Suscettibilità del cane alle illusioni ottiche

In the last decades dogs have gained popularity as a model organism for cognitive studies due their particular evolutionary history and comparative standpoint with humans. Despite the visual tasks being common in cognitive studies, our understanding of dogs’ visual perception is far from exhaustive. Particularly limited knowledge exists about dogs’ post-retinal processing. One way of investigating it is by assessing dogs’ susceptibility to visual illusions. Visual illusions allow us to explore the underlying perceptual mechanisms of several features, such as shape, size, numerosity, color and motion processing. However, up to date, the studies assessing visual illusion susceptibility in dogs are limited. There is particularly limited knowledge about dogs’ susceptibility to contour illusions and the field of numerosity illusions remains completely unexplored. To fill this gap of knowledge, the current PhD project was developed around three independent studies. In the first study we assessed if dogs are susceptible to the Kanizsa’s triangle illusion. In the Kanizsa’s triangle the three angles are visible, but part of the sides are missing. Despite of only some of the physical contour being present, human observers typically perceive the figure as a triangle. To assess if also dogs possess such ability of perceptual completion, we trained six dogs to discriminate a triangle against other geometrical figures. Once dogs reached the learning criterion, they were presented with an illusory triangle and a control stimulus. As a group, dogs chose significantly more often the illusory triangle, demonstrating susceptibility to the illusion. At an individual level, five dogs out of six proved to be susceptible to the illusion. Altogether, this suggests that the underlying mechanisms of contour perception are shared between dogs and humans. In the second study we assessed if dogs are susceptible to the Solitaire illusion. The Solitaire illusion is a robust numerosity illusion, where human observers commonly overestimate the quantity of centrally located items compared to items located in the perimeter. The study consisted of three experiments. In the first experiment we explored dogs’ abilities to discriminate quantities in the ratios of 0.67 and 0.75. The results indicated that dogs’ spontaneous abilities fall somewhere between those two ratios. In the second experiment we presented twenty dogs with the Solitaire illusion pattern using a spontaneous procedure. Dogs chose the illusion at chance level, indicating that they are not susceptible to the illusion. This conclusion was replicated in Experiment 3, where the dogs’ initial experience with the stimuli was manipulated before being presented with the illusion. Altogether, the results suggest that the spatial arrangement of elements, in particular clustering, does not influence dogs’ quantity perception in the same manner as it does in humans. In the third study we assessed if dogs are susceptible to the Connectedness illusion. Connectedness illusion is a numerosity illusion where the elements connected by task-irrelevant lines are commonly underestimated by human observers. Twenty-nine dogs underwent a training phase where they were conditioned to choose a stimulus containing the larger quantity of food. Fifteen dogs reached the learning criterion and were presented with the illusionary stimuli for four trials each. The results revealed that dogs chose the illusory array with connected elements significantly less often that the unconnected control stimulus, implying that they underestimated the connected stimulus similarly to humans. The result indicates that unlike clustering, connecting the elements affects the dogs´ quantitative abilities, and suggesting that different perceptual mechanisms underlie the two illusions. Moreover, this is the first study demonstrating that non-human animals are susceptible to the Connectedness illusion.