Abstract
Introduction: We have previously provided evidence that angles are encoded as part of a global shape and not simply as a difference of local line orientations: discrimination thresholds for angles embedded in isosceles triangles are 2.5 times lower than for those in scalene triangles. Since the shape of the object affects performance, we aimed to ascertain if angle discrimination in the fronto-parallel plane shows the same pattern as “mental rotation” matching tasks, where performance is impaired when the shapes to be matched differ in orientation.
Methods: We measured angle discrimination in a 2AFC paradigm using triangles in four different configurations: 1) No transformation: both stimuli presented at the same location (3° eccentricity) and with the same orientation. 2) Rotation: same location but 90° difference in orientations. 3) Translation: same orientation but different (6° difference) locations. 4) Rotation and translation combined. The triangles (isosceles or scalene) were outlined or defined by a dot at the location of each corner.
Results: 1) Parafoveal angle discrimination thresholds are 30% higher than foveal. 2) Thresholds are 20% lower for isosceles triangles than scalene, compared with 2.5 times lower at zero eccentricity. 3) Thresholds are 70% lower for outlined triangles than for those defined by 3 dots. 4) Most importantly, for all four types of triangle (isosceles and scalene, outlined or dots), discrimination is unaffected whether angles are presented at the same or different location and with the same or different orientation.
Conclusions: The results show no cost to angle discrimination when shapes are rotated or translated. This is consistent with mechanisms that encode angles depending on the shape of the object, but independently of the orientation and position of the object. Our data provide no evidence for the use of “mental transformations” in this task.
GCU studentship grant to GK