Back to TonyÕs CAAM page

Last updated on July 25, 2007 by Anthony Kellems

 

Hippocampal Modeling and Simulation

Also see the hippos group web page

 

This research was conducted as part of the VIGRE Neuroscience PFUG run by Dr. Steve Cox.

 

 

UPDATE (7/25/07):

The parallel code is finished and I have made many improvements on its efficiency and I have cleared up lingering numerical errors. The beta version will be available in early-mid August!

 

UPDATE (6/29/07):

Results from recent weeks have vindicated the quadrature method in my thesis. Not only can I now offer solid proof that it is highly accurate by comparing it to very very fine timestepping simulations, but I have shown that for the CA3 network simulations the formulation of Rangan & Cai is not suitable. Hence the standard quadrature integral formulation and STI algorithm used in my masterÕs thesis is preferred for hippocampal simulations.

 

A beta version of the Integrate-and-Fire code will be available shortly, along with improved GUIÕs for simulation, stimulus editing, and plotting in MATLAB.

 

Documents:

Poster: Efficient and Accurate Neuronal Network Simulation

 

Codes:

Simulation Platform codes

MATLAB GUI's and Visualization Tools

 

 

Research and Goals:

Large-scale network simulations of the hippocampus are more challenging than single-cell simulations. I have built a software platform in C++ that is controlled by a MATLAB interface so that users can efficiently and accurately simulate large networks without having to write code, and I also designed visualization tools to elegantly display the resulting data and to perform some convergence analysis.

 

AppleMark

Top: Oscillatory firing pattern for a 100-cell network.

Bottom: Firing rate for the pattern shown above.

 

 

 

My work focuses on the Integrate-and-Fire neuron model. The software platform simulates networks of these cells using quadrature methods and timestepping schemes. Users retain control over the synaptic connections, the synaptic inputs, model parameters, and the parameters for solution methods through a GUI. See my masterÕs thesis for detailed information and results.

 

The goals for the summer of 2007 are to expand this simulation platform and make it accessible to other researchers, primarily the other students and faculty involved in our research group. Specifically, I aim to accomplish the following: