The possibility that anxiety disorders are partly rooted in the gut—not just the brain—has profound implications for how psychiatry might eventually approach treatment. Most gut-brain research in this space is correlational, leaving open the question of whether microbial imbalances cause anxiety or merely accompany it. This mouse study attempts to push toward causality by transplanting the actual gut ecosystems of humans with social anxiety disorder into antibiotic-cleared rodents.
Fecal samples from five individuals diagnosed with social anxiety disorder (SAD) and five matched healthy controls were transplanted into antibiotic-treated mice. Recipients of SAD-derived microbiota spent significantly more time in the peripheral zones of an open-field test and made fewer entries into the open arms of an elevated-plus maze—both well-validated rodent proxies for anxiety-like behavior. While alpha-diversity metrics were comparable between groups, beta-diversity analysis revealed distinct microbial community structures. SAD-recipient mice showed enrichment of Bacteroidota and Bacteroidales-affiliated genera, including Muribaculum, while control recipients harbored higher abundances of butyrate-producing genera such as Butyricimonas. Functional inference via PICRUSt2 flagged reductions in predicted DNA-repair and biosynthetic pathways in SAD animals, alongside measurable alterations in plasma tryptophan-pathway metabolites—a neurotransmitter precursor axis already implicated in mood regulation.
This study builds on a small but growing body of FMT-transfer research, including prior work linking depression-associated microbiota to passive behavior in rodents. The mechanistic logic is plausible: butyrate shapes gut-brain signaling via vagal afferents and modulates serotonin biosynthesis in enterochromaffin cells, while tryptophan pathway shifts affect both serotonin and kynurenine availability. That said, the limitations are substantial. With only five donor participants per group, this is far too small to represent the heterogeneity of SAD. Mouse behavioral readouts do not map cleanly onto human social anxiety phenotypes, and PICRUSt2 functional predictions are inferences, not direct measurements. This remains an early-stage mechanistic signal—provocative, but requiring replication with larger, more diverse human cohorts before clinical relevance can be assessed.
