Understanding Addiction: Neuroscience and Policy Insights

In the 1970s, Bruce Alexander ran an experiment that challenged everything the scientific establishment thought it knew about addiction. The standard rat addiction experiment had a single rat in an isolated cage with two water bottles - one plain water, one laced with heroin or cocaine.

The rat inevitably chose the drug, escalated its consumption, and died. The experiment was repeated thousands of times and interpreted as evidence that certain drugs are so powerfully addictive that any animal that encounters them will become dependent.

Alexander built something different. He constructed "Rat Park" - a large enclosure with tunnels, platforms, and balls, populated with 16 to 20 rats of both sexes who could socialize, play, and mate. He gave the Rat Park residents the same two bottles.

They overwhelmingly chose plain water. Even rats previously force-fed morphine for 57 consecutive days - rats that were physically dependent on opioids - voluntarily reduced their consumption when moved to Rat Park.^[7]

The experiment was controversial and difficult to publish. It complicated the clean narrative that drugs are addictive by pharmacological necessity. But Alexander's findings aligned with a body of evidence that had been quietly accumulating in human populations.

In Vietnam, approximately 20 percent of US soldiers used heroin. Epidemiologist Lee Robins, studying the veterans' outcomes in a landmark study published in the Archives of General Psychiatry in 1974 and followed up in Drug and Alcohol Dependence in 1993, found that roughly 95 percent of those who used heroin in Vietnam stopped without treatment upon returning home, often without severe withdrawal.^[4]

The context had changed: they were no longer isolated in a combat zone with constant access to cheap drugs and no other means of coping with terror and boredom. They were home, with families, jobs, and social lives.

The question was never simply whether drugs are addictive. The question Alexander was asking - and what a generation of neuroscience and epidemiology has since clarified - is: addictive to whom, under what conditions, and what do we do about it? This is the article about those questions.

"The opposite of addiction is not sobriety. The opposite of addiction is connection." - Johann Hari, Chasing the Scream (2015), summarizing Alexander's argument

Addiction Treatment Options: Evidence Summary

Key Definitions

Addiction - A chronic, relapsing condition characterized by compulsive drug seeking and use despite negative consequences, loss of control over use, and continued use despite a desire to stop. Now formally termed Substance Use Disorder (SUD) in DSM-5.

Substance use disorder (SUD) - The DSM-5 diagnostic category replacing the previous distinction between "abuse" and "dependence." Defined by 11 criteria; presence of 2-3 is mild, 4-5 moderate, 6 or more severe.

Tolerance - A physiological adaptation in which repeated drug exposure reduces the drug's effect, requiring larger doses to achieve the same result. A normal biological process distinct from addiction.

Dependence (physical vs psychological) - Physical dependence is a state in which the body has adapted to a drug's presence such that its removal causes withdrawal. Psychological dependence refers to the emotional and motivational craving for a drug's effects. Both can occur without addiction.

Withdrawal - The physical and psychological symptoms that occur when a physically dependent person reduces or stops drug use. Severity varies dramatically by drug class: opioid withdrawal is profoundly uncomfortable; alcohol withdrawal can be life-threatening.

Craving - The intense desire for a drug, often triggered by cues previously associated with drug use. A core feature of addiction and a major target of treatment.

Mesolimbic dopamine system - The neural circuit running from the ventral tegmental area (VTA) to the nucleus accumbens that underlies reward prediction and motivation. The pathway that addictive drugs primarily hijack.

Nucleus accumbens - A structure in the basal forebrain that receives dopamine input and plays a central role in motivation, reward, and addiction. Often misleadingly called the "pleasure center."

Prefrontal cortex - The brain region involved in executive function, decision-making, and inhibitory control. Consistently shows reduced activity in people with addiction, contributing to impaired control over drug-seeking.

Allostasis (Koob) - The process by which the brain adapts to chronic drug exposure by resetting its motivational baseline (hedonic setpoint), making drug use necessary just to maintain normal function rather than to achieve pleasure.

Incentive salience (Berridge) - The motivational property of wanting that the dopamine system assigns to stimuli. Distinguished from hedonic liking. Addictive drugs powerfully amplify incentive salience.

Opponent-process theory - The theory (Solomon & Corbit, 1974) that pleasurable stimuli generate an opposing negative state; with repetition, the negative state strengthens and the pleasure weakens, requiring the stimulus just to reach baseline.

Harm reduction - Public health approaches that aim to reduce the negative consequences of drug use without requiring abstinence, including needle exchanges, naloxone distribution, and supervised consumption sites.

Neuroplasticity - The brain's ability to change its structure and function in response to experience. Chronic drug exposure produces long-lasting neuroplastic changes in reward, motivation, and cognitive control circuits.

What Addiction Is

The formal diagnostic definition of addiction has been revised significantly over the past four decades.

The DSM-5 (2013) replaced the previous categories of substance abuse and substance dependence with the unified category of Substance Use Disorder, defined by 11 criteria across domains of impaired control (using more than intended, unsuccessful efforts to cut down, much time spent obtaining and using, craving), social impairment (failure to fulfill major role obligations, continued use despite interpersonal problems),

risky use (use in physically hazardous situations, continued use despite physical or psychological harm), and pharmacological criteria (tolerance, withdrawal).

A diagnosis of SUD requires 2 or more criteria within a 12-month period.

According to the Substance Abuse and Mental Health Services Administration's 2022 National Survey on Drug Use and Health, approximately 46 million Americans - about 17 percent of the US population aged 12 and older - met criteria for at least one substance use disorder in the past year.^[6]

This includes approximately 29 million with alcohol use disorder, 8 million with cannabis use disorder, and nearly 6 million with opioid use disorder. These numbers are almost certainly underestimates, because self-report surveys undercount illicit drug use and stigma suppresses disclosure.

The critical distinction between use, misuse, and disorder matters enormously for policy and treatment. The large majority of people who use alcohol, cannabis, opioids, and even cocaine do not develop a use disorder. Most people who try heroin do not become addicted.

The Vietnam data are the most dramatic illustration: 20 percent of soldiers used heroin, but by far the majority stopped without treatment when they returned to a social environment that provided alternative sources of meaning and connection.

The transition from use to disorder is shaped by drug pharmacology, route of administration, frequency and quantity of use, age of first use, genetic vulnerability, psychiatric comorbidity, and - as Alexander's work suggests - the social and environmental context in which use occurs.

The Dopamine Hypothesis

The popular understanding of dopamine as a "pleasure chemical" is a significant simplification that obscures how the brain's reward system actually works and why addiction is so difficult to treat. Two lines of research have fundamentally reshaped the picture.

Wolfram Schultz, a neuroscientist at Cambridge University, published a landmark study in Science in 1997 (doi: 10.1126/science.275.5306.1593) recording the activity of individual dopamine neurons in monkeys performing a conditioning task.

Schultz found that dopamine neurons did not fire in response to rewards themselves - they fired in response to unexpected rewards and, once an association was learned, in response to the cue predicting the reward rather than the reward itself.

When a cue reliably predicted a reward and the reward was delivered as expected, dopamine neurons fired at the cue and showed no particular response at the reward. When an expected reward failed to materialize, dopamine activity dipped below baseline.

Dopamine, Schultz argued, encodes prediction error - the difference between expected and actual outcomes - rather than pleasure.^[3] It is a learning signal, not a pleasure signal.

Kent Berridge and Terry Robinson's research, published in Brain Research Reviews in 1995 (doi: 10.1016/0165-0173(94)00023-I) and developed extensively since, drew a further distinction between "wanting" and "liking." Wanting is the motivational drive to seek and obtain a reward - incentive salience, assigned to stimuli by the dopamine system in the nucleus accumbens.

Liking is the hedonic pleasure of the reward itself - the subjective enjoyment - which is mediated by opioid and endocannabinoid circuits in partially distinct brain regions.

Berridge and Robinson showed that these systems can dissociate: animals whose dopamine systems were selectively depleted still showed pleasure responses to sweet tastes (liking intact) but made no effort to seek food even when hungry (wanting abolished).

They would eat if food was placed in their mouths but would not approach food or work for it.

In addiction, Berridge and Robinson argue, the dopamine system assigns massive incentive salience to drug-related cues - far more than to natural rewards like food or social contact - producing overwhelming wanting without proportional liking.^[1]

This explains what many addicted people describe: the compulsive craving for drugs they no longer find pleasurable, the inability to stop despite knowing the consequences, the way drug-associated environments and cues trigger irresistible urges years into recovery.

The wanting circuit has been sensitized; the liking response has often diminished. The brain has been reprogrammed to assign extreme motivational value to something that no longer reliably delivers the pleasure it once did.

Addictive drugs share the common property of increasing dopamine in the nucleus accumbens, though through different mechanisms. Cocaine and amphetamine block the reuptake of dopamine, flooding the synapse.

Opioids activate mu receptors on inhibitory interneurons in the ventral tegmental area, removing a brake on dopamine neurons and indirectly increasing nucleus accumbens dopamine. Alcohol, cannabis, and nicotine each have distinct mechanisms that converge on the same dopaminergic pathway.

Tolerance and Allostasis

Why does addiction progress from voluntary use to compulsive use that continues despite catastrophic consequences? The allostatic model developed by George Koob and Michel Le Moal, published in Science in 1997 (doi: 10.1126/science.278.5335.52), provides a framework that goes beyond simple tolerance.

The opponent-process theory of Solomon and Corbit (1974) proposed that pleasurable stimuli generate an opposing negative state - a hedonic opponent process - that increases in strength with repeated exposure. The A-process (the pleasure) habituates; the B-process (the opposing negative state) sensitizes.

A novice heroin user experiences intense euphoria and mild subsequent discomfort. A long-term user experiences little euphoria and intense subsequent discomfort (withdrawal). They use, increasingly, not to feel good but to not feel terrible.

Koob and Le Moal extended this to the neurobiological level with the concept of allostasis - a chronic shift in the brain's hedonic setpoint away from its natural baseline.^[2] Normal hedonic function is regulated around a setpoint: the brain maintains a roughly stable baseline of mood and motivation.

Repeated drug use disrupts this regulation. The brain adapts not by restoring the original setpoint but by resetting it at a lower level, requiring drug use just to achieve what used to be the drug-free baseline.

This allostatic shift is driven by downregulation of dopamine receptors in the reward circuit (reducing reward sensitivity) and recruitment of stress-related systems - particularly corticotropin-releasing factor (CRF) and dynorphin - that mediate negative emotional states in withdrawal.

The clinical consequence is that late-stage addiction is characterized not primarily by the pursuit of pleasure but by the relief of negative affect - anxiety, dysphoria, irritability, and anhedonia - that constitutes the withdrawal state. People drink not to feel good but to not feel bad.

This shift has important treatment implications: treating the negative affect of withdrawal and the underlying allostatic dysregulation is as important as blocking the rewarding effects of drugs.

Neuroplasticity underlies these changes. Chronic drug exposure produces lasting alterations in gene expression, dendritic morphology, and synaptic strength in the circuits underlying reward and control.

The prefrontal cortex - the brain region involved in executive function, impulse control, and decision-making - consistently shows reduced activity and gray matter volume in people with severe substance use disorders.

This structural change contributes to the impaired control that defines addiction: the wanting system has been strengthened; the control system has been weakened.

Risk Factors

Not everyone who uses addictive substances becomes addicted. The transition from use to disorder is influenced by a complex interaction of genetic, developmental, psychiatric, and social factors.

Genetic factors account for roughly 40 to 60 percent of the variance in risk for alcohol use disorder, and heritabilities for other substance use disorders are in a similar range.

Twin and adoption studies consistently document this heritable component, and genome-wide association studies have begun identifying specific genetic variants that contribute to risk.

However, no single gene determines addiction risk; polygenic scores explain only a modest fraction of variance, and gene-environment interaction is pervasive.

Age of first use is among the strongest predictors of disorder development. The National Institute on Drug Abuse reports that people who begin drinking before age 15 are four times more likely to develop alcohol use disorder than those who begin at 21.

The adolescent brain - in which the limbic reward system matures earlier than the prefrontal cortex's regulatory systems - appears particularly vulnerable to the lasting neuroplastic changes produced by early drug exposure.

Adverse childhood experiences (ACEs) - including abuse, neglect, and household dysfunction - are strongly associated with substance use disorders.

The CDC-Kaiser Permanente ACE Study, conducted with over 17,000 participants in the 1990s, found that individuals with four or more ACEs were 7 times more likely to report alcohol dependence than those with none.

Trauma activates stress systems that the allostatic model predicts would make drug-induced relief of negative affect especially reinforcing.

Psychiatric comorbidity is the rule rather than the exception in clinical populations: approximately 50 to 60 percent of people with substance use disorders have co-occurring psychiatric disorders, most commonly depression, anxiety disorders, post-traumatic stress disorder, and ADHD.^[9]

The causal relationship is bidirectional and complex: some psychiatric conditions increase substance use (self-medication), some substance use causes psychiatric symptoms, and some shared vulnerability produces both. Effective treatment of addiction almost always requires addressing co-occurring psychiatric conditions.

The Opioid Crisis

The opioid crisis is a case study in what happens when pharmaceutical industry incentives, regulatory failure, and medical culture interact catastrophically. Its origins are traceable to decisions made in the 1980s and 1990s that have cost more than 300,000 American lives.

Through the 1980s, pain was undertreated in American medicine. Opioids were used cautiously, primarily for acute pain and cancer pain, because of well-founded concerns about addiction. A pain advocacy movement, partly organized around genuine patient suffering, argued that addiction risk from therapeutic opioid use had been exaggerated.

The evidence for this claim centered on a brief 1980 letter to the New England Journal of Medicine by Hershel Jick and Jane Porter, reporting that only 4 out of almost 12,000 hospitalized patients who received opioids showed signs of addiction.

This letter - 101 words, no methodology, no controls, describing hospitalized inpatients, not the long-term outpatient chronic pain patients it would later be cited to support - was cited 608 times in subsequent medical literature, frequently as evidence that the addiction risk from opioid therapy was "less than 1 percent."

A 2017 study by Leung and colleagues in the New England Journal of Medicine (doi: 10.1056/NEJMc1716235) traced how this misleading citation had propagated through the literature.^[5]

Purdue Pharma launched OxyContin in 1996 with an aggressive marketing campaign targeting primary care physicians who were not pain specialists.

Sales representatives were paid bonuses for increasing prescriptions and coached to emphasize the extended-release formulation's lower addiction risk - a claim unsupported by clinical evidence.

Purdue paid physicians large fees as speakers at "educational" events and used proprietary prescription data to identify high-prescribing physicians for targeted marketing.

Sales grew from $48 million in 1996 to over $1.1 billion by 2000. In 2007, Purdue and three executives pled guilty to charges of misbranding OxyContin and paid $634 million in fines - a sum that represented a small fraction of their revenues.

Prescription opioid overdose deaths rose steadily through the 2000s. When prescribing restrictions tightened in the early 2010s, many people with opioid use disorder shifted to heroin, which was cheaper and more available.

Then illicitly manufactured fentanyl - 50 to 100 times more potent than morphine - entered the drug supply from Mexico, displacing heroin because of its extreme potency and low cost.

Fentanyl dramatically increased overdose risk: a dose miscalculated by milligrams rather than grams could be lethal. By 2020, fentanyl and fentanyl analogs were involved in the majority of opioid overdose deaths.

Treatment Evidence

The most effective treatments for opioid use disorder are medications - buprenorphine and methadone - yet they are used in a minority of cases due to stigma, regulatory barriers, and the cultural dominance of abstinence-based treatment models.

Buprenorphine is a partial opioid agonist that activates mu-opioid receptors with high affinity but lower intrinsic activity than full agonists like morphine.

Its partial agonism means it produces limited euphoria but effectively eliminates cravings and prevents withdrawal; its high receptor affinity means that if a patient using buprenorphine takes additional opioids, the additional opioids have minimal effect (blocking the reinforcing properties of other opioids).

Methadone is a full opioid agonist with a long half-life that prevents withdrawal and, at adequate doses, blocks the effects of other opioids. Both have been approved by the FDA and studied in randomized controlled trials and large observational studies.

Both consistently reduce opioid use, criminal activity, HIV transmission, and overdose mortality - by 50 percent or more in most studies.

Despite this evidence, fewer than 20 percent of Americans with opioid use disorder receive medication-assisted treatment (MAT).

Until the Mainstreaming Addiction Treatment Act of 2023, physicians needed a special DEA waiver (the "X-waiver") to prescribe buprenorphine - a regulatory barrier that had no parallel for other psychiatric medications and reflected the moral framework that addiction should require suffering to treat, not pharmacological management.

Methadone for opioid use disorder can still be dispensed only through specially licensed clinics, requiring daily attendance that is incompatible with work and family obligations for many patients.

Naltrexone, an opioid antagonist, is FDA-approved for both opioid use disorder (as monthly injectable Vivitrol) and alcohol use disorder. It works by blocking opioid receptors, preventing the reinforcing effects of opioids and reducing the desire to drink in alcohol-dependent patients.

For alcohol use disorder, acamprosate has FDA approval and a reasonable evidence base for reducing craving and maintaining abstinence.

The 12-step model - Alcoholics Anonymous, Narcotics Anonymous, and their variants - is the most widely used recovery framework in the United States.

It provides social support, a recovery community, a structured approach to addressing psychological and spiritual dimensions of addiction, and free access that pharmacological treatments cannot match.

The evidence base is modest but genuine for many people. The Project MATCH study (1997), a large multi-site randomized trial, compared 12-step facilitation therapy, cognitive-behavioral therapy, and motivational interviewing in alcohol use disorder and found no significant differences in outcome between the three - all three helped, and the advantage of any one approach over the others was small.

Cochrane reviews of AA have been complicated by methodological issues, but a 2020 review by Kelly and colleagues concluded that AA participation was associated with meaningful rates of abstinence.^[8]

Policy Approaches

The policy response to addiction in the United States has been shaped more by moral frameworks and racial politics than by evidence. The "war on drugs," declared by Richard Nixon in 1971 and escalated through the Reagan era, framed drug use primarily as criminal behavior to be deterred through punishment.

The US imprisoned more people per capita than any other nation in the world, with drug offenses a major driver. The evidence that criminalization reduces drug use is weak; the evidence that it produces racial disparities, family destabilization, and barriers to employment that worsen outcomes is strong.

Drug courts, which began in Miami in 1989 and spread nationally, represent a partial shift: they divert people with drug offenses into treatment rather than incarceration, with court supervision. Evaluations suggest they reduce recidivism and drug use compared to standard criminal processing, though the comparison is complicated.

The model retains the criminal justice framework - participation is coerced by the alternative of incarceration - and tends to exclude people with the most severe disorders.

Portugal's 2001 decriminalization provides the most studied alternative policy experiment. Portugal decriminalized the personal possession of all drugs (not just cannabis) and shifted resources from criminal enforcement to treatment and harm reduction.

Drug trafficking remained criminal. Evaluations found subsequent declines in drug-related HIV infections, drug-related deaths, and drug-related incarceration, while drug use rates did not substantially increase.

The Portuguese model is not direct legalization - possession can still result in an administrative hearing and treatment requirement - but it represents a fundamental shift from moral to public health framing.

Harm reduction as a policy philosophy has accumulated strong evidence in European countries that adopted it earlier. Needle exchange programs reduce HIV and hepatitis C transmission; the evidence is consistent across settings.

Supervised consumption sites - where people can use pre-obtained drugs under medical supervision - have operated in Canada, Switzerland, the Netherlands, and other countries for decades.

Evaluations of Insite, the Vancouver site operating since 2003, have found reduced overdose deaths in the surrounding area, increased uptake of detoxification services, and no increase in drug use in the neighborhood.

The US, as of this writing, has authorized pilot programs for supervised consumption sites but faces ongoing legal and political barriers.

Naloxone (Narcan) - the opioid overdose-reversing medication that can be administered by bystanders - is perhaps the clearest harm reduction success. Widely distributed in its intranasal form, it is estimated to have reversed hundreds of thousands of overdoses and is now available over the counter.

Its distribution to people who use opioids and their families represents harm reduction in its most direct form: accepting that people will use drugs and ensuring they don't die from it.

Sources & Further Reading

1. Berridge, Kent C., and Terry E. Robinson. "The Mind of an Addicted Brain: Neural Sensitization of Wanting Versus Liking." Current Directions in Psychological Science, 4(3), 71-76, 1995. DOI: 10.1016/0165-0173(94)00023-I

2. Koob, George F., and Michel Le Moal. "Drug Abuse: Hedonic Homeostatic Dysregulation." Science, 278(5335), 52-58, 1997. DOI: 10.1126/science.278.5335.52

3. Schultz, Wolfram. "A Neural Substrate of Prediction and Reward." Science, 275(5306), 1593-1599, 1997. DOI: 10.1126/science.275.5306.1593

4. Robins, Lee N. "Vietnam Veterans' Rapid Recovery from Heroin Addiction: A Fluke or Normal Expectation?" Addiction, 88(8), 1041-1054, 1993.

5. Leung, P.T.M., et al. "A 1980 Letter on the Risk of Opioid Addiction." New England Journal of Medicine, 376, 2194-2195, 2017. DOI: 10.1056/NEJMc1716235

6. Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2022 National Survey on Drug Use and Health. 2023. View source

7. Alexander, Bruce K., et al. "Effect of Early and Later Colony Housing on Oral Ingestion of Morphine in Rats." Pharmacology Biochemistry and Behavior, 15(4), 571-576, 1981.

8. Kelly, John F., et al. "Alcoholics Anonymous and 12-Step Facilitation Treatments for Alcohol Use Disorder: A Distillation of a 2020 Cochrane Review for Clinicians and Policy Makers." Alcohol and Alcoholism, 55(6), 641-651, 2020. DOI: 10.1093/alcalc/agaa050

9. Kessler, Ronald C., et al. "The Epidemiology of Co-Occurring Addictive and Mental Disorders." American Journal of Orthopsychiatry, 66(1), 17-31, 1996.