Getting Started With Realistic Neural Modeling

Some suggestions for using the material on this CDROM for self-paced study

Introductory Background (no programming or knowledge of GENESIS needed)

The genesis/Scripts directory in any of the unpacked GENESIS distribution archive files contains several tutorial simulations (squid, cable, neuron, burster, traub91, purkinje, piriform), described in Part I of "The BoG".

The Tutorials directory on this CDROM contains:

Some Simulation Exercises (non-programming)

  1. Introduction to the Hodgkin-Huxley model. Chapter 4 of the BoG provides the theoretical background and describes the use of the genesis/Scripts/squid simulation. Try exercises 7 and 8 in this chapter. These will help you to understand the process by which action potentials are generated, and the basis for effects such as refractory period and post-hyperpolarization rebound (anode-break). For help (or to peek at the answers), see the Lectures on Computational Neuroscience section on the Hodgkin Huxley model.

  2. The effect of active channels in the dendrites on burst firing. The genesis/Scripts/traub91 directory contains a simulation based on a 19-compartment hippocampal pyramidal cell model. Run the simulation, following the instructions in the README file, and bring up the HELP form. Perform the suggested experiments in order to understand how calcium channels and calcium-dependent potassium channels in the dendrites produce burst firing. Chapter 7 of the BoG provides the background for understanding the types of ionic conductances that lead to this behavior in pyramidal cells. This simulation is also described in the Lectures on Computational Neuroscience.

  3. The De Schutter and Bower Purkinje cell model. The full De Schutter and Bower (1994) detailed Purkinje cell model is available as a tutorial in the genesis/Scripts/purkinje directory. The file help.txt (which can be viewed with the HELP button) describes the model and the various types of synaptic input that you can apply.

  4. The Wilson and Bower piriform (olfactory) cortex model. Run the simulation in the genesis/Scripts/piriform directory, following the instructions in the README.1st file, after reading BoG Chapter 9. Try some of the suggestions in the "help" window and the exercises at the end of the chapter.

Creating Realistic Neural Models with GENESIS

The GENESIS Modeling Tutorial

This multi-part hands-on tutorial in creating GENESIS simulations is the main tutorial of the package, and is intended to be a "quick start" to creating simulations with GENESIS. It should give you the tools and enough information to let you quickly begin creating cells and networks with GENESIS, making use of the provided example simulations. You will find that the modular object-oriented nature of GENESIS makes it easy to create new simulations using parts taken from existing simulations, such as the example scripts in the directories Tutorials/cells and Tutorials/networks.

The GENESIS Reference Manual

The Tutorials directory also contains a copy of the hypertext reference manual for GENESIS commands and simulation object types (classes).

Exercises in GENESIS programming:

There are exercises throughout the GENESIS Modeling Tutorial. Here are some more suggestions:

  1. A simple network model. To get started with network modeling, work through enough of the tutorial to understand the section on "Creating large networks with GENESIS". This section describes the construction of the model in Tutorials/networks/RSnet. This is a simple network, consisting of a grid of simplified neocortical regular spiking pyramidal cells, each one coupled with excitatory synaptic connections to its four nearest neighbors. This might model the connections due to local association fibers in a cortical network. The example simulation was designed to be easily modified to allow you to use other cell models, implement other patterns of connectivity, or to augment with a population of inhibitory interneurons and the several other types of connections in a cortical network. Try some of the modifications suggested at the end of the tutorial section.

  2. Extending the RSNet model to include inhibitory interneurons. This "extended version" of the Tutorials package has three extra network models included in Tutorials/networks.

    The Tutorials/networks/GENESIS_HH_benchmark directory contains a GENESIS implementation of a 4000 neuron version of Vogels and Abbott (2005) model, using neurons with Hodgkin-Huxley dynamics, instead of integrate-and-fire neurons. This was used as a benchmark for a review of neural simulators in a paper in preparation by Destexhe et al. (2006). The README file describes how to run the dualexpVA-HHnet.g simulation, and some of the experiments that can be performed with the GUI.

    From the standpoint of realistic neural modeling with GENESIS, this is not a very interesting simulation. The single compartment neurons have only fast Na and delayed rectifier K conductances, and fire tonically, much like the original integrate-and-fire neurons. The behavior of the simulation is similar to the original, but runs much more slowly. There is little point in using GENESIS for such a model.

    However, this well-documented simulation is offered as a useful starting point for those who would like to perform some "GENESIS script hacking" and try some simple modifications to the script, to use more realistic neuron models and to experiment with modifications of the network connectivity. The README file lists some small modeling projects that could be carried out with modifications to this simulation.

  3. Understanding a detailed network model. The Tutorials/networks/hippo2 directory contains a small hippocampal network model consisting of 72 pyramidal cells (Traub 1994 model) and 18 interneurons (Traub 1995 model) with realistic patterns of connections. The directory contains detailed documentation and papers describing the model, but this is not a tutorial simulation, and will take some work to understand. It runs without a GUI and writes data (a lot!) to files.

    The Tutorials/networks/TurtleVisCortex directory contains another large research simulation, the Nenadic, Ghosh and Ulinsky large scale model of turtle visual cortex. Although it was not intended to be a tutorial, it is very well documented, with modular code. In this simulation, cells are not located on a regular grid, but are placed according to coordinates that are read from files. As an exercise, or the start of a different cortical modeling proect, the simulation could be generalized by using the cell reader to create cell prototypes, instead of creating cells "the hard way", as in BoG Chapters 13-15.

Other resources for realistic neural modeling

The latest edition of the classic "Synaptic Organization of the Brain" contains multi-authored reviews of our current understanding of the "wiring diagrams" of different portions of the brain. It should be the starting place for any modeling study of brain regions.

Shepherd, GM (2003) "The Synaptic Organization of the Brain", 5th ed, edition, Oxford University Press, NY

The November 2005 special issue on Realistic Neural Modeling in the electronic journal Brains, Minds, and Media contains published versions of the modeling tutorials that were presented at the 2005 WAM-BAMM conference and GENESIS users meeting. The Table of Contents for this issue with links to view or download the articles can be found at http://www.brains-minds-media.org/237/. Those who are interested in learning the details of creating realistic single cell models using experimental data may want to read the advanced tutorial by Dieter Jaeger in this issue.


Return to the Ultimate GENESIS CDROM main index

Last updated on: Fri Jul 6 13:28:03 MDT 2007