FIGURE 1.

Training task and neuronal test. (A) Delayed Saccade task for object-reward association training. Subjects completed a Delayed Saccade task where they made saccades to a visual object after an overlap period for the center dot and peripheral object. Each object was associated with a consistent small or large reward delivered after successfully fixating on the object. (B) Passive Viewing task for neuronal test. Subjects fixated on a center dot as 1–6 objects were sequentially displayed in the neuron’s RF. At the end of a trial, subjects received a consistent reward not associated with any objects viewed during trial. This task was done on separate days from the Delayed Saccade task to avoid confounds from short-term learning. (C) Well-learned objects used with monkey R. Each row corresponds to one set of 8 objects. The first four objects (left of the dividing line) were consistently associated with large reward (‘good’ objects) while last four objects (right of the dividing line) with small reward (‘bad’ objects). Each set was trained >10 days before being used in the Passive Viewing task. Box with orange outline shows an example set used for tasks in this figure and Figure 2.