Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster A in Poster Session A: Tuesday, August 4, 9:30 – 11:15 am, Kimmel Center, Shorin & Rosenthal Rooms

A new code for position: flat neural maps in mouse medial frontal cortex

Rajyashree Sen¹, Alexander Medoff¹, Lorenzo Posani¹, Larry Abbott¹, Richard Axel¹, Stefano Fusi¹; ¹Columbia University

Presenter: Rajyashree Sen

Spatial navigation requires both direct representations of location and abstract representations of relational structure that enable extrapolation from explored to unexplored regions of space. Hippocampal place cells in mice encode the animal's position, but a localized code does not support extrapolation. Here we identify a novel neural code for position in the mouse medial frontal cortex that has the geometrical properties of flat map in neural space, and, hence, supports extrapolation. We performed calcium imaging of excitatory neurons in medial frontal cortex, hippocampal dorsal CA1, and lateral orbitofrontal cortex as mice explored a square arena with salient corner landmarks. Position could be decoded from both medial frontal cortex and CA1 populations. In medial frontal cortex, linear decoders trained to predict horizontal or vertical nose position in one half of the arena generalized to the other half (extrapolation), indicating that the two coordinates of the animal's position are represented in a disentangled manner. In contrast, CA1 representations showed markedly reduced extrapolation. At the single-cell level, medial frontal neurons exhibited elongated and broad spatial fields. Our results provide evidence for a low-dimensional, map-like representation in the mouse frontal cortex that supports extrapolation, preserves distances, and allows for planning.

Topic Area: Methods, Tools, Theory & Neural Coding