Epilepsy Etiology: A Study on the Principles Leading to Seizure Events

This repository contains computational models for simulating neuronal activity in epileptic brain tissue, from single-neuron dynamics to population-level behavior. The project was developed as part of the exam of the Brain Modeling course during my Bachelor in Artificial Intelligence.

Overview

The project implements two complementary approaches to modeling epileptic seizures:

1. Microscale (single neuron): Hodgkin-Huxley model with pathological ion channel dynamics
2. Mesoscale (neuronal populations): Wendling neural mass model for local field potentials

Simulated 30-second epileptic seizure event showing smooth transitions through different neuronal states

Models implemented

Hodgkin-Huxley Model

Classic conductance-based model for neuronal action potentials with extensions for seizure-like activity:

• Normal spiking behavior with external current injection
• Pathological activity through time-varying Na⁺ and K⁺ conductances
• Smooth transitions from normal to seizure states using sigmoid functions

Wendling Neural Mass Model

Population-level model simulating interactions between pyramidal cells and inhibitory interneurons:

• Three neuronal populations (pyramidal, slow inhibitory, fast inhibitory)
• Six distinct epileptic states from background to full ictal activity
• Realistic local field potential (LFP) generation

Installation

Requirements

• Python 3.7+
• NumPy
• SciPy
• Matplotlib

See requirements.txt for specific versions.

Setup

# clone the repository
git clone https://github.com/sebapsan/epilepsy-etiology.git
cd epilepsy-etiology

# install dependencies
pip install -r requirements.txt

Usage

Running simulations

Each example script can be run independently:

# Hodgkin-Huxley normal neuron
python examples/hh_normal.py

# Hodgkin-Huxley seizure simulation
python examples/hh_seizure.py

# Wendling basic LFP
python examples/wendling_basic.py

# six epileptic states
python examples/wendling_six_states.py

# complete seizure event
python examples/seizure_concatenation.py

Using the models

from src.models.hodgkin_huxley import HH_ode
from src.models.wendling import WendlingNMM
from src.utils.parameters import get_hh_params, get_wendling_params

# Hodgkin-Huxley simulation
config = get_hh_params()
# run simulation with scipy.integrate.solve_ivp

# Wendling simulation
P = get_wendling_params(A=4, B=15, G=22)
# run simulation with Euler integration

Key parameters

Wendling Model states

The model reproduces six distinct epileptic activity patterns:

Type	A	B	G	Description
1	3.0	26	10	Background activity
2	4.0	26	10	Sporadic spikes
3	5.0	22	10	Rhythmic spikes
4	4.0	15	22	Slow quasi-sinusoidal
5	6.5	10	22	Low voltage rapid activity
6	7.0	18	2	Fast activity (ictal)

Parameters:

• A: Excitatory synaptic gain
• B: Slow inhibitory synaptic gain
• G: Fast inhibitory synaptic gain

References

This work is based on the following papers:

Hodgkin, A. L., & Huxley, A. F. (1952). A quantitative description of membrane current and its application to conduction and excitation in nerve. The Journal of Physiology, 117(4), 500-544. DOI: 10.1113/jphysiol.1952.sp004764

Wendling, F., Bartolomei, F., Bellanger, J. J., & Chauvel, P. (2002). Epileptic fast activity can be explained by a model of impaired GABAergic dendritic inhibition. European Journal of Neuroscience, 15(9), 1499-1508. DOI: 10.1046/j.1460-9568.2002.01985.x

Methodology

Hodgkin-Huxley Model

• Numerical integration: Runge-Kutta 4th order (RK45)
• Time step: Variable (adaptive)
• Simulation time: 120 ms

Wendling Model

• Numerical integration: Euler method
• Sampling frequency: 512 Hz
• Time step: ~1.95 ms

Acknowledgments

I would like to highlight the importance of the seminal papers by Wendling et al. (2002) and Hodgkin & Huxley (1952), fundamental contributions to computational neuroscience and epilepsy research.

Project group members

Pietro Saveri

Sebastiano Pietrasanta

Matteo Salami