- PRG_function.m:
This function implements the Phenomenological Renormalization Group (PRG) method, introduced by Meshulam et al. (2018, 2019). It also can implement the extended version of PRG based on connectivity.
Within this method, the collective activity is iteratively coarse-grained by grouping maximally correlated variables (or maximally coupled variables in the connectivity-based case). At each coarse-graining step k=0,1,…,kmax, clusters of size K=2^k are built, resulting in a system of N/K coarse-grained variables and successively ignoring degrees of freedom. This code computes several observables of the coarse-grained variables as a function of K.
Refs: Meshulam, L., Gauthier, J.L., Brody, C.D., Tank, D.W. & Bialek, W. Coarse graining, fixed points, and scaling in a large population of neurons. Phys. Rev. Lett. 123, 178103 (2019).
Meshulam, L., Gauthier, J.L., Brody, C.D., Tank, D.W. & Bialek, W. Coarse-graining and hints of scaling in a population of 1000+ neurons. arXiv, 1812.11904 (2018).
- metropolis_spin_model.m
Simulates the spin model using the Metropolis algorithm and applies the PRG method to the model's activity.
- Connectivity_matrices.mat
Contains connectivity and distances matrices