Before a child speaks a first word or misses a developmental milestone, clues about autism and ADHD risk may already be hiding in the gut. A wave of longitudinal studies, some tracking children from the delivery room into adolescence, has found that measurable differences in gut bacteria and DNA methylation appear in infancy, years before any behavioral diagnosis. The most recent of these, a large mother-child metabolomics study published in Nature Communications in 2025, traced metabolic trajectories from pregnancy through early childhood and linked them to ADHD and autism diagnoses at age 10, strengthening the case that biology in the earliest stages of life may shape neurodevelopmental risk.
What the major cohorts have found
The evidence draws from several independent studies, each designed to collect biological samples long before any child showed symptoms.
The ABIS cohort, a 20-year Swedish study of 16,440 children followed from birth, analyzed umbilical cord blood and stool samples taken at age one. Researchers at the University of Florida and Linköping University identified early-life microbiome signatures that differed between children who later received autism or ADHD diagnoses and those who did not. “We found that the microbiome of children who went on to develop autism or ADHD was significantly different already at one year of age,” said Eric Triplett, chair of the University of Florida’s Department of Microbiology and Cell Science, who co-led the analysis. The study also flagged antibiotic-treated ear infections during infancy as a correlate of disrupted microbial development.
A U.S.-based birth cohort called WHEALS added a complementary finding: infant gut bacteria sampled at one month and six months of age were associated with ADHD diagnosed a full decade later. Because the microbial differences appeared so early, the researchers argued they were unlikely to be artifacts of later diet changes or medication.
A separate prospective study of infants with and without a family history of autism tracked fecal microbiota and the metabolome across the first three years of life. Compositional and functional differences in gut bacteria appeared before or around the typical diagnostic window for autism spectrum disorder, reinforcing the idea that microbial shifts precede clinical recognition.
On the epigenetic side, the Early Autism Risk Longitudinal Investigation (EARLI) measured genome-scale DNA methylation across multiple tissues: maternal blood in early and late pregnancy, cord blood, and placental compartments. Children who went on to receive an autism diagnosis at 36 months already showed distinct methylation patterns at birth, concentrated in genes previously associated with autism. That these epigenetic signatures were detectable before any behavioral signs appeared is among the strongest evidence that biological changes precede the clinical condition rather than follow it.
The 2025 Nature Communications study tied these threads together. Using mediation analyses, the researchers identified maternal inflammatory proteins, maternal body mass index, and Western dietary patterns as potential bridges between metabolic disruption during pregnancy and later neurodevelopmental diagnoses in children. The implication: a mother’s own microbiome, diet, and metabolic health may shape her child’s risk through inflammatory and metabolic pathways that leave epigenetic traces detectable at birth.
Where the science hits a wall
The central gap is mechanistic. No human study has yet demonstrated the full causal chain from a specific gut bacterium in an infant to a specific epigenetic modification on a gene in the developing brain. The cohorts described above establish strong temporal associations, showing that microbial and methylation differences appear before diagnosis. But the step-by-step biology connecting microbe to methyl group to neural outcome has been mapped primarily in animal models, not in people.
A synthesis paper in the journal Gut Microbes outlined how the maternal microbiome and early-life exposures could interact with genetic susceptibility through microbial metabolism and epigenetic reprogramming. That framework is useful for generating hypotheses, but it remains a theoretical model rather than primary evidence.
One randomized controlled trial of early-life probiotic supplementation, led by Anna Pärtty and colleagues and published in Pediatric Research, did report that infants given Lactobacillus rhamnosus GG in the first six months of life had different fecal Bifidobacterium counts and, at 13-year follow-up, lower rates of ADHD and Asperger syndrome diagnoses compared to the placebo group. That trial is notable because it provides intervention-based evidence rather than just correlation. But it was a single study with a small sample, and no follow-up work has tracked epigenetic changes after a microbiome intervention through adolescence.
Competing explanations also complicate the picture. The ABIS cohort flagged antibiotic-treated ear infections, but it remains unclear whether the antibiotics disrupted the microbiome, whether the infections reflected an already altered immune system, or whether both were downstream effects of some other factor. Maternal diet, inflammation, and BMI all emerged as mediators in the Nature Communications study, but these variables are deeply intertwined, making it hard to isolate which one drives the most risk. Genetic susceptibility overlaps with environmental exposures, blurring the boundary between inherited vulnerability and modifiable influence.
Generalizability is another concern. Several of the most detailed cohorts draw from specific regions or high-risk families, such as those with an older sibling already diagnosed with autism. Their findings may not translate cleanly to more diverse populations with different diets, healthcare systems, or environmental exposures.
What this means for parents and clinicians
The strongest claims in this field rest on prospective designs. Studies like WHEALS, ABIS, and EARLI collected biological samples before any child showed symptoms, then followed participants for years until diagnoses were confirmed. That structure rules out reverse causation: the microbial and epigenetic differences were not caused by treatments or behavioral changes that came after a diagnosis. These are the studies that carry the most weight.
For parents weighing practical steps, the evidence supports paying close attention to the first 1,000 days of life, from conception through roughly age two, as a window when microbial, metabolic, and epigenetic pathways are especially malleable. That does not mean a single course of antibiotics or a particular food choice will determine a child’s trajectory. It does suggest that broader patterns, such as maternal metabolic health, judicious use of antibiotics in infancy, and the establishment of a diverse gut microbiome through breastfeeding and varied early nutrition, may shift probabilities at the population level.
Pediatricians and obstetricians are not yet issuing clinical guidelines based on microbiome or methylation screening, and for good reason: most infants who show a particular microbial or epigenetic pattern will not go on to develop autism or ADHD, and many children with these diagnoses will not share the same early signatures. The value of this research, at this stage, lies in refining risk models and guiding future intervention trials rather than offering families a definitive prediction.
Why diverse birth cohorts and randomized trials are the next frontier
Researchers say the next generation of informative studies will need to combine diverse birth cohorts, serial sampling of the microbiome, metabolome, and epigenome, careful tracking of environmental exposures, and, eventually, randomized interventions introduced during pregnancy or infancy. “We need studies that follow thousands of children from many different backgrounds, collecting samples at multiple time points, so we can see whether these early signals hold up across populations,” said Johnny Ludvigsson, the Linköping University pediatrician who initiated the ABIS cohort more than two decades ago. Only by weaving these strands together will the field be able to distinguish which early-life factors are causal, which are merely correlated, and which can be safely modified. For now, the emerging picture is one of cautious optimism: biology in the earliest weeks of life appears to matter for later neurodevelopment, but it does so in complex, probabilistic ways that resist simple stories of cause and cure.
This article reflects research available as of May 2026. It is not medical advice. Parents with concerns about their child’s development should consult a pediatrician or developmental specialist.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
