ON-Pathway Dysfunction and Timing Properties of the Flicker ERG in Carriers of X-Linked Retinitis Pigmentosa

Abstract

purpose. Carriers of X-linked retinitis pigmentosa (XLRP) frequently show prolonged implicit times of the flicker electroretinogram (ERG). This study tested the hypothesis that a preferential response attenuation within the cone depolarizing (ON) bipolar cell (DBC) pathway is a major contributing factor.

methods. Light-adapted, full-field ERGs were recorded from 10 XLRP carriers and 12 visually normal control subjects. Fundamental amplitudes and phases of ERG responses to sinusoidally flickering stimuli at temporal frequencies ranging from 8 to 96 Hz were analyzed within the framework of a recent vector summation model of the cone system ERG to test for evidence of a response attenuation within the DBC pathway. In addition, ERG responses to sawtooth flicker were examined for a reduced b- to d-wave amplitude ratio, indicative of ON pathway dysfunction.

results. The carriers’ fundamental response phases at 32 Hz correlated significantly with their log ratios of response amplitudes at 32 versus 12 Hz (r = 0.89, P < 0.001) and with their log b- to d-wave amplitude ratios (r = 0.71, P < 0.05), both of which were used as indices of response attenuation within the DBC pathway. A control experiment demonstrated that a reduced sensitivity of cone phototransduction made at most only a minimal contribution to the timing changes in the carriers’ flicker ERG responses.

conclusions. The overall pattern of results indicates that a preferential response attenuation within the DBC pathway is the primary source of timing changes in the flicker ERGs of these carriers of XLRP. These findings illustrate the value of analyzing ERG responses to flickering stimuli at multiple temporal frequencies to evaluate mechanisms of disease action in photoreceptor degenerations.

Retinitis pigmentosa (RP) refers to a group of hereditary retinal degenerations with characteristics that include night blindness, visual field depressions or scotomata, and reductions in the amplitude of the electroretinogram (ERG) of the rod and cone systems. ^{1 2} Prolonged implicit times (time to peak) of the cone ERG have also been noted frequently in patients with RP, both in the response to brief flashes and to 30-Hz flicker. ^{1 3} The explanation for the prolonged implicit time of the cone ERG in RP is not entirely clear. As discussed by Hood and Birch, ³ it is likely that both photoreceptoral and postreceptoral factors play a role. By a quantitative analysis of the a-wave, Hood and Birch ³ concluded that low sensitivity of cone phototransduction could account for part of the timing changes in the patients’ ERGs, consistent with previous suggestions. ^{1 4 5} However, the timing changes were greater than could be accounted for by a photoreceptor abnormality alone, implying a contribution from postreceptoral factors. 

Other investigators have also postulated that a postreceptoral component contributes to the timing changes in the ERG of patients with RP. Birch and Sandberg ⁶ suggested that the increased flicker ERG implicit times might be due to a decrease in the magnitude of normal rod–cone interaction in patients with RP. Massof et al. ⁷ observed that the timing changes in the ERG in patients with RP did not correspond to a simple response delay, but that there appeared to be a smearing out of the flicker ERG waveform, so that the positive peak was more delayed than the negative trough. They suggested that there may be a change in the rate constants of the retinal generators of the flicker ERG response in patients with RP. Falsini et al. ⁸ concluded that there was a postreceptoral component to timing changes in the focal (F)ERG in patients with RP, based on the observation that the second harmonic of the 8-Hz FERG response is more reduced than the fundamental of the 32-Hz FERG response. The underlying assumption was that the second harmonic of the FERG is generated by postreceptoral neurons and the fundamental by the photoreceptors. Hood ⁹ proposed recently that the postreceptoral source of the ERG timing changes in RP may be an abnormal mechanism of adaptation within the outer plexiform layer that alters a time-dependent gain change. 

One possible postreceptoral source of a prolonged flicker ERG implicit time that has not been investigated previously is a selective attenuation of the light response of the cone depolarizing (ON) bipolar cells (DBCs) relative to that of the hyperpolarizing (OFF) bipolar cells (HBCs). If the light response of the DBC pathway of the monkey retina is decreased through the intravitreal injection of l-2-amino-4-phosphonobutyrate (l-AP4, formerly APB), there is an increased phase lag of the ERG response to 30-Hz flicker, with only a minimal effect on response amplitude. ¹⁰ There is suggestive evidence that a response attenuation within the DBC pathway may occur in RP. For example, some patients with RP show a greater reduction in the amplitude of the ERG ON response than in that of the OFF response, ¹¹ consistent with a relative DBC response attenuation. 

The present study focused on the origin of prolonged ERG implicit times in carriers of X-linked retinitis pigmentosa (XLRP). XLRP is a particularly severe form of RP, with affected males exhibiting night blindness, reduced visual acuity, and markedly reduced or nondetectable ERG responses from an early age. ^{12 13} Female carriers of XLRP also manifest signs and symptoms of RP that range from mild to severe, ^{14 15} with the variation in severity likely resulting from both the genetic heterogeneity of XLRP ^{16 17 18} and the variable degrees of lyonization. ¹⁹ As is the case in patients with RP, carriers of XLRP frequently show a prolonged implicit time of the 30-Hz flicker ERG of the cone system. ^{20 21 22 23} Further, it has been reported recently that some carriers of XLRP show an abnormal ERG ON response to a long-duration flash with a relatively preserved OFF response (Bailey CC, et al. IOVS 2001;42:ARVO Abstract 417). This latter finding raises the possibility that the prolonged ERG implicit times of XLRP carriers may result from a response attenuation within the DBC pathway. 

The purpose of the present study was to test this possibility. In the first part of the study, we recorded the ERGs of XLRP carriers in response to a sinusoidally flickering full-field stimulus and analyzed the responses within the framework of a recent vector summation model of the ERG. ¹⁰ According to this model, the fundamental of the ERG response to sinusoidal flicker is the vector sum of the fundamental responses of the cone photoreceptors, the DBCs, and the HBCs. As illustrated in Alexander et al., ²⁴ a response attenuation within the DBC pathway produces characteristic changes in the shapes of the response functions that relate ERG amplitude and phase to temporal frequency. First, the fundamental response amplitude is increased at low temporal frequencies, with little change at high frequencies, thereby producing an amplitude response function that is flatter than normal across the lower frequency range. Second, a DBC response attenuation introduces a phase lag at temporal frequencies of 30 Hz and higher, whereas a phase advance occurs at low temporal frequencies. 

A response attenuation within the DBC pathway should also lead to a reduction in the ERG ON response relative to the OFF response. ²⁵ To evaluate this possibility, we analyzed the carriers’ ERG responses to rapid-on and rapid-off sawtooth flicker. Sawtooth stimuli were used to elicit ERG ON and OFF responses to minimize the possible intrusion of eye movements, which can sometimes obscure the waveform morphology of the OFF response when long-duration stimuli are used. ²⁶ A response attenuation within the DBC pathway should introduce a greater attenuation of the b-wave of the ON response than of the d-wave of the OFF response, with a consequent reduction in the b- to d-wave amplitude ratio. Further, if a DBC response attenuation is the primary determinant of the timing changes in the 30-Hz flicker ERG of carriers of XLRP, then there should be a significant correlation between the b- to d-wave amplitude ratios and the phase of the response fundamentals at 30 Hz. 

An ancillary experiment was performed to evaluate the possible contribution of reduced sensitivity of cone phototransduction to the prolonged implicit times of the flicker ERG of the carriers of XLRP. The approach was similar to that of Hood and Birch. ³ Reduced sensitivity of cone phototransduction is functionally equivalent to reduced stimulus luminance. If reduced sensitivity of cone phototransduction accounts fully for the ERG abnormalities of the XLRP carriers, then it should be possible to mimic those abnormalities in control subjects by reducing the stimulus luminance appropriately. Therefore, we assessed the effect of a reduced stimulus luminance on the phase of the flicker ERG, the shape of the ERG temporal response function, and the b- to d-wave ratio in a subset of the control subjects to determine whether the results would mimic those of the XLRP carriers. The magnitude of the luminance reduction was chosen to produce in the control subjects an a-wave implicit time in response to a brief flash that corresponded to the most prolonged brief-flash a-wave implicit time of the XLRP carriers. 

Methods

Subjects

Ten carriers of XLRP from eight families participated in the study. The carriers’ characteristics are given in Table 1 , with the carriers listed in order of increasing age (mean age, 39.7 years; range, 18–64). Patients 1 and 2 are sisters, and patient 9 is the mother of patient 4. Seven of the women (all except patients 3, 6, and 7) are obligate carriers. The three nonobligate carriers, as well as five of the other seven carriers, showed a tapetal-like reflex, which is considered a diagnostic sign for the carrier state. ¹⁴ The grades of fundus appearance listed in Table 1 are based on a slightly modified version of a fundus-grading scheme used previously. ²⁷ Grade 0 (n = 1) indicates that the fundus appearance is normal, grade 1 (n = 6) indicates a tapetal-like reflex with no peripheral pigmentary changes, and grade 2 (n = 3) indicates pigmentary changes with a regional predilection, with or without a tapetal-like reflex. The carriers had minimal or no lens opacities in the tested eye, which was the left eye in all observers. Patient 8 was taking medication for the treatment of non–insulin-dependent diabetes and low-tension glaucoma. Patient 10 was also taking medication for non–insulin-dependent diabetes. Neither showed signs of diabetic retinopathy. 

Blood samples were obtained for DNA analysis from nine of the carriers (patients 1 and 3–10), representing all eight families. In some cases (patients 3, 5, 6, 8, and 10), a blood sample was also provided by an affected male relative. Mutations were found in the RPGR gene in three of the carriers (patients 4, 9, and 10) from two families (see Table 1 ). Blood samples from carriers 3, 5, and 6 tested negative for mutations in the RP2 and RPGR genes, including exon ORF15. ^{28 29}  

Twelve visually normal control subjects (10 women and two men) participated in the study. Their ages (mean, 46.7 years; range, 34–64) did not differ significantly from those of the carriers (t = 1.41, P = 0.18). All control subjects had best corrected visual acuities of 20/20 or better in the tested eye, clear ocular media, and normal-appearing fundi on ophthalmologic examination. Five of the control subjects also participated in an ancillary experiment in which ERG responses to low-luminance stimuli were recorded. 

The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board of the University of Illinois at Chicago. Informed consent was obtained from all subjects after the nature and possible consequences of the study had been explained to them. 

Stimuli and Instrumentation

ERG responses were measured with instrumentation that has been described in detail previously. ³⁰ In brief, the stimulus consisted of achromatic full-field flicker that was superimposed on an achromatic rod-desensitizing adapting field, both presented within an integrating sphere (Oriel, Stratford, CT). The flickering stimulus and adapting field were provided by two separate optical channels. Each channel had a light source consisting of a 300-W tungsten-halogen bulb (each housed within a projector; Eastman Kodak, Rochester, NY), and infrared blocking filters. The light from the optical channels was combined with a Y fiber-optic light guide (Oriel). The luminance of the adapting field was 17.4 cd/m² (2.9 log td, assuming an 8-mm dilated pupil). The maximum stimulus luminance was 393.0 cd/m² (4.3 log td) and the minimum luminance was 1.4 cd/m² (1.8 log td), as measured with a photometer (LS-110; Minolta, Osaka, Japan). In the absence of the adapting field these luminances produced a modulation of 99%. Against the adapting field, the modulation was 91.2%. 

Temporal modulation of the test field was controlled by a ferroelectric liquid crystal (FLC) shutter (Displaytech, Longmont, CO) and driver (DR-95; Displaytech). The driver was controlled by a signal-processing board (DAS-801; Keithley, Cleveland, OH) housed within a microcomputer. The FLC shutter was driven at a constant frequency of 1 kHz and was pulse-width modulated under computer control. A shutter and driver (Vincent Associates, Rochester, NY) within the second optical channel controlled the adapting field presentation. 

ERG recordings were made in response to sinusoidal flicker presented at temporal frequencies of 8, 12, 16, 20, 26, 32, 40, 50, 64, 80, and 96 Hz (approximately 0.1-log unit steps) at maximum modulation and in sine phase. Because of time constraints, patient 7 was not tested at 80 or 96 Hz. As a measure of the noise level, ERG responses were recorded to a 16-Hz sinusoidal stimulus presented at zero contrast. 

ERG responses were also recorded to sawtooth stimuli at frequencies of 4 and 8 Hz at maximum modulation. Each cycle of rapid-on sawtooth flicker consisted of an abrupt increment in luminance, to emphasize an ON response, followed by a linear decrease in luminance. Each cycle of rapid-off flicker consisted of an abrupt decrement in luminance, to emphasize an OFF response, followed by a linear increase in luminance. In addition, ERG responses were recorded to four brief (10 ms) flashes of 3.9 cd-s/m² (2.3 log td-s), presented at 1-second intervals. 

Procedure

The pupil of the tested eye was dilated with 2.5% phenylephrine hydrochloride and 1% tropicamide drops, and the cornea was anesthetized with proparacaine drops. The observer’s head was held in position with a chin rest and forehead bar. ERGs were recorded with a Burian-Allen bipolar contact lens electrode, grounded with an earclip electrode. Responses were acquired with a signal-averaging system (Viking IV; Nicolet, Madison, WI) that was triggered by a transistor–transistor logic (TTL) signal generated by the signal-processing board and synchronized with the onset of each stimulus cycle. Amplifier band-pass settings were 0.5 to 500 Hz. 

Each subject was light-adapted to room illumination before testing and was then adapted for 2 minutes to the adapting field. Recordings were begun after the subject had adapted to each stimulus for approximately 30 seconds. For each stimulus condition, three 500-ms recordings were obtained to determine reproducibility. Each recording was the average of four sweeps. The three recordings were averaged off-line, so that each waveform included in the analysis consisted of the average of 12 500-ms sweeps. 

Data Analysis

Measures of ERG response amplitudes and timing were obtained as follows. The peak-to-trough amplitude and implicit time of the ERG responses to 32-Hz sinusoidal flicker were derived for each cycle and then were averaged across the 16 cycles in each 500-ms waveform. The amplitude of the b-wave response to rapid-on flicker was measured from the a-wave trough to the b-wave peak. The amplitude of the d-wave response to rapid-off flicker was measured from the onset of the response to the response peak, after the ERG waveform had been digitally low-pass filtered (−3 dB at 97.5 Hz) with a noncausal, zero-phase infinite impulse response filter (Matlab Signal Processing Toolbox; The MathWorks, Natick, MA). Low-pass filtering was applied to minimize the influence of any oscillatory potentials coincident with the d-wave peak. The amplitudes of the b- and d-waves were averaged for the two (4-Hz) or four (8-Hz) cycles in each 500-ms ERG waveform. The amplitude of the a-wave response to a brief flash was measured from the response baseline to the trough of the a-wave. The a-wave implicit time was measured from the flash onset to the a-wave trough. 

ERG responses to sinusoidal flicker were also evaluated by spectral analysis. The fundamental response amplitudes were derived from power spectral densities, and the fundamental phases were obtained from fast Fourier transforms (Matlab Signal Processing Toolbox; The MathWorks). The ERG responses to a zero-contrast stimulus were also analyzed by this method to obtain a noise estimate at each temporal frequency. The fundamental amplitudes that are plotted in the figures represent the full peak-to-trough amplitudes of the derived response fundamentals. The phases are given in cosine phase, as per convention. A subject’s fundamental response at a given temporal frequency was considered distinguishable from noise if the amplitude was at least twice as large as the noise amplitude at that frequency and if there was intertrial consistency in phase. 

The ERG results for the XLRP carriers were analyzed with Spearman correlations. P < 0.05 was considered to be statistically significant. 

Results

ERG Responses to Sinusoidal Flicker

The solid traces in Figure 1 represent the averaged ERG waveforms of the 10 XLRP carriers and three representative control subjects for a temporal frequency of 32 Hz, a frequency typically used clinically. The waveforms of the control subjects (Fig. 1 , left) illustrate the largest (top trace), smallest (bottom trace), and average (middle trace) control responses and also depict the normal variation in waveform shape. The numbers to the right of the carriers’ waveforms are the same designations used in Table 1 . The carriers’ waveforms are presented in order of decreasing amplitude, using the same amplitude scale as for the control subjects. The vertical dashed lines in both panels indicate the mean implicit time of the 12 control subjects’ responses. 

The response amplitudes of the XLRP carriers ranged from normal to markedly subnormal. Although the smallest response was recorded from the oldest carrier, the carriers’ ages and response amplitudes did not correlate significantly (r = −0.31, P = 0.36). The carriers’ implicit times ranged from normal to prolonged, and did not correlate significantly with age (r = 0.33, P = 0.33). However, there was a statistically significant correlation between the carriers’ implicit times and amplitudes (r = −0.79, P < 0.01), so that prolonged implicit times were associated with reduced amplitudes. 

Also shown in Figure 1 are the response fundamentals (dotted traces) for each of the waveforms, derived from spectral analysis. Except for small individual variations in the ERG waveform shape, the response fundamentals captured the overall amplitude and timing of the ERG waveforms. In fact, there was a significant correlation between peak-to-trough amplitude and fundamental response amplitude for the XLRP carriers (r = 0.99, P < 0.001) and control subjects (r = 0.98, P < 0.001), as well as a significant correlation between the implicit times and fundamental response phases for the XLRP carriers (r = −0.93, P < 0.001) and control subjects (r = −0.66, P < 0.05; the negative correlations reflect the fact that a longer implicit time corresponds to a more negative phase). In the remainder of the data presentation, fundamental response amplitude and phase are used as the primary measures of the flicker ERG. 

A more quantitative comparison of the response amplitudes and phases of the XLRP carriers and control subjects at 32 Hz is presented in Figure 2 . Also shown in this figure are the response amplitudes (top) and phases (bottom) at temporal frequencies ranging from 8 to 96 Hz, which were examined to determine whether there were abnormalities in the flicker ERG of the XLRP carriers at temporal frequencies other than 32 Hz. The shaded regions in Figure 2 indicate the ranges of amplitudes and phases for the response fundamentals of the 12 control subjects. The numbers in the legend correspond to the carrier designations in Table 1 . The symbols used in Figure 2 (and in the subsequent figures) were assigned to the carriers as follows: filled and open circles for the two sisters (patients 1 and 2), filled and open diamonds for the daughter and mother (patients 4 and 9, respectively), filled and open squares for the two carriers who had essentially normal ERG response amplitudes and phases (patients 3 and 6), and filled and open upright or inverted triangles for the four carriers with the greatest phase lags at 32 Hz (patients 5, 7, 8, and 10). Data points were not plotted if they were indistinguishable from noise in that subject at that temporal frequency, according to the criterion described in the Methods section. 

Consistent with previous reports, ^{30 31} the amplitude function of the control subjects (Fig. 2 , top, shaded region) had a peak at 32 Hz (indicated by the arrow) and a trough at 12 Hz. All the carriers had amplitudes within the normal range at 12 Hz, but only one (patient 3) had amplitudes within the normal range at all temporal frequencies. The other nine carriers had various degrees of amplitude reduction at frequencies higher than 12 Hz. In six of those carriers (patients 1, 2, 4, 6, 8, and 9), the amplitude reductions were most apparent within the mid-frequency range from 16 to 40 Hz, whereas the remaining three carriers (patients 5, 7, and 10) had substantial amplitude reductions within both the mid-frequency and high-frequency ranges. 

The fundamental response phases of the control subjects (Fig. 2 , bottom, shaded region) were approximately constant or increased between 8 and 16 Hz and then decreased systematically at the higher temporal frequencies, as reported previously. ^{30 31} Only two of the XLRP carriers (patients 3 and 6) had normal phases at all the frequencies tested. The other carriers had various degrees of phase lags (i.e., phases that were more negative than those of the control subjects). These phase lags were most apparent within the mid-frequency range of 16 to 40 Hz. The fundamental response phases were normal or showed a phase advance at 8 Hz, and the phases of all carriers were within the normal range at 12 Hz. The response phases also generally were within the normal range for temporal frequencies of 50 Hz and higher. 

According to the vector summation model, ¹⁰ the amplitude minimum near 12 Hz in the normal ERG temporal response function is due to a relative cancellation between out-of-phase DBC and HBC response components. If the carriers of XLRP have an attenuated DBC response, then the model predicts that there would be a smaller response minimum at 12 Hz, resulting in a relatively flattened response function at low temporal frequencies. To evaluate this prediction, the temporal response functions of the XLRP carriers were first normalized at high frequencies to compensate for any overall reduction in response amplitude. In this normalization, each carrier’s response function was shifted vertically by an amount equal to her average difference from the normal mean amplitudes at frequencies ranging from 50 to 96 Hz (50–64 Hz for carriers 5 and 7; 50–80 Hz for 8). The normalized results are plotted in Figure 3 , in which the symbols represent the values for the individual carriers of XLRP, and the shaded region represents the normal range. 

For all the carriers of XLRP, the shape of the response function at high temporal frequencies corresponded to that of the control subjects. The significance of this relationship will be considered in the Discussion section. In addition, the response functions in two carriers (patients 3 and 6) were similar in shape to those of the control subjects at all temporal frequencies, including the presence of a pronounced response minimum at 12 Hz. However, the curves for the other eight carriers displayed various degrees of flattening across the frequency range from 12 to 32 Hz. This relative flattening is consistent with a response attenuation within the DBC pathway according to the vector summation model. 

If a DBC response attenuation accounts for the phase lags of the carriers’ 32-Hz flicker ERGs, then increased phase lags should be associated with a greater flattening of the amplitude response functions. To investigate this possibility, we derived the ratio of fundamental response amplitudes at 32 versus 12 Hz as an index of curve flatness. The relationship between this response ratio and the response phase at 32 Hz is plotted in Figure 4 . The data for the individual carriers are represented by the symbols; the shaded region indicates the ranges for the control subjects. The solid line in Figure 4 is a bivariate regression line fitted to the carriers’ data. In agreement with the prediction, there was a statistically significant correlation between the 32-Hz phases and the log 32- to 12-Hz amplitude ratios of the XLRP carriers (r = 0.89, P < 0.001), such that increased phase lags at 32 Hz were associated with flatter amplitude functions across the lower temporal frequencies. 

The vector summation model also predicts that a relative response attenuation within the DBC pathway should introduce a phase advance at 8 Hz in addition to a phase lag at 32 Hz. In agreement with this prediction, the carriers’ phases at 8 Hz correlated significantly (r = −0.75, P < 0.05) with their phases at 32 Hz, so that phase advances at 8 Hz were associated with phase lags at 32 Hz. Therefore, this analysis of the ERG temporal response functions of the XLRP carriers is consistent with the hypothesis that a DBC response attenuation is a major determinant of the phase lags of the 32-Hz flicker ERG in this group of XLRP carriers. 

ERG ON and OFF Responses

As a further test of this conclusion, we evaluated the ERG responses of the XLRP carriers to rapid-on and rapid-off sawtooth stimuli that evoked ON and OFF responses, respectively. An attenuated DBC component would be expected to produce a relatively greater reduction in the amplitude of the b-wave of the ON response than of the d-wave of the OFF response. ²⁵ Figure 5 presents the carriers’ ERG responses to rapid-on (Fig. 5 , left) and rapid-off (Fig. 5 , right) flicker at a temporal frequency of 4 Hz, together with the results from a representative control subject (top trace in each panel). A similar pattern of results was obtained at 8 Hz (data not shown). Eight of the carriers (all except patients 3 and 6) had b-wave amplitudes that were below the lower limit of normal, whereas only four carriers (patients 2, 5, 7 and 10) had d-wave amplitudes that were below the normal range. Of the eight carriers who had a reduced b-wave amplitude, seven had b- to d-wave amplitude ratios that were below the normal range, as indicated by the asterisks in Figure 5 . 

If the phase lags of the XLRP carriers at 32 Hz are the result of a DBC response attenuation, as hypothesized, then there should be a significant correlation between their fundamental response phases at 32 Hz and their log b- to d-wave amplitude ratios. The relationship between these two measures is shown in Figure 6 . As predicted, there was a statistically significant correlation (r = 0.71, P < 0.05) between the 32-Hz response phases and the log b- to d-wave amplitude ratios of the carriers, so that increased phase lags were associated with reduced amplitude ratios. Further, the log b- to d-wave amplitude ratios of the XLRP carriers correlated significantly with their log 32- to 12-Hz amplitude ratios (r = 0.83, P < 0.001), as would be expected if both are the result of a reduced DBC response component. In combination with the results shown in Figure 4 , these findings provide support for the hypothesis that a response attenuation within the DBC pathway is the major source of the phase lags of the 32-Hz flicker ERG of the XLRP carriers. 

Control Experiment: Effect of Reduced Stimulus Luminance

Although the results presented thus far indicate that a DBC response attenuation is a major determinant of the timing changes in the flicker ERGs of the XLRP carriers, it is also possible that a reduced sensitivity of cone phototransduction may have contributed. ³ To evaluate this possibility, we assessed the characteristics of the a-waves of the carriers’ and control subjects’ ERG responses to brief flashes. The ERG waveforms were normalized by setting the amplitude of the a-wave of all subjects to unity to provide a direct comparison of the leading edge of the a-wave, an index of cone photoreceptor function, ³ as well as to provide a comparison of a-wave timing. In this normalization, a reduced sensitivity of cone phototransduction would be apparent as a decreased a-wave slope. 

Figure 7 illustrates the normalized waveforms from the carriers, overlaid on shaded regions that represent the control ranges. For clarity, the results from the XLRP carriers were grouped into three plots, based on their a-wave implicit times. Six of the 10 carriers (patients 1–4, 6, and 9) had normal a-wave implicit times and normal a-wave slopes (Fig. 7 , top). This result indicates that there was no reduced sensitivity of cone phototransduction in those carriers. Two of these carriers (patients 1 and 2) had response phases of the 32-Hz flicker ERG that were outside the range of normal. Therefore, it is unlikely that the 32-Hz phase lags for those two carriers were the result of a decreased photoreceptor sensitivity. The a-wave implicit time was prolonged by 1 ms beyond the range of normal in two of the carriers (patients 8 and 10; Fig. 7 , middle), and by 3 and 4 ms in carriers 5 and 7, respectively (Fig. 7 , bottom). This result is consistent with a modest reduction in the sensitivity of cone phototransduction in those four carriers. 

If a reduced sensitivity of cone phototransduction accounts fully for the 32-Hz phase lags of the flicker ERGs of these four carriers, then it should be possible to simulate all their ERG abnormalities in control subjects by reducing the stimulus luminance appropriately. Therefore, five of the control subjects were tested at a stimulus luminance that was reduced by 1 log unit from the standard value. This luminance reduction resulted in a mean a-wave implicit time of 23.4 ± 0.55 ms in the control subjects, which was similar to that in the two carriers with the most prolonged a-wave implicit times (Fig. 7 , bottom). 

Figure 8 compares the control subjects’ results at the lower stimulus luminance with the results for the XLRP carriers at the standard luminance in terms of the fundamental response phase of the 32-Hz ERG (Fig. 8 , left), log 32- to 12-Hz amplitude ratio (Fig. 8 , middle), and log b- to d-wave amplitude ratio (Fig. 8 , right). The control subjects’ results for the lower luminance are represented by the dotted hexagons. The shaded regions indicate the ranges of values in these five control subjects when tested with the standard stimulus luminance. The results in the four carriers with the prolonged a-wave implicit times are represented by triangles. 

The reduction in stimulus luminance introduced phase lags in the fundamental responses of the control subjects at 32 Hz, relative to their response phases at the standard luminance (Fig. 8 , left), but the phase lags of the control subjects were not as great as those of the carriers. Similarly, the 32- to 12-Hz amplitude ratios of the control subjects at the lower luminance were shifted toward the carriers’ values (Fig. 8 , middle), but the ratios were not as reduced as those of the carriers. Finally, the b- to d-wave ratios of the control subjects at the lower luminance remained at or above the values obtained at the standard luminance (Fig. 8 , right), indicating proportional reductions in b- and d-wave amplitudes, whereas the carriers had reduced b- to d-wave ratios. Therefore, although the reduced stimulus luminance produced a-wave implicit times in the control subjects that were similar to those of the four XLRP carriers who had the longest a-wave implicit times, these carriers’ other ERG abnormalities, including the phase lag in the response to 32-Hz flicker, were not reproduced. These findings indicate that a reduced sensitivity of cone phototransduction is not the primary cause of the abnormal ERG findings in any of these carriers of XLRP, although this factor may have contributed slightly to the prolonged b-wave and flicker implicit times in the four carriers with the most prolonged a-wave implicit times. 

Discussion

The majority of the carriers of XLRP in this study showed reduced fundamental response amplitudes (9/10) and increased phase lags (6/10) of the 32-Hz flicker ERG of the cone system compared with the normal subjects. These findings are consistent with previous reports of flicker ERG abnormalities in carriers of XLRP when tested at temporal frequencies at or near 30 Hz. ^{14 20 22} The phase lags of the carriers’ response fundamentals correlated significantly with their fundamental response amplitudes, and the phase lags were most apparent across the midrange of temporal frequencies (16–40 Hz). This result substantiates the value of using temporal frequencies near 30 Hz to identify functional abnormalities in the ERG of carriers of XLRP. ³²  

The main goal of the present study was to test the hypothesis that the abnormality in the timing of the flicker ERG in carriers of XLRP is due primarily to a relatively greater response attenuation within the DBC system than within the HBC system. The pattern of results from the XLRP carriers in this study was consistent with this hypothesis. First, there was a significant correlation between the carriers’ fundamental response phases at 32 Hz and their log 32- to 12-Hz amplitude ratios, with increased phase lags associated with reduced amplitude ratios. According to the ERG vector summation model, ¹⁰ a reduced 32- to 12-Hz amplitude ratio, which corresponds to a flattening of the temporal response function at low frequencies, would be expected if there is a relative attenuation of the DBC response component. The XLRP carriers also showed a significant inverse correlation between their fundamental response phases at 8 and 32 Hz. A phase advance at 8 Hz accompanied by a phase lag at 32 Hz is the result to be expected if there is an attenuation of the DBC response component of the ERG, according to the vector summation model. Further, there was a significant correlation between the carriers’ response phases at 32 Hz and their log b- to d-wave amplitude ratios, with increased phase lags associated with reduced amplitude ratios. Such a decrease in the amplitude of the ERG ON response in relation to the OFF response is consistent with an attenuated DBC response component. ²⁵ Thus, the overall pattern of our results supports the hypothesis that a relative response attenuation within the DBC pathway is the primary source of the 32-Hz ERG phase lags shown by these carriers of XLRP. 

Our data demonstrate further that the phase lags of the flicker ERGs in these carriers of XLRP were not the result of a reduced sensitivity of cone phototransduction. Six of the carriers had normal a-wave implicit times, which is inconsistent with a decreased sensitivity of cone phototransduction. Of those six carriers, two had 32-Hz ERG response phases that were beyond the range of normal. Only 4 of the 10 carriers had a-wave implicit times that were even slightly prolonged, indicative of a modest reduction in the sensitivity of cone phototransduction. An ancillary experiment showed, however, that the phase lags of the 32-Hz flicker ERGs of those four carriers, as well as other features of their ERG responses (reduced b- to d-wave amplitude ratios and reduced 32- to 12-Hz amplitude ratios) could not be simulated in control subjects by reducing the stimulus luminance to mimic a reduced sensitivity of cone phototransduction. Therefore, the timing changes in the flicker ERGs of these XLRP carriers are not due solely to a reduced sensitivity of cone phototransduction or to any factor that is functionally equivalent to a reduced stimulus luminance. We note, however, that in those patients with RP and carriers who have longer implicit times of the flicker ERG than those observed in the present study, the prolonged implicit times are likely to reflect an abnormal light response of the cone photoreceptors as well as a defect in a postreceptoral response component. ³  

According to a previous study, ³³ an abnormal b-wave response to rapid-on flicker, as observed in these carriers of XLRP, could be due to a high-frequency response attenuation at the level of the cone photoreceptors. However, that did not appear to be the case in the current study. The shape of the carriers’ temporal response functions at frequencies at and above 50 Hz matched the shape of the normal response function across that frequency range. Evidence indicates that the shape of this high-frequency region of the ERG temporal response function is governed by the response properties of the cone photoreceptors, ³¹ although the photoreceptors appear to make little direct contribution to the high-frequency ERG response itself. ¹⁰ A more sluggish cone photoreceptor response should produce a lower corner frequency of the temporal response function (i.e., frequency at which the amplitude has decreased by 3 dB from the maximum). ³⁴ This was not observed in the XLRP carriers in this study. Further, the fundamental response phases of the carriers of XLRP were within normal limits at temporal frequencies at and above 50 Hz, consistent with normal timing of the cone photoreceptor response. Thus, this pattern of findings indicates that the ON-response defect and the timing changes in the flicker ERG of these XLRP carriers were not the result of an abnormal temporal response of the cone photoreceptors. 

In addition to a reduced amplitude of the b-wave of the ON response, these carriers of XLRP also tended to show a prolonged b-wave implicit time, particularly those carriers with the smallest overall response amplitudes (Fig. 5) . The prolonged b-wave implicit times could be construed as evidence for a response delay rather than a response attenuation within the DBC pathway. However, as noted previously, ^{25 35} a prolonged b-wave implicit time could be due either to a DBC response attenuation or to a DBC response delay, due to the way in which the responses of the DBCs and HBCs combine to shape the b-wave. These alternatives can be distinguished through an analysis of the ERG temporal response function. A delay in the DBC response relative to the HBC response would produce a response function that has a stronger band-pass shape than normal. ^{10 24} This would cause an increase in the 32- to 12-Hz amplitude ratio, not the decrease that was observed in the present study. Therefore, our results indicate that the prolonged b-wave implicit time of the XLRP carriers is more consistent with an attenuated DBC response than with a DBC response delay. 

An attenuated DBC response component could account for the characteristics of the flicker ERG that have been reported in patients with RP. For example, such a response attenuation could explain the smearing out of the flicker ERG in patients with RP that was observed by Massof et al. ⁷ An attenuation of the response of the DBC system would mean that the peak of the ERG waveform would be shaped by the later-occurring HBC response component. ¹⁰ An attenuated DBC response component could also account for the decreased ratio of second harmonic amplitude to fundamental amplitude in the fERG reported previously in patients with RP. ⁸ As demonstrated recently, ³⁶ an attenuation of the DBC response component of the monkey flicker ERG through the application of l-AP4 increases the amplitude of the fundamental response to sinusoidal flicker but has little effect on the second harmonic, thereby effectively decreasing the ratio of second harmonic amplitude to fundamental amplitude. 

It has been suggested that the postreceptoral component of the timing changes in the ERG in RP may involve an abnormal process of adaptation within the outer plexiform layer. ⁹ This suggestion was based in part on the observation that the second-order kernel of the multifocal ERG, which reflects adaptation to the successive flashes in a flickering stimulus, can be reduced substantially in patients with RP. However, the second-order kernel is also reduced in complete congenital stationary night blindness (CSNB1), ³⁷ a condition that is thought to result from an attenuation in neurotransmission from photoreceptors to DBCs. ³⁵ Therefore, the functional significance of the reduced second-order kernel in RP, and its relationship to ON-pathway dysfunction and to adaptational changes within the outer retina, remain to be determined. 

In conclusion, this analysis of the ERG temporal response function and of ERG ON and OFF responses of the cone system of carriers of XLRP indicates that carriers can have a relatively greater response attenuation within the DBC pathway than within the HBC pathway. Our results show further that this response attenuation is a major factor governing the timing changes observed in the flicker ERG of the XLRP carriers. We speculate that a relative attenuation of the DBC response may thus be the primary source of the postreceptoral component that is thought to contribute to timing changes in the flicker ERG of patients with RP. ^{3 7 8}  

 Supported by National Eye Institute Research Grant EY08301 (KRA) and Core Grant EY01792, a center grant from The Foundation Fighting Blindness (GAF), and an unrestricted departmental grant from Research to Prevent Blindness. KRA is a Research to Prevent Blindness Senior Scientific Investigator.

 Submitted for publication September 25, 2002; revised January 9, 2003; accepted January 22, 2003.

 Disclosure: K.R. Alexander, None; C.S. Barnes, None; G.A. Fishman, 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.

 

Table 1.

 

Patients’ Characteristics

Carrier	Age (y)	Visual Acuity (OS)	Refractive Error (OS)	Genetic Mutation	Fundus Grade (OS)	Goldmann Visual Field (OS)
1	18	20/20	−6.00 + 3.50 × 71	NA	1	Normal
2	25	20/20	−3.75 + 2.75 × 80	NA	1	Normal
3	35	20/15	−2.00 + 0.55 × 100	NA	1	Mild concentric restriction
4	36	20/15	plano + 1.25 × 45	RPGR exon 11; C1344T	0	Normal
5	37	20/25	−8.25 + 0.50 × 135	NA	2	Predominant superior restriction
6	39	20/20	−2.50 + 4.50 × 88	NA	1	Normal
7	41	20/40	−13.50 + 2.50 × 90	NA	2	Midperipheral scotoma in the temporal field for the V/4e target
8	45	20/20	−8.50 + 0.75 × 58	NA	1	Superior restriction
9	57	20/20	+0.50 + 4.25 × 45	RPGR exon 11; C1344T	1	Slight enlargement of the blind spot
10	64	20/30	−1.75 + 0.50 × 75	ORF15Del481-2	2	Concentrically restricted, more so in the superior field

Figure 1.

 

Light-adapted ERG responses to 32-Hz sine wave flicker. Left: ERGs of three representative control subjects, illustrating the largest (top), smallest (bottom), and mean (middle) response amplitudes. Right: ERGs of the XLRP carriers, displayed in order of decreasing amplitude, with the same amplitude scale as for the control subjects. Numbers on the right correspond to the carrier designations in Table 1 . Dotted traces: response fundamental for each waveform derived from spectral analysis. Vertical dashed lines: mean implicit times of the 12 control subjects, as derived from the averages of the 16 cycles in the 500-ms waveforms.

Figure 2.

 

Log amplitude (top) and phase (bottom) of the ERG response fundamentals for the individual XLRP carriers (symbols) as a function of the temporal frequency of sinusoidal flicker, plotted on a log scale. Shaded regions: normal ranges. Arrows: 32 Hz.

Figure 3.

 

Log relative amplitude of the ERG response fundamentals for the individual XLRP carriers (symbols) as a function of the temporal frequency of sinusoidal flicker on a log scale, replotted from Figure 2 and normalized across the high-frequency region to emphasize the shape of the functions at the lower temporal frequencies (see text for normalization details). Shaded region: range of normal amplitudes.

Figure 4.

 

Fundamental response phase at 32 Hz versus log ratio of the fundamental response amplitudes at 32 versus 12 Hz. Symbols: individual XLRP carriers. Shaded region: normal ranges. Line: bivariate regression line fitted to the carriers’ results.

Figure 5.

 

ERG waveforms in response to 4-Hz rapid-on (left) and rapid-off (right) sawtooth flicker for a representative control subject (top trace) and the individual carriers of XLRP, with patient designations indicated to the right of each waveform. The carriers’ waveforms are plotted in the same order as in Figure 1 . Asterisks: carriers with a b- to d-wave amplitude ratio below the lower limit of the normal range. Dotted lines: 50-ms intervals.

Figure 6.

 

Fundamental response phase at 32 Hz versus log ratio of b- to d-wave amplitude. Symbols: individual XLRP carriers. Shaded region: normal ranges. Line: bivariate regression line fitted to the carriers’ results.

Figure 7.

 

ERG waveforms in response to brief flashes, with a-waves normalized to unit amplitude. Traces: individual XLRP carriers. Shaded regions: normal ranges. For clarity, the carriers’ results were grouped according to the a-wave implicit times as follows: top: no prolongation; middle: mild prolongation; bottom: moderate prolongation.

Figure 8.

 

Effect of a 1-log unit reduction in stimulus luminance on the fundamental response phase at 32 Hz (left), the log ratio of the fundamental response amplitude at 32 versus 12 Hz (middle), and the log ratio of b- to d-wave amplitude for 8-Hz sawtooth flicker (right). Dotted hexagons: individual results for five control subjects at the lower luminance. Shaded regions: ranges of values for the five control subjects at the standard luminance. Triangles: results at the standard luminance for the four XLRP carriers who had the longest a-wave implicit times.

The authors thank Deborah J. Derlacki and Sandeep Grover for assistance in performing the ERG testing and Debra K. Breuer, Anand Swaroop, and Beverly M. Yashar for genetic testing of the carriers.