Extended Data Fig. 7: Orientation of object-vector fields is independent of the geometry of the environment but breaks down in the absence of visual input.
a–c, Experiment in which eight cells from three mice were recorded successively in a circular and a square recording box. The boxes were placed at the same location in the recording room and cues external to the box were identical. a, Left, colour-coded rate maps of three cells recorded first in the circle, then in the square. Peak rates are indicated below the rate maps, mouse and cell ID numbers to the left. Colour bar indicates normalized firing rate. Right, directional tuning curves for the same three cells, with firing rates shown as a function of allocentric direction relative to the object. b, Correlation between directional tuning curves in square and circular environments for all eight cells. Black line between box edges indicates median, box edges indicate 25th and 75th percentiles, whiskers extend to the most extreme point that lies within 1.5 × IQR. c, Difference in peak direction of tuning curves between square and circle for all eight cells. d–f, Object-vector cells recorded in light and in complete darkness. d, Colour-coded firing rate maps from five example cells recorded successively in light and darkness (top and bottom, respectively). Mouse and cell ID numbers are shown at the top of each column, peak rates below each rate map. Colour scale is similar for each column of rate maps. e, Distribution of vector-map correlations across pairs of trials recorded either successively in light (yellow) or first in light and then in darkness (grey; n = 21 cells). Stippled line marks the 99th percentile correlation threshold. f, Distributions of spatial information content, spatial coherence, peak firing rate and mean firing rate in light and in darkness. All four measures decreased from light to darkness (spatial information content: Wilcoxon signed-rank test, W = 227, n = 21, P = 1.1 × 10−4; spatial coherence: W = 231, n = 21, P = 6.0 × 10−5; peak firing rate: W = 230, n = 21, P = 6.9 × 10−5; mean firing rate: W = 212, n1 = n2 = 95, P = 8.0 × 10−4). All statistical tests were two-sided. g, Distribution of head-direction scores of all object-vector cells recorded in light and in darkness. Head-direction tuning increased significantly in the absence of visual cues (two-sided Wilcoxon signed rank test: W = 36, n = 21, P = 0.005).
