Experimental design.

<p><b>A</b>. Sequence of task events in a trial. Each black rectangle represents the video screen. The white circle illustrates the central stimulus, the grey horizontal bar is the go cue. The green cross and the purple polygon are the two potential response goals. In this example trial, that could represent the first trial of a session, the response decision (highlighted by the braked rectangle) is toward the purple polygon. We used three potential goals in a session: a purple polygon, a green cross and a blue circle <b>B.</b> In this example sequence of trials, the previous goal was the purple polygon (left yellow box, as in the trial in A). The same goal was, by task design, presented again in the current trial (right green box), together with another potential future goal, which was either: 1) the <i>familiar goal</i> that was the goal discarded in the previous trial (green cross); or 2) the <i>unfamiliar</i> goal that was not presented in the previous trial (blue circle). <b>C.</b> Example sequence of trials with the human partner interacting with the monkey. Numbers indicate the trial position after the trial in A. Each panel represents the response choice. The correct goal (response) was always the goal that differed from the previous goal acquired either by the human agent in the <i>interactive</i> condition or by the monkey in the <i>noninteractive</i> condition. In the human trials the monkeys were required only to monitor the human partner choices. Notice that in this example sequence monkeys in trial number 5 could not choose the purple polygon based on what was their own previous choice (the purple polygon in trial 2), they had instead to choose a goal based on what the human partner chose in trial 4, that was the blue circle.</p>