The Retinal Vasculature and Function of the Neural Retina in a Rat Model of Retinopathy of Prematurity

Abstract

purpose. In a rat model of retinopathy of prematurity (ROP), the retinal vasculature and function of the neural retina were studied longitudinally. Vascular and neural parameters were evaluated for significant relationships.

methods. Retinopathy was induced by exposing newborn rats to alternating 50% and 10% oxygen until age 14 days. To evaluate the function of the neural retina, electroretinographic (ERG) responses to full-field stimuli were recorded from dark-adapted rats at ages 18 and 31 days. Sensitivity and saturated amplitude of photoreceptor and postreceptor activity were derived from ERG a- and b- waves. To evaluate the retinal vasculature, digital fundus photographs were obtained at the same ages, and the tortuosity indices of the arterioles (TI _A) and venules (TI _V) were calculated. ROP rats and room-air–raised control animals were compared. Vascular and response parameters were evaluated for significant relationships.

results. In ROP rats, TI _A was high at 18 days and decreased in every rat to nearly normal levels by 31 days. TI _V was less affected by ROP or age. Deficits in all receptor and post-receptor response parameters were present in 18-day-old ROP rats. Post-receptor sensitivity recovered completely by 31 days. Deficits in other response parameters persisted. No significant correlations between vascular and ERG parameters were found in 18-day-old ROP rats.

conclusions. Noninvasive, longitudinal measures in this model of ROP showed significant abnormalities in both the retinal vasculature and function of the neural retina that were most marked at age 18 days. However, vascular and neural abnormalities did not correlate.

The retinal blood vessels normally grow from the optic nerve head toward the periphery of the retina during the last trimester of pregnancy. ¹ An infant born prematurely, before complete vascular coverage of the retina, is often subjected to high supplemental oxygen for medical management. According to the commonly cited sequence of events leading to retinopathy of prematurity (ROP), the supplemental oxygen halts growth of the retinal blood vessels. Then, with cessation of the supplemental oxygen, the peripheral avascular retina becomes hypoxic. Hypoxia instigates the molecular cascade ^{2 3} that leads to the formation of the abnormal retinal blood vessels that are the clinical hallmark of ROP. ^{4 5} In actuality, even though the premature infant is often subjected to high ambient oxygen, immature lungs and other medical complications may lead to fluctuations in partial pressure of arteriole oxygen and consequent episodes of hypoxia and hyperoxia in the infant’s tissues. ⁶  

The diagnosis of ROP is made by the ophthalmologist’s subjective comparison of blood vessel appearance with established criteria. Dilated venules and tortuous arterioles at the posterior pole are recognized signs of high-risk ROP. ⁷ Only recently have objective image-analysis techniques been used to describe the retinal vasculature in ROP. ^{8 9 10 11 12}  

The vascular abnormalities that characterize ROP appear within a narrow preterm age range. ¹³ It is a provocative coincidence that in this age range the developing rod outer segments elongate rapidly, accompanied by an increase in the rhodopsin content of the retina and burgeoning energy demands in the photoreceptors. ^{14 15 16} Dysfunction of the neural retina, both photoreceptor and post-receptor, has been documented in infants and children with ROP and in rat models. ^{17 18 19 20 21 22} In fact, in children, significant dysfunction of the neural retina may persist even after the vascular abnormalities have resolved. ¹⁷ Resolution of the vascular abnormalities is a common occurrence in patients with ROP. ²³  

Although there is evidence that abnormalities of both the retinal blood vessels and the neural retina occur in ROP, whether and how the neural and vascular abnormalities are related is unknown. As in the preterm infant, the vascular coverage of the retina is incomplete, and the rod photoreceptors are immature in the newborn rat. ^{24 25} Exposure of the newborn rat to alternating high and low ambient oxygen levels produces fluctuating arterial oxygen levels that mimic those encountered in preterm infants and reliably produces retinal vascular abnormalities. ^{6 26} We have used noninvasive procedures to obtain longitudinal measures of both the retinal vasculature and function of the neural retina in this rat model of ROP and in age-matched controls. 

Materials and Methods

Procedures

Induction of Retinopathy.

On the day of birth, newborn Sprague-Dawley rats (Charles River Laboratories, Inc., Worcester, MA) and nursing mother were placed in an oxygen-controlled environment (OxyCycler; Biospherix Ltd., Redfield, NY). Ambient oxygen concentration was monitored continuously, and alternated every 24 hours between 50% ± 1% and 10% ± 1% (Fig. 1) . This exposure paradigm is similar to that used by Penn et al. ⁶ Inside the chamber, the dark–light cycle was 12 hours dark and 12 hours light (10 lux). On postnatal day (P) 14, the rats were moved to room air, and the dark–light cycle was maintained. 

Fundus Photography.

Digital images of the fundus were obtained with a fundus camera (RetCam; Massie Research Laboratories, Inc., Dublin, CA). A corneal contact lens designed to provide a 130° field of view in the premature human infant (the ROP lens) was used. All images had a resolution of 640 × 480 pixels and a color depth of 24 bits. Images, such as the one shown in Figure 2a , were obtained from the right eye. Ophthalmoscopy indicated that the fundus appearance was similar in the left and right eyes. Images were centered on the optic nerve head and extended to the equator, as indicated by the vortex veins (Fig. 2b) . The meridional line demarcated the boundary of dorsal and ventral retina ²⁷ and was used to orient the images uniformly. 

The major retinal vessels formed a radial pattern centered on the optic nerve head (Fig. 2b) . Major arterioles and venules were identified and assigned a tortuosity index (TI). TI is the ratio of the actual length (AL) to the straight-line length (SL) of a blood vessel, ^{28 29} as measured with ImageJ (available at http://rsb.info.nih.gov/nih-image; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). A TI of 1 indicates a perfectly straight vessel. When a bifurcation was encountered (Fig. 2c , arrows), the tracing followed the thicker branch. Measurements were taken in two segments: first to the bifurcation and then from the bifurcation along the selected branch. TI was calculated using the sum of the lengths of these two segments. If the thickness of the two branches appeared the same, one branch was selected by flip of a coin. For each image, three to nine arterioles (median: five) and three to nine venules (median: six) were measured, and the mean tortuosity index for the arterioles (TI _A) and for the venules (TI _V) was calculated. 

Electroretinography.

After at least 16 hours of dark adaptation, the rat was prepared for recording. Under dim red illumination, a Burian-Allen electrode (Hansen Laboratories, Coralville, IA) was placed on the left cornea. The ground electrode was placed on the tail. A heating pad was used to maintain body temperature. When performed in the same session with fundus images, ERGs were recorded first. 

As previously described, ^{20 30 31} ERG responses to a range of brief (<1 ms), white flashes were recorded. The stimuli (Novatron, Dallas, TX) were delivered through a 41-cm integrating sphere and were controlled in intensity by calibrated neutral-density filters (Wratten filters; Eastman-Kodak, Rochester, NY). The stimuli were increased in 0.3-log-unit steps, from dim flashes that evoked a small b-wave (<15 μV), to those that saturated the a-wave amplitude. The records were amplified, digitized, and stored (UTAS-E 2000 system; LKC, Inc., Gaithersburg, MD). Stimuli were delivered at a rate that did not attenuate subsequent response amplitudes. 

The unattenuated flash, measured at the position of the rat’s eye using an integrating radiometer (model S350; United Detector Technology, Orlando, FL) produced 4.6 log μW/cm². Based on the absorbance of the ocular media and the outer segment length and end-on collecting area of the normal adult rat rod, ³² this flash was estimated to produce approximately 135,000 photoisomerizations of rhodopsin per rod (R*). This value was used throughout this study. 

The characteristics of the activation of the rod photoresponse were calculated by fitting the Hood and Birch ³³ formulation of the Lamb and Pugh ^{34 35} model of the activation of phototransduction to the a-wave of the ERG. The model is summarized as  

\[R(i,t)\ {=}\ {\{}1\ {-}\ \mathrm{exp}{[}{-}{\frac{1}{2}}\ {\cdot}\ i\ {\cdot}\ S\ {\cdot}\ (t\ {-}\ t_{\mathrm{d}})^{2}{]}{\}}\ {\cdot}\ Rm_{P3}\ \mathrm{for}\ t\ {>}\ t_{d}\]

where i is the stimulus intensity (R*), S is a sensitivity parameter (R*⁻¹ · s⁻²) based on the time constants involved in the activation of phototransduction, t _d is a brief delay (in seconds), and Rm _P3 is the saturated rod photoresponse amplitude (in microvolts). Rm _P3 is proportional to the number of channels in the outer segment membrane available for closure by light. A least-squares minimization procedure (fmins; MATLAB, The MathWorks, Natick, MA) was used to find the values of S, Rm _P3, and t _d in the equation that best fit the data. Fitting was restricted to the leading edge of the a-wave response, before obvious intrusion of the b-wave or to a maximum of 20 ms after the stimulus. 

The b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave or, in the case of young ROP rats, to a maximum of 160 ms after the stimulus. The stimulus-response function  

\[V(i)\ {=}\ {[}i/(i\ {+}\ {\varsigma}){]}\ {\cdot}\ V_{\mathrm{max}}\]

was fit to the b-wave amplitudes by using an iterative procedure that minimized the mean square deviation of the data from the equation. All parameters were free to vary. In this equation, V is the amplitude of the b-wave (in microvolts), V _max is the saturated amplitude of the b-wave (in microvolts), i is the stimulus intensity (R*), and ς is the stimulus intensity (R*) that evokes a b-wave of half-maximum amplitude. Thus, 1/ς provides a measure of b-wave sensitivity. Responses to high flash intensities, at which a-wave intrusion occurred, were not included in the fit. ³⁶  

Subjects

Albino rats from seven litters were studied. Newborn rats from four litters were oxygen exposed. ⁶ We refer to these animals as ROP rats. Rat pups from the other litters were raised in room air (21% oxygen) as control subjects. Pups were weaned at 25 to 28 days of age. Before all experimental procedures, animals were anesthetized with an intraperitoneal injection of 75 to 100 mg/kg ketamine and 10 mg/kg xylazine. The pupils were dilated with a combination of 2.5% phenylephrine hydrochloride and 1% cyclopentolate hydrochloride. The corneas were anesthetized with a drop of 0.5% proparacaine hydrochloride. All procedures in this study were approved by the Animal Care and Use Committee at Children’s Hospital, Boston, and were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Experimental Design and Statistical Analysis

Data were collected at two ages: P16 to P19 (median: 18 days), and P29 to P35 (median: 31 days). We refer to these as 18-day-old and 31-day-old rats. The control rats (n = 25) participated in all experiments. Some ROP rats were used in multiple experiments. In experiment 1, digital fundus images were obtained from ROP (n = 14) and control (n = 25) rats at both ages. In experiment 2, ERGs were recorded from ROP (n = 11) and control (n = 25) rats at both ages. In experiment 3, both fundus images and ERGs were obtained from each ROP rat (n = 12) at the younger age. 

In experiment 1, a three-factor analysis of variance was conducted to evaluate changes in the tortuosity of the vessels (arterioles versus venules) with age (18 versus 31 days) and group (ROP versus control). In experiment 2, discrete two-factor (age, group) analyses of variance were conducted on the ERG parameters (S, Rm _P3, log ς, and V _max). To evaluate blood vessel and ERG parameters for significant relation in experiment 3, Pearson product moment and Spearman rank order correlations were calculated. The significance level for all tests was P ≤ 0.01. 

Results

Experiment 1

Figure 3shows fundus images of two ROP rats and one control rat photographed at two ages. In both ROP rats, TI _A was higher at the younger age. Indeed, the TI of nearly every arteriole decreased by age 31 days. In the control, TI _A remained nearly the same at both ages. TI _A and TI _V are summarized in Figure 4 . In 18-day-old rats, the TI _A of every ROP rat was greater than in any control, and TI _A was more variable in ROP than in control rats (F = 17.3, df = 1,37, P < 0.01). TI _A decreased in every ROP rat. By 31 days, only five remained above control levels. In contrast to TI _A, TI _V in ROP rats was only slightly higher than in control animals, and changed little with age. Analysis of variance (Table 1)indicated significant main effects of group (ROP versus control), age (18 versus 31 days), and vessel type (arteriole versus venule). 

Experiment 2

Sample ERG records (Fig. 5a)show that the amplitudes of the a- and b-waves were smaller in the ROP rat than in the control and that the stimulus intensity necessary to elicit a just-detectable b-wave was lower in the control. Sample fits of equation 1to the a-wave (Fig. 5b)at the younger age show that the saturated amplitude (Rm _P3) and sensitivity (S) of the photoreceptor response are lower in the ROP rat. Sample fits of equation 2to b-wave amplitudes (Fig. 5c)show that the postreceptor response amplitude (V _max) and sensitivity (1/ς) are lower in the ROP rat. 

In Figure 6and Table 2 , ERG response parameters in 18- and 31-day-old ROP and control rats are compared. All four ERG parameters (S, Rm _P3, log ς, and V _max) differed significantly between ROP and control rats (Table 2) . In ROP rats, only log ς improved to equal that of control subjects (Fig. 6c)at age 31 days. Analyses of variance (Table 2)indicate no interactions of other ERG parameters (S, Rm _P3, V _max) with age. In ROP rats, mild increases in S and V _max paralleled the control, whereas Rm _P3 did not change with age in either ROP or control rats. 

Deficits in receptor and post-receptor parameters are compared in Figure 7 . In 18- and 31-day-old ROP rats, relative attenuation of the saturated amplitude of the post-receptor response (V _max) exceeded that of the receptor response Rm _P3 (18 days: t = 3.54, df = 10, P < 0.01; 31 days: t = 5.96, df = 10, P < 0.01). At 18 days, deficits in log ς were larger than deficits in S (18 days: t = 3.28, df = 10, P < 0.01). At 31 days, the distribution of log ς was similar in ROP and control rats. 

Experiment 3

The TI _A and ERG parameters obtained in the same session from young ROP and control rats are expressed as a proportion of mean control values in Figure 8 . Consistent with the results of experiments 1 and 2, the locus of ROP points is distinct from that of control subjects. However, no ERG parameter in the ROP rats correlated significantly (Pearson product moment, Spearman rank order) with TI _A. Likewise, no correlation of TI _V with any ERG parameter was found in the ROP rats (not shown). 

Discussion

Both the retinal vasculature (Figs. 3 4 ; Table 1 ) and the function of the neural retina (Figs. 5 6 ; Table 2 ) were altered in this rat model of ROP. Although the tortuosity of the arterioles decreased without treatment in every ROP rat, TI _A in approximately a third remained greater than in any control rat at age 31 days (Fig. 4) . The factors that support resolution of ROP are unknown, but, as in this rat model, the vascular abnormalities of ROP resolve spontaneously in most affected infants. ²³ The function of the postreceptor retina, as assessed by log ς, was markedly abnormal at age 18 days, but became normal in every ROP rat by age 31 days. Significant deficits in all other retinal response parameters persisted (Fig. 6 ; Table 2 ). Despite conspicuous abnormalities in TI _A and log ς in 18-day-old ROP rats, these vascular and neural parameters did not correlate (Fig. 8) . Our experiments did not address the role of the choroidal circulation that supplies the photoreceptors. ³⁷  

The low levels of S and Rm _P3 in ROP rats (Fig. 6)may be due to a decrease in the number of rod photoreceptors, short rod outer segments, inefficient phototransduction processes, or even invalid estimation of R*. These explanations are not mutually exclusive. Exposures of adult rats to high or low oxygen, cause apoptotic cell death of mature rods ³⁸ and may cause the death of immature rods in the ROP model studied. Indeed, in another rat model of ROP, ²¹ the thickness of the outer segment and outer nuclear layers at age 18 days was slightly less in ROP than control retinas, suggesting some reduction in the length of outer segments and the number of rods in the ROP retina compared with the control. However, in that ROP model, the rhodopsin content of the retina equaled that in the control; low S and Rm _P3 appeared to be due to inefficient transduction processes in disorganized rod outer segments rather than significant decreases in the number of rods or rod outer segment length. Thus, impaired mobility of the transduction cascade proteins in the disc membranes, or impaired closure of channels in the rod outer segment, cannot be excluded as a possible explanation of low S and Rm _P3 in the ROP rats in this study (Fig. 5) . 

For the calculation of S and Rm _P3 we used a constant value of R* for ROP and control rats because ocular media density, collecting area, and outer segment lengths are not known for ROP rats at the two ages studied. In the 31-day-old control animals average S was ∼10 R*⁻¹ · s⁻², which is in reasonable agreement with published values for dark-adapted, mature rodent rods. ^{31 39 40} In 18-day-old control rats, the slightly lower S and Rm _P3 are consistent with those reported in normal, immature rat rods, which have proportionately shorter outer segment lengths and lower rhodopsin content than do 31-day-old rats. ³¹ Assuming that the stimuli saturated the ROP rod photoresponse, it is unlikely that an error in estimation of R* explains low Rm _P3. In contrast, differences in ROP rod structure could, as in other ROP rats, ²¹ account for both low Rm _P3 and low S. 

The waveform of the ERG response in the young ROP rats (Fig. 5a)shows conspicuous attenuation of the b-wave. In a rat model of ROP, induced by exposure to high ambient oxygen from ages 0 to 14 days, the outer plexiform layer was obliterated, and b-wave amplitudes studied at ages 30 and 60 days were attenuated. ²² Thus, in our young animals, transmission from photoreceptors to second-order retinal cells may be impaired. This is consistent with the finding that postreceptor response amplitudes are more attenuated than receptor amplitudes (Figs. 7a 7c) , and is reminiscent of the dynamic changes in the photoreceptor to bipolar cell synapse proposed in a rhodopsin mutant rat. ⁴¹  

The postreceptor sensitivity, log ς, shows apparent full recovery (Figs. 6c 7d ; Table 2 ). The change in the mean value of log ς between ages 18 and 31 days was large, nearly a log unit, whereas the increment of change in the control rats was only 0.11 log unit. Any error inherent in our method of estimating R* to specify stimulus strength in equation 2is unlikely to explain alone the large improvement in ROP log ς, especially since the ROP rats’ improvement in S between ages 18 and 31 days was rather modest. This leads us to suspect remodeling of the inner retina has occurred in these animals. ⁴² Resolution of blood vessel tortuosity, that is, remodeling of the retinal arterioles, also occurs (Fig. 4 , left panel). Remodeling of vascular and neural components in the ROP retina may share molecular pathways, as is the case in the normal immature retina. ⁴³  

 Supported by a grant from the Massachusetts Lions Eye Research Foundation.

 Submitted for publication January 6, 2006; revised February 3, 2006; accepted April 24, 2006.

 Disclosure: K. Liu, None; J.D. Akula, None; C. Falk, None; R.M. Hansen, None; A.B. Fulton, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Oxygen exposure paradigm. Starting on the day of birth, ambient oxygen concentrations alternated every 24 hours between 50% and 10% for 14 days. ERGs and fundus photographs were obtained at median age 18 and 31 days (arrows).

Figure 2.

 

Fundus image and TI. (a) Sample fundus photograph from a P17 ROP rat. The shape of the image is determined by the camera. (b) The optic nerve head (ONH) is at the center of the image, which extends to approximately the equator as indicated by the vortex veins. The greater thickness and contrast of the venules distinguishes them from arterioles. The meridional line, just inferior to the ONH, demarcates the dorsal and ventral retina. An arteriole is traced from the center of the ONH to the edge of the image. To calculate TI, the actual length (AL) of the arteriole was divided by the straight-line length (SL) from the center of the ONH. (c) Line tracings of the arterioles (thin lines), venules (thick lines), and the meridional line (dotted line). Arrow: indicate bifurcations. (d) The results of the TI measurements are shown for this eye.

Figure 3.

 

Sample fundus images from the right eyes of ROP and control rats selected for TI _A close to the median for the group. The images in the top and middle rows are from two ROP rats, each photographed at both P16 and P31. The images in the bottom row are from a control rat photographed at P17 and P33. A tracing of the arterioles (thin lines), venules (thick lines), and meridional line (dotted line) is presented to the right of each image. The tortuosity index of each arteriole is shown at the end of each tracing. Mean TI _A is as indicated for each image. In both ROP rats, TI _A was higher at the younger age. TI _A in the control rat changed little with age.

Figure 4.

 

For individual ROP and control rats, TI _A (left panel) and TI _V (right panel) are plotted. Lines connect observations from an individual rat at two ages. In every ROP rat, TI _A decreased with age, whereas in control rats TI _A changed little. TI _V changed little with age or group (ROP versus control).

Table 1.

 

Tortuosity Index of Arterioles and Venules: ROP versus Control Rats

Vessel Type	Age (d)	ROP (n = 14)	Control (n = 25)	ANOVA*
Arterioles (TI A)	18	1.100 (0.032)	1.027 (0.008)	F group = 92.4; P < 0.01
	31	1.048 (0.013)	1.032 (0.009)	F age = 45.8; P < 0.01
				F vessel = 37.0; P < 0.01
				F group × age = 76.8; P < 0.01
Venules (TI V)	18	1.048 (0.008)	1.031 (0.007)	F group × vessel = 58.6; P < 0.01
	31	1.042 (0.015)	1.034 (0.005)	F age × vessel = 67.8; P < 0.01
				F group × age × vessel = 86.2; P < 0.01

 Data are the mean (SD).

*  Age and vessel type factors are within subjects. Group factor is between subjects. df ₁ = 1, df ₂ = 37. Significance defined as P ≤ 0.01.

Figure 5.

 

Sample ERG records of an ROP and a control rat. (a) Responses were recorded at the ages indicated above each family of traces. The numbers to the left of each row of traces indicate the stimulus intensity (log R*). For clarity, only every third record is shown. The start of each trace coincides with stimulus onset. The b-wave amplitude is measured from the trough of the a-wave to the peak of the b-wave (vertical line). In the ROP rat at the younger age, the b-wave failed to rise above the baseline. Therefore, the b-wave was measured arbitrarily at 160 ms after the stimulus (arrow). The calibration bars to the bottom right apply to all traces. (b) Sample fits of equation 1(dashed lines) to the a-wave of the P16 ROP and P17 control rats. The equation is fit to the initial 20 ms of responses (solid lines) to the 4.8 to 3.6 log R* stimuli. The control rat’s parameters are S = 6.17 R*⁻¹ · s⁻² and Rm _P3 = 251 μV; the ROP rat’s are S = 4.02 R*⁻¹ · s⁻² and Rm _P3 = 152 μV. The calibration bars at the bottom left apply to both families of a-waves. (c) Fits of equation 2(dashed lines) to the b-waves in the P16 ROP and P17 control rats. For the control rat V _max = 281 μV and log ς = 0.35 log R*. For the ROP rat V _max = 118 μV and log ς = 1.74 log R*. The rightward shift of the ROP curve indicates lower sensitivity. The difference between log ς in the ROP and control rats is 1.39 log units.

Figure 6.

 

The four ERG parameters (S, Rm _P3, log ς, and V _max) in ROP and control rats. Observations for an individual at the two ages are connected by a straight line.

Table 2.

 

ERG Response Parameters: ROP versus Control Rats

Parameter	Age (d)	ROP (n = 11)	Control (n = 25)	ANOVA*
S (R* −1 · s−2)	18	4.65 (2.32)	7.04 (2.47)	F group = 7.06; P = 0.01
	31	8.32 (2.67)	9.88 (3.09)	F age = 25.7; P < 0.01
				F group × age = 0.43; P = 0.52
Rm P3 (μV)	18	144 (60)	266 (74)	F group = 38.3; P < 0.01
	31	151 (39)	246 (48)	F age = 0.25; P = 0.62
				F group × age = 1.32; P = 0.26
Log ς (log R* )	18	1.17 (0.58)	0.37 (0.15)	F group = 37.4; P < 0.01
	31	0.31 (0.07)	0.26 (0.12)	F age = 67.1; P < 0.01
				F group × age = 40.0; P < 0.01
V max (μV)	18	102 (57)	297 (101)	F group = 113; P < 0.01
	31	172 (68)	405 (61)	F age = 20.4; P < 0.01
				F group × age = 0.89; P = 0.35

 Data are the mean (SD).

*  Age factor is within subjects. Group factor is between subjects. df ₁ = 1, df ₂ = 34. Significance defined as P ≤ 0.01.

Figure 7.

 

Deficits in saturated amplitude and sensitivity of photoreceptor and post-receptor ERG response parameters at 18 days (a, b) and 31 days (c, d). The dotted horizontal and vertical lines intersect at (0,0), the control means. The dotted diagonal line has the slope of unity.

Figure 8.

 

Mean arteriolar vessel TI _A and ERG response parameters in 18 day ROP and control rats are shown as a proportion of the control mean. Although the locus of points for ROP rats is distinct from that for the control animals, no significant correlation of TI _A and ERG parameters was found. The proportion of the variability accounted for by the product moment correlation r ² is given at the bottom right.