Regional Variations in Local Contributions to the Primate Photopic Flash ERG: Revealed Using the Slow-Sequence mfERG

Abstract

purpose. To determine the variations with eccentricity of the primate photopic ERG and to separate contributions by different retinal cells by using intravitreal pharmacologic agents.

methods. Slow-sequence multifocal (mf)ERGs were obtained from 19 anesthetized adult rhesus monkeys and 5 normal human subjects. Recordings in monkeys were obtained before and after injections of tetrodotoxin citrate (TTX) to block sodium-dependent spiking; TTX+N-methyl-d-aspartic acid (NMDA)+picrotoxin (PTX) or γ-aminobutyric acid (GABA) to block all inner retinal activity; l-2 amino-4-phosphonobutyric acid (APB) to block the On-pathway; and cis-2, 3 piperidine dicarboxylic acid (PDA) to block the Off-pathway and the otherwise unblocked inner retinal activity. The stimulus consisted of 103 equal-sized hexagons within 17° of the fovea; every 200 ms (15 frames), each hexagon had a 50% chance of remaining at 20 cd/m² or increasing briefly to 4.7 cd-s/m². Oscillatory potentials (OPs; 90–300 Hz) were extracted.

results. The slow-sequence mfERG summed over the stimulated area looked similar to a standard photopic, full-field ERG, with a- and b-waves and OPs. OPs in the foveal and temporal retina were larger than in the nasal retina. This nasotemporal asymmetry was removed by TTX, and the OPs were eliminated, either by blocking inner retina activity or by blocking the On-pathway. The summed mfERG waveform, including OPs, was shaped mainly by the more peripheral retinal regions. The foveal b-wave peak occurred about 5 to 6 ms later than in the periphery, with the depolarizing peak of the On-pathway/bipolar contribution occurring earlier than the depolarizing peak of the Off contribution at all eccentricities. The a-wave was composed of a small photoreceptor contribution and postreceptoral portion originating from hyperpolarizing neurons.

conclusions. The variations in the primate photopic ERG with eccentricity are due to spike-driven oscillatory activity that is more prominent in central and temporal retina than in nasal retina and to the slower timing of all responses in the central, compared with the peripheral, retina. The full-field, photopic ERG most closely resembles the mfERG responses to stimulation of peripheral regions.

The electroretinogram (ERG) is a mass potential that is the sum of electrical activity of all the cells in the retina. It is a useful tool for assessing function in normal and diseased retinas in vivo. The photopic ERG is commonly elicited with a brief flash filling the field of view and presented on a steady background that is strong enough to suppress rod activity. A typical full-field photopic flash ERG consists of a few components the specific retinal origins of which have been investigated in pharmacological dissection experiments in macaque monkey whose retina is very similar to that of humans. The initial negative a-wave originates from cone photoreceptors and off-bipolar cells, ¹ perhaps with contributions from more proximally located cells. ² The a-wave is followed by a positive b-wave, arising from On-bipolar cells and shaped by hyperpolarizing second-order cells (Off-bipolar cells and possibly horizontal cell feedback ³ ) or under some stimulus conditions, spiking neurons of more proximal origin. ⁴ When long-duration flashes are used, the positive d-wave that appears after termination of the stimulus arises from an interplay between the Off-bipolar cell responses and the activity of the depolarizing (On) bipolar cells, ^{3 5} as well as the offset of cone photoreceptor (as reviewed in Ref. ⁶ ). When high-energy flashes are used, small high-frequency rhythmic wavelets called oscillatory potentials (OPs) are superimposed mainly on the b-wave. ^{7 8} The temporal frequency range for OPs has been reported to be between 100 and 150 Hz in the macaque retina ⁹ and between approximately 50 and 200 Hz ¹⁰ or as high-frequency activity above 75 to 100 Hz in humans, ¹¹ although a range of 90 to 160 Hz was reported in earlier literature. ¹² Current source density analyses in macaque retina have shown that photopic OPs originate mainly from the inner plexiform and inner nuclear layers, ⁹ but the exact cellular origins of the OPs are currently unknown. 

Full-field stimulation elicits a single ERG response summed over the entire retina. However, it is known that the cytoarchitecture of the primate retina varies rapidly with changes in eccentricity from the fovea (a densely packed cone-rich area) to the periphery. The number of each of the major cell types to transmit signals (i.e., the cone photoreceptors, bipolar cells, and ganglion cells) decreases with distance from the fovea, and the relative balance in the classes of neuron contributing to different parallel pathways alters. ^{13 14 15 16 17 18 19}  

Previous electrophysiological studies in humans using focal stimulation have shown that the responses from different retinal regions have different waveforms. ^{20 21} For example, in response to focal stimulation, OPs in the temporal retina are larger than those in nasal retina. To understand better these and other variations of the photopic ERG with eccentricity, a technique is required that can stimulate multiple retinal regions and obtain individual responses from each of those regions simultaneously. The multifocal (mf)ERG technique developed by Sutter is a very efficient method of recording focal ERGs from numerous retinal locations in a short time. ^{22 23} mfERGs are typically measured with a multifocal stimulus presented on a CRT that has a frame rate of 75 Hz, with a separation between the m-sequence presentations of 13.3 ms (fast-sequence mfERG). Because of the fast sequence of stimulation, the mfERG obtained with the fast-sequence paradigm are somewhat different in waveform from those of the conventional full-field flash ERGs. However, when the m-sequence presentation is slowed by interleaving seven or more blank frames between the presentations (slow-sequence mfERG), the waveforms of the full-field and mfERG become more similar. ²⁴ Specifically, it is possible in the slow-sequence mfERG to see traditional a- and b-waves and oscillatory potentials, and when the presentation is lengthened, to see d-waves at light offset as well. ^{25 26} Further, with the slow-sequence mfERG, the regional variations of the standard waves of the photopic ERG can be identified. ^{24 25 26}  

Recently, intravitreal pharmacologic agents were used in the macaque eye, which has a retina that is very similar to that of humans, to determine the cellular origins of the components of the fast-sequence mfERG. ^{27 28} It was shown that the monkey’s mfERG is shaped by overlapping contributions from On- and Off-bipolar cells, with small contributions directly from the photoreceptors. Both the relative bipolar and photoreceptor contributions were largest in the central retina. Spiking and nonspiking inner retinal activity of an oscillatory nature also were present. When the spiking activity was eliminated, the monkey mfERG was similar to that of humans, making it a useful model for studying the origins of human mfERGs. 

In the present study we used the slow-sequence mfERG in macaque monkeys to investigate the variations of the photopic flash ERG with eccentricity. Using intravitreal pharmacologic agents, we also separated the contributions of the different retinal cells to the photopic ERG at various eccentricities. Because the slow-sequence mfERG of normal human subjects resembles, to a large extent, that of the macaque, some predictions about the contributions to the human ERG are possible. 

Methods

Subjects

Monkeys.

mfERG recordings were obtained from 22 eyes of 19 adult rhesus monkeys (Macaca mulatta). All experimental and animal care procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care Committee of the University of Houston. All these animals were used in other studies as well (e.g., Refs. ²⁷ , ^{29 30 31} ). 

Humans.

mfERGs were also obtained from five eyes of five normal human subjects (age: 18–25 years). Informed consent was obtained before participation. Procedures adhered to the tenets of the Declaration of Helsinki, and the protocol was approved by the University of Houston Committee for the Protection of Human Subjects. 

Animal Preparation for ERG Recordings

Animals were anesthetized intramuscularly with ketamine (20–25 mg/kg per hour) and xylazine (0.8–0.9 mg/kg per hour) and were treated with atropine sulfate (0.04 mg/kg, injected subcutaneously). The depth of the anesthesia was maintained at a level sufficient to prevent the animals from blinking or moving. Pupils were dilated fully to approximately 9 mm in diameter with topical tropicamide (1%) and phenylephrine hydrochloride (2.5%), and the eye to be studied was refracted retinoscopically for the viewing distance and fitted with appropriate contact lenses. The nonstimulated eye was covered. Heart rate and blood oxygen were monitored with a pulse oximeter (model 44021; Heska Corp., Fort Collins, CO), and body temperature was maintained between 36.5°C and 38°C with a water-circulating heating pad. A modified ophthalmoscopic technique was used to locate the projection of the fovea on the center of the stimulus pattern. The position of the fovea was checked frequently, recentering it when necessary, particularly after intravitreal injections. 

Recordings in Monkeys

ERGs were recorded differentially between DTL ³² electrodes that were placed across the center of the cornea and under a corneal contact lens of both eyes. The DTL fiber was moistened with 1% carboxymethylcellulose sodium. A needle inserted under the scalp served as a ground electrode. Each DTL fiber was anchored with a dab of petroleum jelly near the inner canthus and electrically connected by clip leads at the outer canthus. The low- and high-cutoff frequencies were set at 1 and 300 Hz, with no additional notch filtering. 

Recordings in Humans

ERGs were recorded differentially between the two eyes in human control subjects with DTL electrodes. The human subjects were wired similarly to the monkeys except for the ground electrode, which was an adhesive silver–silver chloride electrocardiogram electrode (Sentry Medical Products, Green Bay, WI), placed on the forehead. The stimulus conditions were as close to those for the monkeys as feasible. The humans wore their own eyeglasses if correction was needed. In all subjects, the recordings were from the right eye, whereas the nontested left eye was covered. 

Intravitreal Injections

Intravitreal injections of 40 to 50 μL were made nasally and temporally in the globe behind the limbus with a sterile 30-gauge needle inserted through the pars plana into the vitreous cavity. Intravitreal concentrations of the pharmacologic agents were estimated by assuming the vitreous volume to be 2.1 mL. The following drugs and concentrations, all in sterile balanced salt solution, were used: tetrodotoxin citrate (TTX; 1.2–2.1 μM), N-methyl-d-aspartic acid (NMDA; 1.4–6.4 mM), γ-aminobutyric acid (GABA; 37–55 mM), picrotoxin (PTX; 0.1–0.4 mM), l-2 amino-4-phosphonobutyric acid (APB; 0.83–3.8 mM), and cis-2, 3 piperidine dicarboxylic acid (PDA; 3.3–3.8 mM). These concentrations of pharmacologic agents were found to be sufficient in other experiments on primates to separate pharmacologically the components of full-field flash ERG (e.g., Refs. ^{3 4 27 31} ). In our experiments, the expected effect of these concentrations on the full-field flash ERG were verified with a Ganzfeld stimulator. ⁴ Recordings were made before and at least 1 hour after injections, when effects had stabilized. 

Stimulation

The stimulus display consisted of 103 equal-sized hexagons, each approximately 3.3° wide, in a field of approximately 35° by 33° (Fig. 1A) . The optic nerve head is approximately 16.5° from the fovea in the monkey, ³³ 17° in humans. The location of the monkey’s optic nerve head is marked X in Figure 1 . Recordings were made with the room lights on. Hexagons appeared at the same luminance as surround (20 cd/m²) or, briefly, as a bright flash (4.7 cd-s/m²) (Fig. 2) . Photopic luminance (cd/m²) of the stimulus was calibrated with a spot photometer (model LS-100; Minolta Camera Co., Ltd., Osaka, Japan). The elements of the sequence were 13.3 ms in duration (corresponding to the frame rate of 75 Hz). However, the actual duration of the incremental light producing a white hexagon was under 2 ms. In the slow-sequence, as illustrated in Figure 2 , the frames in which the hexagons appeared (m-frame) were separated by 14 blank frames (time between flashes being at least 200 ms) for which the entire screen was typically at the background luminance of 20 cd/m². For comparison, we also used seven blank frames (time between flashes being at least 106.7 ms). On every m-frame, a given hexagon had a 50% chance of remaining at the background intensity or increasing to 4.7 cd-s/m². An experimental run, requiring approximately 7 minutes of recording consisted of 2¹¹–1 m-sequences. Thus, each hexagon appeared approximately 2047 times. For recordings in the human subjects, the 7-minute runs were broken into 32 segments to allow for blinking. 

Analysis

After the recording session was completed (Visual Evoked Response Imaging Systems; VERIS, Electro-Diagnostic Imaging [EDI], San Mateo, CA), the program computed the focal ERG for each hexagon as the cross-correlation between the stimulation sequence and the continuously recorded ERG. These first-order responses were analyzed by the software (VERIS ver. 4.0; EDI), as described previously (see Refs. ²² , ²³ ). 

To obtain the portion of the waveform removed by any drug, we subtracted the postdrug recordings from equally weighed control recordings (predrug recordings). 

OPs were present on the positive wave of the slow-sequence mfERG. These OPs were extracted by bandpass filtering the mfERG recordings between 90 and 300 Hz on a computer (Matlab; Mathworks Inc., Natick, MA). We chose to exclude frequencies of less than 90 Hz, because frequency analysis of the mfERGs showed that frequencies greater than 90 Hz were most involved in producing nasotemporal variation (Rangaswamy NV, Frishman LJ, Saszik SM, Hood DC, Harwerth RS, ARVO Abstract 2170, 2002). The amplitude of the OPs in a 50-ms time window, using the 60 points sampled between 10 and 60 ms, was determined by calculating the root mean square (RMS)  

\[\mathrm{RMS}\ {=}\ \sqrt{\frac{1}{60}\ {{\sum}_{t\ {=}\ 10}^{t\ {=}\ 60}}\ (x_{t}\ {-}\ \mathrm{mean})^{2}}\]

An estimate of the noise level was obtained by calculating the RMS between approximately 60 and 80 ms in the records, where oscillatory activity was minimal, and averaging the five values for the four chevrons (N1, N2, T1, and T2) and the foveal (F) record to get the mean OP RMS noise for each eye. 

For quantifying mfERG response amplitudes, depending on the question that we were addressing, we used one of two different methods implemented in the evoked response recording system software: the sum of groups in microvolts or response density in nanovolts per degree. ² For the sum of groups calculation, the responses for each stimulus element (hexagon) in the selected group of elements are added together, providing a cumulative response for that group. This approach provides the absolute amplitude of the mfERG in that group, allowing us to quantify its contribution to the mfERG summed over the entire stimulus array. For calculating the response-density results, the responses for each stimulus element in a group are added, and the result is divided by the total solid angle of all elements in the group. This provides a measure of response per unit area and provides a better picture of the relative prominence of the components of ERG waveforms at different eccentricities. 

Student’s t-test was used for statistical analyses in this study. A paired t-test was used when we had to compare parameters within the same eye of an animal or within the same eye before and after drug treatment. For example, a paired t-test was performed on the following: to compare the asymmetry between the nasal and temporal OP RMS in control eyes (Fig. 5B) , to test for the effect of drugs on the nasal and temporal OP RMS (Figs 5A 5B) , to test for the difference in the time to b-wave peak between the central and peripheral records after inner retinal blockade (Fig. 6) , and to test for significance in the difference between the depolarization peak and hyperpolarization peak times of the On- and Off-pathway/bipolar cells (Figs. 7 8) . 

Results

Control mfERG

The slow-sequence mfERG of the monkey showed clear regional variations. These can be seen in Figure 1B , which shows the records for the entire 103-hexagon array from the normal left eye of one monkey (monkey A). Although looking at the trace array is not the easiest way to resolve all the variations in waveforms, a few general observations can be made. First, there were more oscillations in the central and temporal retina (marked in the figure by the oval) than in the more peripheral and nasal retina. This finding is in qualitative agreement with results of previous human ERG studies of focal ERGs ²⁰ and slow-sequence (three frames interleaved) mfERGs. ³⁴ Also, the amplitude of the major mfERG components decreased with distance from the fovea. A better way to see these variations is to group the responses at different distances from the fovea (Fig. 1C , rings) or in nasal versus temporal regions (Fig. 1D , chevrons) and to compare the waveforms of the different groups. Results obtained by grouping records in these ways will be described in later sections. 

Before examining the regional variations of the slow-sequence mfERG more closely, we refer to Figure 3A which shows the slow-sequence mfERG summed over the whole multifocal array (the central 33° in this case), with the extracted OPs shown below for the normal eyes of two different monkeys, VI and VA. For comparison, Figure 3B shows a typical full-field ERG response to a xenon flash presented on a rod-saturating steady background in a Ganzfeld and the extracted OPs from one of the monkeys, VI. As previously described for the human mfERG, ²⁴ the summed slow-sequence mfERG (at least seven blank frames) and the full-field flash ERG were quite similar in a- and b-waves, with OPs of similar timing superimposed on the b-wave in both cases. These similarities serve to justify the use of the slow-sequence mfERG to look for regional variations in the photopic flash ERG. In the present study, we describe data measured mainly using 14 blank frames rather than 7, although in most cases we measured responses with very similar waveforms for both conditions. We chose to describe the data obtained using 14 blank frames because, in the monkeys, the OPs were slightly larger in those records (data not shown). 

Variation with Eccentricity in Control Subjects

Figure 4B shows records grouped into rings at increasing distance from the fovea in three control eyes of three monkeys (M, VA, and T), with the extracted OPs shown below. The amplitudes at different eccentricities were expressed as sums of groups so that we could compare directly the contribution of OPs at different retinal eccentricities with the whole photopic ERG, as represented by the summed mfERG in Figure 4A . The waveform of the mfERG from the central two rings was substantially different from the waveform of the mfERG from the more peripheral rings. The lower panel of Figure 4B shows that the variation in the OPs, between the central and peripheral rings, made a major contribution to the overall variation with eccentricity. The summed mfERG (Fig. 4A) more closely resembled the responses from rings 3 to 5 than it did the responses from rings 1 and 2. As might be expected because of the greater area involved, the contributions from the outer rings, including the OP component, dominated the summed mfERG. 

Figure 4 also shows that with increasing retinal eccentricity the b-wave rose earlier and its peak time decreased. This was clearly seen in monkey M (dashed line in the left column of B) because the OPs were smallest in this animal. b-Wave peak timing was calculated after filtering the OPs on computer (Matlab; the Mathworks) in slow-sequence records obtained from 16 control eyes. The mfERGs after filtering are shown by the gray lines in Figure 4B . The peak time of the b-wave in the filtered records summed over the whole array was 29.0 ± 1.7 ms, and, for rings 1 and 5, the peak times were 32.7 ± 2.3 ms and 28.0 ± 1.5 ms, respectively (Table 1) . Peak time measurements made after removing the OPs with pharmacologic blockade of inner retina will be presented in a later section. 

Nasotemporal Asymmetries in Control mfERGs

Nasotemporal asymmetries in oscillatory responses have been described previously for the fast-sequence ³⁵ and slow-sequence (with three interleaving blank frames) mfERG. ^{34 36} Pharmacologic blockade of inner retinal activity or induction of experimental glaucoma in monkey reduces or eliminates the asymmetries in the fast-sequence mfERG. ^{29 35} In the present study, as illustrated in the array in Figure 1B , there also were nasotemporal asymmetries in the slow-sequence mfERG records of normal monkey eyes. To investigate the nasotemporal asymmetries, we grouped the responses, by using response-density scaling, into chevrons at equal distance from the fovea in the nasal and temporal retina. The mfERGs for these chevrons, as well as for the seven central hexagons (F) for comparison, are shown in the leftmost column of Figure 5A for three monkeys. N represents nasal retina and T, temporal retina; N1 and T1 are closer to the fovea than N2 and T2. The chevrons were assembled using only hexagons from the central three rows, because OPs in these hexagons were largest (e.g., Fig. 1B ). 

To compare better the nasotemporal variations in the OPs, extracted OPs are shown in the middle column of Figure 5A . In all three monkeys, the temporal OPs (especially T1) were larger than the nasal OPs. The density-scaled foveal OPs also were quite large, particularly at later times in the trace. Amplitudes of OPs were quantified by calculating the RMS of each record (as described in the Methods section) between 10 and 60 ms, as indicated by the two dashed lines in the lower row of Figure 5A , and the averaged results for T1 and N1 for all control eyes are shown as the left-hand bars in the histogram in Figure 5B . For the normal eyes, the RMS of chevron T1 OPs was significantly larger that the RMS of the NI OPs (P < 0.01). OPs in T2 versus N2 tended to be different, but these differences were not significant (P = 0.10). Therefore, in the remaining sections we compared the asymmetry only between N1 and T1 OPs. 

Effect on the OPs of Pharmacologic Suppression of Inner Retinal Activity

To investigate the origin of the OPs, we first used an intravitreal injection of TTX to block sodium-dependent action potentials. In the retina, only the retinal ganglion cells, some amacrine and the interplexiform cells are known to generate sodium-dependent action potentials. ^{37 38} Hence, injecting TTX eliminates any sodium-dependent spiking contribution of these cells to the OPs. The first row in Figure 5A , right column, shows the effect of TTX on the OPs in animal M (effects on the overall mfERG waveform will be considered in Fig. 6 ). TTX decreased the OPs and largely eliminated the nasotemporal waveform asymmetries in them. Figure 5B (group of bars second from the left) shows the average RMS of the OP amplitude for the nasal (N1-black) and temporal (T1-gray) chevrons for four animals injected with TTX. Although TTX reduced the amplitude of the OPs in both the nasal and temporal retina, the RMS for both chevrons was still larger than the RMS of the noise for the extracted OPs (dark gray bar). Further, although the amplitude of the temporal OPs after TTX was significantly reduced (P < 0.05), the amplitude of the nasal OPs was not (P = 0.37). Thus, the N1 and T1 amplitudes were no longer significantly different (P = 0.27) after TTX. In addition to the changes in amplitude, blocking inner retinal activity also modified the timing of the remaining OPs. 

In addition to generating action potentials, retinal ganglion and amacrine cells also produce nonspiking potentials. NMDA is a glutamate agonist that depolarizes cells with NMDA receptors, which are found on ganglion cells and also on at least some amacrine cells. The second row in the right-hand column of Figure 5A shows, for animal VI, that nasotemporal variations were removed, and OP amplitudes were greatly reduced after injection of both TTX and NMDA. The N1 and T1 OP amplitudes (RMS) were not significantly different (P = 0.96). The graph in Figure 5B shows, in four animals, that after TTX+NMDA, both the N1 and T1 OP amplitudes were significantly reduced (N1 and T1 both P < 0.05) compared with the control condition for the same eyes. Although both nasal and temporal OP amplitudes looked smaller after TTX+NMDA than after TTX alone, the difference was small in the two animals in which both measurements were made. In two animals, the effect of adding PTX to TTX+NMDA on the OP RMS was analyzed, and these results for two animals are shown individually in Figure 5B . PTX is a GABA_A&C receptor blocker. GABA_A&C receptors are present in the inner and outer plexiform layers, and we reasoned that blockade of these receptors, in retinas in which the inner retina was blocked would diminish any residual inhibitory feedback contributing to the formation of OPs. PTX did not have any obvious additional effect on the OPs in this case. 

GABA also has been shown to eliminate inner retinal activity (e.g., Ref. ³⁹ ). The left eye of animal VA, for which records are shown in the lowest row of Figure 5A , was injected with GABA. The OPs were practically eliminated, as illustrated for two animals individually in the right most column of Figure 5B . 

Effect on the b-Wave of Blocking Inner Retinal Activity

We also examined the effect of inner retinal blockade on the b-waves of the slow-sequence mfERG. Figure 6 shows density-scaled responses at different retinal eccentricities in four representative animals treated with different pharmacologic agents to block inner retinal activity. The thin gray lines show the control records, and the black lines show the records of the stabilized response at least 1 hour after the drug injection. Figure 6A shows the mfERG from one animal E before and after TTX. The right column shows the difference between the control and after administration of TTX, which includes the oscillatory portion, as described previously, that was removed by TTX. The removed portion in ring 1 also contains a negative-going slow wave at times after 40 ms that may be analogous to the TTX-sensitive photopic negative response (PhNR) ⁴ and some slow oscillations not captured in the extracted high-frequency OP records shown below. Figure 6B shows the effect of TTX+NMDA (left) TTX+NMDA+PTX (middle), and GABA (right). In each case, after inner retinal blockade, regardless of the mode of action of the drug, the mfERGs became smoother, because of removal of oscillations. The mfERGs in the central ring (ring 1) still showed multiple peaks, whereas the records from the peripheral rings generally did not, and the peak of the b-wave of the central records occurred later than in the peripheral records. The dashed line drawn through the peak of the b-wave in ring 1 and extending through the data in the other rings (after GABA in animal VA; Fig. 6B ) highlights the difference in timing between the central and peripheral records. As shown in Table 1 , the mean difference in the peak timing of the b-wave between rings 1 and 5 was 5.8 ± 1.6 ms in nine eyes of seven animals (i.e., both eyes at different times in two of the animals) after inner retinal blockade, and this difference was statistically significant (P < 0.01). The peak time of the b-wave in the records summed over the whole array was 30.3 ± 3.5 ms, in the animals in which the inner retina was blocked, and this peak time was similar to the value obtained in normal control subjects of 30.2 ± 1.4 ms. Thus, the peak time of the summed mfERG was determined mainly by the timing of the more peripheral records (Table 1) . 

On- and Off-Pathway Contributions

We also examined the contributions of the On- and Off-pathways of the retina, and then specifically the contributions of the isolated On- and Off-bipolar cells, and photoreceptors to the slow-sequence mfERG as a function of eccentricity. Two glutamate analogues, APB and PDA, were used to study isolated On- versus Off-responses. ¹ APB blocks signal transfer from photoreceptors to the mGluR6 receptors on On-bipolar cells, which leads to blockade of visual signal transmission in the entire On-pathway. Thus, APB alone can be used to study the two pathways, by examining the contributions removed by APB records (On-pathway) and the post APB records that contain the Off-pathway. The latter also has a direct photoreceptor contribution, which as shown later, was small in these studies. PDA, on the other hand, blocks responses not only of the Off-bipolar cells, but also of all inner retinal cells in both pathways. Therefore, to study the isolated On- and Off-bipolar cell contributions to the ERG, the inner retinal responses must be blocked before APB or PDA are applied. 

The results for one animal after APB alone (Fig. 7A) show the contribution from On-pathway (right) and Off-pathway plus photoreceptor (black traces, middle) in response-density plots. One obvious effect of APB was that it removed almost all the OPs, as can be seen both in the summed ERGs (top) and the extracted OPs (bottom). This result is particularly well illustrated by the superposed records in the right column in the bottom row of Figure 7A , where the control OPs, shown in gray, and the OPs removed by APB, in black, overlap almost completely. The nearly total removal of OPs by APB was similar in all three animals in which APB was injected first. The effect for one of the other animals (DK) is shown in Figure 7B , for the mfERG and extracted OPs summed over all the hexagons. In addition, APB, similar to TTX (compare Fig. 7B ring 1 and Fig. 4B , ring 1), removed a negative-going wave after the b-wave with the timing of the PhNR, as well as some oscillations that were of lower frequency than the filtered OPs. 

APB also removed a large positive potential that formed the leading edge of the b-wave in the normal records (Fig. 7A , top), and revealed a late negative peak of the a-wave, followed by a positive-going potential, presumably reflecting the depolarization of Off-bipolar cells. 

The positive peak of the responses of On-pathway (removed by APB) and Off-pathway (post-APB) were delayed in central retina versus peripheral retina by about the same amount (i.e., ∼6 ms; Table 1 ) when compared with the positive peak when only the inner retina was removed (Fig. 6) . Also, within each ring, the positive peak of the Off-pathway occurred significantly later (P < 0.05) than the positive peak of the On-pathway, and this difference, approximately 7 ms (see Table 1 ), was similar in all rings (only rings 1 and 5 reported). 

Isolation of Bipolar and Photoreceptor Contributions

The logic for isolating bipolar cell responses is illustrated in Figure 8A . To isolate the photoreceptor and bipolar cell contributions to the slow-sequence mfERG, we first blocked the inner retinal responses either with GABA (Fig. 8B , animal VA) or with TTX+NMDA (Figs. 8C 8D , animal DI). This left the photoreceptors and bipolar cells responsive to light. After inner retinal blockade (records shown by thin green lines), APB was applied to block the On-bipolar cell contribution, the remaining waveform, which represents contribution from photoreceptor, Off-bipolar cell and possibly horizontal cells is shown by red traces. Finally, PDA was injected to isolate the cone photoreceptor response. 

Similar to the observation in the previous section after APB alone, the a-wave of the post-APB records in these experiments, where inner retinal contributions already were removed, showed a later peak than in control records. In fact, there appeared to be two phases in the negative components (Fig. 8C , left, two arrows). The origin of these different phases was clarified by injecting PDA to isolate the photoreceptor contribution. The isolated photoreceptor contributions are shown for three animals in Figure 8D expressed in response density (first column) and sum of groups (second column). The solid black line in Figure 8D (third column) is the average of the three (sum of groups), and this average curve appears in Figure 8C (first column), expressed as response density, as well. The figure shows that for our stimulus conditions, the photoreceptors contributed very little to the a-wave. The figures also show that the late-phase peak revealed by APB was determined almost entirely by PDA-sensitive postreceptoral contributions presumably from Off-bipolar cells, although horizontal cell contributions cannot be ruled out. 

The timing of the positive peaks from On- and Off-bipolar cell contributions showed variations with eccentricity similar to those reported above for other conditions. The positive peaks were more delayed in the center than in the periphery for both On- and Off-bipolar cell contributions (Table 1) , and these differences were not significantly different from the values obtained with APB alone (i.e., without inner retinal blockade) reported in the previous section. 

The peak hyperpolarization of the Off-bipolar cell, i.e., the negative peak of the a-wave of the post-APB record (red traces), occurred significantly (P < 0.05) earlier than the depolarization (positive) peak of the On-bipolar cells of APB-isolated responses (blue traces) at all eccentricities (ring 1 = 3.6 ± 2.4 ms and ring 5 = 3.8 ± 0.9 ms). Also, the On-bipolar cell peak depolarization to light occurred significantly earlier (P < 0.05) than the Off-bipolar cell depolarization peak at all eccentricities (4.2 ± 2.2 ms earlier for ring 1 and 3.9 ± 1.3 ms for ring 5). 

Human Slow-Sequence mfERGs

Figure 9 shows human slow-sequence mfERGs, with the same stimulus conditions as we used for the monkeys. The average responses of five normal subjects are shown in the left columns and the results for one individual, in the right columns. The human summed mfERG (Fig. 9A , top row) bears a general resemblance to the monkey’s (Fig. 4 , top row) including a qualitatively similar OP contribution. There are, however, four notable differences between the human and monkey data. First, the human OPs are smaller. This is especially noticeable in the outer rings (compare lower panels of Figs. 9B and 4B ). Second, whereas the monkey’s OPs were largest in the central ring in the response density plots, the human OPs, as previously reported, were largest parafoveally. ²⁰ Third, whereas the timing of the monkey OPs (Fig. 4B , lower panel) from rings 1 and 2 markedly differ from those from rings 4 and 5, there is less difference in timing with eccentricity in the case of the human OPs (Fig. 9B , lower panels). Fourth, there is less nasotemporal variation in the human OPs (compare Fig. 9C , lower panel, with Fig. 5A , middle). When the human responses were grouped into nasal and temporal chevrons, there was no significant difference between N1 and T1 OP RMS (P = 0.61), perhaps due to the small amplitudes. OPs in N2 and T2 were so small that we did not compare them. However, we were able in individual cases to see these variations when we recorded slow-sequence mfERGs, interleaving 7 blank frames instead of 14, so that more trials could be averaged in the same 7-minute recording. 

Similar to the observations in control monkey mfERGs, in the human recordings the peak time in the central retina was later than that from the peripheral retina. This is illustrated by the dashed line in Figure 9B (top) drawn through the b-wave peak in the central ring and extended through all the rings. Measurements of peak times, when the OPs were removed by filtering, were approximately 2.5 to 3 ms slower in ring 1, where it was 31.5 ± 0.8 ms and than in ring 5 where it was 29.1 ± 0.7 ms (P < 0.01), and the summed mfERG time was similar to that in ring 5, 29.1 ± 0.6 ms. 

Discussion

Our results show that the primate photopic ERG varied in waveform with eccentricity. These variations were largely due to the following factors: (1) The amacrine and ganglion cell contributions, which are mainly in the form of OPs, were larger in the central and temporal retina, where they differed in waveform from OPs in other portions of the retina and (2) responses both from On- and Off-pathways and from isolated bipolar cells showed slower times to peak in the central retina than in the more peripheral retina. 

Oscillatory Potentials

Wu and Sutter ³⁴ observed OPs superimposed on the b-wave of the slow-sequence mfERG in humans by interleaving just three blank frames in the multifocal stimulus sequence. Although the initial negative waves are comparable for the mfERG and full-field ERG with three interleaving frames, the positive waveforms are not similar, with less than seven blank interleaving frames. ²⁴ In macaques, in the records with 14 blank frames, the OPs were even slightly more prominent than with 7 blank frames (data not shown for the 7-blank-frame condition). 

As in previous focal and mfERG studies in humans, we found that temporal retinal OPs were larger than nasal retinal OPs. ^{20 34} In fact in our monkeys, the nasotemporal asymmetry was quite pronounced. Sutter and Bearse ⁴⁰ have suggested that there is a nasotemporal asymmetry in the human mfERG, owing to contributions from an optic nerve head component (ONHC). They have developed an algorithm that allows them to separate the human mfERG into an optic nerve head component (ONHC) that increases in latency with increasing distance from the optic nerve head and a locally generated retinal component (RC) with latency that is not related to distance from the nerve head. Recent studies in humans provide support for the ONHC’s involvement in the nasotemporal asymmetry of the OPs. ^{36 41} Using different paradigms to elicit OPs, both studies proposed that the nasotemporal asymmetry is due to the temporal alignment of the ONHC and the RC, causing an augmentation of the OP amplitudes in the temporal retina, and a reduction in amplitude due to cancellation of the components in the nasal retina. 

Our present results in monkeys after TTX also are consistent with the involvement of the ONHC in producing OP asymmetries. TTX reduced the nasotemporal asymmetry, and it did so by reducing the temporal OPs more than the nasal OPs, perhaps due to the absence of spiking activity in the ganglion cell fibers. This interpretation gains support from another study involving some of the same animals, in which TTX reduced or eliminated the ONHC in the fast-sequence mfERG, where it is normally quite large in these rhesus macaques. ³⁰ The finding in the current study that the human showed less nasotemporal variation than the monkey is consistent with the conclusion that the human ONHC extracted from mfERG recordings is relatively smaller than the monkey OHNC. At present, there are no obvious physiological or anatomic explanations for this species difference, although a higher density of ganglion cells has been observed in the macaque macula than in the human. ^{13 15} The photopic negative response, also thought to be generated at the optic nerve head, has been found to be of similar amplitude in the two species in response to full-field flashed stimuli. ^{4 42} We had insufficient data in the present study to make a comparison of the photopic negative response in the slow sequence mfERG. 

Because some amacrine cells and interplexiform cells also produce spiking activity, we cannot attribute the effects of TTX in the present study exclusively to ganglion cells. However, preliminary slow-sequence mfERG data from monkeys with experimental glaucoma indicate a reduction of the nasotemporal asymmetry as visual field defects progress (Rangaswamy NV, Frishman LJ, Saszik SM, Hood DC, Harwerth RS, ARVO Abstract 2170, 2002). Thus, it is important to examine effects of ganglion cell loss in this model or by nerve transection to determine the contribution of ganglion cells and ONHC to the generation of OPs. 

More generally with regard to OPs in the flash ERG, it is important to remember that their origins are not well understood, although they are generally thought to come more from amacrine than ganglion cells ^{9 43 44 45} and from both rod and cone pathways. ^{46 47 48} The findings in the present study, and our preliminary results from animals with experimental glaucoma, are for very specific stimulus conditions, fully photopic, brief flashes, of moderate contrast recurring every 200 ms, and not including retina more peripheral than 33° diameter of the visual field. Furthermore, the amplitude of the OPs was quite sensitive to small variations in stimulus conditions. When we reduced the contrast for three animals, so that the increment was 1.7 cd-s/m², we found that OPs were smaller in the summed mfERG, mainly because of small OPs in the peripheral rings. 

Central Versus Peripheral Timing of the Flash ERG

The most obvious type of variation in the waveform of the photopic ERG in this study was in the form of timing differences at different eccentricities. The OPs, as well as the On- and Off-pathway contributions, were slower in the central retina than in the more peripheral retina. One possible explanation for this timing difference is the longer length of the cone axons in the central retina, known as the Henle’s fibers, than in the peripheral retina. ^{49 50 51} Because of the longer fiber in the central retina, the conduction from the cones to the bipolar cells would be slower than in the periphery. Hsu et al. ⁵² calculated the time for cone signals to reach their terminal from outer segment for a “long” cone axon of 380 μm to be approximately 2.5 ms. Direct measurements of Henle’s fibers in the foveal region found lengths to range from 400 to 600 μm, ^{49 50 51} whereas they were only approximately 50 μm at 2 mm from the fovea. ⁵¹ In the present study, the outermost ring (ring 5) was approximately 4.5 mm from the fovea, which would further shorten the fibers and consequently would increase the delays to peaks of the ERG. We found a difference in the time to peak between rings 1 (foveal) and 5 (33° peripheral, 4.5 mm radius) of approximately 6 ms; estimates based on Henle’s fiber lengths cited in literature could account for at least 4 ms of this difference. 

The longer times to peak in central retina may also be due in part to differences in relative densities of bipolar cells types, midget versus diffuse, in the central versus peripheral retina. There are relatively more midget ganglion cells in foveal regions than in peripheral retina, ^{18 53} and presumably more midget bipolar cells to handle single cone signal transfer to those ganglion cells. ¹⁶ If the midget cells have slower kinetics in the central retina, then this could result in a slower times to peak in the central retina. 

On- Versus Off-Pathway Timing

As in previous studies, ³ we also observed differences in the timing of the On- and Off-bipolar cell contributions to the photopic ERG. The time to negative peak, probably originating from the hyperpolarization of Off-bipolar cells was earlier than the time to peak of the On-bipolar cell (or On-pathway) contribution at all eccentricities. As proposed by Sieving et al., ³ such differences in timing would be expected because of the differences in the glutamate receptors on the two types of bipolar cells. The extra biochemical steps involved in the cascade of the On-bipolar cells with the metabotropic receptors, compared with the directly gated ionic currents underlying the Off-bipolar response, would result in slower kinetics of the On-bipolar cell response (e.g., Ref. ⁵⁴ ). 

Photoreceptor Contribution

It has been shown using pharmacological agents with full-field ERG that the cone-photoreceptor contributes to the a-wave when the stimulus is strong, whereas for weaker photopic stimuli the a-wave is mainly formed by postreceptoral inputs. ¹ Robson et al. ² have further demonstrated that the postreceptoral portion of the response is a slow-sloping wave. Looking at the records after PDA in Figures 8C and 8D it can be said that for the stimulus strengths that we were using, the a-wave had contributions from at least two generators: the cone-photoreceptors that contribute a small portion of the response and the Off-bipolar cells that contribute to the slower sloping portion of the response, forming most of the late negative peak. 

Working Model of the Human Photopic ERG

In Figure 10 , we illustrate a working model of the human photopic flash ERG based on data obtained from macaques after injecting pharmacological agents. Figure 10A shows response-density scaled (ring 1 and ring 5 responses) for the slow-sequence mfERG in animal VA after GABA to remove inner retinal contributions, the On-bipolar cell responses that were removed with APB, the Off-bipolar cell responses, and the cone photoreceptor responses. We decided to use monkey data after removal of inner retinal signals for comparison with human data because inner retinal contributions were relatively smaller in humans than in the rhesus macaques that we studied (e.g., Ref. ²⁷ ). 

In Figure 10B , we show, for density-scaled traces, how the On- and Off-bipolar cell contributions might shape the human slow-sequence mfERG. The initial a-wave is formed by cone photoreceptor and Off-bipolar hyperpolarization with the cones contributing a relatively small influence. The rising limb of the b-wave is formed by the recovery of Off-bipolar cells and also by the depolarization of the On-bipolar cells. The peak of the b-wave forms between the peak depolarization of the Off- and On-bipolar cells. The descending limb of the b-wave is formed mainly by the recovery of the On-bipolar cells. Finally, although not shown in Figure 10 , the inner retina, of course, contributes the OPs. 

In a similar study of the fast-sequence mfERG in many of the same subjects, comparisons of monkey and human recordings were made after removing the monkey’s relatively larger inner retinal contributions to the mfERG. ²⁷ Despite the differences in waveform due to the sequence of presentation, the general findings were essentially the same in the two studies, both with respect to the close resemblance between the macaque and human responses and to the relative timing of On and Off bipolar cell contributions to the macaque mfERG. 

In conclusion, in our study, the primate photopic ERG varied with increasing distances from the fovea. The waveforms were largely shaped by the overlapping contributions from On- and Off-pathways. Inner retinal contributions were mainly of an oscillatory nature and showed a nasotemporal asymmetry. The variations at different retinal eccentricity are largely in the form of timing differences. Because the peripheral rings were larger in area, they generated larger responses, and these responses dominated the timing of the peaks in the slow-sequence mfERG summed over the entire array, suggesting that the timing of the major waves of full-field photopic flash ERG also is determined by the peripheral retina. 

 Supported by National Eye Institute Grants R01-EY09076, R01-EY06671, P30-EY07751.

 Submitted for publication January 4, 2003; revised February 10, 2003; accepted February 17, 2003.

 Disclosure: N.V. Rangaswamy, None; D.C. Hood, None; L.J. Frishman, 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) Unscaled stimulus array used in the multifocal recordings. The region indicated by the oval corresponds to a visual field extent of 33° by 35°. (+) Marks the location of the fovea. (B) Multifocal records obtained from the left eye of one of our control monkeys (monkey A) using the slow-sequence paradigm (14 interleaved blank frames [run # sm420]). The region marked by the oval shows the area of the retina where oscillations were larger, which corresponds to the central and temporal retina of the left eye. (C) Grouping of the responses in rings at different eccentricities. Ring 1 includes the responses in the central 7 hexagons. (D) Grouping of the responses from the central three rows of hexagons where OPs were largest into chevrons. Responses were grouped into chevrons mainly to study the nasotemporal asymmetry in OPs. (X) Corresponds to the location of the blind spot in the visual field of the left eye.

Figure 2.

 

Stimulus presentation for the slow-sequence mfERG used in this study with 14 interleaved blank frames (interflash interval of approximately 195 ms). The flash energy was 4.7 cd-s/m² and the background luminance was 20 cd/m². In any one m-frame, the hexagons have a 50% chance of being at the background luminance or increasing to a luminance of 4.7 cd-s/m².

Figure 3.

 

(A) Slow-sequence mfERG responses summed over the whole field for two control monkey eyes (VA [OS, sm335] and VI [OD, sm378], top) and the OPs extracted from these records (bottom). (B) Full-field flash ERG from one of the control monkeys shown in (A) (VI [OS, sm274]) shown for the same time interval as in (A), along with the extracted OPs for comparison.

Figure 4.

 

(A) Summed slow-sequence mfERG responses for three control monkey eyes (M [OS, sm183], VA [OS, sm335], and T [OD, sm336]). (B). Top: slow-sequence mfERG (represented as sum of groups) grouped by rings (see Fig. 1C ) for the same animals as shown in (A). Gray traces: records obtained after filtering the OPs superimposed on the control records. Bottom: Extracted OPs for the corresponding mfERG records (top).

Table 1.

 

Timing of b-Waves in Rings 1 and 5

	Ring 1					Ring 5					Difference (Ring 1 − Ring 5)	Peak (Summed mfERG)
		Dep. Peak	Hyper. Peak	Difference (D-H)	Dep. Peak	Hyper. Peak	Difference (D-H)
Control (filtered; n = 16)	32.6 ± 2.3			28.0 ± 1.5			4.6 ± 2.6	29.0 ± 1.7
Inner-retinal blockade (n = 9)	35.6 ± 2.9	—	—	29.8 ± 1.7	—	—	5.8 ± 1.6	30.3 ± 3.5
APB-sensitive (On- pathway; n = 3)	30.8 ± 1.7	—	—	24.7 ± 0.9	—	—	6.1 ± 1.9	27.5 ± 0.8
Post APB (Off-pathway; n = 3)	38.0 ± 0.5	29.2 ± 1.7	8.9 ± 2.1	31.9 ± 1.7	22.2 ± 1.3	9.7 ± 2.7	6.1 ± 1.3	35.8 ± 0.9
IR blockade+ APB-sensitive (On-bipolar; n = 3)	33.1 ± 2.7	—	—	26.7 ± 0.1	—	—	6.4 ± 2.7	26.9 ± 0.5
IR blockade+ post-APB (Off-bipolar; n = 3)	37.2 ± 0.9	28.9 ± 1.0	8.4 ± 1.7	30.6 ± 1.3	23.1 ± 1.0	7.5 ± 1.5	6.7 ± 1.4	33.1 ± 1.9

Figure 5.

 

Effect of inner retinal blockade on the OPs (response-density scaled) (A) Left: slow-sequence mfERG responses grouped by chevrons (see Fig. 1D ) for monkeys M (OS, sm183), VI (OD, sm378), and VA (OS, sm335) before injection of pharmacological agents. N1 and N2 are records from the nasal retina, F from the fovea, and T1 and T2 from the temporal retina. N1 and T1 are closer to the fovea than N2 and T2. Middle: Extracted OPs for the three corresponding control traces. Right: Extracted OPs from the same animals after injection of different pharmacological agents. Monkey M was injected with TTX to block sodium-dependent action potentials from the inner retina (retinal ganglion cells and some amacrine cells), monkey VI was injected with TTX and NMDA to block nonspiking activity from the inner retina as well, and monkey VA was injected with GABA to block all inner retinal activity. (B) Comparison of the average nasal (N1) and temporal (T1) OP RMS (calculated response-density–scaled records) in the control and after injection of intravitreal injection of pharmacological agents to block inner retinal activity. The RMS was calculated for the time interval between 10 and 60 ms as represented by the two dashed lines in the lower row of (A). Black: N1 OP RMS; light gray: T1 OP RMS; dark gray: average OP RMS noise, calculated as an average of the RMS (60–80 ms) for all four chevrons, N2, N1, T1, and T2, and the foveal group (F). Error bars, SE. For the animals treated with TTX+NMDA+PTX and GABA, the RMS is shown individually. TN, TTX+NMDA; TNP, TTX+NMDA+PTX.

Figure 6.

 

Effect of inner retinal blockade on the waveform of the slow-sequence mfERG (response-density scaled) (A). Left: slow-sequence mfERGs grouped by rings from monkey E (OS, sm184), before and after intravitreal injection of TTX. Right: difference between before and after TTX injection representing the portion of the waveform removed by TTX. (B) Representative slow-sequence mfERG responses from three animals before and after intravitreal injection of different pharmacological agents to block inner retinal activity. Left: monkey VI, before and after TTX+NMDA (OD, sm378); middle: monkey Z, before and after TTX+NMDA+PTX (OS, sm334); monkey VA, before and after GABA (OS, sm335). Dashed line in the records from monkey VA corresponds to the time to peak of the central records and extendingthrough all rings, after block of inner retinal activity with GABA. The gray traces were recorded before and the black traces after drug injection.

Figure 7.

 

On- and Off-pathway contributions to the slow-sequence mfERG. (A) Slow-sequence mfERGs from monkey T (OS, sm282) grouped by rings (response-density scaled): mfERGs (top) and the corresponding extracted OPs (bottom). Black traces in the left column (gray in the middle column) show the responses before drug injection. Black traces in the middle columns (both top and bottom) were recorded after APB, representing the Off-pathway contribution. The thick black traces in the right column (both top and bottom) are the components removed by APB, representing the On-pathway contributions. The two arrows in the top middle column show the two phases of the a-wave, which have different slopes. (B) Slow-sequence mfERGs grouped over the entire field from another animal, in which OPs were large in the summed mfERG, monkey DK (OS, sm429) before and after APB alone, with extracted OPs. Note that the time to positive peak of the On-pathway (right) occurred earlier than the time to positive peak of the Off-pathway (middle). Also, it is obvious that the summed OPs were mainly contributed by the On-pathway.

Figure 8.

 

Isolation of bipolar and photoreceptor contribution in the slow-sequence mfERG. (A) Illustration of the logic behind the isolation of the contributions from On-bipolar cells after administration of APB, which blocks the photoreceptor On-bipolar synapse. Application of APB to a retina in which inner retinal responses have been suppressed produced responses (red) containing only photoreceptor, Off-bipolar, and possibly horizontal cell contributions. Subtracting these responses from the control condition yielded an estimate of the On-bipolar contribution (blue). (B) Slow-sequence mfERGs grouped by rings (response-density scaled) from monkey VA (OS, sm335) after injection of APB (red), in which the inner retina was blocked with GABA (green). Blue traces in the second column show the isolated On-bipolar contribution removed by APB obtained by subtracting the red traces from the green traces. (C) Same as in (B) (response-density scaled) from animal DI (OD, sm196), in which the inner retina was blocked with TTX+NMDA (green). Solid black trace in the left column is the average record of three animals after PDA (DI [OD, sm196], DK [OS, sm429], and VI [OS, sm284]), showing the isolated photoreceptor component. Arrows: the two phases of the a-wave, which have different slopes. (D) To understand the absolute contribution of the photoreceptors at different eccentricity to the overall photopic ERG, the individual response-density–scaled (left) and sum of groups (middle) records are shown for the three animals (DI, DK, and VI) after PDA. Right: records from animal DI after inner retinal blockade with TTX+NMDA (green) and after APB (red) as sum of groups for rings 1 and 5. The black traces are the average isolated photoreceptor component after PDA from three animals (DI, DK, and VI) and are shown in the middle column as sum of groups.

Figure 9.

 

Human control slow-sequence mfERGs. (A) Slow-sequence mfERG (14 blank frames) summed over the entire field (top) from an individual human control subject (right) and average of five human control subjects (left) along with the corresponding extracted OPs (bottom). The vertical scales are different from that used with control monkeys but the same time scales are used. (B) Slow-sequence mfERGs grouped by rings (top) from an individual (right) and average (left) human control subjects along with their respective extracted OPs (bottom) represented as sum of groups. (C) Slow-sequence mfERGs grouped as chevrons (top) from individual (right) and average (left) along with extracted OPs (bottom) from human control subjects represented as response-density–scaled records.

Figure 10.

 

Working model for the human slow-sequence mfERG. (A) Slow-sequence mfERG from rings 1 and 5 after injection of GABA in monkey VA (black traces); isolated On-bipolar cell responses (thin gray traces); Off-bipolar cells, photoreceptors, and possibly horizontal cell responses (thick gray traces); and average isolated photoreceptor contribution for rings 1 and 5 (dashed traces). (B) A model of the contributions to the human slow-sequence mfERG based on the results obtained in the monkey in (A). The records are all scaled to have the same peak-to-trough amplitude. On BC, On-bipolar cell; Off BC, Off-Bipolar cell; PR, photoreceptor.

The authors thank Suresh Viswanathan and Shannon M. Saszik for assistance with the experiments.