The Yerkes-Dodson Law Experiments (Arousal and Performance) – Robert Yerkes and John Dodson

Few empirical principles within the behavioral sciences have achieved the enduring cultural visibility, multidisciplinary ubiquity, and persistent pedagogical authority of the Yerkes-Dodson Law. Formulated in 1908 by comparative psychobiologist Robert Mearns Yerkes and his graduate student John Dillingham Dodson at the Harvard Psychological Laboratory, the principle arose not from clinical observations of human anxiety, but from an investigation into the associative learning capacities of Mus wagneri, the Japanese dancing mouse. What began as a modest empirical inquiry into how varying intensities of electrical stimulation modified visual discrimination habits gradually metamorphosed across the twentieth century into one of the central dogmas of psychophysiology: the non-linear, curvilinear relationship between physiological arousal and behavioral performance, popularly rendered as the “Inverted-U.”

The conceptual trajectory of the Yerkes-Dodson Law reflects the broader epistemological evolution of experimental psychology itself. Traversing the ideological disputes between late nineteenth-century introspection and early twentieth-century behaviorism, the principle was subsequently reinterpreted through the mid-century drive theories of Clark L. Hull and Donald O. Hebb, ultimately becoming anchored in the contemporary neurobiology of prefrontal catecholaminergic transmission, autonomic nervous system dynamics, and human factors engineering. In contemporary discourse, the law is cited to explain everything from stage fright and athletic choking under pressure to cockpit ergonomics, standardized test anxiety, and executive decision-making within high-stress organizational environments. Yet, beneath this veneer of universal applicability lies a complex, historically fraught, and scientifically contested reality.

To understand the genuine explanatory power and the profound methodological vulnerabilities of the Yerkes-Dodson Law, one must deconstruct both its historical genesis and its modern neurobiological operationalization. The law does not posit a simplistic, monolithic curve; rather, it articulates a dual-faceted hypothesis wherein performance initially ascends with increasing arousal up to an optimal inflection point before degenerating, with this optimal inflection point shifting dynamically as an inverse function of task complexity. This treatise presents an exhaustive examination of the original 1908 Harvard experiments, tracing their experimental design, mathematical formalization, mid-century transformation, neurochemical underpinnings, modern ecological applications, and the robust theoretical critiques that continue to challenge its foundational assumptions.

1. Historical Genesis and Intellectual Context of the 1908 Harvard Experiments

1.1 Foundations of Comparative Psychology at the Harvard Psychological Laboratory

The late nineteenth and early twentieth centuries witnessed a profound paradigm shift within psychological science, characterized by an aggressive departure from the introspective methodologies established by Wilhelm Wundt and Edward Bradford Titchener toward an objective, quantifiable science of comparative animal behavior. At Harvard University, this transition was profoundly shaped by William James and subsequently institutionalized by Hugo Münsterberg, who directed the Harvard Psychological Laboratory in Emerson Hall. Münsterberg championed an applied, physiological approach to mental phenomena, cultivating an academic environment wherein animal behavior was no longer viewed through the lens of anecdotal anthropomorphism, but as a substrate for rigorous mechanistic dissection.

Robert M. Yerkes arrived at Harvard during this transformative era, completing his doctorate under Münsterberg in 1902. Yerkes was deeply committed to elevating comparative psychology to the methodological standing of physical and physiological sciences. Influenced by Jacques Loeb’s theory of tropisms and C. Lloyd Morgan’s canon of parsimony, Yerkes sought to systematically eliminate speculative mentalism from the study of non-human organisms. The prevailing debates of the period centered on evolutionary continuity: if Darwinian principles held true across morphological structures, they must equally govern neurobehavioral adaptations. Comparative researchers faced the imperative of demonstrating how associative processes, sensory thresholds, and behavioral modifications could be tracked without inferring unobservable subjective states.

Habit formation emerged as the premier methodological vehicle for this empirical crusade. Unlike instinctual reflexes, which were viewed as static and phylogenetically hardwired, habits represented behavioral plasticity in action—the concrete, measurable reorganization of motor responses in response to environmental demands. Comparative laboratories raced to construct standardized mazes, puzzle boxes, and discrimination chambers capable of delivering replicable sensory inputs and recording unequivocal behavioral outputs. It was within this specific intellectual climate—marked by an intense desire for mathematical precision, operational control, and the eradication of teleological explanations—that Yerkes turned his attention to the precise quantitative relationship between sensory stimulation and the acceleration of habit acquisition.

1.2 The Collaborative Dynamic Between Robert M. Yerkes and John D. Dodson

By 1907, Robert Yerkes had established himself as a rising authority in the sensory physiology and behavior of lower organisms, having published extensive monographs on the behavioral repertoires of invertebrates, reptiles, and amphibians, alongside his seminal 1907 book, The Dancing Mouse: A Study in Animal Behavior. Yerkes possessed a meticulous, highly structured research agenda aimed at systematically cataloging the visual, auditory, and motor capabilities of non-human subjects. However, the sheer labor demanded by daily animal maintenance, apparatus construction, and the execution of thousands of individual discrimination trials required collaborative assistance, leading to the recruitment of graduate student John Dillingham Dodson.

John D. Dodson entered the Harvard Psychological Laboratory as a graduate student whose contribution to the historic 1908 paper, titled “The Relation of Strength of Stimulus to Rapidity of Habit-Formation,” published in the Journal of Comparative Neurology and Psychology, was profoundly practical and experimental. Dodson undertook the rigorous daily regimen of manual experimentation. This involved preparing the animals, calibrating the electrical induction coils, manipulating the physical trapdoors of the discrimination apparatus, and logging the raw trial-by-trial data. While Yerkes provided the conceptual architecture, theoretical framing, and laboratory infrastructure, Dodson’s experimental diligence generated the granular data tables upon which the entire empirical edifice rested.

Despite the historic immortality achieved by their joint publication, Dodson’s subsequent academic trajectory stands in stark contrast to that of Yerkes. While Yerkes ascended to the presidency of the American Psychological Association, directed the landmark Army Alpha and Beta intelligence testing programs during World War I, and founded the Yale Laboratories of Primate Biology, Dodson drifted into relative historical obscurity. After leaving Harvard, Dodson completed his doctoral work at the University of Minnesota and spent much of his subsequent career teaching at Western Kentucky State Normal School (now Western Kentucky University). Dodson’s relative absence from mainstream psychological literature across the twentieth century generated historical anomalies; for decades, textbooks routinely cited the “Yerkes-Dodson Law” while providing zero biographical context or institutional affiliation for Dodson, with some contemporary scholars even mistakenly assuming “Dodson” was an overlooked female researcher or a typographical error in laboratory historical logs.

1.3 Theoretical Inquiries into Motivation, Reinforcement, and Learning Rates

The primary theoretical impetus for the 1908 investigation was directly tied to emergent formulations of reinforcement, most notably Edward L. Thorndike‘s Law of Effect. In his pioneering 1898 dissertation on animal intelligence, Thorndike had posited that behavioral responses accompanied or closely followed by satisfaction to the animal would be more firmly connected with the situation, whereas responses accompanied or closely followed by discomfort would have their connections weakened. However, Thorndike’s formulations left critical operational questions unanswered regarding the quantitative scaling of aversive stimuli. Psychologists lacked a definitive empirical model describing whether the rate of habit acquisition was a strictly linear function of the magnitude of the motivating stimulus.

The common-sense assumption of the era, embedded within early drive theories, held that if a mild aversive stimulus fostered behavioral avoidance, a stronger aversive stimulus would logically accelerate this avoidance learning proportionally. Under a simple monotonic model, maximal drive should yield maximal learning efficiency. Yerkes and Dodson sought to subject this assumption to controlled empirical stress: Was the relationship between punishment intensity and acquisition velocity uniformly linear, or were there definitive boundary conditions where increasing the severity of negative reinforcement ceased to be pedagogically adaptive?

This theoretical ambiguity carried profound implications for the emerging discipline of experimental psychology. If learning curves were constrained by an upper ceiling of stimulus intensity—or worse, if excessive stimulus intensity actively precipitated behavioral disorganization—then the mechanisms governing associative learning could not be reduced to simple, additive increments of stimulus-response bonding. Yerkes and Dodson recognized that the organism’s physiological state under varying degrees of sensory perturbation constituted a critical, unmapped intervening variable between the presentation of an environmental stimulus and the successful execution of an adaptive behavioral habit.

2. Experimental Design, Apparatus, and Methodology of the Original Study

2.1 Subject Selection: The Japanese Dancing Mouse (Mus wagneri)

The empirical subjects selected for the 1908 experiments were not the standard albino laboratory rats (Rattus norvegicus) that would later dominate twentieth-century psychological research, but rather specimens of the Japanese dancing mouse (classified historically as Mus wagneri var. rotans). These diminutive, pigmented rodents were chosen largely because Yerkes had spent the preceding years extensively documenting their morphological, sensory, and behavioral idiosyncrasies. The mice were renowned for their incessant, circular rapid movements—a behavioral manifestation resulting from a genetic mutation causing inner ear malformations, specifically structural hypoplasia of the semicircular canals and vestibular apparatus, frequently accompanied by congenital deafness.

Yerkes rationalized the use of dancing mice on specific methodological grounds. Their congenital deafness eliminated uncontrolled auditory distractions from the ambient laboratory environment, allowing the experimenters to isolate visual stimuli as the pure discriminatory medium. Furthermore, their continuous, hyperactive exploratory drive provided a relentless baseline of locomotor activity, minimizing the extended periods of resting or motor freezing often observed in common rodents placed within novel environments. However, this physiological uniqueness introduced significant experimental artifacts: the animals’ vestibular dysfunction compromised their motor equilibrium, meaning that spatial navigation was inherently erratic, characterized by sudden whorls, pivots, and spasmodic bursts of movement.

Housing and acclimatization protocols were strictly maintained within the Cambridge laboratory. The mice were kept in temperature-regulated wooden and wire cages, sustained on a standardized diet of cracked corn, oats, bread soaked in milk, and water. Daily trials were executed at regularized intervals to control for diurnal cycles, and handling was systematized to minimize generalized human-handling panic. Despite these controls, the biological fragility of the dancing mice presented chronic challenges; their baseline physiological resilience was notably inferior to wild-type rodents, rendering them acutely vulnerable to physical exhaustion and electrical trauma.

2.2 The Visual Discrimination Chamber: Architecture and Electrical Grid

To quantify discrimination learning, Yerkes and Dodson engineered a custom-built behavioral apparatus constructed from wood, known as the discrimination box. The chamber measured approximately 94 centimeters in length, divided into three distinct operational sectors: an entrance chamber, an intermediate choice box, and two distinct visual exit pathways leading to a common nesting compartment. The apparatus was painted a uniform dull black internally to reduce extraneous optical reflections and ensure that the only salient visual differentials were the experimentally manipulated stimulus targets.

The critical technological innovation within this apparatus was the integration of an electrified floor grid. The floor of the choice compartments was outfitted with a series of parallel copper wires spaced at precise intervals of approximately one-half centimeter. This grid was connected to an electrical circuit designed to deliver a transient electric discharge through the footpads of the mouse whenever it crossed into the non-designated visual pathway. The physical engineering of the entryways, sliding trapdoors, and nesting sectors allowed the experimenter to manually guide the subject into the starting position, isolate it during the decision interval, and grant immediate access to the safe nesting box—which contained food and nesting materials—only when the correct visual doorway was breached.

The operational mechanics required constant manual oversight. Dodson sat positioned directly above the testing chamber, shielded behind an observational screen to prevent his own visual and olfactory cues from biasing the rodent’s spatial selection. When the mouse stood at the bifurcation of the choice chamber, it was confronted by two adjacent visual entryways: one illuminated or marked as the “correct” non-shock path, and the other wired directly to the induction coil. If the mouse crossed the threshold of the incorrect entryway, Dodson closed a key, completing the circuit and delivering an electric shock until the animal retreated or executed a correction. Each complete transit through the apparatus constituted a single trial, with the physical environment reset manually between runs.

2.3 Operationalization of Stimulus Intensity and Task Complexity

The independent variables in the 1908 study were rigorously operationalized across two discrete axes: stimulus intensity (the magnitude of the electrical shock) and task complexity (the perceptual difficulty of the visual discrimination). Stimulus intensity was generated using an induction coil powered by a single storage cell, with voltage calibrated through a sliding secondary coil over a calibrated centimeter scale. Yerkes and Dodson established three distinct operational levels of shock:

• Low Stimulus: A weak electrical current (calibrated around the physical detection threshold of the mouse), sufficient to provoke an orienting response or slight paw withdrawal without inducing gross motor disruption.
• Medium Stimulus: A moderate current that reliably elicited immediate discomfort, audible squeaking, and rapid physical retreat from the electrified grid.
• High Stimulus: A strong, highly aversive current that approached the physiological tolerance limit of the animal, often causing violent muscle twitches, vocal distress, and disorganized panic behavior.

Task complexity was operationalized through the manipulation of visual contrast between the two choice doorways. The discriminanda consisted of interchangeable cardboard panels forming the visual entryways, categorized into three grades of discriminatory difficulty:

• Condition I (Easy): A stark, high-contrast visual discrimination between a pure white doorway and a deep black doorway. The luminance differential was maximal and easily resolvable by the rodent’s visual system.
• Condition II (Medium): An intermediate contrast condition, where the black doorway was replaced with a dark gray panel, narrowing the photometric distance between the positive and negative targets.
• Condition III (Difficult): A subtle, low-contrast visual discrimination between two closely matched shades of light gray, demanding fine perceptual acuity to distinguish the correct pathway from the shock-laden pathway.

Habit acquisition was quantitatively defined by a rigorous mastery criterion: an animal was considered to have successfully acquired the visual discrimination habit only when it achieved three consecutive daily sets of ten errorless trials (thirty consecutive correct runs without a single shock). Errors were recorded whenever a mouse placed both front paws onto the electrified grid of the designated incorrect compartment. By methodically crossing the three levels of shock intensity against the three grades of task difficulty, Yerkes and Dodson constructed a 3×3 experimental matrix designed to illuminate the precise mechanics of learning efficiency.

3. Quantitative Data and Primary Findings of the 1908 Habit-Formation Research

3.1 Acquisition Rates Under Low, Medium, and High Shock Conditions

The quantitative results gathered by Yerkes and Dodson defied the simple, unidirectional expectations of classical reinforcement models. When the data across dozens of subjects and hundreds of trials were tabulated, the number of trials required to reach the mastery criterion varied significantly as a function of electrical shock intensity. However, this variation was not uniform across testing conditions, immediately demonstrating that negative reinforcement did not accelerate habit formation in a simple linear trajectory.

In the moderate discrimination condition, learning efficiency peaked under intermediate electrical voltages. Mice subjected to the weak electrical shock exhibited a slow, protracted learning curve; their rate of acquisition was sluggish, often requiring upward of 150 to 200 trials to achieve the errorless criterion. Because the punitive consequence of an error was minimal, the animals exhibited extensive exploratory behaviors at the choice point, repeatedly sampling the incorrect doorway with little behavioral urgency. Conversely, mice exposed to the moderate shock level demonstrated the steepest acquisition rates, rapidly decreasing their error frequencies and reaching the mastery criterion in substantially fewer trials (often between 50 and 80 runs).

Crucially, when the shock intensity was elevated to the highest setting within this intermediate task, habit formation did not accelerate further. Instead, the learning curve plateaued and began to deteriorate. The animals required an increased number of trials to master the identical visual habit compared to those trained under moderate shock. The data revealed an empirical reality: an optimal band of electrical intensity existed, below which the incentive to eliminate errors was insufficient, and above which the speed of cognitive mastery was actively degraded.

3.2 The Interaction Between Stimulus Strength and Task Difficulty

The paramount theoretical contribution of the 1908 study emerged not from the main effect of shock intensity alone, but from the complex, statistically profound interaction between stimulus intensity and task difficulty. Yerkes and Dodson observed that the specific shock intensity that produced the fastest learning rate shifted systematically depending upon the perceptual challenge presented by the visual doorways. This interaction formed the bedrock of what would later be formally christened the Yerkes-Dodson Law.

When the task was easy (the high-contrast white versus black discrimination), learning efficiency scaled in an almost direct, linear relationship with stimulus strength. Under this condition, the high shock intensity generated the most rapid acquisition, while the low shock produced the slowest. Because the perceptual cues were unmistakable, the mice experienced no confusion regarding which door was safe; the elevated electrical current served purely to galvanize behavioral avoidance of the black doorway without impairing their capacity to identify the target.

However, as the visual discrimination shifted to the difficult condition (fine shades of gray), this relationship dramatically inverted. Under high task difficulty, the strong electrical shock produced catastrophic decrements in learning efficiency. The mice subjected to intense shock in the difficult discrimination condition took an extraordinarily high number of trials to achieve mastery—in several instances failing to acquire the habit entirely within the experimental timeframe. Instead, it was the weak shock that produced the most rapid habit formation in the difficult discrimination task. Yerkes and Dodson explicitly summarized this foundational finding in their 1908 text: an easily acquired habit is formed most rapidly under high stimulus intensity, whereas a habit that is difficult to acquire is formed most rapidly under low stimulus intensity, with moderately difficult habits demonstrating peak efficiency under intermediate stimulus strengths.

3.3 Behavioral Anomalies, Stereotypies, and Stress-Induced Disruptions

Beyond the quantitative trial-and-error logs, Yerkes and Dodson documented qualitative behavioral anomalies that provided vital descriptive clues regarding why high stimulus intensity undermined learning in complex tasks. Under the influence of the highest electrical shock, the behavior of the dancing mice became profoundly disorganized. The animals ceased to engage in calm visual sampling of the stimuli at the choice junction; instead, their behavior devolved into panic-driven motor stereotypies.

Upon receiving severe shocks across their footpads, the mice routinely exhibited frantic, non-adaptive escape behaviors. These included rapid, erratic jumping against the glass ceiling of the chamber, persistent gnawing at the wooden seams, violent bouts of their congenital whirling movements, or, conversely, profound motor freezing wherein the animal huddled into a hyper-tonic ball and refused to advance through the apparatus. When forced into the choice area, subjects driven by extreme shock frequently dashed indiscriminately through whichever doorway was closest, prioritizing immediate physical escape from the grid over deliberate sensory discrimination of the visual panels.

This qualitative disorganization directly compromised the sensory mechanics of the task. For a mouse to distinguish two closely matched shades of gray, it was biologically necessary for the animal to pause at the choice point, orient its retina toward the target panels, and process the subtle luminance discrepancies. The extreme physiological perturbation provoked by the severe electric shock preempted this critical visual inspection phase. The animal’s sensory attention was completely captured by the visceral anticipation of pain and immediate locomotor flight, rendering subtle cognitive discrimination virtually impossible.

4. Mathematical Modeling and the Inverted-U Hypothesis Formulation

4.1 Deconstruction of the Curvilinear Paradigm: Defining the Inverted-U

The empirical findings of the 1908 study established the foundation for the curvilinear conceptualization of behavioral efficiency, universally recognized today as the Inverted-U hypothesis. Mathematically, this construct asserts that the relationship between an activating drive (arousal, stress, or stimulus strength) and performance is not monotonic, but quadratic—resembling a parabolic or normal distribution curve when plotted on a two-dimensional Cartesian plane, with arousal along the horizontal abscissa ($x$-axis) and performance efficiency along the vertical ordinate ($y$-axis).

The operational geometry of this curve can be analytically segmented into three distinct functional zones:

• The Hypotonic Zone (Sub-Optimal Arousal): Located along the ascending left arm of the curve, this zone is characterized by insufficient physiological activation. In this state of behavioral hypoarousal, task engagement is lethargic, attentional focus is diffuse, reaction times are protracted, and habit acquisition rates are profoundly suppressed due to a lack of motivational salience.
• The Homeostatic Apex (Optimal Arousal Zone): The inflection point where the first derivative of the curve equals zero ($dy/dx = 0$). At this critical apex, physiological activation and cognitive resources are harmoniously aligned. Attentional allocation is focused precisely upon task-relevant cues, sensory processing speed is maximized, and behavioral output achieves peak efficiency.
• The Hypertonic Zone (Supra-Optimal Arousal): Located along the descending right arm of the curve, this zone represents excessive activation. Here, escalating physiological tension induces cognitive fragmentation, emotional distress, motor tremor, perceptual narrowing, and systemic degradation of adaptive behavioral responses.

This curvilinear paradigm fundamentally challenged early linear reductionism in psychological mechanics. It established that more stimulation is not inherently better; rather, behavioral efficiency is governed by a principle of non-linear homeostatic balance, wherein deviation in either direction—hypo-activation or hyper-activation—precipitates a catastrophic decay in operational competence.

4.2 The Two Primary Tenets of the Yerkes-Dodson Law

In formal psychological theory, the Yerkes-Dodson Law is structurally codified not as a single empirical statement, but as two distinct, codependent principles. Conflating these two tenets is one of the most common conceptual errors in both popular and academic literature.

Tenet One (The Curvilinear Principle): The quality of cognitive and behavioral performance on any given task is an inverted U-shaped function of physiological arousal or stimulus strength. Performance improves with increasing stimulation up to an optimal apex, beyond which further increments in stimulation yield progressive, non-linear decrements in operational efficacy.

Tenet Two (The Task-Complexity Inversion Principle): The optimal level of arousal necessary to achieve peak performance varies as an inverse function of the cognitive, perceptual, or motor complexity of the task. For cognitively complex, highly nuanced, or intellectually demanding tasks, the optimal arousal apex shifts substantially to the left (requiring low to moderate arousal). Conversely, for simple, automated, highly practiced, or purely gross-motor tasks, the optimal arousal apex shifts substantially to the right (tolerating or requiring elevated levels of arousal).

A vital distinction must also be maintained between task execution efficacy (the speed and accuracy with which an established habit is mobilized) and acquisition velocity (the rate at which a novel habit is initially forged within neural pathways). Yerkes and Dodson’s original experiments specifically measured the latter—the velocity of habit acquisition. In modern applied psychology, however, the law is almost universally applied to execution efficacy. While acquisition and execution share overlapping neurobiological constraints, the cognitive load imposed during the initial acquisition phase of learning is substantially higher than that required to execute a previously automatized psychomotor habit, making the initial learning phase exceptionally vulnerable to hyperarousal-induced collapse.

4.3 Parametric Variations and Asymmetrical Curves

While theoretical textbooks invariably illustrate the Inverted-U as a perfectly symmetrical, Gaussian bell curve, real-world psychophysiological data demonstrate that these functions are rarely symmetrical. In empirical settings, performance curves frequently exhibit pronounced parametric variations, presenting as highly skewed distributions depending upon the nature of the cognitive operations involved.

Tasks heavily contingent upon subtle executive functioning, working memory capacity, and delicate motor stability produce curves that are acutely left-skewed. In these configurations, performance rises gradually through a narrow hypotonic band, achieves its apex at a low threshold of physiological activation, and then crashes abruptly down a steep cliff of hyperarousal. In contrast, gross-motor tasks requiring maximal explosive strength, sheer physical stamina, or simple ballistic trajectories (such as a powerlift or a sprint) generate right-skewed curves. Here, performance ascends across a wide range of mounting arousal, with the apex positioned at near-maximal physiological excitation, exhibiting only a minor degradation when extreme autonomic overload finally induces structural motor fatigue.

Furthermore, psychometricians must account for severe floor and ceiling effects. A ceiling effect occurs when a task is so elemental that baseline performance remains near one hundred percent accuracy across virtually all arousal levels, disguising the descending arm of the curve. A floor effect occurs when a task is so excessively challenging that the subject cannot resolve the discriminanda even under optimal baseline arousal, generating an ostensibly flatline performance profile. Mathematical formalizations attempting to predict the exact coordinates of the optimal inflection point ($A_{opt}$) must therefore incorporate weighting parameters that quantify both intrinsic task complexity ($C$) and the individual subject’s basal stress reactivity ($R$):

$$A_{opt} = f\left(\frac{1}{C}\right) \times \gamma(R)$$

where $\gamma(R)$ serves as an internal biological scaling factor governing the subject’s idiosyncratic autonomic sensitivity.

5. The Moderating Variable of Task Complexity and Cognitive Load

5.1 Perceptual Discrimination versus Automated Sensorimotor Routines

The neurocognitive mechanisms governing the interaction between arousal and performance are fundamentally rooted in the structural distinction between high-level perceptual discrimination and automated sensorimotor routines. Tasks requiring fine perceptual discrimination—such as the fine gray calibrations in the Yerkes-Dodson apparatus, or modern tasks like diagnostic radiological screening, sonar monitoring, and complex financial risk modeling—depend critically on deliberate, top-down cognitive governance. These processes require continuous sensory integration, signal-to-noise optimization, and internal error-checking mechanisms that are acutely susceptible to interference from physiological stress.

In stark contrast, automated sensorimotor routines operate largely via subcortical and cerebellar pathways once they are thoroughly consolidated. When an action sequence has been overlearned to the point of automaticity, its execution requires minimal direct supervisory intervention from the prefrontal cortex. Under conditions of physiological hyperarousal, these hardened, automated habits exhibit exceptional durability. In fact, elevated autonomic activation—characterized by surges in peripheral adrenaline, accelerated heart rates, and heightened muscle spindle excitability—actively facilitates the ballistic throughput of simple motor outputs by lowering the firing threshold of peripheral motor units.

Consequently, when an individual is thrust into an acute survival situation or an environment of high evaluative panic, their capacity to execute deeply ingrained physical habits remains intact or is transiently augmented. However, their capacity to formulate novel mental associations, detect subtle perceptual anomalies, or pivot their cognitive focus away from dominant behavioral instincts collapses completely. Hyperarousal strips the cognitive apparatus down to its most primitive, well-rehearsed behavioral baselines.

5.2 Easterbrook’s Cue-Utilization Hypothesis

The most influential theoretical framework bridging the Inverted-U to cognitive mechanics was formulated by James A. Easterbrook in his seminal 1959 paper, “The Effect of Emotion on Cue Utilization and the Organization of Behavior.” Easterbrook posited that the critical intervening variable linking escalating physiological arousal to altered performance is the progressive, systematic reduction in the range of environmental cues utilized by the organism during task execution—a phenomenon widely referred to as attentional narrowing or “tunnel vision.”

Easterbrook’s model articulates the mechanics of the Inverted-U through a rigorous attentional lens:

• Under Low Arousal: The organism’s attentional aperture is excessively wide. The individual processes not only task-relevant signals but also an expansive array of irrelevant peripheral stimuli. This sensory promiscuity introduces cognitive noise, prolonging decision latencies and compromising task efficiency through distractibility.
• Under Moderate (Optimal) Arousal: The attentional field narrows to an optimal degree. Elevated physiological drive suppresses the sensory processing of irrelevant peripheral distractors while preserving full cognitive access to central, task-critical cues. Performance peaks because the cognitive system is effectively filtered, dedicating its full processing bandwidth exclusively to task-relevant information.
• Under High Arousal (Hyperarousal): The attentional narrowing process continues unabated, breaching the threshold of cognitive safety. The attentional aperture becomes so severely constricted that task-relevant cues are systematically excluded alongside the irrelevant ones. The individual develops profound perceptual tunnel vision, focusing obsessively on a solitary environmental signal (such as a weapon, an error alert, or an electrical shock grid) while becoming utterly blind to surrounding critical data streams necessary for systemic problem-solving.

The Easterbrook hypothesis provides an elegant, mechanistic explanation for the second tenet of the Yerkes-Dodson Law. A simple task requires the monitoring of very few perceptual cues; therefore, severe attentional narrowing does not harm—and may even protect—performance by forcefully eliminating any extraneous environmental distraction. A complex task, conversely, demands the simultaneous integration of a vast, multidimensional array of cues. Any constriction of the attentional field beyond baseline inevitably excises critical task parameters, precipitating immediate cognitive failure.

5.3 Working Memory Capacity and Executive Function Constraints

Modern cognitive psychology has enriched Easterbrook’s framework by mapping the effects of arousal onto the finite resources of working memory capacity and central executive function. As conceptualized by Alan Baddeley and contemporary cognitive neuroscientists, working memory is the active workspace of the human mind, responsible for the transient maintenance, manipulation, and updating of goal-relevant representations in the face of interference.

When an organism experiences escalating physiological stress and high subjective anxiety, an immense volume of working memory bandwidth is seized by intrusive, threat-related cognitive processing. In human subjects, these take the form of worry, iterative catastrophizing, hyper-vigilant scanning for danger, and the conscious monitoring of physiological somatic symptoms (such as palpitations and tremors). These intrusive cognitive streams act as computational “parasites,” directly consuming the limited executive resources of the central executive and the phonological loop.

Because simple cognitive tasks make minimal demands on working memory capacity, they can be executed seamlessly even when a significant portion of cognitive bandwidth is diverted by stress. Complex cognitive tasks, however—such as mental arithmetic, complex spatial navigation, linguistic syntax generation, and multi-step deduction—require nearly 100% of available working memory capacity. Under states of hyperarousal, the invasion of threat-related processing starves the primary task of computational resources. The prefrontal networks lose their ability to sustain task-relevant rules, monitor performance errors, and suppress prepotent, non-adaptive behavioral impulses, resulting in a systemic performance collapse.

6. Mid-Century Rediscovery and Donald Hebb’s Neurological Reconceptualization

6.1 The Decades of Latency: From Comparative Curiosity to Mainstream Theory

Despite its foundational status today, the 1908 publication by Yerkes and Dodson did not achieve immediate, widespread acclaim. For nearly four decades following its appearance in the Journal of Comparative Neurology and Psychology, the paper languished in relative citation obscurity. During the 1920s and 1930s, American psychology was utterly dominated by the classical behaviorism of John B. Watson and subsequently the radical behaviorism of B.F. Skinner, alongside the formal mathematical drive-reduction formulations of Clark L. Hull.

The dominant Hullian paradigm formalized behavioral potential ($SER$) as a direct multiplicative product of habit strength ($SHR$) and drive ($D$):

$$SER = SHR \times D$$

Under this early Hullian equation, drive was conceived as a non-specific, linear energizer of all existing habits. If drive was elevated, the strength of the evoked behavior was mathematically predicted to increase monotonically. Hull’s model struggled conceptually to accommodate the downward arm of the Inverted-U, viewing drive primarily through the lens of homeostasis: hunger, thirst, or shock clearance simply motivated the organism to engage in actions that reduced physiological tension. In this theoretical milieu, Yerkes and Dodson’s observation that high shock actively disintegrated habit acquisition was treated as an empirical oddity or a minor methodological artifact of their unusual rodent subjects, rather than a universal law of behavioral organization.

The citation index of the 1908 paper remained exceptionally flat throughout the interwar period. The broader psychological establishment remained fixated on establishing unified, monotonic mathematical laws of learning. It was not until the mid-twentieth-century collapse of pure drive-reduction theories—and the concurrent birth of physiological psychology and early cognitive science—that researchers were forced to reconsider the fundamental reality that excessively high motivational drive actively poisons cognitive performance.

6.2 Donald O. Hebb’s 1955 Paradigm Shift: Linking Drive and Cortical Arousal

The critical historical turning point that rescued the 1908 findings from comparative obscurity occurred in 1955, when Canadian neuropsychologist Donald O. Hebb published his transformative presidential address to the American Psychological Association, titled “Drives and the C.N.S. (Conceptual Nervous System)” in the Psychological Review. Hebb orchestrated a profound theoretical synthesis, directly unifying the abstract psychological concept of “drive” with newly discovered neurophysiological mechanisms of brainstem cortical activation.

Hebb argued that sensory stimulation serves two entirely distinct, parallel neurological functions when it breaches the central nervous system:

• The Cue Function (Information Transmission): The traditional message-carrying capacity of sensory signals, traveling along specific afferent pathways through sensory thalamic relays directly to primary sensory projection areas of the cerebral cortex, providing explicit data regarding what an environmental event is.
• The Arousal Function (Energizing Tone): A non-specific, general energizing capacity, driven by collaterals branching off the main sensory tracts into the core of the brainstem, specifically activating the reticular formation. This pathway serves to tone up the entire cerebral mantle, dictating how alert or receptive the cortex is to receiving the cue information.

Hebb explicitly utilized this dual-pathway model to resurrect and universalize the Inverted-U curve. He asserted that for any incoming cortical message to be processed, organized, and acted upon effectively, the cortex requires an optimal baseline of non-specific electroencephalographic desynchronization mediated by the reticular system. If arousal is too low, the incoming cue function is lost in a sea of slow-wave, synchronized cortical inertia; the animal falls asleep or fails to process the signal. If arousal is excessively high, the ascending bombardment from the reticular core becomes so intense that cortical networks are overloaded with non-specific sensory noise, culminating in emotional disorganization, behavioral agitation, and cognitive paralysis. Hebb explicitly plotted this neurobiological dynamic as an Inverted-U, linking his theoretical model directly back to the empirical learning curves charted by Yerkes and Dodson nearly a half-century earlier.

6.3 Semantic and Conceptual Evolution: Conflation of Core Constructs

Following Hebb’s 1955 breakthrough, the academic floodgates opened, but this modern resurgence catalyzed a massive semantic and conceptual evolution that significantly detached the theory from its empirical origins. The precise, operationalized independent variable used by Yerkes and Dodson—the physical intensity of an alternating electric shock delivered to the footpads of a rodent—was rapidly subsumed by a wide array of generalized, psychological constructs.

Across the 1960s and 1970s, researchers systematically substituted “electric shock magnitude” with broad umbrella terms: “generalized arousal,” “activation level,” “emotional stress,” “drive,” “trait and state anxiety,” and “evaluative motivation.” While this linguistic transition granted the Yerkes-Dodson framework unprecedented theoretical scope—allowing it to be applied to human clinical pathology, organizational management, and competitive athletics—it simultaneously introduced severe conceptual vulnerabilities. External environmental punishment (nociceptive physical pain) is fundamentally not identical to internal neurochemical arousal, nor is task motivation synonymous with generalized autonomic panic.

Critics began to warn that experimental psychology had fallen prey to an expansive category error. Equating an animal’s physical avoidance of electrical shock trauma with a human student’s internal cognitive apprehension during a differential calculus examination conflated entirely disparate physiological cascades. By treating “arousal” as a monolithic, undifferentiated unitary state, mid-century theorists laid the groundwork for decades of methodological contradictions, obscuring the nuanced, multidimensional neurobiology governing human stress responses.

7. Neurobiological Substrates of Arousal and Behavioral Performance

7.1 The Ascending Reticular Activating System (ARAS) and Cortical Tone

The anatomical validation of Hebb’s arousal framework arrived through the landmark work of Giuseppe Moruzzi and Horace Magoun in 1949, who identified the Ascending Reticular Activating System (ARAS) within the brainstem core. Extending from the rostral medulla through the pontine tegmentum into the midbrain, the ARAS comprises a complex network of interconnected nuclei that project diffusely to the intralaminar nuclei of the thalamus and onward across the entire neocortex.

Electrical stimulation of this reticular core transforms the electrical activity of the cerebral cortex, instantaneously shifting an animal from high-amplitude, low-frequency, synchronized electroencephalographic (EEG) slow-waves (characteristic of drowsiness, lethargy, or deep sleep) to low-amplitude, high-frequency, desynchronized beta and gamma rhythms (indicative of active sensory orientation, mental engagement, and vigilant alertness). This reticular-thalamocortical circuitry functions as the master biological rheostat of the central nervous system, establishing the general baseline of cortical tone upon which all cognitive discrimination depends.

The thalamus serves as the critical regulatory gateway in this network. The thalamic reticular nucleus (TRN) acts as an inhibitory gatekeeper, physically wrapping around the sensory thalamus and modulating the throughput of sensory data destined for cortical analysis. Under optimal ascending ARAS drive, the TRN selectively suppresses background somatic noise while permitting high-fidelity transmission of salient, task-relevant sensory inputs. However, under states of pathological hyperarousal, the ascending monoaminergic and cholinergic bombardment completely destabilizes this thalamic sensory filter. The gate is blown wide open, flooding the neocortex with unprocessed, sensory input that destabilizes executive neuronal assemblies and disrupts attentional integrity.

7.2 Catecholaminergic Dynamics: Norepinephrine and Dopamine in the Prefrontal Cortex

At the highest levels of cognitive processing, the molecular mechanics of the Yerkes-Dodson Law have been brilliantly decoded by modern neurobiologists, most prominently by Amy F.T. Arnsten and her colleagues at Yale University. Arnsten’s research has mapped the precise, non-linear neurochemical actions of catecholamines—specifically norepinephrine (NE) originating from the locus coeruleus and dopamine (DA) from the ventral tegmental area—upon the structural architecture of the prefrontal cortex (PFC).

The prefrontal cortex is the executive epicenter of the primate brain, responsible for sustained attention, working memory representation, abstract reasoning, and behavioral inhibition. Prefrontal pyramidal cell networks sustain mental representations through recurrent excitation via microcircuits known as “delay-period assemblies.” The operational integrity of these delicate assemblies is fundamentally governed by a homeostatic catecholaminergic balance characterized by an inverted U-shaped function, driven entirely by the differential binding affinities of catecholaminergic receptor subtypes:

• Optimal Arousal (Moderate Catecholamines): Under conditions of moderate, non-stressful alertness, modest levels of norepinephrine and dopamine are released into the PFC. Norepinephrine binds preferentially to high-affinity Alpha-2A ($\alpha_2\text{A}$) adrenoceptors located on dendritic spines. Activation of these $\alpha_2\text{A}$ receptors inhibits cyclic adenosine monophosphate (cAMP) production, closing nearby hyperpolarization-activated cyclic nucleotide-gated (HCN) potassium channels. This channel closure prevents signal leakage, dramatically strengthening the network’s synaptic connectivity and boosting the “signal” of goal-directed representations. Simultaneously, moderate dopamine binds to high-affinity D1 receptors, which sculpt the network by pruning away weak, irrelevant background firing (“noise”). Together, this optimal receptor engagement maximizes the neuronal signal-to-noise ratio, fostering peak working memory and executive focus.
• Hyperarousal and Acute Stress (Excessive Catecholamines): When an individual encounters high-threat, high-stress environments, the locus coeruleus and ventral tegmental area fire at extreme frequencies, flooding the PFC with toxic concentrations of NE and DA. Under these conditions, the surplus norepinephrine spills over to engage lower-affinity Alpha-1 ($\alpha_1$) adrenoceptors, while excessive dopamine hyper-activates low-affinity D1 signaling pathways. Activation of $\alpha_1$ receptors initiates intracellular protein kinase C (PKC) cascades, while excessive D1 stimulation drives cAMP and protein kinase A (PKA) pathways. These kinase cascades phosphorylate and force open potassium channels (both HCN and KCNQ channels) along the dendritic shafts of pyramidal neurons. The physical opening of these channels causes massive membrane shunt currents; electrical resistance collapses, and action potentials fail to propagate across synaptic connections. Prefrontal microcircuits essentially “un-couple,” systematically dismantling the working memory networks required for deliberative cognitive processing.

This neurochemical mechanism provides an undeniable, biophysical validation of the Yerkes-Dodson Inverted-U. The prefrontal cortex is structurally engineered to function optimally within a narrow neurochemical window. When catecholaminergic concentrations breach this physiological ceiling, the executive networks do not merely slow down; they undergo an acute, functional disconnection designed to shift control of behavior away from deliberative cortical oversight.

7.3 Amygdalar Modulation and the Subcortical Shift

The catastrophic shutdown of prefrontal microcircuits under hyperarousal is not an evolutionary defect; it represents an ancient, phylogenetically conserved survival adaptation orchestrated by the amygdala. Under baseline conditions of low to moderate arousal, the prefrontal cortex exerts top-down inhibitory governance over subcortical structures, dampening instinctive emotional reflexes through descending projections from the ventromedial prefrontal cortex to the intercalated cell masses of the amygdala.

However, when sensory inputs signify severe, imminent environmental threat, the basolateral amygdala (BLA) becomes hyper-activated. The amygdala projects robustly to the brainstem monoaminergic nuclei, triggering the catastrophic catecholamine surge described above that disengages prefrontal control. Simultaneously, the central nucleus of the amygdala initiates rapid, unmediated motor, autonomic, and endocrine outputs via the hypothalamus and periaqueductal gray. This dynamic induces an immediate, functional subcortical shift: behavioral governance is violently ripped away from the reflective, deliberative prefrontal networks and handed over to the reflexive, habitual, survival-oriented structures of the basal ganglia, amygdala, and brainstem.

From an evolutionary survival standpoint, this subcortical shift is exceptionally adaptive. When a hominid or a rodent is confronted by a lethal predator, slow, deliberative reasoning (e.g., “Which escape route maximizes energy efficiency?”) carries lethal latencies; the organism requires rapid, ballistic, hardwired reflexes (freezing, explosive flight, or aggressive fight). In modern human environments, however—such as taking a standardized licensing exam, landing a malfunctioning jetliner, or executing complex tactical decisions—the survival-oriented subcortical reflexes are functionally maladaptive. The individual is left with high-energy motor drive, acute perceptual narrowing, and zero access to the complex deductive faculties required to navigate the crisis.

8. The Role of the Autonomic Nervous System and Endocrine Stress Response

8.1 Sympathetic-Adrenomedullary (SAM) Axis Activation

The peripheral physiological manifestations that parallel the ascending arm of the Yerkes-Dodson curve are driven by the instantaneous firing of the Sympathetic-Adrenomedullary (SAM) axis. Mediated through the sympathetic preganglionic splanchnic nerves innervating the adrenal medulla, this pathway triggers the immediate, exocytotic release of the catecholamines epinephrine (80%) and norepinephrine (20%) into the systemic bloodstream within hundreds of milliseconds of stress detection.

Peripheral catecholaminergic discharge initiates a systemic physiological mobilization designed to optimize muscular performance: beta-1 adrenergic receptors in the heart increase cardiac inotropy and chronotropy (escalating cardiac output), alpha-1 adrenergic receptors induce peripheral vasoconstriction to non-essential visceral vascular beds while shunting blood directly toward major skeletal muscle groups, and beta-2 adrenergic receptors drive bronchodilation, maximizing alveolar oxygen uptake. Concurrently, sympathetic stimulation triggers pupil dilation (mydriasis) via the pupillary dilator muscle, expanding visual fields, and stimulates eccrine sweat glands, generating the transient increases in skin conductance documented via electrodermal activity (EDA) monitoring.

A critical modern biomarker utilized to index this autonomic tone in human performance settings is Heart Rate Variability (HRV)—specifically the vagally-mediated root mean square of successive differences between normal heartbeats (RMSSD). High baseline HRV reflects robust parasympathetic (vagal) tone, signifying a dynamic, flexible autonomic nervous system capable of rapid, nuanced adjustments between energetic mobilization and rapid recovery. Individuals exhibiting high HRV baselines possess an expanded “optimal arousal zone” along the Yerkes-Dodson continuum, demonstrating elevated resilience against the cognitive degradation typically precipitated by acute sympathetic surge.

8.2 Hypothalamic-Pituitary-Adrenal (HPA) Axis and Glucocorticoid Kinetics

Operating on a vastly different temporal scale than the immediate SAM axis, the endocrine arm of the stress response is mediated by the Hypothalamic-Pituitary-Adrenal (HPA) axis. Initiated by the secretion of Corticotropin-Releasing Hormone (CRH) from the paraventricular nucleus of the hypothalamus, the cascade triggers the release of Adrenocorticotropic Hormone (ACTH) from the anterior pituitary, which travels hematogenously to the adrenal cortex, stimulating the synthesis and systemic release of glucocorticoids (primarily cortisol in humans, corticosterone in rodents).

Cortisol kinetics are characterized by delayed systemic circulation, peaking within target tissues approximately 20 to 40 minutes following stress onset. Because glucocorticoids are lipophilic steroid hormones, they traverse the blood-brain barrier with ease, binding directly to two distinct nuclear receptor subtypes within the central nervous system:

• Mineralocorticoid Receptors (MR): Characterized by an exceptionally high affinity for cortisol (binding the hormone even under low, baseline physiological conditions), MRs are densely expressed within the hippocampus and limbic circuits, maintaining baseline cellular homeostasis, neurogenesis, and normative sensory processing.
• Glucocorticoid Receptors (GR): Characterized by a ten-fold lower affinity for cortisol, GRs remain largely unoccupied during calm, resting states. They become saturated only during significant diurnal peaks or severe, acute HPA-axis stress cascades.

The activation profile of these receptors creates a profound biphasic, inverted U-shaped function governing hippocampal synaptic plasticity and Long-Term Potentiation (LTP). When cortisol elevations are moderate—saturating high-affinity MRs while only partially engaging GRs—LTP is enhanced, accelerating the memory consolidation of significant environmental events. However, when severe stress forces the widespread saturation of low-affinity GRs, hippocampal LTP is actively suppressed, and Long-Term Depression (LTD) is facilitated. This glucocorticoid kinetics directly explains the clinical paradox wherein acute high-arousal stress significantly accelerates the consolidation of a traumatic memory (such as in PTSD etiology) while simultaneously inflicting catastrophic impairments upon the active retrieval of previously learned declarative knowledge (such as during a high-stakes examination).

8.3 Allostatic Load and the Chronic Degradation of Performance

The Yerkes-Dodson Law was originally conceived as a model for acute, episodic adaptations to discrete environmental challenges. However, when physiological arousal is transformed from an acute, transient mobilization into a state of chronic, sustained activation, the biological mechanics undergo a destructive transition termed allostatic load and allostatic overload, a framework pioneered by neuroendocrinologist Bruce McEwen.

Allostasis refers to the active biological process through which the body maintains physiological stability (homeostasis) through dynamic internal change. When the neurochemical mediators of allostasis—cortisol, epinephrine, inflammatory cytokines, and metabolic markers—remain pathologically elevated across weeks or months, the physiological systems undergo progressive structural degradation. Within the central nervous system, persistent high-arousal states drive profound structural remodeling:

• Prefrontal Cortical Atrophy: Chronic glucocorticoid and catecholamine exposure drives the progressive retraction and de-branching of dendrites in prefrontal pyramidal neurons, accompanied by a quantifiable loss of asymmetric, excitatory spine synapses, degrading structural working memory capacity.
• Hippocampal Neurotoxic Remodeling: Sustained GR hyper-activation suppresses adult neurogenesis within the subgranular zone of the dentate gyrus and induces hippocampal dendritic atrophy, fundamentally destabilizing long-term spatial and contextual memory systems.
• Amygdalar Hypertrophy: In direct contrast to the atrophy observed in the PFC and hippocampus, the basolateral amygdala responds to chronic allostatic overload by undergoing structural dendritic growth and hyper-branching. This increases synaptic connectivity within threat-detection networks, permanently lowering the threshold for stress reactivity and locking the organism into an enduring state of pathological hyper-arousal.

Consequently, allostatic overload fundamentally reshapes the geometric morphology of the Yerkes-Dodson curve for that individual. The apex of optimal performance collapses downward, and the entire curve shifts dramatically to the left. The chronic stress survivor enters cognitive testing or high-stakes vocational challenges with their baseline arousal pre-positioned immediately adjacent to the hyper-tonic drop-off point; even minor environmental challenges are sufficient to trigger complete neurobehavioral disintegration.

9. Modern Psychological Applications: Cognitive Function, Education, and Testing

9.1 Evaluative Threat, Test Anxiety, and Academic Attainment

Within educational psychology and psychometrics, the principles of the Yerkes-Dodson Law find immediate empirical expression in the phenomena of evaluative threat and test anxiety. High-stakes academic assessments—such as standardized university entrance examinations, medical board licensing, or legal bar examinations—generate massive autonomic and psychological stress responses in candidates, elevating physiological arousal far beyond normative baseline levels.

In accordance with Easterbrook’s cue-utilization model and working memory depletion paradigms, severe test anxiety directly degrades performance on tasks characterized by high cognitive complexity. When a student is confronted by complex mathematical deductions, dense reading comprehension, or multi-step logic problems, their available working memory capacity is severely constrained. If a substantial fraction of that finite capacity is co-opted by autonomic distress—racing heart, gastrointestinal distress—and recurring intrusive thoughts regarding the punitive life consequences of academic failure, the student experiences a functional drop in operational IQ.

This dynamic is significantly amplified by the phenomenon of stereotype threat, wherein individuals belonging to an identity group stereotyped as possessing inferior ability in a specific domain (such as women in advanced mathematics or minority groups on standardized aptitude metrics) experience an additional, agonizing tier of evaluative hyperarousal. The chronic dread of confirming the negative societal stereotype acts as an autonomic multiplier, forcefully driving these students past the optimal apex of the Yerkes-Dodson curve into the descending hypertonic zone. Modern pedagogical interventions engineered to counter this performance degradation—such as expressive writing exercises prior to testing, mindfulness-based stress reduction, and framing testing as non-evaluative challenge rather than punitive threat—act directly to dampen autonomic over-activation, effectively moving the student’s arousal state backward along the curve to re-align with the optimal performance apex.

9.2 Creative Cognition, Divergent Thinking, and Incubation Effects

The relationship between arousal and creative cognition presents one of the clearest demonstrations of the task-complexity inversion tenet. While convergent thinking—the ability to identify a single, orthodox, mathematically correct solution to a clearly defined problem—can be successfully executed under moderate to high levels of focused arousal, divergent thinking (the generation of novel, unexpected, and remote semantic associations) requires a uniquely relaxed, diffuse neurobehavioral architecture.

Neuroimaging investigations reveal that high-level creative ideation relies upon the dynamic, flexible interplay between the brain’s Default Mode Network (DMN)—associated with spontaneous, unconstrained mental simulations and remote conceptual retrieval—and the Central Executive Network (CEN), responsible for evaluating and formalizing these ideas. Under conditions of high physiological arousal and intense evaluative stress, elevated catecholaminergic tone forces the attentional field to contract tightly around immediate, highly accessible, prepotent mental representations. The individual experiences profound functional fixedness and algorithmic rigidity, falling back on well-rehearsed, conventional cognitive scripts.

To access remote semantic nodes across associative neural networks, cortical tone must be maintained within a low to moderate arousal window, where diffuse attention is preserved. This explains the legendary cognitive utility of the “incubation effect”: individuals wrestling with seemingly intractable creative or scientific challenges frequently achieve breakthrough epiphanies not during intense, high-arousal desk work, but during low-arousal, mundane activities—such as taking a walk, showering, or resting. In these relaxed, hypotonic states, the lack of intense attentional narrowing permits the spontaneous, bottom-up cross-talk of neural networks, allowing novel conceptual associations to coalesce into conscious awareness.

9.3 Instructional Design and Cognitive Load Theory Integration

In contemporary instructional design, the Yerkes-Dodson Law has been seamlessly integrated into John Sweller’s Cognitive Load Theory, which categorizes mental processing demands into three distinct vectors: intrinsic load (the foundational, unalterable complexity of the material itself), germane load (the mental capacity dedicated to processing information, building schemas, and deep learning), and extraneous load (the unnecessary mental friction imposed by poor educational delivery, chaotic environments, or confusing design).

When educational environments are constructed without regard for psychophysiological constraints, extraneous load acts in direct synergy with autonomic hyperarousal. An unorganized classroom environment, highly ambiguous instructions, public cold-calling of students, and intense time pressures impose massive extraneous loads that rapidly accelerate learner arousal into the hypertonic zone. Because the learner’s intrinsic capacity is already taxed by the intrinsic load of the novel educational material, the total cognitive demand breaches the physiological limit of working memory. Learning efficiency disintegrates, resulting in cognitive overload, frustration, and behavioral shutdown.

Instructional scaffolding techniques are engineered precisely to modulate this load-arousal nexus. By deconstructing massive, multifaceted tasks into small, progressive, modular milestones, educators systematically reduce early task complexity. This pedagogical reduction allows students to acquire novel skills while operating safely beneath their individual arousal ceilings. Once the basic structural sub-skills are consolidated and transferred into long-term memory schemas (thereby requiring minimal active working memory bandwidth), the educator can systematically re-introduce task complexity, time constraints, and evaluative pressure—effectively training the learner to maintain high-performance execution across wider, more demanding arousal profiles without triggering cognitive collapse.

10. Performance in High-Stakes Environments: Athletics, Aviation, and Ergonomics

10.1 Sports Psychology and the Individual Zones of Optimal Functioning (IZOF)

The application of the Inverted-U within sports psychology has generated some of the most dynamic theoretical debates in behavioral science. For decades, traditional coaches relied upon the simple Yerkes-Dodson paradigm to prescribe universal relaxation techniques to over-aroused athletes and “pep talks” to under-aroused teams. However, the rigid universality of this single-curve model was systematically dismantled by sports psychologist Yuri Hanin through his development of the Individual Zones of Optimal Functioning (IZOF) model.

Hanin’s empirical investigations into elite athletes revealed that the optimal arousal apex is not an absolute, fixed point determined solely by the mechanical complexity of the sporting discipline; rather, it is deeply idiosyncratic and personal. While fine-motor, high-precision sports (such as competitive archery, biathlon shooting, and tournament golf) broadly demand low-arousal states to prevent micro-tremors in skeletal musculature and maintain wide visual fields, explosive gross-motor sports (such as Olympic weightlifting, rugby scrummaging, and sprint acceleration) frequently demand extreme, supra-physiological arousal to recruit maximal high-threshold motor units. Crucially, even within the identical athletic discipline, Hanin identified elite performers who excelled only when their subjective anxiety was remarkably low, alongside other elite competitors who achieved peak competitive output only when operating under states of extreme, near-rage autonomic fury.

Modern sports science operates by mapping each athlete’s idiosyncratic IZOF through rigorous psychophysiological monitoring—utilizing ambulatory electroencephalography, continuous heart rate variability tracking, and validated state-anxiety inventories. Elite athletic programs train competitors to dynamically navigate into their specific zone of functioning through targeted interventions:

• Down-Regulation Techniques: Slow-paced diaphragmatic breathing (engaging the vagal nerve to lower heart rate), autogenic progressive muscle relaxation, and soothing auditory focus used to drive an over-aroused athlete backward out of the hypertonic choking zone.
• Up-Regulation Techniques: Explosive physical warm-ups, hyperventilatory breathing protocols (such as cyclic hyperventilation to stimulate sympathetic discharge), high-tempo auditory pacing, and aggressive visualization designed to pull a lethargic, hypoaroused athlete up into their optimal operational zone.

10.2 Aviation Cockpits, Military Engagements, and Extreme Decision-Making

High-stakes operational domains—such as commercial and military aviation, nuclear plant operations, and tactical military combat—represent environments where the catastrophic descending arm of the Yerkes-Dodson Law carries mortal consequences. When an unexpected, critical in-flight emergency occurs at cruising altitude—such as an uncontained engine failure, explosive cabin decompression, or multi-sensor instrumentation malfunction—flight crews are subjected to instantaneous, massive autonomic surges triggered by screaming cockpit alarms, flashing Master Caution lights, and the immediate mortal salience of the situation.

Under these conditions, pilots are exceptionally vulnerable to the catastrophic phenomenon of cognitive tunneling (the direct real-world manifestation of Easterbrook’s attentional narrowing). In multiple documented aviation disasters, crews experiencing severe hyperarousal became fixated upon a single, secondary instrument anomaly (such as a burnt-out landing gear indicator bulb) while completely failing to notice central, life-critical cues—such as a decaying airspeed indicator, an excessive pitch attitude, or automated terrain warnings sounding within the cockpit. The extreme sympathetic activation forcefully dismantled their situational awareness, transforming highly trained, experienced aviators into cognitively paralyzed operators incapable of executing multi-layered systems diagnostics.

To defend against this systematic cognitive degradation, modern aviation relies upon the universal institutionalization of Crew Resource Management (CRM) protocols and standardized Quick Reference Handbooks (QRH). CRM is specifically engineered to mitigate arousal-induced cognitive collapse through structural behavioral redundancy: flying duties and diagnostic duties are strictly separated between Pilot Flying (PF) and Pilot Monitoring (PM), cross-verification of all switch throws is legally mandated, and pilots are explicitly trained to force their physical hands away from controls for an initial deliberate, deep-breathing stabilization interval before engaging an emergency checklist. By shifting the emergency response from ad-hoc, panic-driven cognitive problem-solving to highly practiced, algorithmic, standardized checklists, the system systematically lowers the cognitive complexity of the crisis, aligning it safely within the pilot’s stress-tolerance profile.

10.3 Human Factors and Occupational Vigilance Tasks

While the hypertonic descending arm of the Inverted-U captures public attention due to its dramatic disruptions, modern human factors engineering is equally preoccupied with the left-hand, hypotonic arm of the curve: the operational hazards of sub-optimal arousal, monotony, and sustained vigilance decrements. This danger was first systematically quantified during World War II by Norman Mackworth in his pioneering work on the Mackworth Clock Test, designed to explain why British radar and sonar operators progressively missed German U-boat contacts after prolonged periods of monitoring empty screens.

Mackworth demonstrated that when humans are assigned to monitor repetitive, low-event environments requiring continuous attention—such as modern air traffic control radar suites, continuous pharmaceutical manufacturing lines, or remote pipeline automated monitoring—their detection performance suffers a catastrophic decay within 20 to 30 minutes of continuous watch. In these environments, the absence of dynamic, stimulating environmental cues starves the Ascending Reticular Activating System of ascending sensory collaterals. The cerebral cortex drifts into synchronized, slow-wave theta and alpha rhythms. The operator experiences profound subjective boredom, microsleeps, and delayed sensorimotor processing; their performance drops directly down the left-hand hypotonic slope into functional incompetence.

Human factors engineers actively deploy structural, ergonomic interventions to artificially inject arousal back into low-stimulation work environments. These include automated vigilance-checking alarms, forced rotational schedules that cap single-monitoring shifts at 30 minutes, ergonomic sit-stand physical workstations, and the integration of artificial target injections (where the automated system periodically feeds simulated, rare anomalies into the display to force the operator’s nervous system into active orientation and physical verification). The modern industrial struggle is no longer merely preventing employee burnout (hyperarousal); it is equally navigating the dangerous productivity sinkholes of boreout (hypoarousal) within highly automated systems.

11. Critical Evaluation, Methodological Controversies, and Replication Challenges

11.1 Ethical and Methodological Limitations of the 1908 Study

Despite its undisputed canonical status in the history of psychology, the original 1908 experiment by Robert Yerkes and John Dodson falls remarkably short of modern empirical, statistical, and ethical standards of scientific rigor. A cold methodological audit of the original paper reveals vulnerabilities that would render the study unpublishable in any contemporary peer-reviewed journal.

Chief among these flaws was the extreme, unrepresentative sample size. Yerkes and Dodson based their historic conclusions on a total of only forty-four mice distributed unevenly across their complex experimental matrix. When broken down into the specific 3×3 cells (three levels of shock intensity crossed with three levels of visual discrimination difficulty), several critical experimental groups contained as few as two to four animals. Drawing universal, immutable biological laws regarding the complex interactions of sensory drive, cognitive complexity, and learning rates from a handful of individual rodents exhibits profound statistical fragility. A single idiosyncratic or physically sick mouse was sufficient to skew an entire treatment cell’s quantitative learning curve.

Furthermore, the physical apparatus lacked modern objective precision. Electric shocks were calibrated using an archaic sliding induction coil powered by a storage battery, delivering alternating current that varied wildly based on room temperature, humidity, and the individual moisture levels of the animal’s footpads. Illumination of the visual doorways was achieved not through standardized photometric LED arrays, but through natural daylight reflected off adjustable mirrors supplemented by incandescent lamps, meaning that ambient luminance fluctuations throughout the day constantly altered the actual task difficulty of the discriminanda. Finally, the ethical dimension of the research—subjecting structurally defective, vestibularly disabled rodents to repeated, traumatic electric shocks until they displayed frantic panic, motor freezing, and physical convulsions—stands as a stark reminder of the absence of institutional animal welfare oversight in early twentieth-century comparative laboratories.

11.2 The Tautological and Predictive Deficits of the Inverted-U Model

From an epistemological standpoint, the most damning critique directed against the Yerkes-Dodson Law is its persistent vulnerability to tautology and unfalsifiability, a vulnerability extensively critiqued by sports and cognitive psychologists such as Daniel Landers and Alan Neiss. The foundational theoretical defect lies in the chronic absence of any standardized, a priori operational definition of the optimal inflection point ($A_{opt}$).

In practice, the model is frequently utilized as a circular, post-hoc rationalization of experimental outcomes:

• If an increase in environmental stimulation improves an individual’s performance, the researcher claims the subject was operating along the ascending arm of the curve.
• If an increase in environmental stimulation degrades performance, the researcher asserts the subject was driven down the descending arm of the curve.
• If performance remains unchanged, the researcher concludes the subject was resting stably across the broad, flat optimal plateau at the apex.

Because the researcher rarely measures the subject’s internal neurochemical or autonomic state independently of the behavioral output, the model becomes completely non-predictive. One cannot state definitively in advance what precise level of heart rate, skin conductance, or catecholamine concentration constitutes “optimal” for a novel human task or an individual performer. By altering the post-hoc definition of where the apex resides, any empirical dataset—regardless of its shape—can be retrofitted into the Inverted-U paradigm. A scientific model that can explain every conceivable outcome after the fact, but cannot definitively predict an outcome beforehand, lacks true Popperian falsifiability.

11.3 Replication Variations and the Multiplicity of Empirical Curves

When rigorous empirical researchers have attempted to replicate the Yerkes-Dodson phenomena under strictly controlled laboratory conditions, the clean, universal Inverted-U bell curve has frequently failed to materialize. Instead of a single, immutable law, empirical investigations across cognitive psychology and motor learning have yielded an unruly multiplicity of functional relationships.

Depending on the precise experimental parameters, task designs, and physiological indices measured, researchers have documented:

• Pure Monotonic Linear Functions: Where performance continually ascends with increasing arousal without ever displaying a downward inflection, typical of gross-motor tasks, explosive power outputs, or very simple stimulus-response pairings.
• Negative Linear Functions: Where any increase in physiological activation above deep resting baseline immediately degrades performance, common in delicate micro-surgical simulations or ultra-precise motor tasks.
• Step-Function Thresholds: Where performance remains entirely stable across an expansive range of mounting stress until a precise physiological breaking point is reached, whereupon performance collapses instantaneously like a dropped stone, exhibiting no gradual downward curve.
• Asymptotic Curves: Where performance ascends rapidly up to a physiological ceiling and remains locked at that high performance level regardless of massive additional stress increases.

Meta-analytic reviews of the arousal-performance literature have repeatedly concluded that the universal inverted-U is an oversimplification. By compressing the entire multidimensional spectrum of human emotional, cognitive, and autonomic reactions into a single, scalar “arousal” axis, the classical Yerkes-Dodson Law masks the intricate biological reality that different emotional states (such as anger, excitement, panic, and dejection) exert wildly divergent neurochemical impacts upon performance, even when their peripheral autonomic heart-rate profiles are virtually identical.

12. Contemporary Revisions: Catastrophe Models, Multidimensional Arousal, and Future Horizons

12.1 The Cusp Catastrophe Model of Hardy and Fazey

To rectify the severe predictive and conceptual deficits of the smooth, symmetrical Inverted-U, sports psychologists Lew Hardy and John Fazey introduced the Cusp Catastrophe Model of performance in 1987. Grounded in the mathematical principles of René Thom’s catastrophe theory, this model fundamentally rejects the assumption that performance degradation under extreme pressure is always a smooth, gradual, reversible descent.

The Catastrophe Model constructs a three-dimensional topological manifold based on two distinct, interacting independent variables:

• Physiological Arousal: The somatic, autonomic, and peripheral activation of the body (heart rate, respiration, muscle tone).
• Cognitive Anxiety: The cognitive, top-down psychological apprehension (worry, negative performance expectations, threat evaluations).

The dynamic interaction between these two axes creates an asymmetrical, folded topological surface that radically reshapes the performance profile:

• Low Cognitive Anxiety: When the individual experiences very little internal mental worry, changes in physiological arousal yield the traditional, smooth, gentle Inverted-U curve. The performer can move up and down the curve with complete safety and gradual transitions.
• High Cognitive Anxiety: When the individual is simultaneously experiencing intense mental worry and evaluative dread, the surface folds over upon itself, forming a catastrophic cliff (a “cusp”). Under these conditions, escalating physiological arousal initially enhances performance along the upper surface; however, as arousal breaches the critical edge, performance does not experience a gentle, gradual decline. It suffers an instantaneous, catastrophic vertical collapse—plummeting off the upper manifold directly down to the lower failure manifold.

Crucially, the Catastrophe Model accurately captures the clinical reality of the hysteresis effect. When an elite athlete, pilot, or surgeon suffers a catastrophic performance collapse under intense pressure, they cannot simply recover their former elite performance level by lowering their arousal a tiny fraction. Because of the topological fold in the surface, the performer must drop their physiological arousal dramatically—far lower than the point at which the initial crash occurred—in order to “jump” back up to the upper manifold, before slowly rebuilding their performance baseline. The Inverted-U completely fails to predict or explain this irreversible hysteresis dynamic.

12.2 Multidimensional and Valence-Specific Arousal Frameworks

Contemporary psychophysiology has decisively abandoned the antiquated mid-century notion of “unitary arousal.” The modern scientific consensus recognizes that the human nervous system does not possess a solitary, universal activation dial. Instead, activation is inherently multidimensional, valence-dependent, and biologically compartmentalized.

Psychologist Robert Thayer revolutionized this domain by demonstrating that human arousal is split into two functionally and neurochemically distinct biopsychological dimensions:

• Energetic Arousal (EA): Ranging from feelings of energy, vigor, and lively alertness to fatigue, sluggishness, and sleepiness. Energetic arousal is neurochemically supported by moderate dopamine and cellular metabolic vigor, and it correlates almost universally with enhanced cognitive performance, expanded working memory, and positive task engagement.
• Tense Arousal (TA): Ranging from subjective feelings of tension, anxiety, and autonomic dread to quiet calmness and physical placidity. Tense arousal is neurochemically mediated by the severe amygdalar-cortisol-CRH stress cascade, and it correlates directly with cognitive disorganization, perceptual tunnel vision, and behavioral disintegration.

This multidimensional perspective is reinforced by the modern Challenge versus Threat Appraisal framework formulated by Jim Blascovich and Wendy Mendes. When an individual confronts an intense, high-stakes task, their physiological reaction is determined entirely by their cognitive appraisal of their own coping resources:

• Challenge Appraisal (Adaptive): Occurs when the individual evaluates their personal skills and resources as sufficient to meet the environmental demand. This state activates the SAM axis alone: cardiac output surges, peripheral vascular resistance decreases (vasodilation), blood delivery to the brain and skeletal muscles is maximized, and performance is augmented.
• Threat Appraisal (Maladaptive): Occurs when the individual evaluates the environmental demand as overwhelming their personal resources. This state activates both the SAM axis and the HPA axis concurrently: cardiac output increases poorly, peripheral vascular resistance increases (vasoconstriction), systemic cortisol saturates hippocampal receptors, and cognitive performance collapses.

Two human operators exhibiting the exact same heart rate of 140 beats per minute can be in entirely distinct neurobiological universes. One may be in a state of high energetic challenge (maximizing performance at the apex of their capability), while the other is locked in a state of high tense threat (paralyzed at the catastrophic cliff of cognitive failure). The original Yerkes-Dodson Law is structurally blind to this foundational distinction of emotional valence and cognitive appraisal.

12.3 Neuroadaptive Technology, Biomarkers, and Future Horizons

As behavioral science integrates with cutting-edge artificial intelligence, biomedical engineering, and cognitive computing, the foundational questions raised by Yerkes and Dodson in 1908 are being solved through the discipline of neuroergonomics. Rather than treating the Inverted-U as a passive, theoretical model, modern researchers are constructing active, closed-loop neuroadaptive human-machine interfaces designed to dynamically maintain human operators within their optimal performance zones in real time.

These advanced systems utilize arrays of non-invasive, continuous physiological sensors embedded directly into operational environments: high-density dry-sensor electroencephalography (monitoring theta/beta ratios to detect cognitive overload), remote pupillometry and gaze tracking (detecting Easterbrook attentional narrowing via pupil diameter dilation and saccadic fixation dynamics), electrodermal wrist arrays (tracking sympathetic sweat gland discharge), and optical photoplethysmography (measuring vagal heart rate variability). Machine-learning classification algorithms continuously ingest these multi-modal biological telemetry streams, constructing a real-time, dynamic model of the operator’s current coordinates along their multidimensional arousal manifold.

In high-risk settings—such as autonomous combat vehicle operations, air traffic control towers, or precision robotic surgery—these closed-loop architectures can dynamically alter task complexity to prevent human error:

• If the sensor array detects an operator drifting down the left-hand hypotonic slope (hypoarousal, microsleeps, vigilance collapse), the artificial intelligence actively intervenes to up-regulate engagement—introducing auditory alarms, changing display colors, or dynamically re-assigning manual control back to the operator to force mental activation.
• If the system detects the operator rocketing past their optimal apex into the dangerous hypertonic zone (pre-frontal potassium channel opening, catecholaminergic flooding, cognitive tunneling), the AI instantly acts to down-regulate task complexity—stripping non-critical data feeds from the displays, suppressing low-priority auditory warnings, automating secondary flight controls, and presenting only the singular, critical diagnostic decision required to prevent disaster.

From the crude wooden box, copper wires, and Japanese dancing mice of Robert Yerkes and John Dodson’s 1908 laboratory to the algorithmic neuroadaptive cockpits of the twenty-first century, the core empirical quest remains unchanged: deciphering the precise mechanical boundaries where sensory stimulation ceases to be an energizing catalyst for human action and becomes the instrument of its undoing.

Conclusion

The Yerkes-Dodson Law occupies a rare and complex position within the landscape of modern psychology. On one hand, its popular rendering—the ubiquitous, universally applicable Inverted-U curve—often conceals significant historical inaccuracies, conceptual conflations, and predictive limitations. The original 1908 experiment did not investigate generalized human anxiety, nor did it construct a mathematically perfect, universal normal distribution. It was an empirical exploration of how physical punishment affects visual discrimination learning in an atypical rodent model, fundamentally constrained by low sample sizes, archaic apparatus calibration, and questionable statistical generalizability. Furthermore, its post-hoc flexibility and theoretical oversimplification have drawn justifiable criticism from researchers who demand predictive, falsifiable science.

On the other hand, the foundational insight articulated by Yerkes and Dodson—that the relationship between motivational drive and cognitive performance is non-linear, and that the optimal point of activation shifts as an inverse function of task complexity—has survived rigorous theoretical and technological revolutions. What was originally observed as frantic, disorganized locomotion in Mus wagneri has been brilliantly decoded at the level of molecular neurobiology: the differential affinities of prefrontal catecholamine receptors, the biphasic kinetics of glucocorticoid saturation in the hippocampus, and the evolutionary trade-off between reflective neocortical processing and reflexive amygdalar survival circuits. Far from being a historical relic, the underlying mechanics of the law have been refined, multidimensionally segmented, and integrated into contemporary catastrophe models, instructional design theories, and neuroadaptive engineering.

Ultimately, the enduring legacy of the Yerkes-Dodson Law lies in its profound biological humility. It stands as an immutable scientific refutation of both the ascetic belief that human performance flourishes in the total absence of stress, and the industrial hubris that relentless, escalating pressure invariably extracts superior cognitive output. By delineating the finite, delicate neurobiological boundaries within which the human brain can successfully process nuance, maintain working memory, and execute complex reasoning, the law reminds us that optimal performance is not a product of maximal force, but of exquisitely calibrated physiological and cognitive balance.

References

• Arnsten, A. F. T. (2009). Stress signalling pathways that impair prefrontal cortex structure and function. Nature Reviews Neuroscience, 10(6), 410–422. https://doi.org/10.1038/nrn2639
• Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience, 4(10), 829–839. https://doi.org/10.1038/nrn1201
• Blascovich, J., & Mendes, W. B. (2000). Challenge and threat appraisals: The role of affective cues. In J. P. Forgas (Ed.), Feeling and thinking: The role of affect in social cognition (pp. 59–82). Cambridge University Press.
• Easterbrook, J. A. (1959). The effect of emotion on cue utilization and the organization of behavior. Psychological Review, 66(3), 183–201. https://doi.org/10.1037/h0047707
• Hanin, Y. L. (2000). Emotions in sport. Human Kinetics.
• Hardy, L., & Fazey, J. (1987). The Inverted-U hypothesis: A catastrophe for sport psychology? Paper presented at the Annual Conference of the North American Society for the Psychology of Sport and Physical Activity, Vancouver, Canada.
• Hebb, D. O. (1955). Drives and the C.N.S. (Conceptual Nervous System). Psychological Review, 62(4), 243–254. https://doi.org/10.1037/h0041823
• Hull, C. L. (1943). Principles of behavior: An introduction to behavior theory. Appleton-Century-Crofts.
• Landers, D. M. (1980). The arousal-performance relationship revisited. Research Quarterly for Exercise and Sport, 51(1), 77–90. https://doi.org/10.1080/02701367.1980.10609276
• Mackworth, N. H. (1948). The breakdown of vigilance during prolonged visual search. Quarterly Journal of Experimental Psychology, 1(1), 6–21. https://doi.org/10.1080/17470214808416738
• McEwen, B. S. (1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338(3), 171–179. https://doi.org/10.1056/NEJM199801153380307
• Moruzzi, G., & Magoun, H. W. (1949). Brain stem reticular formation and activation of the EEG. Electroencephalography and Clinical Neurophysiology, 1(4), 455–473. https://doi.org/10.1016/0013-4694(49)90219-9
• Neiss, R. (1988). Reconceptualizing arousal: Psychological states in motor performance. Psychological Bulletin, 103(3), 345–366. https://doi.org/10.1037/0033-2909.103.3.345
• Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1207/s15516709cog1202_4
• Thayer, R. E. (1989). The biopsychology of mood and arousal. Oxford University Press.
• Thorndike, E. L. (1898). Animal intelligence: An experimental study of the associative processes in animals. The Psychological Review: Monograph Supplements, 2(4), i–109. https://doi.org/10.1037/h0092987
• Yerkes, R. M. (1907). The dancing mouse: A study in animal behavior. Macmillan.
• Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to rapidity of habit-formation. Journal of Comparative Neurology and Psychology, 18(5), 459–482. https://doi.org/10.1002/cne.920180503