Chapter 15: The Visual System

The visual system conveys more information to the brain than any other afferent system. This information is processed within the brain so as to form a set of maps of the visual world. A relatively large proportion of brain tissue is devoted to vision. The visual system includes the eye and retina, the optic nerves, and the visual pathways within the brain, where multiple visual centers process information about different aspects (shape and form, color, motion) of visual stimuli.

The functions (and clinical correlations) of the cranial nerves (III, IV, VI) involved in moving the eyes have been discussed in Chapter 8, along with the gaze centers and pupillary reflexes. The vestibulo-ocular reflex is briefly explained in Chapter 17. This chapter discusses the form, function, and lesions of the optic system from the retina to the cerebrum.

Anatomy and Physiology

The optical components of the eye are the cornea, the pupillary opening of the iris, the lens, and the retina (Fig 15–1). Light passes through the first four components, the anterior chamber, and the vitreous to reach the retina; the point of fixation (direction of gaze) normally lines up with the fovea. The retina (which develops as a portion of the brain itself, and is considered by some neuroscientists to be a specialized part of the brain, located within the eye) transforms light into electrical impulses (Fig 15–2).

The retina, organized into 10 layers, contains two types of photoreceptors (rods and cones) and four types of neurons (bipolar cells, ganglion cells, horizontal cells, and amacrine cells) (Figs 15–2 and 15–3). The photoreceptors (rods and cones, which are first-order neurons) synapse with bipolar cells (Fig 15–4). These, in turn, synapse with ganglion cells; the ganglion cells are third-order neurons whose axons converge to leave the eye within the optic nerve.

Within the outer plexiform layer of the retina, horizontal cells connect receptor cells with each other. Amacrine cells, within the inner plexiform layer, connect ganglion cells to one another (and in some cases also connect bipolar cells and ganglion cells).

Retinal Rods

Rods are more numerous than cones. These photoreceptors that are sensitive to low-intensity light and provide visual input when illumination is dim (eg, at twilight and at night). Cones are stimulated by relatively high-intensity light; they are responsible for sharp vision and color discrimination. Rods and cones each contain an outer segment, consisting of stacks of flattened disks of membrane that contain photosensitive pigments that react to light. In addition, they each have an inner segment, which contains the cell nucleus and mitochondria and forms synapses with the second-order bipolar cells. The transduction of light into neural signals occurs when photons are absorbed by photosensitive molecules (called visual pigments) in the rods and cones.

The visual pigment within retinal rods is rhodopsin, a specialized membrane receptor that is linked to G-proteins. When light strikes the rhodopsin molecule, it is converted, first to metarhodopsin II and then to scotopsin and retinene₁. This light-activated reaction activates a G-protein known as transducin, which breaks down cyclic guanosine monophosphate (GMP). Because cyclic GMP acts within the cytoplasm of the photoreceptors to keep sodium channels open, the light-induced reduction in cyclic GMP leads to a closing of sodium channels, which causes a hyperpolarization (see Chapter 3). Thus, as a result of being struck by light, there is hyperpolarization within the retinal rods. This results in decreased release of synaptic transmitter onto bipolar cells, which alters their activity (Fig 15–5).

Retinal Cones

Cones within the retina also contain visual pigments, which respond maximally to light at wavelengths of 440, 535, and 565 nm (corresponding to the three primary colors: blue, green, and red). When cones are struck by light of the appropriate wavelength, a cascade of molecular events, similar to that in rods, activates a G-protein that closes sodium channels, resulting in hyperpolarization.

Bipolar, Amacrine, and Retinal Ganglion Cells

Transmission from photoreceptors (rods and cones, first-order sensory neurons) to bipolar cells (second-order sensory neurons) and then to retinal ganglion cells (third-order sensory neurons) is modified by horizontal cells and amacrine cells. Each bipolar cell receives input from 20 to 50 photoreceptor cells. The receptive field of the bipolar cell (ie, the area on the retina that influences the activity in the cell) is modified by horizontal cells. The horizontal cells form synapses on photoreceptors and nearby bipolar cells in a manner that "sharpens" the receptive field on each bipolar cell. As a result of this arrangement, bipolar cells do not merely respond to diffuse light; on the contrary, some bipolar cells convey information about small spots of light surrounded by darkness. (These cells have "on"-center receptive fields, whereas others convey information about small, dark spots surrounded by light ["off"-center receptive fields].)

Amacrine cells receive input from bipolar cells and synapse onto other bipolar cells near their sites of input to ganglion cells. Like horizontal cells, amacrine cells "sharpen" the responses of ganglion cells. Some ganglion cells respond most vigorously to a light spot surrounded by darkness, whereas others respond most actively to a dark spot surrounded by light.

The retinal area for central, fixated vision during good light is the macula (Fig 15–6). The inner layers of the retina in the macular area are pushed apart, forming the fovea centralis, a small, central pit composed of closely packed cones, where vision is sharpest and color discrimination is most acute.

Retinal ganglion cells are specialized neurons that can be grouped into two classes that subserve different functions. Magnocellular ganglion cells have larger diameter axons (faster conduction velocities) and are sensitive to motion but not to color or details of form. Parvocellular ganglion cells have thinner axons (slower conduction velocities) and convey information about form and color. These two information streams converge on different layers of the lateral geniculate nucleus (see Visual Pathways section), an important central target.

Ganglion cell axons, within the retina, form the nerve fiber layer. The ganglion cell axons all leave the eye, forming the optic nerve, at a point 3 mm medial to the eye's posterior pole. The point of exit is termed the optic disk and can be seen through the ophthalmoscope (see Fig 15–6). Because there are no rods or cones overlying the optic disk, it corresponds to a small blind spot in each eye.

A. Adaptation

If a person spends time in brightly lighted surroundings and then moves to a dimly lighted environment, the retinas slowly become more sensitive to light as the individual becomes accustomed to the dark. This decline in visual threshold, known as dark adaptation, is nearly maximal in about 20 minutes, although there is some further decline over longer periods. On the other hand, when one passes suddenly from a dim to a brightly lighted environment, the light seems intensely bright until the eyes adapt to the increased illumination and the visual threshold rises. This adaptation occurs over a period of about 5 minutes and is called light adaptation. The pupillary light reflex, which constricts the pupils, is normally a protective accompaniment to sudden increases in light intensity (see Chapter 8).

Light and dark adaptation depend, in part, on changes in the concentration of cyclic GMP in photoreceptors. In sustained illumination, there is a reduction in the concentration of calcium ions within the photoreceptor, which leads to increased guanylate cyclase activity and increased cyclic GMP levels. This, in turn, tends to keep sodium channels open so as to desensitize the photoreceptor.

B. Color Vision

The portion of the spectrum that stimulates the retina to produce sight ranges from 400 to 800 nm. Stimulation of the normal eye, either by this entire range of wavelengths or by mixtures from certain different parts of the range, produces the sensation of white light. Monochromatic radiation from one part of the spectrum is perceived as a specific color or hue.

The Young-Helmholtz theory postulates that the retina contains three types of cones, each with a different photopigment maximally sensitive to one of the primary colors (red, blue, and green). The sensation of any given color is determined by the relative frequency of impulses from each type of cone. Parvocellular ganglion cells receive color-specific signals from the three types of cones and relay them to the brain via the optic nerve.

Each of the three photopigments has been identified. The amino acid sequences of all three are about 41% homologous with rhodopsin. The green-sensitive and red-sensitive pigments are very similar (about 40% homologous with each other) and are coded by the same chromosome. The blue-sensitive pigment is only about 43% homologous with the other two and is coded by a different chromosome.

In normal color (trichromatic) vision, the human eye can perceive the three primary colors and can mix these in suitable portions to match white or any color of the spectrum. Color blindness can result from a weakness of one cone system or from dichromatic vision, in which only two cone systems are present. In the latter case, only one pair of primary colors is perceived; the two colors are complementary to each other. Most dichromats are red–green blind and confuse red, yellow, and green. Color blindness tests use special cards or colored pieces of yarn.

C. Accommodation

The lens is held in place by fibers between the lens capsule and the ciliary body (Figs 15–1 and 15–7). In the unaccommodated state, these elastic fibers are taut and keep the lens somewhat flattened. In the accommodated state, contraction of the circular ciliary muscle slackens the tension on the elastic fibers, and the lens, which has an intrinsic capacity to become rounder, assumes a more biconvex shape. The ciliary muscle is a smooth muscle that is innervated by the parasympathetic system (cranial nerve III; see Chapter 8); it can be paralyzed with atropine or similar drugs.

D. Refraction

When one views a distant object, the normal (emmetropic) eye is unaccommodated and the object is in focus. A normal eye readily focuses an image of a distant object on its retina, 24 mm behind the cornea; the focal length of the optics and the distance from cornea to retina are well matched, a state known as emmetropia (Fig 15–8). To bring closer objects into focus, the eye must increase its refractive power by accommodation. The ability of the lens to do so decreases with age as the lens loses its elasticity and hardens. The effect on vision usually becomes noticeable at approximately 40 years of age; by the 50s, accommodation is generally lost (presbyopia).

Tests of Visual Function

In assessing visual acuity, distant vision is tested with Snellen or similar cards for persons with fairly normal vision. Finger counting and finger movement tests are used for those with subnormal vision, and light perception for those with markedly subnormal vision. Near vision is tested with reading cards.

Perimetry is used to determine the visual fields (Fig 15–9). The field for each eye (monocular field) is plotted with a device or by the confrontation method to determine the presence of a scotoma or other field defect (see Clinical Correlations under Visual Pathways section). Normally, the visual fields overlap in an area of binocular vision (Fig 15–10).

Anatomy

The visual pathways project from the retina, via the optic nerve, to the brain where they eventually reach the occipital cortex. Since the visual pathways extend over a long course, they are susceptible to injury at multiple points. An understanding of the anatomy of the visual pathways can enable a careful observer to localize lesions in many parts of the visual system on the basis of history and clinical examination.

The optic nerve consists of about a million nerve fibers and contains axons arising from the inner, ganglion cell layer of the retina. These fibers travel through the lamina cribrosa of the sclera and then course through the optic canal of the skull to form the optic chiasm (Fig 15–14). The fibers from the nasal half of the retina decussate within the optic chiasm; those from the lateral (temporal) half do not.

The arrangement of the optic chiasm is such that the axons from the lateral half of the left retina and the nasal half of the right retina project centrally behind the chiasm within the left optic tract. As a result of the optics of the eye, these two halves of the left and right retina receive visual information from the right-sided half of the visual world. This anatomic arrangement permits the left hemisphere to receive visual information about the contralateral (right-sided) half of the visual world and vice versa (see Fig 15–14). After traveling through the optic chiasm, retinal ganglion cell axons travel centrally in the optic tract, which carries the axons to the lateral geniculate body (also termed the lateral geniculate nucleus) as well as the superior colliculus.

The lateral geniculate bodies and the medial geniculate bodies constitute important relay nuclei, for vision and hearing, respectively, within the thalamus. Each lateral geniculate nucleus is a six-layered structure. The different layers have different roles in visual processing. Signals from magnocellular and parvocellular retinal ganglion cells (see earlier discussion of Anatomy and Physiology under The Eye section) converge on different layers within the lateral geniculate. These signals preserve the architectural principle of multiple, parallel streams of visual information, each devoted to analyzing a different aspect of the visual environment. Crossed fibers from the optic tract terminate within laminas 1, 4, and 6, whereas uncrossed fibers terminate in laminas 2, 3, and 5. The optic tract axons terminate in a highly organized manner, and their synaptic endings are organized in a map-like (retinotopic) fashion that reproduces the geometry of the retina. (The central part of the visual field has a relatively large representation in the lateral geniculate bodies, probably providing greater visual resolution, or sensitivity to detail, in this region.) The receptive fields of neurons in the lateral geniculate bodies usually consist of an "on" center associated with an "off" surround or vice versa.

From each lateral geniculate body, axons project ipsilaterally by way of the optic radiation to the calcarine cortex in the occipital lobe. Thus, the right halves of each retina (corresponding to the left halves of the visual world) project by way of the optic radiation to the right occipital lobe and vice versa.

The geniculocalcarine fibers (optic radiations) carry impulses from the lateral geniculate bodies to the visual cortex. Meyer's loop is the sweep of geniculocalcarine fibers that curves around the lateral ventricle, reaching forward into the temporal lobe, before proceeding toward the calcarine cortex. Meyer's loop carries optic radiation fibers representing the upper part of the contralateral visual field. Within the cortex, there is a more extensive representation for the area of central vision (Fig 15–15).

In addition to projecting to the lateral geniculate, retinal ganglion cell axons in the optic tract terminate in the superior colliculus, where they form another retinotopic map. The superior colliculus also receives synapses from the visual cortex. The superior colliculus projects to the spinal cord via the tectospinal tracts, which coordinate reflex movements of the head, neck, and eyes in response to visual stimuli (see Chapter 13).

Other afferents from the optic tract project, via the pretectal area, to parasympathetic neurons in the Edinger–Westphal nucleus (part of the oculomotor nucleus). These parasympathetic neurons send axons within the oculomotor nerve and terminate in the ciliary ganglia (see Fig 15–14). Postganglionic neurons, within the ciliary ganglia, project to the sphincter muscles of the iris. This loop of neurons is responsible for the pupillary light reflex, which results in constriction of the pupil in response to stimulation of the eye with light. Visual axons in each optic tract project to the Edinger–Westphal nucleus bilaterally. As might be expected, when light is shown in one eye, there is constriction of the pupil not only in the ipsilateral eye (the direct light reflex) but also in the contralateral eye (the consensual light reflex).

Anatomy

Visual information is relayed from the lateral geniculate body to the visual cortex via myelinated axons in the optic radiations. As described later, multiple retinotopic maps of the visual world are present within the cortex. The primary visual cortex is the main way station for incoming visual signals. Ultimately, however, visually responsive neurons within at least six parts of the occipital cortex and within the temporal and parietal lobes form separate visual areas, each with its own map of the retina.

A functional magnetic resonance image showing activation of the visual cortex in response to a patterned visual stimulus is displayed in Figure 15–18. When the left half of the visual field is stimulated with a visual pattern, the visual cortex on the right side is activated and vice versa.

The primary visual cortex receives its blood from the calcarine branch of the posterior cerebral artery. The remainder of the occipital lobe is supplied by other branches of this artery. The arterial supply can be (rarely) interrupted by emboli or by compression of the artery between the free edge of the tentorium and enlarging or herniating portions of the brain.

Primary Visual Cortex

The primary visual cortex (also termed calcarine cortex, area 17, or V1) is located on the medial surface of the occipital lobe above and below the calcarine fissure (Fig 15–15). This cortical region is also called the striate cortex because, when viewed in histologic sections, it contains a light-colored horizontal stripe (corresponding to white matter–containing myelinated fibers) within lamina IV. When stained for the mitochondrial enzyme cytochrome oxidase, superficial layers (layers 2 and 3) of area 17 appear to be organized into enzyme-rich regions (termed blob regions on their discovery) and enzyme-poor interblob regions. Within the superficial layers of area 17, parvocellular inputs carrying color information tend to project to the blob regions, whereas inputs concerned with shape as well as color converge on the interblob regions. Magnocellular inputs, carrying information about motion, depth, and form, in contrast, project to deeper layers of the striate cortex.

Visual Association (Extrastriate) Cortex

Beyond the primary visual cortex, several other visual areas (area 18 and area 19) extend concentrically outside the primary visual cortex. These areas are also called the extrastriate cortex or the visual association cortex. Two separate retinotopic visual maps are located within area 18 (V2 and V3), and three retinotopic maps are located in area 19 (V3A, V4, and V5). V2 contains cytochrome-rich stripes, separated by cytochrome-poor interstripes. Continuing the theme of multiple, parallel visual information-processing pathways, magnocellular inputs relay within the thick stripe regions, whereas parvocellular inputs are processed in interstripe and thin stripe zones of V2.

Still another visual area, termed MT, is located on the posterior part of the superior temporal sulcus. This visual area receives and analyzes information about the location of visual stimuli but not their shape or color. The MT area does not provide information about what a stimulus is but does provide information about where it is located.

Histology

The primary visual cortex appears to contain six layers. It contains a line of myelinated fibers within lamina IV (the line of Gennari, or the external line of Baillarger; see Fig 15–19). The stellate cells of lamina IV receive input from the lateral geniculate nucleus, and the pyramidal cells of layer V project to the superior colliculus. Layer VI cells send a recurrent projection to the lateral geniculate nucleus.

Physiology

As noted earlier, there is an orderly mapping (again termed retinotopic) of the visual world onto multiple parts of the visual cortex. The projection of the macular part of the retina is magnified within these maps, a design feature that probably provides increased sensitivity to visual detail in the central part of the visual field.

As visual information is relayed from cell to cell in the primary visual cortex, it is processed in increasingly complex ways (Fig 15–20). Simple cells in the visual cortex have receptive fields that contain an "on" or "off" center, shaped like a rectangle with a specific orientation, flanked by complementary zones. Simple cells usually respond to stimuli at one particular location. For example, an "on"-center simple cell may respond best to a bar, precisely oriented at 458, flanked by a larger "off" area, at a particular location. If the bar is rotated slightly or moved, the response of the cell will be diminished. Thus, these cells respond to lines, at specific orientations, located in particular regions within the visual world.

Complex cells in the visual cortex have receptive fields that are usually larger than those of simple cells (see Fig 15–20). These cells respond to lines or edges with a specific orientation (eg, 608) but are excited whenever these lines are present anywhere within the visual field regardless of their location. Some complex cells are especially sensitive to movement of these specifically oriented edges or lines.

D. Hubel and T. Wiesel, who received the Nobel Prize for their analysis of the visual cortex, suggested that the receptive fields of simple cells in the visual cortex could be built up from the simpler fields of visual neurons in the lateral geniculate. The pattern of convergence of geniculate neurons onto visual cortical cells supports this hypothesis. Similarly, by projecting onto a complex cell in the visual cortex, a set of simple cells with appropriate receptive fields can create a higher-level response that recognizes lines and edges at a particular orientation at a variety of positions.

The visual cortex contains vertical orientation columns, each about 1 mm in diameter. Each column contains simple cells whose receptive fields have almost identical retinal positions and orientations. Complex cells within these columns appear to process information so as to generalize by recognizing the appropriate orientation regardless of the location of the stimulus.

About one half of the complex cells in the visual cortex receive inputs from both eyes. The inputs are similar for the two eyes in terms of the preferred orientation and location of the stimulus, but there is usually a preference for one eye. These cells are referred to as showing ocular dominance, and they are organized into another overlapping series of ocular dominance columns, each 0.8 mm in diameter. The ocular dominance columns receiving input from one eye alternate with columns receiving input from the other (Fig 15–21).