Place Cells, Grid Cells, Attractors, and Remapping
Figure 1
(a) Activity of a CA1 place cell, recorded as a rat foraged for rice grains in a 60 cm-square box for four min. The top plot shows the raw spikes (black squares) superimposed on the path of the rat as it (grey line), and the bottom plot shows a contour plot of firing rate, shown in 20% gradations (black = the peak 20% region). (b) Examples of remapping from two cells in CA1, as a box was changed from black to white. Cell 1 remapped by changing the location of its firing field, while Cell 2 remapped by switching its field off. (c) Evidence for discrete attractors in CA1, adapted from [9]. As an environment was incrementally “morphed” between more circle-like and more square-like (visible as the slightly lengthening corners as the diagrams go from left to right), the cells abruptly and coherently (all together) changed their activity. (d) Cross-sectional attractor landscape for a discrete attractor. The stable states of the system are the hollows into which the system (analogous to a rolling ball) settles and from which it resists perturbation. (e) A continuous attractor which does not resist perturbations but moves smoothly from one state to an adjacent one.
