Extended Data Fig. 9. A small number of subiculum neurons encode egocentric bearing from corners.
(a) Schematic of the linear-non-linear Poisson (LN) model framework with behavioral variables including egocentric corner bearing (E), egocentric corner distance (D), allocentric head direction (H) and linear speed (S) (Model 2). (b) Proportion of subiculum neurons that were classified by Model 2 in large square (green) or convex-1 (orange) environments (n = 10 mice). Neurons combined from all model groups featuring egocentric corner-bearing (E, highlighted in red) account for 6.24 ± 1.20 % (n = 10 mice) of the total recorded subiculum neurons in the square environment. Similarly, 3.1 ± 0.6 % of neurons featuring egocentric corner bearing for convex corners in the convex-1 environment. For the box plots, the center indicates median, the box indicates 25th and 75th percentiles. The whiskers extend to the most extreme data points without outliers (+). (c) Schematics of LN model 3 and 4. Model 3 contains egocentric wall bearing (E), egocentric wall distance (D), allocentric head direction (H) and linear speed (S). Model 4 contains egocentric center bearing (E), egocentric center distance (D), allocentric head direction (H) and linear speed (S) (see Methods). (d) Left: Proportion of egocentric corner cells in the subiculum from the square (mean ± SEM; n = 10 mice), rectangle (n = 10), right triangle (n = 8), and convex-1 (bearing to convex corners; n = 10) environments. Right: pie chart showing the conjunctive coding of egocentric corner cells with other behavioral variables. (e) Representative egocentric corner cells from the square, rectangle, right triangle, and convex-1 (bearing to convex corners) environments. Each column represents a neuron. The first row shows spike raster plots color-coded with egocentric corner bearing on top of the animal’s running trajectory (grey lines). Similarly, the second row is color-coded with allocentric head direction. The third row shows positional rate maps.