Figure 2: Fragmentation of attractors. The figure shows the firing fields of five simultaneously recorded cells from the experiments described in [13] that were recorded as rats were exposed to four different contexts comprising pairwise combinations of a color (black or white) and an odor (lemon or vanilla). The most straightforward version of the attractor hypothesis would predict that when remapping occurs, all cells should remap together. However, it can be seen here that in the lemon box (top panels), when the color was changed from black to white, then cell 1 did not remap while the remaining cells did—cell 2 by shifting its field, cell 3 by switching on a field, and cells 4 and 5 by switching off their fields. When the box was vanilla, however, then cells 1 and 3 responded by remapping, cells 2 and 4 did not fire at all in either vanilla condition, and cell 5 switched off its field. Thus, each cell seemed to act independently and not as part of an attractor, bound to the others. Furthermore, each cell reacted to a combination of color and odor—the response to odor change was conditional upon what the color of the box was, and vice versa.