Figure 1. Representing place in the grid system.
(A) Grid cells (small triangles) in the medial entorhinal cortex (MEC) respond when the animal is in a triangular lattice of physical locations (red circles) (Fyhn et al., 2004; Hafting et al., 2005). The scale of periodicity (the ‘grid scale’, λi) and the size of the regions evoking a response above a noise threshold (the ‘grid field width’, li) vary modularly along the dorso-ventral axis of the MEC (Hafting et al., 2005). Grid cells within a module vary in the phase of their spatial response, but share the same period and grid orientation (in two dimensions) (Stensola et al., 2012). (B) A simplified binary grid scheme for encoding location along a linear track. At each scale (λi) there are two grid cells (red vs blue firing fields). The periodicity and grid field widths are halved at each successive scale. (C) The binary scheme in (B) is ambiguous if the grid field width at scale i exceeds the grid periodicity at scale i + 1. For example, if the grid fields marked in red respond at scales i and i + 1, the animal might be in either of the two marked locations. (D) The grid system is composed of discrete modules, each of which contains neurons with periodic tuning curves, and varying phase, in space. (E) For a simple winner-take-all decoder of the grids in panel D, decoded position will be ambiguous unless li ≤ λi + 1, analogously to panel C (see text). Variants of this limitation occur in other decoding schemes.