizes the body and brain for defense. Psychology distinguishes it from anxiety, the anticipatory unease about uncertain future danger, and studies it through a remarkably tractable laboratory model, Pavlovian fear conditioning, that has mapped much of its circuitry. That circuitry is centered on the amygdala, which evaluates threat and drives defensive reactions through brainstem and hypothalamic outputs, while the conscious feeling of being afraid is increasingly treated as a separate, later product of cortical processing. This article defines fear against its neighbours, surveys its brain basis and the learning and extinction of conditioned fear, examines how fear is measured, and traces the debate over whether fear is a hardwired survival circuit, a basic emotion with a universal expression, or an emotion the brain constructs.

Keywords: fear, threat, amygdala, fear conditioning, extinction

Fear is the emotion of confronting danger. Where anxiety is a diffuse, future-oriented apprehension about threats that may never materialize, fear is the sharp, present-tense reaction to a threat that is identified and close at hand (LeDoux & Pine, 2016). The construct is central to cognitive psychology because its laboratory model, fear conditioning, is one of the best-understood examples of learning in the brain, and because the circuits that detect and respond to threat are conserved across mammals, letting findings in animals inform an account of the human mind. Fear also marks the boundary where normal emotion becomes disorder: the phobias, panic, and post-traumatic states that fill clinics are, in large part, fear learning gone awry.

What Fear Is

Fear is an emotional response to a threat that is recognized and present, organized to deal with the danger through some combination of freezing, flight, fight, and autonomic mobilization. It is distinguished most sharply from anxiety by the nature of the threat: fear answers to a specific, identifiable, usually imminent danger, whereas anxiety is the anticipatory state provoked by potential threats whose occurrence, timing, or form is uncertain (LeDoux & Pine, 2016). The two share circuitry and often co-occur, but the distinction between reacting to a present danger and bracing for a possible one is the organizing contrast of the field.

What fear is at a deeper level is contested, and three views compete. The basic-emotion view holds that fear is one of a small set of discrete, evolved emotions, each with its own signal, physiology, and triggering conditions, and that fear carries a universal, cross-culturally recognized facial expression (Ekman, 1992). A survival-circuit view reframes the matter: the brain's conserved defensive circuits detect and respond to threat non-consciously, and the conscious feeling of fear is a separate cognitive construction layered on top, so that “fear conditioning” is really the conditioning of defensive responses rather than of a feeling (LeDoux, 2014). The theory of constructed emotion goes further, denying that fear has any dedicated biological essence at all: fear is a concept the brain applies to make meaning of core affect and bodily prediction in context, assembled afresh each time rather than triggered as a fixed type (Barrett, 2017).

Defensive behavior is itself graded by how close the threat is, a structure known as the predatory imminence continuum. At low imminence an animal reorganizes ordinary activity to avoid encounters; at moderate imminence, once a threat is detected, it freezes; and at the point of contact it mounts vigorous circa-strike defenses, each mode served by partly distinct neural structures (Fanselow, 1994). Fear is thus not a single state of fixed intensity but a family of defensive modes selected by the proximity and certainty of danger.

Types of Fear

Fear sits within the broader category of emotions in the MeSH thesaurus, which places it at tree position F01.470.361 and records exactly one narrower descriptor beneath it. Table 1 lists that direct subtype. The classification is an indexing scheme for organizing the biomedical literature, not a theory of how fear is built from components, so the single child below reflects how one intense fear state is catalogued for literature searches, not a claim that fear has only one form or that its subtypes must be mutually exclusive.

That panic is the one descriptor filed directly beneath fear shows how a medical indexing vocabulary differs from a psychological taxonomy. A psychological account of the kinds of fear would instead partition it by object and course, the circumscribed fears of specific phobias, the social-evaluative fears, the recurrent surges of panic disorder, and the trauma-cued fear of post-traumatic stress, a structure that cuts across the single branch MeSH records here.

Fear in the Brain

The amygdala is the hub of the brain's threat-processing system. Decades of animal work established that it receives sensory information about a threat, stores the association formed during fear conditioning, and drives the expression of fear through projections to the structures that produce defensive behavior and autonomic change (LeDoux, 2000). Within the amygdala, the central nucleus serves as the principal output: its projections to the brainstem and hypothalamus orchestrate the freezing, startle potentiation, and cardiovascular responses that constitute a fear reaction, so that lesions along these pathways abolish specific components of fear (Davis, 1992).

A key feature of the circuit is that threat reaches the amygdala by two routes. A fast, coarse subcortical pathway from the sensory thalamus delivers a rough signal that can trigger defensive responses before the cortex has finished identifying the stimulus, while a slower cortical pathway delivers a detailed representation that can confirm or cancel the reaction (LeDoux, 2000). This dual-pathway organization explains why a person can flinch from a coiled shape on the path and only afterwards register that it was a stick: the quick route buys speed at the cost of precision. The existence of a functional subcortical “low road” in humans is contested, however: a prominent reappraisal argues that the evidence for a fast thalamo-amygdala shortcut is weak and that the amygdala is better understood as integrating “many roads” of cortical and subcortical input to evaluate biological significance, rather than relying on a single privileged fast pathway (Pessoa & Adolphs, 2010).

The human evidence confirms and extends the animal model. Patients with selective amygdala damage are impaired at recognizing fear in facial expressions while recognizing identity normally, showing the amygdala's specific role in processing threat signals (Adolphs et al., 1994). More strikingly, the patient SM, with complete bilateral amygdala destruction, failed to experience fear across live snakes, a haunted house, and frightening films, while her other emotions remained intact, indicating that the human amygdala is necessary not only for learning about threat but for the feeling of fear itself (Feinstein et al., 2011). A broad synthesis of animal and human work places these findings within the amygdala's wider role in emotional learning, memory modulation, and the guidance of attention (Phelps & LeDoux, 2005).

Fear Learning and Extinction

Fear is learned with extraordinary efficiency through Pavlovian conditioning. A neutral stimulus, the conditioned stimulus, is paired with an aversive unconditioned stimulus such as a mild shock; after even a few pairings the conditioned stimulus alone comes to elicit defensive responses. A meta-analysis of functional imaging studies finds that human fear conditioning engages a consistent network, including the anterior insula, dorsal anterior cingulate, and related regions, with a complementary pattern of deactivation that may track the processing of safety signals (Fullana et al., 2016). Not all associations are learned equally: fear-relevant stimuli such as snakes and angry faces are acquired faster and are more resistant to conscious control than fear-irrelevant ones, which the evolved fear-module account attributes to a preparedness built by natural selection (Öhman & Mineka, 2001).

Conditioned fear does not have to be permanent. In extinction, repeated presentation of the conditioned stimulus without the aversive outcome causes the conditioned response to wane. Crucially, extinction is not the erasure of the original memory but a new, inhibitory learning that competes with it, which is why extinguished fear can return with the passage of time, a change of context, or a reminder of the aversive event. Research spanning rodents and humans has localized the retention of extinction to the ventromedial prefrontal cortex and its regulation of the amygdala, making extinction a model system for translational neuroscience and the basis of exposure therapy (Milad & Quirk, 2012). Because the original association survives beneath the new learning, relapse is a structural feature of fear rather than a failure of treatment.

Measuring Fear

Fear is measured on three fronts that do not always agree, a fact the survival-circuit view treats as central rather than inconvenient. The first is behavioral and physiological: in conditioning experiments fear is indexed by freezing in animals and by the skin conductance response and fear-potentiated startle in humans, the last being the reliable amplification of the eyeblink reflex when a person is in a state of fear (Davis, 1992). These measures tap the output of the defensive circuit and can be obtained without asking the person anything.

The second front is self-report: ratings of how afraid a person feels, how much they expect the aversive event, and how threatening they judge a stimulus. The survival-circuit and two-system frameworks insist that these subjective reports index a different process from the defensive responses, so that a participant may show a robust conditioned skin conductance response while reporting little fear, or the reverse, and conflating the two has impeded progress on fear and anxiety disorders (LeDoux & Pine, 2016). The third front exploits the imminence structure of defense directly: a functional imaging study using a virtual predator that could chase and “capture” participants found that as the threat drew near, activity shifted from the ventromedial prefrontal cortex toward the periaqueductal gray, and that this shift tracked the rated degree of dread, giving a neural readout of fear's intensity as a threat approaches (Mobbs et al., 2007).

Worked Example

Consider a differential fear-conditioning experiment modeled with the Rescorla-Wagner rule, the standard account of how associative strength accrues. On each trial the associative strength V of the conditioned stimulus is updated by ΔV = α(λ − V), where λ is the maximum strength the aversive outcome can support (set to 1 when the shock is present) and α is a learning-rate parameter combining stimulus salience and associability (set to 0.30 here).

Starting from V = 0 for a cue the participant has never encountered, the acquisition trials unfold deterministically. Trial 1 adds 0.30(1 − 0) = 0.30, giving V = 0.300. Trial 2 adds 0.30(1 − 0.300) = 0.210, giving V = 0.510. Trial 3 reaches 0.657, trial 4 reaches 0.760, and trial 5 reaches 0.832. The increments shrink each trial because the term (λ − V) — the amount still left to learn — contracts as V approaches the asymptote of 1, which is why fear learning is fast at first and then levels off.

Now switch to extinction, presenting the cue with no shock so that λ = 0. The same rule runs in reverse from V = 0.832: trial 1 subtracts 0.30(0 − 0.832) = 0.250, giving V = 0.582; trial 2 gives 0.408; trial 3 gives 0.285. The conditioned response decays toward zero, but the model makes the key point explicit: extinction drives the measured association down without restoring the pre-learning state, consistent with the finding that extinction is new inhibitory learning rather than erasure and that fear can therefore return (Milad & Quirk, 2012). The curve the demo draws is this very sequence.

Discussion

The study of fear carries three lessons for cognitive psychology. First, the tractability of fear conditioning has made fear the most completely mapped of the emotions, tracing a path from a sensory stimulus through the amygdala to a defensive response and identifying the prefrontal circuitry that later inhibits it (LeDoux, 2000; Milad & Quirk, 2012). This success is also a caution: because the model measures defensive responses so well, it is tempting to equate those responses with the feeling of fear, a conflation the two-system and survival-circuit frameworks argue has held back the clinical payoff (LeDoux & Pine, 2016).

Second, the human lesion evidence shows that the amygdala is not merely a relay in a reflex but is necessary for the experience of fear, since its destruction can leave a person unable to feel afraid while other emotions persist (Feinstein et al., 2011). Third, the nature of fear remains genuinely open. The basic-emotion, survival-circuit, and constructed-emotion accounts make different predictions about whether fear has a dedicated biological signature, and the evidence that fear-relevant stimuli are learned about preferentially (Öhman & Mineka, 2001) sits more comfortably with an evolved specialization than with a wholly constructed category (Barrett, 2017). Fear is thus both the best-understood emotion mechanically and a live test case for what an emotion fundamentally is.

Current Directions

The most consequential current move is the two-system reframing of fear and anxiety. Its proponents argue that the circuits producing defensive behavior and physiological change are distinct from those producing the conscious feelings reported as fear, and that treatments have underperformed because trials measure and target the former while patients seek relief from the latter; the proposal is to study and treat the two systems separately (LeDoux & Pine, 2016). This reframing has sharpened a long-running debate over what “fear” names in the first place.

A parallel development enriches the behavioral side with a computational taxonomy. Rather than treating defense as a set of reflexes, a hierarchical account distinguishes rigid reactions from goal-directed actions that involve implicit or explicit forecasting of outcomes, and maps each onto partly distinct circuits, bringing the tools of reinforcement learning to the problem of how organisms choose among defensive options (LeDoux & Daw, 2018). At the same time the constructed-emotion program presses the opposite case, using active-inference models of interoception and categorization to argue that fear emerges from domain-general predictive machinery rather than a dedicated module (Barrett, 2017). Whether the field converges on evolved circuits, constructed categories, or some reconciliation of the two is the central open question.

Common Misconceptions

Fear and anxiety are the same thing.

They share circuitry but differ in their object. Fear is the reaction to a specific, present, identifiable threat, whereas anxiety is anticipatory apprehension about an uncertain future danger, and blurring the two has hindered both research and treatment (LeDoux & Pine, 2016).

Extinguishing a fear erases the original fear memory.

Extinction is new inhibitory learning that suppresses the conditioned response, not a deletion of the association. The original memory survives beneath it, which is why extinguished fear returns with time, a change of context, or a reminder (Milad & Quirk, 2012).

A measured fear response proves the person feels afraid.

Defensive responses such as freezing or a skin conductance response index the output of the threat circuit, which can dissociate from the conscious feeling. Someone may show a strong conditioned response while reporting little fear, so physiological and self-report measures must be kept distinct (LeDoux, 2014).

Glossary

Amygdala.

An almond-shaped structure in the medial temporal lobe that evaluates threat, stores conditioned fear associations, and drives defensive responses.

Anxiety.

An anticipatory emotional state directed at uncertain future threats, distinguished from fear's reaction to a present, identified danger.

Conditioned stimulus.

An originally neutral cue that, after pairing with an aversive event, comes to elicit a conditioned fear response.

Extinction.

The decline of a conditioned response when the conditioned stimulus is presented repeatedly without the aversive outcome; a new inhibitory learning rather than erasure.

Fear conditioning.

A Pavlovian procedure in which a neutral stimulus paired with an aversive event acquires the capacity to trigger defensive responses.

Fear-potentiated startle.

The amplification of the startle reflex when a person or animal is in a state of fear, used as an objective index of conditioned fear.

Panic.

A sudden, intense surge of acute fear and autonomic arousal, catalogued in MeSH as the direct subtype of fear.

Periaqueductal gray.

A midbrain region that organizes active defensive responses and whose engagement increases as a threat becomes imminent.

Predatory imminence continuum.

The organization of defensive behavior into modes selected by how close a threat is, from pre-encounter avoidance through freezing to circa-strike defense.

Preparedness.

An evolved bias to acquire fear more readily to stimuli that posed recurrent threats in ancestral environments, such as snakes and angry faces.

Rescorla-Wagner rule.

A model of associative learning in which the change in a cue's associative strength is proportional to the discrepancy between the outcome and current expectation.

Survival circuit.

A conserved neural system that detects and responds to a class of challenges such as threat, on the survival-circuit view distinct from the conscious feeling that may accompany it.

Two-system framework.

The proposal that defensive behavioral and physiological responses and the conscious feelings of fear and anxiety arise from distinct circuits that should be studied and treated separately.

Ventromedial prefrontal cortex.

A prefrontal region that supports the retention of extinction by regulating amygdala output, central to exposure-based therapy.

Key Researchers

Ralph Adolphs

. Bren Professor of Psychology, Neuroscience, and Biology at the California Institute of Technology; showed through lesion and patient studies that the human amygdala is necessary for recognizing fear in facial expressions and for normal social-emotional processing of threat. Google Scholar - Faculty page

Lisa Feldman Barrett

. University Distinguished Professor of Psychology at Northeastern University; advanced the theory of constructed emotion, arguing that fear is not a fixed biological category triggered by a dedicated circuit but a concept the brain constructs from core affect, prediction, and context. ORCID - Google Scholar - Wikipedia

Paul Ekman

(1934-2025). Late Professor Emeritus of Psychology at the University of California, San Francisco; argued that fear is one of a small set of basic emotions marked by a universal, cross-culturally recognized facial expression, grounding the discrete-emotions view of fear. Google Scholar - Wikipedia

Joseph E. LeDoux

. Henry and Lucy Moses Professor of Science in the Center for Neural Science at New York University; mapped the amygdala-centered circuitry of threat processing and fear learning, distinguished a fast subcortical pathway from a slower cortical one, and reframed fear as the conscious feeling layered atop non-conscious defensive survival circuits. ORCID - Google Scholar - Wikipedia

Mohammed R. Milad

. Professor and Head of Psychiatry at the University of Illinois Chicago; established the ventromedial prefrontal cortex as central to the retention of fear extinction, building a translational bridge between rodent extinction circuitry and human anxiety and trauma disorders. Google Scholar - Faculty page

Elizabeth A. Phelps

. Pershing Square Professor of Human Neuroscience in the Department of Psychology at Harvard University; extended animal models of fear conditioning and extinction to the human brain, demonstrating the amygdala's role in emotional learning, memory, and the regulation of fear. Google Scholar - Wikipedia

Arne Öhman

(1943-2020). Late Professor of Psychology at the Karolinska Institute; developed the evolved fear-module account and preparedness theory, showing that fear-relevant stimuli such as snakes and angry faces are learned about faster and processed by automatic, preattentive mechanisms. Google Scholar - Wikipedia

Frequently Asked Questions

What is the difference between fear and anxiety?

Fear is the reaction to a specific, present, identifiable threat, whereas anxiety is anticipatory apprehension about an uncertain future danger. They share brain circuitry and often occur together, but the present-versus-potential distinction is the field's organizing contrast (LeDoux & Pine, 2016).

What is fear conditioning?

Fear conditioning is a Pavlovian procedure in which a neutral cue is paired with an aversive event, so that the cue alone comes to trigger defensive responses. It is the dominant laboratory model of fear because it is simple, fast, and conserved across mammals (LeDoux, 2000).

Which part of the brain is responsible for fear?

The amygdala is the hub of the threat-processing circuit, evaluating danger, storing conditioned associations, and driving defensive responses; in humans it is necessary even for the feeling of fear, which is absent after its complete destruction (Feinstein et al., 2011).

Can a conditioned fear be unlearned?

Fear can be reduced through extinction, in which the cue is presented without the aversive outcome. Extinction is new inhibitory learning rather than erasure, so the fear can return with time or context change, and it depends on the ventromedial prefrontal cortex (Milad & Quirk, 2012).

Why are people more afraid of snakes and spiders than of more dangerous modern things?

Fear-relevant stimuli that posed recurrent threats in ancestral environments are learned about faster and more automatically than modern hazards such as cars, a bias the evolved fear-module account calls preparedness (Öhman & Mineka, 2001).

Is fear a basic emotion?

On the basic-emotion view fear is one of a small set of discrete, evolved emotions with a universal facial expression (Ekman, 1992). Rival views treat it as a non-conscious survival circuit plus a separate feeling, or as a category the brain constructs, so the question is actively debated (Barrett, 2017).

How do researchers measure fear?

Fear is measured by behavioral and physiological signs such as freezing, skin conductance, and fear-potentiated startle, and separately by self-reported feeling. These can dissociate, so the two kinds of measure are interpreted as indexing different processes (LeDoux, 2014).

Does the brain respond differently as a threat gets closer?

Yes. As a threat becomes imminent, activity shifts from the ventromedial prefrontal cortex toward the periaqueductal gray, and defensive behavior moves from cautious avoidance through freezing to vigorous circa-strike responses (Mobbs et al., 2007; Fanselow, 1994).

References