Spatial Representation in Postrhinal Cortex

Animals rely on a variety of internal and external cues to orient themselves when navigating their environments and determining their current spatial context. Information regarding these cues enters the brain from the navigator’s first-person perspective. Information of this type is considered to be egocentric, or self-centered. However, decades of behavioral, electrophysiological, and imaging research suggest that the brain contains a rich collection of spatial representations that are unrestricted by the animal’s first-person perspective, and instead are defined relative to the surrounding environment. These representations are considered allocentric, or world-centered. Despite an abundance of promising modeling work, the specific mechanisms by which first-person sensory information is transformed into an allocentric map-like representation within the brain are just beginning to be elucidated. One potential locus for the transformation from egocentric to allocentric is the rodent postrhinal cortex (POR), which has been implicated in spatial and contextual learning and is densely interconnected with brain regions that process egocentric sensory and allocentric spatial information. POR is also considered to be homologous to the human parahippocampal cortex, which has been strongly implicated in topographic spatial learning and visual scene processing. We therefore sought to determine exactly how egocentric and allocentric spatial variables are represented and combined in POR. We first recorded from POR neurons as rats navigated an open field environment, and found a large proportion of cells that responded to either the egocentric bearing of the environment centroid (center-bearing), the egocentric distance of the environment centroid (center-distance), the animal’s allocentric head direction (HD), or a combination of the three, confirming that POR neurons express a mixture of egocentric and allocentric correlates. Next, we used visual landmark manipulations to demonstrate that POR HD cells are sensitive to the number, distribution, and properties of visual landmarks, such that they fire bidirectionally under certain circumstances (i.e., with two different preferred firing directions), and may produce an estimate of HD based on the constellation of visual landmarks. Finally, we used chemical inactivation of the anterior thalamus to demonstrate that the POR HD signal is at least partially derived from the ‘classic’ vestibular-based HD signal, and is likely to reflect a combination of HD and visual inputs. These experiments provide insight into how egocentric and allocentric spatial information converge in the mammalian brain, and help to elucidate the role of the POR in processing spatial and contextual information