The stochastic diffusion models of nerve membrane depolarization and interspike interval generation

Area of Science:

• Computational neuroscience
• Theoretical neuroscience
• Mathematical biology

Background:

• Deterministic leaky-integrator model provides a foundation for understanding neuronal dynamics.
• Stochastic models are crucial for describing neuronal activity, with intrinsic and external noise sources identified.
• Neuronal coding strategies, particularly rate coding, are essential for information processing in the brain.

Purpose of the Study:

• To provide an accessible review of stochastic diffusion processes for single neuron modeling.
• To analyze rate coding through input-output curves in various stochastic models.
• To compare the limitations and properties of different diffusion models, including additive and multiplicative noise.

Main Methods:

• Review of the deterministic leaky-integrator model.
• Summarization of general principles for simple stochastic models.
• Analysis of rate coding using input-output curves for Wiener, Ornstein-Uhlenbeck, and derived multiplicative noise models.
• Investigation of parameter identification and time-variable input effects.

Main Results:

• The Wiener process model, while simple, exhibits significant limitations.
• The Ornstein-Uhlenbeck process offers a more realistic approach by including spontaneous membrane potential changes.
• Derived multiplicative noise models provide a more nuanced representation of state-dependent neuronal inputs.
• Parameter identification strategies and the impact of time-varying inputs are explored.

Conclusions:

• Stochastic diffusion models, ranging from simple to complex, are vital for understanding single neuron behavior.
• Comparing different noise models (additive vs. multiplicative) enhances our understanding of neuronal responses.
• Accurate parameter identification is critical for validating and applying these neuronal models.