As has been seen, there are a bewildering number of neural modeling programs available for the asking at various university URLs on the Internet. In selecting one, the potential user must weigh and decide:
1. On what platform (type of computer and OS) the program will be run.
2. At what level will the modeling will take place? That is, single neuron, small groups of neurons (e.g., CPGs), large groups of neurons (e.g., the retina), or very large neural networks (e.g., brain and its components), or at all levels.
3. What degree of flexibility is required in the fine structure of the model? That is, how many different types of ion channels are required, and can their parameters be easily edited?
4. Can the user trade off detail for number in simulating networks? For example, can one abandon the HH formalism for spike generation and use RPFM (leaky integrator) spike generators? Are the action of chemical synapses describable by the two-state, alpha model for subsynaptic membrane conductance change, or can one directly generate epsps (or ipsps) from presynaptic spikes, and condition and sum these transients to form the generator potential directly at the SGL?
5. What level of support is available? That is, is there a text or an online tutorial available? Are there example simulations and neuron parameter libraries available online?
The major neural modeling programs such as GENESIS and NSL are complex and detailed; they have a huge effort overhead for the user to learn to use them effectively. They have evolved over time through the efforts of many neurobiologists, computational neuroscientists, and computer scientists. Both GENESIS and NSL have large cadres of users and, better yet, are supported by textbooks with examples, etc. The Shepherd Lab at Yale offers an online, cerebral neuron database (NeuronDB)
with simulation parameters listed for approximatley 20 types of neurons, including spinal type Ia interneurons and motoneurons. It lists voltage-gated conductances, neurotransmitter receptors and substances, canonical forms of neurons (dendrites, soma, axon hillock, axon, and axon terminal), as well as tools for the integration of these properties in a given type of neuron, and comparison of properties across different neuronal types. The URL for the NeuronDB is http://spine.med.yale.edu /neurondb/.
To run many simulation programs, (free) Linux can be easily installed on a Pentium PC. (Actually, it is easier to install Linux if one first removes Windows; the PC is then dedicated to neural modeling and other applications compatible with Linux.)
Because of the intellectual and time overheads associated with mastering the large simulation programs, the entry-level neural modeler should consider first using Simnon to model single neurons and small neural circuits. There is very little overhead in using Simnon, and the model structure is in terms of algebraic, firstorder ODEs. Neural models with Simnon are well suited to the phenomenological locus approach, where individual details such as individual ionic conductance functions and connected compartments are avoided in favor of directly simulating epsps and ipsps, modeling their propagation to an RPFM SGL through low-pass, dendrite transfer functions, and then modeling the spike propagation on the axon as it travels from SGL to synapse with a pure transport lag. Simnon can easily handle the ODEs describing parametric conductance changes, however.
As has been seen, any level of detail (and realism) is possible in neural modeling, given the wide choices of programs available. With all of this preoccupation with modeling, one should not forget from where the numbers come that are used in the models, and the model architecture itself. That is, actual "wet," experimental neurophysiology. This is where it began, and is where it should end.
Was this article helpful?
This guide will help millions of people understand this condition so that they can take control of their lives and make informed decisions. The ebook covers information on a vast number of different types of neuropathy. In addition, it will be a useful resource for their families, caregivers, and health care providers.