Figure 1.

Experimental design. a) Schematic illustration of the two main experimental positions on the retina and the two sets of visual objects. Monkeys first learning to discriminate among the four objects in the “bias test” set (easier than the “restricted experience” objects, see Methods). Monkeys were then trained to discriminate among the four objects in the “restricted-experience” set, whose presentation was restricted to a single retinal position (Monkey M: +2 deg, shown here in red; Monkey P: −2°). The key experimental question was whether selectivity among these newly-learned objects was found at both the trained position (red) and the non-trained position (blue). b) Upper panel: a schematic time sequence of an example trial of discrimination training during the Experience Phase of the experiment (not drawn to scale). Yellow circle indicates gaze position during a typical trial. The animal initiated a trial by fixating a central point, after which a single object was presented. If the animal attempted to saccade toward the object, it was removed from the display, the animal was allowed to report the object's identity by directly saccading to one of the four lighted response targets at the corners of the monitor (yellow arrow), and correct reports were immediately rewarded. The mapping between object identity and the correct response corner was never changed (see Methods). Lower panel: a schematic time sequence of an example trial during the Probe Phase of the experiment in which neuronal responses were recorded (not drawn to scale). The monkey initiated a trial by fixating a central point, and was then only required to maintain that gaze position while 5-10 stimuli were presented at 200 ms intervals. On each trial, these stimuli were objects drawn randomly from both the “restricted experience” object set and the “bias test” object set (see Methods for details).