Abstract
Background: Depth-inversion illusions (DII) involve stimuli for which physically distant points are perceived to be closer to observers than physically near points. Schizophrenia (SZ) patients are less likely to perceive DII as strongly as controls. In particular, there is a negative correlation between the tendency to obtain DII and positive SZ symptoms (Keane et al. 2013). Objectives: First, to test the hypothesis that the tendency to obtain DII correlates negatively with delusional ideation in healthy controls, since delusion is an important positive SZ symptom. Second, to ascertain the frequency with which individuals in the general population perceive DII. Methods: We developed a test in psiTurk to obtain data from hundreds of participants. We assess DII tendency by testing performance with DII stimuli, and we measure delusional ideation adapting Peters Delusions Inventory (PDI; Peters et al. 1999) into an online version. We used two classes of 3D objects: perceptually stable unambiguous objects that serve as “catches”: banana, apple, toy, etc.; and bistable objects that can exhibit DII but can also be perceived without depth inversion: human hollow mask, monkey hollow mask and a reverse-perspective scene. To obtain depth from motion, each object was rotated clockwise or counterclockwise around a vertical axis. We used 3-D probes embedded in strategic locations of the objects to infer whether participants obtained the veridical or illusory depth percept. Results: We conducted a pilot study with 9 participants that validated our approach of using probes to assess the perceptual state of participants. Results from large-scale psiTurk sessions are forthcoming. Discussion: Using the depth-from-motion approach is justified in light of results from a study that obtained evidence for its validity in DII experiments with SZ patients and controls (Keane et al. 2013). Our crowd sourcing experiments can study useful unexplored correlations that require large numbers of participants.