Rule, M. E., Chaudhuri-Vayalambrone, P., Krstulovic, M., Bauza, M., Krupic, J., & O'Leary, T. (2023). Variational log-Gaussian point-process methods for grid cells. Hippocampus, 1–17. doi: https://doi.org/10.1002/hipo.23577.
The reference implementation contained in this repository has not been tested on all platforms; Please report any mistakes, bugs, or difficulties with installation that you encounter, including places were the documentation is inadequate.
The final snapshot of this repository approved during peer review is available at 
We present practical solutions to applying Gaussian-process methods to calculate spatial statistics for grid cells in large environments. Gaussian processes are a data efficient approach to inferring neural tuning as a function of time, space, and other variables. We discuss how to design appropriate kernels for grid cells, and show that a variational Bayesian approach to log-Gaussian Poisson models can be calculated quickly. This class of models has closed-form expressions for the evidence lower-bound, and can be estimated rapidly for certain parameterizations of the posterior covariance. We provide an implementation that operates in a low-rank spatial frequency subspace for further acceleration. We demonstrate these methods on a recording from grid cells in a large environment.
This repository contains example implementations of log-Gaussian Cox process regressions for analyzing grid cells (and perhaps other periodic, densely sampled 2D spatial datasets).
example data/: An example grid cell from Krupic et al. (2018).python/lgcp/: An example implementation as a python packagepython/notebooks/: IPython notebooks to reproduce the figures in the manuscriptmatlab/: A (less-complete) provisional Matlab implementationold_versions/: Previous iterations