Photoreceptors respond to light stimuli by a phototransduction cascade, creating an electrical response that is transmitted through specialized neurons, bipolar horizontal amacrine cells, and subsequently to ganglion cells for transmission to the lateral geniculate and the visual cortex.

Rods and cones serve different functions and have different phototransduction. They also have different neuronal connections and ganglion cells, although the two systems interact to process visual information.

Cone ganglion cells are primarily parvocellular (P) serving color vision, high spatial frequencies, fine two-point discrimination, and stereopsis. Rod ganglion cells are primarily magnocellular (M) responding to direction, motion, speed, flicker, gross binocular disparities, and gross stereopsis; they do not process color (wavelength in the visible spectrum) but only perceive light. Cones are primarily central in the macula, rods peripheral. This separation persists in the lateral geniculate and striate cortex. However, there are neuronal interconnections between these two sets of neurons.

All photoreceptors have the ability to register the direction from which light entering them came. This activity is, in part, due to neuronal integration of rods and cones as receptor fields. These fields are smaller for the centrally placed cones and larger for peripheral rods. They are of two types, on center with inhibitory surround, and off center with excitatory surround. In addition, there is a considerable overlapping of fields. A single retinal receptor influences multiple ganglion cells. It contributes to the receptor field center of some cells and to the surrounds of others.

Photoreceptors in the right and left half of each retina project topographically via the chiasm and appropriate lateral geniculate to the right brain hemisphere for the right half of each retina and to the left brain hemisphere for the left half of each retina. This fact means that fibers from a small given region of the retina project to a given region of the lateral geniculate and, subsequently, to a given region of the primary visual cortex as ocular dominance columns.

At both the lateral geniculate and visual cortex, the input from each retinal half is separated in layers each receiving input from either the right or left retina only, but right and left derived layers are adjacent. Since each small retinal point in the corresponding half visual field of each eye (retina) is now anatomically correlated, binocular interaction (stimulus integration) becomes possible through neuronal interconnections. Disruptions in this system lead to amblyopia, abnormal retinal correspondence, and suppression scotomas in microtropic strabismus.

The fovea registers the visual object as straight ahead while each stimulated neuroepithelial element registers the object as displaced from the fovea by the number of degrees that element is separated from the fovea. The contour of each image is determined by the contrast produced between stimulated and unstimulated photoreceptors. Light waves emanating from objects in the visual field form a miniaturized retinal image. The fovea possessing the highest resolving power for contrast, therefore, has maximal perception of contour achieving the best visual acuity. The fixation point is the spatial location of the object of regard. The fixation line (visual axis) is a line drawn between the fixation point and the fovea.

Binocular summation means the detection threshold for a stimulus is lower with two eyes compared to that of one eye. This ratio is about 1 to 1.41. Clinically we experience this phenomenon as the improved Snellen acuity with both eyes focused on an object rather than one at a time. Some visual cortical cells also show a greater sensitivity when stimulated by both eyes rather than one eye. Binocular improvement better than 1.41 is thought to be due to such neuronal summation.

Best visual acuity is obtained and maintained by an involuntary reflex designed to steer the fixation line onto the object of regard. The development of the fixation reflex is first manifested when the full-term infant is 4 to 5 weeks old. The eyes follow a light or a bright object for a few degrees, but once fixation is interrupted, reestablishment is slow. At 3 months of age, the infant maintains fixation in all fields of gaze and reestablishes fixation instantly after interruption. By 4 months of age, the infant begins to integrate fixation with grasping movements in an attempt to bring the target of fixation to the mouth. Oral identification of the object will continue for the next 4 months before being replaced by visual and tactile identification.

By the time the child is 9 years old, the constant conditioning of the fixation reflex gives it an irreversible quality comparable to an unconditioned reflex; however, before this age, cessation of reinforcing this conditioning process results in its reversal, as manifested by amblyopia. If an infant is totally deprived of the opportunity to initiate stimulation of the fixation reflex before 3 months of age, the fixation reflex will never develop. Corrective treatment that eliminates the stimulus deprivation during approximately the first 3 months of life can facilitate the fixation reflex development. This conclusion has been supported by the clinical experience in treating infants with unilateral cataracts.

Partial deprivation of stimuli, as occurs in intermittent strabismus and anisometropia, results in a less severe amblyopia and a relatively better prognosis for normal development of the fixation reflex with treatment. Treatment of strabismus amblyopia by 4 years of age is almost always successful, but the favorable prognosis decreases as the age at which treatment is initiated increases.

Bilateral macular images below a critical threshold of intensity and clarity in infants results in a “pendular nystagmus” also known as “nystagmus of the blind” and “deprivation nystagmus.” This clinical finding can result from retinal or visual pathway lesions. Deprivation nystagmus does not occur in patients with only a unilateral absence of the fixation reflex even when the normal eye is occluded.

According to clinical observation, the fixation reflex appears to mature by the time the child is approximately 9 years old. Visual acuity remains in a state of flux until that age, decreasing with disuse of the reflex and returning to normal with use of the reflex. Loss of acuity in amblyopia is thought to be a P function. M dysfunction is seen more commonly with strabismus and may contribute to occlusion nystagmus and asymmetric horizontal smooth pursuit as found in infantile esotropia (congenital esotropia).

By definition, binocular vision is vision in which the images from both eyes are used together. There are several facets of binocular function, only one of which is stereopsis. In strabismology, binocularity has been defined in different terminology at different times. Worth used the term “fusion” for what is now called single binocular vision. He subsequently introduced the terms first-, second-, and third-degree fusion. He described first-degree fusion as a simultaneous awareness of dissimilar targets presented in a haploscopic device. For example, targets such as a fish presented to one eye and a bowl to the other was experienced as a single image of a fish within a bowl.

Second-degree fusion was the unifying of the similar portions of the two targets, each target having one minor dissimilarity (e.g., a flying insect with identical body but two sets of wings in disparate positions. The body is fusible, but the wings are simultaneously perceived. The similar portions of each slide incite a motor response allowing fusional vergence amplitudes to be recorded. In simultaneous awareness of dissimilar images (fish and bowl) on each retina, there is no capability of evoking a motor response.

Third-degree fusion, as described by Worth, was the integrating of similar but slightly disparate targets evoking a perception of stereopsis. For example, haploscopic targets of a swing with the frame portion of the swing presented to each eye as a nondisparate image, but the rope and the swing board suspended from the swing frame presenting either bitemporal or binasal retinal image disparity, make the rope and swing board appear swung toward or away from the swing frame when viewed in the haploscope. Polaroid glasses with contour stereopsis targets and red-green glasses with separate red-green targets have also been used to determine stereopsis.

Worth did not imply that all three degrees of fusion invariably occurred at the same time, but rather the following:

Gradually, Worth’s terminology was altered: The term fusion was replaced with the term single binocular vision. Worth’s first-degree fusion became known as simultaneous perception.

Second-degree fusion was simply termed fusion. Third-degree fusion was changed to stereopsis.

Depth perception and stereopsis are not the same thing. There are many monocular clues to depth. Stereopsis requires the integration of the slightly disparate images based on the spatial separation of the two eyes. Hering called this single visual direction orientation of the two eyes together “cyclopean.” The term was used differently by Julesz,¹ who invented the computer-generated random dot stereogram, which produces a complete dissociation between monocularly perceived and binocularly perceived patterns. When the dot pattern presented to one eye is completely random segments may be shifted, but the result is another random pattern not discernable monocularly. The same horizontal shift presented separately to each eye (dichoptically) creates a binocular disparity, in turn creating a stereoscopic shape invisible to either eye alone.

Lang stereo cards are another random dot clinical tool to determine binocularity. The shapes on the card are invisible when the card is held vertically or viewed monocularly. The objects project toward the observer when the card is held horizontally but away from the observer when turned 180°. This is because the dot shift in the one case is to the right retinal half and, in the other, to the left retinal half. This experience of appearing closer in one setting and farther away when the target is rotated 180° shows that the perception is truly stereoscopic without other clues.

Such random dot testing differs in right-left eye dissociation from systems that use red/green or polarized glasses to test for stereopsis. Julesz¹ used his random dot tests to find nondominant brain hemisphere lesions insofar as such patients are unable to perform random dot but continue to have stereo with red/green or polarized image testing. Similarly microtropic strabismus patients may perform to some level on the polarized or red/green but are totally unable to do so with random dot tests.

The Julesz cyclopean performance on random dot testing is anatomically made possible at the level of the visual cortex and beyond. The topographic projection of associated right and left eye retinal receptor fields has limited interactions between lateral geniculate layers containing 1.5 million cells. Within the visual cortex, however, there are now 200 million topographically arranged cells allowing for more complex processing of the retinal lateral geniculate input.

Lateral geniculate cells continue the same center surround and color properties as retinal ganglion cells. At the visual cortex level, there are complex and simple cells that respond to edge orientation, line length, and motion. There is no longer a simple division into excitatory and inhibitory receptive fields. Some neurons are driven only by left eye stimulation, others by right, and some called binocular, by both. Some binocular neurons have exactly corresponding receptive fields whereas others have varying degrees of disparity in all retinal directions. Fusion occurs when the disparity allows integration. Diplopia occurs when the images are disparate beyond that point, falling outside of the Panum fusional area, either too near or too far away.

The Panum fusional area is, therefore, broader than the strictly geometric corresponding retinal points of the empirical horopter. Both fusion and diplopia change with convergence and direction of gaze horizontally and vertically and, therefore, in three dimensions. This visual activity allows for fusion when a person looks vertically although fusional disparity is almost strictly horizontal as evidenced by the lack of stereopsis when the Lang random dot card is held vertically.

The fovea, having the smallest receptor fields, allows for little disparity before diplopia is elicited. The peripheral retina, with larger receptive fields, tolerates more disparity before diplopia is elicited. It is because of separation of the eyes that there are two disparate retinal images fusible within the Panum area that stereopsis occurs. Stereopsis is, therefore, more than just sensory fusion. Motor fusion is a sensory fusion triggered response when for natural (phoria) or induced (prisms) reasons the images fall outside of the sensory fusion region. Fusion is a trainable, only partially involuntary, diplopia avoidance response, and it is not dependent on stereopsis.

Historically, the terminology of Worth and our modern understanding of retina, lateral geniculate, and visual cortex neurophysiology present clinical problems. Many practitioners who were indoctrinated with the haploscopic assessment of single binocular vision continue to think and articulate in terms of Worth’s three levels of fusion rather than the three perceptual phenomena that characterize single binocular vision.

The difficulties created by Worth’s three levels of fusion concept begin at the first level. The simple explanation of simultaneous perception as the ability to simultaneously perceive two dissimilar slides presented to the viewer in a haploscope does not do justice to this complex neurophysiologic component of single binocular vision. Simultaneous perception of dissimilar images in a haploscope is the laboratory equivalent of only one of the two separate visual circumstances occurring in ordinary seeing that demonstrate simultaneous perception. The dissimilar targets in the haploscope simulate the circumstance that occurs in a nonstrabismic person who sees the visual environment at a distance by one eye and the obstruction precluding the same view by the other eye. Being aware of the distinctly different images simultaneously projecting onto the retinas is only one of the visual circumstances that signifies simultaneous perception.