A Multi-Compartment Computational Model of Dendritic Spine Plasticity and Network Hyperexcitability in Epileptogenesis
Department
Department of Biomedical and Translational Sciences
Graduate Program/Concentration
Eastern Virginia Medical School - Doctor of Medicine
Presentation Type
No Preference
Abstract
Background: Epileptogenesis is characterized by a progressive increase in network excitability and synchronization, often accompanied by structural remodeling of dendritic spines and alterations in synaptic connectivity. Experimental evidence suggests that calcium-mediated spine plasticity plays a critical role in modulating synaptic strength and neuronal excitability. However, the mechanistic relationship between spine dynamics, synaptic plasticity, and emergent epileptiform activity remains incompletely understood. Computational modeling offers a powerful framework to explore how dendritic spine remodeling influences network-level hyperexcitability and the transition from physiological to pathological activity states.
Methods: We developed a multi-scale computational model integrating multi-compartment neurons with Hodgkin–Huxley (HH) membrane dynamics, calcium-dependent dendritic spine plasticity, and conductance-based synapses. Our hybrid network consists of detailed neurons (including soma, active dendrites, and spines) and point neurons, connected via a spatially constrained connectivity rule with distance-dependent synaptic strength.
Each detailed neuron incorporates HH sodium and potassium currents, along with a persistent sodium current and a voltage-gated calcium current that directly modulate intracellular calcium concentration]. Spine plasticity is governed by a calcium-dependent rule, where spine neck resistance and head capacitance evolve dynamically based on local calcium levels, reflecting activity-dependent synaptic remodeling. Synaptic interactions are conductance-based and follow a biexponential decay model.
To induce epileptiform activity, we introduced a transient "kindling" current injection (5 µA/cm² from 300 to 700 ms) to a subset of neurons, simulating an external perturbation that drives hyperexcitability. The model was simulated for 2000 ms using an adaptive ordinary differential equation solver, and key metrics—including firing rates, network synchrony (pairwise spike-time correlation), and spine parameter evolution—were computed and analyzed.
Results:
Increased excitability and synchrony during kindling: The transient current injection resulted in a 2.5× increase in network firing rate, with activity spreading from directly stimulated neurons to the broader network. The network synchrony index rose from baseline (0.04) to a peak of 0.32 during kindling, indicating emergent synchronization.
Calcium-driven spine remodeling: In neurons exhibiting sustained high-frequency spiking, intracellular calcium exceeded plasticity thresholds (0.1 µM), leading to a reduction in neck resistance (~10% decrease) and an increase in head capacitance (~15% increase), suggesting synaptic potentiation.
Persistent post-kindling activity: Following the cessation of kindling, firing rates remained elevated (1.7× above baseline), and network synchrony declined only partially, suggesting that transient perturbations can induce long-lasting changes in synaptic structure and excitability.
Conclusion: Our results suggest that dendritic spine plasticity plays a causal role in shaping network excitability, where calcium-mediated structural modifications reinforce synaptic potentiation and facilitate the transition to hyperexcitable states. These findings support the hypothesis that dendritic spine morphology is not merely a passive correlate of epileptogenesis but an active driver of synaptic reorganization and emergent pathological activity. This model provides a computational framework for exploring targeted interventions that modulate dendritic spine plasticity to restore normal network function and mitigate seizure susceptibility. Future work will explore the role of inhibitory synapses, short-term synaptic dynamics, and larger-scale network effects.
Keywords
Computational Neuroscience, Dendritic Spine Plasticity, Calcium-Dependent Plasticity, Epileptogenesis, Multi-Compartment Neuron Model, Conductance-Based Synapses, Network Synchrony
A Multi-Compartment Computational Model of Dendritic Spine Plasticity and Network Hyperexcitability in Epileptogenesis
Background: Epileptogenesis is characterized by a progressive increase in network excitability and synchronization, often accompanied by structural remodeling of dendritic spines and alterations in synaptic connectivity. Experimental evidence suggests that calcium-mediated spine plasticity plays a critical role in modulating synaptic strength and neuronal excitability. However, the mechanistic relationship between spine dynamics, synaptic plasticity, and emergent epileptiform activity remains incompletely understood. Computational modeling offers a powerful framework to explore how dendritic spine remodeling influences network-level hyperexcitability and the transition from physiological to pathological activity states.
Methods: We developed a multi-scale computational model integrating multi-compartment neurons with Hodgkin–Huxley (HH) membrane dynamics, calcium-dependent dendritic spine plasticity, and conductance-based synapses. Our hybrid network consists of detailed neurons (including soma, active dendrites, and spines) and point neurons, connected via a spatially constrained connectivity rule with distance-dependent synaptic strength.
Each detailed neuron incorporates HH sodium and potassium currents, along with a persistent sodium current and a voltage-gated calcium current that directly modulate intracellular calcium concentration]. Spine plasticity is governed by a calcium-dependent rule, where spine neck resistance and head capacitance evolve dynamically based on local calcium levels, reflecting activity-dependent synaptic remodeling. Synaptic interactions are conductance-based and follow a biexponential decay model.
To induce epileptiform activity, we introduced a transient "kindling" current injection (5 µA/cm² from 300 to 700 ms) to a subset of neurons, simulating an external perturbation that drives hyperexcitability. The model was simulated for 2000 ms using an adaptive ordinary differential equation solver, and key metrics—including firing rates, network synchrony (pairwise spike-time correlation), and spine parameter evolution—were computed and analyzed.
Results:
Increased excitability and synchrony during kindling: The transient current injection resulted in a 2.5× increase in network firing rate, with activity spreading from directly stimulated neurons to the broader network. The network synchrony index rose from baseline (0.04) to a peak of 0.32 during kindling, indicating emergent synchronization.
Calcium-driven spine remodeling: In neurons exhibiting sustained high-frequency spiking, intracellular calcium exceeded plasticity thresholds (0.1 µM), leading to a reduction in neck resistance (~10% decrease) and an increase in head capacitance (~15% increase), suggesting synaptic potentiation.
Persistent post-kindling activity: Following the cessation of kindling, firing rates remained elevated (1.7× above baseline), and network synchrony declined only partially, suggesting that transient perturbations can induce long-lasting changes in synaptic structure and excitability.
Conclusion: Our results suggest that dendritic spine plasticity plays a causal role in shaping network excitability, where calcium-mediated structural modifications reinforce synaptic potentiation and facilitate the transition to hyperexcitable states. These findings support the hypothesis that dendritic spine morphology is not merely a passive correlate of epileptogenesis but an active driver of synaptic reorganization and emergent pathological activity. This model provides a computational framework for exploring targeted interventions that modulate dendritic spine plasticity to restore normal network function and mitigate seizure susceptibility. Future work will explore the role of inhibitory synapses, short-term synaptic dynamics, and larger-scale network effects.