On the Perceptual Identity of Depth Vision in the Owl

Whether owls can be trained to discriminate computer generated random-dot patterns purely on the basis of their figure-ground organisation is an open question. Here I investigate this question by (A) providing an operant shaping technique that prepared four barn owls for behavioural measurements employing a remote computer display (Experiment I) and (B) using the random-dot technique together with tests of discrimination-transfer (Experiment II). Tested were two functions of vision that are thought to be involved in figure-ground segregation: (1) texture perception and (2) motion perception. The computer displays consisted of patterns of randomly positioned black and grey dots, that differed either in their position, textural organisation, or common fate (motion). The rapidness with which the owls learned to discriminate varied between the different individuals and depended heavily on the visual function tested. The here described experiments have been instructive in four ways. First, Experiment I has provided an adequate operant shaping method that prepared barn owls to be behaviourally tested with random-dot patterns that were displayed on a CRT screen. Second, Experiment II provided behavioural data that clearly demonstrated that owls can abstract figure from ground in computer displays of random-dot patterns. Third, signal detection analysis of the measured psychometric curves revealed a consistent response criterion for stimuli above or near threshold level, indicating a high degree of stimulus control. Fourth, the transfer tests provided evidence for generalisation discrimination in owls. 40 On the Perceptual Identity of Depth Vision in the owl