A random 50% of trials had an uninterrupted 1.6 s WM delay (Figure 1A, left). In the other 50%, at 800 ms into the delay, the fixation point jumped to the opposite location on the screen (Figure 1A, right). The monkeys were trained to immediately saccade to it and reacquire fixation. Once fixation was acquired again, the WM delay was continued for another 800 ms, equating the full time of fixated delay period with the no-swap condition.
