Fig 1. Experimental paradigm.

(A) A mirror stereoscope was used to present distinct images to each eye. The mirrors in front of the subject’s eyes were transparent to infrared light, allowing for eye tracking. Subjects were instructed to press a key according to their dominant perception: on the left (or right) keypad when the green (or red) gratings appeared to go leftwards (or rightwards). (B) Each block contained 20 repetitions of a 2-s trial with a 1-s blank interval between trials. Non-rivalrous leftwards, rightwards or rivalrous trials were randomly intermixed, composing 25%, 25% and 50% of trials in each block, depicted as red, green and yellow in panel B, C, and D, respectively. (C) An exemplar time course of horizontal eye position in one PD patient for 1 block. Center of the stimulus corresponds to the eye position of 0 deg. In non-rivalrous trials (red and green) there is clear OKN with slow phase moving in the same direction as the stimulus, and fast phase (saccades) moving the eyes back. In this block, horizontal eye position was positively biased towards right. Slight bias is expected as we did not provide a fixation point, which reduces OKN. (D) The button press time course from the same block as above. (E) Magnified button press time course shown in (D). Here, the subject first briefly pressed a left button within the first 0.5 s and then switched to a right button, which they held down for about 0.2 s. Just after 1 s, they briefly let go of the right key before pressing a right key again, until they changed to a left key just before the end of the 2 s trial. In this case, the most consistently endorsed direction (and thus the labelled direction of this trial) is “right”, the first button press time is 0.5 s and the and first consistent button press time is 0.6 s.