Luminance Dependence of Neural Components that Underlies the Primate Photopic Electroretinogram

Abstract

purpose. At lower stimulus intensities, the amplitude of the photopic flash ERG b-wave increases with increasing stimulus intensities, but then plateaus and decreases at higher stimulus intensities (the “photopic hill”). The purpose of this study was to determine the mechanism underlying this unusual phenomenon.

methods. Five adult monkeys (Macaca mulatta and M. fascicularis) were studied. Stimuli were obtained from xenon strobe flashes, and the intensity was reduced by neutral-density filters in 0.4-log unit steps. N-methyl-d-aspartic acid and tetrodotoxin citrate (NMDA+TTX) were used to suppress inner retinal activities and l-2 amino-4-phosphonobutyric acid (APB) and cis-2,3 piperidine dicarboxylic acid (PDA) to block postreceptoral ON- and OFF-pathway activities. The postsynaptic ON- and OFF-components were isolated by subtracting the postdrug ERGs from the predrug ERGs.

results. The intensity–response curve of the photopic b-wave obtained after the intravitreal injection of TTX+NMDA had the same shape as a photopic hill, suggesting that the contribution from the inner retinal neurons to the photopic hill is not significant. At low and intermediate intensities, the photopic b-wave was mainly shaped by the overlapping of two positive peaks from the ON- and OFF-components. However, the amplitude of the positive peak from the ON-component became smaller and broader at higher stimulus intensities. In addition, the timing of the positive peak of the OFF-component was gradually delayed with increasing intensities. After APB+PDA, the remaining cone photoreceptor component contributed only to the negative a-wave at all stimulus intensities.

conclusions. The photopic hill in the primate ERG results mainly from two factors: the reduction of the ON-component amplitude at higher intensities and the delay in the positive peak of the OFF-component at higher intensities.

Photopic electroretinograms (ERGs) are generated by different types of retinal cells ^{1 2 3 4 5 6 7 8 9 10} and are used to assess the functioning of the cone system in normal subjects and patients with various types of retinal diseases. The intensity–response function of the photopic flash ERG b-wave is not a simple function, because at lower and intermediate stimulus intensities, the amplitude of the photopic ERG b-wave increases as the intensity is increased as expected, but at still higher stimulus intensities, the amplitude unexpectedly decreases. This unusual property of the human photopic b-wave was first described by Wali and Leguire, ^{11 12} and has been studied by many investigators. ^{13 14 15 16 17} Because a plot of the b-wave amplitude as a function of the stimulus intensity has an inverted U shape, this phenomenon has been named the “photopic hill.” It has recently been reported that this phenomenon is flash dependent, and similar results are observed at different levels of light adaptation. ¹⁶  

The exact mechanism for the photopic hill has not been determined, although several hypothesis have been proposed: a summation of the negative a-wave and the positive b-waves, ¹¹ an interaction of the depolarizing bipolar cells (DBCs) and hyperpolarizing bipolar cells (HBCs), ^{16 17} and a reduction of the OFF-response (d-wave) at higher stimulus intensities. ¹⁴  

The purpose of this study was to determine the mechanism underlying the photopic hill phenomenon. We asked two questions: first, whether the inner retinal neurons contribute to the photopic hill and, second, how the postreceptoral ON- and OFF-pathways contribute to the photopic hill. To address these questions, we recorded photopic flash ERG b-waves elicited by different stimulus intensities in monkeys. We used pharmacologic agents to isolate the activities from the different retinal cells and pathways. 

Materials and Methods

Subjects

Two rhesus (Macaca mulatta) and three cynomolgus (M. fascicularis) monkeys were studied. The animals were anesthetized with an intramuscular injection of ketamine hydrochloride (7 mg/kg, 5–10 mg/kg per hour maintenance dose) and xylazine (0.6 mg/kg). Respiration and heart rate were monitored, and hydration was maintained by slow, continuous subcutaneous infusion of lactated Ringer’s solution. The cornea was anesthetized by topical 1% tetracaine, and the pupil was dilated with topical 0.5% tropicamide, 0.5% phenylephrine HCl, and 1% atropine. Experiments were conducted in accordance with NIH guidelines on animal use and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

ERGs were also recorded from four normal human subjects (age range, 28–37 years). Informed consent was obtained from all subjects after a full explanation of the procedures. All studies were conducted in accordance with the principles embodied in the Declaration of Helsinki. 

Visual Stimulation

Ganzfeld ERGs were elicited by xenon photostrobe flashes with a nominal flash duration at half height of 10 to 30 μs (color temperature, 5700° K). The Ganzfeld dome also housed the rod-suppressing background light of 40 cd/m². The maximum flash intensity measured in the dome was 2.30 log cd-s/m², which produced a maximum retinal intensity of 4.00 log photopic troland/s. The stimulus luminance was attenuated over a range of 3.2 log with neutral-density filters (Wratten; Eastman Kodak, Rochester, NY) in steps of 0.4 log units. 

To monitor the drug effect on the monkey ERGs, 200-ms flashes were also recorded using a densely packed array of 102 green LEDs (525 nm peak wavelength; 50 nm at half amplitude), which was positioned on the top of the Ganzfeld dome and covered by a diffuser. The LEDs were controlled by a digital function generator (model WF1945; NF Corp., Tokyo, Japan). The stimulus intensity measured in the dome was 2.48 log cd/m², and the stimulus was presented on a white background luminance of 40 cd/m². 

ERG Recording and Analysis

After an initial 10 minutes of light adaptation at 40 cd/m², ERGs were recorded with a Burian-Allen bipolar contact lens electrode (Hansen Ophthalmic Development Laboratories, Iowa City, IA). A ground electrode was attached to the ipsilateral ear. Responses were amplified with a band-pass of 0.3 to 1000 Hz and digitized at 4.3 kHz. Contamination from line noise was reduced with a 60-Hz notch filter. To minimize the adaptational effect of high intensity stimuli, a small number of responses (three to six) were averaged and a long interstimulus interval of 15 to 240 seconds, depending on stimulus intensity, were used (Power Laboratory; AD Instruments, Castle Hill, New South Wales, Australia). 

Drug Application

Drugs were injected into the vitreous with a 30-gauge needle inserted through the pars plana approximately 4 mm posterior to the limbus. Drugs (Sigma-Aldrich, St. Louis, MO) were dissolved in sterile saline and injected in amounts of 0.05 to 0.07 mL. Drugs and intravitreous concentrations were l-2 amino-4-phosphonobutyric acid (APB), 1 mM; cis-2,3 piperidine dicarboxylic acid (PDA) 5 mM; tetrodotoxin citrate (TTX), 4 μM; and N-methyl-d-aspartic acid (NMDA), 4 mM. Recordings were begun approximately 60 to 90 minutes after drug injections, and studies were completed within 5 hours. Although the drug effects are mostly reversible after a recovery period of several weeks, the results that are shown were recorded from the eyes not previously treated. 

We were not able to separate the neural activities of DBCs and HBCs completely. We initially injected TTX+NMDA, APB, and PDA continuously to separate these components in a single experiment ^{10 18} but could not obtain reliable data because of unstable recording conditions due to the duration of the experiment. 

Results

Intensity–Response Function of Photopic ERG b-Wave in Humans and Monkeys

The photopic ERGs elicited by different stimulus intensities and recorded from a representative human and monkey are shown in Figure 1A . The waveforms after filtering the high-frequency OPs (100–300 Hz, gray traces) are superimposed on the actual ERG waveforms (black traces). The means ± SE of the means (SEM) of the b-wave amplitude for four humans and five monkeys are plotted in Figure 1B . 

The waveforms and intensity–response functions of the b-wave were quite similar in the humans and monkeys. As expected, the amplitude of the b-waves increased with increasing stimulus intensities, then reached a maximum at approximately 0.7 cd-s/m². Further increases in the stimulus intensity led to a rapid decline in the b-wave amplitude. These findings agree with previous studies in humans, ^{11 12 13 14 15 16 17} and the stimulus intensity that elicited the maximum b-wave amplitude was approximately the same as the value reported earlier. ^{11 12 13 14 15 16 17} The similarities in the photopic ERGs in humans and monkeys justified the use of monkeys to investigate the mechanism of the photopic hill seen in the human ERGs. 

Effect of Blocking Inner Retinal Neural Activity by TTX and NMDA

First, we asked whether the inner retinal neurons are contributing to the photopic hill. To test this question, we applied TTX and NMDA, because TTX blocks voltage-gated sodium channels and prevents action potentials of ganglion cells and some kinds of amacrine cells. ^{19 20 21} NMDA suppresses synaptic transmission by its antagonistic action (depolarization) on the NMDA subclass of glutamate receptors located primarily on the third-order neurons. ^{22 23} It is thought that the intravitreous injection of TTX+NMDA can suppress most, if not all, electrical activities from inner retinal neurons. ^{10 18 24}  

The photopic ERGs elicited by different stimulus intensities before and after the intravitreal injection of TTX+NMDA are shown in Figure 2 for one monkey (monkey 2). In the right column, the ERGs after TTX+NMDA (black traces) are superimposed on the control waveforms (gray traces). Both the amplitude and implicit times of the photopic ERG b-wave did not change significantly after TTX+NMDA to all stimulus intensities. The intensity–response curves of the photopic ERG b-wave before and after TTX+NMDA were also very similar, and both showed the photopic hill. These results indicated that the contribution from inner retinal neurons to the photopic hill is small. 

We also noted that the amplitude of the OPs was slightly reduced after injection of TTX+NMDA, and the late negative component (photopic negative response, PhNR) became smaller as reported by Viswanathan et al. ⁷ It was also noted that at low and intermediate stimulus intensities, another positive component after the b-wave, the i-wave, became prominent after TTX+NMDA. ⁷  

Effect of Blocking ON-Pathway by APB

Next we asked how the postreceptoral ON- and OFF-pathways contribute to the photopic hill. For this, we first injected APB ²⁵ to block the activities of the postreceptoral ON-pathway and then PDA ²⁶ to block the OFF-pathways. We then isolated the ON- and OFF-components by subtracting the postdrug ERGs from the predrug ERGs. 

The photopic ERGs elicited by different stimulus intensities after APB in two monkeys (monkeys 3 and 4) are shown in Figure 3A . The waveforms after APB (black traces) are superimposed on the control waveforms (gray traces). Even after blocking the postsynaptic ON-pathway by APB, a sharp positive peak (d-wave ^{5 27} ) remained at low and intermediate stimulus intensities (−0.9–0.7 log cd-s/m²). However, the implicit time of this positive peak was 4 to 7 ms later than the implicit time of the control b-wave. 

At higher stimulus intensities of 1.1 to 2.3 log cd-s/m², this positive component became less sharp, and its implicit time was increasingly prolonged (Fig. 3A , arrows). At higher stimulus intensities, the time of this positive peak moved far from that of control b-wave, and a deep, wide negative wave dominated the response. 

The subtracted ON-components (APB-sensitive component) at the different stimulus intensities are shown in Figure 3B . The ON-component had a sharp positive peak at low and intermediate stimulus intensities (−0.9–0.7 log cd-s/m²). With increasing intensities, the amplitude of this positive peak first increased but then decreased. The stimulus intensity which elicited maximal amplitudes of this positive peak was 0.7 log cd-s/m², which was the same stimulus intensity that elicited maximum b-wave amplitude in the control eyes. 

At higher intensities, the amplitude of this positive wave became smaller, and the waveform became wider. These results strongly suggested that the amplitude reduction of the positive ON-component at higher stimulus intensities significantly contributed to the photopic hill. 

Effect of Blocking the OFF-Pathway with PDA

The next step was to test whether the neural activity of the postsynaptic OFF-pathway contributes to the photopic hill. For this, we added PDA after APB in the same two eyes (monkeys 3 and 4). This also allowed us to determine whether the cone photoreceptors contributes to the photopic hill. 

The photopic ERGs recorded after injection of APB+PDA are shown in Figure 4A . The intensity–response function of the isolated photoreceptor response is also plotted in the lower trace of Figure 4A . After APB+PDA, nearly all positive components were lost at all stimulus intensities. These results indicated that the cone photoreceptors do not contribute to the photopic hill directly. It was also noted that the contribution from the cone photoreceptors to the a-wave is minor at lower stimulus intensities. ⁶  

With increasing stimulus intensities, this negative photoreceptor component gradually became deeper and wider. Past studies using whole-cell voltage and current recordings from macaque cone photoreceptors showed a similar delayed recovery of the receptor potential at stimulus intensities that saturate the negative receptor response. ^{28 29 30} This study of single photoreceptors also showed that the amplitude–intensity relation followed a Michaelis function, as our data after APB+PDA show (lower half of Fig. 4A ). ²⁹ However, the duration of the isolated a-wave in our study was not longer than the cone photoreceptor studies would predict. ^{28 29 30}  

The amplitudes of this photoreceptor component were consistently smaller than those after APB alone. These findings also agreed with previous findings that in addition to the cone photoreceptors, other proximal retinal neurons (e.g., HBCs), contribute to the photopic ERG a-wave. ⁶  

The subtracted OFF-component (PDA-sensitive component) at the different stimulus intensities are shown in Figure 4B . As with the ON-component, the OFF-component had a sharp positive peak at low and middle stimulus intensities. Although the time of this positive peak was slightly later (4–7 msec) than that of the control b-wave, it is clear that this positive peak of the OFF-component contributed to the control b-wave at these stimulus intensities. The amplitude of this positive component became larger with increasing intensities at lower stimulus levels (−0.9–0.3 log cd-s/m²), and then became slightly small and broader at higher intensities. In addition, the implicit time of this positive component was gradually prolonged with increasing intensities (arrowheads). At the highest stimulus intensity, a deep, wide negative wave dominated at the timing of the control ERG b-wave. These results indicate that the striking delay in the positive peak of the OFF-component at higher intensities contribute to the photopic hill. 

Effect of PDA Alone

Finally, we wanted to see how the photopic flash ERG b-waves are altered after application of PDA alone to different stimulus intensities. PDA blocks not only the HBCs, but also other neural activity in the outer and inner plexiform layers. ^{26 31 32} This experiment is important because PDA removes the effect of the horizontal cells, and one can test the hypothesis that inhibitory feedback from the horizontal cells is the cause of the photopic hill. 

The intensity response series of the photopic ERG after injection of PDA alone in two monkeys (monkeys 1 and 5) are shown in Figure 5 . After application of PDA alone, the amplitude of the b-wave became larger with increasing intensities at lower and intermediate intensities, then reached a maximum at approximately 0.7 cd-s/m². At higher stimulus levels, the amplitude became small, and the waveform became wider. At maximum intensity, the amplitudes of the positive peak were 51% and 71% of those at 0.7 cd-s/m² in the two eyes. Although the photopic hill was smaller than average (45%, see Fig. 1 ), these results indicated that the photopic hill remained, even after the effects of the horizontal cells were suppressed. This suggests that inhibitory feedback is not the main mechanism for the photopic hill. 

Discussion

Our results are summarized in Figure 6 based on data obtained from monkey 3. The waveforms of the three components—photoreceptoral, ON-, and OFF-components—are shown in different colors at three stimulus levels: low (−0.5 log cd-s/m²), intermediate (0.3 log cd-s/m²), and high (2.3 log cd-s/m²) intensities. 

It has been proposed that the passive addition of negative and positive potentials of the photoreceptoral, ON-, and OFF-pathways may be involved in generating the photopic hill. ^{16 17} Our results partly agree with this proposal. As shown in the right column in Figure 6 , the wide positive wave of the ON-component appeared to interfere with the wide negative OFF- and photoreceptoral components at higher stimulus levels. However, the magnitude of the contribution of this effect to the photopic hill was not as large when compared with two main factors, which will be described later. In addition, it is clear that this passive addition of negative and positive potentials for generating the b-wave exists even at low and intermediate intensities, because there was a timing lag between the two positive peaks (Fig. 6 , left and middle columns). ^{5 10} Thus, although the passive additive effect of negative and positive potentials may also contribute to the generation of the photopic hill, its share is probably relatively small. 

Our results demonstrated clearly that the photopic hill results mainly from the neural activity of the postreceptoral ON- and OFF-pathways. At low and intermediate stimulus intensities, the photopic b-wave is shaped by the overlapping of two positive peaks of the ON- and OFF-components, as expected from earlier studies. ^{5 10} At these stimulus intensities, the amplitudes of these two positive peaks increased with increasing stimulus intensities summing to result in the increase in the amplitude of the b-wave (Fig. 6 , left and middle column). At higher stimulus intensities, however, the positive peak of the ON-component became smaller and broader. In addition, the positive peak of the OFF-component was dramatically delayed with increasing intensity, and no longer contributed to the photopic b-wave (Fig. 6 , right column). Thus, the resultant b-wave decreased. 

From these results, we conclude that the photopic hill results mainly from two factors: the amplitude reduction of the ON-component at higher intensities and the delay in the positive peak of the OFF-component at higher intensities. Our results also indicate that the contribution from inner retinal neurons to the photopic hill is minor, because both the implicit times and amplitude of the photopic b-wave did not change much after TTX and NMDA at all stimulus intensities (Fig. 2) . We also confirmed that the contribution of inhibitory feedback from the horizontal cells to the photopic hill is small, because the photopic hill, although somewhat reduced in amplitude in at least one animal, remained even with application of PDA alone (Fig. 5) . 

The positive component from the APB-sensitive component, presumably generated by DBCs, increased at low and middle intensities, and then became smaller and broader at higher stimulus intensities (Fig. 3B) . Whether these results can predict the shape of the intensity–response function of the cone DBCs in monkeys is still uncertain, because there have been very few reports on the intensity–response function of the light-evoked potentials in the single cone DBCs in the mammalian retina. However, Berntson and Taylor ³³ studied the light-evoked responses from bipolar cells in the mouse, and reported that the amplitude of the photovoltage of cone DBCs increased with increasing stimulus intensity and then reached a plateau at higher stimulus intensity. There was no notable amplitude reduction in the photovoltage of cone DBCs, even at the highest intensities in the mouse. The reason for the discrepancy between their results and ours may be explained by differences in experimental conditions and species. First, they used a long-duration stimuli (390 ms), whereas we used brief xenon flashes (<30 μs). It is known that the photopic hill is seen only when brief-flash stimuli are used. Second, they used mouse retina, whereas we used monkey retina. It has recently been reported that the photopic hill is less prominent in rodents (Joly S, et al. IOVS 2002;43:ARVO E-Abstract 1782). To address the question of whether our results can predict the electrical function of monkey cone DBCs, further studies on the intensity–response function on a single DBC in primate are needed. 

The results obtained with PDA (Fig. 4) indicated that a portion of the photopic hill occurs because the positive HBC response becomes more delayed at higher flash levels, so that it adds less efficiently to the DBC (APB-sensitive) component. This situation can be mimicked at lower flash levels by extending the duration of the stimulus so that the positive OFF-response is shifted in time. Actually, past studies have reported that the amplitude of the b-wave decreases with increasing flash duration (for example, Ref. ¹⁴ ). 

The exact mechanism for the delay in the positive PDA-sensitive component at higher stimulus intensity is also unclear, because there have been no reports on the intensity–response function of single cone HBCs in monkeys. One plausible explanation is that this stimulus-dependent delay could reflect the longer time course needed for cone photoreceptors to recover after strong flashes. The isolated cone photoreceptor responses shown in Figure 4A and data from the single-cell recordings from macaque cone photoreceptors ^{28 29 30} are consistent with this idea. 

In conclusion, the photopic hill in the primate ERG results mainly from two factors: the amplitude reduction of the ON-component at higher intensities and the delay in the positive peak of the OFF-component at higher intensities. To determine the exact cellular mechanism underlying this phenomenon, further studies on the intensity–response function in primate cone bipolar cells are needed. 

 Supported by Grants-in-Aid 14770952 (MK), 13307048 (YM), and 14370557 (HT) from the Ministry of Education, Science, Sports and Culture, Japan.

 Submitted for publication June 26, 2003; revised October 14, 2003; accepted October 31, 2003.

 Disclosure: S. Ueno, None; M. Kondo, None; Y. Niwa, None; H. Terasaki, None; Y. Miyake, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) Photopic flash ERGs elicited by different stimulus intensities recorded from a representative normal human subject and monkey 1. The waveforms after filtering the high-frequency OPs (100–300 Hz, gray trace) are superimposed on the actual ERG waveforms (black trace). (B) The intensity–response curves of the mean photopic b-wave amplitudes in four humans and five monkeys. Means ± SEM are plotted.

Figure 2.

 

Photopic flash ERGs elicited from monkey 2 by different stimulus intensities before and after the injection of TTX+NMDA. ERGs after administration of drugs (black trace) are superimposed on the control ERGs (gray trace).

Figure 3.

 

(A) Photopic flash ERGs elicited by different stimulus intensities before and after APB treatment in two animals (monkeys 3 and 4). ERGs after administration of APB (black traces) are superimposed on the control ERGs (gray traces). (B) Subtracted ON-components to different stimulus intensities. ON-components (APB-sensitive components) were obtained by subtracting the post-APB records from the control records. Results from two animals are shown (monkey 3 and 4).

Figure 4.

 

(A) Top trace: photopic flash ERGs elicited by different stimulus intensities before and after APB+PDA injection in the same two animals. ERGs after injection of APB+BDA (black traces) are superimposed on the control ERGs (gray traces). Bottom trace: response amplitude of two monkeys after APB+PDA is plotted as a function of flash strength on normalized axes. (○) Monkey 3; (▪) monkey 4; r, response amplitude; r _max maximum response to saturating stimulus; i, flash stimulus intensity; i ₀ the flash intensity elicited a peak response of one half r _max. Gray line: Michaelis function, r/r _max = i/(i+i ₀); (B) Subtracted OFF-components to different stimulus intensities. OFF-component (PDA-sensitive components) was obtained by subtracting the post-APB+PDA records from the post-APB records. Results from the same two animals are shown (monkey 3 and 4). Subtracted responses (black traces) are superimposed on the control ERGs (gray traces).

Figure 5.

 

Photopic flash ERGs elicited by different stimulus intensities before and after PDA alone for two animals (monkeys 1 and 5). ERGs after PDA (black traces) are superimposed on the control ERGs (gray traces).

Figure 6.

 

The photoreceptoral (red) and the postreceptoral ON- (blue) and OFF- (green) components of the photopic flash ERGs at three stimulus levels. Each component at low (−0.5 log cd-s/m²), middle (0.3 log cd-s/m²), and high (2.3 log cd-s/m²) stimulus intensities from monkey 3 is shown.

The authors thank Masao Yoshikawa, Eiichiro Nagasaka, and Hidetaka Kudo of Mayo Co. (Nagoya, Japan) and Hiroyuki Sakai of Santen Co. (Osaka, Japan) for technical help.