Psychedelics and cannabis

Almost all randomized trials of cannabis products in autism have been in children and adolescents.¹ The largest gave 150 young people aged 5 to 21 either a whole-plant cannabis extract (20 parts CBD to 1 part THC), purified CBD and THC in the same ratio, or a placebo, for 12 weeks.² It missed one of its two main outcomes. On the other, a clinician's rating of disruptive behavior, 49% on the whole-plant extract were rated much or very much improved against 21% on placebo; the purified product (38%) did not differ significantly from placebo, and the two products did not differ from each other on any measure.² On a secondary scale of autistic social traits, scores fell more on the extract than on placebo (median 14.9 against 3.6 points, in the smaller group who reached that measure).² Both products contained THC. Sleepiness was more common on the cannabis products (28% on the extract, 23% on the purified product, 8% on placebo); reduced appetite (25%, 21% and 15%) did not differ significantly; and no serious adverse event was judged to be caused by treatment.² A later analysis of the same trial found the cannabis products no better than placebo on any sleep measure in a parent questionnaire.¹⁰ Two of the authors had fees and stock options from the company that made the products, which also part-funded the trial, and a single clinician, the paper's lead author, rated all of the improvement scores.² Most participants in these two largest child trials were boys (80% in this one, all of them in the next), so their results say little about autistic girls.^2,3

Later trials have mostly been smaller. In the United States, pharmaceutical CBD given to 39 autistic boys aged 7 to 14 with severe behavior problems did no better than placebo on any main outcome; both phases brought improvement, a strong placebo effect; blinded clinicians rated about two thirds of the boys improved on CBD, against about a third who improved or did not change on placebo.³ An Australian crossover trial in which 29 children aged 5 to 12 completed both phases missed its main outcome, a social responsiveness scale, while some secondary measures (anxiety, social relating, parents' stress) improved.⁴ A Brazilian trial in 60 children aged 5 to 11, using an extract that also contained THC, reported improvements in social interaction, anxiety and agitation.⁵ A later analysis of 19 of those children found the CBD group improved on recognizing facial expressions and on theory of mind within the group, but not more than placebo.⁵² A 2026 trial of an extract with less than 0.3% THC, in 61 children with higher support needs from a specialist neurology clinic (54 analyzed, aged 8 to 17, about half boys), reported scores about one step lower on a seven-point clinician rating of overall autism severity after 8 weeks; the product's developer and a partner company funded it, and several authors were funded by, consulted for or held shares in them.⁶

Reviews agree on how weak this is. A review of randomized trials using Cochrane methods rated the certainty of the evidence very low to low and found that no trial measured quality of life.¹ A meta-analysis of trials in young people found a small effect in autism, but it rested on two trials, both at high risk of bias, and the authors called the evidence insufficient to recommend cannabinoids broadly.⁷ A 2026 meta-analysis of cannabinoids for mental and substance use disorders, which included an autism subgroup, found slightly lower scores on scales of autistic traits, with mostly low-certainty evidence, and judged routine use rarely justified.⁸ A 2026 systematic review of cannabinoid products in children and adolescents with autism, covering 12 studies of which 4 were randomized trials, reached the same conclusion: it rated the certainty of the evidence as limited and supported only cautious use alongside other treatment, under medical supervision, and found no difference from placebo on sleep or overall symptom severity.⁴⁸ A second 2026 review, across autism, fragile X syndrome and Rett syndrome, found 17 studies, only 2 of them randomized trials; pooling them descriptively suggested some behavioral improvement, with mostly mild side effects, but the review concluded the evidence was not enough for routine use.⁴⁹ These reviews and meta-analyses likely draw on many of the same few trials. Some outcomes in these trials, such as the social responsiveness scale and the clinician rating of overall autism severity, score autistic traits themselves, and not every autistic person would count lower scores as a benefit. A 2026 network meta-analysis of 106 randomized trials, comparing autism treatments against one another, found the cannabinoid group had a lower acceptance rate than most others.⁵⁰ One uncontrolled study, of 188 children and young people in an Israeli program, reported improvement in most of those assessed at six months, but only 93 were assessed and there was no comparison group; two authors were employees of the cannabis supplier and a third a paid adviser to it.⁹

Separate from trials, health records show who is prescribed cannabis. In health records for 36,610 autistic people in a large Israeli health plan, only 1.2% had been prescribed medical cannabis, and 4.3% of those prescriptions were for children under 5.⁵¹ Those prescribed were diagnosed earlier on average and were more likely to also have ADHD, an intellectual disability or epilepsy. They came disproportionately from higher-income families, and Arab people were under-represented (2.7% against 11%).⁵¹ Israel allows cannabis for autism only after trials of two approved antipsychotics, but only 28% of prescriptions had documented both, though 69% had tried at least one.⁵¹

In a UK online survey with self-reported diagnoses (166 people aged 16 and over in the matched sample), autistic adults were more likely than non-autistic adults to have used CBD in the past year, about three and a half times the odds, but no more likely to have used cannabis.¹¹ In the archive of a cannabis-tracking app, 111 adults who identified as autistic, with no verified diagnosis, rated their symptoms before and after inhaling cannabis (a median of 2 logged sessions each), and ratings fell by about 73%; there was no comparison group, and only short-term effects were rated.¹² A UK registry of 130 autistic adults prescribed cannabis medicines reported better anxiety, sleep and quality-of-life scores over 18 months, and about 1 in 5 reported adverse events; it had no comparison group, and most authors work for the clinic that prescribes the medicines and owns the registry.¹³

Brain studies in adults have used single doses to see whether autistic brains respond differently. In 17 autistic and 17 non-autistic men, one dose of CBD moved GABA, a chemical messenger that damps brain activity, in opposite directions, mainly in the prefrontal cortex: up in non-autistic men and down in autistic men, measured with a scanner.¹⁴ Complete data were available for only 8 autistic men, and the authors say the results do not speak to whether CBD works as a treatment.¹⁴ Two small studies by different groups, mostly of boys, found lower blood levels of anandamide, one of the body's own cannabis-like molecules, in autistic children and young people; the larger drew its autistic participants from the cannabis trial above.^15,16

Ketamine, an anesthetic also used in lower doses for depression, has been tested against placebo in autistic people in one study, and no further one has been found as of October 2026. That small crossover trial gave two doses of nasal-spray ketamine and two of placebo to 21 people aged 14 to 29 and found no significant effect on social withdrawal, the outcome it targeted; ketamine was well tolerated, with only brief, mild side effects.¹⁷ Since then, reports of esketamine for depression in autistic people are a single case, a girl, and a series of three adults aged 20 to 25, both from the same Italian group.^19,18 In the series, where esketamine was added to an antidepressant, depression scores fell, and the authors note that a case series cannot support statistical conclusions.¹⁸

Trials in genetic conditions that often include autism give a similar picture. In Rett syndrome, a placebo-controlled trial with 22 girls analyzed found no difference in effect between ketamine and placebo after five days of low oral doses; its main aim was safety, and it stopped early.²⁰ An open-label study of one dose in 10 children with ADNP syndrome, with no placebo group, reported no serious adverse events and called its findings hypothesis-generating; two authors hold patent applications on the use.²¹ The work often quoted for how ketamine might help was done in mice, with no autism model: ketamine made new connections on prefrontal neurons form more readily when glutamate was applied.²²

One pilot trial randomized 12 autistic adults with marked to very severe social anxiety to MDMA (8 people) or placebo (4) in two 8-hour therapy sessions with preparation and follow-up sessions.²³ One month after the second session, social anxiety scores had fallen 44.1 points on MDMA and 19.3 on placebo, and a difference was reported to persist at six months.²³ The paper reports both differences as statistically significant, yet the confidence intervals around both effect sizes include zero, so the two measures do not agree.²³ Participants guessed correctly whether they had MDMA or placebo in all but one of 23 sessions, though the authors judged the blinding adequate.²³ There were no serious adverse events; anxiety during sessions was more common on MDMA (75% against 25%).²³ The Multidisciplinary Association for Psychedelic Studies (MAPS) sponsored the trial, and several authors were its employees or funded by it.²³

In a separate survey of 100 autistic adults who had used MDMA outside a clinic, none reported long-term harm from it.²⁴ Its sole author also led the MAPS trial, so the two are not independent, and people recruited through a website may not include those who were harmed.²⁴ No larger trial in autistic people had been published as of October 2026; a small open-label study of a variant of MDMA was recruiting in 2026.⁴⁷ A 2025 commentary records that in 2024 the US Food and Drug Administration declined to approve MDMA-assisted therapy for PTSD.²⁵

In an online survey of 233 autistic adults who had used psilocybin or LSD, about 82% attributed less psychological distress to their most impactful experience, 78% less social anxiety and 70% more social engagement.²⁶ About 58% had a professional diagnosis; the rest identified as autistic.²⁶ In the same survey, about 20% reported more anxiety and about 11% more distress overall, and some described their experience as among the most negatively impactful of their lives.²⁶ The survey was recruited partly through psychedelic organizations and co-produced with an autistic psychedelic group, and its authors say it probably over-represents good experiences.²⁶ A Toronto-based online survey of 261 autistic adults, about half living in Canada and most of the rest in the US, found that 70% had used psychedelics, mostly psilocybin mushrooms. That figure describes this self-selected sample, not autistic adults in general. About 19% of users had taken only small doses or microdoses, and 10% of those asked reported negative effects that persisted, mostly for days to weeks.²⁷

The PSILAUT study at King's College London gave single low doses of psilocybin and placebo to autistic and non-autistic adults to measure how their brains responded, not to treat anything.²⁸ Its first results, in a preprint not yet peer reviewed, report that the higher of the two doses increased connectivity, on average, between certain large brain networks, linked to attention and to self-directed thought, in 37 autistic adults and decreased it in 30 non-autistic adults.²⁹ Two authors of the study protocol were affiliated with COMPASS Pathfinder, the company that makes the psilocybin used, which also part-funded the study and supplied the drug.²⁸ In the registry searched for this entry, one treatment trial is listed: an open-label study in Toronto planning to give psilocybin with psychotherapy to 20 autistic adults with depression that has not responded to other treatment.³⁰ No registered LSD trial in autistic people was found. A 2022 review found that the early studies in autistic children reported some gains in behavior alongside aggression and psychotic reactions, and called for studies that weigh benefits against risks; a 2025 scoping review found 4 studies, all at a low level of evidence.^31,32

Between 1959 and 1974, researchers gave LSD to children diagnosed with autism or "childhood schizophrenia".^33,31 One 1962 study gave it to 12 children, 4 of whom had severe anxiety, and a 1966 study by James Simmons, Ivar Lovaas and colleagues, unusual for being placebo-controlled, gave it to twin boys of about five.^34,35,31 The children's diagnoses do not map neatly onto today's criteria for autism.³¹ A review of this work found that most studies lacked proper controls and offered narrative descriptions, and that LSD was often justified on the grounds that other treatments had failed.³³

Recent authors judge these studies harshly. A 2022 review notes "significant methodological and ethical shortcomings" and warns that informed consent would be a challenge in any future trial with children or people with intellectual disability.³¹ The authors of the 2025 survey call the studies "unethical and methodologically flawed".²⁶

Pooling small studies, about 28% of autistic people have raised levels of serotonin in whole blood, a measure taken in the blood rather than in the brain.³⁶ Psilocybin and LSD act mainly, though not only, on one serotonin receptor, the 2A receptor, so some researchers expect autistic brains to respond to them differently.^31,28 One model of how psychedelics work, called REBUS, proposes that they relax the brain's high-level expectations and let more raw sensory information through.³⁷ A 2025 review set this beside predictive-processing accounts of autism in adults: both may shift weight toward sensory information, but possibly through changes at opposite ends of the brain's hierarchy, and the review's conclusions are conceptual.³⁸ For cannabis, the authors who found lower endocannabinoid levels in autistic children, mostly boys recruited from their own CBD trial, proposed the finding as support for trying cannabinoid treatment.¹⁵

Common psychiatric medicines can change how these drugs act. In small, mostly older studies, SSRI and MAOI antidepressants weakened the effects of LSD, though one trial found escitalopram did not weaken the felt effects of psilocybin; drugs that block the serotonin 2A receptor reduced or blocked them (ketanserin for LSD, risperidone for psilocybin).³⁹ Lithium taken with LSD or psilocybin has been linked to seizures in online trip reports.³⁹ Several trials in healthy adults found that SSRIs reduce the effects of MDMA, and case reports describe MDMA taken with MAOIs causing overheating, seizure-like activity and suspected serotonin syndrome, many of them fatal.⁴⁰

Pharmaceutical CBD is approved in the US for some epilepsies, not for autism. Its label states that it can raise liver enzymes, with higher risk alongside valproate, and that in a study of healthy volunteers it raised the level of a breakdown product of clobazam about threefold.⁴¹ A review links regular use of high-THC cannabis in adolescents and young adults to psychosis and to cannabis use disorder.⁴² Case reports describe suspected mania or psychosis in three young autistic men after regular cannabis use, and a short psychotic episode in one child on a higher-THC product.^43,44 Bladder problems are a recognized harm of recreational ketamine use, though estimates of how common they are vary widely; in 14 trials of ketamine as a psychiatric treatment, urinary symptoms differed little from comparison groups.^45,46

These findings summarize studies; they are not guidance on combining, starting or stopping any medicine or substance. Laws on all of these substances differ between countries and change often, and nothing here is legal advice.

Several autistic writers describe substances, cannabis among them, as a way of coping with masking. Devon Price writes that some autistic people self-medicate social anxiety and "reach for substances to help blunt our sensory issues or help us pretend to be more confident", and that "Alcohol, weed, and other depressants are also an appealing and socially acceptable way of unwinding".^c1 Price also notes that in some places a diagnosis can qualify a person for a medical marijuana card.^c2 Megan Griffith, offering what she calls a guess, links substance use to how overwhelming the world is, to substances working as a social lubricant, and to the identity crisis that masking causes.^c3

Writers also describe harm. Sol Smith tells of River, a client who was "almost always high on everything from weed to meth" to cope with an overwhelming life; after an overdose and getting sober, everyday sounds became unbearable, and Smith presents River's own guess that this was why they had used drugs.^c4 Eric Garcia, in a chapter on the search for a cure, lists LSD among the treatments autistic people were subjected to in the decades after autism was first described.^c5

Psychedelic use has its own autistic writing. Aaron Orsini's Autism on Acid (2019) describes, in its subtitle, LSD helping him understand his autistic perceptions.^c6 He later co-authored the 2025 survey of 233 autistic adults, which the Autistic Psychedelic Community co-produced.²⁶

The neurodiversity writer Nick Walker, a consultant from 2012 to 2018 on the MDMA pilot study described above, writes that the study "wasn't in any way about 'treating autism'", a concept Walker compares to "treating homosexuality". In Walker's account, the study treated social anxiety in consenting autistic adults who wanted to be rid of it, and framed that anxiety as "a symptom not of autism but of traumatic oppression".^c7