www.neuron.yale.edu

Hi NEURON folks.

We're looking for an example of a NEURON model of an unmyelinated cable with an extracellular potential modeled. For example, an HH cable.

We've found some relevant examples in the NEURON src files, but they assume (to use NEURON terminology) that nlayer=2. We basically want a simple example where nlayer=1 (I think).

We think we've worked out the simiplifications, but I'd like a working model to compare ourselves against, and someone must have done this some time somewhere. To me, the way NEURON handles extracellular potentials is not intuitive (nor are extracellular potentials themselves at times!)

Thanks.

Just a piece of information stumbled upon while looking for something completely different something todo with searching for matlab in the NEURON site.

Anyway I obtained the following link:
http://www.neuron.yale.edu/neuron/nsm2004/abstracts.pdf

With this abstract:

Carl Gold: "Simulating extracellular fields with NEURON and the Line Source
Approximation"
The Line Source Approximation (LSA) is an efficient and accurate method for calculating the
extra-cellular electric potential resulting from the currents distributed over the entire 3-D
membrane surface of a neuron. The LSA consists of approximating the 3-D distribution of
membrane currents on a cylindrical neurite surface as a line source of current down the center of
the neurite. With this simplification the calculation of extra-cellular voltage simplifies to an
analytic expression at the cost of a minimal loss of accuracy (Holt and Koch, 1998). To apply the
Line Source Approximation, a precise 3-D neuron morphology is obtained from histological
measurements, the neuron is simulated in the NEURON, and the currents for all compartments
are written to a data file at discrete time steps. The data is then loaded into a Matlab program
which efficiently performs the LSA calculation on matrix representations of the neuron geometry
and membrane currents. The talk will review the theory of the LSA, and discuss practical aspects
of the implementation in NEURON and Matlab. The talk will showcase the application of the
technique in a theoretical comparison of extra- and intra-cellular recording action potentials from
hippocampal neurons (Henze, et. al. 2000).

I hope it is usefull, good luck!
Raj

rbutera wrote:We're looking for an example of a NEURON model of an unmyelinated cable with an extracellular potential modeled. For example, an HH cable.

We've found some relevant examples in the NEURON src files, but they assume (to use NEURON terminology) that nlayer=2. We basically want a simple example where nlayer=1 (I think).

The Programmer's Reference documentation of the extracellular mechanism
http://www.neuron.yale.edu/neuron/stati ... racellular
starts with the sentence "Adds two layers of extracellular field to the section."
The diagram is correct, and there are two layers, but calling this a "field" may invite
misunderstanding--as Orwell said in his essay "Politics and the English Language,"

. . . the slovenliness of our language makes it easier for us to have foolish thoughts.

(available online at many sites including http://www.mtholyoke.edu/acad/intrel/orwell46.htm).

extracellular is exactly what the diagram suggests--just two layers of resistors and
capacitors. The default values are such that it has no effect on a simulation. The
resistive parameters in the membrane-adjacent layer allows nearest-neighbor
extracellular interactions, e.g. mimicking a nonmyelinated axon with a thin extracellular
layer of electrolyte, suspended in a bath of mineral oil. The capacitive parameter in
that layer allows a simple representation of the properties of a myelin sheath. The
outermost layer, which contains e_extracelluar in addition to its own resistors and
capacitors, is useful principally as a hook for whatever extracellular driving force
you might wish to apply to the cell. It can be used to widen the range of interactions,
but they can still only be "nearest neighbors" rather than truly 3 dimensional.

So the extracellular mechanism can be used to implement a reasonably accurate
representation of an axon in oil, or a myelinated axon, but it is not a generic
representation of extracellular space. It is a good way to discover memrane current,
and e_extracellular can be exploited as an extracellular stimulus (Vector play into
e_extracellular, for instance). Carl Gold's work on extracellular recording did this, and
you can find my own strategy for using extracellluar as a means to stimlulate and
record from cells with complex geometry elsewhere in this Forum
Making and using models with NEURON / Modeling networks / extracellular field generation
https://www.neuron.yale.edu/phpBB2/view ... racellular

Some investigators have used extracellular in hybrid simulations that pair NEURON
(to model neurons) with a true field simulator (e.g. MAXWELL), each simulator
computing its own stuff during a time step and then exchanging results before
advancing again.