INTRODUCTION (The NEURON Simulation Environment)

1. INTRODUCTION

NEURON (Hines 1984; 1989; 1993; 1994) provides a powerful and flexible environment for implementing biologically realistic models of electrical and chemical signaling in neurons and networks of neurons. This article describes the concepts and strategies that have guided the design and implementation of this simulator, with emphasis on those features that are particularly relevant to its most efficient use.

1.1 The problem domain

Information processing in the brain results from the spread and interaction of electrical and chemical signals within and among neurons. This involves nonlinear mechanisms that span a wide range of spatial and temporal scales (Carnevale and Rosenthal 1992) and are constrained to operate within the intricate anatomy of neurons and their interconnections. Consequently the equations that describe brain mechanisms generally do not have analytical solutions, and intuition is not a reliable guide to understanding the working of the cells and circuits of the brain. Furthermore, these nonlinearities and spatiotemporal complexities are quite unlike those that are encountered in most nonbiological systems, so the utility of many quantitative and qualitative modeling tools that were developed without taking these features into consideration is severely limited.

NEURON is designed to address these problems by enabling both the convenient creation of biologically realistic quantitative models of brain mechanisms and the efficient simulation of the operation of these mechanisms. In this context the term "biological realism" does not mean "infinitely detailed." Instead it means that the choice of which details to include in the model and which to omit are at the discretion of the investigator who constructs the model, and not forced by the simulation program.

To the experimentalist NEURON offers a tool for cross-validating data, estimating experimentally inaccessible parameters, and deciding whether known facts account for experimental observations. To the theoretician it is a means for testing hypotheses and determining the smallest subset of anatomical and biophysical properties that is necessary and sufficient to account for particular phenomena. To the student in a laboratory course it provides a vehicle for illustrating and exploring the operation of brain mechanisms in a simplified form that is more robust than the typical "wet lab" experiment. For experimentalist, theoretician, and student alike, a powerful simulation tool such as NEURON can be an indispensable aid to developing the insight and intuition that is needed if one is to discover the order hidden within the intricacy of biological phenomena, the order that transcends the complexity of accident and evolution.

1.2 Experimental advances and quantitative modeling

Experimental advances drive and support quantitative modeling. Over the past two decades the field of neuroscience has seen striking developments in experimental techniques that include

high-quality electrical recording from neurons in vitro and in vivo using patch clamp
multiple impalements of visually identified cells
simultaneous intracellular recording from paired pre- and postsynaptic neurons
simultaneous measurement of electrical and chemical signals
multisite electrical and optical recording
quantitative analysis of anatomical and biophysical properties from the same neuron
photolesioning of cells
photorelease of caged compounds for spatially precise chemical stimulation
new drugs such as channel blockers and receptor agonists and antagonists
genetic engineering of ion channels and receptors
analysis of mRNA and biophysical properties from the same neuron
"knockout" mutations

These and other advances are responsible for impressive progress in the definition of the molecular biology and biophysics of receptors and channels, the construction of libraries of identified neurons and neuronal classes that have been characterized anatomically, pharmacologically, and biophysically, and the analysis of neuronal circuits involved in perception, learning, and sensorimotor integration.

The result is a data avalanche that catalyzes the formulation of new hypotheses of brain function, while at the same time serving as the empirical basis for the biologically realistic quantitative models that must be used to test these hypotheses. Some examples from the large list of topics that have been investigated through the use of such models include

the cellular mechanisms that generate and regulate chemical and electrical signals (Destexhe et al. 1996; Jaffe et al. 1994)
drug effects on neuronal function (Lytton and Sejnowski 1992)
presynaptic (Lindgren and Moore 1989) and postsynaptic (Destexhe and Sejnowski 1995; Traynelis et al. 1993) mechanisms underlying communication between neurons
integration of synaptic inputs (Bernander et al. 1991; Cauller and Connors 1992)
action potential initiation and conduction (Häusser et al. 1995; Hines and Shrager 1991; Mainen et al. 1995)
cellular mechanisms of learning (Brown et al. 1992; Tsai et al. 1994a)
cellular oscillations (Destexhe et al. 1993a; Lytton et al. 1996)
thalamic networks (Destexhe et al. 1993b; Destexhe et al. 1994)
neural information encoding (Hsu et al. 1993; Mainen and Sejnowski 1995; Softky 1994)

Address questions and inquiries to
Michael Hines or Ted Carnevale

Digital preprint of "The NEURON Simulation Environment" by M.L. Hines and N.T. Carnevale, Neural Computation, Volume 9, Number 6 (August 15, 1997), pp. 1179-1209. Copyright © 1997 by the Massachusetts Institute of Technology, all rights reserved.
HTML formatting and graphics for page navigation copyright © 1997 by N.T. Carnevale amd M.L. Hines.