Eye, Brain, and Vision
Home
Book
Illusions
Biography
Publications
David Hubel's
Home
Book
Illusions
Biography
Publications
HMS
 
     
                                                                                    
   
 
The receptive fields of two neighboring retinal ganglion cells will usually overlap.
The smallest spot of light we can shine on the retina is likely to influence hundreds of ganglion cells, some off center and some on center. The spot will fall on the centers of some receptive fields and on the surrounds of others.



Before I go on to describe the receptors and other retinal cells, I want to make three additional comments about receptive fields. The first is a general comment about receptive fields as a concept; the other two comments are specifically about the receptive fields of retinal ganglion cells: their overlap and their dimensions.


                                 THE CONCEPT OF                                    A RECEPTIVE FIELD
Narrowly defined, the term receptive field refers simply to the specific receptors that feed into a given cell in the nervous system, with one or more synapses intervening. In this narrower sense, and for vision, it thus refers simply to a region on the retina, but since Kuffler's time and because of his work the term has gradually come to be used in a far broader way. Retinal ganglion cells were historically the first example of cells whose receptive fields had a substructure: stimulating different parts of the receptive fields gave qualitatively different responses, and stimulating a large area resulted in cancellation of the effects of the subdivisions rather than addition. As presently used, receptive field tends to include a description of the substructure, or if you prefer, an account of how you have to stimulate an area to make the cell respond.
When we speak of "mapping out a cell's receptive field", we often mean not simply demarcating its boundaries on the retina or the screen the animal is looking at, but also describing the substructure. As we get deeper into the central nervous system, where receptive fields tend to become more and more complex, we will find that their descriptions become increasingly elaborate.
Receptive-field maps are especially useful because they allow us to predict the behavior of a cell. In the case of retinal ganglion cells, for example, suppose we stimulate an on-center cell with a long, narrow rectangle of light, just wide enough to span the receptive-field center, and long enough to go beyond the whole field, center plus surround. We would predict from the on-center map on page 10 that such a stimulus would evoke a strong response, since it covers all the center and only a small fraction of the antagonistic surround. Furthermore, from the radial symmetry of the map we can predict that the magnitude of the cell's response will be independent of the slit's orientation. Both predictions are confirmed experimentally.


                                THE OVERLAP OF                                    RECEPTIVE FIELDS
My second comment concerns the important question of what a population of cells, such as the output cells of the retina, are doing in response to light. To understand what ganglion cells, or any other cells in a sensory system are doing, we have to go at the problem in two ways. By mapping the receptive field, we ask how we need to stimulate to make one cell respond. But we also want to know how some particular retinal stimulus affects the entire population of ganglion cells. To answer the second question we need to begin by asking what two neighboring ganglion cells, sitting side by side in the retina, have in common.
The description I have given so far of ganglion-cell receptive fields could mislead you into thinking of them as forming a mosaic of nonoverlapping little circles on the retina, like the tiles on a bathroom floor. Neighboring retinal ganglion cells in fact receive their inputs from richly overlapping and usually only slightly different arrays of receptors, as shown in the diagram on this page. This is the equivalent of saying that the receptive fields almost completely overlap.

 

 

 

 

 

 

 

 

You can see why by glancing at the simplified circuit in the diagram above:
the cell colored purple and the one colored blue receive inputs from the overlapping regions, shown in cross section, of the appropriate colors. Because of divergence, in which one cell makes synapses with many others at each stage, one receptor can influence hundreds or thousands of ganglion cells. It will contribute to the receptive-field centers of some cells and to the surrounds of others. It will excite some cells, through their centers if they are on-center cells and through their surrounds if they are off-center cells; and it will similarly inhibit others, through their centers or surrounds. Thus a small spot shining on the retina can stir up a lot of activity, in many cells.


                                 DIMENSIONS OF                                    RECEPTIVE FIELDS
My third comment is an attempt to relate these events in the retina to everyday vision in the outside world. Obviously our vision completely depends on information the brain receives from the eyes; all this information is conveyed to the brain by the axons of retinal ganglion cells. The finer the detail conveyed by each of these fibers, the crisper will be our image of the world.

 

 

 

 


This fineness of detail is best measured not by the overall size of receptive fields, but by the size of the field centers.
We can describe the size of a receptive field in two ways. The more straightforward description is simply its size on the retina. This has the disadvantage of being not very meaningful in the everyday scale of things. Alternatively, you could measure receptive-field size in the outside world, for example, by taking its diameter on a screen that an animal faces, but you would then have to specify how far the screen is from the animal's eyes. The way around this problem is to express receptive-field size as the angle subtended by the receptive field on the screen, at the animal's eye, as shown in the figure on this page.
We calculate this angle in radians simply by dividing the field diameter by the screen distance, but I will use degrees: (radians x 180)/3.14. One millimeter on the human retina corresponds to an angle of about 3.5 degrees. At 54 inches screen distance, i inch on the screen corresponds to 1degree. The moon and sun, seen from the earth, are about the same size, and each subtends one-half a degree.
Receptive fields differ in size from one ganglion cell to the next. In particular, the centers of the receptive fields vary markedly and systematically in size:
they are smallest in the fovea, the central part of the retina, where our visual acuity—our ability to distinguish small objects—is greatest; they get progressively larger the farther out we go, and meanwhile our acuity falls off progressively.
In a monkey the smallest field centers yet measured subtend about 2 minutes of arc, or about 10 micrometers (0.01 millimeters) on the retina. These ganglion cells are in or very close to the fovea. In the fovea, cones have diameters and center-to-center spacing of about 2.5 micrometers, a figure that matches well with our visual acuity, measured in terms of our ability to separate two points as close as 0.5 minutes of arc. A circle 2.5 micrometers in diameter on the retina (subtending 0.5 minutes) corresponds to a quarter viewed from a distance of about 500 feet.

   
 
 
Contents
Next Page
One millimeter on the retina corresponds to 3.5 degrees of visual angle. On a screen 1.5 meters away, i millimeter on the retina thus corresponds to about 3.5 inches, or 89 millimeters.


 
 
 
 
 
As we stimulate a single on-center retinal ganglion cell with ever larger spots, the response becomes more powerful, up to a spot size that depends on the cell—at most about i degree. This is the center size. Further enlargement of the spot causes a decline, because now the spot invades the antagonistic surround. Beyond about 3 degrees there is no further decline, so that 3 degrees represents the total receptive field, center plus surround.



Next Page
Contents
Two neighboring retinal ganglion cells receive input over the direct path from two overlapping groups of receptors. The areas of retina occupied by these receptors make up their receptive-field centers, shown face on by the large overlapping circles.