The gut–brain axis conveys visceral and immune information to the brain predominantly via the vagus nerve. In parallel, the locus coeruleus (LC) mediates phasic "reset" signals that reconfigure large‑scale cortical networks. We propose the Microbiome–Vagus–LC (MVL) hypothesis: dysbiotic microbiome states bias vagal afferent signalling and thereby reduce LC reset efficiency, shaping anxiety/PTSD phenotypes. We define reset efficiency with a pre‑registered composite of LC proxies and behaviour (task‑evoked pupil‑dilation kinetics, EEG P3b amplitude/latency, and task‑switching costs), and index vagal tone primarily with paced‑breathing HF‑HRV (0.1 Hz) and RMSSD, including respiration–HR coherence as a quality metric. We outline (i) a cross‑sectional mediation study using path analysis with bootstrapping and Bayesian SEM as sensitivity, and (ii) an 8‑week, 2×2 factorial trial (high‑fibre/probiotic × sham/active auricular VNS) analysed with ANCOVA on change scores. While individual links (microbiome→vagus, vagus→anxiety/HRV, LC→arousal/reset) are established, the complete M→V→R→Y chain has not been directly tested; this program is designed to make that chain falsifiable.