Hypothesis

Age‑related cognitive rigidity stems not from neuronal loss but from excessively precise priors that suppress prediction‑error signaling. Restoring peripheral serotonin produced by specific gut bacteria (e.g., Limosilactobacillus mucosae, Ligilactobacillus ruminis) will reinstate vagal afferent tone, boost cortical 5‑HT2A‑mediated gain on prediction errors, and selectively improve cognitive flexibility without enhancing reference memory.

Mechanistic Rationale

Aging reduces colonic TPH1 expression in enterochromaffin (EC) cells as microbial diversity and SCFA‑producing taxa decline Gut microbes promote colonic serotonin production through an effect on TPH1. Certain Lactobacilli can directly synthesize serotonin from 5‑HTP, contributing to peripheral 5‑HT pools that activate vagal sensory afferents Specific gut bacteria regulate intestinal motility by producing serotonin. Vagal signaling to the nucleus tractus solitarius modulates locus coeruleus norepinephrine (NE) release, which determines the precision weighting of sensory prediction errors in frontal cortex. When peripheral 5‑HT falls, vagal drive weakens, NE becomes tonically low, and cortical neurons adopt high‑precision priors—manifesting as over‑consolidation. Serotonin, particularly via 5‑HT2A receptors, increases cortical gain and amplifies prediction‑error signals, encouraging belief updating. Thus, boosting gut‑derived 5‑HT should restore the uncertainty signal needed to counteract rigidity.

Predictions

1. Old mice receiving a probiotic cocktail of L. mucosae and L. ruminis (or oral 5‑HTP) will show improved performance in reversal learning and attentional set‑shifting tasks, while performance on spatial reference memory (e.g., Morris water maze probe) remains unchanged.
2. Subdiaphragmatic vagotomy will abolish the flexibility benefits of the probiotic treatment, confirming vagal mediation.
3. Pharmacological blockade of cortical 5‑HT2A receptors (e.g., with ketanserin) will prevent the improvement in set‑shifting despite restored peripheral 5‑HT.
4. Electrophysiological recordings from medial prefrontal cortex will reveal increased neuronal responsiveness to unexpected outcomes (larger prediction‑error evoked spikes) after treatment, without changes in baseline firing rates.
5. Measurements of fecal SCFA, colonic TPH1 mRNA, and plasma 5‑HT will correlate positively with the magnitude of flexibility gains.

Experimental Design

• Subjects: 20‑month‑old C57BL/6J mice (n=12 per group) and young controls (3‑months).
• Groups: (A) Vehicle control, (B) Probiotic cocktail (L. mucosae + L. ruminis, 10^9 CFU each daily), (C) 5‑HTP supplement (5 mg/kg/day), (D) Probiotic + vagotomy, (E) Probiotic + ketanserin (0.5 mg/kg i.p. before testing).
• Procedure: Four‑week treatment, followed by behavioral battery: (i) Y‑maze spontaneous alternation (working memory), (ii) Barnes maze reference memory acquisition and probe, (iii) Attentional set‑shifting (digging paradigm) with reversal stages, (iv) Probabilistic reversal learning to quantify prediction‑error sensitivity.
• Measures: Fecal 16S rRNA sequencing, colonic TPH1 qPCR, plasma 5‑HT (ELISA), vagal afferent firing (extracellular recordings in a subset), mPFC in vivo electrophysiology during unexpected outcome presentation, c‑Fos in locus coeruleus.

Potential Outcomes and Falsifiability

If the probiotic or 5‑HTP improves set‑shifting and reversal learning without affecting reference memory, and these effects are blocked by vagotomy or 5‑HT2A antagonism, the hypothesis is supported. Lack of flexibility improvement, or equal enhancement of reference memory, would falsify the claim that peripheral serotonin specifically rescues over‑consolidation. Conversely, if improvements occur despite vagotomy, alternative pathways (e.g., humoral) would need consideration, prompting revision of the mechanistic model.

This framework directly tests whether re‑introducing controlled uncertainty via gut‑brain serotonin signaling can reverse age‑related cognitive rigidity, shifting the therapeutic goal from restoration of lost function to recalibration of predictive processing.