The Trace Conditioning Experiments – Ivan Pavlov

The history of behavioral neuroscience is anchored by foundational empirical discoveries that transformed philosophical conjectures regarding the nature of the mind into rigorously quantifiable physiological phenomena. Among the most pivotal of these transitions occurred at the turn of the twentieth century in the laboratories of the Imperial Institute of Experimental Medicine in Saint Petersburg. There, Ivan Petrovich Pavlov redirected his scientific inquiries from the gastrointestinal physiology that garnered him the Nobel Prize in Physiology or Medicine in 1904 toward the enigmatic phenomenon of associative learning. While standard classical conditioning established that organisms could associate concurrent sensory inputs with biological imperatives, it was the subtle, temporally decoupled variant known as trace conditioning that unveiled the profound computational capabilities of the mammalian central nervous system. In trace conditioning, an organism forms an associative bridge across a chasm of physical silence, linking a transient, vanished sensory cue to an unconditioned reinforcer that arrives seconds or minutes later.

This temporal disconnection posed an epistemological and neurobiological challenge to classical associationism, which had long posited absolute temporal contiguity as the inviolable prerequisite for associative learning. By demonstrating that animals could reliably anticipate a reinforcer long after the physical termination of the conditioned stimulus, Pavlov compelled the scientific community to confront the necessity of an endogenous neural representation—a lasting functional “trace” or vestige inscribed upon the cerebral cortex. This paradigm transformed the experimental animal from a passive, direct-reflex automaton into an active predictive processor capable of sustaining internal representations of the sensory environment across silent intervals. Trace conditioning laid the empirical foundation for modern inquiries into working memory, neural persistent activity, interval timing, and the complex corticolimbic interactions that govern higher-order cognition.

To comprehend the full scope of Pavlov’s trace conditioning experiments is to trace the trajectory of modern neuroscience itself. It requires exploring the architectural marvel of Pavlov’s “Tower of Silence,” dissecting the intricate balance between excitation and internal inhibition in the cerebral hemispheres, and examining modern discoveries linking these nineteenth- and twentieth-century protocols to hippocampal theta oscillations, medial prefrontal persistent firing, and eligibility traces in artificial intelligence. This treatise provides a rigorous, exhaustive analysis of Pavlov’s trace conditioning paradigm, encompassing its historical evolution, physiological mechanics, neurobiological substrates, and enduring contemporary legacy across cognitive science and psychopathology.

1. Historical and Theoretical Foundations of Pavlovian Conditioning

1.1 From Digestion Physiology to Psychic Secretions

The transition of Ivan Petrovich Pavlov from a celebrated gastrointestinal physiologist to the architect of modern conditional reflex theory represents one of the most profound paradigm shifts in the history of experimental biology. Throughout the late 1880s and 1890s, Pavlov’s primary scientific objective was the systematic elucidation of the regulatory mechanisms controlling digestive glandular secretions. Utilizing ingenious surgical preparations that preserved the autonomic nerve supply of visceral organs, Pavlov investigated how the nervous system coordinates the secretion of saliva, gastric juice, and pancreatic enzymes in response to specific chemical and mechanical properties of ingested foodstuffs. It was during these highly controlled physiological assays that an anomalous, disruptive variable repeatedly interfered with experimental trials: the canines began to salivate prior to the physical contact of food with the oral mucosa. The mere footsteps of the laboratory attendant, the sight of the feeding vessel, or the preparation of the surgical apparatus proved sufficient to trigger copious glandular output.

Initially dismissed as an experimental nuisance termed “psychic secretions,” these anticipatory physiological responses posed a conceptual dilemma. Under the prevailing Cartesian paradigm, reflexes were conceived as deterministic, unvarying neural circuits directly linking peripheral receptor stimulation to muscular or glandular effectors. Psychic secretions, conversely, appeared to be mediated by subjective emotional or mental states, rendering them seemingly inaccessible to the objective methods of experimental natural science. Rather than abandoning physiological rigor for the introspective psychology popular in contemporary European philosophy, Pavlov executed an epistemological leap. He resolved that these psychic phenomena were not mystical operations of an unquantifiable psyche, but rather lawful physiological reflexes of a higher order, occurring within the cerebral hemispheres.

Pavlov recognized that to conquer this psychic domain, he needed to treat the anticipatory response not as an obstacle to digestive physiology, but as the primary subject of empirical inquiry. By substituting uncontrolled subjective variables with meticulously calibrated sensory stimuli, Pavlov began to systematically quantify the salivary response as an objective measure of central nervous system dynamics. The volume, drop rate, and chemical latency of salivation were harnessed as continuous physiological variables that mirrored the otherwise invisible excitation and inhibition coursing through the mammalian brain, establishing an empirical methodology that liberated psychology from the constraints of introspection.

1.2 The Evolution of Associative Paradigms in Early Russian Reflexology

The theoretical framework that facilitated Pavlov’s conceptual breakthrough was deeply indebted to the foundational tenets of early Russian reflexology, most notably formulated by Ivan Mikhailovich Sechenov. In his revolutionary 1863 treatise, Reflexes of the Brain, Sechenov postulated that all conscious and unconscious acts of human and animal life are, at their mechanistic core, complex muscular and visceral reflexes initiated by external sensory stimulation. Sechenov asserted that the most intricate cognitive processes—thought, volition, and emotion—are simply reflex arcs wherein the central inhibitory mechanisms of the brain either delay, attenuate, or modify the ultimate motor manifestation. This mechanistic materialism provided Pavlov with the theoretical justification to treat psychic activities as naturalistic reflexes amenable to laboratory dissection.

Operating within this reflexological matrix, Pavlov systematically constructed a precise taxonomic nomenclature to describe the components of associative learning. He demarcated the biological substrate into the Unconditioned Stimulus (US)—a biologically potent agent such as meat powder or a mild acid solution that invariably and innately provokes an automatic, unlearned response—and the Unconditioned Response (UR), which represents the stereotyped physiological reaction, such as salivary or defensive motor discharge. Against this hardwired biological axis, Pavlov introduced the Neutral Stimulus, an environmental perturbation such as the ticking of a metronome, the illumination of a lamp, or the acoustic sounding of a reed tone, which initially elicited only an orienting reflex but no glandular secretion.

Through systematic pairing regimens, the neutral stimulus was converted into a Conditioned Stimulus (CS), capable of evoking a Conditioned Response (CR)—a learned, anticipatory secretion that closely resembled, yet exhibited distinct temporal and chemical nuances from, the unconditioned response. Initially, Pavlovian theory rested upon the classical assumption of temporal contiguity, derived from associationist philosophy extending from Aristotle through John Locke and David Hume. It was presumed that the physical overlap and simultaneous presentation of the CS and US were the fundamental, non-negotiable drivers of neural connection formation, bridging the cortical analyzers of the senses with the subcortical feeding centers of the brain.

1.3 The Emergence of the Trace Conditioning Paradigm

As Pavlov’s empirical investigations deepened in his Saint Petersburg laboratories, anomalies within the strict temporal contiguity framework became increasingly evident. His assistants noticed that associations could form even when an intentional temporal chasm was introduced between the termination of the neutral stimulus and the delivery of the unconditioned reinforcer. When a tactile cutaneous stimulator or an acoustic tone was activated for a brief duration and subsequently extinguished, leaving the experimental chamber in total silence for tens of seconds or even minutes before the presentation of food powder, the animal nonetheless developed robust conditioned salivation to the preceding, long-extinguished cue.

This discovery marked the official birth of the trace conditioning paradigm (in Russian, sledovoi uslovnyi refleks). Pavlov realized that the physical presence of the external stimulus at the moment of reinforcement was not strictly required for the establishment of a conditional reflex. Instead, he formulated the revolutionary concept of the internal neural representation—postulating that the physical stimulus leaves an enduring, functional physiological “trace” (vestige) within the cortical analyzers of the cerebral hemispheres. The nervous system was not merely reacting to incoming waves of immediate environmental energy; it was capable of preserving, stabilizing, and associating an endogenous representation of a stimulus that existed solely within the physical memory architecture of the brain.

Early exploratory studies within the Pavlovian institute sought to delineate the upper temporal boundaries of this phenomenon. Assistants such as P. P. Pimenov, G. P. Zeliony, and B. P. Babkin systematically stretched the silent interval between stimulus offset and reinforcement from brief intervals of five to ten seconds out to profound delays of several minutes. These experiments revealed that as the temporal gap widened, the nature of the conditioned reflex underwent a qualitative metamorphosis. The animal was required not only to form an association across time, but also to inhibit premature responding, thereby differentiating the trace reflex from simple contiguous or short-delay reflexes and elevating it to a distinct category of neurocognitive operation.

2. Defining Trace Conditioning: Temporal Architecture and Core Mechanisms

2.1 The Chronometric Sequence: CS Onset, Offset, and the Trace Interval

The trace conditioning protocol is fundamentally defined by its rigorous chronometric architecture, which distinguishes it mathematically and operationally from all other forms of classical conditioning. The trial begins at time zero with the Conditioned Stimulus Onset ($CS_{onset}$), during which a sensory stimulus of calibrated intensity and physical quality is presented to the subject. The stimulus is maintained for a defined epoch, terminating at the Conditioned Stimulus Offset ($CS_{offset}$). The duration between $CS_{onset}$ and $CS_{offset}$ constitutes the physical stimulus presentation period ($T_{CS}$), which is kept constant across experimental blocks to prevent temporal variance from confounding associative metrics.

The definitive phase of the paradigm begins immediately upon $CS_{offset}$. From this microsecond forward, all physical emanations of the CS cease entirely, plunging the experimental environment back into its baseline state. The silent gap extending from $CS_{offset}$ to the Unconditioned Stimulus Onset ($US_{onset}$) is designated as the trace interval ($T_{trace}$). It is this empty, stimulus-free temporal expanse that defines the computational uniqueness of the paradigm. The total inter-stimulus interval (ISI) in trace conditioning is therefore expressed mathematically as the sum of the CS duration and the trace interval:

$$ISI = T_{CS} + T_{trace}$$

Upon the termination of the trace interval, the unconditioned reinforcer is administered at $US_{onset}$ and sustained through $US_{offset}$. The dynamics of reinforcement delivery require precise mechanical execution; whether administering desiccated meat powder via pneumatic tubes or introducing a titrating acid solution directly onto the lingual surface, the delivery must possess invariant temporal rise times. Across repeated trials, the stability of $T_{trace}$ governs the acquisition trajectory of the animal. High temporal variance across successive presentations degrades associative strength, while absolute chronometric invariance fosters the emergence of highly accurate, temporally calibrated internal representations.

2.2 Differentiating Trace Conditioning from Other Classical Paradigms

To fully appreciate the cognitive and physiological burden imposed by trace conditioning, it must be systematically contrasted with the alternative permutations of classical conditioning paradigms, which vary strictly along the temporal axis. In *simultaneous conditioning*, the CS and the US are initiated at precisely the same temporal instant and co-terminate simultaneously ($ISI = 0$). Paradoxically, despite absolute temporal contiguity, simultaneous conditioning demonstrates remarkably poor associative efficacy, frequently failing to produce robust conditioned responding because the CS possesses no predictive utility regarding the impending arrival of the biological reinforcer.

In *short-delay conditioning*, widely regarded as the most rapid and robust protocol for associative acquisition, the $CS_{onset}$ precedes the $US_{onset}$ by a brief interval (typically 0.5 to a few seconds), but crucially, the CS persists throughout the intervening duration, co-terminating with or slightly overlapping the US. Here, the sensory organs remain continuously activated by external physical energy until the unconditioned stimulus impacts the peripheral receptors. In *long-delay conditioning*, the interval between $CS_{onset}$ and $US_{onset}$ is significantly extended—often mirroring the total ISI of a trace conditioning protocol—yet the physical CS remains continuously present throughout the entirety of the prolonged anticipatory epoch.

Conversely, *backward conditioning* reverses the biological order, presenting the US prior to the CS, which generally engenders conditioned inhibition rather than excitation. Trace conditioning occupies a unique domain among these paradigms. Unlike delay conditioning, wherein the peripheral nervous system and sensory cortex are continuously driven by ongoing exogenous sensory input, trace conditioning demands that the organism bridge the temporal divide via purely endogenous mechanisms. The complete absence of physical energy in the external environment during $T_{trace}$ imposes a substantial cognitive requirement: the central nervous system must sustain an internal representation of a vanished reality, resisting both decay and external interference, while concurrently computing the metric of elapsed time.

2.3 The Concept of the Memory Trace (‘Vestige’) in Pavlovian Theory

Central to Pavlov’s theoretical resolution of trace conditioning was his physiological conceptualization of the “vestige” or cortical memory trace. Pavlov explicitly rejected the notion that once an external physical stimulus ceases to act upon a sensory receptor, its nervous activity terminates instantaneously. Drawing upon his observations of visual after-images and the prolonged excitability of isolated peripheral nerves, Pavlov hypothesized that the impact of a physical stimulus upon a cortical analyzer initiates a lingering state of physiological excitation that persists long after the physical wave of energy has dissipated from the environment.

This enduring excitation—the vestige—was conceptualized not as a static snapshot, but as a dynamic, decaying neurochemical wave. Pavlov postulated that immediately following $CS_{offset}$, the cortical analyzer enters a state of persistent functional reorganization. The trace undergoes lawful decay kinetics over extended temporal intervals; its intensity progressively wanes, shifting from an initial peak of primary sensory excitation into an attenuated, secondary state of lingering cortical arousal. If the trace interval is short, the unconditioned stimulus collides with a robust, highly energetic cortical vestige, resulting in rapid associative bonding.

However, if the trace interval is systematically extended, the vestige decays toward a critical baseline threshold, demanding that the brain engage active neural mechanisms to maintain the functional connection. Pavlov proposed that the formation of a trace conditioned reflex does not link the unconditioned reflex center directly to the external physical stimulus, but rather connects the subcortical unconditioned reflex center to the decaying endogenous trace located within the cortical sensory analyzers. Consequently, the conditioned response becomes attuned not to the sensation itself, but to the specific internal state of neural decay characteristic of that point in time within the cerebral hemispheres.

3. Pavlov’s Experimental Apparatus and Methodological Innovations

3.1 The Tower of Silence: Engineering an Acoustically Isolated Environment

Ivan Pavlov realized that trace conditioning experiments were susceptible to environmental contamination. Because the trace interval requires an animal to maintain a delicate, stimulus-free internal neural state, the occurrence of any extraneous sensory event—a creaking floorboard, the distant hiss of a steam radiator, or the scent of an adjacent animal—could disrupt the decaying neural trace or induce immediate external inhibition. To eliminate these confounding variables, Pavlov orchestrated the design and construction of a dedicated research facility at the Imperial Institute of Experimental Medicine, universally celebrated as the “Tower of Silence” (Bashnya Molchaniya).

Engineered between 1910 and 1914 with financial support from the Russian government and private philanthropists, this three-story research facility represented a triumph of early twentieth-century sensory isolation. The laboratory chambers were constructed with double-thick walls fabricated from specialized porous stone, brick, and acoustic-dampening lead linings. The foundations were excavated deep into the Saint Petersburg soil and rested upon vibration-absorbing shock absorbers to insulate the chambers from the mechanical tremors of municipal carriage and industrial traffic. Double-paned airtight windows, heavy leaded doors sealed with vulcanized rubber gaskets, and complex baffled air-intake shafts ensured that acoustic, thermal, and olfactory stability was maintained within the interior testing chambers.

Within these acoustically deadened chambers, the experimental canine stood positioned upon an elevated wooden research frame, suspended in a soft leather harness designed to comfortably restrict gross locational movement while permitting normal physiological respirations and postural stability. The experimenter was physically excluded from the testing room; all operational interventions, stimulus controls, and biological reinforcers were operated remotely from an adjacent control corridor via an intricate array of pneumatic tubes, hydrostatic lines, and mechanical pulleys. Through these methodological innovations, Pavlov created an invariant sensory vacuum wherein the temporal gap of the trace conditioning paradigm could be scrutinized with absolute empirical purity.

3.2 Precision Measurement of Salivary Secretions

To quantify the associative dynamics within this isolated environment, Pavlov abandoned crude behavioral approximations in favor of continuous, real-time physiological metrics. He developed an ingenious chronic surgical preparation that exteriorized the duct of the parotid or submaxillary salivary gland. Under aseptic surgical conditions, the natural oral opening of the salivary duct—most frequently the parotid duct of Stensen—was dissected from its native mucosal locus on the inner buccal wall, transposed through an incised channel in the canine’s cheek, and sutured directly onto the exterior skin of the animal’s jaw.

Once healed, this chronic salivary fistula allowed for the unobstructed, pure collection of saliva without subjecting the animal to acute surgical trauma or distress during testing sessions. To capture this glandular output with chronometric precision, Pavlov designed a specialized collection apparatus. A small glass or metal funnel was hermetically sealed over the exteriorized fistula using a specialized adhesive compound consisting of a mixture of resin, wax, and zinc oxide. The funnel terminated in an ultra-fine calibrated glass tube or connected to a closed hydrostatic manometer system.

The movement of salivary fluid through the graduated tube was tracked against an external scale. In its most advanced iterations, the saliva displaced drops across an electrical contact or actuated a delicately balanced mechanical balance lever that etched real-time drops onto a smoked-paper kymograph drum rotating at an invariant velocity. This apparatus enabled the Pavlovian researchers to record three distinct physiological dimensions simultaneously:

• The precise latency of the salivary response (measured to within fractions of a second from stimulus onset or offset),
• The instantaneous rate of salivary secretion (drops per unit time), and
• The cumulative volume and biochemical viscosity of the secreted fluid.

Thus, psychic anticipatory processes were rendered as a continuous, mathematically tractable Cartesian curve.

3.3 Systematic Implementation of Trace Conditioning Protocols

The daily execution of trace conditioning within the Tower of Silence adhered to a rigid, standardized laboratory protocol. Pavlov maintained a stable inventory of conditioned stimuli, each calibrated for sensory intensity and temporal rise times. Auditory stimuli included precision mechanical metronomes calibrated to strike at specific rates (e.g., exactly 100 beats per minute versus 144 beats per minute), pure frequency acoustic reed tones, and electric bells. Cutaneous stimuli were administered via “taktors”—pneumatically operated blunt pins or thermal pads mounted directly to the canine’s shaved flank. Visual stimuli were delivered via transilluminated ground-glass screens presenting geometric forms, rotating discs, or discrete illumination changes.

The administrative routine followed a strict chronometric cadence. An automated master clock or mechanical drop timer initiated the trial by triggering the CS (for instance, a 100 Hz tone) for a precise duration of five seconds. At the five-second mark, the tone cut off abruptly. The chamber then entered the silent trace interval, lasting anywhere from a brief ten seconds to extended durations of up to five minutes. During this interval, the pneumatically controlled feeding apparatus stood dormant. Only at the precise expiration of the trace interval did the experimenter depress a pneumatic lever from the external corridor, forcing a calibrated quantity of desiccated meat powder (often mixed with sugar to enhance palatability) or a 0.5% hydrochloric acid solution into a small food dish directly in front of the dog’s snout.

Crucially, Pavlov recognized that the inter-trial interval (ITI) had to be carefully decoupled from the trace interval itself. ITIs were sustained for long durations—typically ranging from five to twenty minutes—and randomized in length to prevent the animal from establishing a temporal conditioned reflex to the overall cyclic rhythm of the experimental session. Furthermore, contextual habituation was systematically achieved by placing the animal in the apparatus for multiple non-reinforced acclimatization sessions, ensuring that salivation occurred exclusively in response to the specific, decaying vestige of the experimental cue.

4. The Mechanics of the Trace Interval: Bridging the Temporal Gap

4.1 Parametric Variations in Trace Duration

Pavlov and his colleagues conducted extensive, parametric explorations into the limits of the trace interval, systematically mapping how variations in temporal duration dictated the speed of acquisition and the asymptotic magnitude of the conditioned response. In his classic tracts, Pavlov detailed experiments in which the trace interval was systematically manipulated across independent cohorts of canines, beginning with ultra-short intervals of 1 to 5 seconds and progressing through 15, 30, 60, 120, and even 300 seconds of complete physical silence prior to reinforcement.

The empirical results yielded a direct, lawful relationship: an inverse correlation exists between the length of the trace interval and the rate of conditioned reflex acquisition. While a short-delay reflex or an ultra-short trace reflex of 1 to 3 seconds could be reliably established within 10 to 20 pairings of the CS and US, extending the trace interval to 30 seconds expanded the required training regimen to 50 to 100 trials. When the trace interval was stretched to one or two minutes, hundreds of pairings over several months of daily experimentation were frequently required before a stable, statistically significant salivary output was observed.

Furthermore, Pavlov discovered that the ultimate asymptotic response magnitude—the total volume of saliva secreted during the trace epoch—decreased precipitously as a function of trace duration. A 10-second trace might elicit an asymptotic discharge of 30 to 40 drops of saliva, whereas an otherwise identical 60-second trace elicited a meager 5 to 10 drops, even after exhaustive training. Beyond a species-specific critical threshold—which in the domestic canine Pavlov identified as approximately three to five minutes—the associative mechanism broke down entirely. Beyond this limit, the biological significance of the transient stimulus vestige dissipated into the contextual noise of the environment, precluding the cortical analyzers from bridging the temporal expanse.

4.2 Temporal Precision and the Anticipatory Response Profile

One of Pavlov’s most profound discoveries concerning trace conditioning was that the conditioned response did not deploy as an immediate, uncontrolled discharge following the offset of the conditioned stimulus. Instead, the salivary response exhibited an exquisite temporal precision that developed gradually across the course of training. In the initial phases of trace reflex acquisition, after the first 15 to 30 pairings, the canine exhibited an uncalibrated, generalized response: the salivary drops began to fall almost immediately upon the offset of the CS and trickled irregularly throughout the entire duration of the silent trace interval.

However, as training progressed into mature stages, a remarkable chronometric transformation occurred. The salivary flow immediately following $CS_{offset}$ ceased entirely. The early portion of the trace interval became a period of absolute glandular quiescence. Only as the trace interval approached its historical termination point—seconds before the anticipated arrival of the food powder or acid—did the parotid gland activate, producing a sudden, accelerating cascade of saliva that reached its apex precisely at the moment of $US_{onset}$.

Pavlov utilized mathematical recordings from his kymograph drums to model these response latency shifts across consecutive training blocks. He observed that the latency of the first drop of saliva migrated systematically to the right along the temporal axis. This anticipatory response profile demonstrated that the animal was not merely passively remembering that a reinforcer was coming; it was maintaining an internal chronometer capable of measuring the exact duration of elapsed time since the sensory stimulus had vanished. The trace conditioned reflex was revealed to be a double reflex: a reflex to the sensory quality of the original stimulus, coupled with an endogenous reflex to the passage of time itself.

4.3 Role of the Inter-Trial Interval and Contextual Salience

The stability of a trace conditioned reflex is fundamentally governed by the ratio between the trace interval and the inter-trial interval (ITI). In the course of his parametric studies, Pavlov observed that the temporal void of the trace interval introduces an acute vulnerability to contextual conditioning. When an unconditioned stimulus is delivered after a prolonged period of silence, the sensory apparatus of the brain does not operate in an absolute sensory vacuum; rather, it is exposed to the static, tonic background cues of the experimental room—the tactile pressure of the harness, the faint ambient illumination, and the olfactory signature of the testing chamber.

Pavlov recognized that if the ITI is too short relative to the trace interval, these tonic contextual cues compete associatively with the transient, decaying sensory trace of the CS. This competitive dynamic, known in modern associative learning theory as overshadowing or context blocking, can erode the predictive salience of the CS. If the canine receives food after 30 seconds of silence, and the interval between successive trials is only 60 seconds, the continuous context is present for a large proportion of the total time preceding reinforcement, leading the brain to associate the ambient environment rather than the transient auditory or visual cue with the food reward.

To overcome this, Pavlov instituted long, unpredictable inter-trial intervals, often maintaining silence for fifteen to thirty minutes between discrete experimental runs. By maximizing the ITI-to-trace-interval ratio, Pavlov ensured that the transient sensory vestige remained far more informative regarding the imminent delivery of food than the tonic background context. This deliberate temporal spacing prevented contextual cues from overshadowing the decaying sensory trace, thereby forcing the cortical analyzers to anchor the associative bond directly to the memory of the transient stimulus.

5. Internal Inhibition and the Delay of the Conditioned Reflex

5.1 Inhibition of Delay (Retardation of Conditioned Reflexes)

The prolonged period of glandular silence observed during the early phases of a mature trace interval presented Pavlov with one of his greatest theoretical challenges. Why did an animal that had formed a powerful associative bond between a stimulus and food remain entirely motionless and dry-mouthed during the initial 30 or 60 seconds following stimulus offset? A simplistic theory of associative learning would dictate that the presence of the trace should immediately activate the salivary center. Pavlov reasoned that this quiescence could not be explained as a passive failure of transmission or a mere absence of excitation; rather, it represented a profound, active physiological counter-force operating within the cerebral cortex, which he designated as inhibition of delay (retardation of conditioned reflexes).

In Pavlovian neurophysiology, inhibition of delay is classified as a primary form of *internal inhibition*—a process generated intrinsically within the cortical circuits when reinforcement is systematically postponed. Pavlov partitioned the chronometric architecture of the trace interval into two distinct physiological phases:

1. An initial, non-active inhibitory phase, and
2. A terminal, active excitatory phase.

During the non-active phase immediately following $CS_{offset}$, the animal actively suppresses the conditioned response because responding at that juncture is biologically premature, inefficient, and wasteful of metabolic resources.

The behavioral manifestations of this internal inhibitory state were striking. During the silent, non-active phase of a prolonged trace interval (for example, a three-minute trace), experimental dogs did not merely exhibit salivary arrest; they frequently displayed profound motor passivity, muscular relaxation, and even somnolence. The animal’s eyes would glaze over, its head would droop within the leather harness, and it would enter a state resembling light hypnosis or sleep. Yet, as the final seconds of the trace interval drew near, the animal would abruptly awaken, become hyper-alert, orient its head toward the food delivery mechanism, and salivate profusely. The quiescent phase was thus proven to be an active, energy-consuming suppression of neural output that held the underlying excitatory impulse in check.

5.2 The Dynamics of Cortical Excitation and Internal Inhibition

Pavlov conceptualized the central nervous system as a dynamic arena governed by the continuous interplay of two reciprocal, antagonistic processes: excitation and inhibition. To account for the temporal transition from silence to salivation within the trace interval, he formulated his classic wave-like model of *cortical irradiation and concentration*. When a conditioned stimulus is initially presented, it strikes a focal point within the corresponding cortical analyzer (e.g., the auditory cortex in the temporal lobes), generating an immediate focus of cortical excitation.

Following stimulus offset, this excitation does not vanish; rather, as the physical wave ceases, the cortex generates a counter-wave of internal inhibition designed to extinguish premature execution of the reflex. Pavlov hypothesized that this wave of inhibition originates at the focal point of the analyzer and irradiates across the surrounding cortical tissue, temporarily depressing both local and adjacent associative circuits. This accounts for the profound behavioral suppression and drowsiness observed during the early trace interval. Inhibition dominates the cortical landscape, holding the associative memory in a state of suspended animation.

However, this inhibitory wave is not permanent. Driven by an endogenous, physiological timing mechanism that Pavlov struggled to define with contemporary histological tools, the inhibitory wave begins to recede—or “concentrate”—back toward its point of origin as the chronological moment of reinforcement approaches. As the inhibitory blanket contracts, the latent, preserved focus of excitation re-emerges, irradiates rapidly across the cortical mosaic, and travels along established functional pathways to the subcortical salivary centers in the medulla and pons. The trace interval is therefore not a static temporal gap, but a dynamic, undulating cortical battleground where an initial surge of excitation is engulfed by an expansive wave of internal inhibition, which ultimately concentrates to liberate the final, anticipatory excitatory discharge.

5.3 Disinhibition Phenomena in the Trace Period

To prove conclusively that the silent non-active phase of the trace interval was maintained by active internal inhibition rather than a passive lack of neural excitation, Pavlov executed a series of brilliant experimental maneuvers centered on the phenomenon of *disinhibition* (rastormazhivanie). He reasoned that if the silence of the salivary glands was caused by an active inhibitory shield masking an underlying reservoir of excitation, then the sudden introduction of a novel, extraneous stimulus during this silent phase should disrupt the delicate inhibitory process, thereby unmasking the latent excitation.

The empirical verification of this hypothesis was dramatic. Pavlov trained dogs on a long-trace conditioning protocol—for instance, a tactile stimulus applied for five seconds, followed by an empty two-minute trace interval, at the end of which acid was delivered. Once the dog reliably exhibited complete salivary silence during the first 90 seconds of the trace, Pavlov intervened during this quiescent window. At the 30-second mark of the silent interval, when the dog was typically motionless and dry-mouthed, the experimenter introduced a mild, novel sensory distractor—such as the faint visual illumination of a red bulb, the gentle hiss of an unfamiliar air valve, or a soft acoustic bell.

Under ordinary conditions, this novel stimulus would elicit merely an orienting reflex (the “what is it?” reflex) without causing salivation. But when introduced during the silent, inhibitory phase of the trace interval, the novel stimulus immediately provoked an explosive, paradoxical discharge of saliva. The extraneous stimulus excited the cortex, but instead of adding to the inhibition, it disrupted and destroyed the fragile, energy-intensive process of internal inhibition of delay. With the inhibitory brake abruptly shattered by the novel distractor, the underlying, latent excitation that had been silently accumulating was instantly liberated, causing the salivary glands to fire long before the proper temporal arrival of the reinforcer. This disinhibition phenomenon provided irrefutable empirical proof that the trace interval harbored an active, continuous, and robust neural representation that was merely kept under physiological lock and key by cortical inhibition.

6. Comparative Analysis: Trace Conditioning Versus Delay Conditioning

6.1 Structural and Procedural Divergence

The structural divergence between trace conditioning and delay conditioning represents far more than an arbitrary technical adjustment of experimental chronometry; it reflects an entirely different level of neurobiological organization and cognitive computation. In standard delay conditioning, the conditioned stimulus acts as an ongoing, continuous physical presence throughout the entire preparatory period. The peripheral sensory receptors—whether photoreceptors in the retina, hair cells in the cochlea, or mechanoreceptors in the dermis—are continuously bombarded by exogenous physical waves of energy right up until, or through, the onset of the unconditioned reinforcer.

In sharp contrast, the architectural hallmark of trace conditioning is the absolute *temporal void* that separates the offset of the sensory cue from the onset of the biological reinforcer. Once $CS_{offset}$ occurs, the sensory organs are immediately liberated from physical drive. There is no physical energy in the external universe linking the past event to the future biological consequence. This radical procedural divergence fundamentally alters the physiological task confronted by the organism:

Dimension	Delay Conditioning	Trace Conditioning
Physical CS Status at US Onset	Present and active; directly stimulating peripheral receptors	Completely absent; extinct from the external environment
Sensory Engagement	Continuous bottom-up peripheral driving of sensory pathways	Transient sensory drive followed by purely endogenous neural activity
Primary Neural Demand	Direct sensory-to-visceral associative mapping	Active working memory maintenance and temporal interval computation
Resistance to Distraction	Highly robust; continuous external stimulus preserves predictive drive	Exceptionally fragile; extraneous inputs collapse the internal trace
Circuitry Involvement	Largely subcortical and primary sensory cortex-dependent	Strictly requires higher corticolimbic structures (Hippocampus, mPFC)

The structural absence of physical stimuli during the predictive phase shifts the computational burden from a simple stimulus-response integration to an intricate cognitive process. The trace protocol forces the brain to rely on internal neural representations, requiring mechanisms of memory storage, autonomous temporal calculation, and the active shielding of these fragile representations from constant sensory interference.

6.2 Acquisition Dynamics and Asymptotic Performance

When evaluated across identical inter-stimulus intervals (ISIs), the acquisition dynamics of trace conditioning diverge sharply from those of delay conditioning. Under delay protocols—particularly short-delay paradigms where the ISI ranges from 500 milliseconds to a few seconds—learning curves are steep and rapid. An animal often achieves conditioned response criteria within a small number of training blocks. The continuous sensory guidance allows for rapid synaptic reinforcement between the sensory input channels and the unconditioned motor or autonomic nuclei.

In trace conditioning, however, acquisition is markedly retarded. Even when the total ISI of a trace protocol is matched precisely to the ISI of a long-delay protocol, the trace architecture requires a significantly greater number of trials to achieve stable acquisition. Pavlov demonstrated that if a dog is subjected to a 30-second long-delay conditioned stimulus (where a buzzer sounds continuously for 30 seconds before food is delivered), conditioned salivation typically emerges within 20 to 30 pairings. But if the same total 30-second interval is converted into a trace protocol (where the buzzer sounds for 5 seconds, followed by a 25-second silent trace), the number of pairings required to establish consistent conditioning often quadruples.

Furthermore, the asymptotic performance achievable under trace conditioning is quantitatively inferior to that achieved under delay paradigms. Under equivalent motivational states of hunger and identical concentrations of meat powder reinforcement, trace-conditioned subjects consistently secrete a lower total volume of saliva per trial. The salivary flow during a trace reflex is more variable, displaying trial-to-trial fluctuations in latency and volume that are absent in the highly stereotyped, machine-like execution of delay reflexes. This asymptotic disparity highlights the fact that an endogenous, decaying memory trace possesses less associative efficacy and lower functional potency than the raw, vivid, continuous physical excitation driven by an unyielding external stimulus.

6.3 Differential Susceptibility to Extinction and Interference

The differential stability of trace and delay reflexes is further demonstrated by their respective susceptibilities to extinction and sensory interference. In classical extinction regimens, the conditioned stimulus is presented repeatedly in the complete absence of the unconditioned reinforcer. Under these non-reinforced conditions, trace conditioned reflexes extinguish at an accelerated rate compared to delay reflexes. While a robust delay reflex might endure 20, 30, or 50 unreinforced presentations before salivary cessation occurs, a trace conditioned reflex frequently collapses after only a handful of non-reinforced trials.

This rapid extinction kinetics highlights the inherent fragility of the trace mechanism. Because trace conditioning requires the coordinated maintenance of both an internal memory vestige and an active inhibitory timing shield, the failure of the reinforcer to appear rapidly destabilizes the entire neural construction. Furthermore, trace conditioning exhibits acute vulnerability to both *proactive* and *retroactive sensory interference*. If extraneous, irrelevant sensory stimuli occur either immediately before the presentation of the CS or during the silent trace interval itself, the acquisition and execution of the conditioned reflex are severely disrupted.

In delay conditioning, an unexpected ambient noise occurring during the CS presentation may cause a fleeting decrement in salivation, but the continuous presence of the CS rapidly reasserts control over the salivary centers. In trace conditioning, an unexpected sensory event occurring during the silent trace gap permanently obliterates the decaying memory vestige for that trial, resetting the internal clock and plunging the animal into behavioral confusion. Finally, trace conditioning is profoundly vulnerable to fluctuations in the subject’s central nervous system arousal; minor states of fatigue, mild sedation, or subtle environmental shifts that leave delay conditioning entirely intact will cause a complete collapse of trace conditioned responding.

7. The Neurobiological Hypotheses of Ivan Pavlov: Cortical Mosaics and Excitation

7.1 The Pavlovian Model of the Cerebral Cortex

To provide a mechanistic explanation for the phenomena uncovered by his experiments, Ivan Pavlov formulated a theoretical model of the cerebral hemispheres that anticipated modern concepts of functional localization, neural networks, and receptive field organization. Pavlov conceptualized the cerebral cortex as a vast, continuous “cortical mosaic” (korkovaya mozaika) composed of countless millions of cellular elements, each functioning as a specialized receptor or processor of incoming information. This mosaic was not a fixed, rigid map, but a dynamic, undulating surface where microscopic foci of excitation and inhibition constantly shifted, merged, collided, and resolved.

Within this mosaic, Pavlov distinguished specialized structural-functional regions he termed *analyzers*. An analyzer was defined as an integrated neurofunctional apparatus consisting of three continuous components:

• The peripheral sensory receptor (e.g., the retina, the organ of Corti, cutaneous nerve endings),
• The subcortical afferent conduction pathways (the sensory nerves, brainstem relays, and thalamic nuclei), and
• The cortical termination center, which represented the central, highly differentiated core of the analyzer situated within the cerebral cortex.

Pavlov identified specific cortical zones for each sensory modality—the temporal lobes as the auditory analyzer, the occipital poles as the visual analyzer, the postcentral gyri as the cutaneous-tactile analyzer, and the piriform and orbital regions as the chemical (gustatory-olfactory) analyzers.

For Pavlov, the cerebral cortex was not merely an accessory organ of higher intellectual consciousness; it was the obligatory, indispensable organ for the synthesis of all conditional reflexes, and most critically, trace reflexes. He argued that while simple unconditioned reflexes were mediated autonomously by subcortical, brainstem, and spinal circuits, the synthesis of a temporary functional connection bridging two disparate events across time required the plastic, cellular architecture found exclusively within the cortical mosaic of the cerebral hemispheres.

7.2 Cortical Irradiation, Concentration, and Induction

To explain how a sensory event in one analyzer could associate with a biological event in another across a temporal gap, Pavlov posited three fundamental laws governing cortical activity: *irradiation*, *concentration*, and *reciprocal induction*. As previously outlined, when a stimulus stimulates a peripheral receptor, an initial focus of physiological excitation is ignited within the corresponding cortical analyzer. According to the law of irradiation, this excitation does not remain confined to its point of origin; rather, it spreads across the cortical sheet like ripples on the surface of water, traveling through horizontal intracortical fibers and traversing the boundaries of adjacent analyzers.

Following this initial period of irradiation, a counter-acting physiological process is initiated: the law of concentration dictates that the diffuse wave of excitation is gradually drawn back, concentrating into a discrete, highly focused functional locus within the primary analyzer. An identical set of dynamics was ascribed to the process of internal inhibition, which likewise irradiated across the hemispheres before concentrating into distinct inhibitory zones. Around every concentrated focus of excitation or inhibition, a reciprocal process was automatically ignited—a phenomenon Pavlov termed *induction*. A concentrated focus of excitation induced a ring of inhibition around its perimeter (negative induction), while a concentrated focus of inhibition induced an enhanced zone of excitability in the surrounding cortical tissue (positive induction).

Pavlov applied these laws directly to account for trace conditioning. When the CS is presented, it sparks excitation that irradiates and then concentrates within its specific analyzer. Upon $CS_{offset}$, the focus of excitation does not immediately collapse; rather, under the influence of positive induction and intrinsic cellular resonance, it persists as a localized, decaying functional vestige. As the silent trace interval progresses, negative induction and internal inhibition sweep across the cortex, actively suppressing premature activation of the subcortical unconditioned centers. When the trace interval terminates, this inhibition concentrates, and the persistent excitation breaks free, firing down through descending cortico-bulbar pathways to ignite the unconditioned salivary nuclei in the lower brainstem.

7.3 Surgical Ablation and Decortication Studies

Pavlov was not content to leave his cortical hypotheses in the realm of theoretical conjecture. To verify the absolute necessity of the cerebral cortex for the establishment and maintenance of trace conditioned reflexes, he and his surgical associates embarked upon an extensive series of chronic surgical ablation and decortication experiments in canines. Utilizing meticulous surgical techniques, Pavlov’s team performed selective, localized extirpations of specific sensory analyzers, unilateral hemispherectomies, and ultimately, complete bilateral decortications—the total surgical removal of the cerebral neocortex while leaving the underlying basal ganglia, thalamus, and brainstem intact.

The results of these radical surgical interventions were decisive and historically illuminating. When Pavlov subjected completely decorticated dogs to classical conditioning regimens, he discovered a striking functional dissociation. Simple, contiguous, short-delay conditioned reflexes could, with extreme difficulty and extensive training, occasionally be recovered or newly established in decorticated preparations, relying on residual subcortical relays in the thalamus and midbrain. These animals could salivate to the direct, continuous presence of a loud auditory tone that co-terminated with food powder.

However, when it came to *trace conditioning*, the effect of bilateral decortication was absolute, catastrophic, and irreversible. Trace conditioned reflexes were completely abolished. A dog deprived of its cerebral cortex was fundamentally incapable of forming or maintaining an associative connection across even the briefest temporal void. If a stimulus was terminated only five seconds prior to the presentation of food, the decorticated animal never learned to bridge the gap; the unconditioned stimulus arrived as an unexpected, disconnected biological event. Furthermore, localized cortical lesions of the specific analyzer corresponding to the CS modality (e.g., bilateral extirpation of the temporal lobes for auditory stimuli) permanently destroyed trace reflexes to that specific modality, while leaving trace conditioning in alternative, intact analyzers preserved. These surgical extirpation studies provided definitive proof that the preservation of an endogenous stimulus trace across time is a high-level, neocortically mediated cognitive operation.

8. Modern Neuroanatomy of Trace Conditioning: Hippocampus and Prefrontal Cortex

8.1 The Essential Role of the Hippocampal Formation

In the decades following Pavlov’s initial discoveries, the evolution of modern neuroanatomy and behavioral neuroscience has fundamentally expanded and refined our understanding of the anatomical circuits responsible for bridging the temporal void in trace conditioning. The most pivotal discovery in this modern era was the identification of the hippocampal formation as the obligatory, time-dependent neural bridge required for trace, but not delay, conditioning. Beginning with the pioneering investigations of Richard F. Thompson, Solomon, Moyer, and colleagues utilizing the mammalian eyeblink conditioning preparation, a striking double-dissociation was established that firmly separated the neuroanatomical architecture of these two paradigms.

While standard delay eyeblink conditioning (wherein the conditioned tone overlaps with a corneal airpuff) requires only an intact olivocerebellar circuit and can proceed flawlessly in completely decerebrate or hippocampectomized animals, trace eyeblink conditioning requires the structural and functional integrity of the hippocampus. Bilateral lesions or pharmacological inactivations of the hippocampus—specifically targeting the CA1 and CA3 pyramidal cell fields or the dentate gyrus—completely abolish the acquisition and retention of trace conditioned responses, while leaving delay conditioning to the identical physical stimuli entirely unperturbed.

Electrophysiological recordings from hippocampal neurons in awake, behaving animals during trace conditioning protocols have unmasked the precise cellular dynamics that underpin this dependency. As the physical CS terminates and the animal enters the silent trace interval, populations of CA1 and CA3 pyramidal cells do not fall silent; instead, they exhibit robust, highly structured changes in their firing rates. Certain populations of CA1 neurons demonstrate sustained, elevated firing throughout the entire duration of the trace gap, physically bridging the silence with persistent action potentials. Other hippocampal ensembles display complex temporal tiling—firing sequentially in time like a biological bucket brigade, where individual cells discharge at precise chronological offsets (e.g., cell A fires from 100-200 ms post-CS, cell B from 200-300 ms, cell C from 300-400 ms). These “time cells” within the hippocampus provide an endogenous metric of elapsed duration, supplying the downstream cerebellar and motor circuits with the precise temporal signal necessary to calculate the anticipated arrival of the unconditioned stimulus.

8.2 Prefrontal Cortical Circuits in Working Memory Maintenance

Working in tight anatomical and functional reciprocity with the hippocampal formation is the prefrontal cortex, particularly the medial prefrontal cortex (mPFC, encompassing the prelimbic and infralimbic cortices in rodents, corresponding functionally to the dorsolateral and anterior cingulate cortices in primates). While the hippocampus is critical for the initial encoding, consolidation, and temporal indexing of the trace association, the mPFC provides the active working memory architecture necessary to sustain the internal representation of the vanished CS against sensory noise and decay.

During the trace interval, multi-unit recordings within the mPFC demonstrate the phenomenon of *persistent neuronal firing*. Prefrontal pyramidal neurons, supported by dense recurrent collaterals and intrinsic membrane conductances, maintain elevated firing states that outlast the physical duration of the sensory stimulus by several seconds or minutes. This persistent activity serves as the contemporary neurophysiological instantiation of Pavlov’s theoretical “vestige.” The mPFC effectively holds the stimulus representation “in mind,” providing an active top-down executive bias that maintains the memory active across the temporal chasm.

Modern optogenetic and chemogenetic dissections have revealed the exquisite precision of the prefrontal-hippocampal dialogue during trace learning. Monosynaptic projections extending from the intermediate and ventral CA1/subiculum of the hippocampus directly into the prelimbic cortex are dynamically synchronized during the trace interval. Optogenetic inhibition of this pathway specifically during the silent trace gap—leaving the physical CS presentation and the US presentation completely uninhibited—is sufficient to disrupt trace conditioning acquisition. Furthermore, as the trace memory undergoes systems consolidation over weeks and months, the primary neuroanatomical locus for memory storage shifts; while the hippocampus is acutely required for the acquisition and early expression of trace conditioning, remote trace memories eventually become independent of the hippocampus and reside permanently within consolidated prefrontal neocortical networks, vindicating Pavlov’s original insistence on the ultimate neocortical home of the trace.

8.3 Cerebellar and Amygdalar Engagement Across Paradigms

While the hippocampus and prefrontal cortex are indispensable for sustaining the memory representation across the trace interval, they do not act as the ultimate motor or autonomic effectors of the conditioned response. Instead, these higher corticolimbic structures must interface dynamically with subcortical structures that house the core associative and motor-timing machinery: the cerebellum and the amygdala.

In somatic motor conditioning models, such as the classically conditioned eyeblink or nictitating membrane response, the cerebellum serves as the primary site of plastic convergence. The unconditioned stimulus (a corneal airpuff or periorbital shock) is transmitted via the inferior olive through climbing fibers to the cerebellar cortex (Purkinje cells) and the deep cerebellar nuclei (specifically the interpositus nucleus). In standard delay conditioning, the CS is routed directly from the pontine nuclei via mossy fibers to these same cerebellar circuits. In trace conditioning, however, because the physical CS has terminated long before the US arrives, the mossy fiber input from the sensory pontine nuclei decays too quickly to overlap with the climbing fiber signal. To solve this problem, the brain routes the trace representation through the hippocampus and mPFC, which subsequently project back down to the basilar pontine nuclei, providing a sustained, artificial “bridge” of mossy fiber input to the cerebellum. Thus, the cerebellum executes its classic role in motor timing and response topography, but it is entirely dependent on higher corticolimbic top-down drive to supply the missing sensory signal.

A complementary architecture governs emotional and autonomic trace conditioning, such as trace fear conditioning or the visceral salivary paradigms of Pavlov. In trace fear conditioning, where a transient tone is followed by an empty trace interval and a subsequent footshock, the basolateral amygdala (BLA) serves as the primary subcortical locus of associative plastic change. Plasticity within the lateral and basolateral nuclei of the amygdala is essential for attaching affective valence to the conditioned stimulus. Just as the cerebellum requires hippocampal and prefrontal input to time motor execution across a trace gap, the amygdala requires direct synaptic input from the ventral hippocampus and mPFC to sustain the emotional representation of threat across the silent interval. The modern neuroanatomy of trace conditioning thus reveals an integrated, multi-level hierarchy: corticolimbic networks bridge time and preserve memory, while subcortical cerebellar and amygdalar engines orchestrate the millisecond-precise execution of motor, glandular, and autonomic output.

9. Cellular and Molecular Substrates of the Memory Trace

9.1 Synaptic Plasticity and Long-Term Potentiation (LTP)

At the microscopic level, the capacity of the central nervous system to bridge the temporal discontinuity of trace conditioning rests upon specialized mechanisms of synaptic plasticity, most notably Long-Term Potentiation (LTP) and intrinsic neuronal excitability adjustments. Under standard Hebbian learning rules, synaptic strengthening requires the simultaneous, near-instantaneous co-activation of presynaptic and postsynaptic elements—a concept summarized by the adage “cells that fire together, wire together.” Trace conditioning, by its very design, violates naive Hebbian temporal mechanics; presynaptic activity driven by the CS ceases long before postsynaptic depolarization driven by the US takes place.

To overcome this biophysical paradox, the nervous system relies on the activation of the N-methyl-D-aspartate (NMDA) subclass of ionotropic glutamate receptors within the hippocampus and prefrontal cortex. The NMDA receptor functions as a molecular coincidence detector, requiring both presynaptic glutamate binding and concurrent postsynaptic membrane depolarization to dislodge a resting magnesium ($Mg^{2+}$) ion block from its channel pore. In trace conditioning, the induction of NMDA-dependent LTP within the CA1 region of the hippocampus is strictly mandatory. Pharmacological blockade of NMDA receptors via intra-hippocampal infusions of antagonists such as APV (DL-2-amino-5-phosphonovalerate) completely prevents the acquisition of trace conditioned responses while leaving delay conditioning fully intact.

Beyond standard synaptic LTP, trace conditioning induces dramatic changes in the *intrinsic excitability* of hippocampal pyramidal neurons. Intracellular recordings from CA1 neurons harvested from trace-conditioned animals reveal a marked, persistent reduction in the slow afterhyperpolarization (sAHP). The sAHP is a calcium-activated potassium conductance that normally acts as a powerful intrinsic brake on neuronal firing; by downregulating the channels mediating the sAHP, trace conditioning renders these pyramidal cells globally more excitable and capable of sustained, high-frequency burst firing. In parallel, mechanisms of Long-Term Depression (LTD) operating at inhibitory interneuron synapses are recruited to shape the temporal window of excitation, sculpting the precise millisecond-level emergence of the conditioned response and providing a cellular substrate for Pavlov’s “inhibition of delay.”

9.2 Intracellular Signaling Cascades and Gene Expression

The translation of transient calcium influx across the postsynaptic density into an enduring, physical memory trace capable of lasting a lifetime requires the activation of complex intracellular biochemical cascades and de novo gene transcription. Calcium ions ($Ca^{2+}$) entering the dendritic spine via NMDA receptor channels and L-type voltage-gated calcium channels bind to calmodulin, triggering the rapid autophosphorylation of Calcium/Calmodulin-Dependent Protein Kinase II (CaMKII). Activated CaMKII translocates to the postsynaptic density, where it directly phosphorylates AMPA receptor subunits (GluA1), increasing their single-channel conductance and promoting the insertion of additional AMPA receptors into the postsynaptic membrane.

Concurrently, the calcium signal stimulates calcium-sensitive adenylyl cyclases (AC1 and AC8), precipitating a sharp surge in intracellular cyclic adenosine monophosphate (cAMP). This activates Protein Kinase A (PKA) and initiates the Mitogen-Activated Protein Kinase / Extracellular Signal-Regulated Kinase (MAPK/ERK) signaling cascade. Phosphorylated ERK translocates from the dendritic arbor directly into the cell nucleus, where it targets the nuclear transcription factor cAMP Response Element-Binding Protein (CREB) at Serine-133.

Phosphorylation of CREB triggers the transcription of immediate early genes (IEGs)—including c-Fos, Egr1 (Zif268), and Arc (Activity-Regulated Cytoskeleton-Associated Protein)—which orchestrate the structural remodeling of synaptic spines, the synthesis of new scaffolding proteins, and the stabilization of newly expanded synaptic terminals. Pharmacological or genetic disruption of the PKA-ERK-CREB pathway selectively eliminates the long-term consolidation of trace conditioning, leaving the short-term working memory acquisition temporarily functional but completely unable to convert into a durable, permanent trace. Furthermore, recent epigenetic investigations have revealed that trace conditioning alters histone acetylation and DNA methylation profiles within both the hippocampus and the medial prefrontal cortex, structurally altering chromatin architecture to permit the long-term persistence of the association.

9.3 Neurotransmitter Modulation: Acetylcholine, Dopamine, and GABA

The intricate biophysical and transcriptional events supporting trace conditioning do not occur within an isolated neurochemical medium; they are heavily modulated by the concerted actions of ascending neuromodulatory systems: acetylcholine, dopamine, and gamma-aminobutyric acid (GABA).

The cholinergic system, originating in the basal forebrain (specifically the medial septum and the vertical limb of the diagonal band of Broca) and projecting densely to the hippocampus and neocortex, is critical for trace conditioning. Acetylcholine, acting via muscarinic $M_1$ receptors, directly suppresses the potassium conductances that generate the afterhyperpolarization, thereby facilitating the sustained, persistent firing necessary to bridge the trace interval. Furthermore, high cholinergic tone enhances sensory signal-to-noise ratios, preventing extraneous environmental noise from overwriting the fragile decaying trace. Systemic or intra-hippocampal administration of the muscarinic antagonist scopolamine completely abolishes the acquisition of trace conditioning, while having negligible disruptive impact on simple delay conditioning.

The *dopaminergic system*, ascending from the ventral tegmental area (VTA) and substantia nigra compacta, plays a dual role: it provides both a motivational reinforcement signal and an essential computational mechanism for solving the temporal credit assignment problem. Midbrain dopamine neurons project to both the mPFC and the hippocampus, where $D_1$ receptor activation promotes the stability of persistent prefrontal firing states and lowers the threshold for NMDA-dependent synaptic plasticity. During trace conditioning, dopaminergic neurons emit phasic prediction error signals that dynamically update as the animal learns to bridge the trace gap, shifting their activation from the time of the US back to the onset of the CS, thereby signaling the acquired predictive value of the vanished cue.

Finally, the *GABAergic inhibitory network* provides the microcircuit machinery that sculpts the temporal dynamics of the trace interval. Local GABAergic interneurons—specifically parvalbumin-positive ($PV^+$) basket cells and somatostatin-positive ($SST^+$) interneurons—generate rhythmic gamma and theta oscillations that coordinate the precise firing of pyramidal ensembles. $PV^+$ interneurons deliver feedforward and feedback inhibition that prevents runaway excitation during the trace interval, actively establishing the physiological substrate for Pavlov’s “inhibition of delay.” Experimental optogenetic silencing of cortical or hippocampal interneurons destroys the animal’s ability to time its conditioned responding, causing the anticipatory response to lose its chronometric accuracy and revert to premature, uncalibrated behavioral output.

10. Extinction, Disinhibition, and Spontaneous Recovery in Trace Paradigms

10.1 The Dynamics of Trace Conditioning Extinction

Extinction of a trace conditioned reflex occurs when the conditioned stimulus is repeatedly administered in the complete absence of the unconditioned reinforcer ($CS \rightarrow no US$). As the non-reinforced presentations accumulate, the conditioned salivary output undergoes a systematic, progressive decrement until glandular secretion ceases entirely. However, consistent with Pavlov’s original neurobiological assertions, extinction is not a passive process of forgetting, nor does it represent the structural erasure or unlearning of the previously acquired associative memory trace.

Instead, modern neurobiology has confirmed that extinction represents the formation of an active, inhibitory *new learning memory* that competes with and suppresses the original excitatory trace. During trace extinction, the organism must form an association between the $CS_{offset}$, the subsequent silent trace interval, and the novel consequence: biological non-reinforcement. The kinetics of this process are fundamentally different from delay extinction. Because the trace conditioned response is already maintained under a delicate equilibrium of internal inhibition of delay, non-reinforcement rapidly destabilizes this balance.

Extinction of a trace reflex exhibits pronounced *contextual dependency* (the renewal effect). If an animal undergoes trace conditioning in Context A, undergoes extinction in Context B, and is subsequently placed back into Context A, the extinguished trace response instantly and vigorously reappears (ABA renewal). This demonstrates that the original memory trace remained structurally intact within the brain’s circuitry throughout the extinction protocol, held in check solely by a fragile, context-dependent inhibitory mantle coordinated by the infralimbic region of the medial prefrontal cortex and the basolateral amygdala.

10.2 Mechanisms of Spontaneous Recovery and Reinstatement

The non-destructive nature of extinction in trace conditioning is unequivocally confirmed by the phenomena of *spontaneous recovery* and *reinstatement*. If a subject undergoes an exhaustive extinction protocol until salivary secretion has completely ceased, and is then simply removed from the experimental apparatus and allowed to rest in its home environment for twenty-four hours, the reintroduction of the animal to the testing chamber yields a profound behavioral resurgence: the first presentation of the non-reinforced CS evokes an immediate, vigorous conditioned response.

Pavlov accounted for spontaneous recovery by invoking the differential fatigue rates of excitation and inhibition. He hypothesized that internal inhibition is an active, highly demanding physiological state that is intrinsically more unstable and exhaustible than basic neural excitation. Over an extended rest period, the newly formed cortical inhibition that actively suppressed the reflex during extinction naturally wanes, dissipates, or recedes, allowing the underlying, resilient excitatory trace memory to emerge once again into behavioral expression.

In a related manner, the phenomenon of *reinstatement* occurs when an animal with an extinguished trace reflex is exposed to an isolated, non-contingent presentation of the unconditioned stimulus alone (e.g., a single delivery of meat powder or footshock) in the experimental context, without any presentation of the CS. When the CS is subsequently tested shortly thereafter, the extinguished trace conditioned response is instantly restored. The unsignaled US delivery resets the motivational state of the animal, elevating the baseline excitability of subcortical and corticolimbic circuits and stripping away the fragile layer of prefrontal inhibition that was established during extinction training.

10.3 Experimental Neuroses Induced by Trace Conflict

Among Pavlov’s most clinically significant and dramatic discoveries was the realization that pushing the temporal limits of trace conditioning could induce acute psychopathological breakdowns in his animal subjects—a condition he famously designated as *experimental neurosis* (eksperimental’nyi nevroz). These breakdowns were not caused by physical pain, mechanical injury, or chemical toxins, but arose entirely from an irreconcilable, internal functional conflict between the two fundamental processes of the nervous system: extreme, concentrated excitation colliding directly with extreme, taxing internal inhibition.

To induce this experimental neurosis using the trace paradigm, Pavlov would train an animal on an already demanding trace conditioning protocol—such as a three-minute silent trace interval—and then deliberately alter the experimental parameters to increase the cognitive and inhibitory strain. He would either systematically extend the trace interval to five, ten, or fifteen minutes, demand ultra-fine sensory discriminations between two nearly identical trace cues (where one tone signaled food after a three-minute delay, while an adjacent fractional tone signaled non-reinforcement), or suddenly introduce an intense, terrifying sensory distractor during the peak of the inhibitory delay phase.

When the internal inhibitory strain exceeded the biological capacity of the dog’s cerebral cortex, the animal’s behavior fractured catastrophically. Dogs that had previously been docile, cooperative, and highly reliable in the experimental frame suddenly exhibited severe behavioral deterioration. Pavlov documented two distinct clinical profiles of experimental neurosis, which depended upon the animal’s underlying biological temperament:

• The Excitable / Agitated Type: The animal became uncontrollably restless, barked furiously, bit at the leather harness and collection apparatus, showed wild tachypnea and tachycardia, and permanently lost all capacity for internal inhibition. All conditioned reflexes became chaotic, and the dog salivated indiscriminately to any environmental perturbation.
• The Inhibited / Catatonic Type: The animal collapsed into a state of profound motor immobility, stupor, and continuous somnolence. The moment the dog was placed into the experimental frame, its limbs became flaccid or exhibited waxy flexibility; it refused all food, remained entirely unresponsive to conditioned stimuli, and entered a state resembling human catatonic schizophrenia or severe depressive stupor.

These pathological states often persisted for months or years, requiring extensive therapeutic interventions (including prolonged bromide therapy, warm baths, and extended rest in rural environments) to achieve fractional recovery. Pavlov’s trace-induced experimental neuroses provided psychiatry with its first objective, empirical laboratory model of stress-induced cognitive breakdown and psychiatric trauma.

11. Methodological Challenges, Variables, and Vulnerabilities in Trace Conditioning

11.1 Vulnerability to Sensory Interference and Distraction

The methodological execution of trace conditioning is notoriously difficult, presenting empirical challenges that far exceed those encountered in standard delay paradigms. The principal vulnerability of the trace protocol resides in its extreme susceptibility to both internal and external *sensory interference*. Because the bridging of the trace interval relies on an endogenous, decaying neurochemical trace rather than a continuously present external driver, any unexpected sensory intrusion during the silent gap can permanently degrade the memory representation.

In modern laboratories, researchers must confront the reality of sensory contamination. If an auditory trace conditioning experiment is conducted in an unshielded room, transient background events—the cycling of an air conditioning compressor, the distant closing of a door, or the ultrasonic vocalizations of adjacent rodents—act as profound retroactive distractors. In human research, unexpected visual saccades, eye blinks, or wandering thoughts occurring during the trace interval interrupt prefrontal persistent firing and disrupt hippocampal place and time cell coordination.

To insulate against these disruptive forces, experimental designs must enforce absolute standardization. Modern sound-attenuating cubicles equipped with white noise generators are required to mask ambient auditory fluctuations. Automated computerized stimulus generators must control illumination down to the lux and timing down to the microsecond. Furthermore, researchers must implement strict exclusion criteria; trials in which an animal exhibits excessive motor agitation, groom sequences, or spontaneous startle responses during the silent trace interval must be flagged and isolated to prevent mechanical artifacts from contaminating the underlying associative learning metrics.

11.2 Impact of Biological Typology and Arousal States

A central pillar of Pavlovian theory that has found profound validation in modern behavioral genetics is the concept of *nervous system typology* (temperament). Pavlov recognized that the capacity to master trace conditioning is not uniform across all subjects of a given species; it is governed by intrinsic, biologically determined properties of the subject’s central nervous system. Pavlov categorized his experimental canines along three physiological axes: the *strength* of the nervous processes (the capacity of the cortex to endure intense excitatory or inhibitory loads without succumbing to transmarginal inhibition), the *balance* (equilibrium between excitation and inhibition), and the *mobility* (the speed with which the nervous system can switch between excitatory and inhibitory states).

These typological variations dictate an individual’s tolerance for the trace interval:

• Strong, Balanced, and Mobile Types (Sanguine): Rapidly acquired trace conditioned reflexes, exhibited exquisite temporal accuracy in their salivary latencies, and navigated prolonged trace intervals without psychological strain.
• Strong, Unbalanced Types (Choleric): Exhibited powerful, rapid excitatory learning but struggled severely with the inhibition of delay, frequently discharging premature saliva and developing severe agitation when trace intervals were stretched.
• Weak Types (Melancholic): Characterized by low cortical endurance, these subjects were completely incapable of tolerating prolonged trace conditioning; their cortical analyzers rapidly succumbed to “protective” transmarginal inhibition (zapredel’noe tormozhenie), sending the animal into sudden sleep or stupor whenever a trace interval exceeded a few seconds.

In modern contexts, these Pavlovian typologies correspond directly to genetic and epigenetic variations in prefrontal dopamine clearance (e.g., catechol-O-methyltransferase [COMT] polymorphisms) and serotonin transporter dynamics, which govern working memory capacity and anxiety-induced cognitive vulnerability.

11.3 Methodological Pitfalls in Contemporary Replications

Contemporary researchers attempting to implement or replicate trace conditioning paradigms frequently encounter significant methodological pitfalls that can compromise empirical validity. Chief among these is the failure to adequately control for *pseudoconditioning* and *sensitization*. Pseudoconditioning occurs when the presentation of a powerful, emotionally or biologically potent unconditioned stimulus (such as a painful shock or a highly rewarding food delivery) non-specifically elevates the general arousal state of the organism, causing it to respond to any subsequent neutral sensory stimulus even though no genuine associative connection has been formed between them.

To decisively rule out pseudoconditioning and sensitization artifacts in trace setups, modern protocols must employ rigorous *explicitly unpaired control groups* or *discriminative conditioning procedures* ($CS^+$ versus $CS^-$). In an explicitly unpaired control design, the subject receives the exact same number of CS presentations and US deliveries as the experimental group, but they are presented entirely at random or separated by long, variable intervals such that the CS never reliably predicts the US. If an animal in the experimental group exhibits significantly higher response rates during the trace interval than the explicitly unpaired control group, the responding can be definitively attributed to true associative trace learning rather than generalized arousal or sensitization.

Another prevalent pitfall is the failure to distinguish between *context-conditioned responses* and true *cue-specific trace responses*. In setups where the inter-trial interval is insufficiently long or highly predictable, the physical apparatus itself acquires profound excitatory properties. Researchers often mistakenly record baseline contextual motor activity or tonic salivary secretion as an anticipatory trace response to the extinguished CS. The utilization of high-resolution digital cameras, infrared photobeam arrays, and computerized baseline subtraction algorithms is mandatory in contemporary translational models—such as rodent trace eyeblink conditioning or trace fear conditioning—to ensure that the recorded response is strictly driven by the internal memory trace of the CS.

12. Contemporary Legacy and Clinical Applications of Pavlov’s Trace Conditioning

12.1 Trace Conditioning as an Assay for Declarative Memory and Awareness

One of the most remarkable theoretical transformations in the modern legacy of Pavlov’s work was the discovery, spearheaded by Larry Squire, Christian Clark, and colleagues in the late 1990s, that trace conditioning can serve as an objective behavioral assay for *declarative memory* and *conscious awareness*. In a landmark series of human eyeblink conditioning experiments, researchers compared amnesic patients suffering from bilateral medial temporal lobe damage (similar to the famous patient H.M.) with healthy neurotypical control subjects across both delay and trace paradigms.

The findings established a profound cognitive dichotomy: amnesic patients with profound declarative memory deficits could learn standard delay eyeblink conditioning at a normal rate, despite having no conscious recollection of having ever experienced the testing apparatus. However, when switched to a trace conditioning protocol, the amnesic patients failed entirely to acquire the conditioned response. Even more strikingly, in healthy human subjects subjected to trace eyeblink conditioning while concurrently performing an attention-demanding secondary distraction task, only those participants who developed explicit, conscious *declarative awareness* of the CS-US temporal contingency (as assessed by post-experimental questionnaires) succeeded in acquiring the trace conditioned reflex. Those who remained unaware of the contingency failed to acquire trace conditioning, despite mastering delay conditioning unconsciously.

This discovery established trace conditioning as an indispensable translational biomarker for assessing higher-order cognitive function. Today, trace eyeblink conditioning is utilized clinically as an ultra-sensitive, non-invasive behavioral biomarker for the early detection of Alzheimer’s disease and mild cognitive impairment (MCI). Because the transentorhinal and hippocampal circuits are the first to degenerate under the pathological accumulation of tau neurofibrillary tangles and amyloid-beta plaques, trace eyeblink conditioning deficits manifest years before macroscopic declarative memory failures appear on standardized neuropsychological tests. It likewise serves as an operational window into normal neurocognitive development and healthy aging.

12.2 Implications for Psychopathology: PTSD, Anxiety, and Addiction

The physiological mechanisms governing trace conditioning provide crucial insights into the etiology and maintenance of several prevalent neuropsychiatric disorders, most notably Post-Traumatic Stress Disorder (PTSD), severe generalized anxiety disorders, and substance use disorders. In PTSD, the normal temporal boundaries governing fear learning undergo pathological deregulation. When an individual experiences a catastrophic psychological trauma, the intense hyper-arousal and catecholaminergic storm cause an aberrant hyper-consolidation of the fear trace.

Patients suffering from PTSD exhibit pathological *temporal over-generalization* and defective inhibition of delay. Neutral sensory stimuli that occur long before a traumatic event become permanently encoded as imminent harbingers of catastrophe; the brain struggles to deploy the active, prefrontal internal inhibition necessary to restrict the fear response to appropriate contexts or specific temporal windows. Furthermore, trace fear memories are notoriously resistant to standard clinical extinction protocols. In trauma-exposed populations, the prefrontal-hippocampal inhibitory pathways that normally extinguish trace associations are hypoactive, while the basolateral amygdala remains chronically hyperactive, leading to relentless spontaneous recovery, intrusive flashbacks, and persistent autonomic hyper-arousal.

In the domain of *substance use disorders*, trace conditioning mechanisms drive the powerful, insidious phenomenon of cue-induced drug craving. Environmental paraphernalia, geographic locations, and subtle emotional states function as conditioned trace stimuli that precede drug consumption by substantial temporal intervals. Due to the profound dopaminergic surges induced by drugs of abuse such as cocaine, methamphetamine, and opioids, the associative eligibility window is pathologically extended. The central nervous system constructs powerful, durable trace associations between distant environmental cues and subsequent chemical rewards. Even after years of abstinence, re-exposure to a trace cue initiates an internal chain of persistent neural excitation within the ventral tegmental area, nucleus accumbens, and medial prefrontal cortex that drives uncontrollable relapse across the temporal gap. Contemporary translational therapies are currently utilizing memory reconsolidation blockade—administering beta-adrenergic antagonists such as propranolol or NMDA receptor modulators immediately following trace memory reactivation—to pharmacologically disrupt and weaken these maladaptive trace memories.

12.3 Computational Modeling and Artificial Neural Networks

Beyond its clinical and neurobiological ramifications, Ivan Pavlov’s trace conditioning framework has exerted an indelible influence upon the architecture of computer science, computational neuroscience, and artificial intelligence. The fundamental computational challenge posed by trace conditioning—how an agent can assign credit to a transient sensory event that occurred long before a biological reinforcement or reward signal is received—is known in machine learning as the *temporal credit assignment problem*.

To solve this problem within Reinforcement Learning (RL), computational theorists Richard Sutton and Andrew Barto developed the temporal-difference (TD) learning algorithm, which explicitly incorporates the concept of the *eligibility trace* (designated as $TD(lambda)$). In these computational architectures, when a state transition or stimulus occurs, it leaves an internal mathematical footprint—an eligibility trace—that decays exponentially over time according to a decay parameter $lambda$. When a terminal reward or punishment is subsequently encountered at a distant time step, the TD error updates not only the current state, but is distributed backward across all prior states that still maintain an active eligibility trace. This computational mechanism is the direct mathematical descendant of Pavlov’s theoretical cortical “vestige”:

$$e_t(s) = \gamma \lambda e_{t-1}(s) + \mathbf{1}(S_t = s)$$

In contemporary deep learning and artificial neural network (ANN) architectures, the demands of trace conditioning are mirrored in the development of Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks. Standard feedforward networks are incapable of solving trace conditioning tasks because they possess no internal temporal state or memory; once the input vanishes, the hidden layer activations instantaneously collapse. LSTMs and recurrent architectures solve this by introducing specialized recurrent feedback loops, constant error carrousels, and gating mechanisms that allow hidden units to sustain persistent, elevated activation vectors across thousands of temporal cycles in the complete absence of input drive.

More recently, computational neuroscientists have deployed continuous-time recurrent networks (CTRNNs) and spiking neural networks (SNNs) to simulate the precise biological microcircuits of the hippocampus and prefrontal cortex during trace conditioning. These bio-mimetic models reproduce the exact neurophysiological phenomena first glimpsed by Ivan Pavlov over a century ago: persistent attractor states that hold the stimulus vestige, inhibitory feedback interneuron networks that implement the inhibition of delay, and the sudden, burst-like disinhibition triggered by external perturbations. From early twentieth-century canine salivary fistulas to twenty-first-century artificial general intelligence, the trace conditioning paradigm endures as an essential cornerstone of predictive computational theory.

Conclusion

Ivan Petrovich Pavlov’s pioneering investigations into trace conditioning represent one of the most enduring milestones in the annals of science. By methodically engineering an acoustic sanctuary within his Tower of Silence and applying quantitative physiological rigor to the minute drops of canine saliva, Pavlov penetrated beyond the superficial boundaries of simple, mechanical reflexology. In trace conditioning, he unmasked an organism that was not merely reactive to its instantaneous environment, but profoundly predictive—an organism endowed with an internal neural architecture capable of preserving the past, calculating the passage of time, and projecting behavioral strategies into the future.

Pavlov’s brilliant theoretical formulations—the enduring cortical vestige, the dynamic interplay of irradiating and concentrating excitation, and the exhausting, energy-intensive force of internal inhibition of delay—laid the essential conceptual groundwork that modern neuroscience would validate and refine. We now know that the silent temporal void of the trace interval is populated by the rhythmic firing of hippocampal time cells, the sustained persistent activity of prefrontal working memory circuits, and the precisely calibrated plasticity of NMDA receptor cascades. These mechanisms solve the temporal credit assignment problem, bridging past causes and future consequences across the silent chasms of experience.

The contemporary legacy of trace conditioning continues to expand across multiple scientific frontiers. It provides cognitive neurology with a non-invasive diagnostic window into early neurodegenerative decline; it equips clinical psychiatry with mechanistic models to dissect and treat the debilitating pathologies of PTSD, anxiety, and addiction; and it supplies artificial intelligence with the foundational mathematics of eligibility traces that power reinforcement learning engines. More than a century after the master of Saint Petersburg first observed his canines salivating into glass tubes during moments of absolute silence, the trace conditioning experiment remains an immortal testament to the power of empirical science to illuminate the deepest, most complex mysteries of the mammalian brain.

References

• Babkin, B. P. (1949). Pavlov: A Biography. University of Chicago Press.
• Clark, R. E., & Squire, L. R. (1998). Classical conditioning and brain systems: The role of awareness. Science, 280(5360), 77–81. https://doi.org/10.1126/science.280.5360.77
• Kitamura, T., Macdonald, C. J., & Tonegawa, S. (2015). Entorhinal-hippocampal neuronal circuits bridge temporally discontinuous events. Learning & Memory, 22(9), 438–443. https://doi.org/10.1101/lm.038687.115
• Moyer, J. R., Deyo, R. A., & Disterhoft, J. F. (1990). Hippocampectomy disrupts trace eye-blink conditioning in rabbits. Behavioral Neuroscience, 104(2), 243–252. https://doi.org/10.1037//0735-7044.104.2.243
• Pavlov, I. P. (1904). Nobel Lecture: Physiology of Digestion. NobelPrize.org. https://www.nobelprize.org/prizes/medicine/1904/pavlov/lecture/
• Pavlov, I. P. (1927). Conditioned Reflexes: An Investigation of the Physiological Activity of the Cerebral Cortex (G. V. Anrep, Trans.). Oxford University Press.
• Pavlov, I. P. (1928). Lectures on Conditioned Reflexes: Twenty-five Years of Objective Study of the Higher Nervous Activity (Behaviour) of Animals (W. H. Gantt, Trans.). International Publishers.
• Sechenov, I. M. (1965). Reflexes of the Brain (S. Belsky, Trans.). MIT Press. (Original work published 1863).
• Siegel, J. J., Kalmbach, B., Chitwood, R. A., & Mauk, M. D. (2012). Persistent activity in a cortical-cerebellar circuit bridges the interval in trace eyelid conditioning. Journal of Neuroscience, 32(47), 16775–16787. https://doi.org/10.1523/JNEUROSCI.2057-12.2012
• Solomon, P. R., Vander Schaaf, E. R., Thompson, R. F., & Weisz, D. J. (1986). Altered neural activity in the hippocampus of the rabbit during trace conditioning of the nictitating membrane response. Behavioral Neuroscience, 100(5), 729–744. https://doi.org/10.1037/0735-7044.100.5.729
• Sutton, R. S., & Barto, A. G. (2018). Reinforcement Learning: An Introduction (2nd ed.). MIT Press.
• Thompson, R. F. (1986). The neurobiology of learning and memory. Science, 233(4767), 941–947. https://doi.org/10.1126/science.3738519
• Todes, D. P. (2014). Ivan Pavlov: A Russian Life in Science. Oxford University Press.
• Weiss, C., & Disterhoft, J. F. (2011). Trace eyeblink conditioning in the rabbit as a model for cognitive aging and Alzheimer’s disease. In The Organization of Memory (pp. 235–254). Cognitive Neuroscience Society.
• Woodruff-Pak, D. S., & Disterhoft, J. F. (2008). Where is the trace in trace conditioning? Trends in Neurosciences, 31(2), 105–112. https://doi.org/10.1016/j.tins.2007.11.006