Discriminating the integrated orientation of line segments and its statistical efficiencies

Abstract:  We carried out two experiments to investigate both how efficiencies fluctuate due to the experimental conditions and how we determine the representative efficiency among such various efficiencies in the discrimination task of line orientations. In Experiment 1, line segments were displayed simultaneously and dispersed spatially, while in Experiment 2, line segments were displayed sequentially. In both experiments, statistical efficiencies were calculated. In Experiment 1, efficiencies initially increased as stimulus duration increased, and decreased as the number of lines increased to a point where the efficiencies remained almost stable despite additional changes of stimulus duration or number of lines. In contrast, in Experiment 2, efficiencies were stable for almost all numbers of sequentially displayed lines, except for the condition that four lines were displayed in random positions. These stable efficiencies were approximately 50% in Experiment 1 and 30–50% in Experiment 2. We suggest that these stable efficiencies are varied in the context that human and ideal observers would adopt the same strategy regardless of the experimental conditions.