Abstract
Several studies have investigated saccadic eye movements in illusions and most of these have examined simple geometric-optical illusions like the Mueller-Lyer or Judd illusion. In contrast, eye movements in ambiguous figures have rarely been studied (the Necker cube; Einhäuser et al., 2004). In the context of saccadic eye movements Yarbus (1967) showed that saccadic end positions in “natural scenes” can be influenced by prior instruction. Here, we investigated whether the prior instruction to perceive one or the other interpretation of an ambiguous figure affects saccadic end positions. One of four classic ambiguous figures (old woman/ young lady, face/ vase, saxophone player/ woman, and seal/ donkey) was presented randomly 6.5 degrees to the left or to the right of a central fixation point. Prior to the onset of the fixation point subjects (n=11) were instructed to look at one of two possible percepts. Subjects looked at different features of the image when asked to perceive one or the other interpretation (e.g. in the old woman subjects looked at the nose/mouth region and at the cheek and the eye of the young lady). We found a significant difference (p < .05) in landing positions of the first and second saccades depending on the instruction in all four ambiguous figures. We conclude that early saccadic eye movements (first and second saccade) towards ambiguous figures can be strongly influenced by top-down processes (here prior instruction).
Supported by DFG graduate Program ‘Brain and Behavior’ 885/1 (to KH), Bundesministerium fuer Bildung und Forschung: Project Modkog 620 00 177 (to BJW) and by the European Research Training Network: Perception for Recognition and Action - PRA 623-00-099 (to DdG)