Hypothesis

A microbiome‑derived "inflammatory age" predicts cognitive decline by driving MCFA‑mediated GPR84 activation on myeloid cells, which silences vagal anti‑inflammatory signaling before systemic cytokines rise.

Mechanistic Rationale

Age‑related shifts in the gut microbiome increase the proportion of taxa that produce medium‑chain fatty acids (MCFAs) while depleting short‑chain fatty acid (SCFA) producers. MCFAs act as agonists of GPR84 on monocytes and macrophages, triggering TNF and IL‑1β release. These cytokines impair vagal sensory neuron firing, diminishing the cholinergic anti‑inflammatory pathway and allowing low‑grade inflammation to spread to the brain. Because the vagus nerve normally suppresses myeloid activation, its desensitization creates a feed‑forward loop that amplifies inflammaging independently of host chronological age. Thus, the functional age of the microbiota—reflected in metagenomic pathways for MCFA synthesis—acts as an upstream driver of the immune‑neural cascade that precedes measurable rises in circulating IL‑1β or IFN‑γ.

Testable Predictions

1. In a longitudinally followed older cohort, baseline microbiome age (calculated from metagenomic MCFA‑synthesis potential) will predict higher serum MCFA/SCFA ratios at 6‑month intervals, independent of age, diet, and medication.
2. Elevated MCFA/SCFA ratios will precede increased myeloid GPR84 signaling (measured by a novel PET tracer) and reduced vagal tone (lower heart‑rate variability) by at least 3‑month lag.
3. Myeloid GPR84 activation and vagal desensitization will, in turn, predict elevations in plasma TNF and IL‑1β, which then forecast declines in hippocampal‑dependent memory scores over the subsequent 12‑month period.
4. Interventions that lower MCFA production (e.g., targeted prebiotics or GPR84 antagonists) will break the temporal sequence, attenuating vagal loss and cognitive decline even when microbiome age remains high.

Experimental Design

Recruit 300 participants aged 65‑80 with normal cognition at baseline. Collect stool for shotgun metagenomics, serum for metabolomics (MCFAs, SCFAs, bile acids), blood for cytokine panels, and perform resting‑state ECG to derive HRV as a vagal proxy. Subsample 30 participants for PET imaging with a GPR84‑specific ligand at baseline and annually. Administer a standardized memory battery every six months. Use time‑varying Cox models with microbiome age, MCFA/SCFA ratio, GPR84 PET signal, HRV, and cytokine levels as covariates to test whether each layer adds predictive value for incident mild cognitive impairment. Conduct a parallel randomized pilot where half receive a MCFA‑lowering prebiotic for 12 months and the other half receive placebo; compare changes in the biomarker cascade and cognition.

Potential Confounds and Controls

Dietary fat intake influences circulating MCFAs; we will adjust using food‑frequency questionnaires. Antibiotic or proton‑pump inhibitor use can alter microbiota; participants reporting recent use will be excluded or modeled as covariates. Host genetics affecting GPR84 expression (e.g., polymorphisms) will be genotyped and included as interaction terms. Reverse causation—where early neurodegeneration alters gut motility and microbiota—will be addressed by excluding converters to dementia within the first year and by lagging predictor variables.