The Jigsaw Puzzle of Visual Neurophysiology

Elwin Marg

Visual Science, Proceedings of the 1968 International Symposium, J.R. Pierce and J.R. Levene, eds., Indiana University Press, Bloomington, 31-39, 1971.

Visual neurophysiology may seem narrowly specialized compared with visual science in general, but it has a simple general goal: to learn how we see by direct investigation of the very nerve cells involved in vision.

Historically, visual neurophysiology might be said to have started when du Bois Reymond (1849) first measured electrical potentials from the eye and optic nerve of the tench, a European freshwater fish. In a more meaningful and modern sense, the development of our current and rapidly growing understanding of vertebrate visuo-sensory processes can be traced to Hartline (1938). By teasing apart the optic nerve fibers of a frog eye he was able to record from single functional fibers, before the era of microelectrodes. Hartline found that spots of light presented in a certain area of the visual field influenced the firing of the fiber, and this area he termed a "receptive field." Granit and Svaetichen (1939) confirmed Hartline's results with a microelectrode (which was large by today's standards) placed directly on the retina of the open eye.

Thirteen years later Kuffler (1953), working with the newly developed micropipette electrode on the closed eyes of cats, was able to plot the form of the receptive fields of retinal ganglion cells. He found the receptive field topography consisted of a center (disk) and a surround (annulus), one causing the cell to fire more (on-effect which is an excitatory phenomenon) when stimulated by light, and its partner causing it to fire less (off-effect which is an inhibitory phenomenon). In other words, there were on centers with off surrounds or the reverse: off centers with on surrounds. A similar organization of receptive fields of the frog retina was found by Barlow (1953). However, more complex receptive field properties were found in the frog by Lettvin et al. (1959), properties not found in mammals which appear simpler by comparison.

During the last fifteen years, scores of papers have been published on visual receptive fields from the cells at all levels of the system from the ganglion cell layer of the retina on up into the brain. A list of the important papers would be too lengthy to include here and they can be found in various reference and text books (for example, Granit, 1947, 1955; Graham et al., 1965). The most complete and best known investigations are those of Hubel and Wiesel (1960-1963, 1965; Wiesel and Hubel, 1966), who first worked with cats and later with monkeys. They confirmed and extended Kuffler's retinal data, and then went on to make extensive studies of the lateral geniculate body and the visual cortex.

Hubel and Wiesel's results showed that the receptive fields of the ceils of the lateral geniculate body resembled in form those of the retinal ganglion cells. The visual cortex gave a different picture. Here the fields were rectangular or bar-shaped and sensitive to the orientation of a bar-shaped target, for if the stimulus were presented at right angles to the receptive field no response would be elicited. More complex receptive fields were found in which the bar would be effective over a relatively large area or only when one or both ends were limited in length or the target fulfilled other more complex requirements.

The disks, annul), and bars can be constructed into an organization of how we see, a sort of jigsaw puzzle of visual neurophysiology. The pieces may vary in size but not in shape beyond the limitations already mentioned. Unlike the usual jigsaw puzzle, the pieces in the eye and brain are numbered in the millions and at each level (or order of neuron) the picture is different. Unlike a wooden jigsaw puzzle, the visual one has a high degree of overlapping and superposition of the parts. Colors must be added as well as movement and its direction, and at the cortical level, binocularity in animals with overlapping visual fields. Animal experiments will continue to bring the puzzle into better fit and focus. But can it be finally solved without direct information from the human visual system?

It is necessary to get much of our neurophysiological information from animals (under conditions free of pain) because they do not have to volunteer, their physical integrity can be compromised, their lives can be risked and when necessary they can be sacrificed. If ethics clearly allowed, it would be preferable to obtain data from human beings because most of us are actually interested in human function rather than in, for example, cat function. There are, obviously, fundamental species differences. Furthermore there are important physiological functions (equivalent to psychological functions in the brain) which can be studied only in a drug-free, cooperative, and communicating animal, man.

Neurosurgeons must probe the brains of people for the diagnosis and therapy of certain diseases such as medically intractable epilepsy and Parkinson's disease. The consent of the patient undergoing such procedures makes the collection of scientific data as a by-product of the treatment both highly ethical and an opportunity not to be lost in the extension of our knowledge of mankind for mankind.

With this in mind, Dr. John E. Adams, professor of neurological surgery at the University of California, and I developed rugged microelectrodes which could be used for implantation as part of a program of depth elcctrography to aid medically intractable epileptics (Marg, 1964; Marg and Adams, 1967). The electrodes, made of straight tungsten wire of 50 microns diameter, are sharpened to a one-micron tip with a short taper and insulated with IS separately baked coats of an insulating varnish, Isonel 31. A bundle of these electrodes may be used to increase the probability of picking up a unit response, which is further increased by employing an implantable micromanipulator (Fig. 1) controllable from a flexible shaft through the head bandages. The electrodes were connected to the usual follower circuit, amplifiers, loudspeaker, and cathode ray oscilloscope with camera. For use just during surgery deep in the brain at the thalamic level, microelectrodes have been used having a shank diameter of a quarter or a half millimeter (Marg and Dierssen, 1965, 1966).

Figure 1. Micromanipulator for implantation in burr hole. Below: the bundle of microelectrodes project down from a teflon diaphragm on the cylinder which fits in the burr hole. The flange is screwed to the bone and the gear box rests above the scalp. The electrode wires are led off in the small plastic tube through the head bandages. The large plastic tube contains a flexible shaft which may be turned from outside the bandages, lowering the microelectrode bundle 20u per turn for an excursion of up to 3mm. Above: the worm-gear box, lid, diaphragm ring and screws. All metal parts are made of type 316 low-carbon stainless steel.

Figure 2. The supine patient (whose head is just visible at the bottom foreground) is fixating a black dot on the white crdboard sheet 1m from his eyes. A bar-shaped target on a wand is being manipulated to find the receptive field.

Figure 3. The response of this unit gave the receptive field labeled E in Figure 4.

The patient viewed a large piece of cardboard one meter away with an ink dot or grain-of-wheat lamp for fixation (Fig. 2) (Marg, Adams, and Rutkin, 1968). Hand-held wands with disks or bars of various sizes provided the means of plotting the receptive field as the firing of the cell was heard over the loudspeaker and seen on the cathode ray oscilloscope. The receptive field was outlined in light pencil marks on the cardboard.

Cells were found firing spontaneously in an irregular or "bursty" manner (Fig.3). Many of them appeared to be uninfluenced by any kind of visual stimulation we could provide such as random disks and bars, gratings, fingers, etc., or the patient just gazing around the room. Some units did exhibit a visual response to certain targets, and showed inhibition of the spontaneous activity when the eyelids were closed. In some of these we could find the position in the visual field of the receptive field and plot it. The electrode would usually maintain the response for the recording session of several hours and sometimes the same unit would be found unchanged the next day.

First we noted whether the receptive field came from one eye or both and then we looked for any dominance of one over the other in the responsiveness of the cell. We plotted the receptive field, tried various colors and other tests such as stimulation of other sensory modalities, auditory feedback, or voluntary mental image effects. It was soon evident that the receptive field was always the same whether the target was black on white or white on black, or red, yellow, green, or blue on any contrasting back ground. Moreover, all attempts by the patient mentally to influence the firing of a visual cortical cell were fruitless. Hue did not seem to be represented in the cells we recorded, nor did mental imagery or other (nonvisual) sensory modalities.

All the receptive fields were entirely excitatory as far as we could tell, that is, an increase of firing superimposed on the irregular spontaneous activity was evident whenever a target was brought within the receptive field. When the stimulus exactly matched the field, the response was strong. No inhibitory area surrounding the excitatory receptive field was observed but if it were subtle in effect it could have been missed by the simple method of plotting.

Some of the receptive fields could be plotted repeatedly without any apparent decrement of the response. The responses of other fields faded away until only the irregular spontaneous activity remained. Later they might partially recover, only to fade more quickly the second time. This phenomenon has been called progressive attenuation (Hubel and Wiesel, 1965) or habituation (Horn and Hill, 1966). Different visual stimuli or stimuli of other modalities (such as handclapping) did not cause the response to the original stimulus to return, demonstrating a lack of dishabituation (Horn and Hill, 1966; Horn, 1967).

Only a few of the units observed were plotted, primarily because of the limitations of time, and these are shown in Figure 4. One disk-shaped field (E) was binocular, indicating that the unit being recorded was cortical and not directly from an incoming lateral geniculate fiber, if the human lateral geniculate nucleus is similar in organization to that of the cat and monkey. In one unit (D), a horizontal bar target was effective over a 20 cm vertical range (the limits shown by the dashed lines), similar to the "complex" units found in the cat cortex. In these complex units a distinction must be made between the receptive field (the large area between the dashed lines) and the shape of the target which elicits the response (the bar). All of the fields were from units in the contralateral visual hemisphere except a near-vertical thin bar-shaped field whose unit was in the ipsilateral visual cortex. The horizontal bar-shaped fields seemed to give a greater response to movement relative to steady presentation of the targets than did the others. There was some indication of a shift of size and position in some of the units, which could be characterized as plasticity. This was not caused by poor fixation. Generally fixation was steady during the plotting of the receptive fields and in most fields the edges were sharp and did not wander.

By the time the receptive field of one binocular unit (F) was plotted with one eye, it had habituated and no response could be obtained through the other eye. At the next recording session the same unit was still available and the same receptive field could now be obtained from the second eye. This finding would place the habituation phenomenon at the cortical level.

Figure 4. Receptive fields from units in the human visual cortex. "p" is the fixation point at 1m from the eyes.

As mentioned earlier, it was possible to inhibit the spontaneious activity of visually responding cells by closing the lids. No other means of inhibiting the cells was found. Occluding the open eyes with a card or reducing the room illumination did not inhibit the cell.

By merely changing the distance of the cardboard with its fixation point, one can plot the receptive field at various fixation distances, if time allow. It is possible to show if the receptive field or "visual grain" changes its angular size with different fixation distance. This presumably would be directly concerned with the basic mechanism of size constancy or scaling (Richards, 1967, 1968).

These preliminary results demonstrate two points: first, that there are species differences in the visual receptive fields between man and experimental animals that can be discovered only by obtaining data directly from man, under, of course, completely ethical conditions; and second, that there are aspects of the function of the visual system that can be elucidated only by this kind of investigation in conscious volunteer human patients. In fact these points can be extended to the functions of the brain generally, sensory, motor, and integrative.

In the span of thirty years, visual neurophysiology has progressed from the first glimpse of visual receptive fields in frogs to the first glimpse of them in man. They are the jigsaw puzzle of vision. It seems reasonable and safe to predict that, as we put together the pieces of the visual receptive fields puzzle in man, it will form a picture which will allow us to understand how we see.

References

Barlow, H. B. (1953). Summation and inhibition in the frog's retina. J. Physiol. 119, 69-88.

du Bois, Reymond E. (1849). Untersuchungen fiber thierische Elektricitat. G. Reimer, Berlin.

Graham, C. H., N. R. Bartlett, J. L. Brown, Y. Hsia, C. G. Mueller, and L. A. Riggs (1965). Vision and Visual Perception. John Wiley and Son, New York.

Granit, R. (1947). Sensory Mechanisms of the Retina. Oxford University Press, London.

Granit, R. (1955). Receptors and Sensory Perception. Yale University Press, New Haven, Connecticut.

Granit, R., and G. Svaetichin (1939). Principles and techniques of the electrophysiological analysis of color reception with the aid of microelectrodes. Upsala Lakaref. forh. 65, 161 -177.

Hartline, H. K. (1938). The response of single optic nerve fibers of the vertebrate eye to illumination of the retina. Am. J. Physiol. 121, 400-415.

Horn, G. (1967). Neuronal mechanisms of habituation. Nature 215, 707-711.

Horn, G., and R. M. Hill (1966). Responsiveness to sensory stimulation of units in the superior colliculus and subjacent tectotegmental regions of the rabbit. Exper. Neurol. 14, 199-223.

Hubel, D. H., and T. N. Wiesel (1960). Receptive fields of optic nerve fibers in the spider monkey. J. Physiol. 154, 572-580.

Hubel, D. H., and T. Wiesel (1961). Integrative action in the cat's lateral geniculate body. J. Physiol. 155, 385-398.

Hubel, D. H., and T. Wiesel (1962). Receptive fields, binocular interaction, and functional architecture in the cat's visual cortex. J. Physiol. 160, 106-154.

Hubel, D. H., and T. Wiesel (1963). Shape and arrangement of columns in the cat's striate cortex. J. Physiol. 165, 559-568.

Hubel, D. H., and T. Wiesel (1965). Receptive fields and functional architecture in two non-striate visual areas (18 and 19) of the cat. J. Neurophysiol. 28, 229-289.

Kuffler, S. W. (1953). Discharge patterns and functional organization of mammalian retina. J. Neurophysiol. 16, 37-68.

Lettvin, J. Y., H. R. Maturana, W. S. McCullock, and W. H. Pitts (1959). What the frog's eye tells the frog's brain. Proc. Inst. Radio Engr. 47, 1940-1951.

Marg, E. (1964). A rugged, reliable, and sterilizable microelectrode for recording single units from the brain. Nature 202, 601-603.

Marg, E., and J. E. Adams (1967). Indwelling multiple microelectrodes in the brain. Electroencephal. and Clin. Neurophysiol. 23, 277-280.

Marg, E., J. E. Adams, and B. Rutkin (1968). Receptive fields of 'ceils in the human visual cortex. Experientia 24, 348-350.

Marg, E., and G. Dierssen (1965). Reported visual percepts from stimulation of the human brain with microelectrodes during therapeutic surgery. Confin. Neurol. 26 (2), 57-75.

Marg, E., and G. Dierssen (1966). Somatosensory reports from electrical stimulation of the brain during therapeutic surgery. Nature 212, 188-189.

Richards, W. (1967). Apparent modifiability of receptive fields during accommodation and convergence and a model for size constancy. Neuropsychol. 5, 63-72.

Richards, W. (1968). Spatial remapping in the primate visual system. Kybernetic 4, 146-156.

Wiesel, T. N., and D. H. Hubel (1966). Spatial and chromatic interactions in the lateral geniculate body of the rhesus monkey. J. Neurophysiol. 29, 1115-1156.

I thank Miss Nancy Kuwada for her technical assistance. Supported by a grant (tom the National Science Foundation and a Research Professorship from the Miller Institute for Basic Research in Science of the University of California, Berkeley.