The great bulk of the connections between the two sides of the cerebral mantle are made by the interhemispheric commissure (the point of union between the two hemispheres of the cerebrum) called the corpus callosum, which is made up of neurons and their axons and dendrites that make synapses with cortical neurons on symmetrically related points of the hemispheres. Thus, electrical stimulation of a point on one hemisphere usually gives rise to a response on a symmetrically related point on the other, by virtue of these callosal connections. The striate area is an exception, however, and it is by virtue of the connections of the striate neurons with the area 18 neurons that this integration occurs, the two areas 18 on opposite hemispheres being linked by the corpus callosum.

Stereopsis in the midline

Usually stereopsis, or perception of depth, is possible by the use of a single hemisphere because the images of the same object formed by right and left eyes are projected to the same hemisphere; however, if the gaze is fixed on a distant point and a pin is placed in line with this but closer to the observer, a stereoscopic perception of the distant point and the pin can be achieved by the fusion of disparate images of the pin, but the images of the pin actually fall on opposite retinal halves, so that this fusion must be brought about by way of the corpus callosum.

Callosal transfer

In experimental animals it is possible, by section of the chiasma, to ensure that visual impulses from one eye pass only to one hemisphere. If this is done, an animal trained to respond to a given pattern and permitted to use only one eye during the training is just as efficient, when fully trained, in making the discrimination with the other eye. There has thus been a callosal transfer of the learning so that the hemisphere that was not directly involved in the learning process can react as well as that directly involved. If the corpus callosum is also sectioned, this transfer is impossible, so that the animal, trained with one eye, must be trained again if it is to carry out the task with the other eye only.

Superior colliculi

The visual pathway so far described is called the geniculostriate pathway, and in humans it may well be the exclusive one from a functional aspect because lesions in this pathway lead to blindness. Nevertheless, many of the optic tract fibres, even in humans, relay in the superior colliculi, a paired formation on the roof of the midbrain. From the colliculi there is no relay to the cortex, so that any responses brought about by this pathway do not involve the cortex. In humans, as has been said, lesions in the striate area, which would of course leave the collicular centres intact, cause blindness, so that the visual fibres in these centres serve no obvious function. In lower animals, including primates, removal of the striate areas does not cause complete blindness; in fact, it is often difficult to determine any visual impairment from a study of the behaviour of the animals.

Thus, in reptiles and birds, vision is barely affected, so that a pigeon that has been subjected to the operation can fly and avoid obstacles as well as a normal one. In rodents, such as the rabbit, removal of the occipital lobes causes some impairment of vision, but the animal can perform such feats as avoiding obstacles when running and recognizing food by sight. In the monkey, the effects are more serious, but the animal can be trained to discriminate lights of different intensity and even the shapes of objects, provided that these are kept in continual motion. It seems likely, then, that it is the visual pathway through the colliculi that permits the use of the eyes in the absence of visual cortex, although the connections of the optic tract fibres with the pulvinar of the thalamus (an area in the diencephalon), established in some animals, may well permit the use of regions of the cortex other than those denoted as visual.

Relative positions of objects

The perception of the positions of objects in relation to each other is essentially a geometrical problem. Take, for the present, the perception of these relationships by one eye, monocular perception: a group of objects produces images on the retina in a certain fixed geometrical relationship. For the perception of the fact that C is to the left of D, for example, that D is to the left of E, and so on, it is necessary that the incidence of images at c, d, and e on the retina be interpreted in a similar but, of course, inverted geometrical relationship. The neural requirements for this interpretation are (1) that the retina be built up of elements that behave as units throughout their conducting system to the visual cortex, and (2) that the retinal elements have “local signs.” The local sign could represent an innate disposition or could result from experience—the association of the direction of objects in space, as determined by such evidence as that provided by touch, with the retinal pattern of stimulation. In neurophysiological terms, the retinal elements are said to be connected to cortical cells, each being specific for a given element, so that when a given cortical cell is excited the awareness is of a specific local sign. Studies of the projection of the retina on the cerebral cortex have confirmed this.

The retinal stimuli at c, d, and e in are appreciated as objects outside the eye, the retina is said to be projected into space, and the field of vision is thus the projection of the retina through the nodal point. It will be seen that the geometrical relationship between objects and retinal stimuli is reversed; in the retina c is to the right of d, and so on.

Position in relation to observer

The recognition of the directions of objects in relation to the observer is more complex. If the eye is turned to the left, the image of C falls on the retinal point d, so that if d were always projected into the same direction in space, C would appear to be in D’s place. In practice, one knows that C is perceived as fixed in space in spite of the movements of the eye; hence, the direction of projection of a retinal point is constantly modified to take into account movements of the eye; this may be called psychological compensation. It will be seen that correct projection is achieved by projecting the stimulated retinal point through the nodal point of the eye. Movements of the eye caused by movements of the head must be similarly compensated. As a result, any point in space remains fixed in spite of movements of the eye and head. Given this system of compensated projection, the recognition of direction in relation to the individual is now feasible. D may be said to be due north or, more vaguely, “over there”; when the head is turned, since D is perceived to be in the same place, it is still due north or “over there.” In some circumstances, the human subject makes an error in projecting his retinal image, so that the object giving rise to the image appears to be in a different place from its true one; the image is said to be falsely projected. If the eye is moved passively, for example, by pulling on the conjunctiva with forceps, the subject has the impression that objects in the outside world are moving in a direction opposite to that of the eye.

The apparent movement of an afterimage, when the eye moves, is an excellent illustration of psychological compensation. A retinal stimulus, being normally projected through the nodal point, is projected into different points in space as the eye moves; an afterimage can be considered to be the manifestation of a continued retinal impulse, and its projection changes as the eye moves. The afterimage thus appears to move in the same direction as that of the movement of the eye. Whether the drift of an afterimage across the field of view is entirely due to eye movements is difficult to say. One certainly has the impression that the eye is chasing the afterimage.

The directions of lines

So far, consideration has been given to the problem of estimating the positions of points in relation to each other and to the percipient. The estimate of the directions of lines involves no really new principles, since, if two points, A and B, are exactly localized, the direction of the line AB can be appreciated. As will be seen, the organization of the neural connections of the retina and higher visual pathway is such as to favour the accurate recognition of direction; for the moment, the question of the maintenance of a frame of reference must be considered, in the sense that a map has vertical and horizontal lines with which to compare other directions. In fact, the vertical and horizontal meridians of the retina seem to be specialized as frames of reference; the accuracy with which a human subject can estimate whether a line is vertical or horizontal is very great.

An important point in this connection is that of the effects of eye movements on interpretation of the directions of lines because, when the eye moves to positions different from the primary straight-ahead position, the images of vertical lines will not necessarily fall on its vertical meridian. This can be due to an actual torsion of the eye about its anteroposterior (fore and aft) axis or to distortion of the retinal image. This means, then, that the line on the retina that corresponds to verticality in one position of the eye does not correspond to verticality in another, so that, once again, the space representation centre must take account not only of the retinal elements that have been stimulated but also of the corollary motor discharge.

Comparison of lengths

The influence of the movements of the eyes in the estimation of length was emphasized by Helmholtz. An accurate comparison of the lengths of two parallel lines AB and CD can be made, whereas if an attempt is made to compare the nonparallel lines A′B′ and C′D′, quite large errors occur. According to Helmholtz, the eye fixates first the point A, and the line AB falls along a definite row of photoreceptors, thereby indicating its length. The eye is now moved to fixate C, and if the image of CD falls along the same set of photoreceptors the length of CD is said to be the same as that of AB. Such a movement of the eye is not feasible with lines that are not parallel. Similarly, the parallelism, or otherwise, of pairs of lines can be perceived accurately because on moving the eye over the lines the distance between them must remain the same.

Fairly accurate estimates of relative size may be made, nevertheless, without movements of the eyes. If two equal lines are observed simultaneously, the one with direct fixation and the other with peripheral vision, their images fall, of course, on different parts of the retina; if the images were equally long it could be stated that a certain length of stimulated retina was interpreted as a certain length of line in space. It is probable that this is roughly the basis on which rapid estimates of length depend, although there are such complications as the fact that the retina is curved so that lines of equal length in different parts of the retina do not produce images of equal length on the retina.

Optical illusions

Many instances have been cited of well-defined and consistent errors in visual estimates under special conditions. There is probably no single factor by which the errors can be explained, but the tendency for distinctly perceptible differences to appear larger than those more vaguely perceived is important.