Hypothesis

Acarbose extends lifespan and rescues cognitive decline not only by lowering post‑prandial glucose but by creating intermittent metabolic uncertainty that reactivates hippocampal prediction‑error circuits, thereby breaking the over‑consolidated maps that characterize aging brains.

Mechanistic Rationale

1. Glucose fluctuations → astrocytic AMPK activation Acarbose slows carbohydrate absorption, producing lower peaks and occasional troughs in blood glucose. These troughs trigger AMPK in astrocytes, increasing lactate production and release. Lactate acts as a signaling molecule that modulates neuronal NMDA‑receptor subunit composition, favoring GluN2B‑containing receptors that enhance synaptic plasticity and spike‑timing dependent learning.

2. SCFA‑mediated microglial re‑programming The acarbose‑shifted microbiome raises colonic butyrate and propionate. These SCFAs cross the blood‑brain barrier and bind GPR41/43 on hippocampal microglia, shifting them from a synapse‑pruning phenotype to a supportive state that releases BDNF and IGF‑1. This microglial change reduces excessive elimination of nascent synapses, allowing new connections to stabilize.

3. Prediction‑error signal restoration Elevated lactate and BDNF increase neuronal excitability and sharpen theta‑gamma coupling during novel stimulus exposure. The resulting boost in prediction‑error signaling (mediated by locus coeruleus norepinephrine release) signals the cortex that internal models are outdated, prompting map updating. In aged mice, acarbose restores this signaling cascade, which is absent when glucose is stably low or high.

Testable Predictions

• Prediction 1: In 24‑month‑old mice, chronic acarbose treatment will increase hippocampal lactate levels during glucose troughs, measured by microdialysis, and this rise will correlate with improved performance on a reversal‑learning task.
• Prediction 2: Pharmacological blockade of astrocytic AMPK (using compound C) or inhibition of lactate transport (via MCT1 antagonist) will abolish acarbose‑induced cognitive benefits without affecting its glucose‑lowering effect.
• Prediction 3: Genetic deletion of microglial GPR41/43 will prevent the acarbose‑driven shift toward a BDNF‑secreting microglial phenotype and will block memory improvement, despite unchanged peripheral glucose profiles.
• Prediction 4: Simultaneous chemogenetic silencing of locus coeruleus norepinephrine neurons during novelty exposure will erase the acarbose‑enhanced theta‑gamma coupling and the associated behavioral gains.

Potential Confounds and Controls

• Ensure that any observed cognitive changes are not secondary to reduced food intake or weight loss; pair‑fed controls receiving iso‑caloric diets will be included.
• Verify that antibiotic depletion of the microbiome does not acutely reverse acarbose’s glucose‑lowering action, allowing isolation of the SCFA‑dependent arm.
• Use rapamycin‑treated mice as a comparative group to distinguish mTORC1‑dependent lifespan effects from the metabolic‑uncertainty mechanism proposed here.

By framing acarbose as a tool that re‑introduces controlled metabolic noise, this hypothesis links peripheral pharmacology to central computational processes, offering a concrete route to test whether age‑related cognitive rigidity can be loosened by reinstating prediction‑error signaling rather than by merely repairing damaged tissue.