Hypothesis

An elevated cortisol/DHEAS ratio in aging shifts microglial activity toward complement‑mediated pruning of weaker synapses, leaving strongly weighted connections over‑represented. This produces a network that appears over‑consolidated and resistant to new learning, while the underlying cause is excessive removal of labile connections rather than strengthened consolidation.

Mechanistic Rationale

Chronic cortisol exposure reduces glucocorticoid receptor feedback, prolonging glucocorticoid signaling [1]. Simultaneously, age‑related decline in DHEA removes its anti‑inflammatory and neuroprotective actions [2]. The resulting high cortisol/DHEAS ratio upregulates microglial C1q and C3 expression, tagging low‑activity synapses for phagocytosis [3]. Complement‑dependent pruning preferentially eliminates synapses that have not undergone recent potentiation, thereby increasing the average synaptic weight of the remaining ensemble.

This process creates two observable phenotypes: (1) reduced capacity to encode novel information because fewer plastic substrates remain, and (2) heightened stability of existing patterns, which can be mistaken for over‑consolidation. Importantly, the mechanism predicts that blocking complement signaling or restoring DHEA levels will rescue plasticity without necessarily weakening established memories.

Testable Predictions

1. Older individuals with high cortisol/DHEAS ratios will show increased microglial PET signal (e.g., using [11C]PBR28) in hippocampus correlating with lower fMRI‑based pattern separation performance.
2. Post‑mortem tissue from aged donors with high ratios will exhibit elevated C1q/C3 deposition specifically on synapses with low PSD‑95 intensity, while high‑PSD‑95 synapses remain spared.
3. Acute antagonism of C3aR in aged rodents will normalize long‑term potentiation (LTP) magnitude in hippocampal slices and improve performance on a novelty‑preference task, without impairing retention of a previously learned spatial map.
4. Chronic DHEA supplementation in aged animals will lower the cortisol/DHEAS ratio, reduce microglial complement expression, and increase the density of thin, filopodia‑like dendritic spines indicative of nascent synapses.

Experimental Approach

• Human cohort: Recruit 120 participants aged 65‑85, measure morning cortisol and DHEAS, compute ratio, perform resting‑state [11C]PBR28 PET, and administer a pattern separation task (Mnemonic Similarity Test). Use linear mixed models to test whether ratio predicts PET binding and task accuracy, controlling for age, education, and vascular health.
• Rodent validation: Use 24‑month‑old male rats, split into groups receiving vehicle, C3aR antagonist (SB 290157), or DHEA (10 mg/kg/day) for 4 weeks. After treatment, assess hippocampal LTP ex vivo, conduct Golgi staining to quantify spine subtypes, and run a behavioral battery (novel object recognition, water maze reversal).
• Molecular read‑outs: Western blot and immunofluorescence for C1q, C3, iBA1 (microglial marker), PSD‑95, and synaptophysin in hippocampal subfields.

Falsifiability

If high cortisol/DHEAS ratio does not correlate with increased microglial complement activity or synaptic pruning markers, or if complement blockade fails to restore LTP and behavioral flexibility, the hypothesis would be refuted. Conversely, a positive correlation would support the idea that apparent rigidity arises from selective loss of plastic substrates, not from an over‑consolidated, hyper‑stable network.

Broader Implications

This framework bridges the "over‑consolidation" intuition with hormonal dysregulation evidence, suggesting that interventions aimed at restoring neuroendocrine balance (e.g., DHEA supplementation) or modulating microglial pruning may re‑introduce the controlled uncertainty needed for adaptive cognition in later life.