Do gut microbes cause autism? New research says diet plays the bigger role

By Priyanjana Pramanik, MSc.Reviewed by Susha Cheriyedath, M.Sc.Nov 5 2025

A new study reveals that in families with children on the autism spectrum, gut bacteria remain remarkably alike, but diets differ sharply, pointing to selective eating behaviors, not microbial imbalance, as the stronger link to autism traits.

Study: Dietary Patterns, Not Gut Microbiome Composition, Are Associated with Behavioral Challenges in Children with Autism: An Observational Study. Image Credit: QINQIE99 / Shutterstock

In a recent article in the journal Nutrients, researchers examined the differences in dietary habits and gut microbiota between children with autism spectrum disorder (ASD) and their non-ASD family members.

While gut microbial diversity appeared to be remarkably similar between the two groups, distinct dietary patterns emerged. Children with ASD ate fewer vegetables and more sweets, suggesting that in this cohort and design, ASD behaviors could shape the microbiome rather than vice versa.

Rising Autism Rates and Related Health Issues

ASD is a complex neurodevelopmental condition involving social and behavioral difficulties and has become increasingly prevalent, now affecting roughly 1 in 31 children in the U.S.

Alongside its core symptoms, ASD is often associated with gastrointestinal problems, selective eating, anxiety, and immune irregularities, which together reduce quality of life.

Gut-Brain Axis and Microbiome Research Challenges

These comorbidities have drawn attention to the gut-brain axis, a bidirectional system through which gut microbes communicate with the brain via metabolic, immune, and neural pathways.

Animal studies suggest that gut microbiota may affect behavior, but human research has produced inconsistent results, likely due to factors such as diet and geography.

Although some clinical studies indicate that prebiotic or probiotic interventions may modestly relieve gastrointestinal symptoms, their behavioral benefits remain unproven. Consequently, it is unclear whether microbial changes in ASD are a cause or an effect of dietary selectivity and behavioral differences.

Study Design: Comparing Families with and without ASD

This study was designed to disentangle the various confounding factors by combining dietary analysis with microbiome sequencing in children with and without ASD. Researchers aimed to determine whether dietary habits or gut microbial diversity better explain behavioral features of ASD, hypothesizing that diet-driven differences would be the primary factor.

The pilot study examined 17 children with ASD, 9 non-ASD siblings, and 27 parents recruited from an Italian clinical center. Dietary analyses encompassed a larger within-family sample of 79 individuals (26 children with ASD, 12 non-ASD siblings, and 41 parents).

Methods: Microbiome Sequencing and Dietary Assessment

Fecal samples were collected to analyze bacterial and fungal communities, while detailed seven-day food diaries provided dietary data.

Using a within-family design minimized differences in genetics and shared environment. Stool samples were stored at −80 °C and processed using standardized protocols.

Deoxyribonucleic acid (DNA) was extracted, amplified, and sequenced, targeting bacterial and fungal regions.

Resulting sequences underwent quality control, taxonomic classification, and normalization. Microbial diversity was assessed using Shannon and Simpson indices for α-diversity and Bray–Curtis and UniFrac metrics for β-diversity.

PERMANOVA on Bray–Curtis dissimilarities yielded low, non-significant R² values, reinforcing the absence of between-group microbiome differences.

Dietary intake was categorized into eight food groups, and dietary diversity was calculated from food-frequency data. Statistical analyses employed non-parametric tests, such as the Kruskal–Wallis test and the Wilcoxon test, as well as linear mixed-effects models that accounted for family clustering.

Multiple comparisons were corrected using false discovery rate adjustments. The study adhered to established guidelines to ensure methodological transparency and was approved by an institutional ethics committee, with informed consent obtained from all participants.

Microbiome Diversity Shows Minimal Differences

The analysis showed no significant link between gut microbiome composition and autism diagnosis. Bacterial diversity, measured through both α-diversity and β-diversity, was similar among children with ASD, their siblings, and their parents.

Heatmaps and principal component analyses revealed that microbial community structures and species abundances were broadly comparable across all groups. Similarly, the fungal (mycobiome) analysis found no significant group differences. Only two fungal species showed statistical variation, and these differences were between children and parents rather than between ASD and non-ASD siblings.

Dietary Patterns Strongly Linked to ASD Behaviors

In contrast, clear group-level distinctions emerged in dietary patterns. Children with ASD consumed significantly more sweets and sugary foods and fewer vegetables than both siblings and parents. This finding aligns with the statistical results. No differences were found between siblings and parents.

Smaller differences were also seen in fruit and processed food intake, but these mainly distinguished parents from children.

However, the authors note that some of these dietary associations weakened or lost statistical significance when more complex linear mixed-effects models were applied to account for family clustering, likely due to the small sample size.

They also emphasize that the absence of significant microbiome differences may partly reflect limited statistical power rather than a definitive lack of effect.

Overall, the findings suggest that dietary preferences, rather than gut microbial composition, differentiated children with ASD from non-ASD family members, highlighting selective eating behaviors as a defining feature of the ASD group.

Interpretation and Study Limitations

This study found that dietary habits, not gut microbiome composition, best explained group differences in children with ASD. Both bacterial and fungal communities showed minimal variation across participants, indicating that prior inconsistencies in ASD–ASD microbiome studies may stem from small sample sizes, diverse methods, and unaccounted dietary factors.

Children with ASD showed strong preferences for sweet foods and limited vegetable intake, consistent with selective eating behaviors often seen in autism. These habits may contribute to nutritional deficiencies and metabolic issues.

The study’s strengths include its within-family design and integration of bacterial, fungal, and dietary analyses. However, limitations such as the small sample size, cross-sectional nature, and lack of metabolomic data restrict causal interpretation. Differences between child and adult diets also complicate comparisons.

Overall, the findings support a behavior-to-diet-to-microbiome pathway in this cohort; however, causality cannot be inferred from the cross-sectional data. The authors further conclude that larger, multi-omic studies with greater statistical power are required to confirm whether subtle microbiome differences might emerge under more robust analytical conditions.