Visual Attention – Focus And Saccades

Focus

Our visual system differs from a digital camera in the way it detects and processes color but also in its resolution. On a digital camera’s photo censor, photoreceptive elements are spread uniformly in a tight matrix, so the spatial resolution is constant across the entire image frame. The human visual system is not like that.

The spatial resolution of the human visual field drops greatly from the center to the edges. Each eye has approximately six million retinal cone cells. They are packed much more tightly in the center of our visual field – a small region called the fovea – than they are at the edges of the retina. The fovea is only about 1% of the retina, but the brain’s visual cortex devotes about 50% of its area to input from the fovea. Furthermore, foveal cone cells connect 1:1 to the ganglian neuron cells that begin the processing and transmission of visual data, while elsewhere on the retina, multiple photoreceptor cells (cones and rods) connect to each ganglion cell. In technical terms, information from the visual periphery is compressed (with data loss) before transmission to the brain, while information from the fovea is not. All this causes our vision to have much, much greater resolution in the center of our visual field than elsewhere (Lindsay, P., Norman, D.A., Human information processing, 1972; Waloszek, G., Vision and Visual Disabilities: An Introduction, 2005).

To visualize how small the fovea is compared to your entire visual field, hold your arm straight out and look at your thumb. Your thumbnail, viewed at arm’s length, corresponds approximately to the fovea (Ware, C., Visual Thinking for Design, 2008). While you have your eyes focused on the thumbnail, everything else in your visual field falls outside of your fovea on your retina.

In the fovea, people with normal vision have very high resolution: they can resolve several thousand dots within that region – better resolution than many of today’s pocket digital cameras. Just outside of the fovea, the resolution is already down to a few dozen dots per inch viewed at arm’s length. At the edges of our vision, the “pixels” of our visual system are as large as a melon (or human head) at arm’s length.

If our peripheral vision has such low resolution, one might wonder why we don’t see the world in a kind of tunnel vision where everything is out of focus except what we are directly looking at now. Instead, we seem to see our surroundings sharply and clearly all around us. We experience this illusion because our eyes move rapidly and constantly about three times per second even when we don’t realize it, focusing our fovea on selected pieces of our environment. Our brain fills in the rest in a gross, impressionistic way based upon what we know and expect. Our brain does not have to maintain a high-resoltion mental model of our environment because it can order eyes to sample and resample details in the environment as needed (Clark, A., Being There: Putting brain, body, and world together again, 1998).

For example when reading your eyes dart around, scanning and reading. No matter where on the page your eyes are focused, you have the impression of viewing a complete page of text, because, of course, you are. 

Related to this is the fact that the center of our visual field – the fovea and a small area immediately surrounding it – is the only part of our visual field that can read. The rest of our visual field cannot read. What this really means is that the neural networks starting in the fovea, running through the optic nerve to the visual cortex, and then spreading into various parts of our brain, have been trained to read, but the neural networks starting elsewhere in our retinas cannot read. All text that we read comes into our visual system after being scanned by the central area, which means that reading requires a lot of eye movement (Johnson, Jeff, Designing with the Mind in Mind, 2010)

Saccades

Each eye is moved by six muscles. The tendon passes through a ‘pulley’ in the skull, in front of the eyeball. The eyes are in continuous movement, and they move in various ways. When the eyes are moved around, searching for an object, they move quite differently from the way they move when a moving object is being followed with the eyes. When searching, they move in a series of small rapid jerks, but when following they move smoothly. The jerks are known as saccades (after an old French word meaning ‘the flick of a sail’). Apart from these two main types of movement, there is also a continuous small high-frequency tremor.

It turns out that the saccadic movements of the eyes are essential to vision. It is possible to fix the image on the retina so that whenever the eye moves, the images move with it and so remain fixed on the retina. When the image is optically stabilised vision fades after a few seconds, and so it seems that part of the function of eye movements is to sweep the image over the receptors so that they do not adapt and so cease to signal to the brain the presence of the image in the eye. But there is a curious problem: when we look at a sheet of white paper, the edges of the image of the paper will move around on the retina, and so stimulation will be renewed; but consider now the centre of the image. Here the small movements of the eyes can have no effect, for a region of given brightness is substituted for another region of exactly the same brightness, and so no change in stimulation takes place with the small eye movements. Yet the middle of the paper does not fade away. This suggests that borders and outlines are very important in perception. Large areas of constant intensity provide no information. They seem to be ‘inferred’ from the signal borders: the central visual system makes up the missing signals. (Gregory R.L., Eye and Brain – The psychology of seeing, 3rd Ed., 1978)

We constantly make eye movements to seek information. Moving our eyes causes different parts of the visual environment to be imaged on the high-resolution fovea, where we can see detail. These movements are frequent, between two and five jerky movements, called saccades, per second.

In visual search task, the eye moves rapidly from fixation to fixation. The dwell period is generally between 200 and 600 msec, and the saccade takes between 20 and 100 msec. The peak velocity of a saccade can be as much as 900°/sec (Hallett, P.E., Eye movements. In Handbook of Perception and Human Performance, 1986; Barfield, W., Hendrix, et al., Comparison of human sensory capabilities with technical spefications of virtual environment equipment. Presence 4(4), 1995).

Saccadic eye movements are said to be ballistic. This means that once brain decides to switch attention and make an eye movement, the muscle signals for accelerating and decelerating the eye are first programmed, then the program is run to make the eye movement. The movement cannot be adjusted in mid-saccade. During the course of a saccadic eye movement, we are less sensitive to visual input than we normally are. This is called saccadic suppression (Riggs, L.A., Merton, et al., Suppression of visual phosphenes during saccadic eye movements. Vision Research 14, 1974). The implication is that certain kinds of events can easily be missed if they occur while we happen to be moving our eyes.

Another implication of saccadic suppression is that it is reasonable to think of information coming into the visual system as a series of discrete snapshots. The brain is often processing rapid sequences of discrete images. This capacity is being increasingly exploited in television advertising, in which several cuts per second of video have become commonplace. (Ware, Colin, Information Visualization: Perception for Design, 2nd Ed., 2004).