The local motion direction of the dots in the translating RDPs either matched the Pr or the AP direction, but it was always identical in both patterns. The local dots’ speed was the same in all RDPs. Throughout a trial, the translating RDPs followed parallel trajectories at a constant velocity of 3.5°/second, circumventing the RF pattern (Figure 1A). When the initial position of the translating RDPs was between the fixation spot and the RF pattern, they translated toward the periphery (“outward”). When their initial position was eccentric to the RF pattern, they translated toward the fixation spot (“inward”). The RDPs never overlapped. The color

of both translating RDPs was always the same (red or green) but different from the RF pattern’s color (green or red). The two color combinations were randomly intermixed Y-27632 order across trials

to avoid that the animals associated a color with a given stimulus type. During trials, the animals maintained gaze on a fixation spot at the screen center and pressed a button. After 590 ms, the RF and translating patterns selleck kinase inhibitor appeared on the screen (Figure 1A). Three different task conditions were tested. When the fixation spot color matched either that of the RF pattern (attend-RF), or of the translating RDPs (tracking), the animals had to detect a brief (110 ms) change in the corresponding pattern(s) local dots’ speed ( Figure 1C). The change intensity was chosen in such a way that the proportion of correct detections was found 75% or higher. During tracking, speed changes occurred with

equal probability in either one of the translating RDPs. All changes occurred at a random time between 820 and 5,060 ms from trial onset, challenging the animals to sustain attention on the target(s). Releasing the button within 150–600 ms from target change onset was rewarded with juice. We also tested the animals during a third condition in which they attended to the fixation spot and detected a change in its luminance (attend-fixation). The timing of these changes was similar to the one in the other two conditions. The probability that the animal obtained a hit by randomly releasing the lever between trial start and end was “450 ms / 4,020 ms = 0.106” (chance hit rate = 10.6%). During a recording session different trial types were randomly interleaved. Approximately 30% of the trials contained a speed change in the noncued/distracter RDP(s) (e.g., in the RF pattern during tracking, or in one of the translating RDPs during attend-RF), preceding the target change. If the animal released the button in response to this speed change in a distracter, the trial was aborted without reward. This motivated the animals to attend to the target(s) and to ignore the distracter(s). Hit rate in these trials was above 94% in the attend-RF condition and above 90% during tracking. During attend-fixation the hit rate was close to 99%, significantly above chance.