The Conditioned Emotional Response Experiment (Conditioned Suppression) – William Estes and B.F. Skinner

In the annals of twentieth-century experimental psychology, few empirical paradigms have exerted as transformative an influence on the conceptualization of internal affective states as the Conditioned Emotional Response (CER), known contemporaneously across behavioral neuroscience and psychopharmacology as conditioned suppression. Formulated and empirically validated in 1941 by the pioneering behavioral architect B.F. Skinner and his intellectually formidable graduate student William K. Estes at the University of Minnesota, the CER procedure established an objective, quantitative bridge between two historically divergent traditions of learning theory: Pavlovian (respondent) conditioning and Thorndikian-Skinnerian operant conditioning. Before this breakthrough, subjective states such as fear, terror, and anxiety resided predominantly within the qualitative, introspective domains of psychoanalysis or were operationalized merely as crude, uncalibrated autonomic reflexes that resisted precise temporal quantification.

The ingenuity of the Estes-Skinner paradigm lay in its structural architecture: rather than attempting to measure an emotional state through ambiguous visceral proxies or gross motor flailing, the researchers superimposed a classical aversive contingency onto an ongoing, highly stable, appetitively maintained operant baseline. By training a hungry organism to depress a lever for food reinforcers delivered under an intermittent schedule, Estes and Skinner generated a reliable, mathematically steady stream of behavior. When a neutral warning stimulus—such as an auditory tone—was paired with a brief, inescapable, response-independent electric shock and presented during this ongoing operant performance, the warning stimulus acquired the capacity to drastically decelerate or completely arrest lever-pressing behavior. This disruption occurred without altering the physical apparatus, the mechanical availability of the food reward, or the organism’s underlying physiological state of hunger.

This suppression of steady-state operant behavior provided the behavioral sciences with an exquisitely sensitive, non-invasive behavioral assay for quantifying an internal affective state—conventionally labeled “anxiety”—strictly through observable, reproducible changes in response rates over time. The 1941 paper, titled “Some Quantitative Properties of Anxiety,” published in the Journal of Experimental Psychology, not only resolved long-standing methodological impasses regarding the experimental study of aversive motivation, but also paved the way for modern translational neuroscience, computational models of associative learning such as the Rescorla-Wagner model, psychotropic drug screening, and contemporary neurobiological dissections of the fear-conditioning circuitry centered on the amygdala. This comprehensive treatise explores the historical, theoretical, mathematical, neurobiological, and translational dimensions of this landmark scientific contribution.

1. Historical and Theoretical Foundations of the 1941 Estes-Skinner Collaboration

1.1 The Intellectual Climate of Early 20th-Century Behaviorism

The early twentieth century bore witness to a seismic epistemological realignment within psychological science. The radical iconoclasm inaugurated by John B. Watson in his 1913 manifesto, “Psychology as the Behaviorist Views It,” had successfully dismantled the hegemony of structuralist introspectionism and Titchenerian mental elements. However, by the mid-1930s, first-generation classical behaviorism had fractured into several competing, mathematically rigorous neobehaviorist frameworks. Foremost among these was the ambitious hypothetico-deductive drive-reduction system formulated by Clark L. Hull at Yale University. Hull sought to formalize all mammalian learning into an elaborate matrix of mathematical postulates, intervening variables, and mechanistic equations linking habit strength, reaction potential, and generalized drive states to overt muscular movement.

Concurrently, a radically distinct epistemological perspective was coalescing under the stewardship of B.F. Skinner. Rejecting both Hull’s deductive theorizing and his commitment to physiological intervening variables, Skinner advanced a philosophy of science grounded in descriptive, radical behaviorism and inductive functional analysis. Skinner posited that behavior could be scientifically classified into two discrete taxonomic categories: respondent behavior, which is involuntarily elicited by preceding antecedent stimuli (mirroring Ivan Pavlov’s classical salivation paradigm), and operant behavior, which is spontaneously emitted by the organism and governed, strengthened, or extinguished primarily by its subsequent environmental consequences. Skinner’s publication of The Behavior of Organisms in 1938 formally decoupled operant conditioning from traditional reflexology, establishing the rate of an emitted response as the primary dependent variable of a genuinely autonomous behavioral science.

Into this fertile intellectual laboratory at the University of Minnesota stepped William Kaye Estes. Possessing an exceptional analytical mind and an innate inclination toward mathematical formalization, Estes arrived with a keen interest in quantifying behavioral probabilities. Where other contemporaries saw Skinner’s operant chambers as simplistic mechanical curiosities, Estes recognized them as precise, real-time kinetic measuring engines capable of yielding data amenable to deterministic and stochastic mathematical modeling. The intersection of Skinner’s empirical instrumentation with Estes’s budding quantitative sophistication created the ideal crucible for resolving one of the most stubborn theoretical dilemmas confronting behaviorism: the objective status and measurement of human and animal emotion.

1.2 The Problem of Emotion in Pure Operant Psychology

Within the orthodox operational behaviorism of the 1930s, the construct of “emotion” was an intellectual minefield. To traditional psychology, emotions were subjective, conscious feeling states—introspective experiences of dread, rage, sorrow, or euphoria that were inherently inaccessible to direct third-person empirical verification. For Skinner, invoking internal mental states to explain outward behavioral phenomena committed a profound category error, generating vacuous explanatory fictions that halted scientific inquiry. If a rat ran rapidly across an open field away from a predator, attributing its locomotion to “fear” explained nothing; it merely substituted a mentalistic label for the physical phenomenon requiring functional explanation.

Nevertheless, neither Skinner nor the wider neobehaviorist school could deny that organisms undergo profound, observable shifts in their behavioral repertoires when confronted with environmental threats or life-threatening emergencies. The critical challenge was to reconceptualize emotion not as an unobservable, mentalistic event occurring within an ethereal Cartesian theater, nor merely as a chaotic constellation of visceral-autonomic disturbances, but as a dynamic operation—a functional alteration in the broad distribution and probability of an organism’s behavioral repertoires. Skinner sought to demonstrate that an emotional state could be defined rigorously as a temporary change in the strength of an entire class of operant behaviors induced by an environmental operation.

Prior to 1941, methodological attempts to study aversive states in the laboratory suffered from severe experimental confounds. Early researchers typically introduced intense electric shocks directly into the test apparatus while an animal was engaged in a task, resulting in violent motor agitation, erratic jumping, vocalizations, and apparatus damage. These chaotic unconditioned reactions inevitably destroyed the baseline stability of the experimental task, making it impossible to separate the direct mechanical interference of the shock from the anticipatory psychological state preceding it. There existed an acute, unmet operational need for an experimental methodology that could isolate the anticipatory affective state from the physically disruptive impact of the unconditioned aversive stimulus itself, while preserving the mathematical standardization of continuous recording.

1.3 Publication and Context of the 1941 Seminal Paper

The programmatic solution to this methodological bottleneck emerged through intense, systematic laboratory experimentation conducted by Estes and Skinner between 1939 and 1941 in the basement laboratories of the psychology building at the University of Minnesota. The collaboration functioned through a harmonious division of intellectual strengths: Skinner provided the refined operant chambers, the automated continuous cumulative recording technology, and the underlying conceptual framework of operant behavior; Estes contributed rigorous procedural execution, meticulous record-keeping, and the foundational quantitative analyses that transformed raw response counts into formal functional relationships.

Their findings culminated in the historic paper titled “Some Quantitative Properties of Anxiety,” submitted to the Journal of Experimental Psychology and published in the autumn of 1941. The timing of the publication was profound. As the global cataclysm of World War II intensified—with the United States on the precipice of entering the conflict following the events at Pearl Harbor just months later—the academic world was intensely preoccupied with issues of human psychological breakdown, industrial stress, military combat fatigue, and the debilitating impacts of severe anticipatory trauma. While Estes and Skinner’s paper maintained a detached, rigorously objective tone focused exclusively on albino rats pressing levers, the scientific community immediately grasped its profound translational implications.

The reception of the 1941 paper within experimental psychology was marked by immediate recognition of its methodological elegance. Rather than generating controversy through sweeping philosophical broadsides, Estes and Skinner presented pristine, undeniable graphic tracings produced automatically by mechanical cumulative recorders. These tracings demonstrated with mathematical regularity that the steady rate of an appetitively reinforced operant behavior could be selectively, reversibly, and reliably driven to zero by the presentation of an auditory tone that merely signaled an impending, non-contingent shock. The paper bypassed decades of speculative debate regarding the mental architecture of fear, replacing philosophical ambiguity with an objective, reproducible laboratory model that would govern the study of aversive conditioning for the next eight decades.

2. Conceptual Framework: Defining Conditioned Emotional Response (CER) and Conditioned Suppression

2.1 The Operational Definition of the Conditioned Emotional Response

At its core, the Conditioned Emotional Response (CER) constitutes an acquired, anticipatory affective reaction elicited by an initially neutral environmental stimulus (the Conditioned Stimulus, or CS) that has been repeatedly and systematically paired with an inherently noxious, biologically threatening unconditioned stimulus (US), such as an electrical shock. However, what distinguished the Estes-Skinner operational definition from conventional Pavlovian fear conditioning was their radical departure from measuring traditional, discrete skeletal-motor reflexes (such as leg flexion or eyelid closure) or peripheral autonomic readouts (such as pupillary dilation, salivation, or galvanometric skin resistance).

Instead, Estes and Skinner defined the CER functionally, as a significant, measurable perturbation in an independently established, ongoing baseline of voluntary, goal-directed operant behavior. Under this operational architecture, “anxiety” was not treated as a somatic feeling or a neurological state to be inferred; it was defined explicitly as the temporary reduction in the rate of an appetitive operant response during the temporal presence of a warning signal. The emotional state was entirely synonymous with the functional relationship between the warning stimulus and the subsequent behavioral deceleration.

Crucially, this operational paradigm enforced an unyielding structural distinction between direct operant punishment and classical aversive conditioning. In a standard punishment paradigm, the delivery of the aversive stimulus is fundamentally response-dependent (contingent): the animal receives a shock if and only if it emits the designated target response. In stark contrast, the Estes-Skinner CER paradigm employs an entirely response-independent (non-contingent) Pavlovian schedule. The warning stimulus terminates with an electric shock regardless of whether the organism presses the lever, freezes, grooms, or sleeps. The shock is inescapable and unavoidable. This foundational separation ensured that any subsequent deceleration in lever-pressing could not be attributed to an operant avoidance strategy designed to prevent the shock, but represented instead a pure, involuntary emotional interference with the organism’s appetitive motivational system.

2.2 Mechanisms of Conditioned Suppression

The behavioral phenomenon that serves as the quantitative readout of the CER is termed conditioned suppression. Conditioned suppression is an indirect metric: rather than measuring the emergence of a novel motor response, the investigator measures the temporary cessation of an existing, well-learned behavioral repertoire. When the conditioned warning tone sounds, an animal engaged in vigorous, steady lever-pressing for food abruptly ceases pressing the lever. The animal does not abandon the task because food has lost its nutritive value, nor because the mechanical properties of the lever have altered; rather, the ongoing baseline behavior is overridden by an incompatible, prepotent defensive behavioral state.

Ethologists and modern behavioral neuroscientists clarify that this cessation is driven primarily by the activation of an innate, evolutionarily conserved defensive reaction: unconditioned and conditioned freezing (often referred to in the ethological literature as behavioral arrest or tonic immobility). When a wild rodent detects an imminent, unavoidable predatory threat signaled by distal sensory cues, its primary adaptive strategy is profound behavioral immobility—suppressing all voluntary movement and respiration velocity to minimize acoustic and visual detection by the predator. In the confined environment of an operant chamber, the conditioned warning stimulus activates this ancient defensive circuitry, precipitating a state of motor freezing characterized by muscular rigidity, hyper-vigilance, and crouched posture.

From a behavioral-analytic perspective, this freezing response is physically and topographically incompatible with the skeletal-motor actions required to approach, depress, and release an operant lever. Thus, conditioned suppression represents the disruptive intrusion of an evolutionary survival program upon an appetitively reinforced behavioral stream. The investigator observes this competition as response deceleration. Depending on the intensity of the aversive unconditional stimulus, the duration of the conditioned warning signal, and the depth of conditioning, this deceleration can range from a minor, transient hesitation to absolute, protracted behavioral arrest that persists until the warning stimulus ceases.

2.3 Terminology and Semantic Variations Across Disciplines

The terminology surrounding this behavioral phenomenon exhibits fascinating taxonomic variations across different branches of the psychological, ethological, and neuroscientific literature. In classical experimental behavior analysis and radical operant psychology, the phenomenon is traditionally designated as the Conditioned Emotional Response (CER), honoring the precise terminology established by Estes and Skinner in 1941. This term emphasizes the operationalization of an internal emotional construct within a rigorous behaviorist framework.

Conversely, within cognitive psychology, experimental learning theory, and mainstream animal cognition, the descriptive phrase conditioned suppression is universally preferred. Authors such as Leon Kamin, Robert Rescorla, and Allan Wagner adopted “conditioned suppression” because it provides an agnostic, strictly descriptive label that specifies precisely what is observed empirically—the suppression of an operant baseline—without smuggling in speculative mentalistic assumptions regarding the subjective presence of an “emotional” state. In these domains, conditioned suppression is viewed as an exquisitely calibrated behavioral titration procedure for measuring associative strength between environmental events.

In modern translational neuroscience, molecular psychiatry, and neurobiology, the nomenclature has evolved further. Researchers studying the cellular mechanisms of synaptic plasticity in the rodent brain frequently categorize the Estes-Skinner procedure as a specialized variant of auditory fear conditioning or fear-potentiated behavioral arrest. While contemporary neurobiologists frequently bypass the operant baseline entirely—preferring to measure freezing directly via automated infrared beam-breaks or high-speed video tracking algorithms—the term “conditioned suppression” remains the gold standard whenever an investigator needs to evaluate how threat anticipation disrupts executive function, decision-making, reward pursuit, and cost-benefit computation in mammalian models.

3. Apparatus and Experimental Design: The Operant Chamber and Baseline Stabilization

3.1 Architecture of the Estes-Skinner Operant Chamber

The empirical execution of the conditioned emotional response paradigm necessitated sophisticated, highly standardized mechanical instrumentation. The experimental apparatus deployed by Estes and Skinner represented a specialized modification of the standard operant conditioning chamber that Skinner had developed during his doctoral and post-doctoral work at Harvard University. Fabricated from wood, sound-dampening insulation panels, and heavy-gauge wire mesh, the inner dimensions of the chamber were meticulously engineered to provide an isolated, controlled micro-environment that decoupled the experimental subject completely from the uncontrolled sensory fluctuations of the external laboratory.

At the mechanical core of the chamber was a responsive operant manipulandum: a low-inertia metal lever projecting horizontally from one side wall, positioned at a precise height accessible to a standing or crouching albino rat. The lever was mechanically coupled via a fulcrum to an electrical micro-switch mounted on the exterior of the testing compartment. Depressing the lever with a force exceeding a pre-calibrated resistance threshold (typically between 10 to 15 grams) closed the electrical circuit, instantaneously transmitting an electrical impulse to an electromechanical recording bank and reinforcing magazine. Adjacent to the lever sat a concave food cup into which uniform, dehydrated food pellets could be precisely discharged via an automated, solenoid-driven rotary dispenser.

The most critical hardware modification required for the CER protocol was the electrification of the chamber floor. The standard solid or wire-mesh floor was replaced with a series of parallel, cylindrical stainless-steel rods spaced approximately 1.0 to 1.5 centimeters apart. These grid bars were wired into an alternating current (AC) circuit controlled by a variable autotransformer (such as a Variac) and a high-voltage, high-resistance ballast circuit. This electrical infrastructure allowed the experimenter to deliver a uniform, unconditioned electrical shock across the grid floor. Because an ungrounded animal moving across parallel rods could potentially orient its paws onto bars of identical polarity to avoid current flow, advanced iterations incorporated high-speed electrical commutators or “grid scramblers” that rapidly cycled the polarity of individual rods, ensuring that the animal could not adopt a postural strategy to evade the electrical shock.

3.2 Establishing Stable Appetitive Baselines

The absolute prerequisite for conducting a rigorous conditioned suppression experiment is the establishment of an exceptionally stable, predictable, and resilient baseline rate of operant responding. If an animal’s rate of lever-pressing fluctuates wildly due to internal motivational shifts, mechanical unreliability, or volatile reinforcement schedules, it becomes statistically impossible to determine whether a transient reduction in response rate is caused by the presentation of the conditioned warning stimulus or merely represents stochastic baseline drift.

To establish this homeostatic behavioral baseline, Estes and Skinner implemented rigorous appetitive deprivation protocols. Healthy adult male albino rats were placed on a restricted dietary regimen, maintaining their body weights at a standardized 75% to 80% of their free-feeding, age-matched biological control weights. This chronic, sub-maximal caloric restriction established a profound, unvarying appetitive drive state, motivating the animals to work diligently for standardized food pellets throughout experimental sessions spanning several hours.

Crucially, Estes and Skinner did not reinforce the animals on a simple continuous reinforcement (CRF) schedule, where every single lever press yields a food pellet. Continuous reinforcement produces fragile behavioral streams that extinguish rapidly and are characterized by frequent, self-directed post-reinforcement pauses. Instead, the researchers maintained their subjects under an intermittent reinforcement schedule, specifically a Variable-Interval (VI) schedule. Under a variable-interval schedule (for example, a VI 4-minute schedule), a lever press is reinforced with a food pellet only after a variable, unpredictable interval of time has elapsed since the previous reinforcement, with the intervals distributed around a constant mathematical mean.

The profound operational advantage of a variable-interval schedule is that it generates an exceptionally steady, moderate, and uniform rate of responding that persists unabated for hours at a time, entirely devoid of predictable pauses. Because the animal cannot predict precisely when the next reinforcement will become available, it emits a sustained, rhythmic cascade of lever presses at a nearly constant velocity. Skinner and Estes established strict statistical criteria for baseline stabilization: an animal was not subjected to aversive conditioning trials until its session-to-session rate of lever-pressing demonstrated statistical stationarity, exhibiting less than a 5% coefficient of variation across multiple consecutive days of baseline testing.

3.3 Selection and Calibration of Experimental Stimuli

The physical parameters of both the Conditioned Stimulus (CS) and the Unconditioned Stimulus (US) were calibrated with rigorous precision to ensure unequivocal sensory perception while preventing physical tissue damage or absolute behavioral destruction. For the conditioned warning stimulus, Estes and Skinner selected an acoustic stimulus: an auditory tone generated by an electronic audio oscillator or an electromechanical buzzer mounted within the insulated chamber. The auditory CS was calibrated to a specific, salient acoustic profile—typically an 800 Hz or 1000 Hz pure tone, or a distinct, rhythmic clicking sound, delivered at an intensity of approximately 70 to 80 decibels (dB). This intensity was sufficiently elevated above the ambient white noise of the chamber ventilation system (typically 60 dB) to ensure immediate detection, yet quiet enough to avoid precipitating an acoustic startle reflex or unconditioned fear reactions prior to any pairing with shock.

The unconditioned aversive stimulus was an electric shock delivered through the grid floor. The shock parameters were calibrated to achieve a delicate balance: the electrical current had to be intensely aversive, reliably evoking an immediate unconditioned escape-and-flinch response, yet of sufficiently short duration and moderate amperage to avoid causing physical burns, tetanic muscular damage, or prolonged sensory exhaustion. In the 1941 experiments, the shock typically consisted of a 60 Hz alternating current with an intensity ranging from 0.5 to 2.0 milliamperes (mA), delivered for an exact temporal duration of 1.0 to 2.0 seconds.

Finally, the temporal pacing of the experimental session was mathematically structured through the establishment of prolonged Inter-Trial Intervals (ITI). If conditioned warning signals and shocks are delivered in rapid succession, the entire experimental environment becomes saturated with generalized contextual fear. Under such conditions, the animal develops pervasive contextual conditioning to the chamber itself, extinguishing the appetitive baseline entirely and rendering the specific warning stimulus meaningless. Estes and Skinner mitigated this by spacing CS presentations widely across time—utilizing ITIs ranging from 15 to 45 minutes of undisturbed baseline lever-pressing. This temporal architecture ensured that the contextual cues of the chamber remained predominantly associated with appetitive food reinforcement, isolating the conditioned suppression phenomenon strictly to the explicit temporal duration of the auditory warning signal.

4. The Core Experimental Protocol: Classical-Operant Interaction

4.1 Phase-by-Phase Experimental Progression

The empirical execution of the Conditioned Emotional Response experiment unfolded through a sequence of four discrete, meticulously controlled experimental phases, designed to systematically isolate and verify every component of the classical-operant interaction:

• Phase 1: Baseline Operant Stabilization. The animal is introduced to the operant chamber and undergoes lever-press shaping via successive approximations. Once the response is acquired, the subject is transitioned to the intermittent reinforcement schedule (e.g., VI 4-minute) until a resilient, steady-state rate of responding is achieved across daily sessions. No aversive stimuli or warning signals are present during this phase.
• Phase 2: Pre-Conditioning Stimulus Habituation (CS-Alone Testing). Before any pairings with shock occur, the neutral auditory stimulus (CS) is presented repeatedly during ongoing baseline operant performance. This critical control phase verifies that the auditory tone possesses no intrinsic, unconditioned disruptive properties. Typically, upon the initial presentation of a novel 70 dB tone, a rat displays a transient “orienting reflex”—pausing for a few seconds, raising its head, and sniffing the acoustic source. However, within two to three presentations, the animal completely habituates to the benign acoustic cue, continuing to depress the lever at an uninterrupted, constant rate during the tone presentation.
• Phase 3: Pavlovian Aversive Conditioning (CS-US Pairings). The classical aversive contingency is superimposed upon the experimental environment. At predetermined, widely spaced intervals during the session, the auditory tone is activated for a fixed duration—classically four to five minutes in the original 1941 protocol. At the exact millisecond the tone terminates, the grid floor is energized, delivering the 1.0-to-2.0-second electric shock. The shock termination is strictly coterminous with the tone offset (delay conditioning). Critically, the delivery of food reinforcers on the VI schedule continues to operate according to its own independent temporal clock throughout this entire phase; food pellets remain mechanically accessible should the animal choose to press the lever during the tone.
• Phase 4: Assessment of Conditioned Suppression and Associative Dynamics. As repeated pairings of the tone and shock accumulate across successive daily sessions, the experimenter measures the precise rate of lever pressing during the presence of the tone versus the rate observed during equal temporal windows immediately preceding the tone onset. With successive pairings, the tone acquires robust inhibitory control over the operant baseline.

4.2 Non-Contingency of the Unconditioned Stimulus

The structural keystone of the Estes-Skinner paradigm is the absolute non-contingency of the unconditioned aversive stimulus relative to the subject’s operant behavior. In classical conditioning terms, the delivery of the electric shock is entirely independent of the animal’s actions. If an experimental subject presses the lever 100 times during the four-minute tone presentation, it receives the brief shock at the termination of the tone. If the subject presses the lever zero times, curls into a tight ball in the corner, or grooms its vibrissae, it receives the exact same shock at the exact same millisecond.

This non-contingent relationship is theoretically paramount. If the animal could prevent the shock by pressing the lever (as in an active avoidance paradigm) or by refraining from pressing the lever (as in a passive avoidance or punishment paradigm), the deceleration of the baseline could easily be classified as an intentional, rationally acquired instrumental avoidance strategy. The subject would simply be operating under a rule: “Do not press the lever during the tone in order to avoid being shocked.”

Because the shock is physically unavoidable and structurally inescapable, no instrumental manipulation of the lever can alter the subject’s physical fate. Consequently, the complete cessation of lever pressing during the tone yields no evolutionary, utilitarian advantage regarding shock evasion. The suppression of responding represents a pure, unconditioned and conditioned emotional disruption—an involuntary takeover of the animal’s motor apparatus by an amygdala-driven defensive system that forces the organism into tonic immobility in anticipation of an inevitable physical trauma. By severing the contingency between behavior and shock delivery, Estes and Skinner succeeded in isolating pure emotional anticipation from instrumental problem-solving.

4.3 Cumulative Recorder Readouts and Graphical Traces

The empirical data generated by the 1941 experiment were immortalized through the physical output of the cumulative response recorder, an ingenious electromechanical instrument invented by Skinner that served as the foundational polygraph of early behavior analysis. The cumulative recorder functioned by continuously feeding a wide roll of grid-lined paper over a motor-driven drum rotating at an unvarying, constant temporal velocity. A mechanical ink pen rested upon the moving paper. Every single time the experimental rat depressed the lever, an electrical relay energized an escapement mechanism that advanced the pen upward by a minute, discrete step perpendicular to the direction of paper travel.

Under this mechanical arrangement, the slope of the resulting line represented the precise, instantaneous rate of operant responding. A high, rapid rate of responding produced an exceptionally steep, nearly vertical graphical trajectory; a slow, sluggish rate of responding produced a shallow, low-angle slope; and a complete cessation of responding resulted in a perfectly flat, horizontal pen line parallel to the direction of the paper drum. In addition to the vertical stepping mechanism, the pen was equipped with a lateral marker circuit that produced an immediate downward or inward displacement (“blip”) whenever external environmental events occurred, such as the activation of the auditory tone or the delivery of a food pellet.

The cumulative records published by Estes and Skinner in 1941 provided undeniable, visual proof of the conditioned emotional response. During baseline periods prior to tone presentation, the pen tracings climbed at a steady, uninterrupted, uniform 45-degree angle, demonstrating the rock-solid consistency of the VI appetitive schedule. The moment the tone was activated, indicated by a marker blip, the pen trajectory abruptly flattened into an absolute horizontal plateau. For the entire four-minute duration of the tone, the animal emitted zero lever presses; the line remained dead flat. At the termination of the tone, a violent vertical excursion of the pen occasionally appeared, reflecting the mechanical flinch of the shock, followed almost immediately by a post-shock behavioral rebound wherein the pen resumed—and frequently exceeded—its original 45-degree baseline angle. These mechanical traces permitted immediate visual inspection of both the acquisition of the CER over trials and its asymptotic stabilization.

5. Quantitative Measurement: The Mathematical Formalization of the Suppression Ratio

5.1 The Original Estes-Skinner Measurement Metrics

In their historic 1941 publication, Estes and Skinner approached the mathematical quantification of conditioned suppression through an empirical analysis of raw response rates and direct arithmetic difference scores. To capture the extent of behavioral disruption, they systematically partitioned the experimental session into discrete, comparable temporal bins. Specifically, they designated the four-minute period immediately preceding the onset of the auditory warning signal as the control or “pre-CS” period, and the subsequent four-minute period during which the tone was active as the experimental or “CS” period.

The initial metric developed by Estes involved comparing the absolute number of lever presses emitted during the tone against the baseline rate calculated during the preceding control period and during non-shock control days. In their early formulations, Estes calculated percentage reduction scores and tracked the raw mathematical decline in responses across successive days of conditioning. For instance, if an animal consistently emitted 60 responses during the four-minute pre-CS period (a rate of 15 responses per minute) and dropped to 3 responses during the four-minute CS period (a rate of 0.75 responses per minute), the researchers documented a raw behavioral reduction of 95%.

However, Estes, whose mathematical intellect was already gravitating toward formal probability theory, quickly recognized the methodological vulnerabilities of relying solely on raw response counts or uncalibrated difference scores. Animals exhibited substantial individual differences in their raw baseline response rates: a hyper-active rat might depress the lever 200 times per interval, whereas a slower, more deliberate subject might emit only 40 presses under the identical variable-interval schedule. A raw reduction of 30 responses meant something profoundly different for a low-rate subject than for a high-rate subject. To transform conditioned suppression into a universally applicable scientific bioassay, the behavioral science community required an invariant, normalized mathematical index that could scale baseline variance across different individual animals, laboratories, and species.

5.2 The Modern Annau-Kamin Suppression Ratio

The definitive mathematical standardization of conditioned suppression was formulated two decades later by Zaven Annau and Leon J. Kamin in their landmark 1961 paper on the parameters of fear conditioning. Kamin formalized the calculation into a normalized, non-dimensional metric known universally today as the Annau-Kamin Suppression Ratio, traditionally represented by the variable R (or SR):

Suppression Ratio (SR) = B / (A + B)

Where the mathematical terms are defined rigorously as follows:

• B (or CS Responses): The total number of operant responses emitted during the temporal presentation of the Conditioned Stimulus.
• A (or Pre-CS Responses): The total number of operant responses emitted during an equivalent temporal baseline period immediately preceding the onset of the Conditioned Stimulus.

The mathematical properties of the Annau-Kamin suppression ratio are exceptionally elegant, mapping all possible behavioral outcomes onto a continuous, bounded numerical interval ranging from 0.00 to 1.00:

• SR = 0.50 (Zero Suppression / No Conditioning): If the animal exhibits absolutely no fear or conditioned emotional response to the warning stimulus, its rate of responding during the tone remains entirely unchanged from its baseline rate. Under these circumstances, response rate B equals response rate A (e.g., 50 responses during pre-CS, 50 responses during CS). Plugging these values into the formula yields: 50 / (50 + 50) = 50 / 100 = 0.50. A suppression ratio of 0.50 signifies that the warning stimulus possesses zero inhibitory associative control over the operant baseline.
• SR = 0.00 (Complete Suppression / Asymptotic Fear): If the conditioned emotional response is maximal, the animal undergoes total behavioral arrest or continuous freezing throughout the entire presentation of the warning tone. In this state of complete conditioned suppression, B equals 0. The formula yields: 0 / (50 + 0) = 0 / 50 = 0.00. A suppression ratio of 0.00 represents total behavioral inhibition and maximal fear conditioning.
• 0.00 < SR < 0.50 (Graded / Intermediate Suppression): Any intermediate numerical value falling between 0.00 and 0.50 reflects a graded, partial emotional response. For example, a suppression ratio of 0.20 indicates that the animal has substantially decelerated its rate of lever-pressing during the warning signal, but has not frozen completely. This continuous distribution allows investigators to detect subtle, incremental changes in associative strength during acquisition, extinction, or pharmacological drug intervention.
• SR > 0.50 (Conditioned Acceleration / Behavioral Potentiation): While rare in standard aversive protocols, if the presentation of the CS actively elevates responding above the baseline rate (B > A), the ratio yields a value greater than 0.50. This phenomenon, known as conditioned acceleration, occurs under specialized contingencies where the CS signals an appetitive reward or a positive behavioral contrast effect.

A frequent point of conceptual confusion for students encountering the Annau-Kamin metric for the first time is its inverse mathematical relationship to the magnitude of the emotional response: a lower numerical suppression ratio corresponds to a higher magnitude of fear conditioning. Zero represents maximal anxiety, while 0.50 represents total absence of fear.

5.3 Alternative Formulations and Quantitative Critiques

While the Annau-Kamin metric remains the standard across behavioral neuroscience, alternative quantitative indices have been formulated to address specific mathematical and statistical limitations inherent to the ratio. One primary critique raised by biostatisticians involves the behavior of the Annau-Kamin metric under “edge cases”—specifically, when an animal emits zero responses during the pre-CS baseline period (A = 0). If an animal is momentarily inactive during the pre-CS window and subsequently emits zero responses during the CS (B = 0), the formula yields an undefined mathematical expression (0 / 0), forcing the investigator to discard the trial or interpolate data.

To circumvent this and establish a metric with linear intuitive scaling, Howard S. Hoffman and Morton Fleshler proposed the Relative Rate Index or percentage of baseline suppression:

Suppression Index = (A – B) / A

Under this linear metric, a value of 1.00 represents complete suppression (maximal fear), 0.00 represents zero disruption (no fear), and negative values indicate behavioral acceleration. However, this metric suffers from severe mathematical instability whenever the baseline rate (A) is exceptionally low, as small fluctuations in the denominator produce massive, artificial mathematical swings that violate assumptions of homoscedasticity in parametric analyses of variance (ANOVA).

Modern computational behavioral analysts frequently apply logarithmic or square-root transformations to Annau-Kamin suppression ratios prior to conducting parametric statistical hypothesis testing. Because suppression ratios are bounded between 0.00 and 0.50 under standard aversive conditions, their raw sampling distributions exhibit significant positive skewness and truncated variances near the lower bound (asymptotic floor effects). Applying logit transformations or utilizing generalized linear mixed-effects models (GLMM) ensures that the behavioral data meet rigorous standards of statistical normality and continuous mathematical modeling.

6. Detailed Behavioral Dynamics: Extinction, Spontaneous Recovery, and Temporal Patterns

6.1 Extinction Kinetics of the CER

One of the most theoretically profound chapters of Estes and Skinner’s 1941 monograph detailed the systematic extinction kinetics of the conditioned emotional response. Once an animal has acquired an asymptotic CER—exhibiting a suppression ratio approaching 0.00 whenever the auditory warning signal is activated—the underlying associative memory can be systematically extinguished by presenting the Conditioned Stimulus repeatedly over time in the total absence of the Unconditioned Stimulus shock.

To characterize the temporal trajectory of extinction, Estes and Skinner exposed stabilized, conditioned subjects to prolonged sessions where the auditory warning tone was presented continuously or in closely spaced intervals while the animal operated on its standard variable-interval food baseline, without delivering the terminal shock. The extinction curves yielded by the cumulative recorder revealed remarkable behavioral dynamics. Extinction did not occur instantaneously or via a linear, mechanical rebound; rather, it manifested as a gradual, cumulative re-emergence of operant responding during the tone presentation.

During the initial presentations of the non-reinforced tone, the animal remained entirely suppressed, exhibiting profound freezing and depressing the lever zero times. However, across successive minutes of continuous non-reinforcement, minute bursts of operant behavior began to pierce the freezing state. An animal would hesitantly approach the food aperture, emit a single lever press, retreat to the perimeter of the chamber, freeze momentarily, and then emit a short cluster of three or four responses. Gradually, the cumulative slope during the tone tilted upward, progressively aligning with the baseline slope established during pre-CS periods.

Critically, Estes and Skinner demonstrated that the resistance to extinction of the CER is a precise mathematical function of the initial conditioning parameters. Animals that had received high-intensity shocks (e.g., 2.0 mA) or an extensive number of reinforced training trials required hundreds of minutes of non-reinforced CS exposure before the suppression ratio returned to 0.50, whereas animals conditioned with mild shock intensities extinguished rapidly. Furthermore, the velocity of CER extinction was markedly slower than the extinction of standard appetitive operants, demonstrating that aversive pavlovian associations possess extraordinary biological persistence, reflecting their evolutionary significance in preserving organismic survival.

6.2 Spontaneous Recovery and Reinstatement Phenomena

The behavioral re-emergence of the baseline rate during extinction did not, however, signify that the underlying fear association had been permanently erased from the organism’s central nervous system. In their 1941 experiments and subsequent follow-up investigations, Estes observed the classic Pavlovian phenomenon of spontaneous recovery within the conditioned suppression paradigm.

If an animal whose conditioned suppression had been successfully extinguished to a ratio of 0.50 was removed from the experimental apparatus and returned to its home cage for a rest period of 24 to 48 hours, the re-introduction of the animal to the chamber resulted in the immediate, spontaneous resurgence of conditioned suppression. Upon the very first presentation of the auditory tone in the new session, the animal once again ceased lever pressing, yielding a suppression ratio dropping back toward 0.10 or 0.00. Only through repeated, massed extinction sessions across multiple consecutive days did the spontaneously recovered suppression permanently diminish. This established empirically that extinction does not constitute the destruction or unlearning of an associative memory trace, but rather represents the acquisition of a new, inhibitory memory—an active “safety” trace that competes with and suppresses the original excitatory CS-US fear association.

Related phenomena, such as reinstatement and contextual renewal, were subsequently mapped with extraordinary precision using the Estes-Skinner baseline. If an extinguished animal is exposed to a single, unexpected, unsignaled electric shock in the chamber without the tone, the subsequent presentation of the harmless tone hours later immediately reinstates robust conditioned suppression. Similarly, if extinction is conducted in an operant chamber with distinct olfactory and visual cues (Context B) and the animal is returned to the original conditioning chamber (Context A), the animal exhibits immediate ABC/ABA contextual renewal of suppression. The high-resolution stability of the operant baseline made the CER the premier methodological tool for demonstrating that extinguished fear remains latent within neural architecture, poised to re-emerge whenever context, stress, or time alter the organism’s neurochemical equilibrium.

6.3 Temporal Scaffolding: The Role of CS Duration

An exceptionally subtle behavioral dynamic identified in the CER literature involves the temporal scaffolding of the warning signal itself. In their foundational work, Estes and Skinner utilized relatively long CS durations—classically lasting four minutes. When warning signals of such extended duration are deployed, the animal does not necessarily maintain a flat, uniform state of behavioral suppression across every second of the stimulus.

Instead, as conditioning progresses, the animal begins to exhibit a sophisticated temporal discrimination known as the scalloping effect or the inhibition of delay. Because the unconditioned shock is delivered exclusively at the exact termination of the four-minute tone, the initial seconds immediately following tone onset represent a period of objective safety. Highly trained animals gradually learn to exploit this temporal contingency: during the first 30 to 60 seconds of the tone, the rat continues to depress the lever at a moderate, near-baseline rate. As the temporal clock progresses through the second and third minutes, the rate of responding decelerates in a monotonic, downward curve. During the final 30 seconds immediately preceding the shock delivery, the suppression becomes absolute, with the animal freezing entirely.

This scalloped pattern of conditioned suppression demonstrates that the internal emotional state is not an all-or-nothing, static switch, but a dynamic, temporally modulated cognitive-affective process. The internal clock of the organism continuously integrates elapsed time during the warning signal, modulating the descending motor commands to the skeletal musculature in direct proportion to the proximity of the biological threat. Shortening the CS duration to standard modern parameters (e.g., 20 or 30 seconds) compresses this temporal dynamic, yielding near-instantaneous suppression upon stimulus onset, whereas prolonged warning signals unmask the exquisite temporal precision of the mammalian nervous system.

7. Theoretical Implications: Reconciling Pavlovian Conditioning and Skinnerian Operant Behavior

7.1 Two-Process Theory and the Two-Factor Model

The publication of the 1941 Estes-Skinner experiment occurred at a critical theoretical juncture, providing the decisive empirical foundation for what would become the dominant framework in mid-century learning theory: the Two-Process (or Two-Factor) Theory of learning, championed most prominently by O. Hobart Mowrer in 1947. Theoretical psychology had long been locked in an ideological battle between monistic models of learning: Guthrie and Hull argued that all learning was governed by a single mechanism (contiguity or drive-reduction), while Skinner insisted on the structural autonomy of operant and respondent conditioning.

Estes and Skinner’s CER paradigm demonstrated conclusively that classical Pavlovian conditioning and instrumental operant conditioning do not operate within sealed, isolated behavioral silos; rather, they interact dynamically within the same organism in real time. The CER procedure provided the empirical bedrock for Mowrer’s two-factor model by elucidating the precise mechanism through which internal affective states mediate behavioral modifications:

• Factor 1 (Classical Respondent Conditioning): Through Pavlovian pairing with an unconditioned aversive event (US), an initially neutral warning stimulus (CS) acquires conditioned aversive properties, becoming capable of eliciting an involuntary conditioned emotional response (the internal visceral and neural state of “fear” or “anxiety”).
• Factor 2 (Operant Instrumental Interaction): This elicited internal emotional state alters the organism’s motivational priorities, directly competing with and suppressing ongoing appetitive operants, while simultaneously providing the internal drive state necessary to reinforce new instrumental escape or avoidance behaviors.

By showing that a respondent conditioning procedure (CS-US pairing) could systematically control the output rate of an operant response (lever-pressing for food), Estes and Skinner provided the empirical proof that a complete scientific psychology must account for the cross-talk between these two foundational conditioning processes. The CER became the definitive laboratory paradigm for probing the operational interfaces between classical emotional associations and voluntary, goal-directed behavioral streams.

7.2 Competition Between Incompatible Behavioral Topographies

A contentious theoretical debate sparked by the 1941 paper centered on the underlying behavioral mechanism of conditioned suppression: does the conditioned emotional state actively suppress the underlying motivation (the appetitive operant drive) to press the lever, or does it simply elicit an involuntary motor response that is physically and mechanically incompatible with lever pressing?

Skinner himself favored a broad, functional interpretation: emotion was not a physiological muscle twitch, but a systemic change in response predispositions across the organism’s entire behavioral repertoire. When an animal is afraid, the entire class of appetitively reinforced behaviors declines in probability, while the class of defensive behaviors escalates. Conversely, more mechanistically inclined behaviorists, including Estes in his later theoretical writings, emphasized the concept of competing behavioral topographies. From this perspective, the animal ceases to depress the lever simply because it is engaged in an incompatible motor program: tonic immobility (freezing).

To test these competing hypotheses, subsequent investigators constructed ingenious operant chambers where the required operant response was topographically similar—or even identical—to defensive posturing. In famous experiments conducted in the 1960s and 1970s, researchers trained animals to stand completely still or maintain contact with a designated sensor pad to receive food rewards. When an aversive CER warning tone was presented to animals engaged in this stationary operant, the warning tone did not always suppress the behavior; under certain parameters, it actually accelerated or reinforced the stationary posture, because freezing and the target operant shared the identical motor topography. These empirical refinements supported the view that conditioned suppression is mediated by the physical competition between evolutionarily hardwired defensive motor circuits and the somatic motor networks responsible for executing the operant lever press.

7.3 Impact on the Concept of Punishment

The findings of the 1941 CER experiment exerted a profound, lasting impact on Skinner’s conceptualization of punishment, leading to conclusions that would shape his social philosophy and educational writings for decades to come, most visibly in his 1953 magnum opus, Science and Human Behavior. Prior to this work, lay society and traditional psychology assumed that punishment was the symmetrical opposite of reinforcement: while reinforcement strengthened a behavior, punishment was believed to permanently erase or stamp out an undesirable response from an organism’s behavioral repertoire.

Estes and Skinner challenged this assumption directly. Through a related series of experiments published by Estes in his 1944 monograph, “An Experimental Study of Punishment,” the authors demonstrated that when a response is punished (shock delivered upon lever depression), the resulting suppression of the behavior is functionally identical to the conditioned emotional suppression observed in the CER. The physical act of emitting the punished response generates internal and external proprioceptive stimuli that instantly become conditioned stimuli (CSs) paired with shock. The animal does not “unlearn” the response; rather, the incipient execution of the response elicits an immediate conditioned emotional response (anxiety/freezing) that temporarily arrests the completion of the motor act.

Crucially, Estes and Skinner proved that once the aversive contingency is removed, the punished behavior inevitably re-emerges in full strength as the conditioned emotional response gradually undergoes extinction. Punishment, Skinner concluded, does not eliminate behavior; it merely temporarily suppresses its emission through the generation of competing emotional states. This insight led Skinner to become a ferocious public opponent of punitive control in schooling, penal systems, and governance, arguing that punishment is an inefficient, highly destructive behavioral control mechanism that produces pervasive, unintended emotional pathologies without permanently altering underlying operant probabilities.

8. The Neurobiological Substrates of Conditioned Suppression and Fear Conditioning

8.1 The Amygdaloid Circuitry in Conditioned Emotional Suppression

While Estes and Skinner formulated the CER within a strictly descriptive, behavioral-analytic framework that deliberately abstained from neurological speculation, modern translational neuroscience has spent the past four decades mapping the physical brain structures that instantiate their behavioral observations. Today, it is definitively established that the neural engine of conditioned emotional suppression is centered within the complex nuclear architecture of the amygdaloid complex, located deep within the medial temporal lobes.

Neuroanatomical tracing and optogenetic dissections spearheaded by researchers such as Joseph LeDoux and Michael Davis have delineated the precise neurocircuitry through which an auditory CS and a somatosensory US converge to produce conditioned suppression:

• Sensory Convergence in the Lateral Amygdala (LA): The auditory CS travels from the cochlea through the auditory thalamus (medial geniculate nucleus, MGN) and the auditory cortex directly into the lateral nucleus of the amygdala (LA). Simultaneously, the somatosensory footshock US ascends through the spinal spinothalamic tracts and the posterior intralaminar thalamic nuclei to terminate upon the exact same individual pyramidal neurons within the LA.
• Synaptic Plasticity and Long-Term Potentiation (LTP): The coincident arrival of the auditory CS input and the powerful, depolarizing footshock US input induces robust, NMDA-receptor-dependent Long-Term Potentiation (LTP) at the auditory sensory synapses within the lateral amygdala. This synaptic modification permanently increases the synaptic efficacy of the auditory pathway, so that subsequent presentations of the harmless tone alone can depolarize LA neurons and drive downstream defensive networks.
• Output Execution via the Central Nucleus (CeA): The lateral amygdala projects, both directly and via the basal and intercalated (ITC) cell clusters, to the central nucleus of the amygdala (CeA), the primary executive output hub for fear expression.
• Somatic Arrest via the Periaqueductal Gray (PAG): Neurons in the medial division of the central amygdala send heavy, direct, inhibitory GABAergic and peptidergic projections into the midbrain ventrolateral periaqueductal gray (vlPAG). Activation of the vlPAG directly coordinates the cessation of all voluntary movement, driving spinal motor pools into tonic immobility (freezing) and arresting the skeletal-motor commands emerging from the primary motor cortex.

Classical lesion and pharmacological inactivation studies have repeatedly demonstrated that micro-infusions of the GABA-A agonist muscimol directly into the basolateral amygdala or the periaqueductal gray completely obliterate conditioned suppression in the operant chamber. Under amygdalar inactivation, an animal presented with the warning tone exhibits a suppression ratio of exactly 0.50, continuing to depress the lever smoothly for food without hesitation, completely deaf to the impending biological threat.

8.2 Prefrontal-Striatal Interactions in Operant Interruption

Conditioned suppression is uniquely complex because it requires an active, competitive cross-talk between two distinct neuroanatomical systems: the appetitive, goal-directed operant loop that maintains lever-pressing, and the aversive Pavlovian circuit that enforces freezing. This interaction is mediated primarily through dense reciprocal connections between the prefrontal cortex, the ventral striatum, and the amygdala.

Operant lever-pressing for food reinforcers is driven by a cortico-basal ganglia-thalamo-cortical loop comprising the medial prefrontal cortex, the nucleus accumbens (ventral striatum), and the ventral tegmental area (VTA), governed by rhythmic phasic releases of the neurotransmitter dopamine. When the animal is happily pressing the lever on a VI schedule, this mesolimbic dopamine circuit fires tonically, signaling reward expectancy and invigorating motor output.

When the conditioned warning tone sounds, the central amygdala fires intensely, profoundly altering this cortico-striatal balance. Projections from the amygdala directly inhibit dopamine-producing neurons within the VTA and disinhibit medium spiny neurons within the nucleus accumbens shell, causing an immediate collapse of the appetitive reward-seeking signal. Concurrently, distinct subdivisions of the rodent medial prefrontal cortex (mPFC) orchestrate the expression and suppression of the response:

• Prelimbic Cortex (PL): Neurons within the prelimbic cortex project directly to the basolateral amygdala and are critical for the expression of conditioned fear. Optogenetic activation of PL neurons immediately drives freezing and enforces conditioned suppression of lever-pressing.
• Infralimbic Cortex (IL): In stark contrast, the adjacent infralimbic cortex is indispensable for the extinction of conditioned suppression. IL neurons send excitatory projections to the intercalated inhibitory neurons (ITC) of the amygdala, which act as a physiological brake upon the central amygdala. During extinction trials, as the animal learns that the tone no longer signals shock, the infralimbic cortex fires vigorously, suppressing CeA output, silencing the vlPAG freezing commands, and releasing the nucleus accumbens to restore the operant baseline.

8.3 Neuroendocrine and Autonomic Correlates

Conditioned suppression is not an isolated somatic motor phenomenon; it is accompanied by a massive, immediate systemic mobilization of the peripheral autonomic nervous system and the endocrine apparatus, reflecting a total homeostatic pivot toward survival. This neuroendocrine cascade is driven by parallel projections emanating from the central nucleus of the amygdala to the brainstem and hypothalamus.

Direct projections from the central amygdala to the lateral hypothalamus trigger an immediate, massive sympathovagal shift. Post-ganglionic sympathetic noradrenergic neurons fire violently, provoking peripheral vasoconstriction, elevated mean arterial blood pressure, and profound tachycardic or bradycardic cardiac responses depending on the precise posture of the organism. Concurrently, projections to the paraventricular nucleus (PVN) of the hypothalamus activate the Hypothalamic-Pituitary-Adrenal (HPA) axis. Neurons in the PVN release Corticotropin-Releasing Factor (CRF), which stimulates the anterior pituitary to secrete Adrenocorticotropic Hormone (ACTH) into the systemic circulation, driving the adrenal cortex to dump massive quantities of glucocorticoids—principally corticosterone in rodents, and cortisol in primates—into the bloodstream.

Biomedical researchers tracking plasma corticosterone levels throughout the Estes-Skinner paradigm have documented an almost perfect inverse mathematical correlation between circulating corticosterone concentrations and the Annau-Kamin suppression ratio. During pre-CS baseline periods, corticosterone levels hover at low, basal circadian concentrations. Within 90 seconds of warning tone activation, circulating corticosterone spikes dramatically, reaching peak concentrations that directly match the depth of operant suppression. If an animal exhibits a suppression ratio of 0.00, its endocrine system reveals maximal HPA-axis engagement, proving that the cessation of lever-pressing is accompanied by a profound physiological storm of metabolic mobilization.

9. Methodological Variations and Evolution of the CER Paradigm Post-1941

9.1 The Kamin Blocking Effect and Associative Learning Theory

Perhaps the most intellectually monumental theoretical discovery ever achieved using the Estes-Skinner conditioned suppression methodology was the demonstration of the blocking effect by Leon J. Kamin at McMaster University in 1969. Until Kamin’s work, modern psychology assumed that classical conditioning was governed strictly by Pavlovian temporal contiguity: if a conditioned stimulus was paired repeatedly with an unconditioned stimulus in close temporal proximity, an associative link would automatically and inevitably be forged.

Kamin utilized the Annau-Kamin suppression ratio within the CER baseline to shatter this foundational assumption. In a legendary series of experiments, Kamin trained three groups of rats under an identical CER operant baseline:

• Phase 1: Experimental animals (the Blocking Group) received standard CER training where a single stimulus—a light (Stimulus A)—was repeatedly paired with an inescapable electric shock until asymptotic conditioned suppression was achieved (Suppression Ratio = 0.00). The control group received no training.
• Phase 2: In the second phase, both the blocking group and the control group were presented with a compound stimulus consisting of the light and a novel tone simultaneously (Stimulus A + Stimulus B), which terminated in the identical electric shock. Temporal contiguity between the novel tone (B) and the shock was 100% absolute and identical for both groups.
• Phase 3 (Testing): Kamin then presented the novel tone (Stimulus B) alone to assess its ability to suppress the operant lever-pressing baseline.

The results fundamentally revolutionized cognitive and behavioral science. The control group exhibited robust conditioned suppression to the tone (SR = 0.05). However, the blocking group exhibited absolutely no conditioned suppression to the tone, continuing to press the lever uninterrupted (SR = 0.48). Even though the tone had been paired perfectly with the shock, learning about the tone had been entirely blocked by the prior conditioning of the light.

Kamin concluded that conditioning does not occur merely because two events occur close together in time; learning occurs only if the unconditioned stimulus is surprising. Because the shock in Phase 2 was already completely predicted by the presence of the light, it generated no “prediction error,” no cognitive surprise, and therefore no new associative strength could be assigned to the redundant tone. This profound empirical discovery directly inspired Robert Rescorla and Allan Wagner to formulate the legendary Rescorla-Wagner Model (1972) of associative learning, which parameterized Kamin’s prediction error into a formal mathematical algorithm that serves as the direct conceptual ancestor of modern reinforcement learning and backpropagation algorithms in modern artificial intelligence.

9.2 Sensory Preconditioning and Higher-Order CER

The exquisite sensitivity of the conditioned suppression baseline also enabled behavioral scientists to uncover highly intricate, latent cognitive architectures that remain completely hidden in standard motor reflex tests. Foremost among these are higher-order conditioning and sensory preconditioning.

In a sensory preconditioning protocol conducted via the CER, two entirely neutral stimuli—for example, a visual flashing light (Stimulus 1) and a high-frequency tone (Stimulus 2)—are repeatedly paired together in the absence of any electric shock, while the animal presses a lever for food. Because neither stimulus carries biological consequence, the animal continues to press the lever normally, displaying no conditioned suppression. Next, Stimulus 2 alone is paired with an electric shock until it produces a profound CER (SR = 0.00). Finally, the investigator presents Stimulus 1 alone to the animal.

Even though Stimulus 1 was never once paired with an electric shock in the history of the organism’s life, presenting Stimulus 1 produces immediate, powerful conditioned suppression of the operant baseline. The animal demonstrates that it had silently acquired an internal cognitive map linking Stimulus 1 to Stimulus 2 during the initial preconditioning phase. When Stimulus 2 became threatening, the affective value transferred backward across the latent associative link to Stimulus 1. The Estes-Skinner baseline provided the high-resolution quantitative canvas required to detect these subtle, multi-tiered cognitive architectures, permanently dismantling the simplistic view of behaviorism as a crude stimulus-response telephone switchboard.

9.3 Species Variations: Beyond the Laboratory Rat

Although the 1941 experiments were executed exclusively using Rattus norvegicus, the conditioned suppression paradigm was rapidly recognized as a universal mammalian and avian behavioral assay, displaying remarkable cross-species validity. Over subsequent decades, experimental psychobiologists adapted the Estes-Skinner protocol across a wide taxonomic spectrum.

In avian laboratories, researchers modified the apparatus for pigeons (Columba livia). Rather than depressing a horizontal metal lever, pigeons maintained an appetitively reinforced baseline of visual key-pecking, reinforced with grain delivered via an automated hopper on a variable-interval schedule. When an auditory or visual warning signal terminating in a mild electrical shock (delivered through fine wires attached to the pubic bones) was presented, the pigeons displayed profound conditioned suppression of key-pecking, demonstrating suppression ratios identical in mathematical properties to those observed in rodents.

Primatologists extended the CER to non-human primates, including rhesus macaques (Macaca mulatta) and squirrel monkeys (Saimiri sciureus), testing them in specialized primate chairs where steady manual pull-lever baselines were maintained. In human psychological laboratories, researchers established ethical analogues of the CER. Human participants operated a manual joystick or keyboard task to earn monetary rewards or points on a computer screen under an intermittent schedule. A distinctive visual cue signaled an impending, unpleasant aversive event—such as a painfully loud blast of acoustic white noise (95–100 dB) or a mild, localized cutaneous electric shock to the wrist. Just as in the 1941 rat experiments, human subjects exhibited significant, involuntary reaction-time decelerations and complete operant suppression during the warning signal, accompanied by synchronized spikes in skin-conductance responses (SCR), validating the CER as a phylogenetically conserved defensive phenotype across the animal kingdom.

10. Pharmacological and Clinical Applications of Conditioned Suppression

10.1 Anxiolytic and Psychotropic Drug Screening

During the mid-twentieth century, the pharmaceutical industry underwent a dramatic revolution characterized by the synthesis of the first major classes of synthetic psychoactive agents, including meprobamate, chlorpromazine, and the benzodiazepines (such as chlordiazepoxide and diazepam). A major scientific bottleneck facing early psychopharmacology was the total absence of objective, standardized animal models capable of differentiating genuine anti-anxiety (anxiolytic) drugs from simple muscle relaxants, tranquilizers, or motor sedatives.

The Estes-Skinner conditioned suppression paradigm emerged as the premier behavioral bioassay that cracked this pharmaceutical dilemma. When an experimental animal is administered a non-specific sedative (such as a high-dose barbiturate or ethanol), its overall motor ability is impaired: its baseline lever-pressing during the pre-CS interval collapses, and its performance during the CS remains depressed. The drug merely incapacitates the organism physically.

In stark contrast, when an animal is administered a true clinical anxiolytic—such as chlordiazepoxide (Librium) or diazepam (Valium)—the drug produces a spectacular, highly selective behavioral phenomenon known as the release from suppression. Under the influence of a benzodiazepine, the animal’s pre-CS baseline lever-pressing rate remains entirely normal or displays minimal alteration. However, the moment the conditioned warning tone sounds, the animal fails to freeze: it continues to press the lever vigorously, driving the suppression ratio from a pathological 0.00 back toward the normal 0.50 baseline, despite fully recognizing the impending shock. The drug selectively uncouples the Pavlovian emotional fear circuitry from the descending motor suppression systems, providing medicinal chemists with an unambiguous behavioral signature for testing novel anxiolytic compounds.

Subsequent psychopharmacological investigations expanded this methodology to screen Selective Serotonin Reuptake Inhibitors (SSRIs) and specific GABA-A receptor subtype modulators. By establishing dose-response curves within the CER framework, neuropharmacologists mapped how alterations in monoaminergic and GABAergic neurotransmission selectively alleviate conditioned emotional pathology without disturbing general motivational or motor competence.

10.2 Translational Paradigms for Anxiety Disorders and PTSD

In contemporary translational psychiatry, the conditioned suppression paradigm serves as one of the most conceptually robust animal models for dissecting the etiology and treatment of severe human stress pathologies, particularly Generalized Anxiety Disorder (GAD) and Post-Traumatic Stress Disorder (PTSD). Human trauma pathologies are characterized clinically by intrusive, paralyzing emotional responses to trauma-related cues that completely disrupt everyday occupational, executive, and goal-directed pursuits.

In the laboratory, the persistence of conditioned suppression provides an exact operational analog of this human clinical failure. In a typical rodent model of PTSD, animals are exposed to an acute, severe stressor (such as underwater trauma, prolonged immobilization, or unpredictable shock) prior to standard CER training. These traumatized animals exhibit hyper-suppression: their suppression ratios plummet to 0.00 even at negligible shock intensities, and they exhibit profound, pathological resistance to extinction. When normal animals have fully extinguished their fear response across 20 non-reinforced CS exposures, the PTSD-modeled animals continue to display complete behavioral arrest, paralyzed by fear cues that are no longer dangerous.

Psychiatrists utilize this CER extinction framework to evaluate the therapeutic efficacy of novel behavioral and pharmacological interventions designed to accelerate extinction learning. Exposure therapy—the primary cognitive-behavioral treatment for PTSD—is structurally identical to the extinction phase of the Estes-Skinner experiment: the patient is repeatedly exposed to the conditioned fear stimulus in a completely safe, controlled environment until the conditioned emotional response diminishes. Researchers use the CER to identify pharmacological cognitive enhancers, such as the partial NMDA-receptor agonist D-cycloserine, which, when administered prior to extinction sessions, significantly accelerates the return of normal operant responding, translating directly into enhanced protocols for clinical exposure therapy in human combat veterans and trauma survivors.

10.3 Neurobiology of Addiction and Aversive Relapse

The intersection of conditioned emotional suppression with substance use disorders represents one of the most vibrant frontiers in modern neurobiology. Drug addiction is characterized by powerful, compulsive operant baselines: an addicted individual spends an immense proportion of their waking life engaged in goal-directed drug-seeking and drug-taking behaviors maintained on rich, intermittent schedules of reinforcement.

By substituting an intravenous drug-delivery catheter (administering cocaine, morphine, or fentanyl) for the standard food pellet magazine, modern addiction researchers can establish a steady, resilient baseline of operant drug self-administration in laboratory rodents. Once an animal is actively depressing a lever for intravenous infusions of cocaine, investigators superimpose the Estes-Skinner CER protocol onto the drug baseline to evaluate the competing motivational valences of chemical reward pursuit versus biological threat avoidance.

These experiments reveal profound, alarming dynamics: in animals exhibiting high, compulsive drug-seeking phenotypes (modeling severe human substance dependence), conditioned aversive warning stimuli fail to suppress drug-seeking behavior. While a normal animal immediately abandons a food lever when a threat tone sounds, an addicted animal will continue to press the drug-delivering lever relentlessly, completely ignoring the imminent electric shock. The compulsive pursuit of the drug reinforcer fundamentally overrides the amygdaloid defensive freezing program. Furthermore, the subsequent unsignaled delivery of stress-inducing shocks or aversive cues reliably provokes stress-induced reinstatement of drug-seeking, demonstrating that the activation of the HPA-axis and amygdalar stress pathways serves as a potent biochemical trigger driving relapse into drug-seeking behavior.

11. Comparative Analysis: CER Versus Other Paradigms of Aversive Learning

11.1 Conditioned Emotional Response vs. Active and Passive Avoidance

To fully grasp the unique theoretical and methodological identity of the Estes-Skinner CER paradigm, it must be systematically contrasted against the other classic laboratory models of aversive conditioning: Active Avoidance and Passive (Inhibitory) Avoidance.

In an Active Avoidance paradigm—traditionally executed in a two-way shuttle box divided into two compartments by a low hurdle—an animal is placed in the apparatus and presented with a warning signal (tone or light). If the animal physically runs across the hurdle into the opposite compartment within a specified temporal window (e.g., 10 seconds), the warning signal terminates immediately, and the scheduled electric shock is completely avoided. If the animal fails to run, the shock is energized through the grid floor, and the animal must scramble across the hurdle to escape the shock. Here, the behavioral demand is fundamentally active and instrumental: the organism must emit a high-energy motor action to achieve safety. In the CER, the animal has no instrumental control whatsoever; the shock is inescapable, and safety cannot be earned through motor exertion.

In a Passive (Inhibitory) Avoidance paradigm (frequently referred to as the step-down or step-through task), an animal is placed on an elevated, brightly illuminated platform facing a dark, enclosed chamber. Rodents possess a powerful, innate unconditioned preference for dark, enclosed spaces over brightly illuminated open areas. The moment the animal steps down off the platform into the dark compartment, it receives an inescapable electric shock to the paws. During subsequent testing sessions, the animal is returned to the illuminated platform, and the investigator measures the latency of the animal to step into the dark. If the animal has learned the aversive association, it passively suppresses its innate preference, remaining frozen upon the small platform for hundreds of seconds.

The critical structural difference between these paradigms is summarized below:

• Instrumental Contingency: In both active and passive avoidance, the shock is strictly response-dependent. The animal’s behavior directly determines whether the shock is delivered or avoided. In the CER, the shock is strictly response-independent (non-contingent).
• Baseline Requirement: Active and passive avoidance tasks do not require an ongoing operant baseline; they measure discrete trial latencies or shuttle crossings. The CER cannot exist without a continuous, independently maintained, appetitive operant baseline acting as a dynamic measuring scale.
• Motor Requirements: Active avoidance requires rapid, coordinated skeletal locomotion; passive avoidance requires behavioral withholding driven by explicit spatial-contextual cues; the CER captures an involuntary cross-system interference where Pavlovian freezing disrupts an unrelated operant repertoire.

11.2 Conditioned Emotional Response vs. Conditioned Taste Aversion (CTA)

Another monumental paradigm in the taxonomy of aversive conditioning is Conditioned Taste Aversion (CTA), famously documented by John Garcia in the 1950s and 1960s (frequently termed the “Garcia Effect”). In a CTA experiment, an animal is allowed to consume a novel, sweet-tasting solution (such as saccharin-flavored water, the CS). Hours later, the animal is injected with a chemical agent that induces profound visceral gastrointestinal distress and nausea, such as lithium chloride (LiCl, the US). When presented with the saccharin solution days later, the animal completely refuses to drink it, displaying intense rejection responses.

The comparative distinctions between the Estes-Skinner CER and Conditioned Taste Aversion illuminate the biological boundaries and evolutionary specializations of learning mechanisms:

• Temporal Dynamics and Delay Intervals: In the CER, the optimal inter-stimulus interval (ISI) between the onset/offset of the warning tone and the delivery of the electric shock is extraordinarily brief, measured in seconds or minutes. If an investigator delays the delivery of an electric shock by four hours following an auditory tone, absolute zero conditioning occurs. In sharp contrast, Conditioned Taste Aversion operates across massive temporal delays spanning four, eight, or even twelve hours between the ingestion of the flavored water and the onset of chemical nausea. CTA is biologically hardwired to handle the delayed metabolic kinetics of biological poisoning.
• Sensory and Cue-To-Consequence Specificity: The CER relies predominantly upon exteroceptive sensory modalities—acoustic tones, visual lights, and cutaneous footshocks that simulate external predatory attacks. CTA relies exclusively on interoceptive gustatory and visceral modalities. In their famous “bright-noisy water” experiments, Garcia and Koelling proved that an animal cannot easily associate an auditory-visual cue with internal nausea, nor can it easily associate a sweet taste with a cutaneous footshock. The CER taps into the external defensive survival module, while CTA taps into the internal nutritional/poison-avoidance survival module.
• Underlying Neurocircuitry: While the CER is heavily dependent upon the basolateral amygdala, the central amygdala, and the midbrain periaqueductal gray, Conditioned Taste Aversion bypasses the vlPAG entirely, relying on the nucleus of the solitary tract (NTS), the parabrachial nucleus (PBN) of the pons, the gustatory insular cortex, and the bed nucleus of the stria terminalis (BNST).

11.3 Conditioned Emotional Response vs. Direct Operant Punishment

The formal differentiation between the Conditioned Emotional Response and Direct Operant Punishment represents one of the most mathematically and philosophically rigorous boundaries in behavioral analysis. Because both procedures involve electric shocks, levers, and rats, lay observers and novice students frequently conflate them as being functionally interchangeable.

In a direct operant punishment paradigm, the delivery of the aversive stimulus is mechanically and explicitly response-dependent (contingent). The organism is pressing a lever for food reinforcers. Suddenly, the experimenter introduces a punishment contingency: whenever the animal depresses the lever, it receives a brief electric shock directly through the lever or through the floor. The contingency can be summarized logically: Response → Aversive Stimulus. In this scenario, the lever itself becomes a “punished manipulandum,” and any environmental cue signaling that the punishment contingency is active is designated as a discriminative stimulus for punishment (S-delta or S-pun).

In sharp contrast, the Conditioned Emotional Response involves no response-dependent punishment whatsoever. The shock is delivered via a strictly classical, response-independent Pavlovian contingency: Stimulus 1 (CS) → Stimulus 2 (US). The logical rule is: Warning Tone → Shock, regardless of behavior. The animal is never punished for pressing the lever; the animal is simply exposed to an unavoidable, inescapable environmental catastrophe that coincides with the termination of an auditory cue.

The behavioral consequences of these two methodologies diverge significantly over time. Under direct operant punishment, the animal actively alters its motor topography to interact with the environment differently—it may approach the lever hesitantly, tap it with a single claw to minimize current flow, or completely abandon the manipulandum to pursue alternative appetitive activities. Under conditioned emotional suppression, the animal does not modify its lever-pressing topography; it undergoes a total biological systemic shutdown, dropping into profound motor freezing that extinguishes all voluntary behavior indiscriminately until the warning cue passes.

12. Enduring Legacy and Epistemological Impact of Estes and Skinner (1941)

12.1 Foundational Contribution to Mathematical and Quantitative Psychology

The 1941 collaboration with Skinner served as the direct intellectual launching pad for William K. Estes’s subsequent ascent as one of the preeminent mathematical psychologists of the twentieth century. Witnessing the extraordinary quantitative precision with which an operant baseline could record behavioral probabilities, Estes realized that psychology could transcend descriptive, qualitative narratives and achieve the mathematical rigor of theoretical physics.

Less than a decade after the CER publication, Estes published his historic 1950 paper, “Toward a Statistical Theory of Learning,” in the Psychological Review, formally inaugurating Stimulus Sampling Theory (SST). In this theoretical masterpiece, Estes conceptualized the experimental environment not as a monolithic stimulus, but as a vast, finite population of discrete sensory and contextual “elements.” On any given conditioning trial, the organism randomly samples a subset of these elements. Estes translated the raw behavioral shifts observed in the CER operant chamber into formal difference equations and stochastic transition matrices that mapped the exact probability of an element becoming associated with a defensive versus an appetitive state.

The mathematical formalization begun in the 1941 CER paper reverberated across generations of quantitative psychologists. It directly influenced the mathematical formulation of the Bush-Mosteller stochastic learning models of the 1950s, paved the way for the Rescorla-Wagner model’s algebraic error-correction equations in the 1970s, and provided the continuous-time empirical datasets that anchor contemporary dynamic computational models of temporal difference learning (TD learning) deployed in advanced artificial intelligence systems.

12.2 Philosophical Reconciliation of Mentalistic Terms

From an epistemological standpoint, the Estes-Skinner CER experiment achieved something thought impossible within the rigid strictures of early radical behaviorism: it scientifically legitimized the empirical investigation of internal, affective, subjective states—specifically “anxiety” and “fear”—without succumbing to mentalistic, unobservable Cartesian metaphysics.

Before 1941, an experimentalist who spoke of studying an animal’s “anxiety” was viewed by strict behaviorists as harboring unscientific, introspective leanings. Skinner and Estes permanently altered this paradigm by demonstrating that an internal emotional state can be operationalized with uncompromising behavioral rigor. They established that one does not need to crack open the soul of an organism to measure anxiety; anxiety can be defined unequivocally as the functional deceleration of a steady-state operant baseline in response to a warning signal. The internal state was translated into a mathematically reproducible, functional environmental relationship.

This operational paradigm provided the behavioral sciences with a rigorous methodology for studying complex affective and cognitive phenomena. It demonstrated that behaviorism, far from being a simplistic reflexology that denied the existence of emotional depth, possessed the methodological and conceptual tools to systematically investigate the most complex motivational and affective interactions within the mammalian repertoire, setting a benchmark for experimental control that remains an exemplar of the scientific method.

12.3 Current Horizons and Contemporary Re-evaluations

More than eight decades after its initial publication, the Estes-Skinner paradigm remains vibrant, undergoing continuous technological and conceptual rejuvenation at the cutting edge of twenty-first-century neuroscience. In contemporary laboratories, the crude electromechanical micro-switches and ink-pen cumulative recorders of 1941 have been superseded by ultra-high-resolution machine-learning computer vision systems (such as DeepLabCut), which track every fine-grained postural micro-movement, vibrissal deflection, and respiratory expansion of an animal undergoing conditioned suppression with sub-millimeter precision.

Furthermore, modern neuroscientists combine the Estes-Skinner baseline with breathtaking neurobiological tools that could scarcely have been imagined in 1941. Utilizing in vivo two-photon calcium imaging, researchers now record the simultaneous, real-time firing of thousands of individual neurons within the basolateral amygdala, the prefrontal cortex, and the nucleus accumbens while an animal presses a lever for sucrose and experiences a warning tone. Concurrently, using cell-type-specific optogenetics and chemogenetics (DREADDs), neuroscientists can selectively flash laser light down fiber-optic implants into the infralimbic or prelimbic cortex, instantaneously erasing conditioned suppression or reinstating it at will with single-spike temporal precision.

In computational neuropsychiatry, artificial agents governed by deep reinforcement learning algorithms are trained on simulated CER baselines, allowing researchers to mathematically simulate how human anxiety disorders, maladaptive avoidance, and executive functional breakdown emerge from subtle mathematical disruptions in prediction-error computation and value-function discounting. From its modest origins in a Minneapolis basement laboratory during the dawn of the Second World War, the Estes-Skinner conditioned emotional response experiment stands as a monumental intellectual achievement—a timeless, elegant paradigm that fundamentally transformed humanity’s scientific comprehension of the profound, tragic, and adaptive architecture of the emotional brain.

Conclusion

The Conditioned Emotional Response experiment executed by William K. Estes and B.F. Skinner in 1941 represents an unassailable milestone in the history of behavioral science. By superimposing an inescapable, response-independent classical aversive contingency onto an ongoing, stable, appetitively maintained operant baseline, Estes and Skinner dissolved the artificial barrier separating Pavlovian reflexology from Skinnerian operant analysis. In doing so, they provided the scientific community with an objective, continuous, and exquisite quantitative assay—conditioned suppression—capable of measuring an internal emotional state through pure functional alterations in response kinetics.

The legacy of this breakthrough is woven into the very fabric of contemporary neuroscience and clinical psychology. It provided the empirical foundation for Two-Factor learning theory, established the experimental baseline that enabled Leon Kamin to discover the blocking effect and inspire the Rescorla-Wagner model, served as the indispensable pharmacological bioassay that unlocked the development of modern anxiolytic medications, and laid the anatomical groundwork for mapping the mammalian fear circuitry centered upon the amygdala and prefrontal cortex. As modern neuroscience pushes into the frontiers of optogenetics, computational neuropsychiatry, and deep artificial intelligence, the elegant methodology formulated by Estes and Skinner in 1941 continues to serve as an enduring, foundational compass for unraveling the intricate mechanics of fear, anxiety, and behavioral choice.

References

• Annau, Z., & Kamin, L. J. (1961). The conditioned emotional response as a function of intensity of the US. Journal of Comparative and Physiological Psychology, 54(4), 428–432. https://doi.org/10.1037/h0042199
• Davis, M. (1992). The role of the amygdala in fear and anxiety. Annual Review of Neuroscience, 15(1), 353–375. https://doi.org/10.1146/annurev.ne.15.030192.002033
• Estes, W. K. (1944). An experimental study of punishment. Psychological Monographs, 57(3), i–40. https://doi.org/10.1037/h0093551
• Estes, W. K. (1950). Toward a statistical theory of learning. Psychological Review, 57(2), 94–107. https://doi.org/10.1037/h0058559
• Estes, W. K., & Skinner, B. F. (1941). Some quantitative properties of anxiety. Journal of Experimental Psychology, 29(5), 390–400. https://doi.org/10.1037/h0062283
• Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4(3), 123–124. https://doi.org/10.3758/BF03342209
• Hoffman, H. S., & Fleshler, M. (1961). Stimulus factors in conditioned suppression. Journal of the Experimental Analysis of Behavior, 4(4), 307–313. https://doi.org/10.1901/jeab.1961.4-307
• Hull, C. L. (1943). Principles of Behavior: An Introduction to Behavior Theory. Appleton-Century-Crofts.
• Kamin, L. J. (1969). Predictability, surprise, attention, and conditioning. In B. A. Campbell & R. M. Church (Eds.), Punishment and Aversive Behavior (pp. 279–296). Appleton-Century-Crofts.
• LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23(1), 155–184. https://doi.org/10.1146/annurev.neuro.23.1.155
• Mowrer, O. H. (1947). On the dual nature of learning—a re-interpretation of “conditioning” and “problem-solving”. Harvard Educational Review, 17(2), 102–148.
• Pavlov, I. P. (1927). Conditioned Reflexes: An Investigation of the Physiological Activity of the Cerebral Cortex (G. V. Anrep, Trans.). Oxford University Press.
• Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory (pp. 64–99). Appleton-Century-Crofts.
• Skinner, B. F. (1938). The Behavior of Organisms: An Experimental Analysis. Appleton-Century-Crofts.
• Skinner, B. F. (1953). Science and Human Behavior. Macmillan.
• Watson, J. B. (1913). Psychology as the behaviorist views it. Psychological Review, 20(2), 158–177. https://doi.org/10.1037/h0074420