Assessing the effect of shading on dogs’ perception of depth, in a 2D picture

Perception of depth and tridimensionality is very important to navigate the environment and detect objects’ properties. Tridimensionality can be conveyed also by 2D images through pictorial depth cues, including shading. Following a previous study (Broseghini et al., 2024), that proved dogs can use a combination of linear perspective and shading to perceive tridimensionality in a 2D image, here we aimed at assessing whether the same perception could be elicited by shading alone. Fifty-two dogs were tested in two conditions: the “Real hole” condition, with a ball rolling on a plastic board, and falling in a square-shaped hole, and the “Depicted hole” condition, with the ball passing over a photograph of the same hole and disappearing in a box opposite to the starting point. Each dog was presented with both conditions 15 minutes apart and in a randomly assigned order, counterbalanced within the sample. In the “Depicted hole” condition, for half of the sample the photograph presented a low intensity shading, while for the other half a high intensity shading. In accordance with violation of expectation paradigm, we predicted that, if dogs can use shading to perceive tridimensionality, when the ball continued rolling onto the photograph, without falling, they should show surprise, indicated by an increase in attention towards the area of the unexpected event (i.e.: end of the apparatus). Data about the dogs’ orientation were collected in the 30 seconds after the disappearance of the ball, and attention to the end of apparatus compared between the “Real hole” and the “Depicted hole” conditions, with separate analysis for the low intensity and high intensity shading. For the low intensity shading condition, dogs' attention to the end of the apparatus was not affected by condition (“Depicted hole”: mean ± SE = 5.9 ± 0.8 s, “Real hole”: 6.5 ± 0.8 s, Wald Chi-square = 0.532, p = 0.466). For the high intensity shading, condition had a significant effect on dogs’ attention to the end of the apparatus, with dogs remaining oriented towards it for longer in the “Depicted hole” condition (“Depicted hole”: 8.7 ± 1.2 s, “Real hole”: 6.1 ± 0.8 s, Wald Chi-square = 8.296, p = 0.004), indicating that such shading elicited perception of the hole tridimensionality. Our results support the effect of shading as a key aspect on dogs’ ability to perceive depth, in line with many other species, but underlie that the effectiveness of the cue is affected by its intensity.