MOVE

Object recognition in the human medial temporal lobe: latencies from single neurons and local field potentials

Currently, there is no theory that relates the macroscopic structure of human cortex to dynamics in our environment. This is astonishing because it is generally accepted that our brains represent the environment to model incoming sensory input and optimise behaviour. Here, we propose a simple organizational principle for large-scale cortical structure. In this view, cortex can be regarded as a hierarchy, where each level represents trajectories in the environment at a specific time-scale. Low levels, i.e. levels that are close to sensory input, represent fast trajectories, whereas the top levels encode slow environmental dynamics. Note that this hierarchy is not about neuronal dynamics (e.g. nested oscillations in brain activity) but about representations of environmental dynamics that have an inherent hierarchical structure. We suggest that this cortical hierarchy exploits dependencies among environmental dynamics at different time-scales to implement brain function. We first present simulations that provide a proof-of-principle that perception can be understood in terms of a temporal hierarchy. We then discuss the evidence, from a wide range of neuroscientific observations, that cortical structure and function reflect an anatomic-temporal hierarchy.