The visual system is a complex organization that helps us behave successfully in a world full of stimuli. The visual system is unable to measure the properties that define the physical world. What we see is just our subjective perception of the world. Consequently, we can say that we don’t see the world the way it really is, but as a mixture of our sensory input and our past experiences.
A challenge we face in our perception of images is that we cannot specify the objects and conditions in the real world that create our sensory inputs. That is, we cannot resolve the “inverse problem”. From an evolutionary point of view, our visual system has evolved to maximize “reproductive success” when interpreting the physical world. In this sense, evolution of the human species can explain why our visual system had to improve and manage correctly light, color, depth, etc. Humans and other animals needed better information to find food or detect predators that could pose a threat.
Due to the different ways in which our visual system solves the inverse problem, we all are aware of many different illusions and paradoxes with which we can trick the visual system: colors that are the same but look different or lines that are straight but look curved. And paradoxes can also happen when the visual system combines two input images (one from each eye) into a single one. A nice example of this kind of problems is the case when the images presented to each eye are completely different: one consists in vertical stripes and the other in horizontal stripes. Surprisingly, the way that the visual system deals with this situation is by alternating the images from each eye.
Similar geometrical problems appear when dealing with ambiguous images. An example of these images can be the Necker cube, which is simply a wire frame cube with no depth information. By focusing on different parts of the cube, one can force a more stable perception of the cube going in or going out. Other examples may be the face/vase or the duck/rabbit drawings. In these cases the observer has some degree of voluntary control over the switching process, and many researchers have investigated whether the switching process can be triggered by selective attention or by shifts in fixation position. The results of their studies were not clear, as they were very dependent on the image and other parameters.
One must note that perceptual rivalry affects other senses and happens, for example, when there are conflicting inputs into the two ears or two nostrils. Thus, at this point it is worth remembering that successful perceptions and behavior depends not on the actual properties of the world but on the perceptual values assigned by the frequency of occurrence of a stimuli and the reproductive success. How the visual system has evolved to deal with the inverse problem raises an interesting question: why does an organism need the capacity to rival between alternative perceptions when more than one possibility exists?
Some authors have proposed that searching for food or hunting could benefit from a fast disambiguation of conflicting visual, auditory and olfactory stimuli. Having some degree of free will can offer behavioral flexibility but, at the same time, an entirely voluntary strategy might also predispose an organism to taking too long to switch to an alternative threat or food source.
In these situations, the main aim of the brain is to combine all the given inputs to give the more plausible solution according to its experience. But when there are rivalry inputs, that is, when there is no unique answer to the inverse problem, the brain cannot give a good solution. In my opinion, and using the example presented above, a possible reason why the brain alters between the conflicting inputs and sometimes sees vertical stripes and sometimes horizontal stripes may be that it is presenting the two possibilities alternatively to see which fits better with the rest of sensory information received in a short period of time. Maybe a small change of perspective, a sound or some interaction can decide which solution is the correct one. In the end, what we try to resolve is the whole inverse problem, not in a specific time frame. The examples presented in this essay reflect situations that are not common in real life. Thus, my conclusion is that conflicting inputs in real life can be resolved after a short amount of time. What do you think?