The third and more recently discovered category of photosensitive cells is probably not involved in image-forming vision. These are a small proportion, about 2% in humans, of the retina's ganglion cells, themselves photosensitive through the photopigment melanopsin, which transmit information about light through the RHT (retinohypothalamic tract) directly to the SCN (suprachiasmatic nucleus) and other brain structures. Signals from these ganglion cells are used to adjust the size of the pupil, entrain the body's circadian rhythms and acutely suppress the pineal hormone melatonin, processes which in fact function in many blind people who do not have functioning rods and cones.
While rods and cones respond maximally to wavelengths around 555 nanometers (green), the light sensitive ganglion cells respond maximally to about 480nm (blue-violet). There are several different photopigments involved.
Neural signals from the rods and cones undergo complex processing by other neurons of the retina. The output takes the form of action potentials in retinal ganglion cells whose axons form the optic nerve. Several important features of visual perception can be traced to the retinal encoding and processing of light.
In vertebrate embryonic development, the retina and the optic nerve originate as outgrowths of the developing brain. Hence, the retina is part of the central nervous system (CNS). It is the only part of the CNS that can be imaged directly.
The vertebrate retina has ten distinct layers. From innermost to outermost, they include:
In section the retina is no more than 0.5 mm thick. It has three layers of nerve cells and two of synapses. The optic nerve carries the ganglion cell axons to the brain and the blood vessels that open into the retina. As a byproduct of evolution, the ganglion cells lie innermost in the retina while the photoreceptive cells lie outermost. Because of this arrangement, light must first pass through the thickness of the retina before reaching the rods and cones. However it does not pass through the epithelium or the choroid (both of which are opaque).
The white blood cells in the capillaries in front of the photoreceptors can be perceived as tiny bright moving dots when looking into blue light. This is known as the blue field entoptic phenomenon (or Scheerer's phenomenon).
Between the ganglion cell layer and the rods and cones there are two layers of neuropils where synaptic contacts are made. The neuropil layers are the outer plexiform layer and the inner plexiform layer. In the outer the rod and cones connect to the vertically running bipolar cells and the horizontally oriented horizontal cells connect to ganglion cells.
The central retina is cone-dominated and the peripheral retina is rod-dominated. In total there are about seven million cones and a hundred million rods. At the centre of the macula is the foveal pit where the cones are smallest and in a hexagonal mosaic, the most efficient and highest density. Below the pit the other retina layers are displaced, before building up along the foveal slope until the rim of the fovea or parafovea which is the thickest portion of the retina. The macula has a yellow pigmentation from screening pigments and is known as the macula lutea.
The cephalopod retina does not originate as an outgrowth of the brain, as the vertebrate one does. This shows that vertebrate and cephalopod eyes are not homologous but have evolved separately.
The cones respond to bright light and mediate high-resolution vision and colour vision. The rods respond to dim light and mediate lower-resolution, black-and-white, night vision. It is a lack of cones sensitive to red, blue, or green light that causes individuals to have deficiencies in colour vision or various kinds of colour blindness. Humans have three different types of cones (trichromatic vision) while other mammals lack cones with red sensitive pigment and therefore have poorer (dichromatic) colour vision.
When light falls on a receptor it sends a proportional response synaptically to bipolar cells which in turn signal the retinal ganglion cells. The receptors are also 'cross-linked' by horizontal cells and amacrine cells, which modify the synaptic signal before the ganglion cells. Rod and cone signals are intermixed and combine, although rods are mostly active in very poorly lit conditions and saturate in broad daylight, while cones function in brighter lighting because they are not sensitive enough to work at very low light levels.
Despite the fact that all are nerve cells, only the retinal ganglion cells and few amacrine cells create action potentials. In the photoreceptors, exposure to light hyperpolarizes the membrane in a series of graded shifts. The outer cell segment contains a photopigment. Inside the cell the normal levels of cyclic guanosine monophosphate (cGMP) keep the Na+ channel open and thus in the resting state the cell is depolarised. The photon causes the retinal bound to the receptor protein to isomerise to trans-retinal. This causes receptor to activate multiple G-proteins. This in turn causes the Ga-subunit of the protein to bind and degrade cGMP inside the cell which then cannot bind to the Na+ cyclic nucleotide-gated ion channels (CNGs). Thus the cell is hyperpolarised. The amount of neurotransmitter released is reduced in bright light and increases as light levels fall. The actual photopigment is bleached away in bright light and only replaced as a chemical process, so in a transition from bright light to darkness the eye can take up to thirty minutes to reach full sensitivity (see Adaptation (eye)).
In the retinal ganglion cells there are two types of response, depending on the receptive field of the cell. The receptive fields of retinal ganglion cells comprise a central approximately circular area, where light has one effect on the firing of the cell, and an annular surround, where light has the opposite effect on the firing of the cell. In ON cells, an increment in light intensity in the centre of the receptive field causes the firing rate to increase. In OFF cells, it makes it decrease. In a linear model, this response profile is well described by a Difference of Gaussians and is the basis for edge detection algorithms. Beyond this simple difference ganglion cells are also differentiated by chromatic sensitivity and the type of spatial summation. Cells showing linear spatial summation are termed X cells (also called "parvocellular", "P", or "midget" ganglion cells), and those showing non-linear summation are Y cells (also called "magnocellular, "M", or "parasol" retinal ganglion cells), although the correspondence between X and Y cells (in the cat retina) and P and M cells (in the primate retina) is not as simple as it once seemed.
In the transfer of visual signals to the brain, the visual pathway, the retina is vertically divided in two, a temporal (nearer to the temple) half and a nasal (nearer to the nose) half. The axons from the nasal half cross the brain at the optic chiasma to join with axons from the temporal half of the other eye before passing into the lateral geniculate body.
Although there are more than 130 million retinal receptors, there are only approximately 1.2 million fibres (axons) in the optic nerve; a large amount of pre-processing is performed within the retina. The fovea produces the most accurate information. Despite occupying about 0.01% of the visual field (less than 2° of visual angle), about 10% of axons in the optic nerve are devoted to the fovea. The resolution limit of the fovea has been determined at around 10,000 points. The information capacity is estimated at 500,000 bits per second (for more information on bits, see information theory) without colour or around 600,000 bits per second including colour.
The retina, unlike a camera, does not simply send a picture to the brain. The retina spatially encodes (compresses) the image to fit the limited capacity of the optic nerve. Compression is necessary because there are 100 times more Photoreceptor cells than ganglion cells as mentioned above. The retina does so by "decorrelating" the incoming images in a manner to be described below. These operations are carried out by the center surround structures as implemented by the bipolar and ganglion cells.
There are two types of center surround structures in the retina -- on-centers and off-centers. On-centers have a positively weighted center and a negatively weighted surround. Off-centers are just the opposite. Positive weighting is more commonly known as excitatory and negative weighting is more commonly known as inhibitory.
These center surround structures are not physical in the sense that you cannot see them by staining samples of tissue and examining the retina's anatomy. The center surround structures are logical (i.e., mathematically abstract) in the sense that they depend on the connection strengths between ganglion and bipolar cells. It is believed that the connection strengths between cells is caused by the number and types of ion channels embedded in the synapses between the ganglion and bipolar cells. Stephen Kuffler in the 1950s was the first person to begin to understand these center surround structures in the retina of cats. See Receptive field for figures and more information on center surround structures. See chapter 3 of David Hubel's on-line book (listed below) for an excellent introduction.
The center surround structures are mathematically equivalent to the edge detection algorithms used by computer programmers to extract or enhance the edges in a digital photograph. Thus the retina performs operations on the image to enhance the edges of objects within its visual field. For example, in a picture of a dog, a cat and a car, it is the edges of these objects that contain the most information. In order for higher functions in the brain (or in a computer for that matter) to extract and classify objects such as a dog and a cat, the retina is the first step to separating out the various objects within the scene.
As an example, the following matrix is at the heart of the computer algorithm that implements edge detection. This matrix is the computer equivalent to the center surround structure. In this example, each box (element) within this matrix would be connected to one photoreceptor. The photoreceptor in the center is the current receptor being processed. The center photoreceptor is multiplied by the +1 weight factor. The surrounding photoreceptors are the "nearest neighbors" to the center and are multiplied by the -1/8 value. The sum of all nine of these elements is finally calculated. This summation is repeated for every photoreceptor in the image by shifting left to the end of a row and then down to the next line.
The total sum of this matrix is zero if all the inputs from the nine photoreceptors are the same value. The zero result indicates the image was uniform (non-changing) within this small patch. Negative or positive sums mean something was varying (changing) within this small patch of nine photoreceptors.
The above matrix is only an approximation to what really happens inside the retina. First, the above example is called "balanced". The term balanced means that the sum of the negative weights is equal to the sum of the positive weights so that they cancel out perfectly. Retinal ganglion cells are almost never perfectly balanced. Second, the table is square while the center surround structures in the retina are circular. Third, neurons operate on spike trains traveling down nerve cell axons. Computers operate on a single number that is essentially constant from each input pixel. (The computer pixel is basically the equivalent of a biological photoreceptor.) Fourth, the retina performs all these calculations in parallel while the computer operates on each pixel one at a time. There are no repeated summations and shifting as there would be in a computer. Finally, the horizontal and amacrine cells play a significant role in this process but that is not represented here.
Here is an example of an input image and how edge detection would modify it.
Once the image is spatially encoded by the center surround structures, the signal is sent out the optical nerve (via the axons of the ganglion cells) through the optic chiasm to the LGN (lateral geniculate nucleus). The exact function of the LGN is unknown at this time. The output of the LGN is then sent to the back of the brain. Specifically the output of the LGN "radiates" out to the V1 Primary visual cortex.
Simplified Signal Flow: Photoreceptors ==> Bipolor ==> Ganglion ==> Chiasm ==> LGN ==> V1 cortex
Treatment depends upon the nature of the disease or disorder. Transplantation of retinas has been attempted, but without much success. At MIT, The University of Southern California, and the University of New South Wales, an "artificial retina" is under development: an implant which will bypass the photoreceptors of the retina and stimulate the attached nerve cells directly, with signals from a digital camera.
Robert MacLaren and colleagues at University College London and Moorfields Eye Hospital in London showed in 2006 that photoreceptor cells could be transplanted successfully in the mouse retina if donor cells were at a critical developmental stage.