Short-Term Synaptic Plasticity Modulates the Outcome of Neurodegenerative Diseases

Alzheimer's disease (AD) is characterised by progressive neuronal loss and synaptic dysfunction that disrupt memory-encoding neural circuits. Short-term synaptic plasticity (STP), which governs transient activity-dependent changes in synaptic efficacy on millisecond-to-second timescales, represents a potentially targetable mechanism for dynamically compensating circuit-level deficits without requiring structural repair. Using the NEST Simulator, we constructed a spiking neural network of 1,000 Hodgkin--Huxley excitatory and 250 leaky integrate-and-fire inhibitory neurons with embedded memory engram assemblies, modelling AD pathology through accelerated neuronal loss and pathological hyperexcitability. Synaptic dynamics were governed by the Tsodyks--Markram model, and compensatory interventions were evaluated by systematically sweeping the facilitation time constant \((\tau_{\text{fac}})\) and recovery time constant \((\tau_{\text{rec}})\) across a two-dimensional parameter space. Engram recall fidelity was quantified by the selectivity ratio \((\Gamma)\), and network-level oscillatory dynamics were assessed via simulated Local Field Potential power spectral density. Increasing \((\tau_{\text{fac}})\) from 500\,ms to 1,250\,ms rescued engram selectivity in the AD model at intermediate synaptic weight regimes inaccessible at baseline, and partially or fully restored theta- and alpha-band LFP power respectively. The \((\tau_{\text{rec}})\) sweep revealed a fundamental trade-off: increasing \((\tau_{\text{rec}})\) suppressed pathological hyperexcitability, whereas decreasing it maximised \((\Gamma)\), with the two objectives requiring opposing parameter changes. These results demonstrate that STF and STD modulation exert mechanistically distinct compensatory effects, and identify bounded therapeutic windows within which dynamic re-tuning of surviving synaptic circuits can meaningfully mitigate AD-related functional deficits.