Drug-Seeking Behavior

Abstract

Drug-seeking behavior is a type of behavior: the organized pursuit and procurement of a drug, which cognitive psychology studies as the behavioral expression of the learning and motivational systems that addiction recruits. This article sets out that account: how a drug cue acquires control over action through dopamine-signaled reward prediction, how repeated use sensitizes wanting independently of liking, how goal-directed seeking hardens into habit and compulsion, and how drug cues capture attention and trigger craving through largely automatic processes. The same reward-learning machinery that normally guides adaptive choice is, on this view, hijacked, so that seeking persists against intention and cost. Three demonstrations make reward-prediction transfer, the wanting–liking dissociation, and cue reactivity concrete.

Keywords: drug-seeking behavior, incentive salience, reward prediction error, cue reactivity

Drug-seeking behavior is the heading the National Library of Medicine uses for the activities directed toward obtaining a drug, and it is the pivot on which the psychology of addiction turns, because what makes addiction a disorder is not the drug’s immediate effect but the behavior that comes to surround it: the search, the procurement, the relapse after abstinence. Cognitive psychology approaches this behavior not as a moral failure or a simple matter of pleasure-seeking, but as the product of ordinary learning and motivational systems operating on an extraordinary reinforcer. A drug that releases dopamine engages the same machinery that evolved to make animals pursue food and mates, and the central question of the field is how that machinery, acting on a drug, produces a pattern of seeking that outlasts pleasure, resists intention, and is triggered by cues the person may not consciously notice.

What Drug-Seeking Behavior Is

Within the Medical Subject Headings, Drug-Seeking Behavior sits beneath the broad parent heading behavior as the class of activities an organism directs toward obtaining a drug. The placement is apt, because the defining feature of the phenomenon is behavioral: addiction is diagnosed and studied less through what a person feels than through what a person does—the effort expended, the alternatives forgone, the reinstatement of seeking after a period of abstinence. Drug-seeking behavior is therefore the observable expression of the motivational state that the drug has produced, and it is the dependent variable that nearly every theory of addiction ultimately tries to explain.

The cognitive account treats that behavior as the output of the brain’s systems for learning about and pursuing reward—systems that long predate drugs and that normally serve adaptive ends. On this view addiction is a pathology of motivation and choice: the neural circuits that assign value to outcomes, that translate value into the drive to act, and that arbitrate between competing goals are progressively distorted by a drug that acts directly on them (Kalivas & Volkow, 2005). The sections that follow trace that distortion from the bottom up: first the reward-prediction signal that lets a cue acquire control over action, then the sensitization of the motivation to seek, then the shift from deliberate action to compulsive habit, and finally the cue-triggered, largely automatic cognition through which the behavior is set in motion.

Reward Learning and the Dopamine Prediction Error

The foundation of the modern account is a discovery about dopamine. Phasic dopamine neurons do not simply signal pleasure or reward; they signal a reward prediction error, the difference between the reward received and the reward expected (Schultz, Dayan, & Montague, 1997). An unexpected reward drives a burst of firing; a fully predicted reward drives none; and a predicted reward that fails to arrive drives a dip below baseline. Critically, as a cue comes reliably to predict a reward, the dopamine burst migrates backward in time from the reward itself to the earliest predictor of it. The signal that began as a response to the drug becomes a response to the sight of the paraphernalia, the street, the hour of day.

This backward transfer is the formal basis for a drug cue acquiring motivational power, and it connects dopamine to the broader psychology of learning and motivation: dopamine is not the pleasure of the drug but the teaching signal that stamps in the association between a predictor and the value it predicts, and the drive to pursue that predicted value (Wise, 2004). Because most addictive drugs raise dopamine pharmacologically, they generate a prediction error that does not habituate the way a natural reward’s does: the drug keeps teaching the brain that it was better than expected, overvaluing its own cues trial after trial. The demonstration below shows this transfer, with the dopamine response shifting from the drug to its predictor as pairings accumulate.

Where the dopamine signal fires

Early in learning an unexpected reward drives a large prediction error. As a cue comes to predict that reward, the error transfers backward from the reward to the cue. Drag the learning rate and the trial number to watch the reward-time signal shrink while the cue-time signal grows. The model is value = 1 − (1 − α)^t.

Incentive-Sensitization: Wanting versus Liking

If reward-prediction learning explains how a cue gains control over behavior, the incentive-sensitization theory explains why the resulting motivation can become so pathological. Its central claim is a dissociation between two psychologically and neurally distinct components of reward that are ordinarily fused: liking, the hedonic pleasure a reward produces, and wanting, the motivation to obtain it, technically called incentive salience (Robinson & Berridge, 1993). In normal reward they rise and fall together, but they are produced by separable brain systems, and a drug can drive them apart.

The theory holds that repeated drug use sensitizes the dopamine-dependent wanting system, making it hyper-reactive to the drug and its cues, while the liking system does not sensitize and often shows tolerance. The consequence is the defining paradox of addiction: the person comes to want the drug more and more intensely even as they enjoy it less and less, so that craving can be dissociated from pleasure entirely (Berridge & Robinson, 2016). Incentive salience is also what makes cues themselves attractive and pursued, not merely informative. In animals this appears as sign-tracking: some individuals approach and work for a reward-predicting cue as though it were the reward itself, and this tendency, which is dopamine-dependent, marks vulnerability to cue-triggered drug seeking (Flagel et al., 2011). The demonstration contrasts the sensitizing trajectory of wanting with the flat or declining course of liking across exposures.

Wanting pulls away from liking

Incentive-sensitization theory separates two systems the everyday word “reward” conflates. Repeated drug exposure sensitizes wanting (incentive salience) while liking (hedonic pleasure) stays flat or erodes. Drag the exposure history, or switch sensitization off, to see the gap between craving and pleasure open or close.

From Goal-Directed Action to Habit and Compulsion

Drug seeking does not stay the same kind of behavior over the course of an addiction; its psychological control shifts. Early seeking is largely goal-directed action: it is performed because the drug is valued and the action is known to produce it, and it is therefore sensitive to the drug’s current value. With extensive repetition, control migrates to habit: a stimulus–response disposition that is triggered by the situation and run off automatically, relatively insensitive to whether the outcome is currently wanted. An influential account proposes that addiction involves a progression from actions to habits to compulsion, underpinned by a shift in control from ventral to dorsal striatal circuits (Everitt & Robbins, 2005).

Cues, Craving, and Automatic Cognition

The moment-to-moment trigger of drug seeking is usually a cue: a person, place, object, or internal state that has been associated with the drug. Cue reactivity is the complex of responses—physiological, attentional, and subjective—that such a cue evokes, and craving is its most salient subjective component. An early and durable cognitive model proposed that drug urges and much drug-use behavior are controlled by automatic processes: the skilled, habitual action schemas that run off with little cognitive effort, with conscious craving arising mainly when those automatic schemas are blocked and nonautomatic, effortful processing is engaged (Tiffany, 1990). The model’s enduring lesson is that craving and drug use are only loosely coupled, because the behavior can proceed automatically without strong conscious urges.

One of the best-studied automatic phenomena is attentional bias: drug cues capture and hold the attention of drug users, who are faster to orient to drug-related stimuli and slower to disengage from them. Reviews establish that this bias develops alongside the incentive salience the cues acquire, correlates with self-reported craving, and can itself motivate seeking, forming a feedback loop between attention and desire (Field & Cox, 2008). More broadly, the field has come to see such effects as instances of implicit cognition—automatic memory associations and appraisals that operate outside deliberate awareness and predict behavior over and above what people consciously intend (Stacy & Wiers, 2010). The demonstration below shows how the craving and attentional pull of a cue grow with the strength of its learned association and its salience.

Cue reactivity and the pull of attention

A drug cue grabs attention in proportion to its learned salience. Here subjective craving and the attentional-bias cost (extra milliseconds to disengage from the cue) both scale with cue salience × attentional bias. Raise either and watch the reaction toward the cue intensify.

Worked Example

Consider how the dopamine reward prediction error transfers from a drug to a cue that predicts it, using the standard learning rule behind the first demonstration. Let the cached value of the cue be V, updated after each pairing by V ← V + α(R − V), with learning rate α = 0.3 and the drug’s reinforcing magnitude set to R = 1. The dopamine response at drug delivery is the prediction error R − V at that moment, and the response at the cue is the value V the cue has come to hold. Starting from V = 0, the value after t pairings is V_t = 1 − 0.7^t, and the prediction error at drug delivery on trial t is R − V_t−1 = 0.7^t−1. On the first pairing the cue has no value, so the drug elicits the full error of 1.0 while the cue elicits nothing. By the fifth pairing the cue has acquired value V₅ = 1 − 0.7⁵ = 0.832, and the error at drug delivery has fallen to 0.7⁴ = 0.240. In the limit the fully predicted drug elicits no dopamine response at all, while the cue elicits the full response of 1.0—the signal has moved entirely from the reward to its predictor. If the expected drug is then unexpectedly omitted, the prediction error is 0 − 0.832 = −0.832, the dip that encodes “worse than expected.” The lesson is quantitative: the capacity of a drug cue to trigger seeking is the value the prediction-error signal has transferred onto it, and that value grows as one minus a decaying exponential in the number of pairings. This is why a cue encountered after long abstinence can still command the behavior—the value stamped onto it does not simply vanish.

Discussion

Drug-seeking behavior is a clarifying case for cognitive psychology because it shows what happens when ordinary, adaptive mechanisms of learning and motivation are driven by a reinforcer they did not evolve to handle. Each piece of the account is a piece of normal psychology turned pathological: the dopamine prediction error that lets any cue predict any reward; the separable systems for wanting and liking that let motivation and pleasure diverge; the automation of repeated action into habit; the capture of attention by motivationally significant stimuli. Addiction, on this reading, is not a foreign process but a familiar one amplified and distorted—which is precisely why it is so instructive and so hard to treat.

Table 2

Cognitive mechanisms in drug-seeking behavior and what each explains.

Note. Each mechanism is drawn from a section above; together they portray drug-seeking as the joint product of reward learning, sensitized motivation, automated control, and cue-triggered automatic cognition.

The cognitive account has limits and tensions. The habit-and-compulsion view has been challenged by evidence that much drug choice remains goal-directed and responsive to costs and alternatives, pointing instead to the role of negative reinforcement—drug taking to relieve a negative emotional state. And no single mechanism is sufficient: reward learning, sensitized wanting, habit, and automatic cue reactivity are complementary descriptions of a behavior that recruits all of them at once and in different proportions across individuals and stages.

Current Directions

The contemporary field is organized partly by a debate over how automatic drug seeking really is. One prominent line integrates the mechanisms above into the brain-disease model, in which repeated drug exposure produces lasting changes in reward, motivation, and executive-control circuits that make seeking progressively less voluntary (Volkow, Koob, & McLellan, 2016). A complementary neurocircuitry framework organizes addiction into a three-stage cycle—binge/intoxication, withdrawal/negative affect, and preoccupation/anticipation—each tied to distinct circuits and to a shift from positive toward negative reinforcement as the disorder progresses (Koob & Volkow, 2016).

That framing has drawn a pointed challenge. A translational critique argues that the evidence for genuinely habitual or compulsive, value-insensitive drug use in humans is weaker than often assumed, and that much seeking is better explained as goal-directed drug choice under negative affect—a rational, if self-defeating, selection of the drug to relieve distress when better alternatives are unavailable (Hogarth, 2020). The disagreement matters for treatment: an automatic-habit view favors disrupting cues and retraining responses, while a goal-directed view favors changing the value of competing options and the affective states that drive choice. The field is also increasingly applying implicit-cognition methods—attentional-bias and approach-avoidance retraining—to test whether modifying automatic processes directly can change the behavior (Stacy & Wiers, 2010).

Cognitive Implications

Read as cognition, drug-seeking behavior illustrates three themes that reach well beyond addiction. First, it is among the clearest demonstrations that motivation and pleasure are separable: the wanting that drives behavior is not the liking that rewards it, and a system can come to pursue what it no longer enjoys, a dissociation with implications for every theory that equates wanting with expected pleasure (Robinson & Berridge, 1993; Berridge & Robinson, 2016). Second, it shows how a single teaching signal—the dopamine prediction error—can let an arbitrary cue inherit the motivational significance of a reward, a principle that governs normal learning as much as pathological seeking (Schultz, Dayan, & Montague, 1997). Third, it is a sustained case study in the limits of deliberate control: behavior can be driven by overlearned habit and automatic cue reactivity against a person’s explicit intentions, so that knowing better is not the same as doing better (Tiffany, 1990; Everitt & Robbins, 2016). In each, drug seeking is an unusually stark window onto the machinery of ordinary motivated behavior.

Common Misconceptions

People seek drugs because the drugs feel so good.

Seeking is driven by wanting, not liking. The incentive-sensitization account holds that repeated use sensitizes the motivation to obtain the drug while hedonic pleasure stays flat or declines, so craving and seeking can intensify even as enjoyment fades (Robinson & Berridge, 1993; Berridge & Robinson, 2016).

Dopamine is the brain’s pleasure chemical.

Phasic dopamine signals a reward prediction error—a teaching signal about outcomes being better or worse than expected—and the motivation to pursue predicted value, not pleasure itself. Its role in drug seeking is to stamp value onto cues and energize their pursuit (Schultz, Dayan, & Montague, 1997; Wise, 2004).

Craving is what makes addicts use, so without craving there is no use.

Craving and drug use are only loosely coupled. Much seeking runs off automatically through overlearned action schemas, and strong conscious craving arises mainly when those automatic routines are blocked—so use can occur with little felt urge (Tiffany, 1990).

Glossary

Attentional bias.

The tendency of drug cues to capture and hold a user’s attention, producing faster orienting toward and slower disengagement from drug-related stimuli; it tracks craving and can itself motivate seeking.

Compulsion.

Drug seeking that persists despite explicit knowledge of adverse consequences and feels beyond deliberate control; the proposed endpoint of the progression from action to habit.

Craving.

The subjective desire for a drug, the most salient conscious component of cue reactivity; loosely coupled to actual drug use, which can proceed automatically without it.

Cue reactivity.

The complex of physiological, attentional, and subjective responses evoked by a drug-associated cue; a principal trigger of craving and relapse.

Dopamine.

A neuromodulator whose phasic signaling encodes reward prediction error and drives the pursuit of predicted value; the principal chemical target through which most addictive drugs act on reward learning.

Drug-seeking behavior.

The organized pursuit and procurement of a drug; the behavioral expression of addiction and the dependent variable most theories aim to explain.

Goal-directed action.

Behavior performed because its outcome is valued and known to follow from the action; sensitive to the current value of the outcome, and contrasted with habit.

Habit.

A stimulus–response disposition triggered by a situation and run off automatically, relatively insensitive to whether the outcome is currently wanted; the intermediate stage between action and compulsion.

Incentive salience.

The motivational property, “wanting,” attributed to a reward and its cues that makes them attractive and pursued; mediated by dopamine systems and dissociable from hedonic pleasure.

Incentive-sensitization theory.

The account that repeated drug use sensitizes the wanting system while leaving liking unchanged, so that motivation to seek the drug grows independently of the pleasure it provides.

Liking.

The hedonic pleasure a reward produces, a component of reward distinct from wanting and supported by separate brain systems; it does not sensitize with repeated drug use.

Negative reinforcement.

The strengthening of drug taking by the removal of an aversive state, such as withdrawal or negative affect; increasingly emphasized as the disorder progresses from the early positive-reinforcement stage.

Reward prediction error.

The difference between reward received and reward expected, signaled by phasic dopamine; the teaching signal that transfers value from a reward to the cues that predict it.

Sign-tracking.

A dopamine-dependent tendency to approach and work for a reward-predicting cue as though it were the reward itself; a marker of vulnerability to cue-triggered drug seeking.

Wanting.

The motivation to obtain a reward, the everyday name for incentive salience; the component of reward that sensitizes in addiction and drives seeking even as liking fades.

Key Researchers

Kent C. Berridge

. Professor of Psychology at the University of Michigan; with Terry Robinson he originated the incentive-sensitization theory, drawing the distinction between the brain systems for wanting and liking that reframed addiction as sensitized motivation rather than pursued pleasure. Wikipedia · Google Scholar · Faculty page

Barry J. Everitt

. Emeritus Professor of Behavioural Neuroscience at the University of Cambridge; his work established the account of addiction as a progression from goal-directed actions to habits to compulsion, tied to a shift in control across striatal circuits. ORCID · Wikipedia · Google Scholar · Faculty page

Trevor W. Robbins

. Professor of Cognitive Neuroscience at the University of Cambridge; co-author of the actions-to-habits-to-compulsion framework, he linked the loss of top-down prefrontal control to the compulsive character of drug seeking. ORCID · Wikipedia · Google Scholar · Faculty page

Terry E. Robinson

. Emeritus Professor of Psychology and Neuroscience at the University of Michigan; co-originator with Kent Berridge of the incentive-sensitization theory, and a central figure in the study of sign-tracking and individual differences in cue-triggered seeking. ORCID · Wikipedia · Google Scholar · Faculty page

Nora D. Volkow

. Director of the National Institute on Drug Abuse; her neuroimaging research established the dysfunction of prefrontal control circuits in addiction and shaped the brain-disease model of drug-seeking as a disorder of motivation and choice. ORCID · Wikipedia · Faculty page

Frequently Asked Questions

What is drug-seeking behavior in cognitive psychology?

It is the organized pursuit and procurement of a drug, the sense in which the National Library of Medicine uses the term, studied as the behavioral output of the brain’s reward-learning and motivational systems acting on a drug. Its defining feature is behavioral persistence: seeking that continues against intention and cost and reinstates after abstinence (Kalivas & Volkow, 2005).

How does a drug cue gain the power to trigger seeking?

Through reward-prediction learning. Phasic dopamine signals the error between received and expected reward, and as a cue comes to predict the drug, that signal transfers backward from the drug to the cue, stamping value onto it and giving it the capacity to energize seeking (Schultz, Dayan, & Montague, 1997; Wise, 2004).

What is the difference between wanting and liking?

Liking is the hedonic pleasure a reward produces; wanting, or incentive salience, is the motivation to obtain it. They are supported by separable brain systems, and in addiction the wanting system sensitizes while liking does not, so a person can want a drug intensely while enjoying it less and less (Robinson & Berridge, 1993; Berridge & Robinson, 2016).

Why does drug seeking continue even when it is clearly harmful?

Because control over the behavior shifts with repetition from goal-directed action to automatic habit, and in addiction to compulsion, while prefrontal circuits that would inhibit and revalue the behavior are weakened. The habitual pull toward the drug meets a diminished brake (Everitt & Robbins, 2005; Goldstein & Volkow, 2011).

What is cue reactivity?

It is the set of physiological, attentional, and subjective responses that a drug-associated cue evokes, with craving as its most salient conscious part. Cue reactivity, including attentional bias toward drug stimuli, is a principal trigger of craving and relapse (Field & Cox, 2008).

Is craving necessary for drug use?

No. An influential cognitive model holds that much drug use runs off automatically through overlearned action schemas, with strong conscious craving arising mainly when those routines are blocked. Craving and use are therefore only loosely coupled (Tiffany, 1990).

Is addiction a brain disease or a choice?

The field is actively divided. The brain-disease model stresses lasting drug-induced changes that make seeking progressively less voluntary, while a translational critique argues that much seeking remains goal-directed drug choice under negative affect, a self-defeating but value-sensitive selection (Volkow, Koob, & McLellan, 2016; Hogarth, 2020).

What is implicit cognition in addiction?

It refers to automatic memory associations and appraisals that operate outside deliberate awareness and predict drug use over and above conscious intentions. Attentional biases and automatic approach tendencies are examples, and retraining them is being tested as a treatment (Stacy & Wiers, 2010).

References