The hypothesis is that psilocybin reduces microglial-mediated phagoptosis of viable but metabolically stressed neurons in the aging brain, thereby counteracting activity‑dependent pruning that follows efficiency logic. Psilocybin’s acute 5‑HT2A receptor agonism increases neuronal firing and metabolic demand in cortical hubs, which should raise ATP production and lower stress signals such as surface calreticulin. Simultaneously, psychedelic‑induced downstream signaling shifts microglia toward a homeostatic phenotype, decreasing complement component C1q and C3 deposition and reducing phagocytic receptor expression (e.g., MerTK, LRP1). The net effect is a lower probability that stressed neurons are tagged for engulfment, preserving neuronal numbers and network efficiency.

Key predictions:

1. In aged mice (20‑24 mo), a single psilocybin dose (1 mg/kg, i.p.) will decrease the proportion of Iba1+ microglia co‑localizing with neuronal calreticulin or C3b markers in prefrontal cortex and hippocampus compared with vehicle controls.
2. This reduction will correlate with increased neuronal survival measured by NeuN+ cell counts and reduced caspase‑3 activation in the same regions.
3. Electrophysiological recordings will show heightened spontaneous firing rates and glucose uptake (via 2‑DG autoradiography) in neurons that survive phagoptosis, confirming a link between metabolic efficiency and protection.
4. Behavioral assays (e.g., novel object recognition, delayed alternation) will reveal improved cognitive performance in psilocybin‑treated aged mice, and this improvement will be abrogated by pharmacological blockade of 5‑HT2A receptors or by microglial depletion with PLX5622.

Mechanistically, we propose that psilocybin elevates intracellular cAMP in neurons via 5‑HT2A‑Gαs coupling, boosting mitochondrial oxidative phosphorylation and lowering AMP‑activated protein kinase (AMPK) activity. AMPK suppression reduces the exposure of “eat‑me” signals like calreticulin and phosphatidylserine on the neuronal surface. In parallel, psilocybin triggers microglial 5‑HT2A‑β‑arrestin pathways that inhibit NF‑κB translocation, curbing pro‑inflammatory cytokine release and complement synthesis. The combined shift creates a microenvironment where microglia interpret neurons as metabolically competent, thus sparing them from phagoptosis.

This model is falsifiable: if psilocybin fails to alter microglial phagocytic markers or neuronal survival in aged animals, or if blocking neuronal AMPK mimics the protective effect without psychedelic exposure, the hypothesis would be refuted. Conversely, confirming the predicted cascade would bridge the fields of psychedelic neuroplasticity and activity‑dependent neuronal eviction, suggesting that therapeutic benefits of psilocybin in part stem from preserving neuronal efficiency rather than merely inducing novel connectivity.