Study of Rod- and Cone-Driven Oscillatory Potentials in Mice

Abstract

purpose. To characterize rod- and cone-driven oscillatory potentials (OPs) in mice.

methods. Dark- and light-adapted electroretinograms (ERGs) were obtained in three mouse models: wild-type C57BL/6J mouse, cone photoreceptor function loss 1 (cpfl1) mouse, and rhodopsin knockout (rho ^{− / −}) mouse. A Butterworth filter was used to extract OPs from ERG signals. Latencies were calculated from the extracted OPs. Major frequency components were determined from OP power spectra computed using fast Fourier transform (FFT). The total power of the OP signal (an alternative measurement of amplitude) was calculated by numerically integrating the area enclosed by its frequency spectra, which is analogous to the total energy of mechanical vibration.

results. In C57BL/6J mice, dark- and light-adapted OPs had distinctly different peak frequencies (100 to 120 Hz and 70 to 85 Hz, respectively). In cpfl1 mice which possess pure rod ERGs, dark-adapted OPs had a peak frequency similar to those of the wild-type mice, whereas light-adapted ERGs and OPs were not detectable. In rho ^{− / −} mice with pure cone functions, both dark-adapted and light-adapted OPs had peak frequencies of 70 to 90 Hz, which were similar to those obtained from light-adapted OPs in wild-type mice. The total power of cone-driven OPs was less than 5% that of rod-driven OPs. In time-domain, cone-driven OPs occurred approximately 13 ms after rod-driven OPs.

conclusions. Cone- and rod-driven OPs exhibit significantly different characteristics in peak frequency, latency, and total power. By using these characteristics, it is possible to differentiate cone- and rod-driven OPs in mouse models. Understanding these OP features is essential for analyzing OPs.

Oscillatory potentials (OPs) are important components of the electroretinogram (ERG). ^{1 2} They appear as a group of wavelets superimposed on the leading edge of ERG b-wave. Although the exact origin of OPs is unclear, ³ it is commonly believed that they are mainly generated in the proximal retina by neural interactions among bipolar cells, amacrine cells, and ganglion cells. ^{1 2} The oscillations may be initiated by activities of inhibitory feedback circuits in the inner plexiform layer. ^{1 2} OPs may be useful in the diagnosis of a variety of eye diseases, including elevated intraocular pressure, ^{4 5} central retinal vascular occlusion, longstanding systemic hypertension, retinopathy of prematurity, retinal toxicity, and different types of retinal dystrophy and degeneration. ^{1 2 4 5 6 7 8} In addition, both human ^{1 2} and animal ^{1 2 4 5 6 7 8} studies indicate that diabetic retinopathy can lead to reduced OP amplitudes. Because such changes occur earlier than other ocular symptoms such as abnormal vision and other pathologic fundus signs, OPs may be useful in the early diagnosis of diabetic retinopathy. 

For diagnostic purposes, it is important to separate rod- and cone-driven Ops, because the rod and cone pathways are often affected differently during the pathologic process. Abnormalities in one or the other or both may have different implications. ⁹ For example, a recent study demonstrated that the amplitudes of cone-driven OPs are more sensitive than rod-driven OPs to acute IOP elevation. ⁴ However, in practice, OPs are often elicited with high-intensity light stimulus under dark adaptation, according to the current protocol for human OP recording, ¹⁰ recommended by International Society for Clinical Electrophysiology of Vision (ISCEV). Inevitably, OPs extracted from ERG responses recorded under such conditions contain signals generated from both rod and cone pathways. Thus, a thorough understanding of the different characteristics of rod- and cone-driven OPs is critical for interpretation of OP signals. 

The objective of this study was to characterize and compare cone- and rod-driven OPs by using different mouse models. Mice have become increasingly important, not only in the study of a variety of human retinal diseases, but also in the development of new treatments for these diseases. ^{11 12} There are several advantages to performing OP studies in mice. First, mice have higher OP responses than many other animals that have been studied. (Hofeldt KJ et al. IOVS 2004;45:ARVO E-Abstract 820) Secondly, unlike other mammals that have multiple peaks in their OP frequency spectrum, mouse OPs usually show a major single frequency peak (with a subtle frequency peak at the ascending limb of the frequency spectrum), which makes the OPs easier to analyze. (Hofeldt KJ et al. IOVS 2004;45:ARVO E-Abstract 820) To isolate and study OPs generated from either the rod or the cone systems, we used mouse models that possess pure rod and cone functions, as indicated by ERGs elicited with visible light. 

Material and Methods

Animals

Wild-type C57BL/6 mice and cone photoreceptor function loss (cpfl1) ¹² mice were obtained from The Jackson Laboratory (Bar Harbor, ME). The original functional pure cone ^{13 14} rhodopsin knockout mice (rho ^{− / −}) were generously provided by Peter Humphries (Trinity College of Dublin, Ireland). Both cpfl1 and rho ^{− / −} mice have the same genetic background (C57BL/6). Mice were bred and housed under a 12-hour light–dark cycle, with free access to food and water. Because cone functions of the rho ^{− / −} mouse start to deteriorate after 49 days, ^{13 14} all mice used in this study were ∼35 to 42 days old. Extensive ERG tests have confirmed that cpfl1 mice have normal rod functions but lack cone functions. The study was conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and was approved by the Animal Care and Use Committee of the University of Missouri-Columbia. 

Electroretinogram Recording

Bilateral ERGs were recorded by using a procedure similar to that used in previous studies. ^{15 16 17} Mice were dark adapted overnight and anesthetized with a mixture of ketamine (75 mg/kg intramuscularly) and xylazine (13.6 mg/kg intramuscularly). Pupils were dilated with 1% tropicamide. A gold wire loop electrode was placed on the surface of the cornea. A differential electrode was placed under the skin on the forehead, and a neutral electrode was inserted subcutaneously near the tail. Electrical signals were amplified using an amplifier with 10⁴ gain and a bandwidth of 0.1 to 1 kHz (−3 dB points). Signals were digitized at a 5.12-kHz rate using a data-acquisition board (National Instrument, Austin, TX). ERG signals were averaged three to six times to reduce noise. Mice were placed on a heating pad to keep the body temperature at 38°C. 

A white light xenon visual stimulator (Grass PS22; Grass-Telefactor, Inc., West Warwick, RI) was used to illuminate the Ganzfeld. The light flash had a duration of 10 μs. Light luminance was calibrated with an integrating radiometer-photometer (model IL-1700; International Light, Newburyport, MA). The system can achieve a maximum intensity of 0.65 log cd-sec/m². Neutral density filters (Kodak, Rochester, NY) were used to attenuate light intensities over a 6-log range. An electronic timer (Uniblitz, Rochester, NY) was used to control the flash sequence. For dark-adapted ERG recording, the interstimulus interval (ISI) was at least 12 and 30 seconds for low intensities and high intensities, respectively. For light-adapted ERG recording, a background light of 30 cd/m² was applied to suppress rod responses. 

OP Analysis

ERG traces were analyzed off-line using custom-compiled programs (developed in LabView 7; National Instrument, Austin, TX). The OPs were extracted digitally by using a fifth-order Butterworth filter in a manner similar to that described by other groups. ^{7 18 19 20 21} The frequency spectra of extracted OPs were then analyzed by using a fast Fourier transform (FFT). The filter design and FFT operation were implemented with the software program. 

A correct filter bandwidth is important for extracting clean OPs from ERGs. Figure 1shows the extracted OP signals at 30 to 300 and 65 to 300 Hz bandwidths. Spectrum analysis (Fig. 1B)revealed that the OPs from the two pass-band settings were similar at high frequencies (>50 Hz), whereas a large difference existed at frequencies smaller than 50 Hz. When the high-pass frequency was set at 30 Hz, OPs were contaminated by the residual low frequency ERG a- and b-waves. To eliminate the a- and b-wave interference and to avoid the 60-Hz line noise, we chose a pass-band of 65 to 300 Hz. 

Different filter orders were found to have little effect on frequency components at frequencies of 50 Hz and higher; whereas the low-frequency components were gradually removed by using higher-order filters (Fig. 2) . Filter order is a standard parameter in filter design. It primarily determines the width of the transition band: the higher the order, the narrower is the transition between the pass-band and stop band, giving a sharper cutoff in the frequency response. The a- and b-wave contaminations were not completely eliminated by using a Butterworth filter at the fourth order or lower. The residual a- and b-waves distorted the OP signals significantly. A filter at the fifth order provided a good compromise between high stop-band attenuation and minimal phase shift. The amplitude and phase-delay responses of this filter are shown in Figure 3 . The phase delay that was introduced by filtering was frequency dependent. However, the variation of phase delays was small (in the frequency range of 60∼200 Hz) and did not affect our OP analysis results. 

To evaluate OP amplitudes objectively and efficiently, we calculated the areas enclosed by the single-sided frequency spectrum curve, to assess the total power of OPs. This method is analogous to that is used in the calculation of mechanical vibration energy. 

Results

Conventional ERG

In the wild-type C57BL/6 mice, the dark- and light-adapted ERG waveforms and amplitudes measured in our studies were comparable to those obtained in previous studies over a wide range of intensities (Figs. 4 5) . ^{17 22} At the highest stimulus intensity used in this study (0.65 log cd-s/m²), the averaged amplitudes of the dark-adapted a-wave, dark-adapted b-wave, and light-adapted b-wave were 204 ± 31, 356 ± 68, and 106 ± 18 μV, respectively. These values are similar to those reported in previous studies. ^{16 17} The ERG waveforms and amplitudes confirmed that rod and cone functions were normal in C57BL/6 mice. 

The cpfl1 mouse bears a mutation (116-bp insertion) in cGMP-phosphodiesterase α subunit (PDE6C) gene of the cone photoreceptor. ¹² Although cones still physically exist at the age of 35 to 42 days, as demonstrated by PNA staining (data not shown), the aforementioned genetic defect causes failure of the cone phototransduction cascade. However, rod functions remain intact until at least 18 months of age. ERGs elicited with visible light were similar in cpfl1 mice (Figs. 4 5)and CNGA3 knockout mice. ²³ Specifically, no ERG response was recorded in cpfl1 mice at light-adaptation when rod functions were suppressed, indicating that the medium wavelength (M)-cone system is not functional. The dark-adapted ERG responses of cpfl1 mice were similar to those of C57BL/6 control mice at low to middle intensities (Fig. 5 ; intensity, ≤−1.35 log cd-s/m²). When the stimulus intensity was greater than the cone responses threshold (∼ −0.85 log cd-s/m²), the b-wave amplitude of the cpfl1 mice remained stable and formed a plateau (Fig. 5) , as in CNGA3 knockout mice. ¹⁴ However, the b-wave amplitude of the control C57BL/6 mice increased continuously at these intensities. Such different responses in dark-adapted ERG b-wave amplitude and in light-adapted ERG clearly indicate the lack of M-cone function in cpfl1 mice. At 0.65 log cd-s/m², the average b-wave amplitudes were 356 and 261 μV in C57BL/6 and cpfl1 mice, respectively. The contributions of a normal cone system to the dark-adapted ERG b-wave amplitude were approximately one fourth to one third of the contributions from rods. This finding is in agreement with previous studies. ²³  

Pure cone rho ^{− / −} mice and the light-adapted wild-type mice had similar ERGs. Although the raw ERG waveforms showed some variations, their a-waves are small and the b-wave amplitudes were similar. Consistent with previous reports, the dark-adapted responses at high intensities were cone driven, and their amplitudes were similar to those of the light-adapted b-wave recorded at the same stimulus intensities. ^{13 14} Cone system functions appeared normal in rho ^{− / −} mice as evidenced by their light-adapted ERG b-wave amplitudes being comparable to those of the control mice (Fig. 4) . 

Latency of Rod- and Cone-Driven OPs

Figure 6Ashows the representative OPs extracted from the ERGs displayed in Figure 4 . The OP waveforms were similar to those previously obtained using digital filtering. ^{7 24 25} At the highest intensity of 0.65 cd-s/m² the mixed dark-adapted ERG consisted of inputs from both rod- and cone-pathways. ¹⁰ Therefore, the OPs extracted at this intensity level contained contributions from both rod and cone pathways. In light-adapted conditions, the cone OPs were extracted from ERGs recorded at the highest intensity. In time-domain, the latencies (the time between the stimulation onset and the peak of the individual OP with the highest absolute amplitude) of the light-adapted cone OPs in C57BL/6 mice were 57.60 ± 5.32 ms (mean ± SD, n = 6) which is approximately 13 ms longer than those of the dark-adapted OPs (44.56 ± 4.32 ms, n = 6) measured at the same stimulation intensity (P < 0.001, t-test). 

The dark-adapted OP latency of the pure rod cpfl1 mice (44.29 ± 4.20 ms, n = 6) were similar to those of the C57BL/6 mice (P > 0.05). No OPs were found in the light-adapted ERG in cpfl1 mice, because these mice lack M-cone functions. In the pure cone rho ^{− / −} mice, cone-driven OPs with similar latencies were identified at the highest intensity under both dark- and light-adapted conditions. In rho ^{− / −} mice the latencies of dark- and light-adapted OPs were (55.92 ± 7.06 ms, n = 6) and (56.08 ± 2.75 ms, n = 6) respectively, which were approximately 12 ms longer than those of rod-driven OPs in cpfl1 mice (P < 0.01, P < 0.001), but similar to those of cone-driven OPs in wild-type mice (P > 0.05, P > 0.05). These data provided evidence that the rod-driven OPs are faster than the cone-driven in the mouse. 

In Figure 6A , the pure cone rho ^{− / −} mouse has cone-driven OPs in dark-adapted state and the pure rod cpfl1 mouse show rod-driven OPs. These findings suggest that both rod- and cone-driven OPs obtained with high-intensity stimuli under dark-adaptation are integrated parts of the scotopic OPs in a normal mouse. In addition, the OP spectra in Figure 7demonstrate that cone-driven OPs appear in OPs of C57BL/6 with high-intensity stimuli under dark adaptation. 

Frequency Spectrum of the OPs

Figure 7shows OP frequency power spectra extracted from dark- and light-adapted ERGs in C57BL/6, cfpl1 and rho ^{− / −} mice. The light intensity was 0.65 log cd-s/m². In wild-type mice, the peak frequency of the light-adapted, cone-driven OPs was approximately 75 Hz. However, two peak frequencies were identified in the dark-adapted OP spectra. The subtle low-frequency peak exhibited smaller amplitude and was coincident with the cone-driven OPs extracted from light-adapted ERGs, suggesting it was related with the cone system. Because cone- and rod-driven OPs frequency overlap at this frequency region, the first peak should be a combination of both rod and cone system activities. Therefore, the difference between the light- and dark-adapted curves reflects the contributions from the rod pathway. The other peak had a higher frequency of 110 Hz and larger amplitude. Because rod-driven OPs were the major component in the dark-adapted ERG, the dominant frequency component at 110 Hz must be generated by the rod system. Actually, it is almost identical with that of the pure rod mouse. The frequency spectrum of the cone-driven OPs in dark-adapted ERG, with a peak frequency at 75 Hz, may be engulfed in the huge rod-driven OP spectrum. 

The frequency properties of rod- and cone-driven OPs obtained from C57BL/6 mice were further confirmed by measurements in cfpl1 and rho ^{− / −} mice. In pure rod function cfpl1 mice, the peak frequency of the power spectrum was 110 Hz under a 0.65 log cd-s/m² stimulation, similar to the high-frequency component in C57BL/6 mouse. In contrast, the power spectrum for light-adapted ERG was absent. In rho ^{− / −} mice with pure cone function, both the light- and dark-adapted ERGs showed small OP power spectra with the peak frequencies ∼75 Hz. The data demonstrated that the peak frequencies of rod- and cone-driven OPs are different. With a given light intensity which is bright enough to activate rod and cone pathways, the peak frequency of rod-driven OPs is approximately 30% higher than that of cone-driven OPs. Although the peak frequencies changed slightly with the light intensity (Fig. 8A) , rod-driven OPs had consistently higher peak frequencies than did cone-driven OPs. 

Total Amplitude of the OPs

The total OP power (or energy, an alternative measurement of OP amplitude) was quantified by calculating the area enclosed by the frequency power spectrum (Fig. 7) . Figure 8Bshows that the total OP energy increased with the stimulation intensity. The amplitudes of rod-driven OPs were significantly stronger than those of cone-driven OPs. In C57BL/6 mice, the total OP energy (including both rod- and cone-driven OPs) was 418.7 μV² · s at the highest stimulation intensity, but the power of the cone-driven OPs was only 21.5 μV² · s (Fig. 8B) . The energy of cone-driven OPs was approximately 5% of the total OP energy. In cpfl1 mice, the energy of the OPs was 328.2 μV² · s, which was comparable to the total OP energy of the C57BL/6 mice. In rho ^{− / −} mice, the energy of the cone-driven OPs was only 5.93 μV² · s, which is approximately 1.5% of the total energy of the C57BL/6 mice. These results suggest that in the rod-dominant mouse, OPs are mainly rod-driven and the contribution from the cone-driven OPs is minimal. 

Discussion

To explore fully the feasibility of OPs as a measure for clinical diagnosis and research in retinal diseases, an effective and reliable method for extracting OPs from ERGs is essential. Methods for time domain signal analysis have been designed ²⁶ to extract OPs by using a well-established ERG a-wave model, ²⁷ as well as other signal-conditioning methods. ²⁸ Digital filtering by means of a pass-band filter ^{7 18 19 20 21} is another frequently used method because of its reliability and simplicity of implementation. However, an inappropriately designed filter may either introduce contamination from a- and b-waves or corrupt the true OPs. Compared with OP signals, the ERG a- and b-waves consist of lower frequency components (20–50 Hz). If not eliminated, a- and b-waves appear as the initial big negative wavelet in the extracted OPs. ^{26 28} Such contaminations distort the early OPs (especially OP₁) and may lead to misinterpretation of the OPs. The bandwidth and orders of the digital filter are important parameters. We found that the high cutoff frequency of 300 Hz recommended in the ISCEV protocol is enough to preserve OP components. However, the low cutoff frequency of 75 to 100 Hz may be inappropriate for mice. The fifth-order Butterworth filter, ^{18 21 24} with pass-band of 65 to 300 Hz, was effective in removing the a- and b-wave contaminations and preserving all essential OP components. 

Using the digital filtering technique, we studied the different characteristics of rod- and cone-driven OPs. We found that cone- and rod-driven OPs have distinctly different peak frequencies of 70 to 90 Hz and 100 to 120 Hz, respectively. These results were obtained by measuring the timing between two adjacent individual OPs and analyzing the frequency spectra. Further studies are necessary to understand the mechanisms that cause this discrepancy. Such a difference may have practical applications. Because dark-adapted OPs contain contributions from both rod and cone systems, signal-processing techniques could be developed to separate these two contributions based on their frequency difference. Alternatively, a filtering system could be designed for specific extraction of rod- or cone-driven OPs. 

Another important difference between rod- and cone-driven OPs is that they have different time latencies. Contrary to previous results, we found that the cone-driven OPs in mice had longer latencies (∼13 ms) than the rod-driven OPs. Early studies on human subjects assumed that OPs with short latencies (OP₂ and OP₃) probably signal the activation of the cone pathway, and OPs with long latencies (OP₄ and OP₅) result from the activation of the rod system. ^{29 30 31} However, OP measurements in dark-adapting human subjects show that every individual OP could be generated postreceptorally through either a rod or a cone pathway, ³¹ implying that it is impossible to assign a class of retinal photoreceptors to the genesis of one or more OPs. Because both rod- and cone-driven OPs are superimposed in the composite scotopic OPs, our results suggest that short-latency OPs are rod-driven, whereas long-latency OPs contain both rod- and cone-driven signals. These findings are supported by other evidence. First, scotopic OP₄ has been shown to be abolished in human patients with cone-system abnormalities, ³⁰ indicating that the OP₄ is associated with cone functions. Second, Hancock and Kraft ⁷ as well as Dong et al. ³ found that the long latency components of the extracted OPs had lower frequencies than did the earlier ones. ^{3 7} Data from our study demonstrate that cone-driven OPs have lower frequencies than rod-driven OPs. Third, early OPs (OP₁ and OP₂) have been found to react more to the scotopic background light, whereas late OPs (OP₃ and OP₄) have been shown to be more affected by the relatively brighter mesopic conditions. ³² Together, this evidence supports our conclusion that short-latency OPs are mainly rod-driven and long-latency OPs contain input from both rod and cone systems. 

OP amplitude is an important parameter in assessing the inner retina function in healthy and diseased retina. ^{2 7} We have developed a reliable and efficient method to study the OP amplitude by calculating the total power of OPs in the frequency domain. With the traditional OP measurement method, ¹⁰ the amplitude of each individual OP can be easily influenced by different recording and filtering conditions. However, the total energy of the OPs (after removing the a- and b-wave contaminations) is relatively consistent, regardless of the OP phase shift and individual OP amplitude change induced by different filter settings (Fig. 2) . Using this method, we found that rod-driven OPs are dominant in mice; whereas cone-driven OPs only account for less than 5% of the total OP signal. 

 Supported in part by a grant from the Research Board of the University of Missouri-Columbia, NIH/NEI core Grant EY014795, and an unrestricted grant from Research to Prevent Blindness to the Department of Ophthalmology, University of Missouri-Columbia.

 Submitted for publication November 15, 2005; revised January 17, 2006; accepted March 13, 2006.

 Disclosure: B. Lei, None; G. Yao, None; K. Zhang, None; K. J. Hofeldt, None; B. Chang, 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.

 

The effects of different digital filter pass-band selections on OPs extracted from dark-adapted ERGs of wild-type mice. (A) The ERG a-wave contaminated the extracted OPs (arrow) when the pass-band was 30 to 300 Hz. When the band-pass was 65 to 30 Hz, the a-wave contaminations were greatly decreased. (B) The corresponding frequency spectrum of the OPs in (A). There were two peaks in the spectrum: a higher peak frequency ∼100 Hz, which represented the OPs and a low peak frequency ∼25 Hz, which represented a- and b- wave (mainly a-wave) contaminations. Use of the 65 to 300 Hz filter significantly removed the low-frequency component (a-wave), whereas the higher frequency OPs were not significantly affected (filter order was one).

Figure 2.

 

The effects of different filter order settings on OPs in the mouse. (A) OPs extracted from the dark-adapted ERG using a series of filter orders. When filters of lower order were used, there were substantial a-wave contributions (arrow) to the OPs. The a-wave contamination was removed with high order filers. (B) Frequency spectrum of the OPs in (A). Higher-order filters (>4) removed the lower frequency a-wave contaminations, but they did not affect the OP signals that had higher frequencies of ∼110 Hz (filter pass-band, 65–300 Hz).

Figure 3.

 

The amplitude and phase delay responses of the fifth-order digital Butterworth filter. The phase delay introduced by filtering was frequency dependent. However, the variation of phase delay was small in the frequency range of 60 to 200 Hz and did not affect the results of OP analysis results.

Figure 4.

 

Representative dark- and light-adapted ERGs of 6-week-old C57BL/6, cpfl1, and rho ^{− / −} mice. The C57BL/6 mice exhibited normal ERGs. The cpfl1 mice had subnormal rod-driven ERGs and extinguished cone system responses. Rho ^{− / −} mice had normal cone ERGs, but no rod-driven responses. The OPs superimposed on the ascending limb of the ERG b-wave can be clearly recognized. The numbers in the graphs indicate the time (ms) between two consecutive OPs. The results indicate that rod-driven OPs have higher frequencies than cone-driven OPs (stimulation light intensity, 0.65 log cd-s/m²).

Figure 5.

 

ERG response versus intensity curve (V-log I curve) of the dark-adapted ERG a- and b-waves. The a-wave responses, which represent mainly the rod functions, were similar in the two strains. The b-wave responses, which represent the bipolar cell responses, were similar at low stimulus intensities (lower than −1.85 log cd-s/m²) where only the rod system can be activated. At higher intensities, the b-wave amplitude of cpfl1 mice did not increase and formed a plateau, whereas that of C57BL/6 mice continued to increase. At the highest intensity, the b-wave of C57BL/6 mice was approximately 25% higher than that of cpfl1 mice. This difference is attributable to the cone system.

Figure 6.

 

(A) OPs extracted from dark- and light-adapted ERGs of 6-week-old C57BL/6 mice, cpfl1, and rho ^{− / −} mice, by using a digital Butterworth filter (bandwidth, 65–300 Hz, fifth order). The rod-driven OPs had much higher amplitudes and shorter latencies than the cone-driven OPs. (B) OP latency measurements in (left) dark- and (right) light- adapted OPs from three mouse strains. The latencies of cone-driven OPs were approximately 13 ms slower than those of the rod-driven OPs (stimulation light intensity, 0.65 log cd-s/m²).

Figure 7.

 

The frequency spectra of the OPs extracted from dark- and light-adapted ERGs in 6-week-old C57BL/6, cpfl1, and rho ^{− / −} mice. The dark-adapted OPs of the C57BL/6 mice had a major peak at ∼110 Hz and a smaller bump at 75 Hz. Coincidently , the cone system light-adapted OPs presented a small single peak at 75 Hz but the major 110-Hz component was missing. In the cpfl1 mice the dark-adapted OPs showed a single peak at 110 Hz, but no signal was observed in light-adapted measurements. In rho ^{− / −} mice both the dark- and light-adapted OPs showed similar peak frequency ∼75 Hz. The higher frequency component in C57BL/6 and cpfl1 mice was missing. These results indicate that in the mouse the rod-driven OPs have a higher frequency ∼110 Hz and the cone-driven OPs have a lower frequency ∼75 Hz.

Figure 8.

 

(A) The effects of stimuli intensities on the frequencies of the dark- and light-adapted OPs in 6-week-old C57BL/6, cpfl1, and rho ^{− / −} mice (six animals in each group). The peak frequency of the OP changed with the intensity. However, the rod-driven OPs always showed a higher frequency than did the cone-driven OPs. (B) Curves showing total OP energy (amplitude) versus intensity in the three mouse strains. The rod- and cone-mixed OPs in C57BL/6 mice and the rod-driven OPs in cpfl1 mice had similar amplitude. The cone-driven OPs in dark-adapted rho ^{− / −} mice and in light-adapted C57BL/6 and rho ^{− / −} mice showed similar amplitude. In rod-dominant mice, the amplitude of rod-driven OPs accounted for more than 95% of the total OP energy. DA, dark adaptation; LA, light adaptation Bars, SE.

The authors thank Yuanfang Gao, Daniel D. Green, and Kechen Zhang for help in compiling programs and the anonymous reviewers for their valuable suggestions and comments.