The routes
There are at least four channels: the vagus nerve, which carries signals from gut to brain; the immune system, via cytokines and inflammation; microbial metabolites, including short-chain fatty acids and neurotransmitter precursors; and the hypothalamic–pituitary–adrenal axis, which responds to stress and changes gut permeability.
That the routes exist is not in dispute. Germ-free mice, which have no microbiome, show differences in anxiety-like behaviour, stress response and brain development — and those differences partly reverse with colonisation.
Where the evidence gets thin
Mouse to human. Most mechanistic work is in rodents. Human studies are largely observational: people with depression differ in microbiome composition from controls, but diet, medication, sleep and stress all differ too.
Probiotic trials. Effects on mood are small, inconsistent, and often not replicated. The strongest claims come from small trials with multiple outcomes measured.
Specific diseases. Parkinson's has the most intriguing link — alpha-synuclein pathology appears in the gut years before motor symptoms, and vagotomy is associated with lower risk in some cohorts. That is suggestive, not causal.
The microbiome is not a fixed organ; it changes with diet, antibiotics, geography and age. That makes 'the microbiome causes X' hard to test, because the exposure is unstable and the direction of causation is often bidirectional.
Faecal transplant for recurrent C. difficile is the one place where the causal chain is clear and the therapy works. It is also the exception that shows how high the bar is: an infection, a defined pathogen, a measurable cure.
How to read the claims
Treat microbiome claims like nutrition claims: plausible mechanisms, weak human trials, and a strong commercial incentive to overstate. The archive's open problems page files the related 'leaky gut' hypothesis under live debate for the same reason.
What would move it: large, preregistered trials with defined strains, objective endpoints and long follow-up. Some are running. Until they report, the honest answer is that the axis is real and its clinical importance is not established.
Sources for this page
- Wikipedia, “Gut–brain axis” — mechanisms and evidenceen.wikipedia.org
- Cryan et al. (2019), the microbiome and brain function, Physiological Reviewsjournals.physiology.org
- Sampson et al. (2016), gut bacteria and Parkinson's motor symptomswww.cell.com
- U.S. NIAID, faecal microbiota transplantation evidencewww.niaid.nih.gov