NEURAL ACTIVITY IN THE RETINA 



705 



above. In most eyes there is some spontaneous firing, 

 generally greater in the scotopic state. 



The salient point with regard to the general prob- 

 lem of response types is, as shown by Granit (71, 72) 

 and Kuffler (100), that the retina contains two an- 

 tagonistic systems, the on-system and the off-system, 

 which, when made to clash on to the same ganglion 

 cell, are mutually exclusive. One system is excited by 

 light (the on-system), the other is inhibited by light 

 (the off-system) and the latter behaves as if the longer, 

 within limits, the duration during which inhibition is 

 piled up, the more did this favor the subsequent 

 off-discharge. Thus, during the time the off-discharge 

 is inhibited by light something happens that makes 

 it prone to respond when ultimately light is exchanged 

 for darkness. Short exposures tend to give very brief 

 off-effects. When a definite off-effect is seen in the 

 ERG, e.g. in frogs, it also behaves similarly. 



The anatomical convergence means that each 

 fiber has a receptive field, first measured by Adrian 

 & Matthews (3, 4, 5) and shown to be of the order of 

 I mm in the eye of the conger eel, then more pre- 

 cisely with the single-fiber technique by Hartline 

 (87). Figure 1 3 shows the exploring spot used by 

 him and the field sizes obtained in the frog eye when 

 stimulus strength was varied. Just as convergence 

 varies from fiber to fiber, so do the dimensions of the 

 receptive fields. In cats Kuffler (100) found them 

 beautifully organized so that sometimes the on-, 

 sometimes the off-effect occupied the center, the 

 opposite response then occupying its periphcrv and 

 on-off-responses occurring between the two. This 

 provided Kuffler with a good opportunity to make 

 on- and off-spots of the receptive field clash in \arious 

 combinations and thus elegantly to demonstrate the 

 antagonism between the two systems relative to the 

 ganglion discharge. A very complete discussion of 

 receptive fields and on-off systems has been given by 

 Granit (73). Both principles of organization recur in 

 the central structures to which receptors in other 

 sense organs project. 



Why then is the lining of receptors inside the e)e 

 connected to an intricate nervous center just behind 

 it, while other receptor systems mostlv have their 

 first neural organization at the spinal cord level? 

 Apparently receptors cannot do much by themselves; 

 their mes.sagcs must be organized somehow for dis- 

 crimination and integration; and, since the little 

 brain behind the rods and cones moves with the eye, 

 it can because of its place in the retina aid better in 

 the interpretation of the ever-changing boundaries of 



FIG. 13. Chait of the receptive field of a single optic nerve 

 fiber of the frog. Each fine encloses a retinal region within which 

 the exploring spot light (relative size shown above left) — the log 

 of the intensity is given on the line — produced a response from 

 the fiber. On each line the indicated intensity was the threshold; 

 the set of curves constitutes a contour map of the distribution 

 of the retinal sensitivity to light with reference to this par- 

 ticular fiber. [From Hartline (89).] 



light, darkness and color out of which the visual 

 world is synthesized. The eye, as stated above, is 

 never still and, if by artificial means the image is kept 

 stationary (39, 125), it tends to fade out quickly, 

 as if it needed the on-off differentials sharpening up 

 contours during oscillations. Movement of an object 

 or a point across the retina will light up a trail of on- 

 off sparks, as well shown by Barlow (15, 16) in frog 

 experiments set up to illustrate the biological sig- 

 nificance of such factors for perceprion. Apparently 

 also, it is necessary for this highly developed organ 

 not to be forced to one single mode of working in 

 coping with a range of illumination from dusk to 

 bright sunlight. It has been shown that the receptive 

 fields of the retina vary in width with state of adapta 

 tion (17). 



From what has l)een stated it is clear that at least 

 within a receptive field interaction can occur, as 

 first shown by Adrian & Matthews (5) and then 

 studied in detail by Hartline (87, 88). When spike 

 frequency or latency is used as an indicator, area and 

 intensity are found to he interchangeable within the 

 field, whether from excitation or inhibition being 

 unknown. These facts provide a likely explanation 

 of the many old psychophysical observations on the 

 interchangeability of area and intensity in vision. 

 From what has been stated it will be realized that 

 on-off interaction adds to the complexity so that over 



