Similarly the ‘Necker Cube’ illusion (figure 3) can also be explained through top down processing. In this illusion the cube flips between two different interpretations of the picture whereby one of the faces of the cube can appear to be at the front of the cube but also can appear to be at the back of the cube. Work carried out by Wheatstone illustrates how this illusion can be explained through top down processing. He conducted experiments with geometrical solids and three-dimensional models and made the following observations. He found that by following the lines of the cube with the intention to see the cube one way or another causes the cube to shift to be perceived in the chosen way and so the effect of reversal was a matter of will. He also stated that when looking at the cube it could be perceived in two ways, however the figure encountered more frequently in previous experience would be observed more readily than the other perception. If the figure is unfamiliar then the illusion will flip more readily between the two perceptions. He concluded by saying that the sensation of flipping between the two different perceptions of the cube was nothing more than psychological and was based on sensations stored in memory and compared on previous experience. Wheatstone’s explanation for the Necker cube illusion clearly illustrates how top down processes can be used to explain visual illusions as it shows how the perception of the illusion is not simply based on physiological factors but also how psychological factors (such as previous experience of the figures) play an important role.
However bottom up processing is also used to perceive visual illusions. Examples of this include the Herman Grid (figure 4) and Mach Bands (figure 6). Firstly the Herman Grid which is simply a grid consisting of twenty-five black squares intersected by white lines. However when looking at this grid black smudges seem to appear in the intersection of the white lines. This however doesn’t occur at the intersection that is being fixated upon. This illusion can be explained through bottom up processes whereby judgements are made solely on the information that enters the eye and knowledge of the world has no effect. Figure illustrates how the Herman Grid can be explained. It shows how when a receptive field that is over-excitatory is aligned with an intersection there will be a larger amount of inhibition than when the receptive field is centred in the space between two squares. This results in the ganglion cells registering there being less light in the intersections compared to the other areas between the squares. This shows how the illusion is physiological and so does not rely on any previous experience or knowledge thus showing how the illusion can be explained through bottom up processing. Similarly the fact that the grey smudge appears to disappear at the intersection that is being fixated upon can also be explained through bottom up processing. This is because the receptive field is smallest in the foveal region of the retina and then increase in size as the distance from the fovea increases. Therefore the receptive field in the periphery will be larger than at the fixation point and so resulting in there being grey smudges in the peripheral area but not at the intersection that has been fixated on directly. Again this requires no prior knowledge and shows how bottom up processes are used in perception.
Mach Bands (figure 6) also demonstrate how bottom up processes are used in perception. Here where lighter and darker tone merges there appears to be a darker bar in the darker area (its position is indicated by α in figure 6) and a lighter bar in the lighter area (its position is indicated by β in figure 6). This is shown in figure 7 as despite the fact that the measured luminance simply changes from dark to light, the perceived brightness shows how just before the colour merge there is a decrease in brightness and after the colours merge there is an increase in brightness. This illusion has a physiological explanation similar that of the Herman Grid, explained by figure 8. This diagram shows four receptive fields in the Mach Band illusion. A is positioned completely over a dark band and so perceives the area as being dark. Similarly B is completely positioned over a light area and so perceives the area as being light. D however is positioned between the dark and lighter band, with the light photoreceptors in the darker band and the dark photoreceptors partially in the lighter band. This results in lateral inhibition and so a darker band is perceived. Likewise this occurs at the site of receptive field C. Here the light photoreceptors are located in the light band and the dark photoreceptors are partially located in the darker band, resulting in a lighter bar perceived in the lighter band. This explanation of the illusion contrasts with the explanations for the Necker Cube and Muller Lyer illusion, as it requires no previous experience or knowledge and is simply the product of the location of different receptive fields. Therefore this again shows how bottom up processes can be used to explain visual illusions.
Despite the large amount of evidence to explain the above illusions as being perceived through either top down or bottom up processing there is also evidence to suggest that it is a combination of both processes which lead to the overall perception. An example of this can be demonstrated using the Necker Cube. It was strongly believed by Necker himself and was later supported by the work of Brewster that the illusion was caused simply through physiological optics and was not merely a mental operation. Necker stated that since a sensation was felt in the eye when the illusion flipped the illusion was caused by the eye rather than drawing on top down processes. Both the argument by and Necker and Brewster and that of Wheatstone have evidence to support them however it is rather subjective as to which is correct, thus showing how there is no clear explanation as to how the illusions are perceived.
This suggestion of a combination of both top down and bottom up processing has been supported by the work of Bruner, Postman and Rodrigues. They conducted a study whereby participants were briefly shown playing cards. Instead of presenting conventional playing cards they showed one card consisting of black hearts. This resulted in some participants stating that they saw purple hearts and others brown hearts. This combination of the expected red (caused by the previous knowledge of playing cards) and the observed black (caused by the actual colour seen by the participant) shows clearly how both top down and bottom up processes can both be involved in the perception of an illusion.
Therefore despite the more obvious explanations stated in the first four examples there are also alternative explanations which can involve the use of both bottom up and top down processes combined to perceive an illusion. And so as it is hard to provide an adequate and clear-cut explanation for an illusion and these explanations remain different for each illusion, the perception of illusion becomes subjective.
References
S Coren, L M Ward & J T Enns 2004 Sensation and Perception 6th edn
W N Dember & J S Warm 1979 Psychology of Perception 2nd edn
Purves & Andrews 1997 The Perception of Transparent Three Dimensional Objects Proc. Natl. Acad. Sci. USA Vol 94
Mach Bands 1965 Quantitative Studies on Neural Networks in the Retina Holden Day
M W Eysenck 2004 Principles of Cognitive Psychology 2nd edn Psychology Press
Bruner, J. S., Postman, L., & Rodrigues, J. (1951). Expectations and the Perception of Colour. American Journal of Psychology, Vol 64