| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 From the Department of Ophthalmology and Visual Sciences, and the 2 Department of Psychological and Brain Sciences, University of Louisville, Kentucky.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
METHODS. Noninvasive monocular mfERGs were recorded in anesthetized albino (Sprague–Dawley) and pigmented (Long Evans) rats. Achromatic stimuli subtending a 49° by 53° field consisted of 61 hexagons that were generated and presented (at varying rates and luminances) using a Visual Evoked Response Imaging System (VERIS; EDI, San Mateo, CA). The VERIS also was used to calculate individual responses and for analysis.
RESULTS. mfERGs were recorded from pigmented and albino rats by slowing the rate of stimulus presentation to allow for the slow recovery time of the rod system. In each rat strain, responses varied systematically with changes in stimulus parameters. Peak response amplitude increased as the rate of stimulation was slowed and as stimulus luminance was increased. Response latency decreased as stimulus intensity was increased. The local nature of the response was assessed by several independent measures.
CONCLUSIONS. The present work demonstrated the feasibility and limitations of using mfERG to assess topographical changes in the rat retina. It showed that despite the problems of the unavoidable self-adapting nature of the stimulus, the small eye of the animal, and the high potential for light scatter within the retina, multifocal responses with a good signal-to-noise ratio can be obtained from the rat.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Conventional electroretinograms (ERGs) have been useful for assessing changes in retinal function. The ERG is noninvasive, allows repeated assessment, and can be performed in a relatively short period. Clinical ERGs recorded in response to a full-field flash provide a quantitative assessment of global retinal function but do not provide local response information. Focal ERGs permit assessment of localized retinal activity but require detailed procedures to minimize the effects of stray light and require considerable time and effort to obtain responses from multiple retinal regions.1 2 Recently, a technique using nonlinear systems analysis to extract local responses from a continuous ERG was developed.3 This approach, termed multifocal electroretinography (mfERG), permits simultaneous assessment of local responses from a large number of retinal areas. Because it can provide topographical analysis of retinal function in a relatively short period, this technique has already been used in a variety of basic and clinical studies of human vision.4 5 6 7 8 9 10 11 12 13
In brief, the mfERG technique involves the presentation of multiple local flashes in a predefined order (m-sequence) and subsequent extraction of the corresponding ERGs from the field potential by computation of the cross correlation between the stimulus m-sequence and the response cycle (for detailed description, see References 3 and 4). Typically, a single recording session lasts only a few minutes and can provide a topographical map of local retinal responses, making it a practical way to evaluate retinal function in the clinic.
Direct comparisons of mfERGs with full-field ERGs have provided evidence that the two are comparable.5 When stimulus and recording conditions are closely matched for the two methods, changes in stimulus parameters result in comparable changes in "b-wave" amplitude and latency, suggesting that the waveforms represent similar cellular responses.5 In a study examining the rod mfERG in humans, waveforms were found to contain an early component not seen in full-field ERGs.6 Although slightly shorter, this early component more closely matched the full-field ERG in latency compared with the later peak response. When the stray-light response was reduced by adding a surround, the larger peak response decreased, whereas the early component increased in amplitude. Thus, the small early component was attributed to a local rod response. In patients with diseased retinas, multifocal, full-field, and focal ERGs yield compatible results in area and type of dysfunction observed.7
Measurement of mfERGs has been useful in providing a topographical and quantitative functional assessment of normal and diseased retinas. In normal humans, the amplitude of the individual mfERG responses varied with the density of cone receptors.4 Clinically, the mfERG has been used to map the topographical pattern of deterioration in patients with retinitis pigmentosa,8 9 myopia,10 diabetic retinopathy,11 12 and white dot syndrome.13
Application of mfERGs to evaluate topographical changes in animal models of various retinal diseases would be extremely useful for describing the progression of these diseases and for testing the efficacy of various treatments. However, mfERG was developed as a photopic test to assess local cone-related function, whereas most animal models have been developed in rod-dominant rodents. Until recently, the application of mfERGs to test rod function had been equivocal because of the self-adapting nature of the stimulus (repeated, high-luminance flashes). However, by adjusting stimulus parameters to accommodate for the slow recovery of rod photoreceptors, Hood et al.6 have successfully applied the technique to evaluate human rod function. Because mfERGs have not been demonstrated in any small-eyed and rod-dominant animals, the primary objective of the present study was to determine whether the mfERG technique can be further developed to assess retinal function in the rat. First, optimal stimulus and recording parameters were determined, the local nature of the responses was assessed, and the potential influences of dark adaptation and anesthesia type were evaluated. The results showed that with the use of appropriate temporal rates and stimulus intensities, a consistent and easily measurable mfERG response can be obtained from the rod-dominant retinas of both pigmented and albino rats.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Stimulus
Figure 1A
shows an example of the 61-hexagon stimulus array. The stimulus was
generated by the Visual Evoked Response Imaging System (VERIS; EDI, San
Mateo, CA) and presented on a 21-in. monochrome display monitor (P104
phosphor; Nortech Imaging, Plymouth, MN) using a customized Macintosh
(Apple Computer, Cupertino, CA) video card. Stimulus luminance was
calibrated at a 15° angle using a photometer (model 350; United
Detector Technologies, Hawthorne, CA) with a photometric filter. Note
that because of the small size of the rat eye and its physiological
optics, the effective retinal illuminance of the rat eye would be
approximately 10 times higher than that of a human. More specifically,
after the approach of Remtulla and Hallett,15
retinal
illuminance varies as the square of the posterior to nodal difference
in the rat eye (3.33 mm)16
and human eye (16.70
mm)17
(e.g., 16.702/3.332 =
25.15 times greater in the rat). However, based on dilated pupil areas
for human (50.27 mm2)18
and rat (19.63
mm2),17
the human eye should capture
approximately nine times more photons (50.27/19.63 = 2.56). Thus,
a fully dilated rat retina should receive 9.82 times the irradiance of
a similarly stimulated human retina (i.e., 25.15/2.56 = 9.82).
|
The number of blank frames following a stimulus frame determined the maximum rate at which an individual hexagon flashed white. The addition of each F added a minimum of 13.3 msec and an average of 26.6 msec between white stimulus frames. For example, for a stimulus with 0F the average time between white frames (e.g., interstimulus interval, ISI) was 13.3 msec, whereas 6F added a minimum of 6 x 13.3 msec and an average of 6 x 26.6 msec. When blank frames are added to each element in an m-sequence, a smaller number of stimuli are presented per unit time, and a smaller number of responses are recorded per unit time. Thus, as F was increased, the stimulation rate was slowed, and a longer recording period was required to collect an equal number of responses. Values greater than 12F were not used, because this would have required a recording time of almost 30 minutes.
To determine the temporal parameters necessary to obtain an optimal response, recordings were made over a range of temporal rates. The following rates were used: 12F, mean ISI = 332.5 msec; 10F, mean ISI = 279.3 msec; 8F, mean ISI = 226.1 msec; and 6F, mean ISI = 172.9 msec. The signal-to-noise ratio was unsatisfactory for recordings made at less than 6F. In a given recording session, rats were tested at each stimulus rate beginning with the slowest rate, followed by progressively increasing rates. The photopic luminance of each white hexagon was kept at 81 candelas [cd]/m2 for all temporal rates, whereas the background and blank frames were set to 0.4 cd/m2. A rate of 10F was found to elicit large responses, yet permitted a relatively short recording time (10 minutes).
Once the optimal temporal rate for rod stimulation was determined, responses at four different luminances were measured using the 10F rate. The intensity of the white hexagons was varied, whereas the background and blank hexagons remained black (0.04 cd/m2) resulting in a mean screen luminance during presentation of the stimulus elements equal to approximately one eleventh that of the white hexagons. All luminance levels reported here refer to the luminance of the white hexagons and not the mean luminance. To obtain an intensity–response series, the photopic luminance of the stimulus frames was increased in increments of 0.5 log steps (i.e., by a factor of 3) beginning with 3 cd/m2 and increasing successively to 9 cd/m2, 27 cd/m2, and 81 cd/m2.
Recording
Retinal responses were recorded using DTL
electrodes19
placed on the vertical midline of the cornea
of each eye. Recordings from the stimulated eye were referred to the
contralateral eye. The animal was grounded using a needle electrode.
Responses were amplified (x10,000) and filtered (1 Hz/1 kHz; amplifier
model 12, Neurodata Acquisition System; Grass, West Warwick, RI), and
sampled 15 times within each frame interval (15 times every 13.3 msec)
with an analog-to-digital board. The ERG signal was continuously
monitored during recording to ensure signal quality. Signals with high
noise levels were rejected and rerecorded. Recording duration varied
from 2 to 11 minutes depending on temporal parameters. All recording
sessions were divided into two overlapping segments.
Partial Retinal Bleaching
To evaluate whether the responses represented localized retinal
activity, a small retinal area was bleached by presentation of a
full-spectrum light (1000-W xenon source; 240
J/cm2) for 35 seconds. The light beam was focused
on a localized portion (approximately 2 mm in diameter) of the retina,
by using a Maxwellian view system. An mfERG was recorded before
bleaching and several times after bleaching.
Local Lesions
After prior VERIS testing, a pigmented rat was anesthetized with
sodium pentobarbital (50 mg/kg) and positioned so that its right eye
would be aligned in the optical axis of an argon laser beam. A pulse
(0.15 W) was then delivered to the right eye for 1 second. Postlesion
recordings were made approximately 2 hours after delivery of the
lesion. The m-sequence of the stimulus was composed of one white
hexagon (81 cd/m2) followed by 10 blank frames.
Data Analysis
A first-order response was calculated that reflected activity at
each retinal area. Response amplitudes were measured from baseline set
to 0.0 mV at the beginning of the 150-msec averaged response to
individual stimulus flashes. Latency was measured from stimulus onset
to time of peak response amplitude. One iteration of an artifact
rejection manipulation3
was performed on the obtained
signals by using the VERIS software. This manipulation identified
segments that contained significant deviations not temporally related
to the stimulus and replaced such segments with a response estimate
computed from the entire record.3
Each response was
averaged with 17% of all surrounding responses with the VERIS
software. To generate a composite ERG, the 61 responses comprising a
trace array were summed. All response arrays shown are from a single
recording session, not averages of multiple sessions. Although averages
of multiple sessions may have provided a cleaner response array, we
chose to use each day’s recording session to obtain a complete series
of recordings (e.g., an intensity series or blank-frame series) rather
than repeat individual conditions. Because minor deviations in eye
position and electrode location were likely to occur from day to day
(unlike humans, our subjects did not fixate the target) averaging
across days would not have produced valid topography. (Note:
Individual responses are local only if the subject is fixated on the
stimulus. Any eye movements during a recording session would result in
a blurring of the responses. For recordings in humans, a fixation point
is often added to the center of the stimulus matrix. Because the rats
could not be instructed to fixate and were not paralyzed, a stable eye
position was dependent on anesthetic. To test for eye movements, a
1-mm2 thin mirror was secured to the cornea of the
anesthetized rat. A laser beam directed to the mirror was used to track
any positional changes in the eye over a period of 15 minutes. Eye
movements were found to be negligible. Another indicator that eye
movements were not a cause of concern came from online monitoring of
the continuous ERG. If the depth of anesthesia began to decrease during
a recording session, eye movements quickly returned, resulting in
obvious artifacts and noise in the recording.)
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Temporal Parameters
In Figure 2
, summed responses are shown for stimuli incorporating a range of
blank frames: 6F, 8F, 10F, and
12F. The fastest stimulation rate shown
(6F) yielded the lowest response amplitudes. Rates
less than 6F resulted in an unacceptable signal-to-noise
ratio (data not shown). Figure 3
plots the relationship between number of blank frames and response
amplitude for five individual albino rats (Fig. 3A)
and for five
pigmented rats (Fig. 3C)
. As more blank frames were used, producing a
slower stimulation rate, the amplitude of the peak showed little
consistent change, but often slightly increased. Figures 3B
and 3D
show
that, in general, response latency decreased as the number of blank
frames increased.
|
|
|
|
|
|
Effect of Anesthesia Type
Initially, all animals were anesthetized with sodium
pentobarbital. In later sessions, ketamine-xylazine was used because it
allows a faster recovery time. As shown in Figure 8
, recordings obtained in animals under sodium pentobarbital anesthesia
were similar to those obtained in those under ketamine-xylazine
anesthesia, for both albino and pigmented rats.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The self-adapting nature of the mfERG stimulus provided a major challenge to the application of this technique to the rat. Because light adaptation has been shown to greatly reduce the amplitude of the rat full-field ERG,21 rod responses have been typically recorded from dark-adapted rats using single full-field flashes.22 This has produced optimal responses because the rod system is very slow to recover from light adaptation caused by prior stimulation, although adaptation, even to intense lights, does not saturate the rod responses.23 The standard recording conditions for the mfERG stimulus (successive flashes presented at high frequency) were designed to evaluate faster, photopic cone responses, and thus were inappropriate for use with rods. To maximize the rod response in the multifocal paradigm, the stimulus rate was slowed to allow for recovery of the rod system by the addition of dark blank frames to each element of the m-sequence (as described in Reference 6). This procedure resulted in easily measurable, positive b-wave–like responses that varied systematically with the intensity and temporal frequency of the stimulus. Prior dark adaptation had no effect on responses. The stimulus–response relationships and resultant waveforms were quite similar to rod mfERGs obtained in humans by Hood et al.6 The exception was that in rats, response latency decreased as the number of blank frames was increased. This was not reported in human rod mfERGs. In comparison with full-field ERGs, rat and human mfERG data were similar, in that the negative-going "a-wave" was either missing (rats) or negligible (humans), and in that the amplitude of the maximal response was reduced. The "a-wave" was most likely absent because the stimulus luminance was below a-wave threshold. The overall reduction in amplitude likely reflects the continued adaptation of some portion of rod system, as well as reflecting the smaller retinal area that is activated by the multifocal stimulus compared with full-field stimulation.
The waveform obtained in rats is most certainly reflective of rod-mediated activity, which is not surprising, given the small proportion of cones in the rat retina.24 The absence of response to stimulus rates faster than 6F (cone mfERGs can be recorded at 0F) and the slow latency and small amplitude of the peak support the conclusion that this response originates from rods and not cones. The response latencies obtained (50–80 msec) were comparable to human rod mfERGs (60–100 msec) and were clearly longer than those obtained in human cone mfERGs5 (26–33 msec) or in mfERG recordings from the cone-dominated retina of the tree shrew25 (ca. 20–30 msec.). The particular cell type(s) that generate the positive-going responses cannot be identified from this study. It has been suggested that the positive component of the full-field ERG is likely to be generated by rod bipolar cells.26
Light scatter presented a second major concern regarding the application of the mfERG to the rat. This problem is prevalent even in large eyes. For example, in humans, an mfERG response can be obtained "from" the optic disc,4 presumably representing activity in adjacent areas due to light scatter from nerve fibers at the disc. Also, because rods, unlike cones, are not sensitive to stimulus direction,27 they are especially sensitive to intraocular stray light. Consequently, both focal and multifocal human rod ERGs contain a large, slower response elicited by stray light.6 28 Although a slower stray-light component was not observed in rats, the small size of the rat eye increases the prevalence of light scatter, and an absence of pigmentation in the albino eye exacerbates the problem. It is likely that this is why no obvious reduction in amplitude was seen in the hexagon corresponding to the location of the optic nerve head. Because of these limitations it is not possible to define the exact size of a prescribed retinal area responding to an individual stimulus hexagon. However, the partial retinal bleaching experiment and the laser lesion experiment reported here in the pigmented rat showed that these manipulations produced clear localized changes in the response array. In fact, the appearance of the trace arrays resulting from these manipulations resembled published mfERG recordings from patients with retinitis pigmentosa (see Fig. 3 in Reference 9).
In sum, the present work demonstrated the feasibility and limitations of using mfERG to assess topographical changes in the rat retina. It showed that despite the problems of the unavoidable self-adapting nature of the stimulus, the small eye of the animal, and the high potential for light scatter within the retina, multifocal responses with a good signal-to-noise ratio can be obtained. The addition of blank frames increased recording time, but even with the addition of 10 blank frames and an m-sequence of 212 - 1 elements, a complete recording was obtainable in only 10 minutes.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by Grant EY08519 from the National Institutes of Health; the Murray Foundation, Inc., Princeton, New Jersey; the Vitreoretinal Research Foundation, Louisville, Kentucky; the Kentucky Lions Eye Foundation; an unrestricted grant from Research to Prevent Blindness; and funds from an anonymous donor.
Submitted for publication July 7, 1999; revised September 15, 1999; accepted September 29, 1999.
Commercial relationships policy: N.
Corresponding author: Heywood M. Petry, Department of Psychological and Brain Sciences, Life Sciences Building, Room 317, University of Louisville, Louisville, KY 40292. woody.petry{at}louisville.edu
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. T. Pardue, M. J. Phillips, H. Yin, B. D. Sippy, S. Webb-Wood, A. Y. Chow, and S. L. Ball Neuroprotective Effect of Subretinal Implants in the RCS Rat Invest. Ophthalmol. Vis. Sci., February 1, 2005; 46(2): 674 - 682. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Paskowitz, G. Nune, D. Yasumura, H. Yang, R. B. Bhisitkul, S. Sharma, M. T. Matthes, M. A. Zarbin, M. M. LaVail, and J. L. Duncan BDNF Reduces the Retinal Toxicity of Verteporfin Photodynamic Therapy Invest. Ophthalmol. Vis. Sci., November 1, 2004; 45(11): 4190 - 4196. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
