The scripts in these subdirectories were contributed by Adam Strassberg to
demonstrate his GENESIS objects, "Napores" and "Kpores", which are
documented in Doc/Napores.doc and Doc/Kpores.doc.  These objects are used to
simulate a population of ion channel proteins (pores) embedded in a patch of
membrane over an isopotential region.  Each individual pore undergoes
standard Markov kinetics.  Further details are given in:

Strassberg, A. F., DeFelice, L. J. (1993) "Effects of single channel kinetics
upon transmembrane voltage dynamics" in Computation and Neural Systems,
109-113, F. H. Eeckman and J. M. Bower (eds.), Kluwer Academic Press.

Adam F. Strassberg and Louis J. DeFelice (1993) "Limitations of the
Hodgkin-Huxley Formalism: Effects of Single Channel Kinetics upon
Transmembrane Voltage Dynamics", Neural Computation, 5:6.

UnClamp - simulates the response of a space-clamped patch of squid giant
    axon membrane to a step current injection, using Napores and Kpores.

VClamp - simulates the response of a space-clamped patch of squid giant axon
    to a step change in membrane voltage, using Napores and Kpores.

ConClamp - simulates the continuous model with standard Hodgkin-Huxley
    channels for comparison with the two cases above.  The main simulation
    scripts are ConClamp.g and ConVclamp.g.

------------------------------------------------------------------------------

Addition by Dave Beeeman for GENESIS 2.2:

We have discovered an error in the units used to calculate current densities
in these scripts.  The value shown as 100 pA/m^2 is actually 0.1 pA/m^2.  As
the email address given for Strassberg is no longer valid, please send any
other comments, suggestions or corrections to genesis@bbb.caltech.edu.

The paper by Strassberg and DeFelice reveals that the continous dynamics of
the Hodgkin-Huxley equations is not always sufficient to describe neuron
firing when the density of channels is low or the cell is small.

There has recently been a great deal of interest in the effect of the 
stochastic opening of channels on firing, even with moderate numbers of
channels.  Near the threshold for firing, the number of open channels
is small, and stochastic behavior can be important.  For example, see:

E. Schneidman, B. Freedman, and I. Segev (1998) Ion channel stochasticity may
be critical in determining the reliability and precision of spike timing,
Neural Comp. 10: 1679-1703.

The pore objects are useful for modeling this kind of behavior.  However,
the cell reader does not recognize these objects.  In order to make them
more similar to other GENESIS channel objects, and to make them usable with
the readcell "*compt" option, we have created the the extended objects
Na_squid_markov and K_squid_markov from the pore objects Napores and Kpores.
The script tut3markov.g in the markov directory illustrates the use of these
with a simulation based on tutorials/tutorial3.g.   It uses the prototype
Markovian (stochastic) channels defined in markovchan.g instead of the
hh_channels defined in hhchan.g, and uses readcell to build the cell.

This simulation is necessarily rather slow for larger compartments, because
of the large number of random numbers that are generated when there are many
pores.  It is likely that the algorithm used in the Schneidman et al. paper
is more efficient than the one implemented in the GENESIS pore objects.  If
you should improve upon these objects, please contact us at
genesis@bbb.caltech.edu.

