Open Access
Visual Neuroscience  |   October 2017
Mesopic and Photopic Rod and Cone Photoreceptor-Driven Visual Processes in Mice With Long-Wavelength–Shifted Cone Pigments
Author Affiliations & Notes
  • Tina I. Tsai
    Department of Ophthalmology, University Hospital Erlangen, Erlangen, Germany
    Department of Biology, Animal Physiology, FAU Erlangen-Nürnberg, Erlangen, Germany
  • Anneka Joachimsthaler
    Department of Ophthalmology, University Hospital Erlangen, Erlangen, Germany
    Department of Biology, Animal Physiology, FAU Erlangen-Nürnberg, Erlangen, Germany
  • Jan Kremers
    Department of Ophthalmology, University Hospital Erlangen, Erlangen, Germany
    Department of Anatomy II, University of Erlangen-Nürnberg, Germany
    School of Optometry and Vision Science, University of Bradford, Bradford, United Kingdom
  • Correspondence: Jan Kremers, Department of Ophthalmology, University Hospital Erlangen, Schwabachanlage 6, 91054 Erlangen, Germany; Jan.Kremers@uk-erlangen.de
Investigative Ophthalmology & Visual Science October 2017, Vol.58, 5177-5187. doi:10.1167/iovs.17-22553
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Tina I. Tsai, Anneka Joachimsthaler, Jan Kremers; Mesopic and Photopic Rod and Cone Photoreceptor-Driven Visual Processes in Mice With Long-Wavelength–Shifted Cone Pigments. Invest. Ophthalmol. Vis. Sci. 2017;58(12):5177-5187. doi: 10.1167/iovs.17-22553.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: The clearer divergence in spectral sensitivity between native rod and human L-cone (L*-cone) opsins in the transgenic Opn1lwLIAIS mouse (LIAIS) allows normal visual processes mediated by these photoreceptor subtypes to be isolated effectively using the silent substitution technique. The objective of this study was to further characterize the influence of mean luminance and temporal frequency on the functional properties of signals originating in each photoreceptor separately and independently of adaptation state in LIAIS mice.

Methods: Electroretinographic (ERG) recordings to sine-wave rod and L*-cone modulation at different mean luminances (0.1–130.0 cd/m2) and temporal frequencies (6–26 Hz) were examined in anesthetized LIAIS (N = 17) and C57Bl/6 mice (N = 8).

Results: We report maximum rod-driven response with 8-Hz modulation at 0.1 to 0.5 cd/m2, which was almost four times larger than maximum cone-driven response at 8 Hz, 21.5 to 130 cd/m2. Over these optimal luminances, both rod- and cone-driven response amplitudes exhibited low-pass functions with similar frequency resolution limits, albeit their distinct luminance sensitivities. There were, however, two distinguishing features: (1) the frequency-dependent amplitude decrease of rod-driven responses was more profound, and (2) linear relationships describing rod-driven response phases as a function of stimulus frequency were steeper.

Conclusions: Employing the silent substitution method with stimuli of appropriate luminance on the LIAIS mouse (as on human observers) increases the specificity, robustness, and scope to which photoreceptor-driven responses can be reliably assayed compared to the standard photoreceptor isolation methods.

For many electroretinographic (ERG) studies in mice, the mainstay animal model for studying normal human visual processing and disease-related pathophysiology, the separation of signal processes driven by the cone and rod photoreceptors is of great importance. This may be to better our understanding of the panoply of visual pathways that serve the visual system. Furthermore, retinal diseases that primarily affect rods or cones can be studied more easily if their signals can be separated. Generally, this separation is achieved on the basis of the mean luminance adaptation (where rods and cones are more active under scotopic and photopic conditions, respectively),1 or in mice, in which one of the two is a genetic knock-out (KO) (e.g., Cpfl−/−, Rpe65−/−, and Rho−/− mice)25 or modified to be nonfunctional (e.g., Gnat2cpfl3 and Gnat1−/− mice).6,7 The latter method has the disadvantage that the animals cannot be considered as physiologically normal. Therefore, the measured ERG signals from KO mice should be considered with some caution. For instance, Cpfl−/− and Rpe65−/− mice, both thought to be devoid of cones, still show functional activity at photopic conditions or have rod signals that assume cone-like properties,8 respectively. Similarly, rodless Rho−/− mice are considered to have normal functioning cones only between age 4 and 6 weeks.9 Furthermore, even if the photoreceptors themselves are functionally normal, the inherent interactions between rods and cones (i.e., via mutual gap junctions)10 detectable in the mouse ERG of ex vivo eye preparations are missing.11 With the former method, a direct comparison of rod- and cone-driven signals is not possible because the retinae are in different states of adaptation, which also influences the mode of operation of postreceptoral circuitries. A direct comparison would be possible if the separation could be obtained using the silent substitution method that is based on differences in the absorption spectra of the rods and cones (reviewed in Ref. 12). Because this method allows for independent modulation of each photoreceptor type around the mean luminance and chromaticity, factors such as temporal frequency, waveform, and contrast can be studied independently. 
Unfortunately, in the wild-type (WT) mouse, the spectral separation of rod and M-cone absorption spectra is only a few nanometers,13,14 thereby limiting the modulation of excitation in the isolated photoreceptor type. A solution to this problem comes in the form of mice strains, in which the native M-cone pigments are replaced by a variant of the human L-cone (L*-cone) pigment.1517 The appeal of the these strains lies in the increased spectral separation between rods (absorption spectrum maximum at 498 nm) and the substitute expression of medium- to long-wavelength L*-cones (peak absorption at 561 nm) in place of the native mouse M-cone opsin (508 nm). Otherwise, these strains are structurally and physiologically identical to WT animals. Because of this, each photoreceptor subtype can be individually modulated at a greater depth (i.e., to obtain larger responses) and over a greater dynamic range under identical retinal states of adaptation using silent substitution ERG paradigms12 to allow a more comprehensive study of their physiological properties. 
A handful of studies to date have demonstrated the functional advantage of using mice possessing long-wavelength–shifted cone opsins (Opn1lwLIAIS [LIAIS] or a related strain Opn1mwR)15,16 over the WT mouse for studying visual processes pertaining to specific photoreceptor-driven activity using the ERG.1820 We previously demonstrated the use of the silent substitution method for separating rod- and cone-driven signals in the LIAIS mouse.18 Human studies showed that rod activity can be obtained at frequencies (up to about 20 Hz) and retinal illuminances (up to 284 photopic candela per square meter [phot cd/m2]),2125 at which they were conventionally thought to be insensitive.26 It is therefore of importance to explore the range of luminances and frequencies at which rod- and cone-driven signals can be obtained reliably in the LIAIS mouse. In a previous study by Allen and Lucas19 on Opn1mwR mice, the influence of luminance was studied using flashed stimuli. They found that rod- and cone-driven responses have clearly separated luminance regions of maximal activity, with the change between rod- and cone-driven ERG responses occurring at about 104 isomerizations per rod per second (R*/rod/s). Much larger rod-driven compared to L(*)-cone–driven flash ERGs were also duly isolated.19 However, with flashes, a change in stimulus strength or flash frequency also changes the state of adaptation. Furthermore, flashes contain higher harmonics that also elicit ERG responses, thereby precluding a study of the temporal frequency on ERG signals. 
The aim of the present study was to expand the functional profiling of rod- and L*-cone–driven flicker ERG characteristics using sine-wave modulation of the excitation of single photoreceptor types at different temporal frequencies in the LIAIS mouse (as first performed in Tsai et al.18) and at adaptation conditions between 0.1 and 130 cd/m2 spanning three orders of magnitude in luminance levels. A frequency run from 6 Hz (instead of from 3 Hz) was adopted here, as it is generally high enough that the ERG waveforms will not be disrupted by the low-frequency respiratory movements of the animals18 (also evident in Fig. 1C). Moreover, we presume that the rod/L*-cone isolated responses to flicker stimuli above 4 Hz would contain minimal input and influence19,27 from the sluggish responses mediated by intrinsically photoreceptive retinal ganglion cells that contain melanopsin (ipRGCs),2832 which are also active at mesopic to photopic light levels.20,28,33 As opposed to stimulating either the rod/ipRGC or whole cone populations (i.e., cones expressing either and both S- and L-opsins) in Allen and Lucas's Opn1mwR mice study,19 the stimuli here were calibrated to modulate either the rod or L*-cone subtypes alone. As mentioned above, employing sinusoidal modulation instead of flashes allows the independent control of the state of adaptation and of temporal frequency on the magnitude (i.e., response amplitude) and time delay (i.e., response phase) of each response type to be studied under specified modes of operation in the retina. A further objective here was also to explore the narrow19,34,35 mesopic illumination range in detail to find the point at which processes underlying the flicker ERG transitions from being rod dominated to L*-cone dominated. Lastly, we aimed to propose optimal rod- and L*-cone–isolation conditions for mouse flicker ERG recordings and compare them to available data for humans.2123,36 
Figure 1
 
Representative flicker ERG waveforms evoked by mesopic to photopic (A) rod- and (B) M/L*-cone–isolating stimuli (photopic candela per square meter; see Table 1 for conversions to scotopic candela per square meter and response per rod per second) from LIAIS (black traces) and WT mice variants (gray traces; at the optimum luminances for rod and cone isolation only). Displayed are identical 500-ms episodes of responses (i.e., averages of left and right eyes of one animal and the two 500-ms episodes that make up the 1-second recording epoch) to sine-wave stimuli of 6 to 26 Hz (indicated at far left) and at the highest attainable contrast (i.e., 55% cone and 75% rod contrasts for LIAIS, 5% rod and cone contrasts for WT animals; see Materials and Methods). (C) Amplitude plots of 8 Hz rod (top panels) and L*-cone ERGs (bottom panels) after FFT (indicated by the arrow), as presented in (A) and (B) to demonstrate the respective changes in their signal-to-noise ratio with mean luminance.
Figure 1
 
Representative flicker ERG waveforms evoked by mesopic to photopic (A) rod- and (B) M/L*-cone–isolating stimuli (photopic candela per square meter; see Table 1 for conversions to scotopic candela per square meter and response per rod per second) from LIAIS (black traces) and WT mice variants (gray traces; at the optimum luminances for rod and cone isolation only). Displayed are identical 500-ms episodes of responses (i.e., averages of left and right eyes of one animal and the two 500-ms episodes that make up the 1-second recording epoch) to sine-wave stimuli of 6 to 26 Hz (indicated at far left) and at the highest attainable contrast (i.e., 55% cone and 75% rod contrasts for LIAIS, 5% rod and cone contrasts for WT animals; see Materials and Methods). (C) Amplitude plots of 8 Hz rod (top panels) and L*-cone ERGs (bottom panels) after FFT (indicated by the arrow), as presented in (A) and (B) to demonstrate the respective changes in their signal-to-noise ratio with mean luminance.
Materials and Methods
Ethical Approval
All protocols and methods described in this study were approved by the local animal welfare authorities (Regierungspräsidium Mittelfranken, Ansbach, Germany) and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the Society for Neuroscience. 
Experimental Animals
ERG measurements were recorded from 25 mice: 17 LIAIS (13 hemizygous male, 4 homozygous female, 9.5–13 months) and 8 WT littermates of C57Bl/6J origin (male, 10–12 months). The LIAIS mice were created in embryonic stem cells from 129/SJ mice to generate the targeted replacement. The initial targeted replacement mice were mated to a cre line made on the C57Bl background (unspecified if 6J or 6N strains; Ozgene, Perth, Australia). LIAIS refers to the transgenic Opn1lwLIAIS strain originally described by Greenwald et al.15 and later by Tsai et al.,18 which were sourced from the University of Washington (Seattle, WA, USA). Homozygous females and hemizygous males of this strain feature a common L*-cone opsin (with the amino acids leucine, isoleucine, alanine, isoleucine, and serine located at positions 153, 171, 174, 178, and 180, respectively, hence termed “LIAIS” and hereafter denoted as L*-cone) instead of the endogenous mouse M-cone opsin in WT littermates with a normal C57BL/6J background. All animals otherwise expressed normal mouse rhodopsin and S-opsin. This opsin substitution has been shown to induce only a spectral sensitivity shift of the cones toward longer wavelengths, with a maximum sensitivity at about 561 nm instead of 508 nm,13,14,37,38 and not influence cone structure nor function.15,18 
Electroretinography
Mouse flicker ERG procedures were the same as those described previously.18 Animals kept in a 12-hour dark-light cycle were dark adapted overnight before the experiment session. With the aid of a dim red light, anesthesia was induced by an intramuscular injection of 50:10 mg/kg ketamine/xylazine (Ketavet; Pfizer, Karlsruhe, Germany; Rompun 2%; Bayer AG, Leverkusen, Germany). A subcutaneous injection of saline solution (10 mL/kg, 0.9%) protected them from dehydration while under sedation. Topical mydriatics (1 drop [gtt] tropicamide; Mydriaticum Stulln, 5 mg/mL, Pharma Stulln, Stulln, Germany, and 1 gtt phenylephrin-hydrochloride; Neosynephrin POS 5%, Ursapharm, Saarbrücken, Germany) were given to dilate the pupils (to approximately 3-mm diameter). ERG electrodes consisted of ground and reference needles (subcutaneously at the base of tail and on forehead medial to eyes, respectively) and binocular corneal-active contact lenses (Mayo Corp., Inazawa, Japan) internally covered with Corneregel (Dr. Mann Pharma, Berlin, Germany). These were arranged while the mouse was lying on a sliding, heated platform that faced the Ganzfeld bowl (Q450SC; Roland Consult GmbH, Brandenburg, Germany). Throughout the recording sessions, core body temperature was maintained. Each mouse underwent up to five recording sessions (see protocol below), with each session limited to an hour and two sessions at least 1 week apart. Once the study was completed, the animals were sedated with isoflurane and humanely euthanized. 
Silent Substitution Stimuli
All stimuli employed were generated on the Ganzfeld stimulator controlled using the manufacturer-scripted RetiPort system (Roland Consult Q450SC). Visual stimulation constituted double silent substitutions used in Tsai et al.,18 whereby ERG responses from rod, native mouse M-cone, or L*-cone were obtained in isolation using sine-wave temporal modulations of three light-emitting diodes (LEDs; red: peak wavelength of 625 nm; green: 525 nm; and blue: 470 nm). A spectroradiometer (CAS 140; Instrument Systems, Munich, Germany) measured the spectral outputs of the LEDs, and their overall outputs were checked using a spot luminance meter (LS-110; Konica Minolta, Munich, Germany). Nomograms of Lamb39 were used to estimate the absorption spectra (using λmax: 355 nm for the S-cones,13 498 nm for the rods, 506 nm for the native M-cones, and 565 nm for L*-cones). These were corrected for preretinal absorption (except for the L*-cones, which were considered to be sensitive in a wavelength range where preretinal absorption has a negligible influence). The integral of the multiplication of the photoreceptor corrected absorption spectra (equivalent to the fundamentals in humans), and the emission spectra of the LEDs determined the excitation for each photoreceptor type by each of the LEDs.40 The total photoreceptor excitation thus equaled the sum of the photoreceptor excitations to each of the LEDs. Rod and cone contrasts (C) were calculated as Michelson contrast using the maximal and minimal excitations (Emax and Emin, respectively) during stimulation:  
\(\def\upalpha{\unicode[Times]{x3B1}}\)\(\def\upbeta{\unicode[Times]{x3B2}}\)\(\def\upgamma{\unicode[Times]{x3B3}}\)\(\def\updelta{\unicode[Times]{x3B4}}\)\(\def\upvarepsilon{\unicode[Times]{x3B5}}\)\(\def\upzeta{\unicode[Times]{x3B6}}\)\(\def\upeta{\unicode[Times]{x3B7}}\)\(\def\uptheta{\unicode[Times]{x3B8}}\)\(\def\upiota{\unicode[Times]{x3B9}}\)\(\def\upkappa{\unicode[Times]{x3BA}}\)\(\def\uplambda{\unicode[Times]{x3BB}}\)\(\def\upmu{\unicode[Times]{x3BC}}\)\(\def\upnu{\unicode[Times]{x3BD}}\)\(\def\upxi{\unicode[Times]{x3BE}}\)\(\def\upomicron{\unicode[Times]{x3BF}}\)\(\def\uppi{\unicode[Times]{x3C0}}\)\(\def\uprho{\unicode[Times]{x3C1}}\)\(\def\upsigma{\unicode[Times]{x3C3}}\)\(\def\uptau{\unicode[Times]{x3C4}}\)\(\def\upupsilon{\unicode[Times]{x3C5}}\)\(\def\upphi{\unicode[Times]{x3C6}}\)\(\def\upchi{\unicode[Times]{x3C7}}\)\(\def\uppsy{\unicode[Times]{x3C8}}\)\(\def\upomega{\unicode[Times]{x3C9}}\)\(\def\bialpha{\boldsymbol{\alpha}}\)\(\def\bibeta{\boldsymbol{\beta}}\)\(\def\bigamma{\boldsymbol{\gamma}}\)\(\def\bidelta{\boldsymbol{\delta}}\)\(\def\bivarepsilon{\boldsymbol{\varepsilon}}\)\(\def\bizeta{\boldsymbol{\zeta}}\)\(\def\bieta{\boldsymbol{\eta}}\)\(\def\bitheta{\boldsymbol{\theta}}\)\(\def\biiota{\boldsymbol{\iota}}\)\(\def\bikappa{\boldsymbol{\kappa}}\)\(\def\bilambda{\boldsymbol{\lambda}}\)\(\def\bimu{\boldsymbol{\mu}}\)\(\def\binu{\boldsymbol{\nu}}\)\(\def\bixi{\boldsymbol{\xi}}\)\(\def\biomicron{\boldsymbol{\micron}}\)\(\def\bipi{\boldsymbol{\pi}}\)\(\def\birho{\boldsymbol{\rho}}\)\(\def\bisigma{\boldsymbol{\sigma}}\)\(\def\bitau{\boldsymbol{\tau}}\)\(\def\biupsilon{\boldsymbol{\upsilon}}\)\(\def\biphi{\boldsymbol{\phi}}\)\(\def\bichi{\boldsymbol{\chi}}\)\(\def\bipsy{\boldsymbol{\psy}}\)\(\def\biomega{\boldsymbol{\omega}}\)\(\def\bupalpha{\unicode[Times]{x1D6C2}}\)\(\def\bupbeta{\unicode[Times]{x1D6C3}}\)\(\def\bupgamma{\unicode[Times]{x1D6C4}}\)\(\def\bupdelta{\unicode[Times]{x1D6C5}}\)\(\def\bupepsilon{\unicode[Times]{x1D6C6}}\)\(\def\bupvarepsilon{\unicode[Times]{x1D6DC}}\)\(\def\bupzeta{\unicode[Times]{x1D6C7}}\)\(\def\bupeta{\unicode[Times]{x1D6C8}}\)\(\def\buptheta{\unicode[Times]{x1D6C9}}\)\(\def\bupiota{\unicode[Times]{x1D6CA}}\)\(\def\bupkappa{\unicode[Times]{x1D6CB}}\)\(\def\buplambda{\unicode[Times]{x1D6CC}}\)\(\def\bupmu{\unicode[Times]{x1D6CD}}\)\(\def\bupnu{\unicode[Times]{x1D6CE}}\)\(\def\bupxi{\unicode[Times]{x1D6CF}}\)\(\def\bupomicron{\unicode[Times]{x1D6D0}}\)\(\def\buppi{\unicode[Times]{x1D6D1}}\)\(\def\buprho{\unicode[Times]{x1D6D2}}\)\(\def\bupsigma{\unicode[Times]{x1D6D4}}\)\(\def\buptau{\unicode[Times]{x1D6D5}}\)\(\def\bupupsilon{\unicode[Times]{x1D6D6}}\)\(\def\bupphi{\unicode[Times]{x1D6D7}}\)\(\def\bupchi{\unicode[Times]{x1D6D8}}\)\(\def\buppsy{\unicode[Times]{x1D6D9}}\)\(\def\bupomega{\unicode[Times]{x1D6DA}}\)\(\def\bupvartheta{\unicode[Times]{x1D6DD}}\)\(\def\bGamma{\bf{\Gamma}}\)\(\def\bDelta{\bf{\Delta}}\)\(\def\bTheta{\bf{\Theta}}\)\(\def\bLambda{\bf{\Lambda}}\)\(\def\bXi{\bf{\Xi}}\)\(\def\bPi{\bf{\Pi}}\)\(\def\bSigma{\bf{\Sigma}}\)\(\def\bUpsilon{\bf{\Upsilon}}\)\(\def\bPhi{\bf{\Phi}}\)\(\def\bPsi{\bf{\Psi}}\)\(\def\bOmega{\bf{\Omega}}\)\(\def\iGamma{\unicode[Times]{x1D6E4}}\)\(\def\iDelta{\unicode[Times]{x1D6E5}}\)\(\def\iTheta{\unicode[Times]{x1D6E9}}\)\(\def\iLambda{\unicode[Times]{x1D6EC}}\)\(\def\iXi{\unicode[Times]{x1D6EF}}\)\(\def\iPi{\unicode[Times]{x1D6F1}}\)\(\def\iSigma{\unicode[Times]{x1D6F4}}\)\(\def\iUpsilon{\unicode[Times]{x1D6F6}}\)\(\def\iPhi{\unicode[Times]{x1D6F7}}\)\(\def\iPsi{\unicode[Times]{x1D6F9}}\)\(\def\iOmega{\unicode[Times]{x1D6FA}}\)\(\def\biGamma{\unicode[Times]{x1D71E}}\)\(\def\biDelta{\unicode[Times]{x1D71F}}\)\(\def\biTheta{\unicode[Times]{x1D723}}\)\(\def\biLambda{\unicode[Times]{x1D726}}\)\(\def\biXi{\unicode[Times]{x1D729}}\)\(\def\biPi{\unicode[Times]{x1D72B}}\)\(\def\biSigma{\unicode[Times]{x1D72E}}\)\(\def\biUpsilon{\unicode[Times]{x1D730}}\)\(\def\biPhi{\unicode[Times]{x1D731}}\)\(\def\biPsi{\unicode[Times]{x1D733}}\)\(\def\biOmega{\unicode[Times]{x1D734}}\)\begin{equation}C = {{E_{ max}-E_{ min}}\over {E_{ max}+E_{ min}}}\end{equation}
 
See Table 1 in Tsai et al.18 for details on the LED and corresponding photoreceptor contrasts of the rod-, M-cone– and L*-cone–isolating stimuli for WT and LIAIS mice. Each LED set was modulated around a specific mean luminance given in photometric units, photopic candela per square meter, that are weighted with the human photopic spectral luminosity function, Vλ. Photopic candela per square meter is not a useful metric for quantifying luminances in absolute terms for mice because the human Vλ is likely to be of no relevance for the mouse. But it is meaningful for giving luminance in relative terms. We also give the luminance expressed in scotopic candela per square meter (scot cd/m2) and photoisomerizations per rod per second (R*/rod/s), which are also more meaningful for quantifying mean luminance for the rod system in mice because it is probably similar compared to the human scotopic spectral luminosity function, V′λ. Conversions from scotopic candela per square meter to photoisoerizations per rod per second are also provided in Table 1 to aid comparisons.41 Mean luminances are nevertheless given in photopic candela per square meter throughout the text, unless otherwise stated. 
Table 1
 
Generation of the Mean Stimulus Luminance Levels With the Aid of ND Filters and the Number of LIAIS and WT Animals (N) Recorded per Setting
Table 1
 
Generation of the Mean Stimulus Luminance Levels With the Aid of ND Filters and the Number of LIAIS and WT Animals (N) Recorded per Setting
The ratio of red to green to blue LED photopic luminances was always 6:6:1. Therefore, the mean chromaticity of the stimulus was the same in all conditions. At each mean luminance condition near maximal LED contrasts achievable were employed for a clean isolation (i.e., to achieve 75% rod and 55% L*-cone contrasts in LIAIS mice and 5% rod and M-cone contrasts in WT mice; for actual LED contrasts, see Tsai et al.,18 Table 1). Double silent substitution means that (1) for the rod-isolating conditions, contrasts in the M-/L*- and S-cone excitations were 0%, and (2) for M-/L*-cone–isolating conditions, contrasts in the rod and S-cone excitations were 0%. The calculations may not be completely correct for cones that double-express S- and M-/L*-opsins because the S-cone pigment and the L*- and particularly the M-pigments may mutually screen each other and thus alter their absorption spectra. However, the number of double-expressing cones is considered too small and the screening effects relatively too insignificant in the LIAIS mice to influence the data. Furthermore, the LED emission spectra hardly overlapped with the S-opsin absorption spectra. ERG data representing those evoked by M- or L*-cone stimulation are therefore referred to simply as those from WT and LIAIS cones, respectively, unless otherwise stated. 
Here, the rod- and cone-isolating ERG recordings performed at different frequencies (6, 8, 12, 14, 18, 22, and 26 Hz) over a range of mean luminance levels between 0.1 and 130 cd/m2 (see Table 1) are considered to cover mesopic to photopic conditions.42 In most conditions, two of the four LED arrays were covered, and neutral density (ND) filters of 1 or 2 optical density were placed in front of the remaining two LED arrays. These were held in position with custom-made, detachable cardboard frames to minimize variability in how the LEDs were masked after each ND filter change. Resulting photopic mean luminance levels were 0.1, 0.2, 0.5, 1.4, 7, 21.5, and 73 cd/m2 with these filters (Table 1). On up to five randomly chosen LIAIS and WT mice, rod responses were additionally recorded to a 39 cd/m2 setting, and cone responses were measured to 39 and 130 cd/m2 stimuli in order to compare the results with those obtained previously at the same luminances without filters.18 Given only three observations at the 130 cd/m2 setting, here however, only qualitative comparisons were made. Table 1 details the ND filter configurations and number of animals (N) recorded for the eight rod- and cone-isolating mean luminance conditions employed in this study. At the beginning of the experiments, the animals were dark adapted. As such, each mouse was adapted to each mean luminance level for 5 minutes before measurements commenced. At least 40 cycles, each lasting 1 second, were averaged for every stimulus frequency. 
Signal Analyses and Statistics
Electrical signals collected were amplified (100,000×), band-pass filtered (1–300 Hz), and digitized at a sampling frequency of 2048 Hz. Sweeps of 1-second epochs were averaged. Averaged ERGs from the two eyes were Fourier analyzed to extract amplitudes and phases of the signals' first harmonic (fundamental) component (i.e., at the stimulus frequency). The noise level was defined as the mean amplitude at frequencies ±1 Hz of the stimulus frequency. Response phases were only accepted if the signal-to-noise ratio (SNR) of their corresponding amplitudes were larger than two. Bach et al.43 recommend a SNR of 2.85 for the use of the responses. However, that estimate is exclusively based on amplitude data. Also taking phase data into account, we often noticed that reliable responses could also be obtained at lower SNRs. A 180° phase shift was applied to the cone phase results to take into account that they were modulated in counterphase relative to the rods (see Tsai et al.,18 Table 1). All amplitude values were included once signals from each eye were corrected for noise by subtraction of the mean noise. Very small amplitudes that became negative after noise correction were made zero. All signals shown are therefore an average of the left and right eyes of each animal, and where specified, the average of these for each group. 
Since all experimental data exhibited Gaussian distributions (Komogorov-Smimov normality test, P > 0.1 (Prism version 5.00; GraphPad Software, Inc., La Jolla, CA, USA), parametric methods (t-tests and repeated measures ANOVA) were employed for statistical comparisons. The level of significance during pairwise comparisons of parameters between multiple conditions and measures was adjusted using a Bonferroni correction. Phase relationships as a function of temporal frequency are typically linear, and thus their profiles were fitted with a linear regression (y = y0 + ax).25 The slope (a) described the phase change with frequency in degrees per Hertz at photoreceptor-isolating conditions. All regression fits returned a coefficient (r2) >0.80. The slopes were compared using 1-way ANOVA (i.e., set up with mean = a, SEM = SD of a, N = dF + 1) and subsequent post hoc tests. Comparing parameter y0 (i.e., y-intercept) of the different regressions revealed whether phases were shifted under the different conditions. 
Results
Comparison of Photoreceptor-Specific ERGs From LIAIS and WT Mice
Figure 1 displays samples of the isolated rod- (Fig. 1A) and cone-driven flicker ERG waveforms (Fig. 1B) recorded under the mesopic to photopic mean luminance conditions tested from representative LIAIS and WT mice. Signals to all seven temporal frequencies are shown for select retinal illuminances used. Amplitude plots of rod- and L*-cone 8-Hz data after fast Fourier transformation (FFT) are also provided (Fig. 1C) to demonstrate the change in SNR of each response type with mean luminance. The original data may look noisier than those obtained with trains of flashes. This is likely because sine-wave stimuli, in contrast to flashes, do not contain higher harmonics. Furthermore, the stimuli are generally less strong because the maximal contrast is much smaller than with flashes and because only subpopulations of photoreceptors are stimulated (S-cones are not stimulated at all; cones coexpressing S- and L*-pigments are only stimulated by photoisomerizations in the L*-pigment). However, by applying a Fourier analysis, the whole recording period is used, whereas with flashed stimuli often maxima and minima in restricted time windows are analyzed, thereby discarding the rest of the recording period. The amplitude plots (Fig. 1C) show that data with large SNRs can be obtained. The FFTs also show that the fundamentals are substantially larger than the higher harmonics, indicating that harmonic signal distortion is limited. 
As expected from our previous study,18 larger and more regular flicker ERG signals could be recorded from the LIAIS compared to the WT group, since both the rods and cones could be modulated at higher contrasts in the LIAIS strain (i.e., LIAIS cone contrast, 55%; rod contrast, 75%; WT cone and rod contrasts, both 5%). As mentioned in our previous paper,18 the cone-isolating conditions in LIAIS and in WT animals were nearly physically identical: the actual LED contrasts were slightly different, while the ratios of the LED contrast were identical (i.e., rods and S-cones were assumed to have identical fundamentals in WT and LIAIS mice so that identical conditions were needed for silencing them). Another initial observation is the clear distinction between the physiological properties of ERGs mediated by the two photoreceptor types, specifically, that rod-driven activity was greater at lower mean luminances, whereas more luminant conditions evoked larger cone-driven activity. From this it is also possible to point out that the luminance at which ERG responses switch between rod- and cone-dominant processing (i.e., a transition point or range) is likely to be between 1.4 and 7 cd/m2 with the mean chromaticities used in the present study. Moreover, rod ERGs at higher temporal frequencies (>12 Hz) appear to increase again in amplitude when mean luminance was higher than 21.5 cd/m2
The features described above were also reflected in LIAIS (filled symbols) and WT data (unfilled symbols) presented in Figures 2 and 3. There, the fundamental amplitudes (top panels) and phases (bottom panels) of rod- and cone-driven response waveforms, respectively, are presented as a function of mean luminance (rod isolation range: 0.1–73 cd/m2; cone isolation range: 0.2–130 cd/m2). Since the WT data were very small under all mean luminances (i.e., by up to 100% and 70% for rod and cone amplitudes, respectively, even at their preferred conditions described below, P ≤ 0.036), there was not enough dynamic range in their data set to measure the influence of luminance nor that of temporal frequency. Response amplitudes of WT mice (Figs. 2, 3) were often close to zero. The limited WT phase data set also precluded a clear phase relationship to be statistically determined because the phases of noise-like amplitudes were omitted. Still, the limited number of possible comparisons between response phases of WT and LIAIS groups indicated that they were similar under most rod- and cone-isolating conditions (P > 0.05). This sits well with previous studies, which supports the interpretation that the replacement of M-cone pigments in the LIAIS greatly improves the SNR of the recordable photoreceptor-driven signal and does not lead to different functional features from WT mice.15,18 Thus, apart from noting the few exceptions below, further analyses will only be carried out on data from LIAIS mice. 
Figure 2
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by rod-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 2
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by rod-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 3
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by cone-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 3
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by cone-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Influence of Mean Luminance on Rod- and L*-Cone–Driven Responses
LIAIS responses to putatively rod-isolating stimuli (Fig. 2) exhibited clear trends. Their amplitudes were much larger at the lower mean luminance conditions (0.1–1.4 cd/m2) compared to those elicited at high luminances in the present study (7–73 cd/m2) and in the previous study (13–130 cd/m2, P < 0.05 for comparisons with rod data recorded with the same 39 cd/m2 stimuli in Tsai et al.18). In fact, rod amplitudes were largest between 0.2 and 0.5 cd/m2 at most temporal frequencies. Between 0.5 and 1.4 cd/m2, rod amplitudes decreased with increasing luminance by 28% to 46%. Further increases in mean luminance resulted in a more severe drop in rod amplitude by up to 91% (e.g., 0.1 vs. 73 cd/m2 rod-driven amplitudes, P < 0.001). Above about 7 cd/m2, however, there was a slight increase in size again, concomitant with a dramatic shift in phase that is particularly noticeable for the 6-Hz responses (i.e., phase delay decreased by approximately 199° between 7 and 21.5 cd/m2, P = 0.028). This secondary amplitude increase (by approximately 12% to 52% between 7 and 73 cd/m2), however, did not reach significance (P > 0.05). Interestingly, WT response phases were consistently smaller than those of LIAIS responses at 7 cd/m2, where there is a proposed transition between rod- and cone-driven responses (P ≤ 0.007 at all frequencies). This phase difference possibly reflects interactions between rod and residual cone-driven responses to high-luminance rod-isolating stimuli, as proposed by Maguire et al.21 Otherwise, LIAIS rod-driven response phase increased (indicating shorter response delays) when luminance increased from 0.1 to 7 cd/m2 (i.e., by between 41° and 94°, P ≤ 0.002, for 6- to 22-Hz data, with the exception of ERGs at 26 Hz, where rod phases decreased significantly, P = 0.015). When the luminance was higher than 7 cd/m2, rod response phases decreased, in agreement with the phases of ERGs recorded at the 39 cd/m2 setting in the previous study.18 
Taken together, these data strongly indicate that there is a transition in mechanism underlying rod ERG responses at about 7 cd/m2. We propose that the small responses at high luminances in the mouse are determined by residual cone signals, as recently reported for humans. The large responses at low luminances are, in comparison, truly rod driven. In the further analysis, we considered only rod-driven responses at luminances below 7 cd/m2
As touched upon above (Fig. 1), mean luminance had the reverse effect on isolated cone-driven response amplitudes compared to its influence on the amplitudes of rod-driven signals (Fig. 3 versus Fig. 2). That is, cone amplitudes stimulated by modulations at 7 cd/m2 and higher were significantly larger than those elicited by lower stimulus at 1.4 cd/m2 and lower luminances (P ≤ 0.002), with the maximum cone-driven amplitude seen at 39 cd/m2. Cone-driven response amplitudes were therefore much smaller than those of rod-driven responses when evoked with stimuli of 0.1 to 1.4 cd/m2 (i.e., by >50 times) but much larger than rod responses under higher mean luminances (i.e., up to five-fold, all P ≤ 0.015). Still, the maximum cone-driven response amplitude (i.e., 2.59 ± 0.35 μV at 8 Hz, 39 cd/m2) was almost four times smaller than the maximum rod response (i.e., 9.82 ± 3.32 μV at 6 Hz, 0.2 cd/m2). Another distinguishing feature between the responses driven by the two photoreceptor response types was the change in amplitude between luminance conditions at which each photoreceptor type is more sensitive or insensitive (i.e., rod amplitudes up to 22× larger to ≤1.4 cd/m2 versus ≥7 cd/m2, and cone amplitudes up to 2× larger to ≥7 cd/m2 versus ≤1.4 cd/m2) (Fig. 4). Moreover, cone response phases were generally more stable under the more cone-sensitive conditions, as the slight phase increase with increasing luminance between 7 and 130 cd/m2 were insignificant (i.e., the phase delay of cone-driven responses did not decrease as much as rod, except at 6 and 14 Hz). 
Figure 4
 
Summary of LIAIS cone- versus rod-only ERG amplitudes (microvolt, top panels) and phases (degree, bottom panels, mean ± SD) from Figures 2 and 3, given as a function of temporal frequency (Hertz). Responses to the various mean luminances (photopic candela per square meter) are color-coded (see legend and Table 1 for conversions to scotopic candela per square meter and responses per rod per second). Only plots at luminances where we are certain that the indicated photoreceptor elicited measureable responses (i.e., cone: high luminances of 7–130 cd/m2; rod: low luminances of 0.1–1.4 cd/m2), as well as one luminance level outside this range (i.e., cone: 1.4 cd/m2, yellow plots: rod: 7 cd/m2, dark blue plots) are displayed. Phase-frequency relationships are modeled with a linear regression (see Table 2 for the parameters of each fit).
Figure 4
 
Summary of LIAIS cone- versus rod-only ERG amplitudes (microvolt, top panels) and phases (degree, bottom panels, mean ± SD) from Figures 2 and 3, given as a function of temporal frequency (Hertz). Responses to the various mean luminances (photopic candela per square meter) are color-coded (see legend and Table 1 for conversions to scotopic candela per square meter and responses per rod per second). Only plots at luminances where we are certain that the indicated photoreceptor elicited measureable responses (i.e., cone: high luminances of 7–130 cd/m2; rod: low luminances of 0.1–1.4 cd/m2), as well as one luminance level outside this range (i.e., cone: 1.4 cd/m2, yellow plots: rod: 7 cd/m2, dark blue plots) are displayed. Phase-frequency relationships are modeled with a linear regression (see Table 2 for the parameters of each fit).
Curiously, the present cone data to 39 cd/m2 isolation settings were on average quantitatively somewhat smaller compared to those previously recorded at these settings (see Tsai et al.18; P < 0.05 only for 26-Hz comparisons, phases comparable). In regard to the effect of temporal frequency (Fig. 4), and to some extent luminance, however, they were in the same ballpark. Age-related changes may be in play on the present data, as the LIAIS mice were measured at a substantially older age (9.5–13 months) than the cohort in the previous study (4–5 and 9 months old). 
Influence of Temporal Frequency on Rod- and L*-Cone–Driven Responses
We further examined how the rod- and cone-driven response profiles are influenced by the stimulus frequency. Figure 4 replots LIAIS rod and cone data from Figures 2 and 3 as a function of temporal frequency to better illustrate the influence of frequency on individual photoreceptor response types. Since we proposed that high-luminance rod data were driven by residual cone responses and low-luminance cone data were hardly above noise levels, these data sets were not included in the figure. Data sets at 1.4 and 7 cd/m2 were included in both cone and rod profiles respectively only to provide an overlap and to show the rod/cone switch in visual processing mentioned above. 
The most characteristic finding was the low-pass relationship of both rod- and cone-driven response amplitudes (top panels) and linearly decreasing phases (bottom panels) with increasing stimulus frequency (P < 0.001, seen also in Figs. 2 and 3, left to right panels). A low-pass relationship between frequency and cone ERG amplitude to very high-luminance white light on a rod-desensitizing adapting field (800 and 3200 cd/m2 on 5 cd/m2) was described previously in mice.44 In general, our L*-cone data were qualitatively similar to Krishna et al.'s44 study over the corresponding frequency range used in this study. Here, we additionally show that, temporal frequency had a stronger influence on rod-driven responses than on cone-driven responses. From 6 to 26 Hz, rod-driven response amplitudes decreased by a factor between 5- to 50-fold, whereas cone-driven response amplitudes decreased by a factor of maximally 5. At 26 Hz, both rod and cone responses were of similar size (t-test, P = 0.294). 
The corresponding rod and cone frequency-phase plots (Fig. 4; bottom panels) also showed a negative relationship with temporal frequency. The relationship between phase and temporal frequency could be described with a linear regression.25 Table 2 summarizes the mean parameters and goodness-of-fit of each linear model describing the response phases at the different mean luminance conditions. When slopes describing the frequency-phase dependency of rod- and cone-driven responses at their respective preferred luminance conditions were compared, the rod relationship exhibited steeper slopes (i.e., rods for 0.1–0.4 cd/m2, 14.4°–19°/Hz versus cones for 7–130 cd/m2, 11.8°–13.2°/Hz, corresponding to apparent latencies of 40–53 ms versus 33–37 ms, P = 0.002). 
Table 2
 
Summary of Linear Equation Parameters (f = y0 + a × x) Describing LIAIS Rod- Versus Cone-Driven Response Phases as a Function of Temporal Frequency
Table 2
 
Summary of Linear Equation Parameters (f = y0 + a × x) Describing LIAIS Rod- Versus Cone-Driven Response Phases as a Function of Temporal Frequency
Discussion
The improved capacity to noninvasively assess ERG responses elicited by single photoreceptor types in the LIAIS mouse over the WT mouse using the sinusoidal excitation modulation combined with the silent substitution method is once again highlighted in this study, as in our last.18 The current findings revealed that much larger rod-mediated flicker ERGs can be recorded from the lower mesopic luminance limb (<7 cd/m2) compared to the large L*-cone–driven signals at high photopic luminances (>7 cd/m2) when using this technique. Although the prominent difference in the luminance range over which rods and cones are most responsive is a well-known aspect of rod versus cone physiology, the fact that signals originating from the rod versus cone photoreceptor can now also be reliably assessed in controlled isolation (or for a given interaction) and in more detail in the mouse under a neutral adaptation state is otherwise not afforded by flash ERG or flicker ERGs obtained with pulses. In this respect, our results demonstrated that there are distinctions between the frequency characteristics of low-luminance rod compared to high-luminance L*-cone–driven signals. Thus, we conclude that the luminance and frequency characteristics of rod and L*-cone–driven sinusoidal flicker responses are indeed distinct in the LIAIS and thus likely in the WT mouse. In color-normal humans, rod versus L-cone–driven response profiles previously elucidated exhibit luminance characteristics similar to that of mice, whereas comparatively more tangible differences exist in humans in regard to the influence of stimulus frequency on the visual processes of the two response types. 
Luminance Dependency of Rod- and L*-Cone–Driven Responses
The regions of maximal rod or L*-cone activity are at opposite ends of the luminance range examined here, as anticipated from flash ERG results from the analogous Opn1mwR strain.19 
Very large rod-driven responses could be recorded between low luminances of 0.1 to 0.5 cd/m2 (i.e., mesopic conditions, maximum at 0.2 cd/m2, 8 Hz, corresponding to approximately 2 × 102 to 5 × 102 R*/rod/s; Table 1).41 Mean luminance conditions higher than 0.5 cd/m2 gave rise to increasingly smaller rod-driven responses until their amplitudes reached a minimum between 7 and 21.5 cd/m2 (i.e., up to 90% smaller at photopic conditions). In the analogous Opn1mwR strain,19 a response maximum was found for about 103 ∼ 104 R*/rod/s, which is a factor of 5 to 50 higher than the range found here. A reason for this difference may be related to the difference in the type of stimulation (i.e., sinusoidal modulation versus flashes19). Although the maxima of rod-mediated responses were relatively clear in our 12-Hz and 14-Hz amplitude functions (Fig. 2), we cannot rule out that lower luminances may evoke even larger rod-driven responses for lower temporal frequencies. However, since rod data from Allen and Lucas19 show no multiple shoulders or maxima, we are confident that the maximum rod-driven amplitude indeed is around 0.2 cd/m2. It appears, therefore, that the rod-driven response maxima is achieved under mesopic rather than scotopic states of adaptation. 
The luminance boundaries for a maximal and selective rod-driven response with rod-isolating silent substitution sine-wave stimuli in the LIAIS mice (i.e., approximately 0.1–0.5 cd/m2) corresponds well with a the retinal illuminance for a maximal rod-driven response as reported previously in human observers21,23,36,45 using analogous modulations at 8 to 15 Hz (which we calculated to be approximately 2 × 102 to 8 × 102 R*/rod/s). That both Park et al.45 and Cao et al.36 found larger rod-driven responses to reach a peak at low mesopic luminances (e.g., compared to more scotopic luminance ≤ 2 × 101 R*/rod/s) also indicates that the maximal rod-driven response can be found at similar mesopic luminances in mice and human observers. Of course this quantitative comparison makes sense only when similar luminance values also result in similar retinal illuminuances in the two species. This is the case when the ratio of pupil area and illuminated retinal area is similar. Although mouse and human eyes are not completely isometric, the deviations are probably relatively small.41 
In contrast to rod-driven responses that decreased in amplitude with increasing luminance, fundamental responses mediated by L*-cones exhibited the opposite trend: minimal amplitudes to mesopic stimuli of 0.2 to 1.4 cd/m2, followed by a gradual amplitude increase to reach a maximum between 21.5 to 130 cd/m2 (maximum at 39 cd/m2, 8 Hz, photopic conditions). Not surprisingly, the optimal flicker luminance window for obtaining robust L*-cone–driven flash ERG responses in Allen and Lucas's study19 is also higher than the luminance bracket revealed here for flicker ERGs (i.e., approximately ≥106 R*/rods/s versus 2 × 104 to 105 R*/rod/s, higher by a factor of at least approximately 10–50), similar to comparisons of the abovementioned rod isolation conditions. Be aware, however, that Allen and Lucas's data19 deals with flashes, whereas we work with sine-wave modulation. As such, the number of photoisomerizations per L*-cone might be different between our studies, depending on the spectral composition of the backgrounds used. 
Our window defined for maximal L*-cone–driven responses in LIAIS mice also largely overlaps with the L-cone sensitive boundaries previously described for human participants (i.e., approximately 3 × 104 to 2 to 105 R*/rod/s).21,45 Given that the maximal L-cone–driven flicker responses in humans peaked around 54 cd/m2 for 8-Hz modulation,21 we are confident that our protocol was able to reveal the maximum L*-cone–driven flicker amplitude for the LIAIS mouse, considering the similarities described above. 
A difference worth highlighting between the photoreceptor-specific flicker and flash ERG findings in long-wavelength–shifted mice is the ratio of rod-to-L*-cone–driven response amplitudes. Namely, that the maximal rod-driven flicker responses were larger than maximal L(*)-cone–driven ones by almost 4:1 in LIAIS mice and about 1.4:1 in humans.21 Maximal flash ERG b-wave amplitudes to rod-isolating stimuli in the analogous Opn1mwR mice, on the other hand, were smaller (i.e., rod-to-L*-cone of 1:2).19 Since maximal rod-mediated flicker and flash response amplitudes recorded from the two long-wavelength–shifted mice strains were similar in size (∼10 μV) despite being elicited at different luminance conditions (i.e., flash ∼50 × higher than flicker), the rod-to-L*-cone ratio difference between the ERG modalities suggests that the contribution of mouse L*-cone–driven activity to the flash ERG (i.e., at also 50× higher luminances than flicker) is greater. The difference between LIAIS and human rod-to-L(*)-cone flicker amplitude ratios is likely attributed to the higher rod-to-cone numbers in the rodent retina.46 
As for the phases of the flicker responses, there is an interesting difference between the dependency of rod and cone response phases on luminance. While rod-driven response phases increased with increasing luminance (up to 7 cd/m2, where it is assured that the responses are truly rod driven), the phases of L*-cone–driven responses are more stable for luminances above 7 cd/m2 (where cone-isolating stimuli elicit reliable responses). This may indicate fundamental differences in the signal processing of rod and cone signals. In humans, a monotonic shortening of phase delay has also been described for both rod- and L-cone–driven flicker ERGs over the respective (mesopic and mesopic to photopic) luminance ranges, where their amplitudes sharply increased to a maximum.45 Although Park et al.45 did not statistically compare these slope increases, qualitative comparison of their rod- and L-cone–driven luminance-phase profiles suggests that the rod response slope is steeper. They also reported the rod-driven response phase delays to be otherwise more prolonged at lower and higher luminances where their amplitudes are small, the latter of which is also evident in our LIAIS rod-driven ERGs to stimuli higher than 7 cd/m2. Interactions between fast and slow rod pathways at low-luminance levels and general rod insensitivity at high luminance levels have been proposed to be the cause of the more complex rod luminance-phase relationship.45 
Frequency Dependency of Rod- and L*-Cone–Driven Responses
Electrophysiological studies using nonspecific and photoreceptor-specific stimuli have shown the mouse to have low-pass frequency-response profiles with a frequency limit of about 30 Hz and have shown that their phases are negatively correlated with temporal frequency in a linear manner, not only for sine-wave modulation18,44,47 but also for trains of flashes.47 Here, we report the same for low-luminance rod and high-luminance L*-cone ERGs. What we additionally demonstrated in the current study is that rod- and L*-cone-driven frequency-response functions yielded under their respective preferred luminance conditions exhibit two distinguishing characteristics. Hence, rod- and L*-cone–driven signals in the mouse probably reflect differences in the physiology of the receptors themselves as well as in the postreceptoral retinal circuitry. 
For one, the amplitude decrease is much steeper for rod-driven signals (i.e., up to 50-fold amplitude decrease in rod-driven responses versus up to 5-fold decrease in cone-driven response amplitudes). Similarly, the phases of rod-driven responses also depend more strongly on stimulus frequency compared to L*-cone–mediated ones. The steeper rod frequency-phase relationships indicate that their phases become significantly more delayed relative to L*-cone response phases at higher temporal frequencies. This suggests that their processing mechanisms have different temporal characteristics, which was not manifest in our last study18 due to the abovementioned use of only high-luminance rod-isolating conditions. 
Human L-cone-driven ERG responses are band-pass with a maximum at about 30 to 40 Hz and thus respond strongly to much higher frequencies22,48 compared to <18 Hz for mice (see also Tsai et al.18 and Krishna et al.44). This indicates substantial differences in the processing of cone signals in the human and mouse retina. The low-pass frequency-response profiles of rod-driven responses in the mouse were found to be similar to those reported for human recordings with rod-isolating stimuli.21 We report here a 20- to 50-fold amplitude decrease with increasing frequency between 6 and 24 Hz (see Fig. 4). Maguire et al.21 report about a 10-fold amplitude decrease in that frequency range for human observers (their Fig. 321). 
Similar to our results in the mouse, the response phases of isolated photoreceptor-driven responses in humans also decrease linearly with temporal frequency. The frequency dependency of mouse rod-driven response phases suggested an apparent latency of 40 to 53 ms. This is slightly less than the apparent latencies of 60 to 75 ms found in human observers.21,36 The apparent latency of mouse cone-driven responses was found to be between 33 and 37 ms compared to 15 to 35 ms found in human observers.22,48 
In conclusion, the response amplitudes mediated by cone activity in human observers and mice reveal quite different signal processing, although the phase characteristics may be relatively similar. Rod-mediated responses may be similarly processed in mice and humans. 
For future studies, it may be worth examining whether the same results could be achieved without the full (overnight) dark adaptation (typically recommended for recording rod-dominant scotopic flash ERGs in rodents,1 as adopted here) or with a shorter period of darkness exposure prior to ERG recordings (e.g., 2 hours is sufficient to increase rod sensitivity).4952 If this were possible, as demonstrated in humans (i.e. circumventing the customary 30-minute dark-adaptation time21), the total experimental time to record rod-only signals in mice would be greatly shortened. 
Conclusions
Mean luminance and temporal frequency of a stimulus influences the amplitudes and phases of rod- and L*-cone–mediated flicker ERG responses in different ways. In the mouse, as in humans, only rod-driven signals are generated using stimuli within the defined low-luminance range and cone signals evoked at the high luminances specified can be attributed to visual activity in their respective circuitries. Given the lower (i.e., cone) frequency threshold in mice compared to that in humans, we propose rod-isolating flicker stimuli to be optimal at 2 × 102 R*/rod/s and L*-cone–isolating flicker stimuli at 4 × 104 R*/rod/s using 8-Hz sine-wave modulation for both types. By employing this paradigm on long-wavelength–shifted mice (like the LIAIS strain), mouse photoreceptor-specific visual function can be assayed with improved specificity, responsivity, and efficiency compared to the current standard methods. It also offers the prospect of examining individual photoreceptor characteristics to a broad range of fixed stimulus conditions in more detail, which is otherwise not permitted when the adaptation state is not kept constant, thereby also improving the prognostic value of the ERG in translational animal research. Finally, prospective studies on hybrid mouse strains crossbred from mice with long-wavelength–shifted cones and those that currently serve as models of human retinal disorders have the potential to improve current insights now that underlying pathophysiological changes can be isolated to a specific photoreceptor subtype–driven circuitry. 
Acknowledgments
The authors thank Maureen and Jay Neitz for providing the LIAIS mice. 
Supported by the German Research Foundation Grant KR1317/13-1 (JK). 
Disclosure: T.I. Tsai, None; A. Joachimsthaler, None; J. Kremers, None 
References
Weymouth AE, Vingrys AJ. Rodent electroretinography: methods for extraction and interpretation of rod and cone responses. Prog Retin Eye Res. 2008; 27: 1–44.
Redmond TM, Yu S, Lee E, et al. Rpe65 is necessary for production of 11-cis-vitamin A in the retinal visual cycle. Nat Genet. 1998; 20: 344–351.
Lei B, Yao G, Zhang K, Hofeldt KJ, Chang B. Study of rod- and cone-driven oscillatory potentials in mice. Invest Ophthalmol Vis Sci. 2006; 47: 2732–2738.
Humphries MM, Rancourt D, Farrar GJ, et al. Retinopathy induced in mice by targeted disruption of the rhodopsin gene. Nat Genet. 1997; 15: 216–219.
Toda K, Bush RA, Humphries P, Sieving PA. The electroretinogram of the rhodopsin knockout mouse. Vis Neurosci. 1999; 16: 391–398.
Chang B, Dacey MS, Hawes NL, et al. Cone photoreceptor function loss-3, a novel mouse model of achromatopsia due to a mutation in Gnat2. Invest Ophthalmol Vis Sci. 2006; 47: 5017–5021.
Calvert PD, Krasnoperova NV, Lyubarsky AL, et al. Phototransduction in transgenic mice after targeted deletion of the rod transducin alpha-subunit. Proc Natl Acad Sci U S A. 2000; 97: 13913–13918.
Cachafeiro M, Bemelmans AP, Canola K, et al. Remaining rod activity mediates visual behavior in adult Rpe65−/− mice. Invest Ophthalmol Vis Sci. 2010; 51: 6835–6842.
Jaissle GB, May CA, Reinhard J, et al. Evaluation of the rhodopsin knockout mouse as a model of pure cone function. Invest Ophthalmol Vis Sci. 2001; 42: 506–513.
Asteriti S, Gargini C, Cangiano L. Mouse rods signal through gap junctions with cones. eLife. 2014; 3: e01386.
Heikkinen H, Vinberg F, Nymark S, Koskelainen A. Mesopic background lights enhance dark-adapted cone ERG flash responses in the intact mouse retina: a possible role for gap junctional decoupling. J Neurophysiol. 2011; 105: 2309–2318.
Kremers J. The assessment of L- and M-cone specific electroretinographical signals in the normal and abnormal retina. Prog Retin Eye Res. 2003; 22: 579–605.
Lyubarsky AL, Falsini B, Pennesi ME, Valentini P, Pugh ENJr. UV and midwave-sensitive cone-driven retinal responses of the mouse: a possible phenotype for coexpression of cone photopigments. J Neurosci. 1999; 19: 442–455.
Nikonov SS, Kholodenko R, Lem J, Pugh ENJr. Physiological features of the S- and M-cone photoreceptors of wild-type mice from single-cell recordings. J Gen Physiol. 2006; 127: 359–374.
Greenwald SH, Kuchenbecker JA, Roberson DK, Neitz M, Neitz J. S-opsin knockout mice with the endogenous M-opsin gene replaced by an L-opsin variant. Vis Neurosci. 2014; 31: 25–37.
Smallwood PM, Olveczky BP, Williams GL, et al. Genetically engineered mice with an additional class of cone photoreceptors: implications for the evolution of color vision. Proc Natl Acad Sci U S A. 2003; 100: 11706–11711.
Jacobs GH, Williams GA, Cahill H, Nathans J. Emergence of novel color vision in mice engineered to express a human cone photopigment. Science. 2007; 315: 1723–1725.
Tsai TI, Atorf J, Neitz M, Neitz J, Kremers J. Rod- and cone-driven responses in mice expressing human L-cone pigment. J Neurophysiol. 2015; 114: 2230–2241.
Allen AE, Lucas RJ. Using silent substitution to track the mesopic transition from rod- to cone-based vision in mice. Invest Ophthalmol Vis Sci. 2016; 57: 276–287.
Allen AE, Storchi R, Martial FP, et al. Melanopsin-driven light adaptation in mouse vision. Current Biol. 2014; 24: 2481–2490.
Maguire J, Parry NR, Kremers J, Kommanapalli D, Murray IJ, McKeefry DJ. Rod electroretinograms elicited by silent substitution stimuli from the light-adapted human eye. Trans Vis Sci Tech. 2016; 5 (4): 13.
Kremers J, Pangeni G. Electroretinographic responses to photoreceptor specific sine wave modulation. J Opt Soc Am A Opt Image Sci Vis. 2012; 29: A309–A316.
Maguire J, Parry NR, Kremers J, Murray IJ, McKeefry D. The morphology of human rod ERGs obtained by silent substitution stimulation. Doc Ophthalmol. 2017; 134: 11–12.
Pangeni G, Horn FK, Kremers J. A new interpretation of components in the ERG signals to sine wave luminance stimuli at different temporal frequencies and contrasts. Vis Neurosci. 2010; 27: 79–90.
Kremers J, Scholl HPN. Rod-/L-cone and rod-/M-cone interactions in electroretinograms at different temporal frequencies. Vis Neurosci. 2001; 18: 339–351.
Aguilar M, Stiles WS. Saturation of the rod mechanism of the retina at high levels of illumination. Opt Acta. 1954; 1: 59–65.
Ecker JL, Dumitrescu ON, Wong KY, et al. Melanopsin-expressing retinal ganglion-cell photoreceptors: cellular diversity and role in pattern vision. Neuron. 2010; 67: 49–60.
Barrionuevo PA, Nicandro N, McAnany JJ, Zele AJ, Gamlin P, Cao D. Assessing rod, cone, and melanopsin contributions to human pupil flicker responses. Invest Ophthalmol Vis Sci. 2014; 55: 719–727.
Cao D, Nicandro N, Barrionuevo PA. A five-primary photostimulator suitable for studying intrinsically photosensitive retinal ganglion cell functions in humans. J Vis. 2015; 15 (1): 15.
Fukuda Y, Tsujimura S, Higuchi S, Yasukouchi A, Morita T. The ERG responses to light stimuli of melanopsin-expressing retinal ganglion cells that are independent of rods and cones. Neurosci Lett. 2010; 479: 282–286.
Brown TM, Tsujimura S, Allen AE, et al. Melanopsin-based brightness discrimination in mice and humans. Current Biol. 2012; 22: 1134–1141.
Joyce DS, Feigl B, Cao D, Zele AJ. Temporal characteristics of melanopsin inputs to the human pupil light reflex. Vision Res. 2015; 107: 58–66.
Dacey DM, Liao HW, Peterson BB, et al. Melanopsin-expressing ganglion cells in primate retina signal colour and irradiance and project to the LGN. Nature. 2005; 433: 749–754.
Nathan J, Reh R, Ankoudinova I, et al. Scotopic and photopic visual thresholds and spatial and temporal discrimination evaluated by behavior of mice in a water maze. Photochem Photobiol. 2006; 82: 1489–1494.
Umino Y, Solessio E, Barlow RB. Speed, spatial, and temporal tuning of rod and cone vision in mouse. J Neurosci. 2008; 28: 189–198.
Cao D, Pokorny J, Grassi MA. Isolated mesopic rod and cone electroretinograms realized with a four-primary method. Doc Ophthalmol. 2011; 123: 29–41.
Jacobs GH. Variations in colour vision in non-human primates. In: Foster DH, ed. Inherited and Acquired Colour Vision Deficiencies. Boca Raton, FL: CRC Press; 1991: 199–214.
Sun H, Macke JP, Nathans J. Mechanisms of spectral tuning in the mouse green cone pigment. Proc Natl Acad Sci U S A. 1997; 94: 8860–8865.
Lamb TD. Photoreceptor spectral sensitivities: common shape in the long-wavelength region. Vision Res. 1995; 35: 3083–3091.
Kremers J, Usui T, Scholl HPN, Sharpe LT. Cone signal contributions to electroretinograms in dichromats and trichromats. Invest Ophthalmol Vis Sci. 1999; 40: 920–930.
Lyubarsky AL, Daniele LL, Pugh ENJr. From candelas to photoisomerizations in the mouse eye by rhodopsin bleaching in situ and the light-rearing dependence of the major components of the mouse ERG. Vision Res. 2004; 44: 3235–3251.
Stockman A, Sharpe LT. Into the twilight zone: the complexities of mesopic vision and luminous efficiency. Ophthalmic Physiol Opt. 2006; 26: 225–239.
Bach M, Hawlina M, Holder GE, et al. Standard for pattern electroretinography. Doc Ophthalmol. 2000; 101: 11–18.
Krishna VA, Alexander KR, Peachey NS. Temporal properties of the mouse cone electroretinogram. J Neurophysiol. 2002; 87: 42–48.
Park JC, Cao D, Collison FT, Fishman GA, McAnany JJ. Rod and cone contributions to the dark-adapted 15-Hz flicker electroretinogram. Doc Ophthalmol. 2015; 130: 111–119.
Jeon CJ, Strettoi E, Masland RH. The major cell populations of the mouse retina. J Neurosci. 1998; 18: 8936–8946.
Tanimoto N, Sothilingam V, Kondo M, Biel M, Humphries P, Seeliger MW. Electroretinographic assessment of rod- and cone-mediated bipolar cell pathways using flicker stimuli in mice. Sci Rep. 2015; 5: 10731.
Kommanapalli D, Murray IJ, Kremers J, Parry NR, McKeefry DJ. Temporal characteristics of L- and M-cone isolating steady-state electroretinograms. J Opt Soc Am A Opt Image Sci Vis. 2014; 31: A113–A120.
Pinto LH, Vitaterna MH, Shimomura K, et al. Generation, characterization, and molecular cloning of the Noerg-1 mutation of rhodopsin in the mouse. Vis Neurosci. 2005; 22: 619–629.
Pinto LH, Invergo B, Shimomura K, Takahashi JS, Troy JB. Interpretation of the mouse electroretinogram. Doc Ophthalmol. 2007; 115: 127–136.
Behn D, Doke A, Racine J, Casanova C, Chemtob S, Lachapelle P. Dark adaptation is faster in pigmented than albino rats. Doc Ophthalmol. 2003; 106: 153–159.
Liu CN, Mathialagan N, Lappin P, et al. Crizotinib reduces the rate of dark adaptation in the rat retina independent of ALK inhibition. Toxicol Sci. 2015; 143: 116–125.
Figure 1
 
Representative flicker ERG waveforms evoked by mesopic to photopic (A) rod- and (B) M/L*-cone–isolating stimuli (photopic candela per square meter; see Table 1 for conversions to scotopic candela per square meter and response per rod per second) from LIAIS (black traces) and WT mice variants (gray traces; at the optimum luminances for rod and cone isolation only). Displayed are identical 500-ms episodes of responses (i.e., averages of left and right eyes of one animal and the two 500-ms episodes that make up the 1-second recording epoch) to sine-wave stimuli of 6 to 26 Hz (indicated at far left) and at the highest attainable contrast (i.e., 55% cone and 75% rod contrasts for LIAIS, 5% rod and cone contrasts for WT animals; see Materials and Methods). (C) Amplitude plots of 8 Hz rod (top panels) and L*-cone ERGs (bottom panels) after FFT (indicated by the arrow), as presented in (A) and (B) to demonstrate the respective changes in their signal-to-noise ratio with mean luminance.
Figure 1
 
Representative flicker ERG waveforms evoked by mesopic to photopic (A) rod- and (B) M/L*-cone–isolating stimuli (photopic candela per square meter; see Table 1 for conversions to scotopic candela per square meter and response per rod per second) from LIAIS (black traces) and WT mice variants (gray traces; at the optimum luminances for rod and cone isolation only). Displayed are identical 500-ms episodes of responses (i.e., averages of left and right eyes of one animal and the two 500-ms episodes that make up the 1-second recording epoch) to sine-wave stimuli of 6 to 26 Hz (indicated at far left) and at the highest attainable contrast (i.e., 55% cone and 75% rod contrasts for LIAIS, 5% rod and cone contrasts for WT animals; see Materials and Methods). (C) Amplitude plots of 8 Hz rod (top panels) and L*-cone ERGs (bottom panels) after FFT (indicated by the arrow), as presented in (A) and (B) to demonstrate the respective changes in their signal-to-noise ratio with mean luminance.
Figure 2
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by rod-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 2
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by rod-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 3
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by cone-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 3
 
Fundamental amplitudes (microvolt, top panels) and phases (degree, bottom panels) of responses elicited by cone-isolating stimuli. Data are group averages (±SD) given as a function of mean luminance (photopic and scotopic candela per square meter) for LIAIS (filled circles) and WT groups (unfilled circles).
Figure 4
 
Summary of LIAIS cone- versus rod-only ERG amplitudes (microvolt, top panels) and phases (degree, bottom panels, mean ± SD) from Figures 2 and 3, given as a function of temporal frequency (Hertz). Responses to the various mean luminances (photopic candela per square meter) are color-coded (see legend and Table 1 for conversions to scotopic candela per square meter and responses per rod per second). Only plots at luminances where we are certain that the indicated photoreceptor elicited measureable responses (i.e., cone: high luminances of 7–130 cd/m2; rod: low luminances of 0.1–1.4 cd/m2), as well as one luminance level outside this range (i.e., cone: 1.4 cd/m2, yellow plots: rod: 7 cd/m2, dark blue plots) are displayed. Phase-frequency relationships are modeled with a linear regression (see Table 2 for the parameters of each fit).
Figure 4
 
Summary of LIAIS cone- versus rod-only ERG amplitudes (microvolt, top panels) and phases (degree, bottom panels, mean ± SD) from Figures 2 and 3, given as a function of temporal frequency (Hertz). Responses to the various mean luminances (photopic candela per square meter) are color-coded (see legend and Table 1 for conversions to scotopic candela per square meter and responses per rod per second). Only plots at luminances where we are certain that the indicated photoreceptor elicited measureable responses (i.e., cone: high luminances of 7–130 cd/m2; rod: low luminances of 0.1–1.4 cd/m2), as well as one luminance level outside this range (i.e., cone: 1.4 cd/m2, yellow plots: rod: 7 cd/m2, dark blue plots) are displayed. Phase-frequency relationships are modeled with a linear regression (see Table 2 for the parameters of each fit).
Table 1
 
Generation of the Mean Stimulus Luminance Levels With the Aid of ND Filters and the Number of LIAIS and WT Animals (N) Recorded per Setting
Table 1
 
Generation of the Mean Stimulus Luminance Levels With the Aid of ND Filters and the Number of LIAIS and WT Animals (N) Recorded per Setting
Table 2
 
Summary of Linear Equation Parameters (f = y0 + a × x) Describing LIAIS Rod- Versus Cone-Driven Response Phases as a Function of Temporal Frequency
Table 2
 
Summary of Linear Equation Parameters (f = y0 + a × x) Describing LIAIS Rod- Versus Cone-Driven Response Phases as a Function of Temporal Frequency
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×