NOTE: There is an updated version of the simulation code as well as novel implementations in the Python and Julia languages here. It implements, as an example, the cortical microcircuit model of Potjans and Diesmann, Cereb. Cortex (2014).

This project provides simulation code associated with the publication

T. Schwalger, M. Deger, and W. Gerstner. Towards a theory of cortical columns: From spiking neurons to interacting neural populations of finite size. PLoS Comput. Biol., 13(4):e1005507, 2017.

The core simulation code is written in C. There are two libraries:

• glm_netw_sim_0.8.c for the microscopic (neuron-based) simulation
• glm_popdyn_1.1.c for the mesoscopic (population-based) simulation

The C libraries can be used either via a Python interface or directly in a C program.

Prerequisites for compiling c code

• gcc
• fftw3 (libfftw3-3 package in ubuntu)
• gsl (libgsl2 package in ubuntu)

Create python modules from c libraries

In the top level do:

make -Bf Makefile_fasthazard

This is a modification of the original pseudocode implementation in the paper by using a lookup table for calculating the hazard function. This yields slightly better performance. For the original implementation w/o lookup table use

make -Bf Makefile_orig

Examples

• For running a simulation of an excitatory-inhibitory network (both microscopic and mesoscopic, as in Fig.5 of the paper), change to EI_net_py/ and run

python ei_net_N000_p0.2.py

• For running a simulaion of a population of adapting neurons, change to uncoupled_py/ and run

python lif_adap.py