Quantifying rod photoreceptor-mediated vision in retinal
degenerations: dark-adapted thresholds as outcome measures
Alejandro J. Roman1, Sharon B. Schwartz1, Tomas S. Aleman, Artur V. Cideciyan,
John D. Chico, Elizabeth A.M. Windsor, Leigh M. Gardner, Gui-shuang Ying,
Elaine E. Smilko, Maureen G. Maguire, Samuel G. Jacobson*
Department of Ophthalmology, Scheie Eye Institute, University of Pennsylvania, Philadelphia, PA 19104, USA
Received 20 July 2004; received in revised form 15 September 2004; accepted 21 September 2004
Available online 18 November 2004
Pre-clinical trials of treatment in retinal degenerations have shown progress toward preventing loss or restoring function of rod
photoreceptors. In anticipation of human clinical trials, we assessed two psychophysical methods of quantifying rod photoreceptor-mediated
function as potential outcome measures. Modified automated perimeters were used to deliver focal or full-field light stimuli and dark-adapted
thresholds were measured. Patients with retinal degeneration were studied in two experimental protocols. Experiment 1 (nZ35 patients)
studied dark-adapted focal chromatic stimuli in central retinal locations along the horizontal meridian. Experiment 2 (nZ146 patients)
studied dark-adapted responses to a full-field stimulus test (FST) using white and chromatic stimuli. Patients in both experimental groups had
testing on two different visits to determine inter-visit variability. In Experiment 1, two subgroups of patients were identified: a group with a
majority of test loci detected by rod photoreceptors and a group with only cone-mediated detection. Inter-visit variability (95% confidence
interval) was G3$1 dB for normals, G3$0 dB for patients with rod-mediated function and G2$8 dB for patients with only cone-mediated
function. In Experiment 2, the dynamic range of the FST using white stimuli was sufficient to quantify sensitivity in all patients studied,
including those with severe retinal degenerations. Chromatic stimuli in the FST were detectable by 85% of patients and rod- or cone-
mediation could be determined. Regional retinal sources of FST were explored by comparing FST and dark-adapted perimetry in the same
patients; there was a strong correlation between FST level and the loci with highest sensitivity by perimetry. Inter-visit variability (95%
confidence interval) in the patients was G3$9 dB compared to G3$5 dB in normals. Dark-adapted focal threshold measurements with an
abbreviated protocol in retinal degeneration patients with stable fixation may be useful as an outcome measure for therapies that can affect
rod vision. FST measurements were feasible and reproducible in a large spectrum of retinal degenerative diseases and will be most applicable
as a psychophysical outcome measure for treatment trials of very severe disorders in which fixation is lost and there is need for a large
dynamic range of stimulus intensity.
q 2004 Elsevier Ltd. All rights reserved.
Keywords: cone; Leber congenital amaurosis; perimetry; retinitis pigmentosa; rod
Progress toward therapy for retinal degenerative diseases
has accelerated in recent years and clinical trials of novel
therapeutic strategies are ongoing or planned (Bessant et al.,
2001; Margalit et al., 2002; Humayun et al., 2003; Weleber
et al., 2003; Delyfer et al., 2004). Some future therapies will
target the rod photoreceptor or are expected to affect rod
function, thus making rod photoreceptor-mediated vision
worth measuring to determine efficacy as well as safety. To
date, the primary outcome measures in larger trials of
nutrient therapies in retinitis pigmentosa (RP) have been
cone function measurements (Berson et al., 1993, 2004;
Hoffman et al., 2004).
Rod-specific measures of visual function that are
potentially useful for assessing therapeutic effects in retinal
degenerative diseases, such as RP, include the rod
0014-4835/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
Experimental Eye Research 80 (2005) 259–272
* Corresponding author. Dr Samuel G. Jacobson, Scheie Eye Institute, 51
North 39th Street, Philadelphia, PA 19104, USA.
E-mail address: firstname.lastname@example.org (S.G. Jacobson).
1These authors contributed equally to the work.
electroretinogram (ERG) and rod-mediated psychophysics.
The rod ERG b-wave, a surrogate measure of rod vision
has been shown to be a robust measure (Berson et al., 1985,
1993; Birch et al., 1991, 1999, 2002; Grover et al., 2003;
Hoffman et al., 2004). Once peripheral retinal function
becomes impaired in RP, there may be no measurable rod
ERG and this can occur at relatively early disease stages.
Although bright light stimuli could elicit residual rod ERG
function, the recorded signal is complex and tends to have
bothrodandcone components.Unlikethe coneflickerERG,
the rod ERG is not amenable to the strategy of delivering
scores of stimuli at short inter-stimulus intervals to record
small signals by averaging methods because of complexities
introduced by light adaptation.
Dark-adapted psychophysics has a long history of study
by visual scientists seeking knowledge of the limits of visual
perception (Cornsweet, 1970) and has a traditional role in
clinical settings to determine dark-adapted thresholds,
usually to a relatively large achromatic stimulus within
the central field (Marmor et al., 1983). Dark-adapted
chromatic perimetry has been used in many investigations
of retinal degeneration to distinguish rod- from cone-
mediated function across the visual field of patients of
unknown genotypes and subsequently in those with known
molecular causes (e.g. Gunkel, 1967; Massof and Finkel-
stein, 1979; Lyness et al., 1985; Jacobson et al., 1986, 1991,
2000; Birch et al., 1987; Cideciyan et al., 1998). It would
seem logical to try to transfer a version of the technique
from this research role to that of a clinical outcome measure.
Progress toward this goal has already been made (Birch
et al., 1999).
The current work addresses the need for psychophysical
methodology to assess rod photoreceptor-mediated vision
for future clinical trials in patients with retinal degeneration.
Two different dark-adapted psychophysical strategies are
assessed for feasibility as candidate outcome measures: one
using focal stimuli which provide spatial information about
function in the central retina, and a second that employs a
full-field stimulus and thereby no spatial information.
2. Materials and methods
Patients with inherited retinal degenerations (Table 1)
and visually normal subjects participated in the exper-
iments. In Experiment 1, 35 patients (16 female, 19 male)
and 20 normal subjects (11 female, 9 male; ages 19–55)
underwent dark-adapted testing with focal chromaticstimuli
in the central retina in one eye on two visits separated by
no more than 6 months. In Experiment 2, 146 patients
(75 female, 71 male) and 12 normal subjects (6 female, 6
male; ages 10–56) underwent a full-field stimulus test (FST)
in one eye. A subset of 41 of these patients and 10 normal
subjects returned within 6 months for repeat testing. Testing
and retesting were performed at similar times of the day in
both experimental groups. All subjects had complete
ophthalmic examinations. Also performed in all patients
prior to entry into the studies were kinetic perimetry (V-4e
and I-4e test targets) and ERGs using a standard protocol
(Jacobson et al., 1989); in all but the patients with the most
severe retinal degenerations, dark-adapted two-color peri-
metry on a 128 grid throughout the visual field was also
performed (Jacobson et al., 1986). All studies were
approved by the Institutional Review Board and informed
consent was obtained from all subjects after explanation of
the nature of the studies. The research procedures were in
accordance with institutional guidelines and the Declaration
2.2. Testing procedures
Experiment 1 Thresholds were obtained in the dark-
adapted (R45 min) state using 500 and 650 nm (1$78
diameter; 200 ms duration) stimuli on a Humphrey Field
Analyzer (model 600 series; Zeiss-Humphrey Instruments,
Dublin, CA) modified for this purpose (Jacobson et al.,
1986). The pupils of the subjects were fully dilated
(tropicamide 1%; phenylephrine 2$5%). Light-adapted
testing with the same instrument but using a white stimulus
early in the visit served as an explanation of the test
and provided practice at performance. The full threshold
strategy, a staircase algorithm available on the perimeter,
was used for two-color dark-adapted static perimetry. With
this bracketing strategy, stimulus intensity is initially varied
in steps of 4 dB and then in steps of 2 dB. The final
threshold estimate is the intensity of the last presentation
seen by the patient. Details and methods of analysis have
been described (Jacobson et al., 1986). Horizontal profiles,
representing 25 loci, were measured across the central 608 at
3516/19 11–61 (34) 35
aFifteen subjects are included in both Experiments 1 and 2.
bIncludes retinitis pigmentosa (nZ24), Usher syndrome types I, II
(nZ3), cone–rod dystrophy (nZ2), choroideremia (nZ6).
cIncludes retinitis pigmentosa (nZ78), Leber congenital amaurosis or
early onset retinal degeneration (nZ15), Usher syndrome types I, II (nZ8),
Bardet-Biedl syndrome (nZ3), choroideremia (nZ8).
dIncludes Stargardt disease (nZ13), autosomal dominant macular
degeneration (nZ3), early onset macular drusen (nZ2).
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272260
28 intervals (excluding five loci at and around the blindspot).
Testing time for each profile in normal subjects and patients
with mainly rod-mediated function averaged 8$6 min
(range, 7–12 min). The number of stimulus presentations
per profile averaged 220 (range, 170–300). Testing time for
those with only cone-mediated function averaged 4$5 min
(range, 3–5$5 min) and number of stimulus presentations
averaged 118 (range, 75–150). Fixation could not be
monitored in the dark-adapted state with the Heijl-Krakau
blindspot method because targets projected in the blindspot
can elicit false responses to stray light. Fixation was
monitored throughout examinations by infrared viewing of
the patient’s test eye and frequent reminders were given to
maintain eye position; pauses in the testing (approximately
every 2 min) occurred to avoid fatigue. False positive
responding in patients and normals was similar to each other
and similar on the two visits: 80–90% of subjects had no
false positives. False negative responding in patients and
normals followed the same pattern on both visits: 80–90%
had no false negatives.
Experiment 2. Thresholds to a full-field stimulus were
obtained in the dark-adapted state using white, red or blue
flashes (200 ms duration) and the same full threshold
strategy as in Experiment 1. The stimuli were delivered
with an automated perimeter (Humphrey Field Analyzer;
model 750i; Zeiss-Humphrey Instruments, Dublin, CA)
modified for this purpose (see Appendix A for details of
instrument modifications). Chromatic testing was per-
formed to determine whether there was rod- or cone-
mediation of detection. The color filters in the perimeter
were used unmodified and the spectral characteristic of
these broadband filters was measured (Appendix A,
Pre-testing with the white stimulus was used for
explanation of the test and practice at performance.
Measurements were made for each test eye in the following
sequence: white, white, blue, red, white, blue, red, white.
Testing time per eye for this sequence did not exceed 8 min.
A subset of normal subjects (nZ2) underwent dark
(bleaching) adaptation functions following exposure to
light (full-field white flash expected to isomerize 97% of
rhodopsin; Cideciyan et al., 2004) to determine what the
sensitivity difference would be between blue and red stimuli
on the cone plateau (w4–13 min after the flash) and on rod
portions of the function; this enabled interpretation of
results of chromatic testing in the patients.
The relationship of FST level (using white stimuli) to
results of dark-adapted two-color perimetry (72 loci) on a
128 grid (Jacobson et al., 1986; Apathy et al., 1987) from the
same patients was determined. Sensitivity loss was used to
compare the two sets of data. FST loss was defined as the
difference between mean normal and patient FST sensi-
tivity. Rod sensitivity losses at individual loci in the visual
field were at those loci with rod- or mixed- (500 nm detected
by rods; 650 nm detected by cones) mediation and results
(from the 500 nm responses) were plotted as a function
of eccentricity in the field for each patient. Mean field rod
(500 nm) sensitivity losses were also calculated and plotted
for comparison with FST losses. For the entire patient
population, FST sensitivities were plotted versus the highest
sensitivity (to 500 nm) among the 72 loci tested in dark-
2.3. Data analyses
Experiment 1. Difference between dark-adapted sensi-
tivity to 500 and 650 nm stimuli at each locus was used to
determine whether both stimuli were being detected by rods
or by L/M (long/middle wavelength) cones, or there was
mixed rod- and cone-mediation (Jacobson et al., 1986).
Briefly, loci with sensitivity differences (500 nm minus
650 nm) greater than or equal to 28 dB were classified as
mediated by rods, whereas differences equal to or lower than
12 dB indicated cone-mediation. Values between these
limits were considered to have mixed rod- and cone-
mediation. In the calculation, 16 dB was added to the
measured sensitivity levels of the 500 nm stimulus to equate
the energy of this stimulus to that of the 650 nm stimulus.
Complete details of the calculations leading to determining
photoreceptor mediation have been published (Jacobson
et al., 1986; Apathy et al., 1987). Patients were divided into
two subgroups for further analyses: those with cone-
mediated fields and those with most loci having rod-
mediation to the 500 nm stimulus; the latter subgroup
usually also showed a minority of cone-mediated loci
around fixation. Inter-visit variability was examined using
the mean sensitivity of the dark-adapted profiles. For the
rod-mediated group (nZ24), mean sensitivity was derived
from sensitivities at extrafoveal loci using the 500 nm
stimulus. Loci were categorized as having no measurable
sensitivity if rod sensitivity loss was greater than 30 dB
based on mean locus-specific normal data. For the cone-
mediated group (nZ11), the mean sensitivity was derived
from sensitivities at extrafoveal and foveal loci using the
650 nm stimulus because of the greater dynamic range of
results with this stimulus color (Jacobson et al., 1986).
Previous work emphasizing the artifactual truncation of
variability estimates from floor effects (i.e. unmeasurable
vision due to limited dynamic range of perimeter or large
defect depth; Chauhan and Johnson, 1999; Spry et al., 2001;
Blumenthal et al., 2003) prompted us to analyse the dark-
adapted profile results using not only the full number of loci
(608 field) but also an abbreviated set of central loci (168
field) with mostly measurable vision. Inter-visit variability
was quantified and graphically displayed as twice the pooled
standard deviation of the difference distribution (visit 2
minus visit 1) with the Bland-Altman analysis (Bland and
Altman, 1986). Measurement variability of patient and
control groups was also compared with an F-test of equal
variance. Differences in the mean between the two visits
were examined with a paired t-test (PC SAS; version 8.01,
SAS Institute, Cary, NC).
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272 261
Experiment 2. Within-visit variability of FST data was
session with the white stimulus. Variability was quantified as
the within-subject standard deviation or measurement error
of variance. To examine the effect of sensitivity level on
within-visit variability, we plotted the individual subjects’
standard deviations against their means and evaluated the
association by Spearman rank correlation coefficient. Inter-
visit variability of the mean of FST sensitivities to the white
stimulus was determined with the Bland–Altman analysis
effect of sensitivity level on inter-visit variability, we divided
average sensitivity of the two visits (using the median as the
cutpoint) and compared the groups with an F-test for equality
of variances. Measurement variability of patient and control
groups was also compared using an F-test of equal variance;
with a paired t-test.
Table 1 lists clinical diagnoses of the patients included in
the two experimental protocols. For Experiment 1, all
patients had central fixation. For Experiment 2, the test
strategy did not necessitate stable central fixation, so
patients with maculopathy and severe retinal degenerations
that involved both central and peripheral retina were also
studied. The large number of patients with different retinal
degenerations and at different disease stages included in
Experiment 2 provided answers to questions about regional
retinal sources and variability of this newly designed dark-
adapted full-field psychophysical test.
3.1. Experiment 1: dark-adapted focal stimulus test results
Fig. 1A shows representative dark-adapted two-color
profiles of sensitivity across the horizontal meridian in a
normal subject and two patients with retinal degeneration.
In a 22-year-old female normal subject (Fig. 1A, left),
Fig. 1. Dark-adapted focal testing at loci along the horizontal meridian in retinal degenerations. (A) Representative sensitivity profiles in a normal subject and
two patients with retinal degeneration (Patients 1 and 2) using 500 nm (squares) and 650 nm (triangles) stimuli. The photoreceptor mediation at each locus,
based on the sensitivity difference between the two colors, is given: R, rod-mediated; M, mixed rod- and cone-mediated; C, cone-mediated. Gray zone
represents the central 168 of visual field used for calculating mean sensitivities. Hatched area is region around and including the physiological blindspot. F,
fovea; N, nasal; T, temporal. (B) Mean sensitivities (central gray zone; 500 nm stimulus) in the 35 patients compared with a group of 20 normal subjects.
Patients are subdivided by whether their function is mainly mediated by rod photoreceptor pathways (rod-mediated; filled bars) or L/M cone pathways (cone-
mediated; unfilled bars). Bars on vertical axes show mean normal sensitivity (G2SD) for rod-mediation (R, left) and cone-mediation (C, right). Patients 1 and 2,
whose data are shown in (A), are identified.
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272262
extrafoveal sensitivities for both 500 and 650 nm stimuli are
mediated by rod (R) photoreceptors. A 30-year-old female
patient with Usher syndrome (Fig. 1A, Patient 1) has
reduced sensitivities to the 500 nm stimulus and detection is
mediated by rod photoreceptors except at the foveal locus,
which is cone-mediated. Beyond 128 and 188 in the nasal
field, there is a transition from high sensitivity to no
measurable vision; after 88 in the temporal field, there is the
designated blindspot region and thus no data. A 38-year-old
male patient with RP, in contrast, has no measurable rod-
mediated function and the loci with detectable sensitivity
are cone-mediated (Fig. 1A, Patient 2). Peak sensitivity is at
the foveal locus and there is a decline in sensitivity to no
measurable vision within 108 nasal and temporal from the
fovea. We quantified the horizontal extent of functioning
retina in the two patient subgroups to limit further data
analyses to zones of measurable vision. Among the patients
with rod-mediated detection, R80% of eyes had measurable
function at eccentricities up to 48; the percentage declined to
R45% by 88 and was lower at further eccentricities. Among
the patients with only cone-mediated function, there was
also a decline in percentage of eyes with measurable
function from R90% at 48 to R45% at 68 eccentricity; by
88, only about 20% of eyes had responses. The central 168 of
field tested thus had the most measurable function in both
patient subgroups and was used for further analyses (gray
zone in Figs. 1 and 2A). The spectrum of dysfunction within
the central 168 is shown for patients (Fig. 1B). Within the
rod-mediated group, there are patients whose data fall
within normal limits (mean normal, 47$1 dB; SD, 1$72 dB),
but 22 of 24 patients are abnormally reduced in sensitivity.
There is also a range of dysfunction within the subgroup of
patients with only cone-mediated function; all patients have
mean data that are subnormal.
Dark-adapted profiles from two different visits for a
34-year-old normal male subject and the previously
depicted patients with retinal degeneration show that, by
inspection, the results of testing and repeat testing appear
similar (Fig. 2A). Inter-visit variability analyses in normal
subjects and the two subgroups of patients are shown
(Fig. 2B). We used the mean sensitivity of the central test
region with highest percentages of measurable loci
(central 168 of field; gray zone in Figs. 1 and 2A), so as
not to introduce floor effects from the unmeasurable zones
that would cause artifactually reduced estimates of
variability (Chauhan and Johnson, 1999; Spry et al.,
2001; Blumenthal et al., 2003). For the 20 normal subjects
(Fig. 2B, left), the mean sensitivity difference (GSD) is K
0$01 (G1$55) dB. The difference in mean sensitivity is
not significantly different from zero (pZ0$97). For the
subgroup of 24 patients with mainly rod-mediated
function (Fig. 2B, middle), the mean sensitivity difference
(GSD) is 0$52 (G1$52) dB, which is not different from
zero (pZ0$11). For the subgroup of 11 patients with only
cone-mediated function (Fig. 2B, right), inter-visit varia-
bility analyses (using 650 nm stimulus data) have a mean
sensitivity difference (GSD) of 0$14 (G1$40) dB and this
is not significantly different from zero (pZ0$74). A
comparison of measurement variability between the three
groups shows that rod-mediated and cone-mediated groups
do not differ (pZ0$80); there is also no difference
between inter-visit variability in the normal group and
either the rod-mediated group (pZ0$93) or cone-mediated
group (pZ0$76). If all 24 extrafoveal loci from the dark-
adapted profiles are used in the calculation of inter-visit
variability, the following mean sensitivity differences (G
SD) are obtained: normal subjects, K0$34 (G1$01) dB;
rod-mediated patients, 0$20 (G0$75) dB; cone-mediated
patients, 0$06 (G0$51) dB.
3.2. Experiment 2: dark-adapted full-field stimulus
test (FST) results
There was sufficient dynamic range of the FST using
white stimuli to quantify visual function in all patients
studied. As expected from such a wide spectrum of diseases
and severities, there were different levels of visual function
ranging from normal to nearly six log units subnormal
(Fig. 3, upper panel). In the group of 20 patients with
Fig. 2. Inter-visit variability of dark-adapted focal testing in retinal
degenerations. (A) Representative sensitivity profiles across the horizontal
meridian on two different visits in a normal subject and the two patients
with retinal degeneration (from Fig.1). Gray zone delimits central 168 of
visual field used for inter-visit variability analyses. Visit 1, continuous
lines; visit 2, dashed lines. F, fovea; N, nasal; T, temporal. (B) Sensitivity
difference between visits (visit 2 minus visit 1) as a function of mean
sensitivity of the two visits in 20 normal subjects (unfilled circles; 500 nm
data), 24 patients with rod-mediated function (filled circles; 500 nm data)
and 11 patients with cone-mediated function (filled triangles; 650 nm data).
Dotted line is mean sensitivity difference; dashed lines representG2SD.
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272263
sensitivities that were within normal limits (mean normal,
63$9 dB; SD, 2$29 dB), there were 12 patients with the
clinical diagnosis of inherited maculopathy, five patients
with cone-rod dystrophy (CRD), and three with forms of
RP. Among the 126 patients with subnormal FST results, the
most frequent clinical diagnosis was RP (75 patients; 60%).
FST testing was feasible in all patients studied.
Chromatic FST results were also obtained to determine
the feasibility of estimating photoreceptor mediation of
the full-field responses (Fig. 3, lower panel). The
difference in sensitivities to blue and red stimuli (e.g.
Massof and Finkelstein, 1979; Ernst et al., 1983; Jacobson
et al., 1986; Birch et al., 1987) is used to determine rod
versus L/M cone mediation. Normal dark-adapted sub-
jects, who would be expected to use rod-mediated vision
to detect these stimuli, show higher sensitivity to blue
than to red; the difference in sensitivities averages 10$2
(SD, 1$3 dB). The sensitivity difference for L/M cone
mediation of FST chromatic stimuli was determined in
two normal subjects tested on the cone plateau of a dark
adaptation function; both showed red sensitivity greater
(by 10 and 8 dB) than that of blue. In the 125 patients
who were able to detect both blue and red stimuli, 105
patients showed differences that were positive and this is
taken as evidence that rods were the predominant
photoreceptor type mediating FST detection (assuming
that responses to the two stimuli were from the same
retinal region). Patients with the lowest FST results to
white tended to show a different response pattern to the
chromatic stimuli than those with higher sensitivities.
These patients had higher sensitivity to red than blue,
making it likely that their residual vision was L/M cone-
mediated (also confirmed in many such patients by dark-
adapted two-color perimetry; data not shown).
What are the FST responses in patients with more
severe retinal degenerative disease, defined as having both
a non-detectable ERG and very limited visual field extent
by kinetic perimetry? Among the 30 patients with severe
retinal degenerations so defined, all had detectable FST
responses using white stimuli (Fig. 3). In the 17 patients
who were able to detect both blue and red stimuli, 9
patients (53%) were rod-mediated and 8 (47%) were cone-
mediated; a further 10 patients only detected the red
stimulus and were likely to have cone-mediation. The
subgroup of patients with the diagnosis of Leber
congenital amaurosis or early onset retinal degeneration
(nZ15; ages 11–47) is worthy of specific mention. Some
patients in this clinical diagnostic category had markedly
reduced but detectable ERGs and therefore were not
considered severe by the above definition. By kinetic
perimetry (V-4e test target), the 15 patients could retain
detectable small central islands or peripheral islands or
both. FST results to the achromatic stimulus were all
abnormally reduced. Chromatic FST showed that nine
patients detected both blue and red stimuli and eight of
these were rod-mediated and one cone-mediated; six only
Fig. 3. Dark-adapted full-field stimulustest (FST) results in retinal degenerations. Upper. Spectrumof FST sensitivities in normal subjects (nZ12) and patients
withretinaldegenerations (nZ146),rankedfromhigh(left)to low (right)sensitivity.Patientswiththemost severeretinaldegenerations (nZ30),definedby no
detectable clinical ERG and very reduced extent of visual fields to both I-4e and V-4e test targets, are marked (black lines). Hatched bar on vertical axis
represents mean normal FST G2SD Lower. Chromatic FST results are shown below each individual’s white FST result as the difference between blue and red
sensitivities. Like the normal results, most patients with higher white FST results have rod-mediation by the chromatic comparisons; the more severely affected
patients may have rod- or cone-mediation. Normal dark-adapted cone results were obtained by performing the test on the cone plateau of dark adaptation
following light exposure.
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272 264
detected the red stimulus and were likely to have cone-
The regional retinal sources of FST were explored to test
the hypothesis that FST responses originate from the most
sensitive retinal areas, independent of where in the retina
these areas are located (Fig. 4). Data from three patients
illustrate the relationship between FST results for white
stimuli (all rod-mediated by blue-red sensitivity differences
of 8–10 dB) and the sensitivity of individual rod-mediated
loci determined by dark-adapted two-color perimetry. To
compare perimetric and FST data, we graph rod sensitivity
losses (RSL) against eccentricity and also displayed the FST
result as loss. The average RSL across the visual field is also
shown, mainly to illustrate the lack of relationship of this
parameter with FST (Fig. 4A–C). A 16-year-old man with
autosomal recessive CRD and visual acuity (VA) of 20/60
had a large relative central scotoma and some limitation of
peripheral function by kinetic perimetry (Fig. 4A). A map of
RSL and the graph of RSL versus eccentricity show that
beyond the prominent central dysfunction, there was less
severe loss in the mid-peripheral field but increasing
dysfunction again in the far periphery. FST loss falls
among the loci with least RSL on the eccentricity plot
but differs by about one log unit from the average RSL.
A 33-year-old man with simplex RP and VA of 20/20 had
kinetic perimetry that showed a central island separated
from far temporal peripheral islands by a complete annular
mid-peripheral scotoma (Fig. 4B). The map of RSL and the
graph indicate little if any central dysfunction but severe
losses in the mid- and far-peripheral field. FST loss is
accounted for by the cluster of loci with minimal or no RSL
within the central 20 degrees; average RSL differs by at
least 2 log units. A 45-year-old man with X-linked RP and
VA of 6/200 retained an incomplete peripheral island of
vision by kinetic perimetry (Fig. 4C). A map of RSL and
comparison of FST loss with RSL from different eccentri-
cities indicates that the loci in the peripheral field with
minimal RSL relate most closely to the level of FST
Fig. 4. Regional retinal sources of FST results. Spatialdata from the visual fields of three representative patientswith retinal degeneration (A–C) are shown and
comparedwith FSTresults.For each patient,thereis a kinetic fieldusingtwo target sizes (upperleft in each datagroup),grayscale mapofrod sensitivity losses
from dark-adapted perimetry (upper right), and graph of rod sensitivity losses versus eccentricity (lower). Relative scotomas are shown as gray in kinetic fields.
Gray scale for maps of rod sensitivity loss is shown to the right of (B). AvgZaverage loss based on 72 rod sensitivity loss measurements from the dark-adapted
perimetry. FSTZFST loss based on comparison of mean normal and patient FST sensitivity. (D) Testing the hypothesis that FST results are derived from the
highest sensitivities in the retina. FST sensitivity for each patient is plotted against the highest sensitivity to 500 nm among the 72 loci measured in dark-
adapted perimetry. Linear correlation shown.
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272265
disturbance. In the patients with data from both FST and
dark-adapted perimetry, a comparison was made between
FST sensitivity level (white stimulus) and the locus with
maximum sensitivity from dark-adapted perimetry (500 nm
stimulus). Strong linear (slopeZ1$00) correlation (rZ
0$96) was found in this comparison (Fig. 4D) as well as
between FST sensitivities and each of four clusters of loci
(averaged loci within 2, 4, 6 or 8 dB from highest
sensitivity; all show rO0$95), thus providing support for
the hypothesis that FST is measuring the response of the
most sensitive retinal areas.
Within- and inter-visit variability of FST measurements
ranked by the mean FST for that session (Fig. 5A, upper
panel). The differences of each measurement from the
vary with FST level. The individual subjects’ standard
deviations plotted against their means demonstrated they
were unrelated (rZ0$10; pZ0$25). The within-subject
1996) is 1$63 dB for normal subjects and 1$61 dB for the
patient group; the variances are not significantly different
(pZ0$83). Inter-visit variability for the subgroups of 10
normal subjects and 41 patients is also shown (Fig. 5B).
Mean sensitivity difference (GSD) for normals is K0$82
(G1$76) dB and for patients is 0$05 (G1$95) dB, both not
significantly different from zero (pZ0$18 and 0$86,
respectively). Inter-visit variability in the patients is not
dependent on mean sensitivity level (pZ0$93).
Fig. 5. Within- and inter-visit variability of FST. (A) Upper. FST white stimulus results (between 2 and 6 measures) within a single session are shown (small
diamonds) for each of the 12 normal subjects and 146 patients, ranked by mean FST (from high to low). Lower. Differences of each FST measure from session
mean are plotted (C) for normals and patients. Dashed lines represent meanG2SD of the within-visit variability so calculated for the normal group and the
patient group.(B) Inter-visitvariabilityforFST white stimulusresults.Sensitivitydifference (visit 2 minusvisit1) is plottedas a functionof mean sensitivity in
10 normal subjects (unfilled symbols) and 41 retinal degeneration patients (filled symbols). Dotted line is the mean sensitivity difference and dashed lines
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272266
Dark-adapted threshold measurements to quantify rod-
mediated vision are a recommended part of the ocular
examination of patients with RP and related disorders
(Marmor et al., 1983). Natural history studies of inherited
thresholds, but more data are available on cone measures,
such as visual acuity, color vision, light-adapted perimetry
and cone ERGs (for example, Berson et al., 1985, 2002,
2004; Holopigian et al., 1996; Birch et al., 1999; Hirakawa
et al., 1999; Caruso et al., 2001; Flynn et al., 2001; Lodha
et al., 2003). Only cone function, specifically the cone
flicker ERG and light-adapted perimetry, has been used as
the primary outcome measure in RP treatment trials to date
(Berson et al., 1993, 2004; Hoffman et al., 2004). Why
should there be a preference to measure cone function when
rod-mediated vision would seem equally appropriate in
these diseases? There are at least four contributing factors.
First, cone vision testing is less time consuming and testing
apparatus is more available: there is no need to wait lengthy
times to dark adapt the eyes, and visual acuities and visual
fields do not even require pupillary dilation. Second, light-
adapted perimetry to measure cone vision across the visual
field has a rich history of use in glaucoma (e.g. Spry and
Johnson, 2002) and neuro-ophthalmological disease (e.g.
Keltner et al., 1999), and this experience can be drawn upon.
Third, a submicrovolt cone flicker ERG, after much
discussion about methodology and signal origins (Andreas-
son et al., 1988; Dagnelie and Massof, 1994; Birch and
Sandberg, 1997; Sieving et al., 1998), is generally accepted
as a surrogate measure of integrated cone vision in late-
stage retinal degeneration. Finally, inclusion of patients
with measurable cone vision but severely diminished rod
vision in studies expands the available data set for greater
Future trials of treatment in hereditary retinal degener-
ations will benefit from having rod-specific as well as cone-
specific outcome measures in order to understand how each
photoreceptor type is affected by the intervention. As
therapies emerge and become useful at earlier rather than
later disease stages, there may be less need to concede the
loss of rod photoreceptors and only concentrate on retaining
cone function (e.g. Wong, 1997). Furthermore, successful
murine preclinical experiments based on retained outer
nuclear layer thickness (LaVail et al., 1998) may more
appropriately translate into treatment of rod-based vision in
human patients (assuming post-receptoral connectivity is
intact, Marc et al., 2003) since the great majority
of photoreceptor nuclei in the mouse retina are rods.
The present work studied the feasibility and variability of
two types of dark-adapted psychophysical threshold
measurement that may be used to assess rod vision: (1) a
focal stimulus method that samples a group of predefined
central retinal locations and requires stable fixation; and
(2) a full-field stimulus method not requiring stable fixation.
Dark-adapted two-color profile testing, as performed in
Experiment 1, builds upon experience gained from earlier
studies that used such profiles or greater numbers of test loci
to characterize phenotypes of retinal degenerations
(e.g. Massof and Finkelstein, 1979; Lyness et al., 1985;
Birch et al., 1987; Apfelstedt-Sylla et al., 1992; Cideciyan
et al., 1998; Jacobson et al., 2000). Considering two profiles
using different wavelength stimuli were needed to deter-
mine photoreceptor-mediation, we limited the number of
test loci per profile to reduce test time and avoid patient
fatigue. False positive and false negative rates were similar
in patients and normal subjects and like those reported for
glaucoma (Katz and Sommer, 1988, 1990). Such profiles,
however, may be an inappropriate choice to assess outcome
in many future experimental treatments in disease subsets.
Flexibility in the method, though, allows ready replacement
by clusters of loci or other strategies to accommodate the
specific goals of a trial. For example, transition zones
between good and poorly functioning retina, as in the nasal
field in Patients 1 and 2 (Figs. 1A and 2A) or along the
vertical meridian in class B1 rhodopsin gene mutations
(Cideciyan et al., 1998), may be worth monitoring after a
focal intervention placed to avoid degenerate retina but not
to threaten well-functioning areas. Rod-mediated dark-
adapted sensitivity abnormalities (Owsley et al., 2000) and
rod photoreceptor pathology (Curcio et al., 1996) have been
documented in age-related macular degeneration (AMD)
making strategies with a macular grouping of test loci
potentially useful for determining the effects of early or
preventive treatment in AMD patients when fixation is
present. As occurs in glaucoma assessment (Gordon and
Kass, 1999; Reus and Lemij, 2004), there can be
coordinated use of objective structural measurements (e.g.
in vivo cross-sectional retinal imaging) in the same
locations chosen for dark-adapted functional measurements
(Cideciyan et al., 2004; Jacobson et al., 2004).
Sources of variability in automated light-adapted peri-
metry have been well-studied in relation to glaucoma (Spry
and Johnson, 2002). In hereditary retinal degenerations,
most published information is concerned with the variability
offull-field ERG measures (Berson et al., 1985, 1993; Birch
et al., 1991, 1999, 2002; Hoffman et al., 2004). By data
inspection, the dark-adapted profile tests appeared quite
repeatable in all patients with retinal degenerations, but we
sought to quantify the variability. Inter-visit variability in
the group of patients with rod function is comparable to that
in the normal group and the cone-mediated group. There is
only one previous report of variability measurements for
rod-mediated function using automated dark-adapted two-
color static perimetry (Birch et al., 1999). Inter-visit
variability of mean rod sensitivity was determined in 29
patients using 74 retinal locations; a threshold criterion for
change (at the 95th percentile) was reported as 5$26 dB.
This result could be compared with 3$05 dB in our 24
patients with rod-mediated function, sampled from eight
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272267
central field loci. The many differences between the two
studies make comparisons difficult.
The full-field test of dark-adapted thresholds, as
performed in Experiment 2, was born of the necessity to
quantify vision in patients with severe retinal degenerations
and unstable fixation, based on promising therapeutic
directions (de Juan et al., 1999; Acland et al., 2001;
Humayun et al., 2003). The concept of full-field dark-
adapted perceptual testing is not novel and has previously
been explored for glaucoma screening (Glovinsky et al.,
1992) and in retinal degenerations to evaluate results of
retinal transplantation or a visual prosthesis (de Juan et al.,
1999; Humayunet al., 2000, 2003). The instrumentation and
method in the current work was more like the published
design using a Tubingen bowl perimeter (Glovinsky et al.,
1992) than the modified ERG apparatus (Humayun et al.,
2000); there is also a commercially available LED-based
instrument but the limited dynamic range would not
accommodate severe retinopathies (Peters et al., 2000).
We tried to keep modifications to the automated perimeterat
a minimum (see Appendix A) to enable other investigators
to repeat and extend the work using equipment that would
be readily available in most ophthalmic departments and
private offices. The protocol was feasible in a large cohort of
patients with retinal degenerations, including those with
severe retinopathy. The intuitive hypothesis that FST results
would emanate from the most sensitive retinal regions was
tested using dark-adapted perimetry in many of the patients
able to do both tests and found to be valid. Additional
testing strategies will need to be devised to gain spatial
information about the function in severely affected patients.
A formal method and scoring system to determine light
projection in different regions of the visual field, accom-
plished usually with a penlight (Humayun et al., 2000), may
be of value. A suprathreshold static perimetric strategy with
a large stimulus in the same automated perimeter used for
FST may suffice. Use of broad-band chromatic filters in the
perimeter provided an opportunity to determine the
mediation of the light being perceived. If the goal of a
clinical trial is to affect rod retinal pathways and the baseline
pre-treatment results are all cone driven, a significant
increase in sensitivity to the achromatic stimulus
accompanied by a change in mediation from cone to rod
could be an important observation. Phase I safety testing
that would lead from baseline higher sensitivity with rod-
mediated result to lower sensitivity with cone-mediated
result should raise concerns about adverse effects. A caveat
is that no change in achromatic sensitivity and mediation
from baseline may be less readily interpreted. It may
actually mean there was no change, or it could mean that
there were regional increases or decreases in function that
escaped detection because they occurred at sensitivities
below the most sensitive retinal regions (which drive the
FST response; Fig. 4).
Within- and inter-visit variability was studied for the
FST. These results have no ready comparison in
the literature. The closest comparison of techniques may
be with the Goldmann–Weekers adaptometer (G–W), the
time-honored and conventional clinical instrument that can
be used to measure dark-adapted thresholds. G–W
thresholds are usually determined with a 118 diameter
target viewed centrally or eccentrically. Like most instru-
ments using a diffusing bowl, including projection per-
imeters, the light source of the G–W cannot be completely
confined to the target and there is some light scattering
across the bowl. In a patient with unstable fixation and
severe loss of visual sensitivity, the examination with the G–
W may approximate that with the FST. Variability was
reported as 6$75 dB (nZ29) for dark-adapted thresholds
measured with an 118 target located 78 below fixation (Birch
et al., 1999), while there was an earlier report of 1$51 dB
(nZ26) for the same test but located centrally or in the
superior field (Berson et al., 1985). The inter-visit
variability results for FST of G3$90 dB fall between
these published G–W estimates.
FST appears to be a suitable candidate perceptual
outcome measure for clinical trials in subsets of patients
that cannot perform dark-adapted perimetry. The technique
should complement ERG and other measures, such as
pupillometry (Aleman et al., 2004), and permits seriously
impaired rod function to be quantified. Although the
methods used in Experiments 1 and 2 are different and
will serve different patient populations, it is of interest that
the inter-visit variability of G3$90 dB from FST (repre-
senting the highest sensitivity in the retina under investi-
gation) is not vastly different than G3$04 dB from the focal
testing (representing the highest sensitivity loci in the
profile). This suggests that such values may be the current
best estimate of inter-visit variability for these types of
psychophysical rod vision testing. Further advances in
protocols and greater understanding of sources of variability
may make dark-adapted threshold measurements an even
less variable and more predictable monitor of rod function
than previously considered.
This work was supported by the NIH/NEI, Macula
Vision Research Foundation, Foundation Fighting Blind-
ness, F.M. Kirby Foundation, Macular Disease Foun-
dation, Research to Prevent Blindness and the Mackall
Trust. We are grateful to Jessica Emmons, Alexander
Sumaroka, Ilya Kniznik, Paul Schied, Mary Nguyen,
Michael Pianta, Alexander Pantelyat, Andy Cheung,
Boram Kim, Marisa Roman and Malgorzata Swider for
critical help. Special thanks are due to Mr F. Letterio and
members of Zeiss-Humphrey (Dublin, CA) who provided
generous assistance and advice about how to modify the
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272 268
Appendix A. Modifications to an automated perimeter
to perform the Full-field Stimulus Test (FST)
(Humphrey Field Analyzer, model 750i, Zeiss-Humphrey,
Dublin, CA) was modified to record psychophysical
thresholds to full-field light stimuli under dark-adapted
conditions (Fig. A1). A uniform full-field stimulus was
produced by commanding the perimeter to project the
light beam to a single inferior field locus (xZ0, yZK608)
where a circular first surface mirror (25 mm diameter;
Edmund Optics 32-945) was attached (Fig. A1(A) and
(B)). The beam reflected by the mirror was pointed to a
white diffuser at the top of the bowl which served to
illuminate the bowl (Fig. A1(A) and (C)). The white
diffuser (50!55 mm, cardboard painted matt white Sears
#66108) was affixed to the yellow background illuminator
window, covering it completely (Fig. A1(C)). A matt
black cardboard baffle was installed below the chin rest
(Fig. A1(A) and (B)) to block any spurious light from the
mirror reaching the eye.
In order to perform dark-adapted testing, the back-
ground light of the instrument was turned off using a
setup code provided from the manufacturer. This setup
code works only under ‘blue on yellow background’ test
condition which inserts the instrument’s blue filter in the
stimulus light path. To test with white light, the blue
filter has to be moved out of the light path by lifting
Fig. A1. Modifications to an automated perimeter to perform FST. (A) Schematic of perimeter bowl and light path from projector to mirror to diffuser. (B)
Mirror and black baffle near and below chinrest (viewed from above). (C) Diffuser fixed to original yellow illuminator at top of bowl (viewed from below). (D)
View with top cover of perimeter removed showing two adjacent wheels: aperture wheel with enlarged aperture is at left, and the colour filter wheel at right. A
custom-made plastic repositioning cap is attached to the colour filter wheel to ensure that the instrument will test with white when programmed to test with
blue. (E) A holder for a neutral density filter for the main light path. (F) Spectral characteristics of the broadband blue and red stimuli compared with the white
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272269
the hinged cover at the top of the instrument and rotating
the filter wheel (Fig. A1(D)). The perimeter, however,
automatically replaces the blue filter in the light path
every time the operator cycles the machine to perform a
new test. To avoid delays associated with manual
rotation of the filter wheel, a custom-built plastic snap-
on cap was designed to increase the preset rotation angle
and make the instrument use an open filter slot instead of
the blue filter. This repositioning cap can be fitted
manually by pressure to the filter wheel body, thereby
permitting repeated testing with white light and avoiding
manual realignment of the filter wheel (Fig. A1(D)).
The dynamic range of the stimulus light was increased
to be able to accommodate the extreme loss of visual
sensitivity in patients with severe retinal degenerative
diseases and still be able to test absolute rod sensitivity
in normal subjects. The maximum luminance (0 dB
stimulus level) was increased by enlarging the available
aperture. The aperture wheel in the perimeter light path
normally includes six openings: five Goldmann target
sizes (I–V), and a wider opening, which is not used for
conventional testing. An increase of available light level
of about 6$5 dB can be obtained by using this larger
opening instead of size V, but the software does not have
an option to select this opening. Therefore, size II target
was mechanically enlarged (Fig. A1(D)) recognizing that
this modification eliminates future testing with this target.
Using this enlarged aperture, mirror and diffuser, the
maximum (0 dB) luminance was 3$7 cd mK2, uniform to
within 2 dB. To extend the low end of the stimulus
dynamic range, a 3$0 log unit neutral density filter
(Melles Griot #03 FNA 226, Carlsbad, CA) on a custom-
built plastic filter holder was installed coaxial to the light
beam (Fig. A1(E)). This filter was manually inserted
prior to testing subjects with anticipated higher light
Stray light leaking from the projector enclosure was
baffled by covering bowl openings with black masking tape
(Jacobson et al., 1986). The lateral bowl opening used for
the chin rest mechanism was shielded using a cardboard
template and black felt. Rear ventilation openings were
light-shielded with a hanging black curtain leaving ample
room for air flow. The original yellow fixation LED was
replacedby a red LED and attenuated untilit was just visible
to the subject, or turned off completely in subjects with no
fixation (Jacobson et al., 1986). The infrared illuminators
(for eye visualization) were disabled after head positioning,
since part of the illumination fell within the visible range
and may have disturbed dark adaptation in subjects with
Custom test parameters used were as follows: strategyZ
full threshold; speedZslow; gaze trackingZoff; fixationZ
central; stimulus sizeZII (corresponding to the enlarged
aperture, see above); colorZblue; backgroundZyellow;
single point at xZ0, yZK608.
Acland, G.M., Aguirre, G.D., Ray, J., Zhang, Q., Aleman, T.S.,
Cideciyan, A.V., Pearce-Kelling, S.E., Anand, V., Zeng, Y.,
Maguire, A.M., Jacobson, S.G., Hauswirth, W.W., Bennett, J., 2001.
Gene therapy restores vision in a canine model of childhood blindness.
Nat. Genet. 28, 92–95.
Aleman, T.S., Jacobson, S.G., Chico, J.D., Scott, M.L., Cheung, A.Y.,
Windsor, E.A., Furushima, M., Redmond, T.M., Bennett, J.,
Palczewski, K., Cideciyan, A.V., 2004. Impairment of the transient
pupillary light reflex in Rpe65(–/–) mice and humans with Leber
congenital amaurosis. Invest. Ophthalmol. Vis. Sci. 45, 1259–1271.
Andreasson, S.O., Sandberg, M.A., Berson, E.L., 1988. Narrow-band
filtering for monitoring low-amplitude cone electroretinograms in
retinitis pigmentosa. Am. J. Ophthalmol. 105, 500–503.
Apathy, P.P., Jacobson, S.G., Nghiem-Phu, L., Knighton, R.W., Parel, J.-
M., 1987. Computer-aided analysis in automated dark-adapted static
Field Symposium. Martinus Nijhoff Publishers, Dordrecht, pp. 279–
Apfelstedt-Sylla, E., Kunisch, M., Horn, M., Ruther, K., Gal, A.,
Zrenner, E., 1992. Diffuse loss of rod function in autosomal dominant
retinitis pigmentosa with pro-347-leu mutation of rhodopsin. Ger.
J. Ophthalmol. 1, 319–327.
Berson, E.L., Sandberg, M.A., Rosner, B., Birch, D.G., Hanson, A.H.,
1985. Natural course of retinitis pigmentosa over a three-year interval.
Am. J. Ophthalmol. 99, 240–251.
Berson, E.L., Rosner, B., Sandberg, M.A., Hayes, K.C., Nicholson, B.W.,
Weigel-DiFranco, C., Willett, W., 1993. A randomized trial of vitamin
A and vitamin E supplementation for retinitis pigmentosa. Arch.
Ophthalmol. 111, 761–772.
Berson, E.L.,Rosner, B.,Weigel-DiFranco,
Sandberg, M.A., 2002. Disease progression in patients with dominant
retinitis pigmentosa and rhodopsin mutations. Invest. Ophthalmol. Vis.
Sci. 43, 3027–3036.
Berson, E.L., Rosner, B., Sandberg, M.A., Weigel-DiFranco, C., Moser, A.,
Brockhurst, R.J., Hayes, K.C., Johnson, C.A., Anderson, E.J.,
Gaudio, A.R., Willett, W.C., Schaefer, E.J., 2004. Clinical trial of
docosahexaenoic acid in patients with retinitis pigmentosa receiving
vitamin A treatment. Arch. Ophthalmol. 122, 1297–1305.
Bessant, D.A., Ali, R.R., Bhattacharya, S.S., 2001. Molecular genetics and
prospects for therapy of the inherited retinal dystrophies. Curr. Opin.
Genet. Dev. 11, 307–316.
Birch, D.G., Sandberg, M.A., 1997. Submicrovolt full-field cone electro-
retinograms: artifacts and reproducibility. Doc. Ophthalmol. 92, 269–
Birch, D.G., Herman, W.K., deFaller, J.M., Disbrow, D.T., Birch, E.E.,
1987. The relationship between rod perimetric thresholds and full-field
rod ERGs in retinitis pigmentosa. Invest. Ophthalmol. Vis. Sci. 28,
Birch, D.G., Anderson, J.L., Fish, G.E., 1991. Longitudinal measures in
children receiving ENCAD for hereditary retinal degeneration. Doc.
Ophthalmol. 77, 185–192.
Birch, D.G., Anderson, J.L., Fish, G.E., 1999. Yearly rates of rod and cone
functional loss in retinitis pigmentosa and cone-rod dystrophy.
Ophthalmology 106, 258–268.
Birch, D.G., Hood, D.C., Locke, K.G., Hoffman, D.R., Tzekov, R.T., 2002.
Quantitative electroretinogram measures of phototransduction in cone
and rod photoreceptors: normal aging, progression with disease, and
test–retest variability. Arch. Ophthalmol. 120, 1045–1051.
Bland, J.M., Altman, D.G., 1986. Statistical methods for assessing
agreement between two methods of clinical measurement. Lancet
Bland, J.M., Altman, D.G., 1996. Measurement error. BMJ 313, 744.
Blumenthal, E.Z., Sample, P.A., Berry, C.C., Lee, A.C., Girkin, C.A.,
Zangwill, L., Caprioli, J., Weinreb, R.N., 2003. Evaluating several
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272 270
sources of variability for standard and SWAP visual fields in glaucoma
patients, suspects, and normals. Ophthalmology 110, 1895–1902.
Caruso, R.C., Nussenblatt, R.B., Csaky, K.G., Valle, D., Kaiser-
Kupfer, M.I., 2001. Assessment of visual function in patients with
gyrate atrophy who are considered candidates for gene replacement.
Arch. Ophthalmol. 119, 667–669.
Chauhan, B.C., Johnson, C.A., 1999. Test-retest variability of frequency-
doubling perimetry and conventional perimetry in glaucoma patients
and normal subjects. Invest. Ophthalmol. Vis. Sci. 40, 648–656.
Cideciyan, A.V., Hood, D.C., Huang, Y., Banin, E., Li, Z.Y., Stone, E.M.,
Milam, A.H., Jacobson, S.G., 1998. Disease sequence from mutant
rhodopsin allele to rod and cone photoreceptor degeneration in man.
Proc. Natl Acad. Sci. USA 95, 7103–7108.
Cideciyan,A.V., Aleman,T.S., Swider,M., Schwartz, S.B., Steinberg, J.D.,
Brucker, A.J., Maguire, A.M., Bennett, J., Stone, E.M., Jacobson, S.G.,
2004. Mutations in ABCA4 result in accumulation of lipofuscin before
slowing of the retinoid cycle: a reappraisal of the human disease
sequence. Hum. Mol. Genet. 13, 525–534.
Cornsweet, T.N., 1970. Visual Perception. Academic Press, New York pp.
Curcio, C.A., Medeiros, N.E., Millican, C.L., 1996. Photoreceptor loss in
age-related macular degeneration. Invest. Ophthalmol. Vis. Sci. 37,
Dagnelie, G., Massof, R.W., 1994. Sub-microvolt electroretinograms:
negotiating the pitfalls of photoelectricity and noise, in: Vision Science
and Its Applications, 1994 Technical Digest Series, vol. 2. Optical
Society of America, Washington, DC pp. 354–357.
de Juan Jr., E., Cooney, M.J., Humayun, M.S., Jensen, P.S., 1999.
Treatment of retinal disease in the new millennium. Ophthalmol. Clin.
North Am. 23, 539–562.
Delyfer, M.N., Leveillard, T., Mohand-Said, S., Hicks, D., Picaud, S.,
Sahel,J.A., 2004. Inheritedretinaldegenerations:therapeuticprospects.
Biol. Cell 96, 261–269.
Ernst, W., Faulkner, D.J., Hogg, C.R., Powell, D.J., Arden, G.B., Vaegan,
1983. An automated static perimeter/adaptometer using light emitting
diodes. Br. J. Ophthalmol. 67, 431–442.
Flynn, M.F., Fishman, G.A., Anderson, R.J., Roberts, D.K., 2001.
Retrospective longitudinal study of visual acuity change in patients
with retinitis pigmentosa. Retina 21, 639–646.
Glovinsky, Y., Quigley, H.A., Drum, B., Bissett, R.A., Jampel, H.D., 1992.
A whole-field scotopic retinal sensitivity test for the detection of early
glaucoma damage. Arch. Ophthalmol. 110, 486–490.
Gordon, M.O., Kass, M.A., 1999. The ocular hypertension treatment study:
design and baseline description of the participants. Arch. Ophthalmol.
Grover, S., Fishman, G.A., Birch, D.G., Locke, K.G., Rosner, B., 2003.
Variability of full-field electroretinogram responses in subjects without
diffuse photoreceptor cell disease. Ophthalmology 110, 1159–1163.
Gunkel, R.D., 1967. Retinal profiles, A psychophysical test of rod and cone
sensitivity. Arch. Ophthalmol. 77, 22–25.
Hirakawa, H., Iijima, H., Gohdo, T., Imai, M., Tsukahara, S., 1999.
Progression of defects in the central 10-degree visual field of patients
with retinitis pigmentosa and choroideremia. Am. J. Ophthalmol. 127,
Hoffman, D.R., Locke, K.G., Wheaton, D.H., Fish, G.E., Spencer, R.,
Birch, D.G., 2004. A randomized, placebo-controlled clinical trial of
docosahexaenoic acid supplementation for X-linked retinitis pigmen-
tosa. Am. J. Ophthalmol. 137, 704–718.
Holopigian, K., Greenstein, V., Seiple, W., Carr, R.E., 1996. Rates of
change differ among measures of visual function in patients with
retinitis pigmentosa. Ophthalmology 103, 398–405.
Humayun, M.S., de Juan Jr., E., del Cerro, M., Dagnelie, G., Radner, W.,
Sadda, S.R., del Cerro, C., 2000. Human neural retinal transplantation.
Invest. Ophthalmol. Vis. Sci. 41, 3100–3106.
Humayun, M.S., Weiland, J.D., Fujii, G.Y., Greenberg, R., Williamson, R.,
Little, J., Mech, B., Cimmarusti, V., Van Boemel, G., Dagnelie, G., de
Juan, E., 2003. Visual perception in a blind subject with a chronic
microelectronic retinal prosthesis. Vis. Res. 43, 2573–2581.
Jacobson, S.G., Voigt, W.J., Parel, J.M., Apathy, P.P., Nghiem-Phu, L.,
Myers, S.W., Patella, V.M., 1986. Automated light- and dark-adapted
perimetry for evaluating retinitis pigmentosa. Ophthalmology 93,
Jacobson, S.G., Yagasaki, K., Feuer, W.J., Roman, A.J., 1989. Interocular
asymmetry of visual function in heterozygotes of X-linked retinitis
pigmentosa. Exp. Eye Res. 48, 679–691.
Jacobson, S.G., Kemp, C.M., Sung, C.H., Nathans, J., 1991. Retinal
function and rhodopsin levels in autosomal dominant retinitis
pigmentosa with rhodopsin mutations. Am. J. Ophthalmol. 112, 256–
Jacobson, S.G., Cideciyan, A.V., Iannaccone, A., Weleber, R.G.,
Fishman, G.A., Maguire, A.M., Affatigato, L.M., Bennett, J.,
Pierce, E.A., Danciger, M., Farber, D.B., Stone, E.M., 2000. Disease
expression of RP1 mutations causing autosomal dominant retinitis
pigmentosa. Invest. Ophthalmol. Vis. Sci. 41, 1898–1908.
Schwartz, S.B., Roman, A.J., McInnes, R.R., Sheffield, V.C.,
Stone, E.M., Swaroop, A., Wright, A.F., 2004. Nuclear receptor
NR2E3 gene mutations distort human retinal laminar architecture and
cause an unusual degeneration. Hum. Mol. Genet. 13, 1893–1902.
Katz, J., Sommer, A., 1988. Reliability indexes of automated perimetric
tests. Arch. Ophthalmol. 106, 1252–1254.
Katz, J., Sommer, A., 1990. Reliability of automated perimetric tests. Arch.
Ophthalmol. 108, 777–778.
Keltner, J.L., Johnson, C.A., Spurr, J.O., Beck, R.W., 1999. Comparison of
central and peripheral visual field properties in the optic neuritis
treatment trial. Am. J. Ophthalmol. 128, 543–553.
LaVail, M.M., Yasumura, D., Matthes, M.T., Lau-Villacorta, C., Unoki, K.,
Sung, C.H., Steinberg, R.H., 1998. Protection of mouse photoreceptors
by survival factors in retinal degenerations. Invest. Ophthalmol. Vis.
Sci. 39, 592–602.
Lodha, N., Westall, C.A., Brent, M., Abdolell, M., Heon, E., 2003. A
modified protocol for the assessment of visual function in patients with
retinitis pigmentosa. Adv. Exp. Med. Biol. 533, 49–57.
Lyness, A.L., Ernst, W., Quinlan, M.P., Clover, G.M., Arden, G.B.,
Carter, R.M., Bird, A.C., Parker, J.A., 1985. A clinical, psychophysical,
and electroretinographic survey of patients with autosomal dominant
retinitis pigmentosa. Br. J. Ophthalmol. 69, 326–339.
Marc, R.E., Jones, B.W., Watt, C.B., Strettoi, E., 2003. Neural remodeling
in retinal degeneration. Prog. Retin. Eye Res. 22, 607–655.
Margalit, E., Maia, M., Weiland, J.D., Greenberg, R.J., Fujii, G.Y.,
Torres, G., Piyathaisere, D.V., O’Hearn, T.M., Liu, W., Lazzi, G.,
Dagnelie, G., Scribner, D.A., de Juan Jr., E., Humayun, M.S., 2002.
Retinal prosthesis for the blind. Surv. Ophthalmol. 47, 335–356.
Marmor, M.F., Aguirre, G., Arden, G., Berson, E., Birch, D.G.,
Dowling, J., Enoch, J., Fishman, G.A., Fulton, A.B., Garcia, C.A.,
Gouras, P., Heckenlively, J., Hu, D., Lewis, R.A., Niemeyer, G.,
Parker, J.A., Perlman, I., Ripps, H., Sandberg, M.A., Siegel, I.,
Weleber, R.G., Wolf, M.L., Wu, L., Young, R.S.L., 1983. Retinitis
pigmentosa: a symposium on terminology and methods of examination.
Ophthalmology 90, 126–131.
Massof, R.W., Finkelstein, D., 1979. Rod sensitivity relative to cone
sensitivity in retinitis pigmentosa. Invest. Ophthalmol. Vis. Sci. 18,
Owsley, C., Jackson, G.R., Cideciyan, A.V., Huang, Y., Fine, S.L.,
Ho, A.C., Maguire, M.G., Lolley, V., Jacobson, S.G., 2000.
Psychophysical evidence for rod vulnerability in age-related macular
degeneration. Invest. Ophthalmol. Vis. Sci. 41, 267–273.
Peters, A.Y., Locke, K.G., Birch, D.G., 2000. Comparison of the
Goldmann-Weekers dark adaptometer and LKC technologies scotopic
sensitivity tester-1. Doc. Ophthalmol. 101, 1–9.
G.E., delMonte, M.,
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272 271
Reus, N.J., Lemij, H.G., 2004. The relationship between standard
automated perimetry and GDx VCC measurements. Invest. Ophthal-
mol. Vis. Sci. 45, 840–845.
Sieving, P.A., Arnold, E.B., Jamison, J., Liepa, A., Coats, C., 1998.
Submicrovolt flicker electroretinogram: cycle-by-cycle recording of
multiple harmonics with statistical estimation of measurement
uncertainty. Invest. Ophthalmol. Vis. Sci. 39, 1462–1469.
Spry,P.G., Johnson,C.A.,2002. Identification of progressiveglaucomatous
visual field loss. Surv. Ophthalmol. 47, 158–173.
Spry, P.G., Johnson, C.A., McKendrick, A.M., Turpin, A., 2001.
Variability components of standard automated perimetry and fre-
quency-doubling technology perimetry. Invest. Ophthalmol. Vis. Sci.
Weleber, R.G., Kurz, D.E., Trzupek, K.M., 2003. Treatment of retinal and
choroidal degenerations and dystrophies: current status and prospects
for gene-based therapy. Ophthalmol. Clin. North Am. 16, 583–593.
Wong, F., 1997. Investigating retinitis pigmentosa: a laboratory scientist’s
perspective. Prog. Retin. Eye Res. 16, 353–373.
A.J. Roman et al. / Experimental Eye Research 80 (2005) 259–272272