Intrinsic Dynamics in Neuronal Networks. I. Theory

Peter E. Latham, Barry J. Richmond, P. G. Nelson, Sheila NirenbergJournal of Neurophysiology 2000 · DOI

TL;DR

This paper proposes that the presence of "endogenously active" neurons (neurons that fire without external input) is the key to generating stable, low-rate firing patterns observed in many brain networks. They show, through theoretical models and simulations, that a critical fraction of these neurons is needed to prevent networks from either being silent or bursting, and that spike-frequency adaptation can induce bursting.

What Problem Does This Paper Address?

The mammalian brain is remarkably active even without external stimulation. This intrinsic activity typically falls into two main patterns: a steady, low-frequency firing (1-5 Hz) and rhythmic bursting. While theoretical models have had some success explaining bursting, understanding how networks maintain stable, low firing rates has been a persistent challenge. Previous attempts using large, interconnected networks often resulted in much higher firing rates or silence, suggesting that standard assumptions about neuronal interactions might be missing a crucial element. This paper aims to bridge this gap by identifying the core mechanisms that generate these distinct intrinsic firing patterns and the conditions under which networks transition between them.

Methods

The authors employ a multi-pronged approach:

  • Semianalytic Model (Mean Firing Rate Dynamics): They develop a theoretical framework based on the average firing rates of excitatory and inhibitory neuronal populations. This model simplifies the complex dynamics of individual neurons into a set of differential equations, allowing for analytical exploration of network behavior. They specifically incorporate spike-frequency adaptation, a phenomenon where neurons reduce their firing rate during sustained activity.
  • Large-Scale Network Simulations: To validate their theoretical predictions, they simulate large networks of spiking neurons. These simulations use a detailed model of individual neurons, including their membrane potential dynamics, synaptic currents, and afterhyperpolarization currents responsible for adaptation.
  • Parameter Variation: They systematically vary key parameters, such as the fraction of endogenously active cells, network connectivity (infinite range vs. local), and synaptic potential amplitudes, to understand their impact on network dynamics.

Datasets

Not discussed in the paper. The study relies on simulated data generated from their models and simulations.

Key Insights and Ideas

  • Endogenous Activity is Key: The presence of even a small fraction of "endogenously active" neurons (those that fire spontaneously) is critical for establishing stable, low firing rates. Without them, networks tend to be either silent or fire at high rates.
  • A Phase Transition: The fraction of endogenously active cells acts as a critical parameter. As this fraction increases, networks transition from silence to bursting, and then to steady low-rate firing.
  • Spike-Frequency Adaptation Induces Bursting: While not strictly necessary for low firing rates, spike-frequency adaptation plays a crucial role in inducing bursting. It can temporarily silence endogenously active cells, leading to a crash in firing rate and subsequent re-ignition.
  • Connectivity Matters, But Less So: Network connectivity influences the stability and oscillatory nature of firing patterns, but the fraction of endogenously active cells is the primary determinant of the basic firing mode (steady vs. bursting vs. silence).
  • Experimental Predictions: The theory yields testable predictions: isolated low-firing networks must contain endogenously active cells, and reducing this fraction should induce bursting.

Key Results

  • Theoretical Predictions: The authors predict a three-regime system controlled by the fraction of endogenously active cells: silence (zero fraction), bursting (intermediate fraction), and steady low-rate firing (high fraction). They also show that spike-frequency adaptation can induce bursting.
  • Simulation Validation: Large-scale simulations with spiking neurons confirmed these predictions. They observed transitions between silence, bursting, and steady firing as the fraction of endogenously active cells was varied.
  • Robustness: The model's predictions held across different network architectures (infinite range vs. local connectivity) and parameter choices, suggesting the underlying principles are general.
  • Experimental Corroboration (Mentioned): The paper references accompanying experimental work that validated these predictions, particularly the presence of endogenously active cells in low-firing networks and the induction of bursting by reducing their fraction.

Limitations and Future Work

  • Simplified Models: While the mean-field model captures essential dynamics, it is a simplification of the complex behavior of individual spiking neurons. The authors acknowledge that quantitative details might differ between the simplified model and full simulations.
  • Parameter Sensitivity: While the model is generally robust, certain parameters (like the strength of inhibitory-inhibitory coupling) can influence stability and lead to oscillations.
  • External Input: The current work focuses on intrinsic dynamics. Future work could explore how external inputs interact with these intrinsic properties to shape network activity in more realistic scenarios.
  • Structured Connectivity: The study primarily uses random connectivity. Investigating how more biologically realistic structured connectivity (e.g., cortical circuits) affects these dynamics is a natural next step.
  • Synaptic Adaptation: The paper focuses on spike-frequency adaptation. The role of synaptic adaptation in generating bursting patterns was briefly discussed and deemed less likely to produce the observed phenomena. Further exploration of its interplay with other mechanisms could be valuable.

Summary updated

Abstract

Many networks in the mammalian nervous system remain active in the absence of stimuli. This activity falls into two main patterns: steady firing at low rates and rhythmic bursting. How are these firing patterns generated? Specifically, how do dynamic interactions between excitatory and inhibitory neurons produce these firing patterns, and how do networks switch from one firing pattern to the other? We investigated these questions theoretically by examining the intrinsic dynamics of large networks of neurons. Using both a semianalytic model based on mean firing rate dynamics and simulations with large neuronal networks, we found that the dynamics, and thus the firing patterns, are controlled largely by one parameter, the fraction of endogenously active cells. When no endogenously active cells are present, networks are either silent or fire at a high rate; as the number of endogenously active cells increases, there is a transition to bursting; and, with a further increase, there is a second transition to steady firing at a low rate. A secondary role is played by network connectivity, which determines whether activity occurs at a constant mean firing rate or oscillates around that mean. These conclusions require only conventional assumptions: excitatory input to a neuron increases its firing rate, inhibitory input decreases it, and neurons exhibit spike-frequency adaptation. These conclusions also lead to two experimentally testable predictions: 1) isolated networks that fire at low rates must contain endogenously active cells and 2) a reduction in the fraction of endogenously active cells in such networks must lead to bursting.
BibTeX
@article{latham2000intrinsic,
  author = {Peter E. Latham and Barry J. Richmond and P. G. Nelson and Sheila Nirenberg},
  title = {{Intrinsic Dynamics in Neuronal Networks. I. Theory}},
  journal = {Journal of Neurophysiology},
  year = {2000},
  doi = {10.1152/jn.2000.83.2.808},
}
Figure