Synapse with exponential decay

Physical System

A "fast" onset synapse.

Model

	g = gmax*exp(-t/tau)
	i = g*(v - e)

Simulation

expsyn.mod

text of expsyn.hoc file

This simulation consists of three spike train generators connected to one synapse.

These files are copies of the expsyn example files (along with pregen.mod in nrn/examples/nrniv/netcon. By running

	special expsyn.hoc -
the example will start with a PointProcessGroupManager that allows you to vary the parameters for the three spike train generators and the synapse. Another panel shows the weight and delay for the three connections.

Exercises

1) Compare the conductance change when the synaptic decay time is 100 (ms), 10(ms), and 1(ms)

2) When spikes arrive at the same time, does summation hold?

3) Is it safe to use a Network Connection delay of 0?

4) Compare results with fixed step and global variable time step method. With the variable time step method the default local absolute error tolerance can be obtained by typing

	cvode.atol
Compare this value with the synapse conductance state g. The absolute local step error tolerance can be changed with
	cvode.atol(1e-6)

5) With a spike generator short interval of PI/2 (ms) print the times of the synaptic conductance discontinuities for the fixed and variable time step methods as a function of discontinuity number.

Note: If you pick a Graph line into the vector clipboard, you can then use the Vector methods. Here are examples of some pertinent Vector methods applied to the clipboard:

hoc_obj_.c.printf
prints a copy of the vector

hoc_obj_.c.deriv(1,1).printf
prints the euler derivative

hoc_obj_.c.deriv(1,1).indvwhere(">", 1e-5).printf
prints the indices

hoc_obj_.ind(hoc_obj_.c.deriv(1,1).indvwhere(">", 1e-5).add(1)).printf
prints the peak values

hoc_obj_[1].ind(hoc_obj_.c.deriv(1,1).indvwhere(">", 1e-5).add(1)).deriv(1,1).printf
prints the time intervals between synaptic discontinuities

SubNotes:
1. These emacs commands work for command line editing from the console (MSWin: NEURON's interpreter window).
^P     previous line (a crude history function)
^A     front of line
^E     end of line
^B     backward character
^F     forward character
2. Long command lines can be constructed by revision + accretion (recall a prior line, make changes and add new stuff to it).


NEURON summer course
Copyright © 1998, 1999 by N.T. Carnevale and M.L. Hines, all rights reserved.