Orientation analysing mechanisms in the detection of symmetrical dot patterns

This thesis investigated the possible role of orientation analysing mechanisms in the detection of symmetry in dot patterns. It was proposed that the mechanisms that detect axes of symmetry are similar to those that determine orientation in luminance contours, and that the mechanisms are linked.;Evidence of similarities in the mechanisms of symmetry detection and orientation perception was provided using an orientation illusion called the tilt aftereffect (TAE). It was demonstrated that axes of bilaterally symmetrical dot patterns produce TAEs that were similar to TAEs produced by luminance contours. Evidence of a link in the mechanisms was also provided. Significant interaction effects were produced in conditions where participants adapted either to an axis of bilaterally symmetrical patterns or a grating and then were tested with either an axis or a grating. The manner in which the mechanisms of orientation perception and symmetry detection may be linked was discussed in terms of a common neural site at which orientation information from different contours classes may converge. Despite similarities and a possible link in the underlying mechanisms, it was argued that the mechanism that detects axes of symmetry is different from the mechanism that detects the orientation in luminance contours and the empirical evidence supported this. The visual system does not process axes of symmetry and luminance contours as equivalent stimuli.;The thesis also examined the conditions and processes by which orientation components in symmetrical dot patterns (axes of symmetry and imaginary x axes) were determined. It examined how adaptation to these orientation components produced the reported TAEs, and how differences in salience of orientation information in these components determined the size of the TAE.;Finally, the processes involved in determining the orientation of axes of symmetry in a neural network were examined, and used to generate hypotheses about similar processing in human symmetry detection.