More than three decades ago, I moved from Kenneth ("Kacy") Cole's lab at the National Institutes of Health to develop my own lab at Duke University. Over this period my students and I have used the experimental and computational tools available to us to work at understanding how bits and pieces of nerve cells work. My major focus has been on how a neuron's morphology (form) and ionic channels interact to generate and propogate nerve impulses.
Over the years, we gradually improved our simulation tools but when Michael Hines joined us, they took on a professional elegance. His first tool, NEURON 1. (originally called CABLE), was designed to provide the convenience of an interpreter (with a built-in editor) and several built-in channel types for the user. The user wrote interpreter statements to control the input, output, parameters, etc. Some of these intrepreter lines called for compiled "number crunching" routines which ran at much higher speeds. The compiled code itself incorporated a major new algorithm by Hines, giving a breakthrough in the speed of solution of the matrix equations describing branched nerves. With NEURON 1, we were able finally to assemble all of the bits and pieces together and describe a realistic nerve cell in full detail, including synaptic inputs.
In NEURON 2. he made additional improvements:
Now, with the recently added graphical interface called "InterViews", a new level of elegance and convenience has been achieved and we call this version NEURON 3.0. It allows the user to step back from the details of programming and to concentrate on the functioning of the nerve or network under investigation.
This book, written in "digiscript", is in response to increasing and incessant requests for a manual on how to use NEURON 3.0 for simulations. In order generate the energy for this task, I decided that it must include additional items which I consider important for modeling.
As I approached the writing of this volume, I realized also that I had been in an unusual position as the fields of precise voltage clamp measurements of nerve membrane electrical characteristics and computational neuroscience developed. Not only had I worked with "Kacy" Cole but had also known and or worked with many of the persons involved in the development of the field of computational neurosciences since its inception with the work of Hodgkin and Huxley. Kacy arranged for the first digital computer calculations of their famous equations and I had plotted out the results from printed tables. I could present a unique and personal view of this history.
In addition, I had the unique privilege of working under Art Vance at the RCA Labs in Princeton, NJ, for a few years following my PhD in physics. A man of amazing intuitive feel for the performance of dynamic systems, Art independently developed negative feedback circuits and operational amplifiers which became the basis of analog computers and many precision instruments. He taught me that appropriate application of negative feedback can produce almost ideal devices to measure current or voltage. Although analog computers have been largely outmoded by digital computers, operational amplifiers live on in a wide variety of precision measurement instruments.
Using such operational amplifier techniques learned from Vance, I was able
to design both very high impedance feedback circuits for measurement of
membrane potentials via microelectrodes and extremely low impedance
clamp circuits in Kacy Cole's lab.
Operational amplifiers in analog computers are used in the current measuring mode; electronic feedback makes them nearly perfect devices for measuring current because they have almost no voltage generated at the current summing point (in contrast to most circuits which measure current by virtue of measuring the voltage across a small known resistor). Recognizing this, I introduced these "virtual ground" circuits as current-to-voltage converters in the above noted circuits to voltage clamp the squid giant axon. Although at that time it was difficult to convince others of their power, such circuits are now routinely used for current measurement. In fact, this is the basic circuit in the now ubiquitous patch clamp providing ever mode data for simulations to fit.
I hope to be forgiven for taking the liberty of presenting my own experience in the field of nerve modeling and for failing to point out the contributions of many others in its development.