A large government analysis of Minnesota’s medical cannabis program has found that many patients with autism reported meaningful reductions in anxiety, depression, sleep problems and several other symptoms after beginning medical marijuana treatment. The research, released by the Minnesota Office of Cannabis Management, examined 1,811 children and adults who enrolled in the state program with autism spectrum disorder as a qualifying condition between August 2018 and July 2024. Of those patients, 1,645 went on to purchase medical cannabis, giving researchers an unusually large real-world population for studying cannabis use among autistic patients.
The findings are particularly significant because anxiety and depression were common when patients entered the program. Researchers found that 88.3 percent of cannabis-purchasing patients reported moderate-to-severe anxiety at enrollment, while 59.8 percent reported moderate-to-severe depression. Substantial percentages subsequently reported clinically meaningful improvements, with some maintaining those improvements for months. The results do not prove that cannabis caused the changes because the research did not include a placebo or untreated control group, but the size of the study adds important new evidence to a field that until recently has relied heavily on small clinical trials and observational studies.
Anxiety Was the Most Common Symptom Among Patients With Autism
Anxiety was by far the most prevalent symptom measured in the Minnesota population. Of the 1,645 patients who purchased medical cannabis, 1,450 entered the program with anxiety scores considered moderate to severe. Patients participating in the program were required to complete self-evaluations that rated eight symptoms on a numerical scale from zero to 10. Those symptoms included anxiety, depression, disturbed sleep, fatigue, pain, nausea, vomiting and lack of appetite. Researchers considered a reduction of at least 30 percent from a patient’s baseline score to represent a clinically meaningful improvement.
Among patients who started with moderate-to-severe anxiety, 57 percent experienced at least a 30 percent reduction in their anxiety score within four months of making their first medical cannabis purchase. Researchers were also interested in whether improvements lasted rather than appearing temporarily. Among patients with adequate follow-up information after initially improving, 67.4 percent maintained their reduction for at least another four months. Overall, 34.3 percent of all patients who initially reported moderate-to-severe anxiety both achieved a reduction of at least 30 percent and maintained that improvement for at least four additional months. The pattern was remarkably similar between younger and older patients, with 33.5 percent of pediatric patients and 35.1 percent of adults achieving and maintaining meaningful anxiety reductions.
Depression Showed Some of the Strongest Improvements
Depression was another frequently reported problem among patients entering Minnesota’s medical cannabis program. Of the patients included in the analysis, 982—or 59.8 percent—reported moderate-to-severe depression at baseline. Depression was more common among the adult participants, but improvement after patients started purchasing cannabis was observed across age groups. Among those who initially had moderate-to-severe depression, 63.4 percent reported at least a 30 percent reduction in their depression score within the first four months.
The durability of those reductions was also notable. Among patients with follow-up information after initially improving, nearly three-quarters maintained their improvement for at least another four months. Overall, approximately 41 percent of patients who began with moderate-to-severe depression achieved and maintained a clinically meaningful reduction. Pediatric and adult patients had nearly identical sustained improvement rates. These findings are important because depression and anxiety can significantly affect functioning and quality of life in autistic people, yet the Minnesota data should not be interpreted as demonstrating that cannabis treats autism itself. Instead, the study examined symptoms and conditions that commonly occur alongside autism.
Sleep Problems and Fatigue Also Declined
Mental health symptoms were not the only areas where participants reported improvement. Disturbed sleep was present at moderate-to-severe levels in 70.2 percent of the patients when they entered the program. Among those patients, 60.7 percent reported at least a 30 percent reduction in sleep-related symptoms within four months of beginning medical cannabis use. Ultimately, 37.3 percent of patients with significant sleep problems both reached the improvement threshold and maintained that improvement for at least four months. Adults appeared somewhat more likely than pediatric patients to achieve sustained improvement in disturbed sleep.
Fatigue followed a similar pattern. More than half of the patients—56.2 percent—reported moderate-to-severe fatigue when entering the program. Approximately 57 percent of those patients initially improved by at least 30 percent, while 34.4 percent achieved and maintained the improvement. Other symptoms tracked by researchers, including nausea, pain, vomiting and lack of appetite, also improved among portions of the study population. Because these symptoms can interact with one another, improvements in areas such as sleep or pain could potentially influence anxiety, mood or daily functioning as well. The observational design, however, does not allow researchers to determine which changes were directly attributable to cannabis.
Researchers Also Examined Autism-Associated Behaviors
Minnesota researchers went beyond general symptoms by examining a smaller group of participants using the Aberrant Behavior Checklist Second Edition, commonly known as the ABC-2. The questionnaire measures several behavioral domains frequently evaluated in autism research, including irritability, hyperactivity, social withdrawal or lethargy, repetitive behaviors and inappropriate speech. Participants younger than 26 were invited to complete the questionnaire when entering the program and again approximately 10 weeks later.
Only 46 patients completed both assessments, making this portion of the study much smaller than the main symptom analysis. Among those participants, scores decreased significantly across all five behavioral categories after medical cannabis treatment began. Lower scores represented a reduction in the presence or severity of the measured behaviors. Researchers cautioned that the finding should be interpreted carefully because the 46 participants represented only a small fraction of those eligible to complete the questionnaire. Patients who believed cannabis was helping may also have been more likely to complete follow-up assessments, potentially exaggerating the apparent benefit. Even so, the results add another area of interest for future controlled studies.
Medical Cannabis Use Varied Dramatically by Age
One of the most revealing parts of the Minnesota analysis involved the types of cannabis patients actually purchased. Researchers examined 18,102 transactions involving more than 47,000 individual medical cannabis products during patients’ first year in the program. There was no single standardized cannabis treatment. Participants purchased products with widely varying concentrations and ratios of THC and CBD, including oral solutions, capsules, gummies, vapor products and cannabis flower.
Age produced especially large differences. Among patients younger than 18, 85.3 percent of purchased products were taken through the digestive system, including gummies, capsules and oral solutions. Younger patients were also much more likely to use CBD-dominant or relatively balanced cannabinoid formulations. Adults followed a very different pattern. Approximately 68 percent of products purchased by adult patients were inhaled products such as flower or vapor formulations, and high-THC products were substantially more common. These differences make it impossible to identify one particular cannabis product, THC level, CBD concentration or cannabinoid ratio as responsible for the improvements seen in the overall population.
Controlled Cannabis Studies in Autism Have Produced Mixed Results
The Minnesota findings join a growing body of clinical research examining cannabinoids in people with autism. One of the best-known studies was published by Adi Aran and colleagues in Molecular Autism in 2021. The randomized, double-blind, placebo-controlled study involved 150 children and young adults between the ages of five and 21. Researchers compared placebo with two CBD-rich cannabinoid preparations containing CBD and THC in a 20-to-1 ratio. A whole-plant extract improved disruptive behavior on one of the study’s two primary outcome measures and also produced improvement on a secondary measure of autism-related symptoms, although the second primary behavioral measure did not show a significant advantage.
Other controlled research has produced similarly complicated results. A 2024 randomized, double-blind, placebo-controlled trial involving 60 children between five and 11 years old examined a CBD-rich cannabis extract and reported significant improvements in anxiety, social interaction and psychomotor agitation. Another randomized crossover study involving 29 autistic children found that CBD did not significantly improve its primary measure of social responsiveness compared with placebo, although secondary measures suggested reductions in anxiety and improvements in some aspects of social functioning. Taken together, the studies suggest that cannabinoids may affect certain symptoms without necessarily producing broad improvements across every measurement used to evaluate autism.
Not Every Trial Has Found Cannabis Superior to Placebo
Research published in 2025 provided another reminder of why placebo-controlled studies remain essential. Doris Trauner and colleagues studied purified pharmaceutical cannabidiol in autistic boys between seven and 14 years old who experienced significant behavioral difficulties. Thirty-nine participants were included in the final analysis of the double-blind crossover trial. Children received both CBD and placebo during separate treatment periods, allowing researchers to compare responses within the same participants.
Both CBD and placebo were associated with improvements on several measurements, but researchers did not find a statistically significant advantage for CBD on the trial’s primary standardized behavioral outcomes. The result demonstrated the importance of placebo effects, caregiver expectations and natural variations in behavior when evaluating treatments for autism. Similarly, a 2022 analysis of sleep outcomes from the earlier 150-participant Aran trial found that CBD-rich cannabinoid treatments were not superior to placebo for measures including bedtime resistance, delayed sleep onset and sleep duration. These findings do not eliminate the possibility of benefits for particular symptoms or subgroups, but they show why observational improvement alone cannot establish treatment effectiveness.
New Research in Autistic Adults Supports Further Investigation
Cannabis research involving autism has historically focused heavily on children and adolescents, but newer research is beginning to include more adults. A 2026 analysis of the UK Medical Cannabis Registry examined 130 autistic adults who had been prescribed cannabis-based medicinal products. Researchers followed patient-reported outcomes involving anxiety, sleep and health-related quality of life for as long as 18 months. Anxiety scores, sleep quality and measures of overall quality of life improved after treatment began, with some changes persisting during long-term follow-up.
Like the Minnesota study, however, the UK research was observational. There was no placebo group, patients knew they were receiving cannabis-based treatment and other changes in health or treatment could have contributed to the results. The similarity between the two datasets is nevertheless noteworthy. Two separate real-world populations have now produced associations between medical cannabis use and improvements in anxiety and other quality-of-life measures among autistic adults. Controlled clinical trials will be needed to determine whether those associations translate into a genuine therapeutic effect and which cannabinoid formulations are most appropriate.
Side Effects Were Usually Mild but Still Important
The Minnesota report also examined adverse effects during patients’ first year of cannabis use. About one in five patients reported at least one side effect. Nearly three-quarters of all reported side effects were classified as mild, while approximately 5.8 percent were considered severe. Dry mouth was the most commonly reported adverse effect overall, followed by problems including increased appetite, drowsiness, mental clouding and fatigue.
Anxiety provides an especially important example of cannabis producing different effects in different individuals. Although anxiety was the symptom that affected the largest number of patients and many participants experienced meaningful reductions, anxiety was also the most frequently reported severe side effect. Panic attacks were reported as well. THC can increase anxiety in some people, particularly at higher doses or in susceptible individuals, while CBD has been investigated for potentially different effects on anxiety. Because Minnesota patients used products containing widely different amounts of both cannabinoids, the study cannot establish which formulations offered the best balance between symptom improvement and adverse effects.
Why the Minnesota Study Matters
The scale of the Minnesota analysis distinguishes it from most previous cannabis research involving autism. Randomized clinical trials have generally included dozens or roughly 150 participants, while the state registry captured information from more than 1,600 patients who actually purchased medical cannabis. That allowed researchers to observe product preferences, symptom patterns, sustained improvement and adverse effects under normal medical cannabis program conditions rather than within the highly controlled environment of a clinical trial.
At the same time, a large observational study cannot answer the same questions as a randomized controlled trial. Minnesota researchers did not compare cannabis patients with an untreated or placebo group, and the analysis could not fully account for medications, behavioral therapy, psychotherapy or other treatments being used at the same time. Patients or caregivers supplied many of the symptom ratings, and researchers tracked cannabis purchases rather than confirming exactly how much of each product patients consumed. These limitations mean the study demonstrates an association between medical cannabis use and symptom improvement rather than proving that cannabis caused the improvements.
What Comes Next for Cannabis and Autism Research
The Minnesota findings provide a stronger foundation for future clinical trials designed to determine which autistic patients, if any, are most likely to benefit from cannabinoid treatment. Research will need to distinguish THC from CBD, compare different ratios and doses, evaluate oral and inhaled products separately and track patients for longer periods. Studies should also examine specific treatment targets rather than treating autism as a single symptom. Reducing severe anxiety, improving sleep or treating co-occurring depression represents a very different medical goal from attempting to change the core characteristics of autism.
For now, the evidence remains encouraging in some areas but incomplete. Minnesota’s unusually large dataset found that significant portions of autistic patients experienced sustained reductions in anxiety, depression, sleep disturbance and fatigue after beginning medical cannabis. Controlled trials have also identified possible improvements in anxiety and certain behavioral or social measures, while other studies have found little difference between cannabinoids and placebo. The emerging picture is therefore more complex than either calling cannabis an autism treatment or dismissing the reported benefits entirely. Medical cannabis may ultimately prove useful for selected patients experiencing specific co-occurring symptoms, but determining which cannabinoids work, who is most likely to benefit and how treatment can be delivered safely will require larger and more rigorous clinical trials.
