Key Takeaways
- Recent research reveals that addiction risk stems more from brain architecture than from specific substances.
- A study analyzed data from over 2.2 million people, identifying genetic factors linked to substance use disorders.
- Behavioral disinhibition, a heritable trait, connects addiction risk to ADHD and impulse control issues.
- Polygenic scores may help predict addiction vulnerability, but they are not yet ready for clinical use.
- This study suggests that understanding the brain’s reward processing is crucial for treatment and prevention strategies.
For most of the past century, the question researchers asked about addiction was the wrong one. Not wrong in a trivial sense — wrong in a way that shaped treatment, policy, and the quiet shame carried by millions of people who couldn’t seem to stop. The question was: what is it about alcohol, or heroin, or cannabis, that makes people dependent? The better question, it turns out, is: what is it about certain people?
A study published today in Nature Mental Health, drawing on genetic data from more than 2.2 million individuals, has answered that question with unusual precision. Most of the heritable risk for developing a substance use disorder has little to do with the substance. It comes, instead, from a cluster of genes that shape how the brain processes rewards, regulates impulses and weighs up consequences — the same genes, broadly speaking, that also raise risk for ADHD, conduct problems in childhood, and a pattern of risk-taking behaviour that researchers call behavioral disinhibition. The drug, in a sense, is almost incidental.
That framing may sound like an overstatement. It isn’t, quite. The team at Rutgers Health, led by Holly Poore and senior author Danielle Dick, used a genomic method called Structural Equation Modelling to analyze four substance use disorders simultaneously — alcohol, tobacco, cannabis and opioids — alongside related traits including ADHD, risk tolerance and early initiation of substance use. Twin and family studies had already suggested this general picture; genetic variance shared between addiction disorders accounts for somewhere around 74 to 80 percent of the heritable component in alcohol use disorder, and 62 to 74 percent in others. What Poore’s team did was demonstrate it at the genomic level, identifying specific variants, specific pathways, with something that at least resembles mechanistic precision.
“Most of the genetic predisposition to substance use disorders isn’t about how bodies respond to drugs; it’s about how brains are wired,” said Dick, director of the Rutgers Addiction Research Center. “Specifically, risk is mostly related to genes that broadly impact how our brains process rewards and regulate behavior.”
The architecture they found works on two levels. First there is the broad externalizing factor: a set of genetic variants (708 genome-wide significant loci in the single-factor model) that influence reward processing, self-control and neural signalling in ways that cut across many outcomes simultaneously. These are variants mapped to genes involved in postsynaptic protein localization, vesicle-mediated transport, and synaptic signalling — the low-level infrastructure of how neurons communicate. Sitting on top of that, and largely distinct from it, are substance-specific variants. For alcohol, these include familiar genes in the alcohol dehydrogenase family (ADH1B, ADH4, ADH5 and ADH6), which govern how efficiently the body breaks down ethanol. For tobacco, the signal clusters in the nicotinic acetylcholine receptor genes, CHRNA3 and CHRNA5 among others. These aren’t about wanting; they’re about metabolizing, about how the body tolerates or processes the specific chemical in question.
Genetic risk for addiction is mostly rooted in how the brain handles rewards and self-regulation, not how the body processes any particular substance. The same variants that raise addiction risk also increase risk for ADHD, conduct problems and impulsive behaviour more broadly. Substance-specific genes, which do exist, appear to work mostly by affecting how efficiently the body metabolises or responds to a given chemical, rather than by driving the core pattern of problematic use.
Behavioral disinhibition describes a heritable tendency toward difficulty regulating impulses, higher risk tolerance and reward-seeking behaviour. It’s the shared underlying factor linking addiction disorders with childhood conduct problems, adult antisocial behaviour and ADHD. Research suggests this underlying trait is itself about 80 percent heritable, more so than any individual disorder it influences. Understanding it as a central feature of addiction risk, rather than a side effect, changes how prevention and treatment might be targeted.
That’s the translational aim, though the science isn’t there yet clinically. Currently, a broad externalizing polygenic score explains roughly 3 to 7 percent of variance in substance use disorder diagnoses in independent samples — meaningful at the population level, but not predictive enough for individual clinical use. The researchers suggest a two-level approach, one broad score for general vulnerability and separate substance-specific scores, could eventually provide more actionable information about who might benefit from preventive intervention.
No. Heritability in the range of 40 to 60 percent means environment still accounts for a substantial portion of risk. The researchers are explicit that genetic risk scores describe vulnerability, not destiny. They also point out that many environmental factors known to reduce addiction risk — regular exercise, healthy weight management, interventions that reduce conditions like sleep apnoea — may plausibly influence some of the same biological pathways.
Because most of the genetic signal is shared. When researchers study alcohol use disorder in isolation, they’re treating the genetic variance it shares with opioid use disorder, cannabis use disorder and ADHD as statistical noise. By modelling all of these together, the shared signal becomes much easier to detect — in this study, the joint approach identified 187 genetic variants that had not been found in any previous study. The analogy is rough but workable: looking for a common thread in many fabrics at once is easier than squinting at each fabric separately and hoping the thread shows up.
“Those same genes show up across many outcomes — things like ADHD, conduct problems and other risky behaviors — and then layered on top of that are genes that are more specific to each substance,” said Dick. “What this paper does, for the first time, is tease apart those pathways at the genomic level.”
The methodological innovation here is perhaps what matters most. Genome-wide association studies for individual psychiatric conditions have had a difficult decade — expensive, sample-hungry, and returning effect sizes so small they strain credulity. Drug-by-drug searches for addiction genes faced the same problem in miniature, compounded by the fact that substance use disorders almost never occur in isolation. Someone dependent on alcohol is substantially more likely than average to also have cannabis problems, opioid problems, conduct disorder — all of which the traditional disease-by-disease approach treats as confounders to be controlled away rather than signal to be harvested. By modelling the shared architecture directly, Poore’s team found 187 novel genetic variants not identified in earlier externalizing or addiction studies. More than half of all loci they identified had never previously been linked to any substance use trait.
That matters for treatment as well as for discovery. The team mapped their gene findings to a drug interaction database and identified 118 druggable targets associated with the broad externalizing pathway, connecting to more than 1,000 FDA-approved drug-gene interactions. Some of the treatments already in use for substance use disorders — naltrexone, methadone, varenicline, acamprosate — showed up in this analysis, suggesting these drugs may be working partly through mechanisms that cut across addictions rather than through purely drug-specific biology. There’s some intriguing territory here around repurposing existing medications, though that work is still at the discovery phase.
The polygenic score findings are perhaps the most directly clinically relevant part. A score built from the broad externalizing factor predicted risk for multiple substance use disorders simultaneously, explaining between 3 and 7 percent of variance in disorder diagnoses in a large independent cohort. Substance-specific scores were also informative, but in a different and complementary way — better at distinguishing who is specifically vulnerable to problems with alcohol versus nicotine, for instance. “A broader metric can tell us who is generally more vulnerable to addiction and other externalizing problems, while more specific scores can help us understand who is at higher risk for problems with different substances,” said Dick. “That doesn’t mean genes determine someone’s destiny, but they can help us identify who might benefit most from targeted prevention or earlier intervention.”
There are real limits to what the study can claim. The sample is restricted to individuals of European ancestry, a consequence of where the large genetic databases currently exist rather than a design choice. Generalising findings across populations remains a significant outstanding problem in psychiatric genomics. The residual genetic effects for opioid use disorder — the variance left over after removing what opioids share with the externalizing factor — were also underpowered; no genome-wide significant hits emerged there. And polygenic scores currently explain a small fraction of variance, useful for research but not yet ready for clinical implementation.
Still. The picture that emerges from 2.2 million people’s worth of data is striking in its coherence. What makes someone vulnerable to addiction is, mostly, a brain architecture that processes rewards intensely, regulates impulses incompletely, and gravitates toward immediate sensation over deferred consequence. Whether alcohol or cannabis or opioids becomes the medium for that vulnerability depends on genetics of metabolism, receptor sensitivity, and social exposure — a second layer, almost a modifier, applied to an underlying neural predisposition. The question was never really which drug. The question was always how the brain was built.
Source: Poore et al., Nature Mental Health (2026). DOI: 10.1038/s44220-026-00608-6
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.