The contribution of the different retinal elements to the fast

Transcription

1 Primate Photopic Sine-Wave Flicker ERG: Vector Modeling Analysis of Component Origins Using Glutamate Analogs Mineo Kondo and Paul A. Sieving PURPOSE. To study how the photoreceptoral and postreceptoral ON- and OFF-components contribute to the photopic sinewave flicker ERG in the monkey by isolating the components with glutamate analogs. METHODS. Monkey photopic flicker ERGs were elicited with sine wave stimuli (mean luminance, 2.66 log cd/m 2 ; 80% modulation depth, on a 40 cd/m 2 white background) and were recorded for stimulus frequencies of 4 Hz to 64 Hz, before and after intravitreal injection of DL-2-amino-4-phosphonobutyric acid (APB) and cis-2,3-piperidinedicarboxylic acid (PDA) that block ON- and OFF-bipolar activity, respectively. The amplitude and phase of the fundamental component were analyzed. RESULTS. The flicker response amplitudes increased after APB, for frequencies of 6 Hz to 32 Hz. The further addition of PDA to isolate the photoreceptor component resulted in a relatively small residual response that decreased monotonically from 4 Hz to 32 Hz. The postsynaptic APB (ON-) and PDA (OFF-) sensitive components were isolated by subtraction and were characterized by amplitude and phase vectors. The ON- and OFF-components were larger than the initial control responses for stimuli of 8 Hz to 40 Hz. These two components had a frequency-dependent phase difference of 160 to 230 ; normally, they interfere with each other and reduce their net contribution. The phase difference between ON- and OFFcomponents was nearly 180 for a 10-Hz stimulus, and the phase cancellation caused a prominent dip in amplitude at this frequency. CONCLUSIONS. These results indicate that postreceptoral ONand OFF-components contribute substantially to the sine-wave flicker ERG, especially at higher stimulus frequencies. Because of phase cancellation, they mask each other in the net response in a frequency dependent fashion. The photoreceptor contribution is greater than the net postsynaptic component only for frequencies of approximately less than or equal to 10 Hz. These results can be summarized by a vector model that may be useful for interpreting changes resulting from retinal disease. (Invest Ophthalmol Vis Sci. 2001;42: ) The contribution of the different retinal elements to the fast flicker ERG has not been determined precisely in the primate. Earlier studies indicated an outer retinal locus for the fast flicker ERG signals. 1 7 Baron, Boynton, and colleagues 3 5 From the Department of Ophthalmology and Visual Sciences, University of Michigan, Ann Arbor. Supported by NIH Grant R01-EY06094 and Vision CORE Grant EY07003; The Foundation Fighting Blindness, Hunt Valley, Maryland; Research to Prevent Blindness, Inc., New York; Alcon Research Institute Award. Submitted for publication May 22, 2000; revised August 31, 2000; accepted September 22, Commercial relationships policy: N. Corresponding author: Paul A. Sieving, Center for Retinal and Macular Degeneration, The W. K. Kellogg Eye Center, University of Michigan, 1000 Wall Street, Ann Arbor, MI psieving@umich.edu used sine-wave stimuli and recorded intraretinal local ERGs with a microelectrode inserted into the cone-rich fovea of the primate. They reported that the response characteristics of the local flicker ERG did not change much after restricting the activity to the photoreceptors by applying sodium aspartate. In a later study, Donovan and Baron 2 compared ERG responses recorded globally at the cornea to intraretinal local ERG responses; they concluded that for sinusoidal stimuli of 4 Hz to 12 Hz, the corneal ERG originated from the same retinal cells as the foveal local ERG, that is, the cone photoreceptors. These results encouraged others to study the flicker response with sine wave stimuli with the assumption that this procedure yielded photoreceptor-enriched signals. 6 8 However, Donovan and Baron 2 found a disparity between intraretinal local ERG (LERG) responses and the corneal flicker response, which led them to wonder whether additional components might be contributing to the corneal flicker ERG beyond the intraretinal receptor potentials. Chang et al. 9 also suggest the possibility of postreceptoral contribution to the sine-wave flicker ERGs. It is known that the retinal response to sine-wave flicker stimuli is dominated by a component at the same temporal frequency as the stimulus, termed the fundamental component. However, under some conditions, a component is seen at twice the stimulus frequency (i.e., a second harmonic component) Current source density analysis 14 suggested that the fundamental component was dominated by photoreceptor activity, yet the second harmonic component originated from multiple retinal sources and included contributions from the middle and inner retina. A systems analytic approach in humans 12,13 also agreed with a model of the sine-wave flicker ERG that attributed the majority of the linear (or fundamental) component to photoreceptor-related processes, particularly at medium and high frequencies. Some clinical studies 15,16 have used the fundamental and second harmonic component of the sine-wave flicker ERG as an indicator of outer and inner retina function, respectively. Previously, the authors studied the photopic fast flicker ERG using square waves and photostrobe flashes 17 and found that the corneal response was virtually eliminated by aspartate or by the combination of DL-2-amino-4-phosphonobutyric acid (APB) 18 plus cis-2,3-piperidinedicarboxylic acid (PDA). 19 These drugs block transmission to the second-order neurons postsynaptic to cones. 20 This implied a significant contribution from the inner retina to the corneally recorded photopic fast flicker ERG. What was curious was the apparent difference on flicker origins reported for sine waves versus findings in this study using brief-flash or square stimuli. The authors previous study, however, did not employ sine-wave stimuli and was limited to a 33.3-Hz flicker. This study revisited the question of the origins of the primate flicker ERG and recorded sine-wave flicker responses from monkeys over a wide range of temporal frequencies on a photopic background before and after applying glutamate analogs. Particular attention was given to the nature of the fundamental frequency component. Photoreceptoral and postreceptoral ON- and OFF-components were obtained by Investigative Ophthalmology & Visual Science, January 2001, Vol. 42, No. 1 Copyright Association for Research in Vision and Ophthalmology 305

2 306 Kondo and Sieving IOVS, January 2001, Vol. 42, No. 1 waveform subtractions. The amplitude and phase of these separated components were plotted and characterized by amplitude and phase vectors. The results indicated that postreceptoral neural activity contributes substantially to the fundamental component of the sine-wave flicker ERG. The magnitude of the relative contributions depended on the stimulus frequency. Results of this study also showed that the relative phase between postsynaptic ON- and OFF-components dramatically affects the flicker ERG, and that phase cancellation of these two components creates an amplitude dip near 10 Hz that is normally observed in the primate sine-wave flicker ERG. MATERIALS AND METHODS Animal Preparation Experiments were conducted in accordance with NIH guidelines on animal use and with the ARVO Statement on the Use of Animals in Ophthalmic and Vision Research. Three rhesus (Macaca mulata) monkeys were used in this study. The animals were sedated with ketamine hydrochloride (7 mg/kg IM; 5 to 10 mg/kg/h IM maintenance dose) and xylazine (0.6 mg/kg IM). Supplemental oxygen was given by nasal canula during the entire anesthesia period. Respiration and heart rate were monitored, and hydration was maintained with continuous, slow s.c. infusion of lactated Ringer solution. Corneal anesthesia (topical proparacain 0.5%) and full pupil dilation (topical phenylephrine HCl 2.5% and atropine 1%) were used. Drug Application The drugs and intravitreal injection techniques have been described previously. 21 Drugs were injected into the vitreous by a 30-gauge needle inserted through the pars plana approximately 5 mm to 6 mm posterior to the limbus. Drugs (Sigma Chemical Co., St. Louis, MO) were dissolved in sterile saline and injected in amounts of 0.05 ml to 0.07 ml of 40 mm APB or 200 mm PDA. Drug effects were monitored by tracking the photopic b-wave after APB and the d-wave after PDA. Recordings were started after full drug effects were seen, normally approximately 60 minutes to 90 minutes after drug injections, and studies were completed within 5 hours thereafter. Although these drug effects are mostly reversible after a recovery period of several weeks, the results shown were recorded from three eyes not treated previously. Visual Stimulation Sine-wave stimuli were produced by a inch diameter, densely packed array of 102 red LEDs (623 nm peak wavelength; 8 nm halfwidth) that was positioned 10 cm from the eye. This illuminated a ping-pong ball hemisphere (40 mm in diameter) that was placed immediately in front of the eye to give wide-field stimulation. Sinewave modulation of the red LEDs was controlled by a digital function generator (model 39; Wavetek, San Diego, CA), which fed into a linear power amplifier. Light output was determined as a function of input voltage and was found to be highly linear at the light levels used in our study. The sinusoidal light modulation was symmetrical, and the nonlinearity attributable to the light output under these conditions was measured as less than 2%. The maximal and minimal stimulus intensities were 2.90 log cd/m 2 and 1.70 log cd/m 2 on a constant white background of 40 cd/m 2. The net result was 2.66 log cd/m 2 mean luminance and 80% modulation depth. Recording and Analysis After an initial 10 minutes of light adaptation at 40 cd/m 2, flicker ERGs were recorded using a Burian Allen bipolar contact lens electrode (Hansen Ophthalmic Development Laboratories, Iowa City, IA). A ground electrode was placed on the ipsilateral ear. Responses were amplified with a band-pass of 0.1 Hz to 1000 Hz at 3 db, and digitized at a 10 khz-rate. The system incorporates a narrow-band, 60-Hz line FIGURE 1. Monkey photopic ERG to single flashes before and after intravitreal application of glutamate analogs. The stimulus was a 200- millisecond square-wave pulse of 623 nm red light from the red LEDs that gave 2.90 log cd/m 2 on a 40 cd/m 2 white background. frequency analog notch filter. In preliminary studies, it was first confirmed that the analog notch filter did not affect either amplitude or phase of the flicker ERG fundamental component for the stimulus frequencies that were used. Fifty to 100 responses were averaged for each condition. The amplitude and phase of the fundamental components were analyzed using Fourier coefficients of the response. 22,23 The phase lag relative to the stimulus sine wave was presented on polar plots with positive values drawn counterclockwise from the 0 polar axis, as done in earlier studies. 23,24 Because Fourier analysis gives phases only within a 0 to 360 range, the authors extrapolated absolute response phases beyond this limit by comparison with phases of adjacent temporal stimulus frequencies. RESULTS Single Flash Responses The effects of the APB and PDA on the ERG are illustrated in Figure 1 for a 200-millisecond flash. These responses were recorded on one animal during the same session as the sinewave flicker ERG recordings shown in Figure 2. APB injection eliminated nearly 95% of the b-wave but had no effect on the a-wave amplitude and timing; the d-wave was enhanced at stimulus termination. The subsequent addition of PDA resulted in some decrease of the a-wave amplitude; this effect was noted previously and is thought to indicate a contribution of OFF-bipolar cell activity to the a-wave. 25 Sine-Wave Flicker ERGs before and after Glutamate Analogs The left column of Figure 2 shows control flicker ERG elicited at different stimulus frequencies before drug application. The shape of the responses is not sinusoidal for all stimulus frequencies. The responses show partial frequency-doubling for stimuli near 10 Hz and 48 Hz, indicating a contribution of the second harmonic component at these frequencies. The middle column shows that the waveform after APB is simplified and becomes more sinusoidal; the amplitude becomes larger for stimuli of 10 Hz to 32 Hz. Because APB suppresses transmission from photoreceptors to depolarizing bipolar cells rather selectively, this indicates that the ONbipolar cells contribute to the control sine-wave flicker ERG

3 IOVS, January 2001, Vol. 42, No. 1 Primate Photopic Sine-Wave ERG 307 than 16 Hz. After APB PDA, the amplitude of fundamental component decreased monotonically with increasing frequency from 4 Hz to 32 Hz and was quite small at frequencies higher than 24 Hz. Assuming that APB PDA essentially eliminates signal transmission to cells postsynaptic to the photoreceptors, this result indicates that the control sine-wave ERG contains considerable contributions from postsynaptic neuronal activity, especially at higher stimulus frequencies. Note that even after isolating the photoreceptoral activity by APB PDA, the frequency amplitude curve did not show a simple low-pass function, but it had a small local maximum near 50 Hz that was consistently observed for all three monkeys. Component Analysis The next step was to determine how the postsynaptic ON- and OFF-components participated in the sine-wave flicker ERG of the normal eye. ON- and OFF-components were isolated by subtracting the postdrug records from the predrug records; the APB-sensitive (ON-) component was derived by subtracting the post-apb records from the controls, and the PDA-sensitive (OFF-) component was obtained by subtracting the post-(apb FIGURE 2. Photopic flicker ERGs to sine-wave stimuli (red LEDs at 623 nm; 2.66 log cd/m 2 mean luminance; 80% modulation depth). Numbers at far right indicate the trace duration in milliseconds. Hatched line is the sinusoidal light output as monitored by a photodiode. response. Subsequent addition of PDA (right-hand column) caused further changes to the response, with a considerable amplitude reduction for stimuli of 10 Hz and above. As the combination of APB PDA effectively blocks transmission from photoreceptors to all postsynaptic cell types, the responses remaining after APB PDA represent activity of the cone photoreceptors themselves. This cone cell activity contributes substantially to the responses at 4 Hz but to a progressively lesser degree for higher stimulus rates, and direct cone activity represents less than half of the control waveform for stimuli of 16 Hz and above. Figure 3 shows a compilation of the amplitude and phase of the fundamental component of these responses. The mean and SE bars are shown for three animals. The fundamental amplitude goes through a minimum at approximately 10 Hz, and is maximum at approximately 40 Hz to 48 Hz. The phase of the fundamental component changes smoothly across frequencies, but it has a small deflection at approximately 10 Hz. For higher frequencies, the phase increases rather linearly with stimulus frequency. These control ERG results agree well with previous reports of human sine-wave flicker ERG. 8,11 13 Response amplitudes after application of APB were significantly larger than controls (P 0.05) for frequencies of 6 Hz to 32 Hz and had a broad maximum around 20 Hz (Fig. 3, open circles). The amplitude dip near 10 Hz in the control was diminished greatly after APB. After APB, the phase of the remaining response became monotonic, unlike the control. With the further addition of PDA, the amplitude of the fundamental was greatly attenuated (Figure 3 top, crosses). The change was modest compared to controls at frequencies less than 10 Hz but was considerable at frequencies higher FIGURE 3. Plots of amplitude (top) and phase (bottom) of the fundamental component of the flicker ERG response to sinusoidal stimulation before and after glutamate analogs. Mean SEM is shown for three animals. Dashed line: control condition without drugs; open circles: after intravitreous injection of APB; crosses: after intravitreous injection of APB PDA.

4 308 Kondo and Sieving IOVS, January 2001, Vol. 42, No. 1 PDA) records from the post-apb records. The amplitude and phase of these components were plotted as a function of stimulus frequency in Figure 4. The mean and SEM are shown for three animals. The postsynaptic ON- and OFF-components had relatively large amplitudes for a wide range of frequencies of 8 Hz to 40 Hz. These two components behaved somewhat different from each other and had different frequency-tuning characteristics. The PDA-sensitive (OFF-) component amplitude had a maximum around 10 Hz to 30 Hz, yet the APBsensitive (ON-) component rose steadily up to 48 Hz and then fell. The ON- and OFF-components had a relatively large phase difference of 160 to 230 that was frequency dependent (Fig. 4, bottom). Thus, the ON- and OFF-components are normally partially out of phase and interfere with each other, thereby causing the net postsynaptic component to be smaller than either alone. The responses are maximally 180 out of phase near 10 Hz stimulation. The bold line (Fig. 4, top) shows the amplitude of net postsynaptic component, obtained by taking the fundamental of the difference of the (APB PDA) records from the control ERG. For frequencies less than 10 Hz, the photoreceptoral component was larger than the postsynaptic component. However, at stimulus frequencies of 16 Hz and above, the postreceptoral component became significantly larger (P 0.05) FIGURE 5. The constituent component vectors (amplitude and phase) for the fundamental frequency component of the 4 Hz, 10 Hz, and 32 Hz sine-wave flicker ERG. Red arrow: APB-sensitive (ON) component; blue arrows: PDA-sensitive (OFF) component; cyan arrow: postsynaptic component (ON OFF); black arrow: control ERG. Positive values of phase delay are plotted counterclockwise. Vector amplitudes are shown in absolute scale to the 10- V calibration bar. than the photoreceptoral component, and postreceptoral activity essentially dominates the response at frequencies greater than 24 Hz. At 24 Hz and above, amplitude of the photoreceptoral component was less than 20% of the postreceptoral component for all three monkeys. FIGURE 4. Plots of amplitude (top) and phase (bottom) of the fundamental component for isolated components. Mean SEM is shown for three animals. Dashed line: control conditions without drugs; filled circles: APB-sensitive (ON) component; open circles: PDA-sensitive (OFF) component; crosses: photoreceptoral component; bold line: postsynaptic component (ON OFF). Origins of the Fundamental Frequency Component Sine-Wave Flicker ERG Using Constituent Vectors Figure 5 shows the constituent phase and amplitude vectors for the fundamental frequency components of 4, 10, and 32 Hz sine-wave flicker responses. For these vector figures, we used the results from the representative monkey with waveforms shown in Figure 2. At 4 Hz, the amplitude of the ON-component (red arrow) is smaller than and 160 out of phase from the OFF-component (blue arrow); consequently, the postsynaptic activity vector (cyan arrow) is dominated by the OFF-component. This sums with the photoreceptor component (green arrow) that is shifted approximately 90. The net resultant ERG (black arrow) is slightly larger than either the photoreceptor or postsynaptic components alone and has an intermediate phase. Note that the photoreceptor activity provides a substantial contribution to the 4 Hz sine-wave flicker ERG (as was also

5 IOVS, January 2001, Vol. 42, No. 1 Primate Photopic Sine-Wave ERG 309 FIGURE 6. The vector model of the 32-Hz, sine-wave flicker ERG for postsynaptic dysfunction. All vectors are shown in absolute scale to the 10- V bar near the top right. Phase delay is counterclockwise. (A) Simulation of the ON-component delay. A 23 -delay of the ON-component (equivalent to 2 milliseconds delay at 32-Hz stimulation) causes a 7 -delay, but the amplitude increases by 48%. (B) Simulation of the ON-component reduction. A 50% reduction in the ON-vector causes a minimal change in the ERG amplitude but introduces a 35 -phase delay. (C) Simulation of the OFF-component delay. A 23 -delay of the OFF-component causes a 31 -phase delay and reduces the amplitude to approximately 1 2 of the control. (D) Simulation of the OFF-component reduction. A 50% reduction of the OFFcomponent amplitude speeds the ERG by 49 and reduces the amplitude to 83% of control. observed in Fig. 2), although the phase is advanced slightly ahead of the photoreceptor vector. At 10 Hz, the postsynaptic ON- and OFF-amplitudes are each larger than the underlying control amplitude (Fig. 3, top). However, the two vectors are close to 180 out of phase (174, 185, and 186 in the three monkeys in this study). This phase cancellation causes an amplitude cancellation in the net postsynaptic component, which consequently is smaller than the photoreceptor vector. The 10-Hz ERG again is dominated by the photoreceptor activity but at a phase that is advanced slightly ahead of the photoreceptor vector, similar to the 4 Hz result. However, the major cancellation of the postsynaptic elements caused a net reduction in the overall ERG amplitude, and the 10-Hz flicker is relatively small. At 32 Hz, the PDA-sensitive component and the APB-sensitive components both are relatively large, and the phase difference between them is 220 to 225. They add together to give a large postsynaptic vector that is nearly 180 out of phase with the small photoreceptor vector. Consequently, the flicker ERG at 32-Hz stimulation is dominated nearly entirely by postsynaptic activity. These examples at three different stimulus frequencies demonstrate the necessity of accounting for both phase and amplitude of the three main ingredient components of flicker ERG. The absolute sizes of the vectors alone are insufficient to describe the results, as illustrated at 10 Hz where the ON- and OFF-components are both large but are diametrically out of phase and nearly cancel each other, resulting in a considerable contribution from the much smaller photoreceptor vector to the resultant ERG. Vector Model of Flicker ERG for Postsynaptic Diseases This vector model of the flicker ERG led us to consider how postsynaptic dysfunction can change the amplitude and phase of the sine-wave flicker responses. The 32-Hz ERG was selected for this simulation, because many clinical laboratories use stimuli near 30 Hz to assess photopic flicker function in patients. Figure 6 shows the effects of altering the ON- or the OFFcomponent. The uppermost diagram in Figure 6 shows the normal result at 32-Hz stimulus frequency. Panels (A) and (B)

6 310 Kondo and Sieving IOVS, January 2001, Vol. 42, No. 1 show a phase delay or amplitude reduction in the ON-component, respectively. Abnormalities in the ON-pathway through the depolarizing bipolar cells result in a simulation of some forms of congenital stationery night blindness (CSNB). A 23 delay (equivalent to 2 milliseconds delay at 32 Hz stimulation) in the ON-pathway causes only a 7 delay (i.e., 0.6 millisecond timing change) of the ERG, but the amplitude increases by nearly 48%. A 50% reduction in the ON-vector causes a minimal change in the ERG amplitude but introduces a 35 -phase delay, equivalent to approximately 3 milliseconds. Panels (C) and (D) show the comparable results for phase or amplitude changes in the OFF-component. A 23 delay (2 milliseconds at 32 Hz) in the OFF-component causes a 31 slowing of the ERG and reduces the amplitude to approximately 1 2 of the control. A 50% amplitude reduction in the OFF-component amplitude speeds up the ERG by 49 (nearly 4 milliseconds) but reduces the amplitude by only 17%, to 83% of control. These results indicate that dysfunction of the postsynaptic components can change the flicker ERG response considerably in both amplitude and phase, and that disease could result in responses larger or faster than controls under some specific conditions. DISCUSSION Contribution of Postsynaptic Component to the Sine-Wave Flicker ERGs These results demonstrate that, in addition to the cone photoreceptors, activity of the postsynaptic neurons contributes to the fundamental frequency component of the photopic sinewave flicker ERG in the primate, particularly at higher stimulus rates. Results were depicted mainly from one monkey eye, but they were confirmed in two additional eyes, representing responses from a total of three different animals. Only minimal variations were seen, and the descriptions here are limited to the reproducible findings. The balance of contributions from the cones versus inner retinal neurons depended on stimulus frequency. At lower frequencies, the magnitude of the photoreceptor component was relatively larger, but for stimuli of approximately 24 Hz and above, the postreceptoral components became the major element in the overall response. These results indicate that postphotoreceptoral neuronal activity is critical for the fullfield sine-wave flicker ERG, particularly at higher stimulus frequencies. Thus the functional integrity of the inner retina is essential for developing normal fast flicker responses that are recorded at the cornea. This study s results indicate that reductions of the photopic fast flicker ERG should not necessarily be interpreted as indicating cone cell dysfunction, because flicker changes can also arise from postreceptoral dysfunction as shown in our simulation model. These results with sine-wave stimuli, along with previous flicker studies using square-wave and strobe-flash stimuli, 17 demonstrate that primate photopic flicker responses are dominated by proximal retinal activity at higher stimulus frequencies, including near 30 Hz, which is used commonly in assessing clinical patients. This holds true even for the fundamental component of the response to sine-wave stimulation and casts doubt on using the fundamental and second harmonic components to deduce photoreceptor versus inner retinal activity in the primate. 15,16 Baron and coworkers 2 5 previously studied the origins of the flicker ERG by intraretinal recordings with a microelectrode, and they found similarities between the local ERG (LERG) and the corneal ERG in phase and amplitude as a function of frequency when they recorded between 4 Hz to 12 Hz. 2 They concluded that the photoreceptors were the dominant origin of the flicker response. Findings in this study are not in disagreement, as the isolated photoreceptor contribution is greater than the postsynaptic component for frequencies below approximately 10 Hz. Baron et al. 2-5 also demonstrated that the response characteristics of LERG to sinusoidal flicker did not change much after isolation of the photoreceptors by application of aspartate. 3 However, they did not record simultaneously at the cornea. It has been shown previously that the flicker ERG at the cornea is essentially abolished after application of aspartate, 17 leading to the conclusion that the LERG and corneally recorded flicker ERG (at 30 Hz) have different origins. One qualifier is that Bush and Seiving 17 used a pseudosine-wave stimulus created by a mechanical shutter. A further example of the disparity between intraretinal and corneal recordings is found in the M-wave, 26 which is a large LERG response in the proximal retina but is minimally evident in the corneal ERG. ON- and OFF-Component Phase Lag Dramatically Affects the Flicker ERG The component analysis using waveform subtraction unmasked the presence of large postsynaptic ON- and OFF-components that were embedded in the control ERG. These two major components normally interfere with each other to some extent, due to phase lags of 160 to 230 that make the net postsynaptic component smaller than either component alone. This is not surprising, however, because it has been shown already that blocking either ON- or OFF-components by APB or PDA caused single-flash photopic ERGs to have larger negative or positive responses than control ERGs, indicating a push pull interplay between the ON- and OFF-component. 21 The phase differences between ON- and OFF-components depended on stimulus frequency. As stimulus frequency increased, the phase difference increased from 160 to 230 and was approximately 180 out of phase for 10 Hz stimuli. The destructive interference in these postsynaptic components coincides with the amplitude dip at 10 Hz frequency. Clearly, other factors are also necessary for the amplitude dip at 10 Hz, including the relative amplitudes of the ON- and OFF-component, and interactions between photoreceptoral and postreceptoral components. If the amplitudes of the ON- and OFFcomponents are comparable, their summed vector results in nulling the postsynaptic contribution. However, at 10 Hz, the OFF-component was consistently slightly larger than the ONcomponent, causing incomplete cancellation. At other frequencies the phase delay is either less than or more than 180, and incomplete cancellation should occur. Stockman et al. 27 observed similar destructive phase cancellation between the fast and slow rod signals in the human scotopic flicker ERG. One might hope, the cone photoreceptoral activity could be captured in clinical situations by recording the sine-wave ERG at approximately 10 Hz, where the post-receptoral ONand OFF-components effectively cancel each other and thereby minimize the amplitude contributed by postreceptoral activity (Fig. 4). However, it should be emphasized that these findings apply only to the normal retina. Results may be quite different for diseased retinas in which pathologic changes may alter the relative amplitudes and/or phases of the ON- and OFF-components. Photoreceptoral Contributions at Lower Frequencies In contrast to the fast flicker ERGs, the slow sine-wave ERG had a relatively larger contribution from the photoreceptors. The 4-Hz sine-wave ERG was reduced only 25% in fundamental amplitude from the control after APB PDA. This indicates that the major portion of this slow sine-wave response origi-

7 IOVS, January 2001, Vol. 42, No. 1 Primate Photopic Sine-Wave ERG 311 nates at the photoreceptoral level. This finding may help explain why, in past studies, the conclusions have been that with relatively slow sine-wave stimuli, the sine-wave flicker ERG originated predominantly from near the photoreceptors. 2,14 Even at low-stimulus frequencies, however, the postreceptoral ON- and OFF-components participate in the flicker ERG response. It can be noted that even after blocking output from the photoreceptors using APB PDA, the response was not a simple sine wave (Fig. 2). Further, the frequency-amplitude curve of the photoreceptor component was a low-pass function across 4 Hz to 32 Hz, but above 32 Hz showed a small relative maximum at approximately 50 Hz in all three monkeys (Fig. 3). These results suggest a complex response of the cone photoreceptors, or that more than the cones alone are contributing even after APB PDA. One can wonder whether these responses are mediated, in part, through glial cell reaction to cone activity, because both retinal pigment epithelium cells and Müller cells responded to extracellular ionic changes in the subretinal space that result secondarily from photoreceptor light responses. 28 However, evidence from the mudpuppy shows that glial cells are unable to follow fast flicker stimuli, 29 although this is a cold-blooded amphibian rather than a primate. Simulation Model for Postreceptoral Disease Results of our simulation model indicated that retinal pathology of the postreceptoral components can alter the flicker ERG. Both phase delays and amplitude reductions of postsynaptic ON- or OFF-components can change the phase and the amplitude of the final flicker ERGs. Results showed that the dysfunction of postreceptoral components might even result in an increased amplitude or faster phase under some conditions. This suggests that, in clinical recordings of flicker ERG responses, neither large amplitude nor fast phase necessarily means that the retina is healthy. Thus care is advised in drawing conclusions about retinal pathology from flicker changes. These simulations may eventually aid in understanding the mechanism for an interesting phenomenon in patients with the complete type of congenital stationary night blindness (ccsnb). These subjects are believed to have selective dysfunction of the postsynaptic ON-pathway Kim et al. 24 found that ccsnb patients had a delayed fundamental component without much amplitude change. Although that study used strobe-flash 30-Hz stimuli, this study s simulation model of the ON-component reduction for sine-wave stimuli (Fig. 6B) roughly mimicked the flicker ERG change seen in the ccsnb patients. The present study would predict that, for sine-wave stimuli, ON-pathway disease may affect flicker ERG timing in preference to reducing the amplitude. The ccsnb data suggest that the simulation model in Figure 6 may be useful for interpreting flicker ERG change in presumed postsynaptic retinal diseases, although further work is needed to develop models specifically for xenon-pulse flicker, as the Fourier spectrum of the response to a brief-flash, 30-Hz flicker may be considerably different from that of a sine-wave, 30-Hz flicker. Contribution of Nonlinear or Subharmonic Components In this study, only the fundamental component was considered. However, second and higher harmonic components also contribute to the response for some stimulus rates (data not presented). For example, at 10 Hz, the second harmonic amplitude is as large as the fundamental amplitude (7.2, 6.5, and 6.2 V for the second harmonic versus 7.7, 6.2, and 6.0 for the fundamental in the three monkeys studied) and results in a frequency-doubled appearance for the response to 10 Hz sinewave stimulation. Interestingly, glutamate analogs appeared to have different effects on the different harmonic components. For instance, the application of APB increased the fundamental amplitude across almost all frequencies (Fig. 3), but APB decreased the second harmonic amplitude to variable degrees (data not shown). This means that the fundamental component becomes dominant after APB with relatively less second harmonic contribution, thus causing the waveforms to appear more sinusoidal after APB (Fig. 2). Further analysis of the second and higher harmonic components is needed to interpret systematically the origin of the sine-wave flicker ERG. Crevier and Meister 33 found a phenomenon of period-doubling of responses to intense square-wave stimuli of 30 Hz to 70 Hz in the salamander and the human. They attributed this to nonlinear feedback interactions, possibly involving cone photoreceptor and OFF-bipolar cells. Period-doubling causes a onehalf harmonic response that could fold back into and contaminate the true fundamental component if responses are averaged out of synchrony. We examined responses elicited at 32 Hz to 64 Hz by single trains of 15 responses without averaging. Period-doubling was evident only for intense squarewave stimuli. No doubling occurred with sine-wave stimuli even at high intensities, and specifically none was found for the stimulus intensities used in the present study. CONCLUSIONS The sine-wave flicker ERG results presented here were obtained at higher mean luminance and modulation depth than has been used in some previous clinical studies. 15,16 Also, flicker responses were recorded at lower mean luminance (2.24 log cd/m 2 ) and lower modulation depth of 40% and confirmed that this study s main findings did not change much (not shown). Thus it is concluded that the majority of the fast-flicker responses requires participation of proximal retinal neurons and that activity of cones alone is a minor part of the normal fast flicker responses at the cornea, irrespective of pulse, square-wave, or sine-wave stimulation. These results also show that the relative phase of the postsynaptic ON- and OFF-components dramatically affects the sine-wave flicker ERG and that phase cancellation of these two components contributes to the amplitude dip normally seen at approximately 10 Hz. Acknowledgments The authors thank Ronald A. Bush, Naheed W. Khan, and Bo Lei for valuable comments, and Eric B. Arnold and Jeffrey A. Jamison for software and hardware help. References 1. Granit R. Sensory Mechanisms of the Retina. London: Oxford University Press Donovan WJ, Baron WS. Identification of the R G-cone difference signal in the corneal electroretinogram of the primate. J Opt Soc Am. 1982;72: Baron WS, Boynton RM. Response of primate cones to sinusoidally flickering homochromatic stimuli. J Physiol. 1975;246: Baron WS, Boynton RM, Hammon RW. Component analysis of the foveal local electroretinogram elicited with sinusoidal flicker. Vis Res. 1979;19: Baron WS, Boynton RM, van Norren D. Primate cone sensitivity to flicker during light and dark adaptation as indicated by the foveal local electroretinogram. Vis Res. 1979;19:

8 312 Kondo and Sieving IOVS, January 2001, Vol. 42, No Abraham FA, Alpern M. Factors influencing threshold of the fundamental electrical response to sinusoidal excitation of human photoreceptors. J Physiol. 1984;357: Abraham FA, Alpern M, Kirk DB. Electroretinograms evoked by sinusoidal excitation of human cones. J Physiol. 1985;363: Seiple WH, Holopigian K, Greenstein V, Hood DC. Temporal frequency dependent adaptation at the level of the outer retina in humans. Vis Res. 1992;32: Chang Y, Burns SA, Kreitz MR. Red green flicker photometry and nonlinearities in the flicker electroretinogram. J Opt Soc Am A. 1993;10: Gouras P, Gunkel RD. The frequency response of normal, rod achromat and nyctalope ERGs to sinusoidal monochromatic light stimulation. Doc Ophthalmol. 1964;18: Baker CL, Hess RF. Linear and nonlinear components of human electroretinogram. J Neurophysiol. 1984;51: Odom JV, Reits D, Burgers N, Riemslag FC. Flicker electroretinograms: a systems analytic approach. Opt Vis Sci. 1992;69: Burns SA, Elsner AE, Kreitz MR. Analysis of nonlinearities in the flicker ERG. Opt Vis Sci. 1992;69: Baker CL, Hess RR, Olsen BT, Zrenner E. Current source density analysis of linear and non-linear components of the primate electroretinogram. J Physiol. 1988;407: Falsini B, Iarossi G, Porciatti V, et al. Postreceptoral contribution to macular dysfunction in retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1994;35: Falsini B, Iarossi G, Fadda A, et al. The fundamental and second harmonic of the photopic flicker electroretinogram: temporal frequency-dependent abnormalities in retinitis pigmentosa. Clin Neurophysiol. 1999;110: Bush RA, Sieving PA. Inner retinal contributions to the primate photopic fast flicker electroretinogram. J Opt Soc Am A. 1996;13: Slaughter MM, Miller RF. 2-amino-4-phosphonobutyric acid: a new pharmacological tool for retina research. Science. 1981;211: Slaughter MM, Miller RF. An excitatory amino acid antagonist blocks cone input to sign-conserving second-order retinal neurons. Science. 1983;219: Xu XJ, Xu J, Huang B, et al. Comparison of pharmacological agents (aspartate vs. aminophosphonobutyric plus kynurenic acids) to block synaptic transmission from retinal photoreceptors in frog. Exp Eye Res. 1991;52: Sieving PA, Murayama K, Naarendorp F. Push pull model of the primate photopic electroretinogram: a role for hyperpolarizing neurons in shaping the b-wave. Vis Neurosci. 1994;11: Brodie SE, Naidu EM, Goncalves J. Combined amplitude and phase criteria for evaluation of macular electroretinograms. Ophthalmol. 1992;99: Sieving PA, Arnold EB, Jamison J, et al. Submicrovolt flicker electroretinogram: cycle-by-cycle recording of multiple harmonics with statistical estimation of measurement uncertainty. Invest Ophthalmol Vis Sci. 1998;39: Kim SH, Bush R, Sieving PA. Increased phase lag of the fundamental harmonic component of the 30 Hz flicker ERG in Schubert Bornschein complete type CSNB. Vis Res. 1997;37: Bush RA, Sieving PA. A proximal retinal component in the primate photopic ERG a-wave. Invest Ophthalmol Vis Sci. 1994;35: Sieving PA, Frishman LJ, Steinberg RH. M-wave of proximal retina in cat. J Neurophysiol. 1986;56: Stockman A, Sharpe LT, Ruther K, Nordby K. Two signals in the human rod visual system: a model based on electrophysiological data. Vis Neurosci. 1995;12: Steinberg RH, Frishman LJ, Sieving PA. Negative components of the electroretinogram from proximal retina and photoreceptor. In: Osborne N, Chader G, eds. Progress in Retinal Research. New York: Pergamon Press; 1991: Miller RF, Dowling JE. Intracellular response of the Müller (glial) cells of the mudpuppy retina: their relation to b-wave of the electroretinogram. J Neurophysiol. 1970;33: Miyake Y, Yagasaki K, Horiguchi M, Kawase Y. On- and offresponses in photopic electroretinogram in complete and incomplete types of congenital stationary night blindness. Jpn J Ophthalmol. 1987;31: Houchin K, Purple RL, Wirtschafter JD. X-linked congenital stationary night blindness and depolarizing bipolar system dysfunction [ARVO abstract]. Invest Ophthalmol Vis Sci. 1991;32(4): S1229. Abstract nr Young RSL. Low-frequency component of the photopic ERG in patients with X-linked congenital stationary night blindness. Clin Vis Sci. 1991;6: Crevier DW, Meister M. Synchronous period-doubling in flicker vision of salamander and man. J Neurophysiol. 1998;79: