Visual illusions have been a source of fascination since the ancient Greeks period when visual compensatory methods were employed in architectural design to combat illusory effects. The Parthenon temple of Athens illustrates this. In an effort to stop the Parthenon temple from giving the appearance of sagging in the middle, the temples designers made use of linear perspective in the construction of the temples columns and also raised the middle of the structures floor. Linear perspective is when two parallels lines give the appearance of converging. When the temple is viewed from any angle, an observer sees a normal temple structure but in reality the temples true form is slightly curved like a dome (Gombrich, 2000). This example demonstrates how easily our visual perceptions can slip-up. Morgan (1996) defined visual illusions as “errors of perception that occurs when our visual system make mistakes”. Johannes Mueller who is often named as the pioneer of modern physiology invented the term visual illusion. He defined it as a distortion that is facilitated by malfunctioning in the visual system (Rock, 1995).

The visual system is a sensory system that enables us to make a representation of our environment by interpreting sensory input from electromagnetic energy like light. Light wavelength reverberates from an objects image and enters the cornea of the eyes before going through the pupil which is monitored by the iris. It travels on to the lens from which an image is projected to the retina at the back of the eyes. This is called retinal-image. When an object is at a far from us, a smaller image is focused on the retina, when it is close, a bigger image is projected on the retinas’ photoreceptor cells which consist of nerve cells called rods and cons that transduce light energy into neurochemicals signals. Neurochemicals signals goes through an intricate processing and encoding in the photosensitive ganglion cell and through the optic nerve to the lateral geniculate nuclei (LGN) in the thalamus. Neural impulses is then synapse to the neurons in the visual cortex and to other regions in the occipital lobe, to the inferior temporal cortex which functions include perception of form and recognition of objects (Goldstein, 2007). From inferior temporal cortex information is also sent to other parts of the brain. This process is known as visual sensory processing or sensation. It allows us to sense visual stimuli. Sensation is the stimulation pattern of incoming energy into our visual system and it usually uses a bottom-up processing. It can be distinguished from perception which is the interpretation of sensory information and involves a longer period of organization. Perception often uses a top-down cognitive processing to interpret incoming information.

Gregory (1997) outlines that illusions can be grouped into three categories. First, illusions that are determined by natural laws of physics without the involvement of perception. i.e. rainbow, lighting, fire, mirage, and stars at night. Second, illusions that uses bottom up processing in the visual perceptual system. Bottom up processing uses characteristics of what we see or what we perceive to form a perception i.e. the Hermann grid illusion which shows dark patches at the intersection. This is directly because of the neural processes known as lateral inhibition which this paper will explore later. The third category is influenced by top down cognitive processing which relies on previous experience and expectation to make sense of the information coming into our visual system i.e. the Muller-Lyer illusion, the Ames room and the figure-ground illusion.

The Gestalt school of thought contributed much to psychologists understanding of how we view the world: how we make sense of it using depth and colour and how we recognise shapes and objects (Goldstein, 2007). A key question that they addressed was how we organize what we see. This is important because we have to be able to segregate an object from everything else around it to make sense of our surrounding. The Gestalt school believed that this ability is almost completely dependent on the observers experience and not innate i.e. when looking at the reversible figure in Figure 1; those who see the black coloured area as representing the background see a vase whilst those viewing the white area as background perceive two faces. This shows that when viewing a picture or a scene, the first thing that we do is focus on one object from the scene or picture before segregating everything else into background (Zimbardo et al, 1995). The gestalt school argues that this perceptual organisational processing is significant because it can affect how a visual stimulus is perceived (Goldstein, 2007).

Gregory & Wallace (1963) tested a man named S. B. after he had undergone correctional surgery to reverse early born blindness. Several visual illusions like Muller-Lyer, Ponzo illusion and the Poggendorf illusions were administered to measure his visual perceptual capacities. One particular black and white figure-ground illusion known as the face and fan illusion was shown to S. B. The result showed that he could only make out the figure image of fan from which he remembers from the sense of touch. He could not see the background face even when it was pointed out to him several times. The authors concluded that S.B was unable to distinguish between figure-ground because he was still new to vision and did not have enough visual cognitive learning and experience that would allow him to do so. This shows that learning and experience plays a crucial role in perception.

Fig. 1: The Rubin Vase (1915)

Furthermore, the Gestalt school of thought outlines that the visual system uses a set of six principles known collectively as the laws of pragnanz or the law of groupings to recognise and segregate things (Gombrich, 2000). These laws are, first, the law of proximity which asserts that seeing spatial proximity in things tends to make us group them together and see them as a whole. Second, the law of similarity asserts that the visual system has a tendency to group things with similar characteristics like size, shape, texture and colour together. The law of good continuation states that the brain has a preference for seeing things in an unbroken stream; there it will sometimes continue a visual pattern irrespective of reality. The forth law, the law of closure: asserts that we tend to fill in gaps in things because we like to see a whole form. The fifth law, the law of common fate asserts we things moving in same direction have a tendency to be seen as belonging together. The sixth law, the law of symmetry, we tend to see symmetrical images or lines as a whole even when they are far from each other. These principles suggest that we like to group things together because our visual system prefers to simplify complex things but at the same time this system of grouping allows for errors to occur at different levels of processing (Coren & Girgus, 1978).

The Muller-Lyer illusion is one of the most well known illusions. It provides evidence that past knowledge and experiences can influence our perception. The Muller-Lyer illusion as shown in Figure 3 shows two vertical lines with fins. The left vertical line with fins pointing inwards and the right vertical lines with fins pointing outwards. The right line appears to be slightly longer than the left even though they are of the same length. This effect persists even when we are aware that the lines are identical. It is believed that we perceive the lines inaccurately because of inaccurate informational input in the visual system. Gregory (1966) proposed that the inaccuracy in the visual system occurs because of displaced size constancy. “Size constancy is the ability to correctly judge an object’s size despite changes in its distance and hence changes in its retinal image size” (Ungerleider, Ganz & Pribram, 1976).

Furthermore, Gregory (1966; 1968) expounds that we have a tendency of using our three-dimensional world as a quantifying basis when interpreting simple two-dimensional images. Since we measure the three-dimensional world solely on the basis of prior knowledge and experiences, this same principles is automatically applied in our visual system when viewing some things, creating perceptual errors in the visual system. The outlines of the Muller-Lyer lines are similar to the interior and exterior corners of a house as shown in Figure. 3. In our three-dimensional world we perceive the interior corner of a house to be at some distance away from us therefore the image on the retina is smaller (Gregory, 1966). When viewing the Muller-Lyer, our visual system activates size scaling because the two outward fins are seen as being more distant. These results in a perceptual error; seeing the two lines as different. However, this idea is contrasted by Ginsburg (1984) who makes the case that the Muller-Lyer effect is determined by neural blurring. Others have argued that size constancy does not fully explain the Muller-Lyer because when the lines are replaced with dotted lines the illusory effects is reduced from in between thirty to seventy percent (Coren, 1970; Wenderoth & wade, 1981).

Segall, Campbell & Herskovitz (1963; 1966) support the idea of knowledge and experience influencing perception. They examined the role that the environment plays in perception. Members of the Bête tribe in South Africa were tested. The Bête people normally lived in round mud huts and were unfamiliar with an environment that has rectangular corner house. The experimenters found that the effect of the Muller-Lyer illusion was significantly reduced because the tribal people didn’t have the prior learning or experience of a rectangular world to illuminate the illusion. This study suggests that cultural variability does indeed influences perception in the same way that learning and experience does.

Gregory & Wallace (1963), during the testing of S. B’s perceptual abilities administered an adjustable hand-held Muller-Lyer. Four readings were recorded. The result showed that the effect of the illusion was much reduced for S.B than for normal people, although a small percentage of normal people with the same apparent equality of length measurement could be found.

Fig. 2: Muller-Lyer illusion

Additionally, there are other types of perceptual constancy processes such a size, colour, shape and brightness that can influence visual perception. Morgan, hole & Glennerster (1990) argued that we have an inherent bias towards the lines with outward fins because we cannot segregate the lines from the context which in case are the inward and outward fins. This would seem to correlate with the gestalt concept of figure and background. This next paragraph shall look at how psychologists have tried to explain the Ames room illusion which is also categorized as a top-down cognitive processing illusion.

The Ames room illusion is another illustration of how the visual system often makes perceptual errors by not seeing what is really there. When the Ames room is viewed through a viewing point, it appears like a normal rectangular room (see Fig.3). A person on one corner of the room appears to have shrunken compared with the person on the other side of the room. In Fig. 4, the girl on the left looks about half the size of the girl on the right corner. In reality the room is constructed of trapezoidal floor, walls and ceiling. In fact the viewed from certain angle, Ames room is slightly similar to the Ponzo illusion which is caused by incorrect constant scaling. Gregory (1987) argues that the illusion might be the consequence of the rooms’ odd and irregular construction. However, Seckel and Karke (1995) demonstrated that the room itself does not contribute much to the illusion. The authors found that the size illusion is due largely to the horizontal line of the room and the perspective of the background. This relates to the Gestalt perceptual organisation theory of figure and ground.

Fig. 3: Ames Room shape

The Ames room illusion is effective because the girl on the right appears huge in relation to her twin. The reason for this is that the twin on the left is projecting a smaller image on the retina and the twin on the right is projecting a bigger image. Therefore, the visual system utilizes both of this retina images while assuming that both are equally at the same distance (Seckel & Klarke, 1997). According to Gregory (1987) the visual system interprets this ambiguous illusion through a top down processing system, where previous experience of rectangular rooms is used in the analysis of this new situation. This is supported by Segall et al. (1963) who found that the Muller-Lyer’s illusory effects was significantly reduced when perceived by disingenuous people who are unfamiliar with a rectangular view world. Other studies have reported the same findings (Stewart, 1971; Deregowski, 1989) with the addition that age and gender increases susceptibility to the Ames room illusion. This next paragraph examines in some detail the Hermann grid illusion.

Fig. 4: Ames Room.

The Hermann grid was discovered by Ludimar Hermann in 1870. In this illusion shown in Fig. 5, gray spots are seen at the intersection grids even though the light intensity is constant but they disappear when they are looked at directly. This illusion as mention earlier uses a bottom-up processing because previous learning is not needed to perceive it. Baumgartner (1960) suggested that the illusion occurs because of differences in ganglion cell spike rates when their receptive field looks at the on-centre of the intersection and the non intersecting grid line. He used an oculo-perimetric device to measure the spike rates in neuronal ganglion cells in response to the Hermann grid illusion. He concluded that looking in between the grid line produces a high spike rate in the ganglion cell because there are only 2 white inhibitory space whilst looking straight at the centre of the intersection however produces a lower spike rate because there are four white region in the inhibitory space. This neural process called lateral inhibition. Lateral inhibition is ability of a excited neuron to suppress neighbouring ones activities when excited (Eagleman, 2001).

Furthermore, the Baumgartner (1960 citied in Morgan, 1996) suggest that the reason that the gray spots disappear when they are directly looked at is that the fovea perceptive field is smaller when we direct the fovea centralis which is at the centre of the macula in the retina. The fovea function is to utilize our clearest and most focused vision. Schiller & Carvey (2005) dispute this theory by arguing, first, that the effect of the illusion doesn’t rely on size, and second, that when the illusion is reversed so that the colour is the other way round, so that the ganglion cells fall on the darker colour the effect still remains.

Additionally, Geier et al. (2008) argue that the Hermann grid effect is dependent on straightness by demonstrating how the illusory effect disappears when modifications are made to the lines as shown in Fig. 6. This illusion and others like the Mach Bands illusion has taught psychologists much about the lateral interaction that takes place between cells (Eagleman, 2001).

Fig. 5 Fig. 6

Conclusion

Visual illusions have taught psychologists that seeing can be deceiving. They provide us with a wealth of evidence as to how our visual perception can be affect by the way incoming input is coded during the sensory process plus the ways that different factors such as constancies, depth cues, distance scaling, learning, expectation, experiences and culture can dictate and influence our sensation and perception.

Conclusively, visual illusions demonstrate that our visual system, imperfect though it is, is a very complex system that will continue to challenge our understanding of the human perception processes.