 Neurons in the entohynol cortex, a navigational brain region, are able to maintain their local computations while simultaneously reporting on global context changes. This ability is modelled by a recurrent neural network that performs both navigation and context inference. The network's activity patterns are shown to be qualitatively similar to those observed in the entohynol cortex, suggesting that it may serve as a general solution for remapping. This article was authored by Isabelle Icelot, Lisa M. Giocomo, and Alex H. Williams.