Abstract
Spatial landmarks are critical reference points for navigation. An animal's position relative to spatial landmarks is represented by hippocampal landmark vector cells (LVCs), which exhibit spatial firing fields at an equivalent allocentric angle and distance to landmarks. It is unknown whether visual inputs are necessary for the formation and maintenance of LVCs. In rats exploring a novel platform with landmarks in complete darkness, LVCs were present in CA1. Their vector fields moved along with the landmarks and maintained their spatial selectivity when the landmarks were removed. In an update to McNaughton's model of vector encoding, we demonstrate how changes in entorhinal inputs can result in the phenotypes we observe in place cells and LVCs. This model suggests that these two spatial neurons may be unified by a common computational mechanism, with differences in spatial firing patterns resulting from differences in egocentric inputs.