The effect of luminance variation on the apparent position of an edge.

A gray outline against a white (or black) ground appears to deviate when one of the divided regions turns into black (white). The direction of shift is not predictable on the basis of luminance profile and polarity contrast of this part of contour, called gray edge (to indicate a stepwise gradient from black to gray and from gray to white). Rather, it appears to depend on the luminance profiles of the collinear regions: A gray edge shifts toward the dark side whenever collinear with a gray line traversing a white ground. The same gray edge takes the opposite direction whenever it extends against a black ground. This rule proved to be successful in predicting the illusory convergence of the sides of a square that formed the stimuli of the first experiment, but the magnitude of the phenomenon was affected by luminance ratios and polarity contrasts of the gray edges, in agreement with the findings of the experiments on gray or blurred edge misalignment. A second experiment tested some hypotheses predicting the combined effects of two or more distorting sources. These hypotheses, suggested by the physical theory of vector sum, were partially disproved. A new model is proposed that assumes different ways of integrating local distortions. The third experiment tested predictions of how distorting pulses in opposite directions combine. The illusory misplacement of edge studied in this experiment is proposed as the underlying phenomena of the café wall illusion, the hollow square illusion, and other illusory phenomena observed with blurred areas. A connection with the induction grid phenomena is hypothesized.

The global figural characteristics in the Zöllner illusion.

The Zöllner illusion has been accounted for in terms of local interactions between the vertical lines and the crossing segments. Recently, however, some evidence supporting the importance of global figural characteristics--ie of figural elements that are not directly interacting with the test lines--in the occurrence of orientation illusions has been reported. Three experiments have been conducted with parts of the Zöllner figure to test whether this illusion is affected by the global figural characteristics. The results indicate that, similarly to what has been observed for other orientation illusions, the Zöllner illusion depends on both local and global characteristics of the stimulus configuration. In addition, results suggest a similar weight for both these figural characteristics in determining the occurrence of the illusory effect. Finally, relations among different orientation illusions are also discussed.

Rigid and non-rigid kinetic depth effect with rotating discrete helices.

To examine the conditions in which human observers fail to recover the rigid structure of a three-dimensional object in motion we used simulations of discrete helices with various pitches undergoing either pure rotation in depth (rigid stimuli) or rotation plus stretching (non-rigid stimuli). Subjects had either to rate stimuli on a rigidity scale (Experiments 1 and 2) or to judge the amount of rotation of the helices (Experiments 3 and 4). We found that perceived rigidity depended on the pitch of the helix rather than on objective non-rigidity. Furthermore, we found that helices with a large pitch/radius ratio were perceived as highly non-rigid and that their rotation was underestimated. Experiment 5 showed that the detection of a pair of dots rigidly related (located on the helix) against a background of randomly moving dots is easier at small phases in which the change of orientation across frames is also small. We suggest that this is because at small phases the grouping of dots in virtual lines does occur and that this may be an important factor in the perceived nonrigidity of the helices.

Perceived rigidity and nonrigidity in the kinetic depth effect.

Stroboscopic simulations of three-dimensional rotating rigid structures can be perceived as highly nonrigid. To investigate this nonrigidity effect a sequence of either three (experiment 2 and 3) or thirty six frames (experiment 4) was used, each consisting of a set of dots with location on the horizontal axis corresponding to the parallel projection of a nominally defined helix. Observers were asked to judge the angle of rotation of eighty helices defined by the factorial combination of eight phase (phi) values (ie difference between the sinusoidal path of one dot and its neighbours) and ten different angular displacement values (alpha). When in each static frame the dots can be organized into curved dotted line (small values of phi), the perceived 3-D helices are highly nonrigid. But when shape information is not available in each static frame (high values of phi), the helices are perceived as rigid and rotation judgement is possible providing that alpha < 15 degrees. It appears that at small values of phi observers fail to recover the rigid structure of the helices since the input to the structure from the motion process may be distorted.

Sex-related differences in hemispheric asymmetries in processing simple geomerical figures.

16 (8 men, 8 women) subjects were required to respond same or different to pairs of geometrical figures. Same responses were given to physical identity pairs, to analog identity pairs, and to name identity pairs. Male subjects showed a left visual-field advantage regardless of the level of processing, whereas female subjects did not show a clear-cut hemispheric asymmetry. These results were discussed in terms of different processing strategies employed by the two sexes.

Laterality effects, levels of processing, and stimulus properties.

Two hypotheses of hemispheric specialization are discussed. The first stresses the importance of the kind of processing to which the stimulus is subjected, and the second stresses the importance of the nature of the stimulus. To test these hypotheses, four experiments were carried out. In Experiment 1 verbal material was employed in a same-different classification task, and an overall right visual field superiority was found. Experiment 2, in which verbal stimuli were subjected to visuospatial transformations (i.e. mental rotations), yielded no laterality effect. In Experiment 3 geometrical figures were employed in a classification task similar to that of Experiment 1, and an overall left visual field superiority was found. In Experiment 4 both verbal and geometric stimuli were employed. The results showed a significant interaction between field of presentation and nature of the stimulus and no interaction between field of presentation and level of processing.

Different levels in processing simple geometrical figures.

An experiment has been conducted in order to extend to geometrical figures Posner's (1969) findings of various levels of processing for alphabetical material. Pairs of simple figures (triangles and trapezoids) were tachistoscopically presented to subjects asked to judge if they were "same" or "different". Subjects were instructed to respond "same" either if the two figures had the same shape or the same name, e.g., a scalene triangle paired with a right-angled one. The results support the theory of different levels of processing also for geometrical figures.