Cannabis and Autism: Why Parents Are Trying CBD and What the Trials Found

Search for cannabis and autism and you will find parents describing transformations, and studies that appear to confirm them.

Then you reach the trials that included a placebo group, and the picture changes considerably. This article is about that gap, because anyone making a decision here deserves the second half of the evidence as well as the first.

We sell cannabis seeds. We have no product in this conversation and no reason to sell you an outcome, which is exactly why we can report the null results.

The short answer

The open-label studies are encouraging and the controlled trials are mostly not. Three of the four randomized placebo-controlled trials to date missed their primary endpoint, and every one of them documented a large placebo response. That combination is the single most important fact in the subject.

Why parents are trying it

Start with the scale, because it explains the demand. The CDC's current surveillance estimate is that about 1 in 31 (3.2%) children aged 8 years has been identified with ASD. Many of those families are managing severe irritability, aggression or self-injury, and the approved medication options are antipsychotics with substantial metabolic side effects.

Into that gap came a set of Israeli studies with striking numbers. The largest followed 188 ASD patients treated with medical cannabis between 2015 and 2017, and among those who answered the six-month questionnaire, 30 percent reported significant improvement and 54 percent moderate improvement.

Two things about that study get lost in the retelling. Fewer than half the original cohort provided six-month data, so the widely quoted improvement percentages describe 93 respondents, not 188 enrolled patients. And the authors state the limitation themselves: this is an observational study with no control group, so no causality between cannabis therapy and improvement in patient wellbeing can be established.

That is not a criticism of the researchers. It is what they wrote.

What happened when someone added a placebo

The decisive trial randomized 150 children and adolescents in Israel, double-blind, placebo-controlled, with a crossover design and two active formulations tested against placebo.

Its conclusion is one sentence: evidence for efficacy of these interventions are mixed and insufficient. Changes in the primary outcome score did not differ among groups.

There was one positive signal, and it deserves accurate reporting. On a co-primary clinician-rated measure, disruptive behavior was much or very much improved in 49 percent on whole-plant extract against 21 percent on placebo. The purified cannabinoid arm reached 38 percent, which did not separate from placebo.

So the honest summary of the most rigorous trial in this field: one measure showed a real effect for whole-plant extract, the other did not, purified cannabinoids failed, and the authors declined to call the result sufficient.

The trials since have not rescued it. A 2025 US crossover trial in boys aged 7 to 14 found both groups improved with no significant difference between them, and its authors highlighted a prominent placebo effect. A 2026 Australian trial found no significant effect on its primary social-responsiveness measure. A 2024 Brazilian trial did report significant gains, largely on parent-report instruments the authors created themselves.

Four controlled trials, three primary endpoints missed. Every one of them recorded that parents saw improvement on placebo, too.

Why the placebo effect matters more here than usual

This is the mechanism behind the gap between parent reports and trial results, and it is not about anyone being fooled.

Autism symptoms fluctuate. Outcome measures are mostly parent-rated questionnaires. Families entering these studies have often tried many things and are hoping hard. The Israeli crossover trial detected exactly this, finding change from baseline greater in the first period than the second, which the authors read as a larger initial placebo effect.

An open-label study cannot separate any of that from the drug. Which is why a 44 percent response rate in an uncontrolled 2026 trial tells you very little, and why the 21 percent placebo response in a blinded trial tells you a great deal.

The unexamined day is a wasted opportunity. Reflect on what you did, what you learned, and how you can improve.

John Dewey

The safety side, which is not nothing

CBD is well tolerated in these trials, and that is a real finding. It is also not the same as harmless.

The FDA-approved CBD medicine carries a label showing the incidence of ALT elevations above 3 times the upper limit of normal (ULN) was 13% (10 and 20 mg/kg/day dosages) and 12% (25 mg/kg/day dosage) in EPIDIOLEX-treated patients compared with 1% in patients on placebo. Most of those elevations occurred alongside valproate, and the rate reached 30 percent in patients taking both valproate and clobazam.

Those two drugs turn up frequently in autistic children with co-occurring epilepsy or severe behavioral symptoms, which makes the interaction directly relevant. Somnolence ran 32 percent overall, rising to 46 percent with clobazam.

One further gap worth naming: the large Israeli crossover trial did not test liver enzymes at all. The autism literature has not yet been sized to detect rare hepatic harm.

What is legal, and what is being sold

The legal picture and the retail picture point in different directions. Autism is a qualifying condition for medical cannabis in a number of states, including Pennsylvania, Illinois, Michigan, Minnesota and Texas, whose statute lists amyotrophic lateral sclerosis, autism, cancer, epilepsy, an incurable neurodegenerative disease, multiple sclerosis, post-traumatic stress disorder, a seizure disorder, or spasticity. Georgia qualifies adults with autism spectrum disorder but requires severe autism for anyone under 18.

The unregulated CBD market is a different matter. The FDA has sent warning letters over exactly these claims, one of which quoted a seller stating that possible uses for CBD include helping with skin problems, such as acne, autism, ADHD, and even cancer and found the products to be unapproved new drugs.

The gap between the two is enormous. State programs supply tested product at measured doses under medical supervision. A supermarket CBD oil is a supplement of uncertain content making claims the FDA has already prosecuted.

Where a seed company does not belong

Here is the section where our industry usually inserts itself, and where the honest answer is that we should not.

Nothing in our catalog is a therapy for a child. The trials above used pharmaceutical-grade extracts at doses calculated per kilogram of body weight, monitored by clinicians, in formulations manufactured to a standard no flower can match. Our plants are adult recreational products, bred for potency, terpene profile and grow performance. Those are different objects with different purposes, and blurring them would be the exact thing the FDA warning letters exist to stop.

What we can say plainly is what our seeds are. Papaya Frosting is Papayamosa crossed with Banana Frosting, a 70% indica at 28% THC, flowering in 60 to 65 days at a compact 100 to 110 cm, tropical papaya and mango over pepper and spice, mellow and relaxing. Jelly Cake is Biscotti crossed with Sunset Sherbet, an 80% indica at 25 to 27%, 56 to 63 days at 100 to 120 cm, candied kiwi and creamy vanilla over citrus and nutty cake, relaxing toward sleepy, and one of the earliest outdoor finishers we breed.

Four decades of breeding qualifies us to describe a plant accurately. It does not qualify us to advise a family, and anyone in this position should be talking to a pediatric neurologist instead of reading a seed company's blog.

The short version

About 1 in 31 US children is identified with autism, and the observational cannabis studies that drew parents to CBD were open-label with no control group, a fact their own authors state.

The controlled trials tell a harder story. The largest, a 150-participant placebo-controlled crossover, concluded that evidence for efficacy was mixed and insufficient, with whole-plant extract beating placebo on one co-primary measure and purified cannabinoids failing. Three of the four randomized trials to date missed their primary endpoint, and all documented substantial placebo responses.

CBD is generally well tolerated, but the FDA label for prescription CBD shows liver enzyme elevations at 13 percent against 1 percent on placebo, concentrated in patients also taking valproate and clobazam. Autism qualifies for medical cannabis programs in several states, while the FDA has issued warning letters over autism claims made for retail CBD. This is a conversation for a pediatric specialist.

Barney's Farm has been developing premium cannabis genetics since the 1980s, with over 40 Cannabis Cup wins. Explore our full seed catalog and find strains bred for every climate and skill level.