Autism spectrum disorder (ASD) is a neurodevelopmental condition that varies greatly in terms of symptoms and severity, making it difficult to diagnose and treat. A child with ASD cannot be identified at birth; it’s only as the child gains new skills over time that the development of those skills (especially social skills) appears markedly different from neurotypical children. Diagnosis is based on both deficits in social communication and interaction, and restricted or repetitive behaviors. While the previous decades of scientific research have focused almost exclusively on the genetic basis of autism, the latest research is exploring the hypothesis that  the gut microbiome—the community of microorganisms living in the human digestive tract and enabling important digestive and immune functions—might contribute to ASD through the gut-brain axis.

Children with ASD tend to have different gut microbes from neurotypical children, but it’s not clear if confounding factors are responsible for this. Multiple studies have now shown that different and less diverse communities of microbes live in the guts of children with ASD. But not only do the studies find inconsistency in exactly which microbes are present, but also confounding factors are a major issue: gastrointestinal symptoms and distinct dietary preferences are also likely to alter the gut microbiota in kids with ASD. For comparative studies, scientists struggle to find neurotypical and ASD children who are matched on every parameter, from antibiotic history to habitual diet to frequency of constipation, in order to assess whether it’s really the ASD that could account for any gut microbiome differences.

Individual observations need to be understood and tested in larger groups of children with ASD. By now there are several high-profile anecdotes of parents observing dramatic changes in the behavior of their children with ASD in relation to a gut-focused intervention: for instance, in one published case report, the father of a child with ASD says he observed a dramatic change in some of his son’s autistic behaviors after the child began a course of amoxicillin, a common antibiotic, for strep throat. It’s clear that chronic antibiotic administration is not a good idea, but it’s tempting to interpret these individual reports as proof that gut-related interventions are promising. However, the idea cannot be confirmed unless researchers develop a hypothesis for exactly why this happened and test it in a larger group of children with ASD.

Animal studies help us understand the biology at play, but animal behavior is hard to compare to human behavior. So far, most of the exciting research on ASDs and gut microbiota has been carried out in animal models: for example, one of the earliest relevant findings was that germ-free mice show altered social behavior. And a study showed young mice with socially different behaviors (increased repetitive burying of marbles; reduced vocalizing in response to other mice) who received the human-derived bacteria Bacteroides fragilis corrected many of these behaviors, making them behave more normally. Perhaps the most talked-about work was a study that showed transferring gut microbiota from humans with ASD into germ-free mice was enough to alter the social behaviors of the recipient mice, and that treating the mice with specific microbially-produced molecules corrected these behavioral differences. While all of these data points lend support to the idea of a causal relationship between gut microbiota and ASDs, the animal studies always have somewhat limited applicability to humans—especially because no animal model has a complex enough social life to fully recapitulate all features of human ASDs.

Children with ASD differ in so many ways that it may not be appropriate to group them all under the same label. In the big picture, a high priority for researchers is to identify some kind of biological or psychological marker for ASD because they’re not certain at present whether ASD is a single condition resulting in altered behaviors, or whether it’s a "collection of more or less similar phenotypes with multiple, varying etiologies". The behaviors and abilities and physical parameters of children with ASD vary widely, but if scientists were able to identify sub-groups based on a biological feature it could allow them to test the effects of different treatments targeting the relevant biology. It could be that only specific sub-groups of children with ASD would end up responding to a gut-focused treatment.

Gut-targeted treatments for ASD currently do not have good scientific support. A systematic review showed very limited value for both probiotics and prebiotics—which are common gut microbiota modulation strategies—for children with ASD. And while one fecal transplant study with a two-year follow-up gained a lot of attention, the study had no control group and failed to rule out a host of confounding factors; the lead investigator even reported that with the current level of evidence she wouldn’t give the treatment to her own child. Currently, for safety reasons, experts caution against giving fecal transplants to children with ASD except as part of a clinical trial.