Fig. 1. (left) Place field of a cell with a field near the barrier end of the track; dark indicates high firing rate, grey low. (right) Each panel shows the trajectory of the animal near the end of the track, heavy dots indicate specific times at which the cell fired a spike.

Fig. 2. The temporal correlation matrix. (left) Expected matrix if there were a break in the representation at the turn-around point. Note the sharp pinchpoint in the middle. Simulation from Redish [14]. (right) Average correlation matrix, showing 2 s before and after each turn-around. Note the lack of a sharp pinchpoint. This indicates a continuity of the representation around the turn-around point.

Fig. 3. (top) Trajectory of an animal through part of the task. Time along the x-axis, position along the y-axis. The animal was running out and back. Spikes fired by the single cell in question are indicated by heavy dots. Note that the cell had a very directional field, but continued firing as the animal turned prematurely (such as on the seventh lap shown). (left) Change in directionality across short turn-arounds. Each dot indicates the average of all cells with a certain directionality for an animal. While inbound cells did not change, outbound cells lost their directionality when the animal turned around within the field. (right) Temporal correlation matrix across short turn-arounds. Although the outbound representations were spread across into the inbound spaces, there was still a sharp boundary where the inbound cells began firing.