Fig. 2: Stimulus display and task schematic.

a The first frame of stimuli in all three conditions were identical and contained ≈ 83 randomly oriented Gabors presented within a 12.5° radius of an invisible circle. The position of the Gabors was random, subject to an exclusion area (2.1°) around each Gabor to prevent overlaps. Motion information was delivered through either motion of the carrier (within a static envelope), the envelope (over a static carrier) or both features. b Participants adjusted the speed and direction of a 1/F noise probe to match that of the estimated global motion of the Gabor motion array. The location of the mouse pointer relative to the center was used to determine the direction and speed of probe motion in real time. The probe stimulus was moved in the direction towards pointer location at a speed that increased linearly with the distance of the pointer from the center. The speed of the probe ranged from 0°/s at the inner annulus circle to a maximum speed of 37.5 °/s at the outer boundary circle. c Trial sequence. After fixation for 1250, 1500, or 1750 ms (at random across trials), one of three motion signal conditions were presented for 150, 300 or 500 ms (randomly interleaved across trials), followed by the matching phase. Eye movements were recorded at 1 KHz throughout the experiment.