/CGNetwork_MEPMoment

SZT Dimensional Reduction of Emergent Spatiotemporal Cortical Dynamics via a Maximum Entropy Moment Closure

Primary LanguagePython

CGNetwork_MEPMoment

SZT Dimensional Reduction of Emergent Spatiotemporal Cortical Dynamics via a Maximum Entropy Moment Closure

CG Network for V1 via a Maximum Entropy Moment Closure

The code is tested in Anaconda 3 (Python 3.6 and Python 3.7) on Windows 10.

We also provide the large code (conductance-based I&F spiking neuron model mentioned in our paper), details are in another repository on this homepage, https://github.com/shaonannan/largecode_IFSpiking

The structure of this Python program uses Allen's simulation platform DiPDE for reference. DiPDE (dipde) is also an excellent simulation platform for numerically solving the time evolution of coupled networks of neuronal populations. You can find their algorithms in,

Website: © 2015 Allen Institute for Brain Science. DiPDE Simulator [Internet]. Available from: https://github.com/AllenInstitute/dipde.

We are still refining this code, especially the 'MFEs' parts (simulation.py, function 'def getMFE_ifdyn' in utilities.py and etc.) which account for transient dynamics. Another project we are following up will be updated on this page.

Get started with default configurations

In EXAMPLE.py file

cmd = 'python RunVoltageMoment.py -stimulus 0'

Use network with default parameters to run simulation in counterchanging on/off stimulus

cmd = 'python RunVoltageMoment.py -stimulus 1'

Use network with default parameters to run simulation under ON-/OFF-LMI stimulus

You can also change the parameters by adding other parameters in the command, e.g. changing $S_{EE}$

cmd = 'python RunVoltageMoment.py -stimulus 1 -see 1.20'

Simulating results in Code_directory/... and save as .mat file.

Network and Maximum Entropy Moment

See 'RunVoltageMoment.py' for the architecture of our CG network

See 'internalpopulation.py' and 'utilities.py' for our reduced method (Maximum Entropy Moment closure)

See the source code to know details about how this algorithm works.