Sites of cone system sensitivity loss in retinitis pigmentosa

Sites of Cone System Sensitivity Loss in Retinitis Pigmentosa William H. Seiple, Karen Holopigian, Vivienne C. Greenstein, and Donald C. Hood*

Purpose. To examine the sites of cone sensitivity loss in patients with retinitis pigmentosa by comparing focal electroretinographic and psychophysical modulation thresholds. Methods. Both psychophysical and electrophysiologic increment threshold curves were obtained in retinitis pigmentosa patients and a group of age-matched, normally-sighted adults. Results. The majority of the retinitis pigmentosa data could be accounted for by a vertical displacement of the normal curve. The retinitis pigmentosa patients showed similar patterns of cone sensitivity losses using both techniques. Conclusions. The combined electrophysiologic and psychophysical results provide support for an outer retina locus for these cone sensitivity losses. The data suggest that these deficits may be caused by a spatially independent loss of cone photoreceptors with normal adaptation properties in the remaining photoreceptors. Invest Ophthalmol Vis Sci 1993;34:2638-2645.

From, the Department of Ophthalmology, New York University Medical Center and and * Psychology Department, Columbia University, New York, New York. This xuorh was partially supported by a grant from RP Fighting Blindness and by an unrestricted grant from Research to Prevent Blindness, Inc. to the Department of Ophthalmology, NYU Medical, Center. Submitted for publication: August 10, 1992; accepted February 17, 1993. Proprietary interest category: N. Reprint requests: William Seiple, Department of Ophthalmology, BEL5N15, NYU Medical Center, 550 First Ave, New York, NY 10016.

retinal functioning5 and the psychophysical thresholds provide an assay of both receptor and postreceptor integrity. The various mechanisms that have been proposed to account for sensitivity losses in RP patients will not necessarily have the same effect on these two measures. Differential effects on FERG and psychophysical thresholds might be expected even if the deficit is acting at the level of the receptors. For example, a loss of perifoveal cones and a subsequent reduction of the visual field area might be expected to have a large effect on the FERG, but no effect on psychophysical thresholds if they are foveally determined. Therefore, a comparison between the changes in electrophysiologic and psychophysical thresholds provides a powerful tool for examining hypotheses about the sites and mechanisms of sensitivity loss. In this study, the sites of cone system sensitivity loss in RP patients were examined by comparing electrophysiologic and psychophysical increment thresholds obtained under the same stimulus conditions. If the sites of cone sensitivity losses are at the level of the outer retina, then both FERG and psychophysical thresholds should be increased. Deficits at postrecep-

2638

Investigative Ophthalmology & Visual Science, August 1993, Vol. 34, No. 9 Copyright © Association for Research in Vision and Ophthalmology

JL here are reports in the psychophysical literature that cone system sensitivity is decreased in some patients with retinitis pigmentosa (RP).1'2 The sites and mechanisms of these sensitivity losses, however, remain unclear. Although this group of hereditary degenerative retinal diseases primarily affects the photoreceptors, postreceptoral contributions to these deficits have also been suggested.34 One approach toward identifying the retinal sites of cone sensitivity loss in RP is to compare the results of psychophysical and electrophysiologic techniques. In the current experiment, the changes in focal electroretinogram (FERG) thresholds for high temporal frequency stimuli (>20 Hz) caused by RP were compared to the changes in psychophysically elicited thresholds for the same stimuli. The FERG thresholds provide an assay of outer

Cone System Loss in RP toral sites should produce elevated psychophysical thresholds, but normal FERG thresholds. METHODS Subjects Eleven patients with RP participated in this study (mean age = 37 years). All had nondetectable standard full-field electroretinograms (both photopic and scotopic), central visual fields of at least 10 degrees (as measured on a Goldmann perimeter with a III/4e white target) and corrected visual acuities of 20/40 or better (see Table 1 for clinical characteristics). The patients had no significant media or lens opacities, and had no evidence of macular edema on fluorescein angiography. Data were also collected on nine control observers with corrected visual acuity of 20/20 or better and normal ophthalmologic examinations (mean age = 32 years). All subjects gave informed consent to participate in this study; the research followed the tenets of the Declaration of Helsinki and was approved by the New York University Human Subjects Committee. Stimulus The stimulus was a diffused array of red light-emitting diodes (peak wavelength 660 nm) subtending 9 degrees.6 The array was located at the center of an illuminated (white) ganzfeld bowl and the light-emitting diodes were sinusoidally modulated.* For each adapting level, the mean luminance of the light-emitting diode array and the illuminated background were adjusted to be equal using the photopic filter setting of a Spectraspot photometer (Photo Research Co., Burbank, CA).

FERG Recording Monocular FERGs were recorded using a gold-foil electrode referenced to the ipsilateral ear. The contralateral ear served as ground. The data were amplified * There is a nonlinear relationship between voltage input and light output of light-emitting diodes. A consequence of this nonlinearity is that symmetrical modulations of voltage around a mean voltage level will produce asymmetrical modulations of light output around the mean light level. That is, a voltage deviation above the mean will produce a larger increase in light output than the decrease in light output produced by an equal voltage deviation below the mean. This would result in a small increase in the mean adaptation level which would be greatest for large modulations around low mean levels. For example, at 100% modulation around a mean level of 2.0 log td the increase in the mean adaptation level would be 0.1 log td. However, because we are measuring threshold performance, the change in mean adaptation level would be much less than that observed using 100% modulation. For example, at 2.0 log td mean adaptation level, FERG thresholds for the control subjects averaged 27% modulation. At this threshold modulation, the error in calculating the mean adapting level is only 0.02 log td. At the higher mean levels (3.5 and 4.0 log td), light output was linearly related to voltage; therefore, light modulation was symmetrical and there was no change in mean adaptation level.

2639 TABLE

l. RP Patient Characteristics

No.

Age

Acuity

1

47 24 36 30 20 30 40 39 29 43 39

20/2020/30 20/20 20/25 20/2020/30+ 20/40 20/3020/15 20/20 20/20-

2 3 4 5 6

7 8 9 10 11

Visual Field

Mode

10-15

Usher's

>30 >30

ISO AR ISO

10-20 >30 >30

10-15 10 >30 10 20

Usher's AR ISO ISO ISO ISO AR

ISO, isolated. AR, autosomal recessive.

(Gain = 10K) and band-pass filtered between 1 and 100 Hz. Electrophysiologic Threshold Retrieval A swept-stimulus, lock-in amplifier retrieval method was used to estimate FERG thresholds.57 FERG amplitudes were quantified in real time as stimulus modulation was changed from subthreshold (0.01%) to suprathreshold (100%). FERG modulation thresholds were then obtained by extrapolating these amplitude versus modulation functions to the zero response level. Procedure Threshold data were collected at six levels of retinal illuminance, from 1.5 to 4.0 log trolands (td). Before testing, the subject's pupil was dilated with 1% tropicamide (Mydriacyl, Alcon, Ft. Worth, TX) and 2.5% phenylephrine (Mydfrin, Alcon). The eye was dark adapted for 45 minutes and then light adapted to the lowest mean level of adaptation (1.5 log td for both the stimulus and the background) for 2 minutes. The subject's psychophysical thresholds for flicker detection were measured at temporal frequencies between 20 and 50 Hz. The modulation depth of the stimulus was slowly increased from below threshold until the subject reported seeing flicker. Four estimates of flicker threshold were averaged for each stimulus condition. After this, FERG thresholds for only the 30-Hz flicker rate were recorded. FERG amplitude was measured continuously as stimulus modulation depth varied from 0.01% to 100% during a 20-second period. Three thresholds were obtained for each adapting level. The mean luminance of both the test stimulus and background were then increased to the next adapting level. The subject was light adapted to the new level for 2 minutes and psychophysical and FERG thresholds were again obtained. This procedure was repeated for all six adapting levels.

Investigative Ophthalmology & Visual Science, August 1993, Vol. 34, No. 9

2640

Patient # 1 A. FERG

RESULTS Electrophysiologic and psychophysical thresholds for the 30-Hz stimulus were compared for the RP patients and the normal subjects. The increment threshold data for each subject were fitted using the following equation: log T = log To + log[(A+ Ao)/Ao]

n

•o

log A o

log To

2.49

1.89

2.51

1.26

log A o

log Tc

2.73

1.93

2.77

1.22

2

(1)

where T is the threshold illuminance, n is the slope of the diagonal line, To is the unadapted threshold illuminance, which specifies the vertical position of the function, A is the adapting illuminance, and Ao specifies the horizontal position of the function. Initially, when equation 1 was fitted to the data, the value of n was allowed to vary. However, there were no significant differences between the slopes of the two groups. The average slopes for the psychophysical data were 1.0 ± 0.17 for the normal subjects and 1.1 ± 0.36 for the patients (t(18) = -0.78, P = NS): for the FERG data the slopes averaged 1.0 ± 0.25 and 0.9 ± 0.2 (t(18) = 0.92, P = NS), respectively. In addition, we were concerned about allowing n to vary because the elevations in thresholds that occur with disease result in a decrease in the number of data points that can be used to determine the slope, and this can lead to an artificially smaller n.8 Therefore, all of the data presented below were refitted with n set equal to 1.0. Examples of these fits for two RP patients are presented in Figures 1 and 2. The FERG and psychophysical threshold data for patient 1 are shown in Figure 1. The dashed curve shows thefitof equation 1 to the mean data of the control subjects (circles). The solid curve is the bestfitto the patient's data (squares). The values of log To and log Ao for the control and the patient data are presented to the right of the plot. For the patient's FERG data, log To was increased by 0.63 log units and log Ao was equivalent to normal; the patient's psychophysical data was also best fit by an increase in only log To (by 0.71 log units). In other words, only a vertical shift of the control subjects' curve fits this patient's data. The FERG and psychophysical data of patient 2 are shown in Figure 2. For this patient, the fit to the FERG data required an increase in both log To (by 0.79 log units) and log Ao (by 0.63 log units); the psychophysical data were alsofitby an increase both in log To (by 0.58 log units) and log Ao (by 0.74 log units). That is, the average curve of the control subjects must be shifted "up and to the right" by approximately equal amounts to fit this patient's data. Figure 3 summarizes the changes in log To and log Ao for the 30-Hz FERG data for all subjects. Relative log Ao' values were obtained by subtracting the mean log Ao value of the control group from each subject's

0

1

2

3

4

Adapting Illuminance (log td)

B. Psychophysics

•a 2

Adapting Illuminance (log td)

FIGURE l. Focal electroretinogram (A) and psychophysical (B) thresholds for the 30 Hz target plotted as a function of adaptation level for RP patient 1 (squares). Best-fit values of log Ao and log To to the mean of the control subjects' (circles, dotted line) and to the patient's data (solid line) are indicated.

log Ao value (abscissa). Likewise, relative logT o ' values were obtained by subtracting the normal subjects' mean log To value from each subject's log To value (ordinate). This calculation allows a direct comparison of the elevation in log To with the elevation in log Ao for each subject. The values for control subjects are plotted as squares and the patients' values are plotted as circles. The dotted diagonal line is the locus where the points would fall if equal changes in relative logT0' and relative log Ao' were observed. All of the RP patients' data showed an increase in relative log To', with most of the relative log Ao' changes falling within the normal range (dashed box). Only one patient (patient 2, open circle) had FERG data that showed a larger than normal change in relative log Ao'. A summary of the changes in log To and log Ao for the 30-Hz psychophysical data is presented in a similar format in Figure 4. Again, all of the RP subjects

2641

Cone System Loss in RP

Patient # 2 A. FERG log

2

Ao

log To

3.14

2.05

2.51

1.26

2

0

1

2

3

4

Adapting Illuminance flog td)

B. Psychophysics

2

log A o

log T c

3.51

1.80

2.77

1.22

FERG thresholds provide an assay of outer retinal function,5 the similarity between the two sets of results suggests a receptoral locus for these sensitivity deficits. Several possible mechanisms could result in sensitivity losses at the level of the photoreceptors. For example, it has been proposed that sensitivity losses in RP patients may be caused by a loss of photopigment or misalignment of photoreceptors9"13. A reduction of photopigment density would result in a decrease in quantal catch and would therefore decrease the effective intensity of both the stimulus and the background, similar to placing a neutral density filter in front of the eye. This multiplicative scaling of intensity would lead to the increment threshold curves being shifted "up and over" (i.e., the increase in log To would equal the increase in log Ao) so that sensitivity losses would be greatest at lower background levels and the data would approach normal values at the higher levels.31416 To demonstrate the effects of decreased quantal catch and to determine whether the data collected with our procedure are consistent with a decreased quantal catch hypothesis, we compared 30-Hz FERG and psychophysical thresholds with and without a 0.5

2

0)

Psychophysics 1.00 Adapting Illuminance (log td)

0.75

FIGURE 2. Focal electroretinogram (A) and psychophysical (B) thresholds for the 30 Hz target plotted as a function of adaptation level for RP patient 2 (squares). Best-fit values of ilog Ao and log To to the mean of the control subjects' (circles, o dotted line) and to the patient's data (solid line) are indicated.

showed increases beyond the normal range in relative log To', and most had normal relative log Ao' values. Patient 2 (open circle), however, showed a relatively equivalent increase in both relative log To' and relative log A o \ similar to his FERG data. DISCUSSION Although RP is primarily a disease of the photoreceptors and retinal pigment epithelium, evidence has been presented that implicates postreceptoral deficits as well.3'4 In this study, we compared focal electroretinographic and psychophysical temporal modulation thresholds as a function of adaptation level to investigate sites and mechanisms of central cone sensitivity losses in patients with RP. The sensitivity losses we observed in both the electrophysiologic and psychophysical data were qualitatively similar. Because the

+ JO 0

DC

0.50 0.25 0.00 -0.25 -0.50 -0.50

a

L

a

0.00

1

0.50

Relative Log

1.00

Ac

FIGURE 3. The mean log Tofitto the control subjects' FERG data is subtracted from the log To valuefitto each subject's data. These data are plotted against the difference between each patient's log Ao and the log Ao valuefitto the mean of the control subjects' data. The individual differences for the control subjects are plotted as squares and the patients' data plotted as circles. The dotted box encloses the control subjects' data and the dashed diagonal line is the locus of points for equal changes in relative log To' and relative log Ao'.

Investigative Ophthalmology & Visual Science, August 1993, Vol. 34, No. 9

2642

FERG 1.00 0.75 0.50

Relai:ive

Log

• o \-

0.25 0.00 -0.25 -0.50 -0.50

0.00

0.50

1.00

Relative Log Ac FIGURE 4. The mean log To fit to the control subjects' psychophysical data is subtracted from the log To value fit to each subject's data. These data are plotted against the difference between each patient's log Ao and the log Ao valuefitto the mean of the control subjects' data. The individual differences for the control subjects are plotted as squares and the patients' data plotted as circles. The dotted box encloses the control subjects' data and the dotted diagonal line is the locus where the points would fall if equal changes in relative log To' and relative log Ao' were observed.

losses observed in a group of RP patients. In addition, Kilbride et al17, using a densitometric technique, were unable to account for their RP results based on a quantal catch hypothesis. A second hypothesis for sensitivity losses in RP patients is a change in the photoreceptor membrane properties, which may delay reestablishment of equilibrium.1819 We previously demonstrated that the phase of the FERG as a function of temporal frequency in RP patients is normal despite large losses in responsiveness.6 In addition, RP patients' electroretinographic implicit times were found to be within the normal range, although recovery times were prolonged.18 A delayed recovery should reduce response amplitude in proportion to temporal frequency, independent of the level of retinal illuminance. This would cause the difference between the log To values for the RP patients and the normal subjects to increase with temporal frequency. To examine this, we compared the psychophysical data of the two groups as a function of temporal frequency (Figure 6). These data are not consistent with the above hypothesis. The differences between log Tos of the averaged data for the control subjects and the patients are relatively constant with temporal frequency. Reductions in psychophysical sensitivity observed in RP patients have also been attributed to a loss of

0.5 Neutral Density Filter 1.00 PP

neutral density filter (Model 50962, Oriel Corporation, Stratford CT) in four control subjects. The filter's attenuation was flat over the wavelengths employed in this study. The results of this experiment are presented in Figure 5. Relative log Ao' values were derived by subtracting each subject's log Ao value from their log Ao value obtained with the filter. Similarly, the relative log To' values were calculated by subtracting each subject's log To from the log To obtained with the filter. This yielded a measure of the change in log Ao and log To caused by the neutral density filter. The neutral density filter resulted in relatively equivalent increases in log To and log Ao for both the electrophysiologic and psychophysical measures. The data from only one RP patient (patient 2) are consistent with this reduced 'quantal catch' hypothesis (Figs. 3, 4). For the remaining ten patients, relative log To' was increased and relative log Ao' was unchanged. The patients' data, therefore, are not consistent with a decrease in quantum catch of the functioning cones. Greenstein and Hood 315 also found that a decrease in quantal catch did not account for the psychophysical threshold

0.75 0.50 0.25

XT / T

/ ' FERG

0.00 -0.25 -0.50 -0.50-0.25 0.00 0.25 0.50 0.75 1.00 Relative Log Ao

5. The log To valuefitto the threshold data collected without a neutral density filter have been subtracted from log To value fit to the subject's data collected with a 0.5 neutral density filter. The same was done for the log Ao values. The average of four control subjects' log To difference (± 1 SD) is plotted against the average log Ao difference (± 1 SD) (circle = FERG; square = psychophysical). FIGURE

Cone System Loss in RP

2643

A. Log A o 4.0 r

3.0

— Normals .2.0

-- Patients

1.0

0.0 10

20

30

40

50

Temporal Frequency (Hz)

B. Log T n 3.5

2.6

ing value obtained with the smaller target. Because there have been some questions concerning changes in the slopes of these functions with decreasing target size, our data were initially fitted allowing the slope of equation 1 (n) to vary. The average slopes for the psychophysical data were 1.0 ± 0.2 for the large target and 1.0 ± 0.2 for the small target (t(6) = 0.0, P = NS). For the FERG data the average slopes were 0.9 ± 0.2 and 1.0 ± 0.2, respectively (t(6) = -0.40, P = NS). Because there were no statistically significant differences, the data were refitted with n = 1.0. For the psychophysical data, reducing the stimulus area resulted in an increase in both relative log To' and relative log Ao'. These psychophysical data agree with previous reports of psychophysical sensitivity changes as a function of target size.23'24 For the FERG data, relative log To' also increased for the smaller stimulus; however, the value of relative log Ao' did not change. The RP patients' FERG sensitivity losses are consistent with the results observed in control subjects when the contributions of perifoveal photoreceptors was reduced; that is, both showed an increase in relative log

1.7

10° vs 2° Target 1.00

-0.1

// -1.0 10

20

30

40

Temporal Frequency (Hz)

FIGURE 6. Averaged (± 1 SD) log Ao (A) and log To (B) values o obtained by fitting equation 1 to the control subjects' (circles) and patients' (square) psychophysical thresholds as a function of temporal frequency.

0.50

CD

DC

/ //

FERG

T

<

0.25

/

/ PP

/

_ro

cone photoreceptors. Two patterns have been observed: a loss of the perifoveal cone photoreceptors with a relative sparing of the central cones, and a spatially independent loss of photoreceptors throughout the macular area.9-20"22 A third testable hypothesis, a loss of perifoveal photoreceptors, was mimicked by recording modulation thresholds in response to a smaller diameter stimulus. The use of a smaller diameter stimulus reduces the contributions of perifoveal cone photoreceptors while retaining the contribution from intact central photoreceptors. Thresholds for a 10-degree stimulus were compared to thresholds for a 2-degree stimulus in four control subjects. The results of this experiment are presented in Figure 7. Relative log Ao' and relative log To' values were calculated by subtracting the value obtained with the larger target from the correspond-

/

0.75

50

0.00 -0.25

-0.50 -0.50-0.25 0.00 0.25 0.50 0.75

1.00

Relative Log Ao FIGURE 7. The log To fit to the threshold data from a 10-degree target have been subtracted from log To fit to the subject's data collected with a 2-degree target. These data are plotted against the difference in the log Ao with the 10 degree target subtracted from the log Ao obtained for the 2-degree target. The points plotted are the average differences (± 1 SD) for the psychophysical (square) and FERG (circle) for four control subjects. The dotted diagonal line is the locus where the points would fall if equal changes in relative log To' and relative log Ao' were observed.

2644

Investigative Ophthalmology & Visual Science, August 1993, Vol. 34, No. 9

To' only (compare Figures 3 and 7). The RP patients' psychophysical data are not consistent with this hypothesis. Reducing the peripheral cone contribution in control subjects increased both relative log To' and relative log Ao'; whereas the RP patients' psychophysical thresholds exhibited an increase in relative log To' only (Figure 4). Therefore, reduced peripheral cone contribution cannot solely account for both sets of data. A spatially independent drop-out of cone photoreceptors within the macular area underlying our stimulus provides a fourth possible explanation for our results.25 A decrease in the total number of photoreceptors pooling to produce a massed electrophysiologic response, but with the remaining elements adapting normally, would result in a proportional reduction in amplitude at all temporal frequencies and at all levels of retinal adaptation. This would produce an increase in log To only and is consistent with our FERG data. For drop-out of cone photoreceptors to account for our psychophysical findings, we have to make assumptions concerning both detection and the site(s) and amount(s) of adaptation. We assume that detection is mediated by a pooled contribution of elements, rather than by independent element(s). Regarding adaptation, we assume that the site is at the level of the outer retina. Given these assumptions, a spatially independent photoreceptor drop-out would result in increases in log To only.8 Electrophysiologic evidence supports an outer retinal locus of gain. Valeton and van Norren26 reported significant gain changes at the level of the photoreceptors when measuring local electroretinograms and late receptor potentials. Hood and Birch27 have found evidence supporting a cone photoreceptor locus of gain in the a-wave of the human electroretinogram. It has also been demonstrated that substantial temporal frequency dependent gain occurs at the level of the photoreceptors when high temporal frequency, sinusoidally modulated stimuli are used.528'29 A model of light adaptation that places most of the gain at the level of the receptors is consistent with both our electrophysiologic and psychophysical findings. A spatially independent loss of photoreceptors within the macula not only accounts for our data, but can also account for data obtained from RP patients using a 'Stiles-type' increment threshold paradigm.3 To account for RP patients' increment threshold data, a small postreceptoral gain component must be assumed, in addition to assuming a receptoral site for gain.3 The need for an additional postreceptoral gain component may be attributable to differences between 'Stiles-type' increment threshold paradigms and the paradigm employed in the current study. These include: differences in stimulus duration and wave-

shape, in the interstimulus interval, and in the size of the adapting field. All of these factors have been shown to exert major influences on psychophysically measured sensitivity.30 In summary, RP patients exhibited similar psychophysical and FERG sensitivity losses as a function of background illuminance. This finding suggests that the primary site of cone sensitivity losses observed in RP patients is at the level of the outer retina. By comparing psychophysical and electrophysiologic thresholds, we were able to test and reject several possible mechanisms of sensitivity loss. We have eliminated explanations based solely on decreases in quantal catch, changes in membrane properties, and decreases in perifoveal cone density. The results of this study are consistent with mechanisms based on spatially independent loss of receptors, with the remaining cones having normal adaptation properties. Key Words retinitis pigmentosa, temporal frequency, cone sensitivity, focal electroretinogram References 1. Kayazawa F, Yamamota T, Itoi, M. Temporal modulation transfer functions in patients with retinal disease. Ophthalmic Res. 1982; 14:409-416. 2. Tyler CN, Ernst W, Lyness AL. Photopic flicker sensitivity losses in simplex and multiplex retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1984; 25:1035-1042. 3. Greenstein VC, Hood DC. The effects of light adaptation on L-cone sensitivity in retinal disease. Clin Vis Sci. 1992;7:l-7. 4. Stone JL, Barlow WE, Milam AH. Morphometric analysis of retinal ganglion cells in retinitis pigmentosa. Invest Ophthalmol Vis Sci. ARVO Abstracts. 1992; 33:1397. 5. Seiple W, Holopigian K, Greenstein V, Hood DC. Temporal frequency dependent adaptation at the level of the outer retina in humans. Vision Res. 1992;32:2043-2053. 6. Seiple WH, Siegel IM, Carr RE, Mayron C. Evaluating macular function using the focal ERG. Invest Ophthalmol Vis Sci. 1986;27:1123-1130. 7. Nelson JI, Seiple WH, Kupersmith MJ, Carr RE. Lock-in techniques for the swept stimulus evoked potential. J Clin Neurophysiol. 1984; 1:409-415. 8. Hood DC, Greenstein VC. Models of the normal and abnormal rod system. Vision Res. 1990; 30:51-68. 9. Kolb H, Gouras P. Electron microscope observations of human retinitis pigmentosa, dominantly inherited. Invest Ophthalmol Vis Sci. 1974; 13:487-498. 10. Ripps H, Brin KP, Weale RA. Rhodopsin and visual threshold in retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1978; 17:735-745. 11. Perlman I, Auerbach E. The relationship between visual sensitivity and rhodopsin density in retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1981;20:758-765.

Cone System Loss in RP 12. van Meel GJ, van Norren D. Foveal densitometry in RP. Invest Ophthalmol Vis Sci. 1983;24:1123-1130. 13. Gouras P, MacKay CJ. Light adaptation of the electroretinogram: Diminished in retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1989;30:619-624. 14. Hood DC, Greenstein VC. An approach to testing alternative hypotheses of changes in visual sensitivity due to retinal disease. Invest Ophthalmol Vis Sci. 1982;23:96-101. 15. Greenstein VC, Hood DC. Test of decreased responsiveness hypothesis in retinitis pigmentosa. AmerJ Optom Physiol Optics. 1986;63:22-27. 16. Hood DC. Testing hypotheses about development with electroretinographic and incremental-threshold ddU.JOpt Soc Am. 1988; 5:2159-2165. 17. Kilbride PE, Fishman M, Fishman GA, Hutman LP. Foveal cone pigment density difference and reflectance in retinitis pigmentosa. Arch Ophthalmol. 1986; 104:220-224. 18. Seiple W, Greenstein V, Carr R. Temporal sensitivity losses in heredoretinal disease: Tests of hypotheses. Br J Ophthalmol. 1989; 73:440-447. 19. Dagnelie G, Massof RW. Foveal cone involvement in retinitis pigmentosa progression assessed through llash-on-flash parameters. Invest Ophthalmol Vis Sci. 1993;34:231-242. 20. Flannery JG, Faber DB, Bird AC, Bok D. Degenerative changes in a retina affected with autosomal dominant retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1989;30:191-211. 21. Szamier RB, Berson EL. Retinal ultrastructure in ad-

2645

22.

23. 24.

25. 26. 27.

28. 29. 30.

vanced retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1977; 16:947-962. Szamier RB, Berson EL, Klein R, Meyers S. Sex-linked retinitis pigmentosa: Ultrastructure of photoreceptors and pigment epithelium. Invest Ophthalmol Vis Sci. 1979;18:145-160. Barlow HG. Dark and light adaptation: psychophysics. In: Jameson D, Hurvich LM, eds. The Handbook of Sensory Physiology, Vol 7. Berlin: Springer-Verlag; 1972. Chen B, MacLeod DIA, Stockman A. Improvement in human vision under bright light: grain or gain?y Physiol. 1987;394:41-66. Sandberg MA, Berson EL. Visual acuity and cone spatial density in retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1983;24:1511-1513. Valeton JM, van Norren D. Light adaptation of primate cones: an analysis based on extracellular data. Vision Res. 1983;23:1539-1547. Hood DC, Birch D. Changes in the gain and time constant of human cone photoreceptors with light adaptation. Advances in Color Vision, Technical Digest, Optical Soc Am. 1992;4:10-12. Baron WS, Boynton RM. Response of primate cones to sinusoidally flickering homochromatic stimuli. J Physiol. 1975; 246:311-331 Abraham FA, Alpern M, Kirk DB. Electroretinograms evoked by sinusoidal excitation of human cones.y Physiol. 1985;363:135-150. Brown JL. Flicker and intermittent stimulation. In: Graham, CH, ed. Vision and Visual Perception. New York: John Wiley and Sons, Inc; 1965:215-320.