Arousal

Abstract

Arousal, which MeSH classifies under psychophysiology, is the global state of physiological and psychological activation that ranges from deep sleep to alert excitement and sets the level of an organism’s readiness to respond. Yerkes and Dodson first described its inverted-U relation to performance, and Hebb recast arousal as a nonspecific energizer of behavior. Schachter and Singer then showed that the same undifferentiated arousal is labeled by cognition into distinct emotions, while Russell’s circumplex made arousal one of two dimensions organizing all affect. Modern work locates much of arousal’s regulation in the locus coeruleus–norepinephrine system, whose phasic and tonic modes trade focused engagement against exploratory distractibility. Three interactive demonstrations model the arousal–performance curve, two-factor emotion, and adaptive-gain neuromodulation.

Keywords: arousal, activation, locus coeruleus

What Arousal Is

Arousal is the overall level of activation of the nervous system and the body — a continuum running from unconsciousness and deep sleep, through drowsiness and calm wakefulness, to alertness, excitement, and extreme agitation. It is indexed by a convergent set of measures: cortical desynchronization on the EEG, elevated heart rate and blood pressure, pupil dilation, increased skin conductance, and heightened muscle tone. Where a specific behavior answers the question what an organism does, arousal answers how intensely and how readily it does anything at all — it is the energetic or intensive dimension of behavior rather than its content or direction.

The construct earns its keep because that single level has pervasive consequences. The same rise in activation that sharpens a sprinter also floods a test-taker with anxiety; the same drop that brings restful calm also dulls a night-shift monitor’s vigilance. Because arousal modulates perception, attention, memory, and action all at once, it is one of the few variables in psychology whose manipulation reaches nearly every other process — which is exactly why an account of the mind cannot treat it as background.

Two cautions organize the modern view. First, arousal is not perfectly unitary: cortical, autonomic, and behavioral indices can dissociate, and different situations recruit somewhat different activation systems, so a single number is an idealization. Second, arousal is nonspecific as to content: on its own it supplies intensity but not meaning, and what a given state becomes — a performance, an emotion, a decision — depends on the task, the context, and the interpretation laid over it. The sections that follow trace those three consequences in turn: arousal and performance, arousal and emotion, and the neuromodulatory machinery that sets the level.

Types of Arousal

The Medical Subject Headings (MeSH) vocabulary files Arousal (D001143) under psychophysiology, and lists three narrower descriptors beneath it. MeSH is an indexing classification built for retrieving literature, not a theory of activation, so these subtypes are a librarian’s partition rather than a mechanistic taxonomy: they are orthogonal to the dimensional view developed on this page, in which arousal is a single continuum crossed with valence. Each names a distinct state or process in which the general activation construct is expressed — a focused deployment of it (attention), a motivational instance of it (sexual arousal), or the baseline waking state it presupposes (wakefulness).

Only attention has a dedicated article on this site so far; Sexual Arousal and Wakefulness are named here without links. The upward direction of the hierarchy is equally telling: MeSH places arousal under psychophysiology — the study of the bodily basis of psychological processes — an indexing choice that reflects how thoroughly arousal is defined through its autonomic and cortical signatures rather than any placement in a clean conceptual taxonomy.

The Arousal–Performance Law

The oldest and most durable finding about arousal is that its relation to performance is non-monotonic. Robert Yerkes and John Dodson (1908), training mice on a brightness discrimination under electric shock, found that performance improved with shock intensity up to a point and then declined — an inverted-U curve. Crucially, the location of the optimum depended on task difficulty: on the hard discrimination, the best performance came at a lower level of stimulation than on the easy one. This second clause, often forgotten, is the substantive content of the Yerkes–Dodson law: not merely that moderate arousal is best, but that the arousal that is optimal falls as the task gets harder.

Donald Hebb (1955) gave the curve a psychological rationale. In his account arousal is a nonspecific energizer — a drive-like activation, distinct from the cue or steering function of a stimulus, that the ascending reticular system supplies to the cortex. Too little arousal and the cortex is underdriven and behavior is sluggish; too much and cue function itself deteriorates as activation swamps the specific signals that guide response. The inverted U is thus the sum of two opposing effects of one variable: activation helps until it begins to interfere.

James Easterbrook (1959) supplied the mechanism that still dominates textbook accounts: cue utilization. As arousal rises, the range of cues an organism uses narrows. At low arousal, attention is broad and includes irrelevant cues; moderate arousal prunes the irrelevant ones and aids performance; but high arousal keeps narrowing until relevant cues are excluded too, and performance falls. The same narrowing that focuses a moderately aroused mind tunnels an over-aroused one — which is why complex tasks, needing more cues in play at once, have their optimum at lower arousal than simple tasks that need only a few.

Arousal and Emotion

If arousal is nonspecific, then bodily activation alone cannot say which emotion a person feels — and that gap is the starting point of the most influential theory linking the two. Stanley Schachter and Jerome Singer (1962) proposed a two-factor theory: an emotion is the joint product of (1) a state of undifferentiated physiological arousal and (2) a cognitive label the person applies to it by appraising the situation. In their experiment, participants injected with adrenaline who lacked an explanation for their arousal took on the mood of a confederate — euphoric or angry — whereas those told to expect the arousal symptoms did not. The same racing heart became elation or irritation depending only on the interpretation available.

The theory reframed a century-old debate. Where the James–Lange view made emotion the perception of a specific bodily pattern, and the Cannon–Bard view severed emotion from a too-slow, too-uniform physiology, Schachter and Singer kept physiology central but made it general: arousal sets the intensity, cognition sets the identity. Emotion is neither read off the body nor generated purely in the head, but constructed from both.

Dimensional models absorbed this insight into a geometry. James Russell’s (1980) circumplex model of affect placed every emotional state in a plane defined by two orthogonal axes — valence (pleasant–unpleasant) and arousal (activated–deactivated). Excitement and serenity share positive valence but differ in arousal; fear and boredom share negative valence and differ likewise. On this view arousal is not a by-product of specific emotions but a constitutive dimension of all of them. Lisa Feldman Barrett’s (2017) theory of constructed emotion pushes the logic further: arousal, as a component of core affect, is a low-dimensional interoceptive summary that the brain categorizes into discrete emotions using learned concepts — making the Schachter–Singer labeling step continuous and predictive rather than occasional.

The Neuromodulatory Basis

The modern neural account of arousal begins with the discovery that it has a definite brainstem substrate. Giuseppe Moruzzi and Horace Magoun (1949) found that electrically stimulating the reticular formation of the brainstem desynchronized the cortical EEG and roused a sleeping animal to alertness, identifying the ascending reticular activating system (ARAS) as the diffuse pathway that sets the cortex’s overall level of activation. This was the anatomical reality behind Hebb’s “nonspecific energizer”: a subcortical system whose tonic output drives cortical readiness independent of any specific sensory content.

Arousal is not a single dial, however, but the output of several interacting neuromodulatory systems, of which the best characterized is the locus coeruleus (LC), a small brainstem nucleus that is the brain’s principal source of norepinephrine. Pribram and McGuinness (1975) had already argued on behavioral grounds that “arousal” conceals at least three separable control processes — arousal proper (phasic response to input), activation (tonic readiness to respond), and effort (the coordination of the two) — presaging the modern distinction between the LC’s two firing modes.

Gary Aston-Jones and Jonathan Cohen (2005) formalized that distinction as the adaptive-gain theory. In phasic mode, the LC is moderately active at baseline and fires crisp bursts to task-relevant, decision-triggering events; this transiently raises neural gain, sharpening the very responses that matter and yielding focused, high-performance exploitation of the current task. In tonic mode, baseline LC firing is high and phasic bursts are absent; gain is elevated indiscriminately, responsiveness to everything rises, behavior becomes distractible, and the animal disengages to explore alternatives. The inverted-U of performance falls directly out of this: too-low tonic firing is inattentive drowsiness, the phasic regime is the optimal peak, and too-high tonic firing is the distractible, over-aroused decline. Poe and colleagues’ (2020) review confirms the LC is far more anatomically and functionally differentiated than the old uniform-projection picture assumed, with modular outputs that can bias specific circuits.

Two further lines connect this machinery to cognition. Mara Mather and Matthew Sutherland’s (2011) arousal-biased competition theory shows that arousal does not raise or lower processing uniformly: it amplifies the neural priority of already-salient or goal-relevant representations while suppressing the rest, so arousal sharpens the winner-take-all landscape of perception and memory rather than adding a constant. And Unsworth and Robison (2017) tie tonic LC regulation to individual differences: people with better attention control and working-memory capacity show tighter, more task-appropriate baseline arousal (indexed by pupillometry), consistent with the idea that regulating the LC well is part of what makes some minds more focused than others.

Worked Example

The Yerkes–Dodson law is easiest to see as a curve rather than a slogan. Model performance P as a function of arousal a (scaled 0–1) that peaks at a task-specific optimum a\* and falls off on either side:

P(a) = 100 · exp[ −k · (a − a\*)² ]

The width parameter k captures how sharply performance depends on getting arousal right. A simple task has a high optimum and a forgiving curve (a\ = 0.75, k = 6); a hard task has a low optimum and an unforgiving one (a\ = 0.35, k = 12). Evaluating both across the arousal range:

Read down the columns and both curves are inverted U’s — each rises to its peak and falls. Read across the rows and the law’s real content appears. At low arousal (a = 0.20) the hard task (76.3) vastly outperforms the simple task (16.3), because the simple task’s optimum is far away; at high arousal (a = 0.90) the ordering reverses completely (87.4 vs 2.7), the hard task having collapsed. The two curves cross near a ≈ 0.52 (P ≈ 71.9). The practical moral is the one coaches and clinicians rediscover: the arousal that produces a personal best on a well-learned, simple skill is precisely the arousal that wrecks a demanding, novel one, so there is no single “ideal” level of activation — only a level ideal for a given difficulty. The interactive curve below lets the arousal level and task difficulty be varied and shows the performance marker tracking up and over the inverted U.

Discussion

Arousal sits at an awkward but revealing junction in cognitive psychology: it is at once a physiological variable, a dimension of subjective experience, and a control parameter for cognition, and its history is largely the story of keeping those three in view at once. The Yerkes–Dodson curve, Hebb’s energizer, Easterbrook’s cue utilization, Schachter and Singer’s labeling, and Aston-Jones and Cohen’s adaptive gain are not competing theories of one thing so much as accounts of arousal’s consequences at different levels — behavioral, attentional, emotional, and neural — that turn out to be strikingly consistent. The inverted U that Yerkes and Dodson measured in mice reappears, a century later, as the exploitation–exploration trade-off of a brainstem nucleus.

Two themes recur. The first is nonspecificity: from Hebb’s energizer to Russell’s dimension to Barrett’s core affect, arousal keeps being characterized as supplying intensity without content, so that its behavioral and emotional meaning is always fixed by something else — the task, the cue structure, the appraisal, the concept applied. The second is regulation: the modern emphasis has shifted from arousal’s effects to its control, asking not merely what a given level does but how the brain sets and adjusts that level moment to moment, and why some individuals do so more adaptively than others.

Arousal also binds outward to much of the rest of the field. It is the state variable behind attention and vigilance, a constitutive dimension of the emotions, a determinant of what is encoded into and retrieved from memory, and, through signal detection theory, a lever on the sensitivity and criterion of perception itself. Few constructs in psychology are so peripheral in appearance and so central in reach.

Current Directions

The most active current thread treats arousal as a regulated, individually varying signal rather than a fixed input. Pupillometry has re-emerged as a non-invasive index of locus-coeruleus activity, and work in this vein — exemplified by Unsworth and Robison’s (2017) account — links moment-to-moment and baseline pupil dynamics to differences in attention control and working-memory capacity, recasting “good regulation of arousal” as a measurable cognitive trait. The open question is causal direction: whether tighter LC regulation produces better control, or merely accompanies it.

A second front sharpens how arousal shapes cognition. Mather and Sutherland’s (2011) arousal-biased competition framework has driven experiments showing that arousal selectively strengthens high-priority representations while weakening low-priority ones, rather than acting as a uniform amplifier — a result with consequences for why emotional events are remembered vividly at the expense of their neutral surroundings. Meanwhile Poe and colleagues’ (2020) evidence for a modular, target-specific locus coeruleus challenges the founding assumption that arousal is delivered as a single global broadcast, and raises the possibility of partially independent arousal channels — a reframing whose implications for the century-old “unitary arousal” debate are still being worked out.

Common Misconceptions

“More arousal means better performance.”

The relation is an inverted U, not a straight line: beyond an optimum, further arousal impairs performance, and that optimum is lower for hard tasks than easy ones (Yerkes & Dodson, 1908).

“Each emotion has its own distinct bodily arousal.”

Arousal is largely undifferentiated; the same activation is labeled into different emotions by cognitive appraisal of the situation (Schachter & Singer, 1962).

“Arousal is a single unitary quantity.”

Cortical, autonomic, and behavioral indices can dissociate, and the locus coeruleus itself is now known to be modular, so a single arousal number is an idealization (Poe et al., 2020).

“High arousal boosts every representation equally.”

Arousal biases competition: it amplifies already-salient or goal-relevant representations while suppressing the rest, rather than raising all processing uniformly (Mather & Sutherland, 2011).

Glossary

Activation.

A near-synonym for arousal, often reserved (as by Pribram and McGuinness) for the tonic, baseline readiness to respond, as distinct from the phasic response to a specific input.

Adaptive-gain theory.

Aston-Jones and Cohen’s account in which the locus coeruleus adjusts neural gain, switching between a phasic mode that supports focused task exploitation and a tonic mode that supports exploratory disengagement.

Arousal-biased competition.

Mather and Sutherland’s theory that arousal amplifies the neural priority of salient or goal-relevant stimuli while suppressing low-priority ones, sharpening rather than uniformly raising processing.

Ascending reticular activating system (ARAS).

The diffuse brainstem–to–cortex pathway, identified by Moruzzi and Magoun, whose stimulation desynchronizes the cortical EEG and produces waking arousal; the anatomical substrate of nonspecific activation.

Circumplex model of affect.

Russell’s model arranging emotional states in a plane defined by two orthogonal dimensions, valence and arousal, so that every affect is a point on a circle around a neutral center.

Core affect.

In Barrett’s framework, the ever-present, low-dimensional state of valence and arousal that the brain categorizes, using learned concepts, into discrete emotional experiences.

Cue utilization.

Easterbrook’s principle that rising arousal narrows the range of environmental cues an organism uses, aiding performance by excluding irrelevant cues until, at high arousal, relevant cues are lost too.

Locus coeruleus.

A small brainstem nucleus that is the brain’s principal source of norepinephrine and a central regulator of arousal, attention, and the phasic–tonic control of neural gain.

Nonspecific energizer.

Hebb’s conception of arousal as a general activating drive that powers behavior without specifying its direction, distinct from the cue or steering function of a stimulus.

Norepinephrine.

The neurotransmitter released by the locus coeruleus that modulates arousal and neural gain throughout the cortex; also a peripheral stress hormone.

Phasic mode.

A locus-coeruleus firing pattern of low-to-moderate baseline activity punctuated by crisp bursts to task-relevant events, associated with focused, high-performance engagement.

Tonic mode.

A locus-coeruleus firing pattern of high, sustained baseline activity without phasic bursts, associated with distractibility, disengagement, and exploratory behavior.

Two-factor theory.

Schachter and Singer’s account of emotion as the joint product of undifferentiated physiological arousal and a cognitive label supplied by appraising the situation.

Valence.

The pleasant–unpleasant dimension of affect; together with arousal it forms the two axes of the circumplex model.

Yerkes–Dodson law.

The empirical generalization that performance is an inverted-U function of arousal, with the optimal level of arousal decreasing as task difficulty increases.

Key Researchers

Gary Aston-Jones (Rutgers University). Co-author of the adaptive-gain theory linking phasic and tonic locus-coeruleus–norepinephrine activity to arousal, attention, and the exploitation–exploration trade-off. ORCID - Rutgers Faculty - Google Scholar

Lisa Feldman Barrett (Northeastern University). Her theory of constructed emotion treats arousal, within core affect, as an interoceptive summary the brain categorizes into discrete emotion using learned concepts. ORCID - Wikipedia - Wikidata - Northeastern Faculty

Donald O. Hebb (1904–1985). Recast arousal as a nonspecific energizer supplied to the cortex, giving the inverted-U curve a neural rationale and separating drive from cue function. Wikipedia - Wikidata

Mara Mather (University of Southern California). Developed arousal-biased competition, showing that arousal sharpens the priority landscape of perception and memory rather than amplifying processing uniformly. ORCID - Wikipedia - Wikidata - USC Faculty

James A. Russell (Boston College). Author of the circumplex model of affect, which makes arousal one of two orthogonal dimensions — with valence — organizing all emotional experience. ORCID - Wikipedia - Wikidata - Google Scholar

Stanley Schachter (1922–1997). With Jerome Singer, proposed the two-factor theory of emotion: undifferentiated arousal labeled by cognitive appraisal yields a specific emotional experience. Wikipedia - Wikidata

Robert M. Yerkes (1876–1956). With John Dillingham Dodson, established the inverted-U arousal–performance relation and its dependence on task difficulty — the law that still bears their names. Wikipedia - Wikidata

Frequently Asked Questions

What is arousal in psychology?
Arousal is the overall level of activation of the nervous system and body, running from deep sleep through calm wakefulness to high alertness and agitation. It sets how ready an organism is to perceive, decide, and act, and is measured by EEG, heart rate, pupil size, and skin conductance.

What is the Yerkes–Dodson law?
It is the finding that performance is an inverted-U function of arousal — too little or too much both impair — and that the optimal level of arousal is lower for difficult tasks than for easy ones (Yerkes & Dodson, 1908).

Is arousal the same as emotion?
No. Arousal is the intensity or activation dimension; an emotion also requires content. In two-factor theory, the same arousal becomes different emotions depending on the cognitive label applied to it (Schachter & Singer, 1962).

How are arousal and attention related?
Arousal sets the general level of activation on which selective attention operates, and MeSH even files attention as a focused form of arousal. Rising arousal narrows the range of cues attended to, which helps up to a point and then hurts (Easterbrook, 1959).

What part of the brain controls arousal?
No single structure does, but the locus coeruleus — a brainstem nucleus that is the main source of norepinephrine — is a central regulator, adjusting neural gain through its phasic and tonic firing modes (Aston-Jones & Cohen, 2005).

What are the two dimensions of the circumplex model?
Valence (pleasant–unpleasant) and arousal (activated–deactivated). Every emotional state is located as a point in the plane they define, so arousal is a constitutive dimension of all affect rather than a by-product (Russell, 1980).

Why does high arousal sometimes make memory better and sometimes worse?
Because arousal biases competition rather than acting uniformly: it strengthens high-priority, salient representations while suppressing low-priority ones, so emotional details are remembered vividly at the expense of their surroundings (Mather & Sutherland, 2011).

Is arousal a single thing?
Increasingly, no. Cortical, autonomic, and behavioral measures can dissociate, and the locus coeruleus is now known to have modular, target-specific outputs, suggesting partially independent arousal channels rather than one global signal (Poe et al., 2020).

References