granule

This directory contains a hippocampal granule cell model implemented by Erik De Schutter. This was used in his model of Ca diffusion in the hippocampus granule cell, which was described in:

E. De Schutter and J.M. Bower: Sensitivity of synaptic plasticity to the Ca2+-permeability of NMDA-channels: a model of long term potentiation in hippocampal neurons. Neural Computation 5: 681-694 (1993).

ABSTRACT

We have examined a model by Holmes and Levy (1990) of the induction of associative long term potentiation (LTP) by a rise in the free Ca2+ concentra- tion ([Ca2+]) after synaptic activation of dendritic spines. The previously reported amplification of the change in [Ca2+] caused by coactivation of several synapses was found to be quite sensitive to changes in the perme-ability of the N-methyl-D-aspartate (NMDA) receptor channels to Ca2+. Varying this parameter indi-cates that maximum amplification is obtained at values that are close to Ca2+ permeabilities reported in the literature. However, amplification fails if per-meability is reduced by more than 50%. We also found that the maximum free [Ca2+] reached in an individual spine during synaptic coactivation of several spines, depended on the location of that spine on the dendritic tree. Distal spines attained a higher [Ca2+] than proximal ones, with differences of up to 80%. The implications of this result for the uniformity of induction of associative LTP in spines in differ-ent regions of the dendrite are discussed.

IMPLEMENTATION

The script AUTO_CA_fract.g outputs [Ca2+] concentration in the head of 7 spines in the file 'fname'. The variable CA_FRACT controls the relative Ca2+permeability of the NMDA channel (1 corresponds to permeability as in Mayer and Westbrook (1987)). After loading the script, ampli96 will co-activate 96 spines at 200Hz, ampli1 will activate 1 spine at 200Hz.

The two files are:

granule.p: a sketchy representation of the cell

AUTO_CA_fract.g: a complete script controlling the simulation. Is reasonably well documented inside. Produces only file output. To run the simulation, include this (it loads the cell) and type do_ampli1 (activates 1 spine) or do_ampli96 (activates 96 spines).

NOTE: The script AUTO_CA_fract.g was written for GENESIS 1, and will need to be converted for GENESIS 2. A first pass at conversion may be done with the genesis/bin/convert program (see genesis/Doc/Convert.txt), but some changes will have to be done manually. The variable user_intmethod is set to an obsolete method (4), and should be changed to Exponential Euler (0) or an implicit method (10 or 11) with the use of hsolve.

The GENESIS 1 difpool and poolbuffer objects have been replaced by difshell and fixbuffer in GENESIS 2. However, there are some differences in the naming of some fields. These changes in the script will have to be made by hand. To facilitate this, the old GENESIS 1 documentation (plain ASCII text) files difpool.doc and poolbuffer.doc have been included here. Also, the scripts provided in the GENESIS distribution in genesis/Scripts/examples/spinedemo show how to implement the same Ca diffusion mechanism with GENESIS 2 objects. If you convert this script to GENESIS 2, please submit your script to http://genesis-sim.org.