Here's the Pairwise Ising model, and the Pairwise-Indepeendet simplex model available.
Only the toy-dataset is available, since the other data used in the paper belongs to the "Blue Brain Project". This makes it problematic to use the Pairwise-Independent simplex model due to that the toy-dataset does not contain the connectivity. Hence, the simplices composed of the neurons are unknown. If you have you own data with the known connectivity there is a description in the pairwise_independent_simplex2 how to make the data compatible with the code.
Use py2.7 and conda 4.3.30.
Credit: Jan Humplik, GasÏper Tkačik. See their article, "Probabilistic models for neural populations that naturally capture global coupling and criticality" at http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1005763