Hypothesis

Age‑related decline in autophagy does not merely cause toxic buildup; it reduces a basal level of molecular noise that normally prevents synapses from becoming overly stable. When autophagy wanes, aggregation‑prone proteins accumulate and act as "molecular brakes" that suppress the stochastic fluctuations needed for synaptic remodeling. Restoring autophagy re‑introduces this noise, thereby re‑opening a window for experience‑dependent plasticity without requiring global brain regeneration.

Mechanistic Rationale

• Autophagy clears short‑lived synaptic proteins; its slowdown doubles protein half‑lives across brain regions, leading to accumulation of ~1,700 neuronal aggregates, ~30 % of which co‑localize with pathological deposits【https://pmc.ncbi.nlm.nih.gov/articles/PMC12139995/】.
• These aggregates can bind to scaffolding molecules (e.g., PSD‑95, Gephyrin) and sterically hinder the conformational shifts required for long‑term potentiation (LTP) or depression (LTD).
• Concurrently, the complement‑mediated pruning shift—↑C1q "eat‑me" tags and ↓CD47 "don’t‑eat‑me" signals—suggests that synapses are being tagged for removal rather than stabilized【https://www.bu.edu/kilachandcenter/cognitive-decline-in-old-age-may-be-linked-to-increased-pruning-of-brain-cell-connections/】. In a low‑noise environment, tagged synapses may be eliminated indiscriminately, while protected synapses become rigid.
• Disrupting the BCL‑2/Beclin‑1 interaction (F121A mutation) extends lifespan and healthspan without cognitive deficits, indicating that enhanced autophagy can be decoupled from neurodegeneration【https://pmc.ncbi.nlm.nih.gov/articles/PMC5992097/】.

We propose that the aggregates themselves generate a low‑amplitude, stochastic perturbation of synaptic membrane dynamics—akin to molecular noise—that biases the probability of synaptic weight changes toward intermediate values. When this noise diminishes, the synaptic weight distribution narrows, manifesting as behavioral rigidity.

Predictions & Experimental Design

1. Measure synaptic noise – Use patch‑clamp recordings in hippocampal slices from young, aged, and aged F121A knock‑in mice to quantify variance in miniature excitatory postsynaptic current (mEPSC) amplitude and frequency. Prediction: aged slices show reduced variance; F121A restores variance to youthful levels.
2. Correlate aggregate load with noise – Perform immunofluorescence for ubiquitin‑positive aggregates and PSD‑95, then correlate puncta density with mEPSC variance across individual slices. Prediction: higher aggregate density predicts lower variance.
3. Behavioral test of flexibility – Subject mice to a reversal learning task in the Morris water maze. Prediction: aged mice exhibit prolonged reversal latency; F121A mice perform like young controls, and variance in mEPSC predicts individual performance.
4. Pharmacological rescue – Treat aged wild‑type mice with an autophagy inducer (e.g., spermidine) for four weeks. Prediction: increased LC3‑II/I ratio, reduced aggregate burden, restored synaptic noise, and improved reversal learning.

All predictions are falsifiable: if autophagy enhancement fails to restore synaptic variance or cognitive flexibility despite clearing aggregates, the hypothesis would be refuted.

Potential Implications

If validated, this framework shifts therapeutic focus from merely removing toxic aggregates to tuning the stochastic synaptic environment. Intermittent fasting, exercise, or pharmacological autophagy modulators could be prescribed not to "boost memory" but to re‑introduce controlled uncertainty, thereby counteracting the maladaptive rigidity that masquerades as cognitive decline.