Psychedelics for addiction: promise, limits and caution

The Short Version

Psychedelic-assisted therapy has real, and sometimes striking, results for addiction and related conditions, but every expert in these conversations ties those results to a careful clinical setting, not to the drug alone. Dr. Matthew Johnson of Johns Hopkins describes one to three high-dose psilocybin sessions, with screening, preparation and integration around them, producing lasting change, including in smokers, and points to older ketamine-therapy work with high success rates for heroin and alcohol addiction [1]. Dr. Anna Lembke, an addiction psychiatrist at Stanford, calls herself optimistic about clinical use but warns that addiction is a chronic, relapsing condition that no quick fix is likely to solve, and she already treats patients addicted to MDMA and microdosing [2]. Andrew Huberman's MDMA episode reports large benefits in PTSD trials, with many participants' substance problems also resolving, but lays out the crash afterward, the unresolved neurotoxicity question and the danger of fentanyl in street drugs [3]. The honest verdict: there is real hope for compulsive behaviors, but it is a guided, supervised and still-experimental treatment, not a shortcut past the slow work of recovery.

Table of contents

What They Said

Matthew Johnson: big effects from a few guided sessions

Johnson, a behavioral pharmacologist who at the time of recording was a professor of psychiatry at Johns Hopkins and helped lead one of the world's leading psychedelic research centers, is the most optimistic voice here, and also the most specific about how the treatment is actually delivered [1].

What counts as a psychedelic. Johnson says the word is "more of a cultural term" than a pharmacological one, spanning several drug classes united by their ability to profoundly alter one's sense of reality, including the sense of self [1]. The classic psychedelics, such as LSD, psilocybin, DMT and mescaline, all act as agonists or partial agonists at the serotonin 2A receptor [1]. Ketamine, PCP and dextromethorphan are NMDA antagonists with more dissociative effects; salvinorin A is a kappa opioid agonist; and MDMA stands in a class by itself, releasing the brain's own serotonin along with dopamine rather than mimicking serotonin at the receptor [1]. This matters for the addiction debate, because "psychedelics" is not one treatment but a family of very different drugs.

The addiction evidence he points to. Johnson describes older Russian work by Krupitsky, which used high-dose ketamine delivered as a psychedelic therapy, with preparation and discussion. In his words: "And they found incredibly high rates of success and some pretty well-controlled trials for both heroin addiction and alcohol addiction" [1]. He contrasts that with the way ketamine is mostly used today. Esketamine (Spravato), approved for treatment-resistant depression, is given at low doses without psychedelic therapy, and its antidepressant effects last about a week [1]. Of ketamine clinics using the drug off-label, he says, "A lot of them aren't treating it like psychedelic therapy" [1]. With psilocybin, by contrast, his group sees effects a year or more later, and he wonders whether that gap is "a pharmacological difference" or whether clinics "get a lot more mileage out of ketamine if they treated it like psychedelic therapy" [1]. His point is that the therapy model around the drug may matter as much as the molecule.

He also mentions addiction work directly: the Beckley Foundation funded early smoking-cessation research, and his group had upcoming studies on opioid addiction [1]. A smoker in one of his studies said it felt like he could simply decide not to smoke, like flicking a switch [1].

How a trial works. Johnson walks through the Hopkins model step by step [1]:

• Screening over a couple of days, with structured psychiatric interviews to exclude psychotic disorders such as schizophrenia and bipolar disorder, plus cardiovascular screening.
• Preparation of about four to eight hours, building rapport with the guides and explaining the full range of possible effects. He describes drugs as "uppers, downers, and all arounders," with psychedelics as all-arounders: the experience could be the most beautiful or the most terrifying of someone's life.
• Dosing with pure psilocybin in a capsule, typically in the 20 to 30 milligram range for therapeutic trials, starting around 9 a.m. Participants wear eyeshades, lie on a couch, and are told that every emotional response is welcome. Blood pressure and pulse are checked about every half hour.
• Integration: participants write about the experience that night and return the next day for a one to two-hour session of supportive, non-directive therapy.

He advises against big decisions straight afterward, like ending a relationship or quitting a job: "sit on that two weeks" [1]. He also encourages participants to take the experience seriously, unlike friends who might dismiss it the next day [1].

Why it might work. Johnson's leading hypothesis is that "the common denominator are persisting changes in self-representation" [1]. He adds that "the self is the biggest model," including labels like smoker or failure [1]. He calls the resulting insights "duh experiences": people realize, with real force, something they had told themselves a million times, such as that they could simply decide [1]. He credits philosopher Chris Letheby's book with arguing that changes in self-representation can explain therapeutic effects from a naturalistic point of view, with or without spiritual interpretations [1].

His cautions. Johnson is not a cheerleader. He names two big risks. The first is people with psychotic disorders or the mania of bipolar disorder, or a strong predisposition to them, whom psychedelics can destabilize; trials exclude them, and even people with a close relative who has these disorders [1]. The second, far more common, is a bad trip. Even in the Hopkins setting at around 30 milligrams of psilocybin, "about a third of people will say, essentially at some point they have a bad trip" [1]. He keeps a growing file of cases in which people panicked and were hurt or killed, including an 18-year-old in Oregon who took a large dose, ended up disoriented in a neighbor's house and was killed by police [1]. He also notes that most common psychedelics, including MDMA, stimulate the serotonin 2B receptor, which has been linked to heart valve problems and was why fen-phen was pulled from the market; a few therapeutic sessions are not a concern, but dosing twice a week for years is a different question [1].

On microdosing, he reports that the handful of double-blind studies, four or five at the time, found anything from no effect to slight impairment, and none showed improved creativity, cognition or sustained mood [1]. His own interest lies elsewhere: "I'm interested in big effects," meaning durable change from one to three high-dose sessions [1].

On policy, he says, "I like the word regulation better than legalization," and argues that clinical research would be needed even if these drugs had never been criminalized [1]. In the transcript he explains that with screening, preparation and integration, "the efficacy is going to be increased and the risks are going to be mitigated drastically" [1].

Anna Lembke: optimism for the clinic, alarm about the culture

Lembke, chief of the Stanford Addiction Medicine Dual Diagnosis Clinic and author of Dopamine Nation, comes at psychedelics from the other direction: she spends her working life with people whose brains have been pushed out of balance by high-dopamine substances and behaviors [2].

Her framework for addiction. For Lembke, "pleasure and pain are co located": the same brain regions process both, and they act like a seesaw that always wants to return to level [2]. Every tip toward pleasure is followed by an equal and opposite tip toward pain. If people keep indulging, the brain turns down its own dopamine transmission until they are stuck on the pain side, a state she likens to clinical depression, with anxiety, irritability, insomnia and constant preoccupation with using [2]. Her core treatment is 30 days without the drug or behavior, which in her experience is roughly how long the reward pathway takes to recover, and she warns patients, "you will feel worse before you feel better" [2]. She cites a study by Schuckit and Brown in which depressed men addicted to alcohol were hospitalized with no depression treatment and no alcohol; after four weeks, 80% no longer met criteria for major depression [2].

In severe addiction, she says, the balance itself loses its resilience and stays tipped to the pain side for months or years [2]. Relapse can come from stress, or from things going well, because success lets people drop the vigilance that keeps their use in check [2]. Huberman shares a friend's line, 40 years sober: "no matter how far you drive, you're always the same distance from the ditch" [2]. Lasting recovery, in her view, depends on vigilance, honest connection, telling the truth and taking life one day at a time [2].

The contradiction Huberman raises. Against that backdrop, Huberman points out that psychedelic therapy seems to break every rule they have discussed. He describes the idea as taking people in a pattern of addiction and launching them into an experience that is "also chemical and extreme," which is "supposed to eject the person into a life where drug use isn't as much of interest" [2]. He puts it bluntly: "This violates everything we've talked about in terms of dopamine biology," and "It's a dopamine feast" [2]. About MDMA specifically he adds, in passing, "how neurotoxic if neurotoxic debatable" [2].

Her answer: skepticism, then cautious openness. Lembke replies, "I think it's good that you're skeptical. I think we all should be skeptical" [2]. She acknowledges clinical studies showing benefit, while stressing that "these are small studies and there's short duration, small number of subjects" [2]. The examples she gives involve people addicted to alcohol having a psychedelic experience "in a very controlled setting," typically high-dose psilocybin or MDMA [2]. The benefits have appeared only in carefully selected people and always within psychotherapy [2].

Her metaphor is taking the gondola rather than walking up the mountain: psychedelic therapy as a condensed version of therapy that can help people see their lives from a wider view and reconnect with their values [2]. The metaphor cuts both ways. A gondola gets you the view, but it does not build the legs you need to walk back down. That is the heart of her doubt: addiction is a chronic, relapsing condition, and she doubts a quick fix works for it [2].

What worries her most. Her biggest concern is not the clinic but the culture. She describes "a lot of people who are looking for some kind of spiritual awakening who on their own, not in the context of any kind of therapeutic psychological work," microdose or try psilocybin or MDMA with a friend [2]. Her verdict: "That's a disaster and almost never works out well" [2]. She also rejects the idea that psychedelics can't be addictive: "I have patients clinically who definitely are addicted to, you know, MDMA to microdosing" [2]. She blames the speed at which popular books spread: "the narrative of popular authors can expand and wick out so fast that pretty soon people are essentially taking their mental health into their own hands" [2].

And yet she ends on hope: "I actually have I have great optimism for this business of clinical use of psychedelics, including MDMA," adding that "Matthew Johnson at Johns Hopkins is doing fabulous work on this" [2].

Andrew Huberman as host: splitting trauma from addiction

Across the Lembke conversation, Huberman offers his own observations. He notes that two people with addiction he knows got worse after MDMA-assisted therapy, while "the people I know who had severe trauma, who did this, who took this approach seemed to be doing better" [2]. He objects to lumping all psychedelics, and trauma and addiction, together: "I'm a splitter not a lumper, as we say in science," and "this is definitely not a one size fits all kind of thing, but it sounds like it may have some utility under certain conditions" [2]. He also highlights Johnson's work and research separating the hallucinogenic effects of psychedelics from their antidepressant ones [2]. These are personal anecdotes, not trial data, but they sharpen the question of who benefits.

Huberman's MDMA episode: real benefits, real costs

In his solo episode, Huberman goes deep on one compound, MDMA, and treats it as neither a miracle nor a menace [3].

How it works. Like methamphetamine, MDMA blocks dopamine transporters and causes massive dopamine release, making it a stimulant; unlike methamphetamine, it does the same for serotonin, with serotonin increases at least three and perhaps as much as eight times greater than the dopamine increases [3]. The result is an "empathogen": elevated energy without irritation, emotional warmth and trust, directed toward others and toward oneself [3]. A mouse study from Robert Malenka's lab at Stanford found that dopamine release drives MDMA's rewarding, reinforcing quality, while serotonin acting on 1B receptors in the nucleus accumbens drives its pro-social effect [3]. That split is directly relevant to addiction: the same drug carries both the pull toward repeated use and the social effect that therapy tries to harness.

Huberman also describes imaging work. People on MDMA rate threatening faces as less threatening, with reduced amygdala activity [3]. PTSD is associated with heightened connectivity between the amygdala and the insula, and in people given MDMA alongside therapy that connectivity weakens, with the degree of weakening scaling with symptom relief [3]. MDMA dramatically increases oxytocin, but the evidence suggests oxytocin plays at most a minor role in its social effects [3].

The clinical results. In the MAPS Phase 3 trial for severe PTSD, participants had three 90-minute preparatory sessions with two therapists, three roughly eight-hour sessions with MDMA or placebo, and three 90-minute integration sessions [3]. Doses were 80 milligrams with an optional 40-milligram booster in the first session, and 120 milligrams with an optional 60-milligram booster in the second and third [3]. The results: an 88 percent clinically effective response rate with MDMA-assisted therapy versus 60 percent with therapy and placebo, and 67 percent of the MDMA group no longer met criteria for PTSD by the end [3]. Huberman says he knows of no other psychiatric treatment with success of this magnitude and invites psychiatrists to correct him [3]. For the addiction question, the most relevant finding is that many participants with alcohol or other substance use disorders also saw those problems resolve, and MDMA is now being explored for alcohol use disorder, depression and eating disorders [3]. MDMA therapy did not increase suicidality compared with placebo [3].

He repeats one line as a guardrail: "MDMA taken on its own does not cure PTSD" [3]. It augments talk therapy. He draws on psychiatrist Paul Conti's definition that "trauma is an event that fundamentally changes the way that our brain works for the worse," and explains that under MDMA people are more willing to trust both their therapist and themselves to explore such events [3].

The crash. Many people experience a crash of low mood, lethargy and low motivation after MDMA [3]. Huberman says it is unlikely to be caused by neuron death in the first 24 to 48 hours, and that internet protocols recommending L-tryptophan, 5-HTP or L-tyrosine afterward lack evidence and are illogical, possibly even further depleting serotonin and dopamine [3]. A more plausible contributor is prolactin, which rises after big dopamine increases and is associated with lethargy; some clinicians are exploring P5P, a form of vitamin B6 known to suppress prolactin, but there are no controlled human studies yet [3]. Read alongside Lembke's seesaw, the crash is the pain side of a very large tip toward pleasure [2][3].

Neurotoxicity. Methamphetamine is well known to be neurotoxic, even in a single dose, but MDMA is more complicated [3]. Rodent studies show reductions in serotonin-related proteins after repeated doses, though short-term depletion is not the same as neuron death [3]. MDMA's reputation for "putting holes in your brain" largely traces to a 2002 paper in Science that was later retracted after the authors acknowledged that, due to labeling issues, the monkeys had actually received methamphetamine [3]. A study of long-term ecstasy users in Utah who avoided other drugs, including caffeine and alcohol, found little evidence of decreased cognitive performance in moderate users (22 to 50 lifetime doses) and heavy users (60 to 450 doses), though heavy users showed poorer strategic self-regulation, a finding whose direction of cause is unknown [3]. His overall read: with pure MDMA, at clinical doses, without other drugs, in controlled settings, the risk seems lower than the popular press suggests, but the question is not resolved; risk rises with high or frequent doses, other drugs such as caffeine, and hot settings, since even a three- to four-degree rise in body temperature can begin to kill neurons [3].

Street drugs. His strongest warning is about contamination. He cites estimates that 60 to perhaps 80 percent of gray-market drugs are being repackaged or reformulated with fentanyl, which has caused many deaths [3]. As of June 2023, MDMA remained a Schedule I drug in the United States [3]. He stresses that MDMA is not a miracle cure and that time will reveal any long-term problems [3].

Where They Agree, and Where They Don't

Where they agree

The setting is the treatment. All three sources tie the benefits to a structured clinical model. Johnson's trials use days of screening, hours of preparation and follow-up integration [1]. Lembke says benefits appear only in carefully selected people and always within psychotherapy [2]. The MAPS trial wraps three drug sessions inside nine hours of preparation and integration with two therapists [3]. Huberman's "MDMA taken on its own does not cure PTSD" [3] and Johnson's view that ketamine clinics "aren't treating it like psychedelic therapy" [1] make the same point from two directions.

Unsupervised use is the danger zone. Johnson's file of people hurt or killed during bad trips [1], Lembke's "That's a disaster and almost never works out well" about solo spiritual seeking [2], and Huberman's fentanyl warning [3] all point the same way.

Microdosing is unproven. Johnson reports double-blind studies finding no effect to slight impairment [1]; Lembke has patients addicted to microdosing [2].

Screening matters. Johnson's trials exclude people with psychotic or bipolar disorders and those with close relatives who have them [1]; Lembke stresses carefully selected participants [2].

Where they don't

How much to hope. Johnson is excited about "big effects" and year-plus results from one to three sessions [1], and the MAPS numbers make Huberman say he knows of nothing comparable in psychiatry [3]. Lembke, whose daily work is relapse, is more guarded: the studies are small and short, and addiction is chronic and relapsing [2]. The difference comes partly from what each person looks at. Johnson and the MAPS trial measure outcomes in selected volunteers inside rigorous protocols; Lembke sees the people who didn't get selected, who self-medicated, or who relapsed years after getting well.

Whether a "dopamine feast" can treat a dopamine problem. Huberman's challenge that the approach "violates everything we've talked about in terms of dopamine biology" [2] is never fully answered. Johnson's answer would be a different mechanism: the change comes from lasting shifts in self-representation, not from the dopamine itself, and classic psychedelics act mainly at the serotonin 2A receptor [1]. The MDMA episode offers a partial answer too: serotonin at 1B receptors, not dopamine, drives the pro-social effect, while dopamine drives the rewarding quality [3]. That means the risk Huberman worries about is real for MDMA in particular, since part of its effect is exactly the reinforcement that drives addiction [3].

Can you get addicted to psychedelics? Lembke says the claim that you can't "doesn't make any sense to me," and she has patients addicted to MDMA and microdosing [2]. The MDMA episode supports her on the mechanism, since dopamine release gives MDMA a rewarding, reinforcing quality [3]. Johnson's worry about long-term frequent use centers on heart valves via the 2B receptor rather than addiction [1].

Trauma versus addiction. Huberman's anecdote, that people with addiction he knew got worse after MDMA therapy while people with trauma did better [2], sits uneasily with the MAPS finding that many PTSD participants' substance problems resolved [3]. The two can both be true: in the trial, addiction improved as part of treating trauma, in screened patients, with therapy; Huberman's acquaintances were being treated for addiction itself. None of the sources settles which patients do best.

Neurotoxicity. Huberman's aside in the Lembke conversation, "how neurotoxic if neurotoxic debatable" [2], matches his fuller view in the MDMA episode: lower risk than the press suggests under controlled conditions, but not resolved [3].

Why It Works

The sources offer several mechanisms, none presented as proven.

Loosening the brain's models. Johnson describes people as prediction machines whose perception is largely top-down; psychedelics loosen learned predictions [1]. The biggest model is the self, including labels such as smoker, and a psychedelic session may let people revise it, producing the "duh experience" of realizing they could simply decide [1]. For addiction, where identity and habit are bound up together, that is a plausible route to change.

Letting go. Huberman frames the therapeutic core as letting go, and Johnson agrees that people who try to hang on are more likely to have a bad trip [1]. Huberman adds that neural networks change in response to experience, given the right neuromodulators, not language or affirmations, and that plasticity is a process that may unfold long after the trigger [1]. This is one reason integration after the session matters.

Glutamate and plasticity. Classic psychedelics activate the 2A receptor, which increases glutamate transmission downstream, possibly a commonality with ketamine [1]. Rodent research from David Olson's and Bryan Roth's labs shows growth of dendrites and new connections in the days after psychedelics [1]. Huberman's broader takeaway about MDMA is that it is a tool for accessing neuroplasticity, helping the brain rewire when used safely in the right context [3].

Lower threat, more trust. MDMA reduces amygdala reactivity, weakens amygdala-to-insula connectivity in proportion to symptom relief, and, through serotonin 1B receptors in the nucleus accumbens, rewards social connection itself [3]. Johnson similarly speculates that reduced amygdala reactivity may let people reprocess traumatic memories without a fight-or-flight response, and notes that truly challenging bad trips are much less likely with MDMA than with classic psychedelics [1]. Because the reward attaches to connection, it may make connection more likely even after the drug wears off [3], which echoes Lembke's point that honest human connection is a powerful source of dopamine that can replace what drugs provided [2].

A wider view. Lembke's gondola metaphor describes the same effect in plain terms: a condensed version of therapy that shows people their lives from above and reconnects them with their values [2].

Why the setting boosts it. Johnson's comparison of low-dose esketamine (about a week of benefit) with psilocybin therapy (a year or more) and Krupitsky's ketamine-as-psychedelic-therapy work suggests the therapy model itself may extend and deepen the effect [1].

Where It Applies, and Where It Doesn't

Where it looks promising:

• Alcohol, heroin and nicotine addiction in clinical trials. Krupitsky's ketamine therapy for heroin and alcohol [1], smoking-cessation work and planned opioid studies at Hopkins [1], and small, short studies of high-dose psilocybin or MDMA for alcohol addiction [2].
• Addiction that travels with trauma. Many MAPS participants with substance use disorders saw those problems resolve alongside their PTSD [3], and Huberman's own observations favor people with severe trauma [2].
• Screened, prepared adults. Every positive result here comes from people who passed psychiatric and medical screening and were guided through the experience [1][2][3].

Where it doesn't apply, or is risky:

• People with, or prone to, psychosis or mania, including those with a close relative who has these disorders [1].
• Solo or "with a friend" use for a spiritual awakening, which Lembke says almost never works out well [2].
• Microdosing, which the double-blind evidence doesn't support [1] and which Lembke has seen become an addiction [2].
• Street drugs, given estimates that 60 to perhaps 80 percent of gray-market drugs are being repackaged or reformulated with fentanyl [3].
• Frequent, long-term use, given the 2B receptor heart-valve concern [1] and rising neurotoxicity risk with high or frequent doses [3].
• People taking SSRIs, at least for MDMA: an SSRI taken beforehand blocks MDMA's pro-social and empathogenic effects [3].
• People under 18, where no formal research exists yet [1].
• Anyone hoping to skip recovery work. Addiction is chronic and relapsing [2], and vigilance never ends [2]. Even the best trial results come with the caveat that time will reveal long-term problems [3].

Legal status also limits where it applies. These compounds remain Schedule I federally, and local decriminalization does not eliminate risk [1][3]. If you are considering any of this, talk with a qualified physician or addiction specialist first, particularly if you have a psychiatric history, heart problems, or take other medications.

What To Do With It

1. Treat psychedelics as a possible add-on to treatment, not a replacement for it. Every positive result here came inside psychotherapy, with preparation and integration [1][2][3]. If you are struggling with addiction, start with proven approaches such as Lembke's 30-day abstinence, expecting to "feel worse before you feel better" for about two weeks [2].
2. If you're interested in psychedelic therapy, go through a trial. Johnson points people to the Hopkins psychedelic research website and clinicaltrials.gov, and his group listed upcoming studies on opioid addiction [1].
3. Be honest in screening. Disclose any personal or family history of schizophrenia, bipolar disorder or mania, and any heart problems; these are exactly what trials screen for [1].
4. Don't self-medicate for a spiritual breakthrough. Lembke's clinical experience is that this "almost never works out well" [2], and bad trips outside a protected setting can turn deadly [1].
5. Skip microdosing as an addiction or mood fix. Controlled studies found no benefit to slight impairment [1], and it can become its own addiction [2].
6. Never trust street supply. Fentanyl contamination is widespread in gray-market drugs [3].
7. Plan for the aftermath. Don't make big decisions right after an experience; "sit on that two weeks" [1]. Take the integration work seriously, since plasticity is a process that unfolds over time [1]. After MDMA, don't rely on internet supplement protocols for the crash; Huberman says they lack evidence [3].
8. Keep up the vigilance afterward. A breakthrough does not end a chronic, relapsing condition [2]; build the honest connection, truth telling and one-day-at-a-time habits Lembke describes as the core of lasting recovery [2].
9. Match the drug to the problem. "Psychedelics" covers very different drugs with different effects [1][2]; the strongest data here are for MDMA in PTSD with associated addictions [3], while psilocybin evidence for addiction comes from small, short studies [2].

Sources

1. Dr. Matthew Johnson: Psychedelics for Treating Mental Disorders · Huberman Lab
2. Dr. Anna Lembke: Understanding & Treating Addiction · Huberman Lab
3. The Science of MDMA & Its Therapeutic Uses, Benefits & Risks · Huberman Lab