NEURAL ACTIVITY IN THE RETINA 



703 



cells), it will be necessary to use nonpolarizable pene- 

 trating microelectrodes. Such work was initiated by 

 Tomita (141, 142). Instead of a review of arguments 

 from general electroretinographic work for which the 

 author's summary (69) may be consulted, a brief dis- 

 cussion of the results obtained by such niicromcthods 

 is given. 



Tomita"s work was criticized and ignored by Otto- 

 son & Svaetichin (i 16) on the grounds that his pene- 

 trating microelectrodes were held to be too coarse. 

 Tomita has since (142) repeated his experiments with 

 the Ling-Gerard type of microcapillaries, used by 

 Svaetichin, and confirmed his previous results. At the 

 same time Brindley (28, 29, 30) has also published a 

 careful study using the same technique. Ottoson & 

 Svaetichin's conclusions as far as fundamental ques- 

 tions are concerned were: a) that slow potentials of 

 the ERG type are obtained in the receptors only; 

 this has since been definitely refuted by Brindley (30) 

 and by Tomita & Torihama (142), who found large 

 potential changes mirroring the ERG in the bipolar 

 layer; b) that rods respond only with positive, and 

 cones only with negative retinograms [cf. also Svaeti- 

 chin (137)], disproved by Granit (73) as well as by 

 Forbes ft al. (55) and Brindley (29); c) that the resting 

 potential of the retina is a receptor potential (115), 

 since refuted by Brindley (28); tt) that, from the fact 

 that cocaine slowly attacks the ERG but immediately 

 stops the discharge of impulses through the optic 

 nerve, it is possible to conclude that the ERG is a pure 

 receptor potential [from Kiihne & Steiner (102) 

 onwards the number of agents capable of blocking the 

 impulse discharge without much effect on the ERG 

 has been slowly multiplying, yet without suggesting 

 to anyone such far-reaching conclusions] ; e) that from 

 the size of sudden potential gradients of the order of 

 20 to 30 mv within the fish retina (approached from 

 the receptor end) it is po.ssible to assume that the 

 electrode recorded intracellularly from single cones. 

 Now Brindley (28) has shown that there are charac- 

 teristic steps in the radial resistance to a penetrating 

 microelectrode, the largest one across the external 

 limiting membrane (see below), and extracellular 

 spike potentials of the order of 40 to 60 mv have been 

 recorded by Granit & Phillips (79) by the same tech- 

 nique at the surface of the cerebellar Purkinje cells. 

 Furthermore, an extracellular retinal microelectrode 

 has been shown by Brindley (30) and Tomita & Tori- 

 hama (142) to pick up its response from very distant 

 illuminated regions. This is, of course, what one must 

 expect. Light intensities and techniques of illumina- 

 tion are hardly ever mentioned in the papers by 



Svaetichin & Ottoson but this, in itself, suggests that 

 the whole retina or a large fraction of it was illumi- 

 nated. Focal microillumination would be needed for 

 localized responses and there is in its favor the further 

 advantage that absence of a response within the bi- 

 polar layer shows whether damage has occurred (30, 

 142). Such damage probably explains why Ottoson 

 & Svaetichin have missed the response inside the 

 retina. In fact, Brindley (30) describes two types of 

 responses in excised frog eyes of which the .second 

 agrees with the pictures of Svaetichin. This type 

 Brindley holds to be characteristic of local damage 

 because it alone is seen when the focal intraretinal 

 response is absent. These remarks may suffice to show 

 why it is felt that Ottoson & Svaetichin have under- 

 rated the analytical difficulties of the work they set 

 out to do. For this reason individual good observa- 

 tions in their work, perhaps unjustly, lose their sig- 

 nificance to a reviewer and can only be rescued by 

 those who have undertaken microelectrode work with 

 the same structure and thus can evaluate them criti- 

 cally against a background of specific experience. 



Apparently retinal neurons do not differ from other 

 neurons, all of them producing potential changes. 

 Both Brindley and Tomita have arranged their ex- 

 periments for comparison of precise focal microil- 

 lumination around the microelectrode tip with illu- 

 mination of larger areas. Unless an electrode within 

 the bipolar area responds to focal illumination with a 

 reponse of the intraretinal type, the region around 

 this electrode is not likely to be in a normal state. 

 Brindley suggests that this condition is due to dam- 

 age of the external limiting membrane. The inside 

 focal response has a maximum among the bipolar 

 cells at a depth of 1 00 to 1 40 yii from the ganglion side. 

 The large ERG elicited by general illumination is 

 always obtained and tends to be positive. The two 

 authors differ in that Tomita's response to diffuse 

 light reverses sign within the inner nuclear layer, 

 going from positive at the ganglion side to negative 

 at the receptor side of the retina, while Brindley finds 

 more variability in this regard. Now, is the focal 

 response inside the retina identical with the ERG? 

 Obviously this response is physiologically important 

 but both Brindley and Tomita argue against identi- 

 fication. In the reviewer's opinion it is impossible at 

 the present stage of our knowledge to be certain as to 

 whether or not bipolar cells contribute to the ERG. 

 Tomita, for instance, finds no focal response within 

 the receptor layer. This is no crucial objection to a 

 localization in the receptors, nevertheless it is a fact 

 to be considered. Brindley holds the focal response to 



