Abstract
Introduction: Previous studies have investigated whether angle discrimination can be predicted by the sensitivity of orientation discrimination mechanisms, or whether angles are encoded by specialised detectors. Conflicting evidence has been presented to support both alternatives. In all these studies, it has been assumed that angle discrimination is independent of the shape that defines the angle. We aimed to test the validity of this assumption. Methods: We measured angle discrimination performance for three different conditions. In the first condition, the stimuli were isosceles triangles (sides forming the angle were of identical length), in the second they were scalene triangles (sides were of different lengths) and, in the third condition, each side defining the angle was varied randomly between and within trials. For all conditions, triangles were either outlined or defined by a dot at the location of each corner. Within trials, we varied randomly the stimulus size (up to ±70%) and the overall orientation of the triangle (up to ±10 degrees), in order to eliminate any cue to the task other than the angle itself. Subjects were asked to judge which of two successively presented angles was more obtuse. Results: It was found that angle discrimination thresholds, for both outlined and dot triangles, were lowest when the stimuli were isosceles triangles (outlined: 1.94 degrees; dots: 3.06 degrees). Performance was dramatically poorer for scalene triangles (outlined: 5.99 degrees; dots: 7.11 degrees), and for the random condition (outlined: 7.48 degrees; dots: 8.81 degrees). Conclusions: These results show that performance in angle discrimination tasks is dependant on the stimulus shape. This has strong implications for the mechanisms involved in angle discrimination and casts doubt on all proposals made so far.
Glasgow Caledonian University Studentship grant to GK and GL