Peer-reviewed article
Microbes on the Mind: Multi-modal Neuroimaging Reveals Gut-Brain Axes in Infants.
Tengfei Li, Yue Yang, Seo-Ho Cho, Brittany R Howell, Zhengwang Wu et al. ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This study examined the relationship between gut microbiota and brain development in 223 infants aged 0-3 years using brain imaging and stool samples. Researchers found associations between gut microbes and different aspects of brain structure, white matter, and connectivity patterns. The study identified two main patterns linking microbial composition to brain characteristics, and found that brain imaging measures could predict changes in microbiome composition several months later. The findings suggest connections between gut and brain development in early infancy, though the study cannot prove that microbes directly cause brain changes.
Key points
- Microbiome composition was associated with cortical morphology, white-matter microstructure, and functional connectivity in infants
- Two principal covariance patterns emerged: a structural-metabolic mode and a functional-taxonomic mode
- Brain imaging measures predicted changes in microbiome composition 1-6 months later in autoregressive models
- Study included 223 typically developing children aged 0-3 years with multimodal MRI and fecal metagenomics
- Results show cross-domain covariance and temporal ordering but do not establish causality
Limitations
The study is observational and cannot establish causal relationships between gut microbes and brain development. The findings are correlational and temporal ordering alone does not prove that brain changes cause microbiome changes or vice versa.
Abstract
The gut microbiome undergoes rapid maturation after birth and has been associated with cognition and mental health, yet its relation to early brain development remains poorly defined. We analyzed 223 typically developing children aged 0-3 years with multimodal MRI and paired fecal shotgun metagenomics. Microbiome-wide association analyses identified associations with cortical morphology, white-matter microstructure, and functional connectivity. Ridge-regularized canonical correlation analysis identified two principal covariance patterns: a structural-metabolic mode and a functional-taxonomic mode. Sparse canonical partial least squares recovered the principal CCA score and phenotype patterns within the same sample. Longitudinal canonical-score slopes were not significantly correlated; in autoregressive cross-lag models, however, neuroimaging CCA1 scores predicted microbiome CCA1 scores 1-6 months later. These results indicate cross-domain covariance and temporal ordering within the measured windows but do not establish causality.