Higher Nervous Activity

Abstract

Higher nervous activity is a mental process: Ivan Pavlov's term for the acquired, conditioned activity of the cerebral cortex through which an organism learns the signals that predict biologically important events. Studying salivary reflexes in dogs, Pavlov showed that a neutral stimulus paired with food comes to elicit the response on its own, and that this learning reflects the balance of two cortical processes, excitation and inhibition. Twentieth-century research recast the conditioned reflex as the learning of relationships between events rather than mere co-occurrence, and modern neuroscience located its mechanisms in amygdala fear circuits, cerebellar reflex pathways, dopamine prediction-error signals, and the molecular machinery of synaptic plasticity. This article traces higher nervous activity from Pavlovian physiology to contemporary associative-learning theory, and the three demonstrations let the reader grow an association, weigh contingency against contiguity, and read a generalization gradient.

Keywords: higher nervous activity, classical conditioning, excitation and inhibition, associative learning, prediction error

At the turn of the twentieth century the physiologist Ivan Pavlov, already a Nobel laureate for his work on digestion, turned the salivary reflex of the dog into an instrument for studying learning. He observed that dogs began to salivate not only to food in the mouth but to the signals that reliably preceded it, and he named the study of such acquired reflexes the physiology of higher nervous activity (Pavlov, 2010). The programme was deliberately physiological: Pavlov held that the laws of learning were the laws of the cerebral cortex, and he sought them in the measurable rise and fall of a reflex rather than in introspection (Windholz, 1997). A century later the conditioned reflex remains one of the most heavily studied phenomena in the behavioural sciences, reinterpreted as a window onto how animals and people detect the predictive structure of their world (Domjan, 2005).

What Higher Nervous Activity Is

Pavlov divided the reflexes of the organism into two classes. The unconditioned reflex is inborn and fixed: food in the mouth elicits salivation without any learning, the unconditioned stimulus reliably producing its unconditioned response. The conditioned reflex is acquired: a stimulus that is at first neutral, a bell or a metronome, comes through repeated pairing with food to elicit salivation on its own, now a conditioned stimulus evoking a conditioned response (Pavlov, 2010). Higher nervous activity is the physiology of this second class, the lawful process by which the cortex forms, strengthens, and dissolves such acquired connections. Pavlov treated the term as coextensive with the learned adjustment of behaviour to a changing environment, in explicit contrast to the lower nervous activity of fixed spinal and brainstem reflexes (Windholz, 1997).

The programme was radical for its insistence on objectivity. Where earlier psychology asked what an animal felt or perceived, Pavlov measured only the reflex: how many drops of saliva, after how many pairings, with what delay. This methodological austerity made conditioning a quantitative science and supplied a shared vocabulary, acquisition, extinction, generalization, and discrimination, that the whole of learning theory still uses (Domjan, 2005). It is closely allied with, and often used synonymously with, classical conditioning, the procedure Pavlov invented; higher nervous activity is the broader physiological doctrine within which that procedure sits, and it stands beside operant conditioning and simpler forms of learning such as habituation as one of the basic ways experience reshapes behaviour. Table 1 sets out the core phenomena the rest of the article develops.

Excitation, Inhibition, and Cortical Dynamics

Pavlov explained the whole repertoire of conditioning with two opposed cortical processes. Excitation is the process that builds and expresses a conditioned reflex; inhibition is the process that restrains it. Acquisition is the growth of excitation; extinction, for Pavlov, was not the erasure of the connection but the active accumulation of inhibition over it, a claim later evidence has strongly vindicated (Pavlov, 2010). He marshalled a family of observations under these two headings. A conditioned response spreads from the trained signal to similar ones, an irradiation of excitation he called generalization; training an animal to respond to one tone but not to a neighbouring tone sharpens the reflex through inhibition, yielding discrimination. When the two tones are brought too close together, so that the cortex is required to excite and inhibit almost the same input, the animal's behaviour can break down into agitation or torpor, a phenomenon Pavlov named experimental neurosis and took as evidence that the balance of the two processes has real limits (Windholz, 1997).

From individual differences in the strength, balance, and mobility of excitation and inhibition Pavlov built a typology of nervous systems, mapping his four types onto the classical temperaments. That typology is now of mainly historical interest, but the underlying insight, that conditioning is the joint product of an excitatory and an inhibitory process rather than a single associative bond, survived every later reformulation. Contemporary accounts of extinction, discrimination, and the return of extinguished responses are still, at bottom, theories about how learned excitation and learned inhibition compete for control of behaviour (Bouton, 2004).

The Two Signal Systems

Late in his life Pavlov extended higher nervous activity from animals to the distinctively human case with the doctrine of the two signal systems. The first signal system is the one he had studied throughout: direct sensory signals, the sight of food or the sound of a bell, that stand in for events in the world. The second signal system, which Pavlov held to be unique to humans, is language, words that function as signals of signals, one further remove from the physical stimulus, allowing the conditioning laws to operate over abstractions rather than only over sights and sounds (Windholz, 1997). The proposal was an attempt to bring speech and thought, the province of cognition, inside the same physiological framework as salivation. Its specifics did not survive the cognitive revolution, which showed that language is not simply a second layer of conditioned reflexes, but the ambition marks the reach Pavlov intended for higher nervous activity: a single account of learning from the simplest reflex to human verbal thought.

The Soviet Lineage and the Fate of the Doctrine

Higher nervous activity did not begin with Pavlov. Its intellectual root is Ivan Sechenov, the founder of Russian objective physiology, whose 1863 monograph Reflexes of the Brain argued that even the most elaborate acts of mind are ultimately reflexes of the brain, to be explained by physiology rather than introspection (Sechenov, 1863/1965). Pavlov took this materialist programme as his starting point and credited Sechenov as its father; the conditioned reflex was his way of making Sechenov's claim experimental. The very phrase higher nervous activity is a calque from the Russian vysshaya nervnaya deyatelnost, and it carries that Russian lineage into every language that later borrowed it (Todes, 2014).

That lineage became a matter of state. Pavlov, though no friend of the Bolsheviks, was cultivated by the Soviet government as a scientific figurehead, and after his death in 1936 his doctrine hardened into orthodoxy (Todes, 2014). At the 1950 Joint Session of the USSR Academy of Sciences and the Academy of Medical Sciences — the Pavlovian session — higher nervous activity was declared the official and mandatory framework for Soviet physiology and psychology, rival approaches were denounced, and researchers who had strayed from Pavlovian teaching were pressed to recant (Krementsov, 1997). The episode is a case study in the political capture of a science: a genuine research programme was frozen into dogma and defended by administrative power rather than by evidence.

This history explains the term's peculiar standing today. In the West higher nervous activity is now largely of historical interest, its phenomena absorbed into the vocabulary of associative learning and behavioural neuroscience and seldom named as such (Domjan, 2005). Yet the phrase persists — in the names of Russian institutes and journals, and as a formal MeSH heading under which this literature is still indexed — a reminder that a scientific label can outlive the doctrine that once enforced it. What survived the ossification was not Pavlov's cortical theory but his method: the study of learning as a lawful, measurable adjustment of behaviour, which the next section takes up in its modern form.

From Reflex to Contingency

Pavlov's own theory assumed that contiguity, the near-simultaneous occurrence of conditioned and unconditioned stimuli, was what forged the connection. The decisive twentieth-century correction showed that contiguity is not enough. In a classic experiment, Robert Rescorla held the pairing of a signal and a shock constant while varying how often the shock also occurred without the signal. When shock was equally likely with or without the signal, no conditioning developed despite abundant pairings; conditioning appeared only when the signal raised the probability of shock above its background rate (Rescorla, 1968). What the animal learns, then, is not that two events touched in time but that one predicts the other, a contingency between them. This reframing is why the mature view treats Pavlovian conditioning as the detection of causal and predictive structure in the environment rather than as a mechanical stamping-in of reflexes (Rescorla, 1988).

The insight was formalised in the most influential model in the field. The Rescorla-Wagner model holds that an association gains strength only in proportion to how surprising the outcome is: learning on each trial is driven by the discrepancy between the outcome that occurred and the outcome the currently present signals already predicted. As a signal comes to predict the outcome well, that discrepancy shrinks and learning slows, producing the decelerating acquisition curve Pavlov had charted. The same principle explains blocking, in which a cue already predicting an outcome prevents a redundant new cue from being learned: with no surprise left, there is nothing to drive the new association (Domjan, 2005). Later work added that animals encode not only whether but when the outcome will arrive, building a temporal map in which the informativeness of a cue, how much it reduces uncertainty about the timing of the outcome, governs how readily it is learned (Balsam & Gallistel, 2009).

The Neuroscience of Conditioning

Pavlov's wager that the laws of conditioning are the laws of the brain has been redeemed in detail, though the cortex he emphasised shares the work with older structures. The best-mapped case is conditioned fear. When a tone is paired with a mild shock, convergent evidence locates the plasticity that binds them in the amygdala: sensory pathways carrying the tone and the shock meet on the same neurons there, and strengthening of the tone pathway is what allows the tone alone to trigger the battery of defensive responses (LeDoux, 2000). Lesion, stimulation, and recording studies converge on the same conclusion, that the amygdala is necessary for acquiring and expressing conditioned fear, making it one of the clearest instances of a memory traced to a defined circuit (Maren, 2001). The circuitry has been dissected to the level of specific nuclei and their inputs, distinguishing the pathways that acquire fear from those that store and express it (Kim & Jung, 2006).

Different forms of conditioning recruit different circuits, exactly as one would expect if the brain solves each reflex with dedicated machinery. Discrete, precisely timed motor reflexes such as the conditioned eyeblink depend on the cerebellum, where decades of work have localised the essential plasticity for a conditioned movement (Thompson, 2005). Across these systems the common currency is synaptic change, and the molecular biology of that change, from short-lasting modification of existing proteins to the gene expression that makes a memory permanent, has been worked out most fully in simple conditioning preparations, linking the behaviour Pavlov measured to the switching on of specific genes at the synapse (Kandel, 2001). The broad lesson of this literature is that associative learning is not one thing in the brain but a family of circuit-specific solutions sharing a plasticity mechanism (Fanselow & Poulos, 2005).

Prediction and Reward

The prediction-error idea at the heart of the Rescorla-Wagner model turned out to have a physical signature. Recording from dopamine neurons while animals learned to expect a reward, Wolfram Schultz found that the neurons do not simply report reward: they fire to a reward that is unexpected, fall silent when an expected reward is omitted, and, as a signal comes to predict the reward, transfer their response from the reward to the signal (Schultz et al., 1997). This is precisely the reward prediction error the learning theory required, computed in the brain and broadcast to the structures that update predictions, a striking convergence of a behavioural model derived from salivating dogs with the firing of identified neurons. The dopamine system links Pavlovian learning to reward processing and to the wider machinery of value-guided choice.

The account has been refined rather than overturned. Detailed study of the circuitry shows how the prediction error is assembled from separate streams carrying information about actual and expected reward, and how it is distributed to targets that use it for different purposes (Watabe-Uchida et al., 2017). Current debate concerns just how much the dopamine signal encodes: whether it is a simple scalar teaching signal or carries richer information about the structure of the task and the state the animal believes itself to be in, a question that connects Pavlovian conditioning to the brain's broader effort at predictive coding (Gershman & Uchida, 2019).

Extinction and Clinical Reach

If conditioning is how fears and appetites are learned, extinction is how they are treated, and understanding its true nature has direct clinical stakes. Extinction is not unlearning. An extinguished response returns when the animal is tested outside the extinction context (renewal), after the passage of time (spontaneous recovery), or after a single reminder of the outcome (reinstatement), which shows that the original association was intact all along and that extinction laid down a new, context-dependent inhibitory memory on top of it (Bouton, 2004). The behavioural and neurobiological mechanisms of this inhibitory learning, spanning the amygdala, hippocampus, and prefrontal cortex, are now understood in enough detail to explain why extinction is fragile and context-bound (Bouton et al., 2021). Because stress reopens the balance between learned fear and its inhibition, often at the expense of extinction, the same circuitry is central to how stress disorders develop and relapse (Maren & Holmes, 2016).

The clinical reach of higher nervous activity is broad. Exposure therapy for anxiety and phobia is extinction learning applied deliberately, and the two-system framework distinguishes the conditioned defensive responses that exposure targets from the conscious feeling of fear, which need not track them, clarifying why physiological and subjective fear can be treated somewhat separately (LeDoux & Pine, 2016). The principle reaches even into pharmacology: a drug's cues become conditioned stimuli that elicit anticipatory, opponent bodily responses, so that tolerance is in part a learned, situation-specific conditioned reflex, and an addict is at heightened overdose risk in a novel setting where those protective conditioned responses are absent (Siegel, 1975). The reach extends historically as well: the first demonstration that a human fear could be conditioned to a neutral object, in the ethically indefensible study of an infant known as Little Albert, established that Pavlov's laws applied to human emotion and set the agenda for a century of clinical learning research (Watson & Rayner, 1920).

Worked Example

The Rescorla-Wagner model makes the growth of an association quantitative, and the acquisition demo above computes exactly this. The associative strength V of a conditioned stimulus is updated on every reinforced trial by the rule V → V + αβ(λ − V), where λ is the maximum strength the outcome can support (set to 1 for a reinforced trial), αβ is a combined learning rate fixed by the salience of the stimulus and the outcome, and (λ − V) is the prediction error, the surprise remaining on that trial. Take a learning rate αβ = 0.3, a fully reinforced outcome λ = 1, and a stimulus that starts with no strength, V = 0.

On trial 1 the prediction error is (1 − 0) = 1, so V rises by 0.3 × 1 = 0.3, to V = 0.300. On trial 2 the error has shrunk to (1 − 0.3) = 0.7, so V rises by 0.3 × 0.7 = 0.21, to V = 0.510. Trial 3 adds 0.3 × 0.49 = 0.147, giving V = 0.657; trial 4 adds 0.3 × 0.343 = 0.1029, giving V = 0.760; trial 5 adds 0.3 × 0.2401 = 0.072, giving V = 0.832. The increments, 0.300, 0.210, 0.147, 0.103, 0.072, shrink on every trial even though the learning rate never changes, because the quantity being learned about, the prediction error, is itself consumed by the learning. That decelerating approach to a ceiling is the acquisition curve Pavlov measured drop by drop, and it is why the largest learning happens early, when the outcome is most surprising. Switching the demo to extinction sets λ = 0, so the same rule now drives V downward by 0.3 × (0 − V) each trial, modelling the growth of inhibition rather than its removal.

Discussion

Higher nervous activity began as a physiologist's bet that learning could be studied with the exactness of a reflex, and the bet paid off in a way Pavlov could not have foreseen. The vocabulary he coined survives intact; the phenomena he catalogued are reproduced daily; and his central intuition, that conditioning reflects a competition between excitation and inhibition in the brain, prefigured the modern picture of learned fear and its regulation. What changed is the content poured into that frame. Contiguity gave way to contingency and prediction error; the single cortical association gave way to a family of circuit-specific mechanisms; and the reward prediction error moved from a term in an equation to a measurable dopamine signal. The through-line is that an organism is built to detect the predictive structure of its world, and the conditioned reflex is the simplest visible expression of that capacity.

The frontier now concerns how this ancient learning system meshes with the rest of cognition. Pavlovian conditioning and the episodic memory system have long been studied as separate literatures, yet a single conditioning episode is also an event that can be remembered, and recent work argues the two systems interact more than either tradition assumed (Dunsmoor & Kroes, 2019). Understanding that interaction, how a simple learned association is embedded within, and sometimes overridden by, richer representations of specific past events, is where the century-old study of higher nervous activity rejoins the broader science of memory and cognition it helped to found.

Common Misconceptions

Classical conditioning is just an animal salivating, a curiosity with little to do with real cognition.

The salivation is only the readout. What the procedure measures is how an organism learns the predictive and causal relationships among events, which is a foundational cognitive achievement, not a reflexive quirk (Rescorla, 1988). The belief persists because the textbook image of the dog is memorable while the underlying claim about learning is abstract.

Extinction erases the conditioned response, so once a fear is extinguished it is gone.

Extinction adds new inhibitory learning rather than deleting the original association, which is why the response returns with a change of context, the passage of time, or a reminder (Bouton, 2004). The misconception matters clinically: it predicts that a single successful exposure session should cure a phobia, and the relapse literature shows it does not.

Conditioning happens automatically whenever two events occur close together in time.

Mere contiguity is insufficient. When a signal and an outcome are paired often but the outcome is just as likely without the signal, no conditioning occurs; the signal must actually raise the probability of the outcome (Rescorla, 1968). Contiguity feels sufficient because in everyday life predictive signals are also contiguous ones, so the two are rarely pulled apart outside the laboratory.

Glossary

Acquisition.

The phase in which a conditioned response is established as the conditioned stimulus is repeatedly paired with the unconditioned stimulus.

Blocking.

The failure of a redundant cue to be learned when it is added to a cue that already predicts the outcome, evidence that surprise drives learning.

Conditioned reflex.

An acquired response in which a previously neutral stimulus comes to elicit a reaction through its pairing with an unconditioned stimulus.

Conditioned stimulus.

An originally neutral stimulus that, after pairing, evokes the conditioned response; the signal in a Pavlovian arrangement.

Contingency.

The degree to which the conditioned stimulus raises the probability of the unconditioned stimulus above its background rate; the true driver of conditioning.

Discrimination.

The narrowing of a conditioned response to a reinforced signal while a similar unreinforced signal is inhibited.

Excitation.

In Pavlov's system, the cortical process that builds and expresses a conditioned reflex; the positive pole of learning.

Extinction.

The decline of a conditioned response when the conditioned stimulus is presented without the unconditioned stimulus; new inhibitory learning, not erasure.

First signal system.

Pavlov's term for conditioning to direct sensory signals, the sights and sounds that stand in for events in the world.

Generalization.

The tendency of stimuli similar to the conditioned stimulus to evoke the conditioned response, in a gradient that falls off with dissimilarity.

Higher nervous activity.

Pavlov's name for the conditioned, learned activity of the cerebral cortex, the physiology of acquired reflexes and their adjustment to the environment.

Inhibition.

The cortical process that restrains a conditioned reflex; the negative pole that underlies extinction, discrimination, and the return of extinguished responses.

Pavlovian session.

The 1950 Joint Session of the USSR Academy of Sciences and Academy of Medical Sciences that made higher nervous activity the official, mandatory doctrine of Soviet physiology and psychology and suppressed rival approaches.

Prediction error.

The discrepancy between the outcome that occurs and the outcome predicted by the present signals; the surprise term that drives associative learning.

Reinforcement.

In the Pavlovian sense, the delivery of the unconditioned stimulus that strengthens the conditioned reflex.

Renewal.

The return of an extinguished response when it is tested in a context different from the one in which extinction took place.

Second signal system.

Pavlov's proposed human-specific system of language, in which words act as signals of signals, one remove from the physical stimulus.

Spontaneous recovery.

The partial reappearance of an extinguished conditioned response after a rest interval, showing the original association was retained.

Unconditioned stimulus.

A stimulus that elicits a response without any learning, such as food producing salivation; the outcome a conditioned stimulus comes to predict.

Key Researchers

Mark E. Bouton (University of Vermont). Showed through renewal, spontaneous recovery, and reinstatement that extinction is context-dependent new learning rather than erasure of the original association. Faculty Page - Google Scholar - ORCID

Michael S. Fanselow (University of California, Los Angeles). Behavioral neuroscientist whose work on fear conditioning helped establish that mammalian associative learning is a family of circuit-specific mechanisms sharing a plasticity currency. Faculty Page - Google Scholar - ORCID

Eric R. Kandel (Columbia University). Nobel laureate whose analysis of simple conditioning traced the molecular and cellular basis of memory storage from synaptic modification to gene expression. Faculty Page - Nobel Prize - ORCID

Joseph E. LeDoux (New York University). Mapped the amygdala circuit that acquires and expresses conditioned fear and reframed it within a two-system account separating defensive circuits from conscious feeling. Faculty Page - Google Scholar

Stephen Maren (University of Illinois Urbana-Champaign). Dissected the amygdala, hippocampal, and prefrontal circuits that encode and regulate Pavlovian fear and its extinction. Faculty Page - Google Scholar - ORCID

Ivan Petrovich Pavlov (1849-1936). Russian physiologist and Nobel laureate who founded the study of higher nervous activity and the conditioned reflex, giving learning its first rigorous experimental physiology. Wikipedia - Nobel Prize

Robert A. Rescorla (1940-2020). Demonstrated that conditioning depends on the informativeness of a signal rather than mere contiguity, and co-developed the Rescorla-Wagner model of associative learning. Wikipedia - Biographical Memoir

Frequently Asked Questions

What is higher nervous activity?
It is Ivan Pavlov's term for the conditioned, learned activity of the cerebral cortex, the physiological study of how neutral signals come to control reflexes through experience, in contrast to the fixed, inborn reflexes of the lower nervous system (Pavlov, 2010).

How is higher nervous activity related to classical conditioning?
Classical conditioning is the experimental procedure Pavlov invented; higher nervous activity is the broader physiological doctrine, the theory of cortical excitation and inhibition, within which that procedure and its phenomena are explained (Windholz, 1997).

What is the difference between a conditioned and an unconditioned reflex?
An unconditioned reflex is inborn and needs no learning, as when food elicits salivation, whereas a conditioned reflex is acquired when a neutral stimulus is repeatedly paired with the unconditioned stimulus until it elicits the response on its own (Pavlov, 2010).

Why is contingency more important than contiguity in conditioning?
Because a signal paired often with an outcome produces no conditioning if the outcome is equally likely without it; what the organism learns is that the signal predicts the outcome, raising its probability, not merely that the two occurred together (Rescorla, 1968).

Does extinction erase a conditioned response?
No. Extinction lays down new inhibitory learning over the original association, which stays intact, so the response can return with a change of context, the passage of time, or a reminder of the outcome (Bouton, 2004).

Where in the brain is conditioned fear learned?
Convergent evidence locates the plasticity for conditioned fear in the amygdala, where the pathways carrying the signal and the aversive outcome meet, though other reflexes such as the conditioned eyeblink depend on different circuits (LeDoux, 2000).

How do dopamine neurons relate to Pavlovian learning?
They fire to unexpected rewards, fall silent when an expected reward is omitted, and shift their response to predictive signals, behaving as the reward prediction error that associative-learning theory requires (Schultz et al., 1997).

Why does higher nervous activity matter clinically?
Exposure therapy for anxiety and phobia is extinction learning applied deliberately, and conditioned responses even shape drug tolerance and overdose risk, so the principles of higher nervous activity underlie a large part of clinical learning research (LeDoux & Pine, 2016).

References