Abstracts

Altered dendritic-spine morphology—specifically neck elongation and head shrinkage—has been documented in both animal models and surgical tissue from patients with drug-resistant focal epilepsy. Yet the biophysical consequences of these structural changes are still uncertain because experimental manipulations rarely isolate morphology from concurrent molecular or network alterations. A rigorously constrained computational model allows spine geometry to be varied in isolation, making it possible to quantify whether realistic morphologic shifts alone produce measurable changes in synaptic integration, calcium dynamics, and the propensity for hyperexcitability that could lower the threshold for seizure initiation.

Spine meshes spanning neck lengths 0.2–2.5 µm and head volumes 0.05–0.8 µm³ were attached to a reconstructed layer-5 pyramidal neuron. Passive properties (Rm 28 kΩ·cm², Cm 1 µF/cm²) were fitted to dual-patch recordings; active NaV1.2, KV4.2, and CaV2.3 gradients followed published values (34 °C). Six morphology bins (120 spines each) received single-synapse and 10 Hz Poisson AMPA/NMDA input. A 50-cell network with distance-dependent connectivity assessed population effects. Outcome measures were peak EPSP, spine-calcium decay (τ), four-spike burst probability, and high-frequency oscillation (HFO, 80–200 Hz) power.

Raising neck resistance from 35 MΩ to 110 MΩ (morphology index < 0.3 → > 0.7) increased dendritic EPSP amplitude by 12 ± 2 % and prolonged calcium τ by 33 ± 4 % (mixed-effects model, p < 0.05). Burst probability climbed from 0.19 to 0.32 (OR 1.9, 95 % CI 1.4–2.4). At the network level, a 20 % shift toward high-resistance spines lowered the minimal stimulus for self-sustained firing by 9 % and elevated HFO power by 11 ± 3 % (p < 0.05), without altering channel densities