FEB MAR Apr 26 2004 2005 2006 13 captures 13 Oct 02 - 26 Mar 05

Close Help

Computational Neuroscience and Whole Brain Emulation Related Models

Memory Models with Oscillatory Short-Term Recurrent Synaptic Intermediate-Term Memory

Connectivity Memory Model

This memory model combines a local distribution of connectivity, network-wide interneuronal firing with an independent interneuronal after-hyperpolarization, full-strength synaptic efficacy with weight redistribution, absolute and relative refractory periods, and the basic oscillatory short-term memory with recurrent intermediate-term memory (LTP and LTD) functions proposed by Jensen et al.
In the works: automatic computation of the WCORRECTION parameter; synapse-specific choice of pre- and post-synaptic ITM (the ``no ITM'' option is already supported).
connectivity-0.9.5.tar.gz - compile with: make connectivity
Usage Documentation.

Corrected Recurrent Efficacies

This model corrects the basic recurrent intermediate-term memory (LTP and LTD) functions proposed by Jensen et al. to achieve full-strength synaptic efficacy with weight distribution according to the LTP and LTD dynamics. The model also adds a physiologically plausible absolute refractory period that improves the segmentation of short-term memory cycles.
corrected-1.0.tar.gz - compile with: make corrected

Jensen STM and ITM Model

This model produces the oscillatory short-term memory and intermediate-term memory based on LTP and LTD functions, as proposed by Jensen et al. (Learning and Memory, 1996). Additional code caters to the measurement of cell-specific component functions.
jensen-1.0.tar.gz - compile with: make jensen

connectivity-all-0.9.5.tar.gz - Full collection, compile with: make

proccfg: An architecture generator for the model input format used in connectivity.c. Architecture generation guidelines are obtained from a configuration file "inputname.cfg", using a simple interpreted command language.
Usage Documentation.

Sample input data:

1. fullSTM-nonoise: Demonstrates short-term memory repetition of stimulus patterns in the absence of intermediate-term memory learning and noise for the corrected short-term memory capacity of about four items.
   Usage Documentation.

Spiker Model

• C++ Program: Spiking Neuron Simulation (main-spiker.cpp)
• C++ Include File: Spiking Neuron Functions (spiker.cpp)
• C++ Header File: Spiking Neuron Functions (spiker.hh)
• A voltage clamped spiking neuron (clamp.m)
• Octave script: Interface to ``spiker'' randomly connected spiking neuron network experiment (spiker.m)
• Octave script: Interface to ``spiker'' hand-coded spiking neuron XOR network experiment (spikexor.m)
• Octave script: Interface to several ``spiker'' experiments (neuron.m), and a slightly older version (neuron_bak.m)

Neuronal Structures Emulation (or the LaTeX source), also called Heterogenous Neurological Reconstruction and Emulation: A Review of Physiologically Extended Spiking Neuron Models, Interoperability through Uniform Formulation, uses the following Octave scripts for functional analyses of GESN models.

• Octave script: Simulations of Hodgkin-Huxley equations for the giant squid axon (hhsquid.m)
• Octave script: Analysis of the components of the Jensen-Lisman-Idiart spiking neuron model (jensencomp.m)
• Octave script: Comparing learning window functions of different spiking neuron learning rules (learnwindow.m)
• Octave script: Relative timing of a presynaptic spike at $\Ht_i$ and the onset of an evoked postsynaptic potential at $\tpspi$ (tpspi.m)
• Octave script: Comparison of separate V_m(t) & V_m^thres(t) and V_m(t) combined functions (vmvthres.m)
• Octave script: Spike Response Model: Postsynaptic Potential Response Kernel (srmpsp.m)

Generic Enhance Spiking Neuron (GESN) Models

• The GESN Library
• C++ Program: Source code to ``gesn'', a program interface to several GESN model networks (main.cpp)
• Mekfile for ``gesn'' (and ``spiker'' above) (Makefile)
• Octave script: Demonstrate Jensen and Lisman response and LTP/LTD characteristics (jensencomp.m)
• Octave script: Create a ``.cmd'' file for hippwalk.m to generate an input data set representing regular zig-zag motion through an environment (hippzigzag.m)
• Octave script: Create a ``.ang'' file of angles to cues in an environment, corresponding to locations walked to according to ``.cmd'' command file or standard input (hippwalk.m)
• Octave script: Convert a ``.ang'' file or standard input into a sparse (single bit per cue) binary encoding of angles (hippang2in.m)
• Shell script: Combines input data set generation stages from ``.cmd'' to ``.ang'' to ``.in'' formats, using ``hippwalk.m'' and ``hippang2in.m'' (hippwalk2ang)

Some graphical representations concerning GESN (and hippocampal) modeling.
These are currently referenced here, as they are not yet included in an encompassing text and were not generated by a model or script in ~/src/nnmodels, but manually created with XFig to clarify some modeling considerations.
• Process diagram of $\GESN{HH}{IF0}$ model functions (gesn_hh_if0_algorithm.fig)
• Process diagram of $\GESN{IF0}{Com}$ model functions (gesn_if0_com_algorithm.fig)
• Protocol diagram for the generation of input data sets from angles to cues along a path of motion through an environment (hippwalk-protocol.fig)
• Diagram explaining the system of modular component resources that form the basis of the GESN model family (modular-components.fig)
• Diagram explaining the distribution across components of processes that accomplish plasticity (modular-learning.fig)

Miscellaneous Models

Speech Recognition Neural Network

• Concept: A Speech Recognition Neural Network Model