Event:

Predictive coding theories view the brain as a Bayesian interpreter that computes the difference between the external stimuli and a prior, internal model of the world. An important component of this predictive process is the hierarchical organization of the cortex that enables prediction and prediction-error signals to be processed in multiple feed-forward and feed-back loops. At the end of cortical hierarchy lays the hippocampal formation, an associative cortical area that integrates multimodal sensory-motor information and is crucial for expression of internally-generated representations such as cognitive mapping, memory and imagining. Mammalian navigation uses internal models to predict and represent the spatial-temporal statistical regularity of the sequence of environmental locations by sequential activation of ensembles of place cells in the hippocampus, which together form a cognitive map of the world. However, the predictive capacity of hippocampal networks during sleep/rest and their spontaneous organization into associative attractors that lead to stable and plastic networks active during future spatial navigation have remained unknown. Here, we develop a predictive coding model to reveal the internal structure of spontaneous functional neuronal connectivity during sleep in naive rats and use it to predict the future place cells and their sequence during subsequent encoding of novel spatial experiences.