(A) In the Ambiguous condition, bistable images were presented for 60 s each, and subjects pressed buttons to indicate their current percept. (B) Percentage of time spent in each perceptual state during the Ambiguous condition. (C) In the Unambiguous condition, bistable images which were modified to reduce their ambiguity were presented. (D) Percentage of valid trials for each image type during the Unambiguous condition. Valid trials consisted of the subject pressing the button only once for the intended percept. (E) In the Discontinuous condition, ambiguous images were shown nine times with interleaving blank periods. (F) Percentage of blank periods that were classified as Stable, Unstable, or Missing. ‘Stable’ indicates that perception was the same before and after the blank period, ‘Unstable’ that it was different, and ‘missing’ indicates that no button press was recorded during the pre or post image presentation period. (B, D, F) Dots indicate individual subjects; bars and error bars indicate group mean and standard error of the mean (SEM).