Heroin Basically Vanished From the Street Supply. The Craving It Built Never Got the Memo.

“Opioids made me truly happy for the first time in my life,” a Filter contributor who writes under the pseudonym Alex Mauri wrote in 2019, describing the heroin use they’d decided, deliberately, to keep. Three years later, after quitting, they wrote a different essay that most coverage of “recovery” doesn’t make room for: “Quitting heroin, in short, has removed some inconveniences but not solved my problems.” Both of those sentences are true, from the same person, about the same drug. That contradiction — that stopping the substance doesn’t automatically stop what the substance was managing — is where the actual psychology of heroin lives, and it’s the part that gets flattened every time this story gets told as a before-and-after.

If you’ve ever been told that craving is just a willpower problem you haven’t tried hard enough to solve, here is what the last two years of research on heroin actually found: craving has a specific, measurable shape, it is not the same thing as withdrawal, and neither one predicts who relapses as cleanly as the people warning you about it seem to think.

Heroin itself has nearly vanished from the illicit supply that mattered most — fentanyl and its analogs now dominate — but the population of people whose brains and behavior were shaped by years of heroin use hasn’t vanished with it. Understanding what heroin actually did to craving, withdrawal, and the brain’s own reward circuitry isn’t a historical footnote. It’s the baseline every fentanyl-era treatment decision is still being made against.

Craving isn’t one thing, and knowing which thing it is changes what actually helps

A 2025 factor analysis in Drug and Alcohol Dependence, tracking 128 treatment-enrolled patients with opioid use disorder across four clinical visits, broke heroin craving into three distinct psychological components using the Heroin Craving Questionnaire: Lack of Self-Control (the sense that you couldn’t stop yourself even if you tried), Positive Expectancies (the belief that using would make things better), and Urgency (the raw intensity of the pull, independent of what you believed about it). That distinction matters clinically because the three factors don’t move together — someone can score high on Urgency while scoring low on Positive Expectancies, meaning the pull is intense but they’ve stopped believing it’ll actually help, which is a very different treatment target than someone who still believes using will fix something.

The study’s most useful finding for anyone bracing for relapse, or watching someone they love brace for it, cuts against the standard warning: baseline craving intensity did not predict who stayed abstinent across all four visits and who returned to use. Of the 128 patients, 33 remained abstinent throughout and 44 returned to use — and craving scores alone couldn’t sort them into those two groups in advance. A separate 2025 network analysis of 207 heroin users, drawing on national survey data, found that sleep disturbance functioned as the strongest bridge connecting craving to the rest of the withdrawal syndrome, and that craving itself correlated specifically with depressed mood and sweating rather than the full withdrawal picture. In plain terms: craving isn’t withdrawal’s shadow. It’s traveling with its own separate psychological engine, one that a sleepless night can rev up independent of how many hours since the last dose.

Your brain kept score, and researchers can now watch it happen

None of this is only in people’s heads in the sense of being imaginary — it shows up on a scan. Researchers publishing in Brain in late 2024 put 30 people with heroin use disorder and 25 people without it into an fMRI machine and showed them a drug-cue movie clip, then repeated the scan after 15 weeks of inpatient treatment. At baseline, 28 separate cortical brain regions synchronized differently — more in step with the drug cues — in the heroin group than in controls. After 15 weeks of treatment, 12 of those regions had normalized, and the orbitofrontal cortex specifically — a region involved in weighing the value of a reward against its cost — showed synchronization changes that correlated directly with how much a person’s reported craving dropped.

Two things follow from that, and they cut in opposite emotional directions on purpose. The reassuring one: this is treatable circuitry, not a permanent scar. Three and a half months of treatment measurably rewired regions that had been running in lockstep with drug cues. The harder one: this is circuitry, meaning it took real time under real clinical care to shift, not a decision made once in a moment of resolve. If you are the person who’s relapsed after “should have known better by now,” the honest answer from this research is that you were fighting a wiring pattern that a scanner can see and that takes months of sustained treatment to move — not a character flaw a stronger version of you would have overridden on the first try.

Three and a half months of treatment measurably rewired regions that had been running in lockstep with drug cues.

What changed in the brain when heroin got quietly replaced by something scarier

Here’s the piece almost nobody wrote about the heroin-to-fentanyl transition: users, by their own overwhelming account, did not want this trade. A 2024 scoping review in Health Psychology Research synthesizing the psychosocial literature on illicit fentanyl use found that people who’d used both drugs described fentanyl as scarier and less pleasurable than heroin — not a fentanyl generation that chose an upgrade, but a heroin generation whose supply got adulterated and then replaced out from under them. The same review found a significant correlation between fentanyl use and suicidal ideation, anxiety, and depression: among overdose survivors, roughly 60% reported some level of suicidal motivation behind the overdose, and 36% described it as a high level. People who ended up using pure illicit fentanyl specifically, rather than a fentanyl-heroin mix, showed greater addiction severity and more withdrawal problems than either the mix users or heroin-only users — the drug that replaced heroin on the street didn’t just change the risk of dying. For a meaningful share of the people using it, it changed the psychological experience of using into something closer to punishment than relief.

The withdrawal clock changed too, in a way that reshaped how people manage their own care. A 2022 study in Clinical Toxicology mined 267,136 posts across seven opioid-focused subreddits and manually coded hundreds describing precipitated withdrawal — the sudden, severe withdrawal that buprenorphine can trigger if started too soon after an opioid. In the heroin era, people reported that danger window closing at a median of 16 hours after last use. In the fentanyl era, the same danger window stretched to a median of 40 hours, sometimes up to a week. That is not a trivial difference for someone trying to time their first dose of medication-assisted treatment without medical supervision — it more than doubled the length of the hardest part of the process, and precipitated-withdrawal discussion volume on those forums tracked fentanyl’s rise in the drug supply almost exactly.

A treatment system built for heroin is still catching up to what replaced it — and to the people it left holding the transition

This is where the psychology of the individual and the failure of the system meet, and it’s worth naming out loud: America keeps rediscovering, a few decades late, that the infrastructure built to treat one era’s drug wasn’t built to treat the next one. Filter’s Helen Redmond traced how methadone clinics — the treatment infrastructure built specifically for heroin’s era — spent decades fighting for the right to simply exist in a neighborhood, dating back to the 1970s and 80s, when Dr. Beny Primm faced arson trying to open clinics in Harlem and Brooklyn: “The building was set on fire. We brought guard dogs to the location and they killed the dogs.” That same community resistance to treatment infrastructure was still alive in 2023, when a proposed clinic near a Lynnwood, Washington Boys & Girls Club drew the same fight in a new decade. The lesson isn’t really about methadone. It’s that every time a drug era shifts — heroin to fentanyl, injection to synthetic, one withdrawal timeline to a longer one — the treatment system and the surrounding community take years to catch up to what already changed on the street, and the people using in the gap pay for that lag with their own bodies.

You don’t have to be the exception to this system’s lag to get better inside it. The research above isn’t an argument that recovery from heroin, or from what fentanyl did in heroin’s place, is hopeless or purely mechanical. It’s the opposite: the orbitofrontal cortex changed in 15 weeks; the craving factors that predicted nothing about your future still described something real and nameable about your present; the withdrawal window that got twice as long is still a window, not a wall. Alex Mauri’s second essay didn’t end on defeat, even while refusing to pretend quitting solved everything: “Pressuring people to quit drugs is fundamentally wrong, as well as likely to fail — particularly when nothing is done to alleviate psychological or socio-economic circumstances.” That’s not a slogan. It’s the same finding the brain scans and the craving questionnaires keep arriving at from a completely different direction: the thing that changes outcomes was never the pressure. It was the time, the treatment, and the circumstances underneath — and all three of those are still things you can actually get more of, starting tonight.

It was the time, the treatment, and the circumstances underneath — and all three of those are still things you can actually get more of, starting tonight.