Hypothesis: Age‑related neuronal loss results from a breakdown in the astrocyte‑neuron lactate shuttle that leaves high‑firing neurons energetically stranded, triggering their removal by dysfunctional microglia, rather than from an active pruning program that selects inefficient cells.

Background: The seed idea frames neuronal death as an efficiency‑driven eviction, but 's review shows that loss follows passive degeneration patterns—oxidative protein damage, proteasome/lysosomal failure, and preserved soma despite synapse loss[1][2][3]. Microglia become dystrophic and phagocytosis‑deficient, impairing clearance of damaged elements[6][7]. Meanwhile, neuronal activity shows a biphasic effect: moderate firing is protective, while excess excitation accelerates DNA damage and shortens lifespan via REST/IGF pathways[4][5]. It's becoming clear that energetic support declines with age. These observations point to an energetic mismatch rather than a selective culling.

Mechanistic insight: Astrocytes supply lactate to neurons through monocarboxylate transporters (MCTs), coupling glycolytic output to synaptic demand. With age, astrocytic glycolytic flux declines and MCT1/4 expression drops, reducing lactate availability precisely when neurons need it most during bursts of firing. Energy stress raises the AMP/ATP ratio, activating AMPK and inhibiting SIRT1, which leads to hyperacetylation of PGC‑1α and diminished mitochondrial biogenesis. Concurrently, lactate deficit increases intracellular NADH, favoring lactate dehydrogenase reversal and producing reactive oxygen species that damage mitochondrial DNA. Released mtDNA activates the cGAS‑STING pathway in neurons, inducing a senescence‑associated secretory phenotype (SASP) that tags the cell for microglial removal. However, aged microglia are unable to execute phagocytosis efficiently, so senescent neurons accumulate and are eventually lost through secondary necrosis. We're proposing that the lactate deficit initiates a cascade that tags active neurons for removal.

Testable predictions:

1. In aged mouse cortex, extracellular lactate measured by microdialysis will be significantly lower around layers II/III pyramidal neurons compared with young controls, while glucose levels remain unchanged.
2. Genetic overexpression of MCT1 in astrocytes will restore lactate flux, reduce neuronal SASP markers (p16^INK4a, γH2AX), and preserve neuron counts without altering overall firing rates (measured by in vivo calcium imaging).
3. Chemogenetic reduction of astrocytic lactate release in young mice will reproduce the aged pattern of selective loss of highly active neurons, accompanied by increased cGAS‑STING activation and microglia‑associated senescence markers.
4. Pharmacological boosting of microglial phagocytosis (e.g., with annexin V‑linked liposomes) in aged animals will clear SASP‑positive neurons but will not rescue neuron numbers if the lactate shuttle remains impaired, indicating that the primary driver is energetic deficit.

Falsification: If rescuing astrocytic lactate release fails to prevent neuronal loss or SASP induction despite verified lactate restoration, the hypothesis is falsified and alternative mechanisms (e.g., cell‑intrinsic proteostatic collapse) must be considered.

This framework shifts the focus from "pruning for efficiency" to "metabolic abandonment" and suggests that metabolic support, not synaptic activity alone, determines neuronal survival in the aging brain.