Extended Data Fig. 10: Distinction between object-vector cells and border cells.
a, Experiment with object on suspended wall. Image to the left shows recording box with an object attached to a suspended wall with a 15-cm passage underneath the wall and the object. In this configuration, there is no impediment to the mouse’s movement near the object. Right, colour-coded rate maps for three example object-vector cells recorded on trials without any objects present (top row) and with an object (black square) attached to a suspended wall (white line). The three cells respond robustly to the suspended object. b, Colour-coded firing rate maps of four border cells recorded in the absence or presence of an object. The two cells on the left (one in a square and one in a circular box) showed no response to the object. The two cells on the right produced clear object-vector fields. c, Overlap between populations of object-vector cells and border cells (BCs). Fifteen out of 56 border cells also passed the criteria for object-vector cells. The two cells to the right in b are among those cells. d, Box plot showing, for the 15 overlapping border and object-vector cells, the distance from centres of object-vector fields to the nearest point of the object, and the distance from centres of border fields to the nearest wall. 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. Mean distance from object-vector field to object was significantly greater than the distance from border field to the wall (object-vector field to object: 21.4 ± 2.6 cm; border field to wall: 7.6 ± 0.6 cm; two-sided Mann–Whitney U-test, U = 381, P = 1.6 × 10−4), consistent with the interpretation that border cells and object-vector cells are functionally independent (Fig. 4).
