FIG. 5.

The effect of single-unit rank-order preservation on population recognition performance. A: example single units that maintained rank-order object preference (e.g., IT) or not (e.g., V1) across position. B: averaged population performance on position-invariant task. C: example units from the populations in B. All units in each population had similarly “shaped” response functions but positioned randomly to cover the 2D stimulus space, see methods. D: example single units that maintained rank-order object preference (e.g., AVG) or not [(e.g., (iv)_c] across clutter conditions. E: averaged population performance on clutter-invariant task, same as B.