Abstract
The purpose of our experiments was to investigate whether visual perception could be influenced by extended presence ‘inside’ a 3-D virtual reality of computer-generated autostereograms (AST). In addition to usual ASTs (with stationary hidden figures, words, and 3-D surfaces) we used novel, dynamic autostereograms (DASTs), generated in a 3-D studio as a short film, where consecutive frames were ASTs with small modifications of the disparity values. The percept resembled textured ‘mushrooms’ rising towards the observer from the main frontal plane. When the disparity rose to some limiting value, the growth stopped and was inverted: a convexity was gradually transformed into a concavity and vice versa. The maximum range of local disparity change was ±0.6 deg. Our ASTs and DASTs were arranged in a sequence of increasing complexity. During a ‘training’ session (3 – 15 min), the subject had to actively examine the sequence and recognise all hidden forms.
Performance in threading a needle with one (the right) hand was recorded before and after ‘training’. The needle was set at a 50 cm distance from the subject, turned so that the eye of the needle was not visible. The total time of five successful trials was measured for each of five right-handed subjects with normal or corrected vision. In two sets of experiments, this visuomotor task was carried out under binocular and monocular viewing, respectively.
In both conditions, performance improved significantly (Fisher's criterion) after the ‘training’ session. The total time decreased from 50 – 55 s to 35 – 40 s in the binocular condition and from 70 – 75 s to 47 – 55 s in the monocular condition. The results are discussed in terms of ‘tuning’ of vergence eye movements during autostereogram examination.
Get full access to this article
View all access options for this article.