It’s morning in Nahariya, a tiny Israeli town near the Lebanese border, and 4-year-old Benjamin is repeatedly smashing his head against the wall. He spins wildly in circles, screeching at full volume. As his mother tries frantically to calm him, he pulls down his pants and defecates on the floor.

When they leave their apartment, Benjamin wrestles free of her hand and nearly runs into oncoming traffic. Sharon attempts a trip to the supermarket but leaves before she finishes shopping because her son is screaming while he picks up items and throws them to the floor.

That was in October 2016, and typical of most days at the time. Sharon, a single mother who moved to Israel from the United States one year earlier, was alone and losing control. Benjamin was taking Ritalin, a drug usually associated with attention deficit hyperactivity disorder (ADHD), which he did not have. He’d also tried the antipsychotic ziprasidone and a mix of antidepressants and anti-anxiety drugs. None of them helped, and he often became more hyperactive as they wore off.

All that changed a year ago, when Benjamin started taking marijuana. In the little apartment he shares with his mother, mornings are now relaxed and orderly. His transformation may signal the arrival of a long-awaited and desperately needed healing for the many others just like him: children living with severe autism.

Autism spectrum disorder affects about 1 percent of children around the world, with disproportionately high rates in developed countries. In the United States, the Centers for Disease Control and Prevention estimates that one in 68 children has been identified as having ASD, a wide-cast net of a diagnosis that encompasses several complex brain disorders that make communication and other interactions difficult. Children with milder “high-functioning” autism are often uninterested in making friends, feel uncomfortable when touched and have a hard time making eye contact or reading social cues. These individuals face challenges but can usually navigate building a life within their society.

But in cases of severe “low-functioning” autism like Benjamin’s, the symptoms are more pronounced and often violent. Children engage in repetitive and sometimes harmful behaviors, like rocking and head-banging, and are hypersensitive to sound and light, with exposure often triggering tantrum-like meltdowns. They can’t sleep. They have rages. Some of these children never learn to speak, or they reach their teen years uttering only a few words.

ASD has no cure, and most children’s symptoms are treated with medications approved for depression, anxiety or ADHD. As was the case with Benjamin, these drugs often cause their own set of obsessive behaviors and insomnia, along with weight gain. For many children with severe autism, the drugs help for just a few hours; once they wear off, symptoms like hyperactivity become even more extreme.

A Last Option

By October 2016, Sharon was desperate. She gave birth to Benjamin alone and chose to move to Israel to be part of a close-knit community. But his condition was isolating. She was lost, alone, exhausted, frustrated. More than all that, she was sad for her child.

One day, while filling Benjamin’s regular prescription for Ritalin, Sharon vented about the side effects to the pharmacist. He responded with a surprising suggestion. She should contact Dr. Adi Aran, he said, a pediatric neurologist in Jerusalem who had begun experimenting with medical cannabis as a treatment for children like Benjamin.

Sharon balked at first. Benjamin wasn’t even in first grade. And wasn’t marijuana a dangerous and illegal drug? But at the same time, he was living in a home filled with locks and padded furniture. She couldn’t go shopping or visit friends without worrying that her child would become violent. She had tried every option that conventional medicine had to offer. In other words, she had nothing to lose.

In Israel, cannabis use is legal in a small number of medical cases, such as epilepsy, severe chronic pain and certain forms of cancer. Aran, who directs the pediatric neurology unit at Jerusalem’s Shaare Zedek hospital, had been recommending cannabis for some of the epileptic children he treated. At the end of 2015, he began an informal study of medical cannabis for severe pediatric autism.

In online forums for parents of children with ASD, Sharon read about some new forms of cannabis that were created specifically for young children. She watched an Israeli documentary from that year that showed children with ASD transformed by medical marijuana. The more she learned, the more determined she became to enroll Benjamin in the Israeli study. She sent Aran an email begging him to consider treating him, and he invited the mother and son to Jerusalem.

It was a fraught trip. Sharon doesn’t own a car, and the five-hour journey, by bus and train, included several violent outbursts and meltdowns. At the hospital, Aran reviewed Benjamin’s medical history and observed his behavior. Seeing the severity of his symptoms and the long list of medications that had already been tried without success, the doctor agreed that he was a good candidate. Sharon was sent home with a prescription for an oil made from a specially calibrated strain of Israeli cannabis, along with paperwork to chart her son’s progress.

Autism and Epilepsy: The Overlap

Aran’s colleagues in the global pediatric community were still calling for caution in 2015. The American Academy of Pediatrics, which is staunchly opposed to legalization of marijuana, had just issued a policy statement opposing medical marijuana outside the regulatory process of the U.S. Food and Drug Administration (FDA). (The AAP still maintains that stance.)

But Aran was starting to see evidence in his favor. His first inkling that cannabis could work for autistic kids came from anecdotal reports of parents who had used the drug to treat children with epilepsy. The rationale behind the treatment, and the reason it worked, came down to the marijuana plant’s two primary chemicals: the psychoactive agent tetrahydrocannabinol (THC) and the antipsychotic cannabidiol (CBD).

The brain is filled with cannabinoid receptors, which are named after the plant and function like special locks to which THC is the key. When THC binds to cannabinoid receptors in the brain, several sensations flood the body, what marijuana users call “the high.”

CBD works differently, and often with opposite effects. It doesn’t bind directly to cannabinoid receptors, it’s not psychoactive, and it doesn’t alter how the brain functions. Instead, CBD interacts with the brain indirectly. That process, called modulation, combats psychosis, depression, inflammation, anxiety and depression. While it’s THC that gets people stoned—and poses a potential danger to immature brains—it’s the plant’s CBD that relaxes them and counters anxiety, making it relevant to epilepsy and autism.

A healthy human brain runs on a balance of excitation and inhibition, a push and pull that regulates information as it flows through the chemical synapses in our head. With excitation, cells fire, transmitting information and signals. Inhibition keeps that flow of traffic in check. Like high-pressure water flowing through a narrow hose, these two systems work together to distribute information without overloading the system.

People with epilepsy suffer from reduced inhibition, which causes seizures. Over the past five years, a handful of successful studies on the use of cannabis, all employing specialized strains with little to no THC, have shown CBD is a legitimate treatment for certain forms of severe pediatric epilepsy. Doctors believe the drug works because CBD increases inhibition, thus helping to prevent the firing of seizure-triggering neurotransmitters, the brain’s chemical messengers. And because CBD does not cause a high, it’s believed that it presents little risk to the developing brain of a child when administered on its own.

Israel has been at the forefront of medical marijuana research since modern health care began considering its merits. Raphael Mechoulam, who studies medicinal chemistry at Hebrew University in Jerusalem, first identified THC and CBD by studying 5 kilos of Lebanese hashish in the early 1960s. He was eager to unlock its chemical components in the same way that researchers had studied and mapped cocaine and heroin in the past, and in 1980 his research led to the very first trial on the use of CBD for epilepsy.

The results were promising, but the stigma of marijuana as a dangerous psychoactive drug was too strong to lead to immediate change. The U.S. Drug Enforcement Administration classifies it as a Schedule 1 controlled substance, meaning it’s considered addictive and unsafe and lacks medical use.

Still, Mechoulam and other scientists continued to quietly research CBD and its effects. Partly influenced by their work, which showed marijuana to be a potent pain reliever, California legalized medical cannabis in 1996, with a number of other states following. But it took an 8-year-old girl to convince the medical community that weed is a legitimate treatment for sick children.

Charlotte Figi, who lives in Colorado, has life-threatening epilepsy. Since infancy, she suffered up to 300 grand mal seizures a week. By the age of 5, her heart had stopped several times, and she couldn’t walk or eat on her own. In 2013, her desperate parents convinced a Denver doctor to prescribe cannabis oil for their daughter. The compound, a special strain of cannabis with a 20-to-1 ratio of CBD to THC, saved her life.

Charlotte is now 11. Every day, she takes two doses of cannabis oil, with that same 20-to-1 ratio, in her food. Her seizures have nearly ceased. She is healthy and thriving. Her recovery is so remarkable that a special high-CBD and low-THC strain of medical cannabis produced in Colorado was named Charlotte’s Web.

Last year, a London-based pharmaceutical company brought Epidiolex, a CBD-based drug, to the FDA for approval. In a study released last month, that drug helped slash epileptic seizures by 41 percent, compared with 14 percent among patients taking a placebo. Epidiolex could be approved by the FDA as early as this summer; if that happens, it will be the first time the agency has opened the regulatory gate to a marijuana-derived drug.

When Charlotte’s case came to light in 2013, Aran was one of a handful of neurologists prescribing cannabis to young people with epilepsy. But nearly one-third of children with autism also suffer from epilepsy. As Aran watched his epileptic patients suffer fewer seizures, he noticed that for those who were also autistic, repetitive behaviors, communication difficulties and frustrations with social interactions also improved. Case studies in medical journals across the world noted the same overlap.

“We [in the medical community] saw children with epilepsy and autism really improve, not just in their epilepsy but also in their behavior,” says Aran. “Sometimes, it was only the autism symptoms that improved.”

Aran, 47, was well versed in Mechoulam’s research on CBD and epilepsy, but he began to wonder: Could CBD work in cases where the patient suffered only from autism?

A Trial Run

The parents of his autism patients read online message boards and Facebook posts telling stories of how CBD worked across the epilepsy-autism overlap, and they hammered Aran to try cannabis on their children.

He spent two years hesitating. “At first, I didn’t think it was worth exploring,” he says, sitting at his cluttered desk in his modest Jerusalem office. “Yes, this form of severe autism is a real problem, and the patients and families and the education system are all suffering. But in medicine we have to be cautious.”

It’s the medical community’s role, he says, to protect patients from being swayed into false treatments, especially those that could prove harmful. He was curious about trying CBD for his severely autistic patients, but he wasn’t certain it was the right move ethically.

In Israel, a small country with informal customs, it’s typical for a parent to call a doctor’s personal cellphone to beg for a prescription. Gradually, after talking to dozens of persistent parents, Aran changed his mind, he says. In December 2015, he started the world’s first open-label study on the use of cannabis for pediatric autism, prescribing the drug to a few of his most severely affected children, ranging in age from 5 to 20, and charting and monitoring the results. Benjamin joined this study several months later.

Geography worked in Aran’s favor. In America, despite the legalization of marijuana in a number of states, possession of the drug is still a federal crime. Wide-scale research and cultivation is impossible for American marijuana growers, and the lack of federal regulation means doctors who wish to prescribe marijuana to patients in states like California and Colorado have little control over the product the patient receives from a dispensary.

But in Israel, a nascent medical marijuana industry is thriving. The country’s combination of year-round sunshine and high-tech resources puts it in a unique position to grow and manufacture a number of cannabis-based drugs.

The nation’s Ministry of Health sees Israel becoming a global leader in medical marijuana and has taken dramatic steps to make that happen. In 2016, Yaakov Litzman, an ultra-Orthodox rabbi then serving as the nation’s minister of health, allowed the dispensing of cannabis, similar to what’s done with all other medications. He also joined a commission that opened the Israeli medical marijuana market to export, a move with the potential to inject billions into the country.

Israel is now one of three nations, alongside Canada and the Netherlands, to have a government-sponsored cannabis program. With the global medical marijuana market now surpassing the $30 billion mark, this tiny country of only 8 million people is poised to gain a major chunk of that profit.

The cannabis Aran prescribes for autism and epilepsy is a special strain originally produced for epilepsy patients, with the 20-to-1 ratio of CBD to THC that worked so powerfully for Charlotte Figi. So far, he has prescribed it only for his most severe patients: children who had never responded to traditional autism medications, were mostly nonverbal and quite violent, and whose parents were desperate. He has no interest, he says, in prescribing cannabis to children who suffer from other subtypes of autism, like Asperger’s syndrome, that potentially respond to therapy or traditional drugs. Cannabis, he says, is a last resort.

Aran ultimately enrolled 60 children between the ages of 5 and 21 in that first study. He tracked the results of each patient for six months through a series of parent questionnaires and in-office visits.

A paper that will be published later this year in the journal Pediatrics summarizes the results. Most parents said their children improved from the treatment. Nearly half saw a marked reduction in the core symptoms of autism, and nearly a third said their children either started speaking for the first time or were communicating nonverbally. One child said, “I love you, Mom”—for the first time in his life.

As for Benjamin, within two weeks of filling the prescription from Aran, Sharon says, he was calmer. He responded when she spoke to him. He could sit still and make eye contact. If she took him with her to visit friends, she could sit with the adults drinking tea while he played quietly in the other room. Within months, he was doing so well that his teachers recommended he leave his special-needs school for a standard classroom. “It’s like a miracle. I can leave the house and go out with him and not worry,” says Sharon. “I can breathe.”

The Need for Data

And yet, despite what parents like Sharon were documenting, Aran remained cautious. Parents have touted plenty of other drugs as miracle autism cures over the years—and all have consistently failed to pan out under testing. And the design of his first trial allowed for the possibility that parents would document the changes in their children that they expected to see, a phenomenon known as confirmation bias.

His next step—a larger, double-blind, placebo-controlled study—would address these concerns. Launched in 2017, the trial is the first of its kind in the world. Aran and his team enrolled 150 new patients for the seven-month study. The children receive cannabis oil for three months and a placebo for three months, with a four-week period in between for the first treatment to taper off and the second to start up. Of the cannabis oil the children receive, there are two possible options: an oil made from the whole cannabis plant, including extracts and impurities, or an oil that is 20 parts pure CBD and one part pure THC. The results are expected later this year.

Because the study is blinded, Aran doesn’t know which patients have received which of the three options. But he admits it’s hard not to think the study is on the verge of a breakthrough. He has treated thousands of autistic patients—patients who break furniture, tear out their hair and cannot control their bodily functions. The changes he has seen since he began prescribing cannabis have shocked him.

“I keep seeing the same thing,” he says. He doesn’t know which children are taking the placebo, but the people in their lives seem to. Every time the dose tapers off, he receives calls from parents and schools asking for the child to be put back on the drug. “We decrease the treatment, and suddenly we have a crisis.”

David, a burly, 6-foot-1-inch 20-year-old who spent most of his appointments with Aran wearing physical restraints, became gentle enough to hug his sister. He smiled and said her name out loud for the first time in his life. Eitan, a nonverbal 11-year-old who was obese, after years of medications for his tantrums caused compulsive eating, has lost the weight. Aran can’t be sure yet that these changes are related to CBD, and any data will require an extensive peer review before publication. Still, he says, “something is working.”

The Autism Market

Doctors and parents aren’t the only ones eager to have the study blindfolds removed. The major Israeli cannabis companies, locked in a race for dominion over a market estimated to reach $4 billion in the coming year, are also watching.

Three companies have dominated the fast-growing Israeli cannabis market, each scrambling for dominion. Tikun Olam became the first medical cannabis supplier in Israel when it launched in 2005. Breath of Life is now the world’s largest production, research and development facility for medical cannabis, with 1 million square feet of cultivation fields and 30,000 square feet of labs and greenhouse space. Better is Israel’s only 100 percent organic medical cannabis company. All three are running their own private studies to produce data that will help sell their drugs. (Aran is currently using cannabis from Breath of Life, which created a 20-to-1 strain called Topaz specifically for this study and provided it for free.)

These companies have their eyes on the study in part because it will resolve a debate surrounding what’s known as the entourage effect. That term, first introduced by Mechoulam, refers to the idea that the many compounds in cannabis—THC, CBD and others—work together. Take any part out, and the plant fails to have an effect. Whether that effect is real is unknown. Aran’s trial, which includes an isolated-compound oil and a whole-plant oil, will answer that question.

Breath of Life is hedging its bets, growing whole-plant cannabis and doing laboratory work that extracts pure CBD for pharmaceuticals. Its goal is to bring a cannabis product to the FDA that can be approved for American children with autism. Once the data resolve which oil, if any, works, the company can focus its resources on that version exclusively.

“If the whole plant proves to be statistically significant over the clean version, that means there is an entourage effect, and we understand where we as researchers are going,” says Tamir Gedo, CEO of Breath of Life. “But either way, we are going to take it to the FDA.”

Research in the United States

Since Aran began his research, physicians in the United States have begun to warm up to cannabis as a potential autism treatment as well.

Dr. Eric Hollander, director of the Autism and Obsessive Compulsive Spectrum Program at New York’s Montefiore Medical Center, announced last year that he is running his own study on medical cannabis in pediatric autism—the first in the U.S. to explore cannabis for pediatric autism. “There’s a big unmet need,” he says.

But unlike in Aran’s study, Hollander’s patients are receiving a treatment that contains neither THC nor CBD. They will receive pure cannabidivarin, or CBDV, a cannabinoid derived from the cannabis plant that is very similar in chemical makeup to CBD. It also has a similar track record to CBD in terms of medical application. In studies among patients with epilepsy, CBDV has also been effective in reducing seizures.

The U.S. Department of Defense is funding Hollander’s study. After realizing that military families with autistic children struggle with relocation and placement in remote locations, the department began an active program on autism research. A cannabis grower based in London is providing the drug, administered in capsule form. The goal of his study is to see if CBDV alone can combat the core symptoms of autism.

Whatever the results, change is still a long way off. For American parents hoping to give cannabis to their severely autistic children, it could be several years before any autism drug, either from Israeli or American research, is readily available to them. In the United States, doctors in only three states—Georgia, Oregon and Pennsylvania—can legally prescribe cannabis to autistic children. A law on the books in Minnesota will grant access to doctors in July.

Hollander will start recruitment for his study in March 2018 but won’t have results for another three years. If they show a statistically significant difference between CBDV and the placebo in treating the symptoms of autism, he will likely run a Phase III study, a necessary step toward bringing the drug to market. At that rate, he says, it will be at least six years before the drug he is testing reaches the U.S. market.

But the global medical community is interconnected, and Aran’s study might put pressure on the FDA to speed up the process, or at least approve off-label applications for epilepsy CBD drugs like Epidiolex. Hollander believes if the data out of Jerusalem in late 2018 proves that transformations like Benjamin’s are not rare among patients with severe autism, and if the possible side effects are mild and manageable, some states might amend their medical marijuana laws to grant parents greater access.

In the meantime, Aran stresses the need to change the anti-marijuana stigma that still pervades American medicine and drug regulation. “Giving marijuana to children is unthinkable, but CBD is not marijuana,” he says. “It’s not a drug. It’s a medication.”