Oscillatory Potential Analysis and ERGs of Normal and Diabetic Rats

Abstract

purpose. To identify and characterize the early alterations of the ERG oscillatory potentials (OPs) under conditions of poor glycemic control associated with diabetes in an animal model. To characterize and correlate the a- and b-wave properties of the ERGs of diabetic animals in parallel with the changes in oscillatory potentials.

methods. Blood sugars, weights, and ERGs were measured for a group of rats each week for 3 weeks to obtain baseline information. A single injection of streptozotocin was given to the experimental animals. Weights, blood sugar, glycosylated hemoglobin, and detailed ERGs were recorded weekly for up to 12 weeks in control and experimental animals.

results. OP kinetics were found to be inherently related to amplitude. Amplitude-matched OPs were delayed in diabetic animals when compared with baseline data from the same animal. The kinetics of OPs in control animals stayed the same or were slightly accelerated relative to their baseline values. For a given recording condition, OP kinetics were very stable over time and this stability was not disturbed in diabetic animals. Diabetic animals showed a small but significant delay in the a-wave, but no change in b-wave timing. The sensitivity of the b-wave was reduced twofold, but that of the a-wave was not changed.

conclusions. OPs have been used to evaluate retinal function in both diabetic models and patients. The comparison of amplitude-matched OPs is a robust determinant of changes in kinetics. Researchers and clinicians who use OPs may wish to consider the relationship between OP amplitude and kinetics to avoid confounding assessments of abnormalities. The amplitude versus kinetics relationship does not change form in diabetic animals; it is merely shifted (delayed) on the time axis.

The electroretinogram (ERG) has been used for decades to uncover the mechanisms of retinal physiology and their alterations in disease. Individual components of the ERG were described by Granit in 1933. ¹ Cobb and Morton ² described the oscillatory potentials (OPs) as a new component of the ERG in 1953. Soon after their description, OPs were identified as sensitive indicators of disease in diabetic retinopathy ³ and, as such, have now been studied in a variety of diseases such as glaucoma, ^{4 5 6} vascular occlusions, ^{7 8 9} and congenital eye diseases. ¹⁰ Although the characteristics of the OPs and their transformation in disease is being investigated, the basic stimulus dependence of the OPs and their timing has not yet been explored in detail. 

The OPs are four to six wavelets in the ERG that are present on the rising phase of the b-wave. ¹¹ OPs were first shown to arise in the inner plexiform layer ¹² but to have different retinal depth profiles for individual peaks. ¹³ The individual peaks have also been shown to have distinct pharmacologic properties, with the earlier peaks being selectively diminished by γ-aminobutyric acid (GABA) antagonists and the later peaks being sensitive to glycine blockers. ¹⁴ Which specific cells in the retina are responsible for the OPs is still being debated. It has been suggested that OPs are generated by the amacrine cells, because their retinal depth was shown to be similar to that of the amacrine cells ¹³ and because of pharmacologic studies ¹⁵ showing that dopamine, ¹³ GABA, and glycine blockers ¹⁴ all diminish the OPs. 

As the leading cause of blindness in adults 21 to 75 years old, diabetic retinopathy is an important public health problem. ¹⁶ Strict glycemic control as well as early treatment have been shown to improve outcomes for diabetic patients. ^{17 18} Although clinical exams often focus on the visualization of retinal lesions in diabetic patients, electrophysiologic changes have been shown to occur before the onset of clinically evident retinopathy. ^{19 20 21 22} Thus, an electrophysiological assessment of retinal function could be an extremely valuable monitoring metric for retinal health in diabetic patients. A reduction in the a-wave amplitude has been demonstrated in patients with type 1 diabetes without clinical indices of retinopathy. ^{19 23 24 25} Delays in the a-wave implicit times have also been reported in diabetic patients compared with control subjects. ²⁶ However, the most consistently reported index of disease in diabetes is the OP. It has been repeatedly reported that the OPs have reduced amplitudes in diabetes. ^{19 27 28} Abnormalities in the OPs are of great interest in the assessment of diabetic retinopathy, because OP abnormalities have been shown to predict onset as well as progression of diabetic retinopathy, ^{21 27} but which abnormalities in the OP are predictive is not clear. 

Sakai et al. ²² looked at the timing of ERG changes in young rats treated with streptozotocin (STZ). In their study, a- and b-wave amplitudes and implicit times remained unchanged in diabetes. They examined three of the OP peaks and found they were reduced in amplitude, although this reduction was significant only for the second OP peak. They observed OP delays that appeared as early as 2 weeks after STZ administration and preceded clinically visible lesions. 

It has been independently reported that the peaks of OPs are reduced in amplitude ^{21 22 28 29} and delayed. ^{20 26} However, herein we report that these two variables are inexorable related and one cannot validly consider delays without matching amplitudes and vice versa. Our results confirmed the finding that a-waves are delayed; in addition, we observed that b-wave sensitivity is reduced in diabetic animals. 

Methods

Long-Evans male rats were used and handled according to the principles of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Two groups of Long-Evans rats were studied to correlate retinal function with induced diabetic state. The elevated blood glucose and poor glycemic control found in diabetic patients was produced in young rats by a single injection of STZ (65 mg/kg intraperitoneally). Animals were 8 to 9 weeks old and averaged 222 g at the time of injection. Some injected animals had temporary elevations of blood sugar, but reverted to normal blood sugar levels and were omitted from the study. Because blood glucose measurements are associated with significant variability, we used glycosylated hemoglobin measurements to further establish the distinction between diabetic and control animals. Glycosylated hemoglobin is a representation of blood sugar over time and is used clinically, not to diagnose, but to observe patients once they are identified as diabetic. Fasting blood glucose, glycosylated hemoglobin and ERGs were measured weekly for 8 to 14 weeks. Blood was obtained from the lateral tail vein with a 22-guage needle while the animal was under halothane anesthesia. Blood sugar was measured with a blood glucose meter (Accu-chek Advantage; Roche Diagnostics, Indianapolis, IN). By the same method, blood was obtained for glycosylated hemoglobin measurement (A1c Now Monitor; Metrika, Sunnyvale, CA). Animals were fasted overnight before testing. 

For ERG studies, animals were anesthetized with xylazine (10 mg/kg intramuscularly) and ketamine (50 mg/kg intramuscularly) and kept in darkness for 45 minutes before recording. Body temperature was maintained at 38°C by a heating pad (Braintree Scientific, Braintree, MA) during recording. Corneas were anesthetized with proparacaine (0.5%) and pupils dilated with topical phenylephrine HCl (2.5%) and tropicamide (1%). Only the eye to receive light stimuli was dilated. Most animals recovered from anesthesia, but diabetic animals were more sensitive to the anesthetics; data were discarded from any animal that did not survive the full course of 7 weeks. 

The light source was a 100-W tungsten-halogen lamp focused onto one end of a fiber optic. Stimulus duration, typically 2 ms, was controlled with a shutter with a 6-mm aperture (Uniblitz; Vincent Associates, Rochester, NY). The energy output of the flashes was calibrated daily. The equivalent pulse-width of a square-wave with the energy delivered by the flash would have a duration of 1.85 ms, and we found this to be very consistent day to day. Stimulus strength was controlled by a set of calibrated inconel neutral-density filters that allowed attenuation in steps of approximately 0.3 log units up to a maximum of 6.9 log units attenuation. The unattenuated stimulus was calibrated daily with an optical power meter (Graseby Optronics, Orlando, FL). The wavelength of the stimulus, 505 nm (35 nm bandwidth), was determined by a three-cavity interference filter (Andover Co., Salem, NH). White light was also used to extend the intensity range, and these stimuli were corrected to their effective intensity at 505 nm. We converted the white-light stimulus energy to effective 505-nm photons empirically. We measured the complete intensity–response functions using both white light and 505-nm light in five animals on four to eight separate occasions. For each set of paired intensity–response relations, a correction factor was measured to align the data sets. The white-light photon density was derived by arbitrarily using the energy of a 650-nm photon. The mean of 28 such measures was then used as the correction factor for converting white-light energy into equivalent 505-nm photons. All stimulus intensities are reported as photon density (in photons per square micrometer) emanating from the 4.2 mm diameter opening of the electrode/diffuser in contact with the cornea (see below). The light intensities used were not sufficient to observe saturation of the upper limb of the b-wave clearly; therefore, no conclusions were drawn from this portion of the b-wave data set. 

An electrode designed after that described by Lyubarsky and Pugh ³⁰ was used on the experimental eye. A platinum wire loop was affixed to the tapered end of a Plexiglas rod that had been hollowed out to receive the fiber optic. This arrangement was designed to ensure a constant distance between the fiber optic and the eye and to act as a diffusing element. A second platinum loop served as the neutral electrode and was placed on the other eye. Responses were amplified 5000×, high-pass filtered with a 10-Hz cutoff frequency (AC 0.1), and low-pass filtered at 300 Hz using an amplifier (Astro-med CP122W; Grass Telefactor, W. Warwick, RI). The ERG voltage and stimulus-monitor signals were digitized with hardware (MIO16) and software (LabView) from National Instruments (Austin, TX). Data was recorded at either 0.2 or 0.5 ms/pt. A stimulus set consisted of 3 to 20 responses to the same wavelength and intensity of light. 

The amplitude-intensity relationship of the b-wave and a-wave were plotted and fitted with modified Michaelis function with the form:  

\[r\ {=}\ R_{\mathrm{max}}\ {_\ast}\ (i^{n})/(i^{n}\ {+}\ k^{n})\ {+}\ \mathrm{base}\]

The b-wave functions had an upper and a lower limb that were fit independently to a modified Michaelis function. ^{31 32 33} For the a-wave data and the lower limb of the b-wave data the base was set to zero. 

Results

Two groups of Long-Evans rats were studied to correlate retinal function, as measured by the electroretinogram (ERG), with induced diabetic state. For the purposes of this study, a diabetic animal is one that was treated with STZ and had fasting blood sugars greater than 200 mg/dL within 3 weeks of injection. Control animals had fasting blood glucose levels of 94.7 ± 4.0 mg/dL, not significantly different from the value of 80.1 ± 4.9 mg/dL found during the baseline weeks (before STZ injection) of the diabetic group. After STZ treatment, the blood glucose levels increased to an average of 391 mg/dL. Diabetic animals also had elevated glycosylated hemoglobin values, from 9% to higher than 13%, the upper limit of our tests. The glycosylated hemoglobin level of the control group averaged 3.9% and was never higher than 4.7%, whereas that of the diabetic group averaged 10.9% after STZ treatment. The starting weight of the animals in the two groups was not different, but the control group gained weight (+20 g/week), whereas the diabetic group was relatively static (+2 g/week) after STZ treatment. Clarity of the cornea and lens were evaluated weekly. One animal, in which a cataract developed, was removed from the study. To evaluate visual function, we recorded the dark-adapted flash responses with stimuli covering 4 to 5 log units of intensity. 

A family of ERG responses to brief flashes of light is illustrated in Figure 1 . Responses from a normal rat at 7 of the 14 stimulus intensities presented are shown in Figure 1a . Figure 1b shows similar responses from a representative diabetic rat. Qualitatively, these ERGs are not appreciably different. However, a quantitative evaluation of the timing and amplitude of the a- and b-waves comparing the baseline weeks to the final 2 weeks of study revealed some differences. In control animals the b-wave implicit time and amplitude were not significantly different over the 8 to 11 weeks of the study. For the diabetic groups however, the amplitude of the b-wave declined significantly (P = 0.007) and the implicit time was unchanged. The statistical analysis was performed on paired values for each animal, comparing the average of the baseline weeks to the last 2 weeks of the study for that animal. Thus, the mean values themselves, as given in Table 1 , may not be as informative as the sets of paired values. For example, in eight of nine of the diabetic rats, the implicit times for the a-waves showed a delay, whereas the results for the control animals were evenly mixed: three delayed, three accelerated, and one identical with the baseline value. Therefore, the implicit time of the a-wave was significantly delayed in the diabetic group (P < 0.001) and not significantly changed in the control group. The amplitudes of the a-wave for both groups declined significantly, although the diabetic group’s decline (34%) was somewhat larger than that of the control group (28%). 

The peak amplitudes of the b-wave response were plotted against the log of the intensity of the stimulus that evoked the response. Figure 2 shows the intensity–response relation for the b-wave for a control (Fig. 2a) and diabetic (Fig. 2b) rats. Similar intensity–response functions were recorded each week for each animal. The smooth curves are fittings of a modified Michaelis function (see the Methods section). The upper and lower limbs of the b-wave function were fitted independently after Peachey et al. ³³ The results for the control animal (Fig 2a) show that the maximum amplitude is reduced, but the curve is not shifted along the intensity axis. The K-values of Table 1 are results from the fitting of the lower limb of the b-wave. In the diabetic animals, the amplitude was more significantly reduced and the curve is shifted to the right, indicating desensitization of the b-wave. On average, the diabetic animals showed a 0.3-log unit desensitization, or a 2.1-fold increase in the light intensity required to produce a half-maximum response for the lower limb of the b-wave. 

The OPs were isolated by band-pass filtering the retinal response between 34 and 300 Hz. We chose 34 Hz as a cutoff frequency to avoid any loss of signal power, especially for the slower OPs of diabetic animals. Figure 3 demonstrates that both timing and amplitude of OPs are intensity dependent. Figure 3a shows a family of OPs recorded from the six brightest flash intensities of the ERG series. The highest-intensity stimulus produced the largest, fastest OPs. Then, as the stimulus strength was reduced roughly by factors of two, the OPs were reduced in amplitude, and their peaks were delayed. Figure 3b demonstrates the method used to measure OP amplitude (sum of a₁, a₂, a₃, and a₄) and OP TTP (sum of t₁, t₂, t₃, and t₄). Figure 3c plots the OP timing, as measured by the sum of the TTP for the four largest peaks, versus the log intensity. A 1-log unit reduction in light reaching the eye resulted in an approximately 15-ms delay in TTP. Typically, our signal-to-noise ratio permitted measurement of OPs of as little as 10 μV. Thus, we could observe and measure OPs at five to nine intensities. There is clearly a linear relationship of OP TTP versus log I over the range of intensities investigated. There is also a linear relation between the OP amplitude and log I (Fig. 3d) . Therefore, it is not surprising that combining the TTP versus log I and amplitude versus log I relations results in a linear relationship between the amplitude and the TTP that is shown in Figure 3e . 

Treatment with STZ does not induce a short-term toxic effect manifest in the ERG, but several weeks of poor glycemic control produces delays in the OP timing, as illustrated in Figure 4 . Shown in black is the OP recorded from this animal at baseline, before treatment with STZ. The thin gray trace shows the OP recorded from that animal 1 week after STZ treatment. Comparison of these two traces shows relative similarity, with a slight delay of 4.9 ms in TTP. Thus, the changes in kinetics described are not due to an immediate neurotoxic effect of STZ. The thick gray trace in Figure 4 is an amplitude-matched OP recorded 5 weeks after STZ-treatment. The OPs in this trace are similar in shape and amplitude to the other traces, but the peaks are delayed with respect to the trace from the baseline week (Fig. 4 , arrows). The delay in TTP (OP delay), taken as the sum of the delays for four peaks, was 29.3 ms, close to the mean OP delay seen for all diabetic animals, of 30.8 ms. The latter peaks are somewhat more delayed than the earlier peaks, meaning that the OPs are spread out in addition to being delayed by the increased latency of the first peak. The change in the time between the first and fourth peaks was 4 ms, representing about a 10% spreading of the OPs. For nine diabetic animals the TTP delay averaged 30.8 ms, whereas the OP spreading averaged 2.7 ms. The control animals also showed a small degree of OP spreading (0.7 ms) over the course of the study. Summary data for TTP delay is presented in Figure 5 , where the sum OP delay is plotted versus the final blood sugar level in seven control animals (open triangles) and nine diabetic animals (filled diamonds). The OP delay was measured as the sum of the individual delays of four peaks of the OP. The final blood sugar was taken with the animal fasting and was selected as a measurement to quantify the diabetic status of the individual animal. The control animals showed no delay and, in fact, showed a small acceleration of their OPs of 2.5 ± 3.8 ms (n = 7, mean ± SEM). Diabetic animals had a clear delay of their OPs compared with amplitude-matched OPs from baseline weeks (30.8 ± 5.6 ms, n = 9) with much greater variability of both the delays and the blood sugars. In the final week, the blood sugars of the diabetic group was 494 ± 51.4 mg/dL and that of the control group was 87 ± 6.3 mg/dL. 

We measured the repeatability of the OP kinetics by presenting 20 flashes of white light spaced at 10-second intervals to a dark-adapted animal. Within such a recording session, the OP kinetics generated by any particular stimulus accelerated slightly from the first to the 20th flash response, but were very stable, both in normal and diabetic animals, as shown in Figure 6 . Figure 6a demonstrates the stability of OP kinetics in a control animal for two sets of recordings taken several weeks apart, early (open triangles) and late in the study (filled triangles). The dashed lines give the mean values. The small decrease in the mean value (1.7 ms) that occurred between the two recordings represents a minor acceleration of the kinetics and is typical of the acceleration that we observed in control animals as they reached maturity (e.g., open triangles, Fig. 5 ). In Figure 6b similar data from a diabetic animal are presented. The open circles are from the baseline week, before treatment with STZ, and the filled circles are recordings from the same animal 6 weeks after STZ administration. The considerable increase in the OP TTP (21.8 ms) illustrates the delay in OP kinetics that we observed in all diabetic animals. Note that the variability of the OP kinetics from flash to flash, or across the 20 flashes is not altered after the onset of diabetes. The slight downward slope in both groups indicates a minor acceleration with repeated stimulation, probably due to light adaptation. The delay of the OPs was not accompanied by a significant shift in their fundamental frequency, as demonstrated by the relative constancy of the interval between first and fourth peaks (for example, Fig. 4 ). Whereas the sum of the TTPs shifted over 30 ms, the spread of the peaks increased only 2.7 ms in diabetic animals and 0.7 ms in control animals. 

We have already demonstrated that OP kinetics are inversely related to their amplitude (Fig. 3e) . To further examine this relationship over time and in the presence of diabetes, we plotted the kinetics of the OPs, taken as the sum of the time to peak for the four peaks (TTP), versus the amplitude of the OPs, taken as the sum of the amplitudes of those same four peaks. Figure 7a shows data from a control animal. Pictured are baseline data from the first 2 weeks (open circles) and data from the final 2 weeks of the study (filled circles). The leftward shift represents the acceleration of the OPs that occurred in our animals as they reached sexual maturity (Figs. 5 , open triangles; 6a ). In the other three panels (Figs. 7b 7c 7d) are similar data from diabetic rats. The open circles are baseline data recorded before treatment with STZ. The filled circles are data from the end of the study, 6, 7, and 8 weeks after STZ treatment (Figs. 7b 7c 7d , respectively). The right shift corresponds to a delay in OP kinetics after several weeks of poor glycemic control. The baseline data were fit with a linear regression, and a parallel line was fit to the later data to indicate that, in these animals, any pair of amplitude-matched OPs showed a similar delay. Regression analysis showed that the slopes of the lines formed by baseline and end-of-study data were not significantly different, but that delays represented by the rightward shift in the diabetic animals were significant. 

Discussion

Diabetes is a disease with both neurologic and vascular components. It has been shown that electrophysiologic changes occur before the appearance of visible lesions. ^{19 21 22} Description of these changes could help our understanding of the pathophysiology of diabetic retinopathy (DR) as well as provide the foundation for the design of new screening tests for diabetic patients. Sakai et al. ²² previously reported that OPs become delayed soon after the onset of elevated blood sugar in an animal model of diabetes and before the onset of vitreous fluorophotometric changes. In our work, we found that OPs are delayed after several weeks of experimentally induced diabetes. Numerous studies point to the OPs as the most sensitive electrophysiologic indicator of DR. ^{19 20 21 28 34} Holopigian et al. ⁶ reported that the b-wave is as sensitive an indicator of DR as the OPs. Although the b-wave implicit time was unchanged, we found that the b-wave amplitude was significantly reduced in our diabetic animals. However, this reduction in amplitude, while significant (P < 0.01), was not found in every animal (6/9 animals), indicating that the OP kinetics are a better index of early disease in DR than the b-wave amplitude. The sensitivity of the lower limb of the b-wave declined over twofold in the diabetic animals, but again not in every case (7/9 animals). Thus, it too is less sensitive than the OPs, and it is a more difficult parameter to evaluate because of the greater number of ERG responses to record and time required for fitting the intensity–response function. The reduction in the sensitivity and amplitude of the b-wave could, by themselves, account for the appearance of the delay in the a-wave, because these two components of the ERG have opposite electrical signs. The reductions of the a- and b-wave amplitudes noted in Table 1 may, in part, be due to an aging effect. A similar effect has been previously observed in mouse ³⁵ and human ERGs. ^{36 37}  

Typical clinical ERG recordings (International Society for Clinical Electrophysiology of Vision [ISCEV] 1999) measure the OP with a series of four bright white flashes presented at 15-second intervals; the first response is discarded. ³⁸ We used a similar paradigm, but extended the stimulus to 20 flashes to observe the stability of the OP kinetics. Although the change in timing was most significant between the first and second flashes (Fig. 6a) there was a steady reduction in the TTPs across the entire 20-flash sequence. The fact that the OPs stability and peak-to-peak timing remain unaffected by diabetes suggests that the neural circuit responsible for generating the OPs is probably intact, but that the triggering mechanism may be desensitized. 

It has been separately reported that OPs are delayed and reduced in amplitude in diabetes. In the current study, we show that the kinetics and amplitude of the OPs are related and that neither should be considered independently. The timing of the OPs is dependent on the stimulus intensity, and future studies of OPs should consider this relationship. If a delay in the OP timing is suspected, it is important that the stimulus intensity be considered, because a false appearance of a delay could be given by a lesser light intensity stimulus, resulting in a smaller amplitude OP. For example, the presence of nonretinal disease, such as a cataract or corneal lesion, would reduce retinal illumination and produce a smaller OP having a slower time course. In such a situation, a slower OP would be due to decreased photon count at the retina and not to a change in retinal function. 

The clinical use of the ERG often focuses on the use of a single-intensity bright flash. ³⁸ Based on our findings in animals, if a single-intensity flash ERG was recorded in a patient before and after the onset of diabetic retinopathy, the OPs would be expected to become both smaller in amplitude and slower in their time course. Clarification of the relationships between OP amplitude, kinetics, and stimulus intensity in human patients would be a step forward. When possible, recording of a subject’s OPs over range of light intensities would serve as a useful baseline for later comparisons of amplitude matched OPs as an indicator of retinal health. 

 Supported by Grant EY10573 from the National Eye Institute (TWK). HAH was funded by the MD/MS program in Physiology and Biophysics.

 Submitted for publication September 30, 2003; revised November 14, 2003; accepted November 15, 2003.

 Disclosure: H.A. Hancock, None; T.W. Kraft, 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.

 

ERG response to a 2-ms flash from normal and diabetic rats. Pictured are dark-adapted responses for 7 of 14 light intensities recorded demonstrating the range of response waveforms. Timing and scale bars apply to all traces. Each trace is the average of 3 to 20 responses to the same stimulus. Intensity values beside each trace are the log photons per square micrometer emitted from the surface of a translucent contact lens electrode (4.2 mm in diameter). (a) Responses from a normal rat; (b) responses from a diabetic rat with blood sugar was five times higher than baseline 2 weeks after STZ injection.

Table 1.

 

Comparison of a- and b- Wave Parameters in the Control and Diabetic Groups

	Implicit Times (ms)			Amplitude (μV)			K-Value
		Control	Diabetic	Control	Diabetic	Control	Diabetic
a-Wave baseline	18.3	17.7	227	225	4.345	4.520
a-Wave end value	19.1 NS	19.6*	163*	148*	4.525 NS	4.451 NS
		(P < 0.001)	(P < 0.05)	(P < 0.01)
b-Wave baseline	67.1	67.5	678	628	1.628	1.787
b-Wave end value	72.3 NS	71.5 NS	589 NS	418*	1.875 NS	2.100*
				(P < 0.01)		(P < 0.05)

 Each number represents the mean for the diabetic (n = 9) or control (n = 7) groups. The K-value is the Log of the intensity (photons/μm²) necessary to produce a half-maximal response, the K-value for the b-wave pertains only to the lower limb of the b-wave intensity response function, as shown in Figure 2 .

*  Significant change from baseline, probability based on a paired t-test of baseline versus end value for each animal in each group; NS, no significant difference.

Figure 2.

 

Intensity–response functions for b-waves recorded in control and experimental animals. The maximum amplitudes of the b-waves measured during baseline weeks and at the end of the study are plotted against the stimulus strength. (a) b-Waves recorded from a control animal in the first week of the study and 9 weeks later. The smooth curves are fits of modified Michaelis functions to the lower limb of the b-wave. (b) b-Waves from an animal in the diabetic group before and 12 weeks after STZ treatment.

Figure 3.

 

OP kinetics and amplitude depended on stimulus intensity. (a) OPs from a single animal recorded in one session to stimuli of several intensities. Traces represent the averaged response of 3 to 20 repeated stimuli. The time course of the peaks was slower at lower light intensities. (b) Method of calculating the TTP and amplitude of individual peaks. The amplitude was measured by taking the maximum amplitude for each peak. The TTP was the time after stimulus to the point of maximum amplitude. The summed amplitude is calculated by adding a₁ + a₂ + a₃ + a₄. The OP summed TTP is the sum of t₁ + t₂ + t₃ + t₄. (c) The TTP for OPs varied with stimulus intensity. The sum of the oscillatory TTP was plotted versus the log of the stimulus intensity and showed a linear decline for much of the intensity range used. Log I is given in log photons per square micrometer. (d) Summed amplitude of OP peaks increased with increasing stimulus intensity. (e) OP kinetics and amplitude are linearly related. OPs of smaller amplitude had a slower summed TTP than OPs of larger amplitudes.

Figure 4.

 

OPs were delayed several weeks after STZ treatment. OPs from a representative animal are shown at baseline and after STZ. Note that OPs do not become delayed until several weeks after treatment. Arrows: direction of the change in timing of the individual peaks.

Figure 5.

 

The OPs were delayed after several weeks of elevated blood sugar levels. The OP delay was measured as the difference between the summed OP TTP for four peaks measured at baseline. This delay, (change in TTP) is plotted against the final blood sugar level. (▵) Control animals, which did not receive STZ injections and did not have elevated blood sugar. (♦) Diabetic animals. The average delay for control animals was −2.5 ms and for diabetic animals was 30.8 ms.

Figure 6.

 

Stability of the OP timing within and between trials. Stability of the OP kinetics was measured by presenting 20 flashes of white light spaced at 10-second intervals. (a) The OP kinetics in a control animal at baseline (▵) versus end of study (week 8, ▴). (b) Diabetic animal at baseline (○) and 7 weeks after STZ injection (•). The OP TTP was taken as the sum of the four OP peaks.

Figure 7.

 

OP kinetics and amplitude were related. The kinetics of the OPs, taken as the sum of the TTP for the four peaks, was plotted versus the amplitude of the OPs, taken as the sum of the amplitudes of those same four peaks. (a) Data from a control animal: (○) baseline data from the first 2 weeks; (•) data from the final 2 weeks of the study. The leftward shift represents an acceleration of the OPs. (b, c, d) On the other three panels are similar data from diabetic rats. (○) Baseline data recorded before treatment with STZ; (•) data from the end of the study, 5 to 7 weeks after STZ treatment. The right shift corresponds to a delay in kinetics with diabetes.

The authors thank Derron Allen and Steve Denny for laboratory assistance, Abi Yildirim and Jerry Millican for excellent technical support, and Micheal Loop for constructive critique of the manuscript.