Could the Gut Microbiome Be Linked to Autism? Inside the Fierce Debate
For more than two decades, researchers have searched for any connections between gut bacteria and autism. What they've found is intriguing—but inconsistent.

Jim Adams’ daughter was nearly three years old when she was diagnosed with autism, similarly to other children suspected to have the neurodevelopmental condition: by clinicians observing how she played, spoke, and made eye contact. There was no blood draw, no brain scan, and no confirmation on a lab report, as there are for other diagnosable conditions.
“We didn’t know what caused autism,” says Adams, now the director of the Autism/Asperger’s Research Program at Arizona State University (ASU), as well as the founder and president of the Autism Diagnostics Lab. His search for answers led him on a nearly 30-year quest, culminating in the recent announcement of what he considers the first biological screening test for autism that comes from the Autism Diagnostics Lab, which measures metabolites in urine that are produced by the gut microbiome.
The development of the urine test rests on over two decades of research, including Adams's own, suggesting that, for the most part, autistic people who have been studied have a distinct gut microbiome compared with neurotypical people. The link has led some scientists to theorize that gut bacteria could actually be driving symptoms of the condition, a pathway researchers are also exploring in conditions ranging from depression to Parkinson's disease—so autism is no newcomer to the field.
Although genetics remains the dominant known cause of autism, researchers agree that some non-genetic factors—including prenatal pathogen exposure, air quality, and parental age—likely contribute to a person’s susceptibility. But whether the gut microbiome is one of them is where scientists part ways.
“To my mind, [it’s] quite speculative,” says Kevin Mitchell, an associate professor of developmental neurobiology at Trinity College Dublin. “The microbiome is in that pool of environmental factors where it has a bit more scientific hypothesis, it’s not a pseudoscientific hypothesis.” But in the last decade of studies Mitchell has seen tying the gut microbiome with autism, “I've just never seen any one of them that looked at all convincing.”
Interest in the gut microbiome’s potential connection to autism dates back to the early 2000s, when a small study suggested oral vancomycin, an antibiotic, temporarily improved communication and behavioral issues in some autistic children. The children had bowel problems and had been previously treated with broad-spectrum antibiotics; the researchers theorized that their guts were maybe overrun with "neurotoxin-producing bacteria." Clearing those bacteria out with another antibiotic, the thinking went, might help stabilize the gut flora.
The study was controversial at the time, as the improvements were short-lived and the findings considered “far too weak to back antibiotic use” in autism, as one Johns Hopkins neurologist said at the time. There were also concerns desperate parents would start pressing their doctors for a prescription.
The study was not proof that gut bacteria can cause autism. But the biological mechanism underlying that study isn’t without merit. The gut and brain communicate with each other through a network known as the gut-brain axis, says Alessio Fasano, a professor of pediatrics and nutrition at Harvard Medical School who studies autism and the gut microbiome. The axis uses molecules transmitted or produced by the nervous system, the immune system, and gut microbes to carry those signals. They include neurotransmitters such as serotonin and dopamine, chemicals released during inflammation, and metabolites, such as short-chain fatty acids, that gut microbes make when breaking down food or medications.
Fasano points to research comparing the gut microbiomes of autistic children with those of their neurotypical peers. Some of those studies have found that autistic children tend to have more of certain bacteria, such as Bacteroidetes and Clostridium, and fewer of others, such as Coprococcus and Bifidobacterium, which are believed to benefit both cognition and general health. Coprococcus, for instance, produces a short-chain fatty acid associated with anti-inflammatory effects on the brain.
(Scientists Are Starting to Understand How Autism and ADHD Can Overlap.)
In mouse studies, where animals are given gut microbes from stool samples of autistic people, Bacteroides has been shown to affect social behavior and induce repetitive behavior. And when pregnant rodents are exposed to chemicals, such as valproic acid—known to cause autism-like neurological changes in their offspring—the animals seem to develop gut microbiomes and behaviors resembling autistic symptoms.
But the findings are by no means definitive, say Fasano and Sarkis Mazmanian, a microbiologist at the California Institute of Technology, even though both experts believe that the microbiome may be a factor in autism. Comparison studies are only observations, Fasano says, and show correlation rather than cause. Mouse models, while important to preclinical research, are not a perfect proxy for autism in humans.
To Mitchell, the bigger problem is that studies looking for a link haven't adequately accounted for the wide variation within both autism and the gut microbiome. Autism is known as a “spectrum” condition because its core characteristics—including communication, behaviors, interests, and sensory processing—show up differently from person to person. Tying that variation, both within an individual and across people, back to the gut microbiome is a herculean effort, he says.
In 2025, Mitchell, Dorothy Bishop, a professor of developmental neuropsychology at the University of Oxford in England, and Darren Dahly, a statistician at the University College Cork in Ireland, examined more than 30 studies on autism and the gut microbiome. These studies included animal experiments, observational research, and clinical trials. The three researchers concluded that weak designs, flawed statistics, and conceptual problems undermined the claims. Mitchell notes that many of the studies paired few subjects with many variables, calling the underlying analysis “guaranteed to give you false positives.”
Some scientists also point out that a link between the gut microbiome and autism doesn’t mean gut bacteria is necessarily what’s driving the condition. Sonia Ballal, a pediatric gastroenterologist at Boston Children’s Hospital who cares for children with neurodevelopmental conditions, says food selectivity and sensory sensitivities can drive some of the issues seen in autism.
“Some kids are very extreme about [their diet],” she says. “They will only eat beige foods or chicken nuggets from a certain brand, served on a certain plate. So you create this strict and narrow diet, and as a result of that, you can have things like micronutrient deficiencies.”
A 2014 meta-analysis found children with autism had more than four times the odds of general GI symptoms compared with other children, including more diarrhea, constipation, and abdominal pain. Untreated GI distress in these kids has been linked to worsened sleep, as well as behavioral and psychiatric problems, though the associations don’t establish which condition causes the other.
“Sometimes the GI problems, when they’re worse, can make somebody’s autism-related behaviors worse,” Ballal says. “That we know for sure.”
The lack of consensus hasn't stopped scientists from exploring the gut microbiome as a target for autistic symptoms, though so far with little success.
Probiotics are one avenue of exploration. A 2016 mouse study found a Lactobacillus strain reduced some autism-like behaviors in mice, but human probiotic trials have been mixed. A 2025 randomized trial of 180 children with autism reported improvements in GI and behavioral scores, while a smaller, placebo-controlled Spanish trial the same year found no meaningful effect on core autism traits. In their 2025 review, Mitchell and his co-authors found little support for probiotics across nine meta-analyses.
Scientists have also tested transferring donor gut bacteria into recipients—a more intense approach that’s sometimes called a fecal microbiotic transplant. Adams and his colleague Rosa Krajmalnik-Brown, director of ASU’s Biodesign Center for Health, used a modified version they call a microbiota transfer therapy. In a 2017 open-label trial, in which participants know what treatment they are receiving, 18 autistic children aged 7 to 16 with significant GI problems received vancomycin, a bowel cleanse, and an initial high-dose infusion of screened donor microbes, then smaller daily doses for seven to eight weeks alongside an acid-reducing drug.
Adams and Krajmalnik-Brown say the trial saw major reductions in GI symptoms and gradual improvements on autism-trait assessments, with most holding two years later. But the trial had no placebo control, families knew the participant’s treatment, and many had changed medications, diets, or supplements before follow-up—all limiting how much can be concluded.
Timothy Buie, co-director of the GI-Autism Clinic at Boston Children’s Hospital, is skeptical of ASU’s fecal microbiota transplant, and similar efforts like it, for autism.
“I think they are way ahead of themselves to suggest that they’ve got a nice probiotic mix that’s just going to fix everything,” says Buie. “I’m not sure me giving you a transplant of someone else’s [microbiome] makeup is going to make you better with your health. We definitely see environmental changes, but I think we’re just not at the place to say, hey, you need a little Akkermansia or a little more Bacteroides, or you’ve got too much Pseudomonas.”
Mazmanian acknowledges that scientific missteps have held back efforts to pin down a link between the gut microbiome and autism. He still believes the gut microbiome likely contributes to autistic behaviors in some people, possibly those with genetic susceptibility. Truly understanding the link, he says, will require larger studies that approach the gut microbiome-autism question differently, such as examining how the gut interacts with an individual’s genetic risk for autism and identifying the specific signals, whether proteins or other molecules made by gut bacteria, that may contribute to the condition.
(Scientists Create Most Detailed Map Yet of Proteins Linked to Autism.)
Fasano led one such large-scale study, which ended in December 2025 and tracked how the microbiome changes over time in children at high risk of autism. Called the Genome, Environment, Microbiome and Metabolome in Autism (GEMMA), it followed infants from roughly 300 families in the U.S., Ireland, and Italy over the course of 10 years who already had an older sibling with autism. Studies have found that younger siblings in such families are more likely to develop the condition than children without autistic siblings.
The goal was to see how the microbiome shifts in early life, before any diagnosis. The researchers collected blood, urine, stool, and saliva samples every six months from infancy until age three. Fasano and his collaborators are still analyzing the data and have yet to publish results.
In Hong Kong, gastroenterologist Siew Ng of the Chinese University of Hong Kong is running a similar effort, launched in 2019, following families in Hong Kong and mainland China to assess how early-life environmental exposures can affect a child's health, including the risk of developing autism.
Unlike GEMMA, the study draws from the general population rather than recruiting high-risk families, and it collects samples such as stool, blood, and urine from both parents and children. In a recent study using data from the project, Ng’s team found an interplay between an infant’s epigenome, the collection of chemical modifications on DNA that regulates how and when genes are expressed, and the infant’s gut microbiome.
Babies born by cesarean section showed differences in epigenetic modifications to genes involved in neural development. Children with certain epigenetic modifications to those genes scored higher on measures of autism at age 3. But early colonization of the infant’s gut by Lachnospira pectinoschiza, a member of the Lachnospiraceae family, which is known for making short-chain fatty acids, appeared to lessen that effect.
While that study identified specific bacteria, Ng says her lab’s focus has been on what the gut microbes of autistic people are doing metabolically, such as which metabolites they make, rather than simply which ones are present. "I'm trying to emphasize the importance of functional pathways rather than what [group or] name of bacteria," she says, "because it's like a forest inside of your body."
For Adams, now the father of an adult daughter with autism, the work remains personal. The urine test and microbiota transfer therapy he and his colleagues at ASU developed may not be definitive. But he finds them to be encouraging steps in a field still trying to understand whether, and how, the gut microbiome is tied to autism—research he and other scientists hope may lead to breakthroughs for families living with autism.
“We’re not curing,” Adams said. “But are we helping? Absolutely.”