III. ASSEMBLING THE VISUAL IMAGE
The 100 million receptors in the eye do not each have their own nerve fiber (axon) pathways to the visual cortex. This is fortunate! Imagine an optic nerve 100 times larger: about as thick as your wrist, there would be little room for anything else in the brain!
To send an image based on 100 million receptors to the rest of the brain using only 1/100 as many pathways, the visual system uses a 3-fold strategy:
1. It has high acuity only directly ahead in the central region of the retina – the fovea.
2. It uses eye movements to aim the fovea in various directions of interest.
3. It relies on memory to maintain an overall accurate image of what exists in those directions not being viewed at a particular moment.
Within these limits, differences in photon density are converted into brightness effects, and differences in photon wavelength are converted to distinct color effects. The scatter of these photons as they pass through the eye is reigned in by lateral inhibition. This keeps these effects in their place in the visual image.
The picture in our head may look like a very fine grained photograph. However, mapping the “grain” of retinal photoreceptors and ganglion cell pathways into a visual image is just the beginning.
Examining what happens in the visual cortex by physiological recordings and perception experiments reveals processes far different from what was expected.
A mere neural “projection system” would be too constrained by the small dynamic range of neural impulse rates, the limited number of available connections from retina to cortex, slow transmission speed and would be too inflexible to take advantage of known possibilities about the surfaces reflecting those photons. Evolution has taken advantage of a hundred trillion synapses’ enormous information processing potential to overcome these constraints. (The survival value of doing so may also have helped gain that astronomical number.)
Three processes form those localized differences into shapes:
1. Many neural pathways from adjacent retinal locations converge on single cells in the cortex, making them sensitive to lines and edges.
2. The responses from specific line and edge-sensitive cells converge on other cells, making them sensitive to specific shapes, such as faces.
3. Many line-sensitive cells of the same orientation and equal spacing converge on single cells, making them sensitive to arrays of parallel lines that represent specific spatial frequencies of the retinal image.
Transformation of images into spatial frequencies is the most efficient known means for transmitting and processing any image.
Still, this is not sufficient.
To produce a picture, the shapes must be organized. The neurophysiological mechanisms for this are only partly understood. However, visual perception experiments) can reach beyond the neurophysiology to reveal principles. used to do the organizing. See Appendix E – Psychophysical Methods. Some of these effects are explained in the next section.