NEURAL ACTIVITY IN THE RETINA 



709 



750r Fusion frequency 

 iriashesf^^, 



100 ~ 



50 



Light intensity 



10 



100 



1000 



10000 



FIG. 16. Double logarithmic plot of fusion frequency of the 

 electroretinogram against stimulus intensity in meter candles. 

 O, cat; C, guinea pig (two animals); •, pigeon. [From 

 Granit (73).] 



20' 



SOcylsec 



FIG. 17. Evcephale isole given 

 curare. Collicular stimulating 

 electrodes on contralateral side 

 at depth H 2 in Horsley-Clarke 

 coordinates. C, controls with test 

 light of 3 lux alone. /, and 18-20 

 show first and three last records 

 of stimulation period 22 sec. in 

 duration, at a rate of 47 per sec. 

 Note, no driving, as shown by 

 isolated shock artifacts; after- 

 wards diminution of spike fre- 

 quency. Sweep interval in sec. 

 [From Granit (74).] 



portional to impulse frequency set up by the indi- 

 vidual flashes just before the moment of fusion when 

 impulse frequency still can be measured. Fusion it- 

 self is defined as the flicker frequency at which effects 

 of individual flashes on the spike frequency are no 

 longer discernible. Flicker and fusion in electrical 

 records has recenth' been discussed by Granit (73). 



CENTRIFUG.AL CONTROL 



The inner plexiform layer, which forms a network 

 of dendrites between ganglion cells and bipolar cells 

 densely interspersed with amacrine cells, also re- 

 ceives the terminals of the centrifugal fibers (.see fig. 

 2). Their central station in the brain is unknown. 

 However, experiments have shown that it is possible 

 to obtain diff'erent kinds of centrifugal effects on the 

 ganglion cell discharge (44, 74). These are partly 

 excitatory, partly inhibitory but quite often mixed, 

 excitation followed by inhibition, and generally re- 

 quire an array of antidromic stimuli to the optic 

 nerve before a definite effect is noticed. This is not 

 surprising. We are best informed about centrifugal 

 effects from the brainstem to the muscle spindles (77) 

 through the so-called gamma neurons and these too 

 mostly require iterative stimulation. Similarly the 

 suppression of the cochlear nerve discharge (58) by 



stimulation of the centrifugal olivocochlear bundle is 

 fully developed only after it has been stimulated for 

 about half a second at the optimal rate of 100 per sec. 

 The effect on the retina is very .similar independently 

 of whether the site of stimulation is the optic nerve 

 where it is spread out in the pretectum or the brain- 

 stem reticular substance. In the former case there is, 

 of course, also driving of the ganglion cells by anti- 

 dromic stimulation which seems to facilitate excita- 

 tory components on the driven cells. On the other 

 hand, from lower portions of the optic nerve inhibi- 

 tory effects are quite common. When the effect is 

 excitation, it tends to be an increase of level of excita- 

 bility so that a greater number of impulses are dis- 

 charged between 'on' and 'off. The on-off' differ- 

 entials tend to disappear in this general outburst. 

 Again, when the final result is suppression, the whole 

 effect of light is suppressed. An inhibitory effect from 

 the brainstem reticular formation is shown in figure 

 17. High-frequency on- or off-bursts cannot be much 

 altered by centrifugal stimulation. 



Recently further work by Dodt (44) on the rabbit 

 eye has led to the actual demonstration of a centrifu- 

 gal spike picked up in the retina. This spike is from 

 8 to 20 msec, delayed with respect to antidromic im- 

 pulses recorded from the ganglion cells. These also 

 are positive-negative and much larger than the cen- 

 trifugal spikes which are purely negative as if they 

 started below the recording electrode itself. It is 



