January 2005
Volume 46, Issue 1
Free
Visual Psychophysics and Physiological Optics  |   January 2005
Grating Acuity at Different Luminances in Wild-Type Mice and in Mice Lacking Rod or Cone Function
Author Affiliations
  • Christine Schmucker
    From the Section of Neurobiology of the Eye and the
  • Mathias Seeliger
    Retinal Electrodiagnostics Research Group, University Eye Hospital, Tübingen, Germany; the
  • Pete Humphries
    Department of Genetics, Trinity College, The University of Dublin, Dublin, Ireland; and the
  • Martin Biel
    Institute of Pharmacology, Center for Drug Research, Ludwig-Maximilians University, Munich, Germany.
  • Frank Schaeffel
    From the Section of Neurobiology of the Eye and the
Investigative Ophthalmology & Visual Science January 2005, Vol.46, 398-407. doi:10.1167/iovs.04-0959
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Christine Schmucker, Mathias Seeliger, Pete Humphries, Martin Biel, Frank Schaeffel; Grating Acuity at Different Luminances in Wild-Type Mice and in Mice Lacking Rod or Cone Function. Invest. Ophthalmol. Vis. Sci. 2005;46(1):398-407. doi: 10.1167/iovs.04-0959.

      Download citation file:


      © 2016 Association for Research in Vision and Ophthalmology.

      ×
  • Supplements
Abstract

purpose. The mouse eye has become an important model in vision research. However, it is not known how visual acuity changes with luminance. Therefore, grating acuity of mice was measured at different luminances in an automated optomotor paradigm. Furthermore, mutant mice lacking either rods (RHO−/− and CNGB1−/−) or cones (CNGA3−/−), or both, were studied to determine the rod and cone contribution to visual acuity.

methods. Freely ranging individual mice were automatically tracked at a 25-Hz sampling rate with a self-programmed video system in a large rotating optomotor drum. The drum had a square-wave grating inside with adjustable spatial frequency. The angular speed of the mice with respect to the center of the drum and the angular orientation of the snout-tail body axis were analyzed. In addition, the motor activity of the wild-type mice was recorded at different luminances.

results. The optomotor drum provided reliable data on visual input to the mouse’s behavior and was convenient to use, since the experimenter’s had only to place the mice individually in a Perspex cylinder. Optomotor grating acuity of the wild-type mice was limited to 0.3 to 0.4 cyc/deg. Maximum optomotor responses were obtained at 0.1 to 0.2 cyc/deg. The importance of visual input declined monotonically with decreasing luminance (30 cd/m2, 100%; 0.1 cd/m2, 76.4%; 0.005 cd/m2, 45.9%; and darkness, −9%). Mice lacking functional rods were able to resolve gratings up to 0.1 cyc/deg at 30 cd/m2. Surprisingly, mice lacking functional cones had an optomotor acuity that was similar to the wild-type. Double-knockout mice without rods and cones had no detectable grating acuity.

conclusions. Because the visual system of the mouse is more responsive at bright luminances, experiments in which visual input is important should be performed in photopic conditions (30 cd/m2 or even more). Apparently, spatial vision is governed by the rod system, which is not saturated in the mesopic or low photopic range. Mice lacking both rods and cones have no detectable grating acuity, indicating that the retinal melanopsin system does not contribute to spatial vision.

Currently, mice represent a widely studied mammalian model in vision research, although the mouse is not predominantly a visual animal. However, the advantages of the mouse model prevail and several aspects of visual function have already been examined. Using a force-choice behavioral paradigm, visual acuity was assessed in the “visual water task” by Prusky et al. 1 A behavioral discrimination test was also applied by Gianfranceschi et al. 2 (T-maze behavioral task), to measure visual acuity of wild-type and bcl2 transgenic mice, and by Prusky and Douglas 3 (visual water task), to study the developmental plasticity of mouse visual acuity. The forced-choice behavioral task requires extensive training and is demanding. Another technique used in assessing visual functions, is the optomotor response to a drifting grating. Sinex et al. 4 used this method to study grating acuity in the adult house mouse and the reeler mutant mouse. Furthermore, visual capabilities of mice were examined electrophysiologically 5 (Ridder W, et al. IOVS 2002;43:ARVO E-Abstract 1802) or anatomically, by determining the sampling intervals of the photoreceptors. 6 7 Even color vision was studied in a behavioral discrimination test. 8  
Despite the evidence that mice have some spatial vision, it is not known at present at which luminances it is important. For example, the mouse eye growth responds only sluggishly on deprivation of form vision 9 10 11 12 (Fernandes A, et al. IOVS 2004;45:ARVO E-Abstract 4280) which causes myopia in other animal models. This could suggest that either eye growth is only marginally controlled by visual input, or that the luminances that were used in the experiments were not appropriate. Therefore, in the first part of this study, we measured the visual acuity of wild-type C57BL6/J mice at different luminances. 
Furthermore, it is not known how the rod and cone system contributes to visual acuity. Some studies have been undertaken to examine the relative importance of rod and cone input, using mice lacking rod function (i.e., the RHO−/− or CNGB1−/− mutant) or cone function (i.e., the CNGA3−/− mutant), or even both (i.e., the double-mutant CNGA3−/− RHO−/−). These studies involved ERG recordings 13 (Geiger S, et al. IOVS 2003;44:ARVO E-Abstract 1871), histologic analysis, 14 or recordings of the pupillary light reflex and heart rate. 15  
The same four knockout models were therefore also behaviorally tested in the present study. The Rhodopsin-knockout (RHO−/−) mouse used in this study carries a replacement mutation in exon 2 of the rhodopsin gene. 16 As a result, RHO−/− mice do not build rod outer segments. Within 3 months, these mice loose all their photoreceptors. However, between postnatal weeks 4 and 6, when cone degeneration is not yet substantial, the mice can be used to study cone function in isolation. 13  
The photoreceptor membrane potential hyperpolarizes in response to illumination by closure of the cyclic nucleotide-gated (CNG) cation channels, 17 which, in turn, decreases synaptic glutamate release. In rod photoreceptors, the CNG channels are formed by the subunits CNGA1 and CNGB1 and in cone photoreceptors by CNGA3 and CNGB3. In respective knockouts of one channel subunit (CNGA3 and CNGB1), both the direct effects of the lack of one of these subunits and indirect effects such as problems with cellular trafficking are believed to cause the electrophysiologically observed selective functional loss. Consequently, the CNGB1−/− mouse used in this study completely lacks rod photoreceptor-mediated vision, but in comparison with the RHO−/− mouse, the rods are physically still present until late stages. The CNGA3−/− mouse 18 lacks cone-mediated light responses, which is also associated with a progressive degeneration of cone photoreceptors. Hence, these mice can be used to dissect rod- from cone-mediated signaling pathways. There are even double-knockout mice (CNGA3−/−RHO−/−) available, lacking both functional cones and rods. 14 These mice show a progressive degeneration of all photoreceptors within 3 months after birth. The inner retina remains unaffected. Until postnatal week 7, presynaptic markers and postsynaptic glutamate receptors are expressed, suggesting that neurotransmission can take place. 14 Panda et al. 19 showed that, in mice lacking rods and cones, the circadian rhythm is still regulated via a third retinaldehyde-based visual pigment, melanopsin, which is mostly expressed in a subset of retinal ganglion cells. Furthermore, it has been shown that mice lacking functional photoreceptors in the outer retina still have a light-induced pupil response 15 20 that is mediated by photosensitive ganglion cells containing melanopsin. We used the double-knockout mice to find out whether the retinal melanopsin system also contributes to spatial vision. 
Material and Methods
Animals
All experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The mouse experiments were approved by the University commission for animal welfare (reference AK3/02). The behavioral study included 25 black C57BL/6 wild-type mice, three Rhodopsin-knockout mice (RHO−/−), three rod CNG channel knockout mice (CNGB1−/−), three cone CNG channel knockout mice (CNGA3−/−), and three transgenic mice lacking both rod and cone function (CNGA3−/−RHO−/−). Age ranges of the tested animals were between 30 and 40 days. Black C57BL/6 wild-type mice were obtained from Charles River GmbH (Sulzfeld, Germany) and bred in the animal facilities of the Institute. Two mouse strains were generated by authors at the Institute (RHO−/− by PH and CNGB1−/− by MB), and CNGA3−/−RHO−/− mice were bred in Tübingen (by MS). CNGA3−/− mice on a matching C57BL/6 background were directly produced (by MB). 
Animals were housed with their mothers until weaning at approximately postnatal day (P)21, and then in groups of three to four in standard mouse cages under a 12-hour light–dark cycle. Ambient illuminance was provided by incandescent lights and was approximately 500 lux on the cage floor (measured with a calibrated photograph cell in photometric mode; United Detector Technology; Hawthorne, CA). All experimental procedures were conducted during the light phase (between 10 AM and 4 PM) of the daily cycle. 
Optomotor Experiment
Experimental Setup.
Spatial acuity was measured in an optomotor experiment as shown in Figure 1 . During testing, mice were individually placed in a clear transparent acrylic glass cylinder (diameter: 15 cm; height: 18 cm) that was placed in the middle of a rotating drum. Large and small optomotor drums were tested in the experiments to evaluate the effects of target distance and potential refractive errors in mice. If mice were myopic one would expect a higher grating acuity in the smaller drum, even if the spatial frequencies were adjusted for viewing distance. Furthermore, because the larger drum took more space and was more difficult to handle, a smaller set-up would have been more convenient. In the present study, the large drum had a diameter of 63 cm and a height of 35 cm, and the small one a diameter of 22 cm and a height of 29 cm. Data from the small drum are shown in Figure 5 , all other data are from the large drum. 
The drums provided the mouse with a drifting vertical square-wave pattern as it rotated in the vertical axis. Spatial frequency could be varied by placing stripe patterns of different widths (spatial frequency ranging from 0.03 to 0.6 cyc/deg) inside the drum. Stripe cylinders were made from clear plastic foil on which black stripes were printed with a 600-dpi laser printer. Because the inside of the drums were covered with white paint, the contrast was determined by the density of the print of the black stripes, which was close to 100%. 
The cylindrical container in which the mouse was freely moving was placed on a stationary white platform (diameter: 16 cm) in the center of the rotating drum (Fig. 1) , approximately 2 cm from its bottom. The drum was turned by an electric DC motor (Conrad Electronics, Hirschau, Germany). The direction of rotation could be changed by reversing the polarity of the voltage. The best optomotor responses were obtained for an angular speed of the stripe pattern between 50 and 60 deg/sec. Because the Perspex cylinder containing the mouse was closed, it was unlikely that the mouse was stimulated by air currents that might have been generated by the rotating drum. Furthermore, controls with stationary drums were performed (described later). 
Illumination of the Drum.
Spatial acuity testing was performed at different luminances in the drum (30, 0.1, 0.005, and 0 cd/m2, as measured with a luminance meter (LS-100 LS-110; Minolta, Osaka, Japan), positioned at the center of the acrylic glass cylinder at about the height of the mouse and oriented toward the stripe pattern. The luminance of 30 cd/m2 was generated by a light bulb (60 W; Philips, Eindhoven, The Netherlands). Luminances of 0.005 and 0.1 cd/m2 were produced by a white LED (diameter 10 mm, mcd typ 1200; Conrad Electronics) that was placed above the cylinder at 48 cm distance from the mouse. A frosted plastic diffuser, placed 2 cm below the LED, generated a largely homogenous illumination. To measure behavioral responses under very dim illumination or in complete darkness, the mouse container was illuminated by two high-power infrared LEDs (IR LEDs, VX-301 IR transmission diode, 80 mW/sr; Conrad Electronics) that were inserted in the cover of the acrylic glass cylinder, ∼16 cm above the mouse. 
The luminance meter was also used to estimate the stripe contrast directly. It was focused either on the black or the white stripes, and contrast was calculated by C = (Lmax − Lmin)/(Lmax + Lmin), where C is the contrast and L is luminance of the stripes. The measured contrasts were approximately 90% at 30 and 0.1 cd/m2 and 82% at 0.005 cd/m2
Programming Algorithms and Measured Parameters.
It was impossible to judge by eye whether the mouse followed a stripe pattern or not, since presumed phases of tracking were interrupted by movements in the opposite direction or by complete loss of interest, as the mouse often engaged in long periods of cleaning behavior. It was, therefore, necessary to automate the movement analysis. At this end, the mouse was imaged by a simple IR-sensitive monochrome miniature surveillance video camera (PAL format, 768 × 576 pixels; Conrad Electronics) that was equipped with a lens with a focal length of 5 mm to achieve a large field of view. The camera was mounted in the center of the top cover of the acrylic glass cylinder (Fig. 1) . After digitization of the video frames by a standard video board (Matrox Meteor II; TheImagingSource, Bremen, Germany), the video images were processed at 25 Hz by software written by one of the authors (FS) in Borland C++, version 5.02. The following steps were performed:
  1.  
    Measurement of the average pixel brightness in each video frame.
  2.  
    Detection of all pixels that were >40% darker than the average brightness.
  3.  
    Calculation of the center of mass of these pixels. This procedure reliably marked the center of the mouse body.
  4.  
    Measurement of the mouse’s angular running speed. Angular velocity (degrees per frame) of the center of mass with respect to the center of the cylinder was summed up over time, and the SD of all angular changes was determined after termination of the measurement session (approximately after 20 seconds). A one-sample t-test was automatically performed to find out whether there was a significant trend of the mouse to move in the direction of the drifting stripe pattern. Because the measurement of angular movement occurred in degrees, the 360°-to-0° transition, or vice versa, caused artifactual high speeds. Therefore, the program ignored measurements in which the angular velocity exceeded 2 deg/frame (50 deg/sec).
  5.  
    Measurement of the mouse’s angular body orientation. Because the mouse also turned its snout-tail axis in response to the drifting stripes, its orientation was also evaluated as a second parameter. An orthogonal regression was fit through the pixels marked in step 3. The change of the slope of this regression was tracked over time. Again, transitions from 360° to 0°, or vice versa, caused high angular speed as an artifact. This problem was again solved by excluding velocities above 2 deg/frame.
  6.  
    Because tracking the activity of the mouse was essential for gaining statistically reliable data, the locomotor activity was also recorded. The average absolute angular position change from one frame to the next, as determined in step 4 was taken as a measure of the activity.
The screen output of the software is shown in Figure 2
Measurement Procedures.
In most C57BL/6 wild-type mice, the directional preference of the movement was correlated with the drift direction of the stripes, immediately after the animal was placed inside the drum. In a few mice, the measurements had to be delayed until the animals had adapted to their new environment and had finished their self-cleaning behavior. To minimize habituation of the optomotor response, 21 the direction of rotation of the drum was reversed approximately every 20 seconds. The reversion was repeated five times at each spatial frequency. The initial direction of rotation was randomly chosen. Changing the direction by a mechanical switch on the power supply took about 2 seconds. Angular running speed, angular orientation speed, and locomotor activity were recorded for each direction of rotation. Spatial frequencies of the stripe patterns were exchanged in a random order. 
To assess the baseline noise in the measured parameters (i.e., the effects of spontaneous activity of the mouse), we tested how variable the responses of the animals were when no visual stimulus was present. Wild-type mice (C57BL/6) were therefore measured in a drum that was not moving and in a rotating drum that had no stripe pattern inside. 
All mice were tested at seven different spatial frequencies (0.03, 0.05, 0.1, 0.2, 0.3, 0.4, and 0.5 cyc/deg), using the large drum with a diameter of 63 cm. The wild-type mice were tested at four different light levels. For the measurements in darkness, mice were dark adapted for at least 60 minutes. RHO−/− and CNGB1−/− mice were tested at three light levels (30, 0.1, and 0.005 cd/m2) and both CNGA3−/− and CNGA3−/−RHO−/− mice were tested at two light levels (30 and 0.005 cd/m2). 
Furthermore, C57BL/6 wild-type mice were tested in the much smaller drum with a diameter of only 22 cm. Spatial frequencies of 0.03, 0.05, 0.07, 0.1, 0.2, 0.25, 0.3, 0.4, 0.5, and 0.6 cyc/deg were presented, with the stripe width corrected for the shorter viewing distance. However, because the mice could vary their distance to the stripe pattern, the changing viewing angles introduced large variations in the spatial frequencies. Using this setup, measurements were performed only at 30 cd/m2
Statistical Analysis
The response of the mouse to different stripe patterns was defined as the difference of its angular movement preference when the drum was rotating clockwise versus counter clockwise. This difference was analyzed both for the angular running speed and angular body orientation speed. The more this value differed from zero or the more it differed from the condition when no visual stimulation occurred, the more important the visual input was to the mouse’s behavior. 
Mean responses and standard deviations were plotted against spatial frequency. To estimate the cutoff spatial frequency that the mouse could still see, the responses were tested against zero, using paired Student’s t-test. 
Furthermore, responses at different spatial frequencies, responses under conditions when no visual stimulus was present, responses at different luminances, and responses of wild-type and knockout mice were compared by analysis of variance (one-way ANOVA). Post hoc analysis (the Dunnett test) was performed on factors that were found to be significant in the ANOVA. The significance level was set at 5%. 
Locomotor activity was compared at different luminances only in C57BL/6 wild-type mice by using a variance ratio test. Statistical tests were performed on computer (JMP, ver. 4 software; SAS Institute, Cary, NC). 
Results
Baseline Variability of the Measurement Procedure
Angular running and orientation speeds and the locomotor activity were studied in the C57BL/6 wild-type mouse in the large drum, either stationary or rotating, without any stripe pattern. Figure 3shows the mean optomotor responses and standard deviations in both cases. In addition, the effect of different luminances was tested. The responses were not significantly different from zero (P > 0.15, variance ratio test) under any of the tested conditions, indicating that the movement patterns of the mice were random. For the angular running speed, the response in the stationary drum was 0.002 ± 0.002 deg/frame and in the rotating drum 0.008 ± 0.013 deg/frame. For angular orientation speed, the response was 0.020 ± 0.008 and −0.001 ± 0.030 deg/frame, respectively. The responses were also not different between the stationary and rotating drum without the stripe pattern (P > 0.12, variance ratio test). 
Locomotor activity was not significantly affected by luminance (Fig. 3C; P > 0.10, variance ratio test). The mean locomotor activity was 0.298 ± 0.038 deg/frame. At each light level, locomotor activity was significantly different from zero (P < 0.02, variance ratio test). 
Spatial Vision in C57BL/6 Wild-Type Mice
Optomotor Response in the Large Drum.
Average responses and their standard deviations at different spatial frequencies are shown for both angular running speed and orientation speed in Figure 4 . Furthermore, Figure 4shows the possible uncertainty in the spatial frequency variable, resulting from the fact that the mice could move and vary the angles of viewing the stripes. A potential uncertainty in the spatial frequency variable of approximately ±20% was introduced. On average, the angular running speed was significantly larger than the angular orientation speed, and this difference reached statistical significance (difference: 0.022 ± 0.031 deg/frame, df = 40, t = 2.02, P = 0.003; variance ratio test). Obviously, the angular running speed had more descriptive power. 
The largest responses were obtained when the drum was rotated at 30 cd/m2. At this luminance, the mice displayed significant responses (against the null hypothesis) at spatial frequencies up to 0.3 cyc/deg. Beyond this stripe width, more animals began to move randomly (P > 0.05, variance ratio test). Comparing the responses at 30 cd/m2 with the response when no visual stimuli were present (response when no visual stimuli were present refers to both the response to a stationary drum and to a rotating drum without any stripe pattern inside; see Figs. 3 ) showed significant differences (P = 0.003, one-way ANOVA). This confirmed that the mice were able to resolve gratings up to 0.1 cyc/deg (P < 0.05, Dunnett test). 
The responses also declined when the luminance was reduced. To estimate the importance of visual input at different light levels, the responses at all tested spatial frequencies were added up. Using the sum of the responses at 30 cd/m2 as a reference, we found that visual input lost its importance from the brightest condition (30 cd/m2, 100%) to 76.4% at 0.1 cd/m2, to 45.9% at 0.005 cd/m2, to −9% in complete darkness. This was supported by a one-way ANOVA, which revealed significant differences between the different luminances (P = 0.003). The most compelling results of the post hoc analysis were that the responses at 0.1 cd/m2 did not differ significantly from the response at 30 cd/m2 (P > 0.05, Dunnett test), but there was a significant difference between the response at 30 cd/m2, at 0.005 cd/m2 (P < 0.05, Dunnett test), and in complete darkness (P < 0.005, Dunnett test). 
Optomotor Response in the Small Drum.
To test whether target distance and potential refractive errors had an effect on the measured grating acuity, wild-type mice were also studied in the small drum. 
Figure 5shows the responses of the mice to drifting gratings at 30 cd/m2. As in the large drum, responses reached a peak between 0.07 and 0.25 cyc/deg (angular running speed) or at 0.1 cyc/deg (angular orientation speed). In the small drum, angular running speed was not significantly different from angular orientation speed (difference: 0.012 ± 0.037 deg/frame, df = 18, t = 2.1, P = 0.48; variance ratio test). During these tests, the mice showed responses that were significantly different, from zero up to 0.5 cyc/deg (P < 0.05, variance ratio test). However, at a slightly lower spatial frequency of 0.4 cyc/deg, no significant response was measured. In addition, a one-way ANOVA revealed that the responses were different from the responses without visual stimuli (Fig. 3 ; P < 0.0001). The conclusion drawn by the variance ratio test was supported by a post hoc analysis (P < 0.05, Dunnett test). The slightly higher spatial acuity obtained in this experiment could have resulted either from myopic refractive errors of the mice or from the fact that they were able to approach the stripe patterns, increasing the viewing angle and reducing spatial frequencies. The uncertainty of the spatial frequency variable was calculated by simple geometry and is plotted as horizontal error bars in Figure 5
Spatial Vision in RHO−/− Mice
Figure 6shows the grating acuity in Rhodopsin knockout mice at three light levels. Contrary to wild-type mice in the large drum, there is no significant difference between angular running and orientation speed (difference: −0.004 ± 0.066 deg/frame, df = 38, t = 2.0, P = 0.78, variance ratio test). Comparing the responses of the RHO−/− mouse to the condition in which no visual stimulation occurred (Fig. 3) , significant differences were revealed (P = 0.002, one-way ANOVA). The Dunnett test showed that significant responses at 0.03, 0.05, and 0.1 cyc/deg were only elicited at 30 cd/m2 (P < 0.05). The same conclusion was reached when the responses were tested against zero (P < 0.05, variance ratio test). 
A one-way ANOVA was also performed to identify differences between the responses of the Rhodopsin knockout and the wild-type mice (P = 0.007). The post hoc analysis showed that there was no difference between the responses at 30 cd/m2 in the two strains (P > 0.05, Dunnett test). However, the RHO−/− mouse showed a reduced response at both 0.1 and 0.005 cd/m2 (P < 0.005, Dunnett test). 
Spatial Vision in CNGB1−/− Mice
Grating acuity in the second model lacking rod-mediated vision is shown in Figure 7for three light levels. As in the Rhodopsin knockout mouse, significant responses were only elicited at 30 cd/m2. Again, there was no significant difference between angular running and angular orientation speed (difference: −0.005 ± 0.024 deg/frame, df = 40, t = 2.0, P = 0.36, variance ratio test). 
These mutants showed significant response to gratings of 0.03, 0.05 and 0.2 cyc/deg at 30 cd/m2 (P < 0.05, variance ratio test). In comparison to the response when no visual stimuli were present (Fig. 3) , responses were significant at 0.03 and 0.05 cyc/deg at 30 cd/m2 (P < 0.05, Dunnett test). 
To uncover differences between the responses of the CNGB1−/− and the wild-type mice, a one-way ANOVA was performed (P = 0.001). Similar to the RHO−/− mouse, the post hoc analysis showed no difference between the responses at 30 cd/m2 (P > 0.05, Dunnett test). Again, the responses at 0.1 and 0.005 cd/m2 were significantly reduced (P < 0.005, Dunnett test). A one-way ANOVA did not reveal any differences between the two knockout models lacking rod function (P = 0.4). 
Spatial Vision in CNGA3−/− Mice
Data from mice lacking cone function are presented in Figure 8 . On average, the angular orientation speed was significantly larger than the angular running speed, and this difference reached statistical significance (difference: 0.050 ± 0.048 deg/frame, df = 26, t = 2.1, P = 0.0007, variance ratio test). Surprisingly, both at 30 and at 0.005 cd/m2, the CNGA3−/− mice performed in the optomotor task comparably to the wild-type (no significant difference, P > 0.08, one-way ANOVA). 
Student’s t-test showed significant responses up to 0.1 cyc/deg at 30 cd/m2 (P < 0.05, variance ratio test). Comparing the responses of the mouse lacking cone function with the responses when no visual stimuli were present (Fig. 3)a one-way ANOVA revealed significant differences (P < 0.0001). The most compelling results of the post hoc analysis were that spatial frequencies up to 0.2 cyc/deg were resolved at 30 cd/m2 (P < 0.05, Dunnett test). 
Spatial Vision in CNGA3−/−RHO−/− Mice
The optomotor responses of mice lacking both rods and cones are shown in Figure 9 . Different from C57BL/6 mice but similar to mice lacking rod function, there was no significant difference between angular running and orientation speed (difference: −0.006 ± 0.36 deg/frame, df = 26, t = 2.1, P = 0.59, variance ratio test). Responses were neither significantly different from the null hypothesis (P > 0.05, variance ratio test) nor from the responses without visual stimulation (P = 0.55, one-way ANOVA). In conclusion, these animals were obviously not able to distinguish the black and white stripes. 
Comparisons of Optomotor Responses in Wild-Type and Mutant Mice
To quantify the importance of spatial vision in the wild-type and knockout models, the responses obtained at different spatial frequencies were added up, providing a number that reflects the importance of spatial visual input. This summation was performed at all tested luminances, and the results are shown in Table 1 (response refers to angular running speed) and Table 2(response refers to angular orientation speed). The sum of the responses of the wild-type mice at 30 cd/m2 were used as a reference and set to 100%. The tables illustrate that the importance of visual input declined monotonically with decreasing luminance. In mutant mice lacking rod function (RHO−/− and CNGB1−/−), no visual input was detected at all in dim light (0.1 and 0.005 cd/m2). At 30 cd/m2, the optomotor response was still 24% to 41% of the wild-type when angular running speed was evaluated, and between 60% and 74% when angular orientation speed was evaluated. Mice lacking cone function (CNGA3−/−) displayed an optomotor response that was even larger than that in wild-type mice. Also, in this mutant, the angular orientation speed achieved higher statistical significance. The responses of the double-knockout mice (CNGA3−/−RHO−/−) did not differ from the noise levels found in the wild-type in complete darkness. 
Discussion
Evaluation of the Optomotor Paradigm
The optomotor response to a moving grating can be used to evaluate the spatial acuity in mice, as suggested by the following observations:
  1.  
    Most mice showed a preference to move in the direction of the stripe pattern.
  2.  
    There was no preference, whether the drum was rotated without the stripe pattern inside or was not rotated at all.
  3.  
    The movements were also random in complete darkness.
  4.  
    The differences in movement preferences were not explained by light-dependent activity changes, since the locomotor activity did not change at different luminances (P > 0.10, variance ratio test).
Even though the behavioral procedure offers the advantage that the animals can be tested without prior training and while they are freely moving, a disadvantage might be that they are not perfectly centered in the drum and therefore are viewing the stripe patterns from variable distances. 
A complete grating sensitivity function for a single mouse requires approximately 45 minutes of testing. Hence, relatively rapid screening for spatial vision in mice can be performed, and the experimenter need only place the mouse in the Perspex container. 
Spatial Acuity and Contrast Sensitivity in C57BL/6 Wild-Type Mice, Compared with Other Mammals
In primates, foveal visual acuity is limited by the density of cones, as they have a one-to-one correspondence between cones and ganglion cells in the highest-density region of the retina. 22 In lower mammals, the upper limit is imposed by the peak density of the retinal ganglion cell (RGC) mosaic, as there is a significant degree of convergence of cones onto ganglion cells. 23 Gianfranceschi et al. 2 estimated a visual acuity in wild-type mice of 1 cyc/deg using the equation: estimated visual acuity = (√D/2) · RMF, where D is the peak density of RGCs (4500 cells/mm2), and RMF is the retinal magnification factor (0.015 mm/deg; RMF = 2 · π · posterior nodal distance of the eye/360). In most species, the behavioral visual acuity is very close to the estimated acuity (i.e., horse, dog, cat, rabbit, and dolphin; see overview). 2 However, in mice and rats whose habits do not depend primarily on vision, the behavioral acuity is lower. 
Our optomotor experiment provided a slightly lower spatial resolution limit (0.3 –0.4 cyc/deg), compared with other behavioral paradigms. An acuity of approximately 0.5 cyc/deg was found in wild-type mice in measurements of optokinetic eye movements 4 or forced-choice procedures (visual water task 1 and T-maze behavioral task 2 ). In addition, electrophysiology, both pattern electroretinogram (PERG) 24 and visual evoked potentials (VEPs), 5 25 26 has been extensively used to measure acuity and contrast thresholds in both wild-type and mutant mice, even during development. 27 Both VEP and PERG measurements provide spatial resolution thresholds (visual acuity) of approximately 0.6 cyc/deg. The lower spatial frequency cutoff found in the present study may be because the statistical significances were hidden in higher standard deviations. 
In our study, mice were most sensitive to square-wave gratings between 0.1 and 0.2 cyc/deg. Lower spatial frequencies caused a reduced optomotor response. This result is in accordance with the observations in a behavioral study by Sinex et al., 4 who reported the highest grating sensitivity at 0.125 cyc/deg in mice. Porciatti et al. 5 estimated highest sensitivity to gratings of 0.06 cyc/deg, using pattern VEPs. This sensitivity is also comparable with the average receptive field size obtained from single-unit recordings in the visual cortex of the mouse. 28 29 30  
By comparison, the contrast sensitivity function of Long-Evans and nondystrophic RCS rats peaked near 0.2 cyc/deg. 31 The peak sensitivity to sine-wave gratings for the cat is at∼1 cyc/deg 32 and is at ∼4 cyc/deg in the pigeon, 33 3 to 5 cyc/deg in the squirrel monkey, 34 and 3 cyc/deg in the macaque. 35  
Grating Acuity at Different Light Levels
The importance of visual input decreased monotonically with luminance (30 cd/m2, 100%; 0.1 cd/m2, 76.4%; 0.005 cd/m2, 45.9%; and darkness, −9%), suggesting that the high-acuity system of the mouse requires relatively high light levels, similar to humans (>200 lux, 36 which is equivalent to ∼30 cd/m2 in our test conditions). Porciatti et al. 5 found that the VEP amplitude is at its maximum at a luminance of 25 cd/m2 (low photopic range). Similar to our study, a decreasing VEP response was observed when the light was dimmed. No reliable response was elicited in the scotopic range (< 0.01 cd/m2) in the VEPs. This result is different from our behavioral data, because there was still visual input detected (approximately 45% of the maximum response) in the range where only rod vision was present (0.005 cd/m2). Our findings are supported by the study of Herreros de Tejada et al., 37 who measured absolute visual threshold in albino and pigmented mice, by using an operant method (modified Skinner box) and estimated threshold values of −5.3 and −5.5 log cd/m2, respectively. Because the mouse retina is rod dominated (97%), 38 these results are not surprising. As should be expected, no visual input was measurable in complete darkness. Also, Mitchiner et al. 21 did not observe any eye movements in an optokinetic paradigm when the stripes were rotated in the dark. Visual input was most important at photopic conditions, suggesting that studies on myopia should be done at luminances of 30 cd/m2 or even higher. 
Refractive State and Visual Acuity
Although there are extensive data on the refractive state in mice as measured by optical techniques, 11 it is difficult to evaluate the small-eye artifact, and accordingly, the true subjective refractive state. There was a slight improvement in spatial acuity when a smaller drum was used (large drum: 0.3 cyc/deg, small drum: 0.5 cyc/deg; P < 0.05, variance ratio test). This observation could suggest that the mice were slightly myopic. However, since small eyes with high refractive power have a large dioptric depth of focus, 39 the gratings used in our task were probably in best focus in both drums, and the potential myopia was not limiting. Also electrophysiological recordings in mice did not require refractive corrections to map receptive fields in the visual cortex. 30 40 Therefore, the slight improvement in spatial acuity in the small drum resulted most likely from the fact that the mice could approach the stripe pattern, thereby increasing their viewing angle. 
Spatial Acuity in Mutant Mice
We found that spatial vision in the juvenile RHO−/− and CNGB1−/− mice was limited to the photopic range (30 cd/m2), a finding in line with previous electrophysiological and anatomic work. 13 These authors had shown that during the period of complete absence of rod input but normal or even supernormal cone responses between postnatal weeks 4 and 6, the (cone only) ERG features a right shift, due to the lower sensitivity in cones than in rods. Our behavioral data showed a reduced visual acuity in these cone-only models (0.1 cyc/deg in RHO−/− and 0.2 cyc/deg in CNGB1−/− compared with 0.3 cyc/deg in C57BL/6 wild-type mice; P < 0.05, variance ratio test). This suggests that, in the absence of a macula, the peak visual performance in mice is obtained when they use their rod system. The difference between the RHO−/− and the CNGB1−/− mutants may reflect the more natural morphologic organization of cone outer segments in CNGB1−/− mice, due to the presence of supporting, but nonfunctional rods. The all-rod-mouse (CNGA3−/−) performed the behavioral test, in both photopic and scotopic conditions, similar to the wild-type. This observation suggests that the rod system (without dilation of pupils) is not entirely saturated at luminances of up to 30 cd/m2. It may be speculated that rod vision originates in this case from a midperipheral ring between a central area of desensitization (too much light for the rod system) and a peripheral ring of subthreshold stimulation (too little light for the rod system). 
In mice without any functional photoreceptors in the outer retina (CNGA3−/−RHO−/−), no optomotor response was elicited under our test conditions. This suggests that the melanopsin-containing ganglion cell system does not contribute to spatial vision. A similar conclusion was reached, based on ERG recordings, by Claes et al. 14  
Summary
The study demonstrated that an automated optomotor paradigm with freely ranging mice could be used to study spatial visual functions in both wild-type and mutant mice. It was found that spatial vision in mice is limited to the photopic range. Our results suggest that experiments in which visual input is important should be done at 30 cd/m2 or even higher. Studies with mutants suggest that spatial acuity is governed by rod vision which appears, without dilation of pupils, to function properly, even in the mesopic or low photopic range. Mice lacking both cones and rods have no detectable grating acuity, indicating that the retinal melanopsin system does not contribute to spatial vision. 
 
Figure 1.
 
Setup for the behavioral experiment to test spatial acuity in mice. Mice were placed individually in a Perspex container. To measure behavioral response under dim illumination or in darkness, the mouse was illuminated by two high-powered IR LEDs. An IR light-sensitive video camera imaged the mouse and, after digitization of the video frames, a screen output as shown in Figure 2was obtained. The pattern of vertical black and white stripes that was placed inside the drum was made from clear plastic foil. Spatial frequency could be varied by placing stripe patterns with different stripe width inside the drum. The drum was illuminated either by a light bulb or a white-light LED. In the latter case, a frosted diffuser was placed in front of the LED to generate a homogenous illumination.
Figure 1.
 
Setup for the behavioral experiment to test spatial acuity in mice. Mice were placed individually in a Perspex container. To measure behavioral response under dim illumination or in darkness, the mouse was illuminated by two high-powered IR LEDs. An IR light-sensitive video camera imaged the mouse and, after digitization of the video frames, a screen output as shown in Figure 2was obtained. The pattern of vertical black and white stripes that was placed inside the drum was made from clear plastic foil. Spatial frequency could be varied by placing stripe patterns with different stripe width inside the drum. The drum was illuminated either by a light bulb or a white-light LED. In the latter case, a frosted diffuser was placed in front of the LED to generate a homogenous illumination.
Figure 2.
 
Screen dump of the C++ program that tracked the mouse. The program tracked the movement of the center of mass of the mouse, marked by the cross with the circle. The trace of movement is shown on the right. The average angular velocity of the center of mass of the mouse (Average Running Speed) in relation to the center of the container was summed up over time, and the standard deviation of all angular changes was calculated after termination of the measurement session (approximately after 20 seconds). The angular movement of the snout-tail axis (see dotted line on the mouse image) was also recorded (Angular Orientation Speed). Finally, the Average Activity was recorded as the average absolute angular position change from one frame to the next.
Figure 2.
 
Screen dump of the C++ program that tracked the mouse. The program tracked the movement of the center of mass of the mouse, marked by the cross with the circle. The trace of movement is shown on the right. The average angular velocity of the center of mass of the mouse (Average Running Speed) in relation to the center of the container was summed up over time, and the standard deviation of all angular changes was calculated after termination of the measurement session (approximately after 20 seconds). The angular movement of the snout-tail axis (see dotted line on the mouse image) was also recorded (Angular Orientation Speed). Finally, the Average Activity was recorded as the average absolute angular position change from one frame to the next.
Figure 3.
 
Responses of C57BL/6 wild-type mice in a stationary white drum (A) and in a rotating white drum without the stripe pattern inside (B). There was no significant preference of the mice to move in any particular direction (P > 0.15, variance ratio test). The locomotor activity of the mice was not significantly different at different luminances (P > 0.10, variance ratio test). (C) Stationary drum without stripe pattern. Each bar graph shows the mean ± SD of results from at least three animals.
Figure 3.
 
Responses of C57BL/6 wild-type mice in a stationary white drum (A) and in a rotating white drum without the stripe pattern inside (B). There was no significant preference of the mice to move in any particular direction (P > 0.15, variance ratio test). The locomotor activity of the mice was not significantly different at different luminances (P > 0.10, variance ratio test). (C) Stationary drum without stripe pattern. Each bar graph shows the mean ± SD of results from at least three animals.
Figure 4.
 
Mean optomotor responses and their standard deviations of C57BL/6 wild-type mice at different spatial frequencies. (A) Angular running speed and (B) angular orientation speed. The horizontal error bar illustrates the uncertainty of the spatial frequency variable, originating from the fact that the mice could change the viewing angle by moving closer to the stripe pattern. Data from seven animals are shown, with three or more animals tested at each data point. Angular running speed had more statistical power (P < 0.005, variance ratio test). The responses of the mice were significantly different from the condition in which no visual stimuli were present, for spatial frequencies up to 0.1 cyc/deg at 30 cd/m2 (P < 0.05, Dunnett test). If compared with the null hypothesis, responses were significantly different from zero up to 0.3 cyc/deg (P < 0.05, variance ratio test). The responses at 0.1 cd/m2 were decreased, but not significantly different from the responses at 30 cd/m2 (P > 0.05, Dunnett test). Responses at 0.005 cd/m2 were significantly reduced compared to the response at 30 cd/m2 (P < 0.05, Dunnett test). Responses in complete darkness were neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the response when no visual stimuli were present (P > 0.005, Dunnett test).
Figure 4.
 
Mean optomotor responses and their standard deviations of C57BL/6 wild-type mice at different spatial frequencies. (A) Angular running speed and (B) angular orientation speed. The horizontal error bar illustrates the uncertainty of the spatial frequency variable, originating from the fact that the mice could change the viewing angle by moving closer to the stripe pattern. Data from seven animals are shown, with three or more animals tested at each data point. Angular running speed had more statistical power (P < 0.005, variance ratio test). The responses of the mice were significantly different from the condition in which no visual stimuli were present, for spatial frequencies up to 0.1 cyc/deg at 30 cd/m2 (P < 0.05, Dunnett test). If compared with the null hypothesis, responses were significantly different from zero up to 0.3 cyc/deg (P < 0.05, variance ratio test). The responses at 0.1 cd/m2 were decreased, but not significantly different from the responses at 30 cd/m2 (P > 0.05, Dunnett test). Responses at 0.005 cd/m2 were significantly reduced compared to the response at 30 cd/m2 (P < 0.05, Dunnett test). Responses in complete darkness were neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the response when no visual stimuli were present (P > 0.005, Dunnett test).
Figure 5.
 
Optomotor responses of C57BL/6 wild-type mice at different spatial frequencies, for angular running speed (A) and angular orientation speed (B), as measured in the small drum. Error bars are as in Figure 4 . Data from 15 animals contributed to the curve, with 5 or more animals for each data point. Mice showed significant responses for spatial frequencies up to 0.5 cyc/deg. Eight of 10 responses were significantly different from zero (P < 0.05, variance ratio test) and significantly different from the responses when no visual stimuli were present (P < 0.05, Dunnett test).
Figure 5.
 
Optomotor responses of C57BL/6 wild-type mice at different spatial frequencies, for angular running speed (A) and angular orientation speed (B), as measured in the small drum. Error bars are as in Figure 4 . Data from 15 animals contributed to the curve, with 5 or more animals for each data point. Mice showed significant responses for spatial frequencies up to 0.5 cyc/deg. Eight of 10 responses were significantly different from zero (P < 0.05, variance ratio test) and significantly different from the responses when no visual stimuli were present (P < 0.05, Dunnett test).
Figure 6.
 
Optomotor responses of Rhodopsin knockout (RHO−/−) mice are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At 30 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test). At 0.1 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test) nor significantly different from the responses when no visual stimuli were present (P > 0.05, Dunnett test).
Figure 6.
 
Optomotor responses of Rhodopsin knockout (RHO−/−) mice are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At 30 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test). At 0.1 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test) nor significantly different from the responses when no visual stimuli were present (P > 0.05, Dunnett test).
Figure 7.
 
Optomotor responses of mice lacking rod function (CNGB1−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars as in Figure 4 . At 30 cd/m2, the mice showed significant responses at 0.03 and 0.05 cyc/deg (P < 0.05, Dunnett test). At both 0.1 and 0.005 cd/m2, responses were random (P > 0.05, Dunnett test).
Figure 7.
 
Optomotor responses of mice lacking rod function (CNGB1−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars as in Figure 4 . At 30 cd/m2, the mice showed significant responses at 0.03 and 0.05 cyc/deg (P < 0.05, Dunnett test). At both 0.1 and 0.005 cd/m2, responses were random (P > 0.05, Dunnett test).
Figure 8.
 
Optomotor responses of mice lacking cone function (CNGA3−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . In this mutant, the angular orientation speed provided higher significances (P < 0.0007, variance ratio test). At 30 cd/m2, the mice showed significant responses up to 0.2 cyc/deg (P < 0.05, Dunnett test). At 0.005 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test).
Figure 8.
 
Optomotor responses of mice lacking cone function (CNGA3−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . In this mutant, the angular orientation speed provided higher significances (P < 0.0007, variance ratio test). At 30 cd/m2, the mice showed significant responses up to 0.2 cyc/deg (P < 0.05, Dunnett test). At 0.005 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test).
Figure 9.
 
Optomotor responses of mice lacking both rod and cone function (CNGA3−/−RHO−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At both 30 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the responses without visual stimulation (P > 0.05, Dunnett test).
Figure 9.
 
Optomotor responses of mice lacking both rod and cone function (CNGA3−/−RHO−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At both 30 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the responses without visual stimulation (P > 0.05, Dunnett test).
Table 1.
 
Relative Responses in Terms of Angular Running Speed in Wild-type and Knockout Mice
Table 1.
 
Relative Responses in Terms of Angular Running Speed in Wild-type and Knockout Mice
Response (%) % Response When No Visual Stimulus Was Present (Rotating Drum) % Response When No Visual Stimulus Was Present (Stationary Drum)
30 cd/m2 0.1 cd/m2 0.005 cd/m2 0 cd/m2
Wild-type 100 76 46 −9 18 3
RHO−/− 24 −1 −14
CNGB1−/− 41 −11 −4
CNGA3−/− 70 8
CNGA3−/−RHO−/− 15 15
Table 2.
 
Relative Responses in Terms of Angular Orientation Speed in Wild-type and Knockout Mice
Table 2.
 
Relative Responses in Terms of Angular Orientation Speed in Wild-type and Knockout Mice
Response (%) % Response When No Visual Stimulus Was Present (Rotating Drum) % Response When No Visual Stimulus Was Present (Stationary Drum)
30 cd/m2 0.1 cd/m2 0.005 cd/m2 0 cd/m2
Wild-type 100 48 41 12 −29 33
RHO−/− 74 7 −53
CNGB1−/− 60 −7 −3
CNGA3−/− 139 82
CNGA3−/−RHO−/− 24 24
PruskyGT, WestPWR, DouglasRM. Behavioral assessment of visual acuity in mice and rats. Vision Res. 2000;40:2201–2209. [CrossRef] [PubMed]
GianfranceschiL, FiorentiniA, MaffeiL. Behavioral visual acuity of wild type and bcl2 transgenic mouse. Vision Res. 1999;39:569–574. [CrossRef] [PubMed]
PruskyGT, DouglasRM. Developmental plasticity of mouse visual acuity. Eur J Neurosci. 2003;17:167–173. [CrossRef] [PubMed]
SinexDG, BurdetteLJ, PearlmanAL. A psychophysical investigation of spatial vision in the normal and reeler mutant mouse. Vision Res. 1979;19:853–857. [CrossRef] [PubMed]
PorciattiV, PizzorussoT, MaffeiL. The visual physiology of the wild type mouse determined with pattern VEPs. Vision Res. 1999;39:3071–3081. [CrossRef] [PubMed]
HughesA. The topography of vision in mammals of contrasting life style: comparative optics and retinal organization.CrescitelliF eds. Handbook of Sensory Physiology. 1977;7:615–756.Springer Berlin.
MartinG. Psychophysics. Limits of visual resolution. Nature. 1986;319:540. [PubMed]
JacobsGH, WilliamsGA, FenwickJA. Influence of cone pigment coexpression on spectral sensitivity and color vision in the mouse. Vision Res. 2004;44:1615–1622. [CrossRef] [PubMed]
SchaeffelF, BurkhardtE, HowlandHC, et al. Measurement of refractive state and deprivation myopia in two strains of mice. Optom Vis Sci. 2004;81:99–110. [CrossRef] [PubMed]
SchmuckerC, SchaeffelF. A paraxial schematic eye model for the growing C57BL/6 mouse. Vision Res. 2004;44:1857–1867. [CrossRef] [PubMed]
SchmuckerC, SchaeffelF. In vivo biometry in the mouse eye with low coherence interferometry. Vision Res. 2004;44:2445–2456. [CrossRef] [PubMed]
TejedorJ, de la VillaP. Refractive changes induced by form deprivation in the mouse eye. Invest Ophthalmol Vis Sci. 2003;44:32–36. [CrossRef] [PubMed]
JaissleGB, MayCA, ReinhardJ, et al. Evaluation of the rhodopsin knockout mouse as a model of pure cone function. Invest Ophthalmol Vis Sci. 2001;42:506–513. [PubMed]
ClaesE, SeeligerMW, MichalakisS, BielM, HumphriesP, HaverkampS. Morphological characterization of the retina of a CNGA3−/−RHO−/− mutant mouse lacking functional cones and rods. Invest Ophthalmol Vis Sci. 2004;45:2039–2048. [CrossRef] [PubMed]
BarnardAR, ApplefordJM, SekaranS, et al. Residual photosensitivity in mice lacking both rod opsin and cone photoreceptor cyclic nucleotide gated channel 3 subunit. Vis Neurosci. .In press.
HumphriesMM, RancourtD, FarrarGJ, et al. Retinopathy induced in mice by targeted disruption of the rhodopsin gene. Nat Genet. 1997;15:115–117. [CrossRef] [PubMed]
BielM, ZongX, LudwigA, SautterA, HofmannF. Structure and function of cyclic nucleotide-gated channels. Rev Physiol Biochem Pharmacol. 1999;135:151–171. [PubMed]
BielM, SeeligerMW, PfeiferA, et al. Selective loss of cone function in mice lacking the cyclic nucleotide-gated channel CNG3. Proc Natl Acad Sci USA. 1999;96:7553–7557. [CrossRef] [PubMed]
PandaS, ProvencioI, TuDC, et al. Melanopsin is required for non-image forming photic responses in blind mice. Science. 2003;25:525–527.
HattarS, LucasRJ, MrosovskyN, et al. Melanopsin and rod-cone photoreceptive systems account for all major accessory visual functions in mice. Nature. 2003;424:76–81. [PubMed]
MitchinerJC, PintoLH, VanableJW. Visually evoked eye movements in the mouse. Vision Res. 1976;16:1169–1171. [CrossRef] [PubMed]
PerryVH, CoweyA. The ganglion cell and cone distributions in the monkey’s retina: implications for central magnification factors. Vision Res. 1985;25:1795–1810. [CrossRef] [PubMed]
PettigrewJD, DreherB, HopkinsCS, et al. Peak density and distribution of ganglion cells in the retinae of microchiropteran bats: implications for visual acuity. Brain Behav Evol. 1988;32:39–56. [CrossRef] [PubMed]
PorciattiV, PizzorussoT, CenniMC, MaffeiL. The visual response of retinal ganglion cells is not altered by optic nerve transection in transgenic mice overexpressing Bcl-2. Proc Natl Acad Sci USA. 1996;93:14955–14959. [CrossRef] [PubMed]
RossiFM, PizzorussoT, PorciattiV, MarubioLM, MaffeiL, ChangeuxJP. Requirement of the nicotinic acetylcholine receptor β2 subunit for the anatomical and functional development of the visual system. Proc Natl Acad Sci USA. 2001;98:6453–6458. [CrossRef] [PubMed]
GianfranceschiL, SicilianoR, WallsJ, et al. Visual cortex is rescued from the effects of dark rearing by overexpression of BDNF. Proc Natl Acad Sci USA. 2003;100:12486–12491. [CrossRef] [PubMed]
HuangZJ, KirkwoodA, PizzorussoT, et al. BDNF regulates the maturation of inhibition and the critical period of plasticity in mouse visual cortex. Cell. 1999;98:739–755. [CrossRef] [PubMed]
DragerUC. Receptive fields of single cells and topography in mouse visual cortex. J Comp Neurol. 1975;160:269–290. [CrossRef] [PubMed]
ManginiNJ, PearlmanAL. Laminar distribution of receptive field properties in the primary visual cortex of the mouse. J Comp Neurol. 1980;193:203–222. [CrossRef] [PubMed]
GordonJA, StrykerMP. Experience-dependent plasticity of binocular responses in the primary visual cortex of the mouse. J Neurosci. 1996;16:3274–3286. [PubMed]
McGillTJ, DouglasRM, LundRD, PruskyGT. Quantification of spatial vision in the royal college of surgeons rat. Invest Ophthalmol Vis Sci. 2004;45:932–936. [CrossRef] [PubMed]
BlakeR, CoolS, CrawfordM. Visual resolution in the cat. Vision Res. 1974;14:1211–1218. [CrossRef] [PubMed]
NyePW. The binocular acuity of the pigeon measured in terms of the modulation transfer function. Vision Res. 1976;8:1041–1053.
MeriganWH. The contrast sensitivity of the squirrel monkey. Vision Res. 1976;16:375–379. [CrossRef] [PubMed]
De ValoisR, MorganH, SnodderlyD. Psychophysical studies of monkey vision III. Spatial luminance contrast sensitivity tests of macaque and human observers. Vision Res. 1974;14:75–81. [CrossRef] [PubMed]
DiepesH. Refraktionsbestimmung. 1975; 2nd ed. 82–85.Bode Pforzheim Germany.
Herreros de TejadaP, Munoz TedoC, CostiC. Behavioral estimates of absolute visual threshold in mice. Vision Res. 1997;37:2427–2432. [CrossRef] [PubMed]
Carter-DawsonLD, La VailMM. Rods and cones in the mouse retina I. Structural analysis using light and electron microscopy. J Comp Neurol. 1979;188:245–262. [CrossRef] [PubMed]
PennesiME, LyubarskiAL, PughEN. Extreme responsiveness of the pupil of the dark-adapted mouse to steady retinal lumination. Invest Ophthalmol Vis Sci. 1998;39:2148–2156. [PubMed]
DragerUC. Observations on monocular deprivation in mice. J Neurophysiol. 1978;41:28–42. [PubMed]
Figure 1.
 
Setup for the behavioral experiment to test spatial acuity in mice. Mice were placed individually in a Perspex container. To measure behavioral response under dim illumination or in darkness, the mouse was illuminated by two high-powered IR LEDs. An IR light-sensitive video camera imaged the mouse and, after digitization of the video frames, a screen output as shown in Figure 2was obtained. The pattern of vertical black and white stripes that was placed inside the drum was made from clear plastic foil. Spatial frequency could be varied by placing stripe patterns with different stripe width inside the drum. The drum was illuminated either by a light bulb or a white-light LED. In the latter case, a frosted diffuser was placed in front of the LED to generate a homogenous illumination.
Figure 1.
 
Setup for the behavioral experiment to test spatial acuity in mice. Mice were placed individually in a Perspex container. To measure behavioral response under dim illumination or in darkness, the mouse was illuminated by two high-powered IR LEDs. An IR light-sensitive video camera imaged the mouse and, after digitization of the video frames, a screen output as shown in Figure 2was obtained. The pattern of vertical black and white stripes that was placed inside the drum was made from clear plastic foil. Spatial frequency could be varied by placing stripe patterns with different stripe width inside the drum. The drum was illuminated either by a light bulb or a white-light LED. In the latter case, a frosted diffuser was placed in front of the LED to generate a homogenous illumination.
Figure 2.
 
Screen dump of the C++ program that tracked the mouse. The program tracked the movement of the center of mass of the mouse, marked by the cross with the circle. The trace of movement is shown on the right. The average angular velocity of the center of mass of the mouse (Average Running Speed) in relation to the center of the container was summed up over time, and the standard deviation of all angular changes was calculated after termination of the measurement session (approximately after 20 seconds). The angular movement of the snout-tail axis (see dotted line on the mouse image) was also recorded (Angular Orientation Speed). Finally, the Average Activity was recorded as the average absolute angular position change from one frame to the next.
Figure 2.
 
Screen dump of the C++ program that tracked the mouse. The program tracked the movement of the center of mass of the mouse, marked by the cross with the circle. The trace of movement is shown on the right. The average angular velocity of the center of mass of the mouse (Average Running Speed) in relation to the center of the container was summed up over time, and the standard deviation of all angular changes was calculated after termination of the measurement session (approximately after 20 seconds). The angular movement of the snout-tail axis (see dotted line on the mouse image) was also recorded (Angular Orientation Speed). Finally, the Average Activity was recorded as the average absolute angular position change from one frame to the next.
Figure 3.
 
Responses of C57BL/6 wild-type mice in a stationary white drum (A) and in a rotating white drum without the stripe pattern inside (B). There was no significant preference of the mice to move in any particular direction (P > 0.15, variance ratio test). The locomotor activity of the mice was not significantly different at different luminances (P > 0.10, variance ratio test). (C) Stationary drum without stripe pattern. Each bar graph shows the mean ± SD of results from at least three animals.
Figure 3.
 
Responses of C57BL/6 wild-type mice in a stationary white drum (A) and in a rotating white drum without the stripe pattern inside (B). There was no significant preference of the mice to move in any particular direction (P > 0.15, variance ratio test). The locomotor activity of the mice was not significantly different at different luminances (P > 0.10, variance ratio test). (C) Stationary drum without stripe pattern. Each bar graph shows the mean ± SD of results from at least three animals.
Figure 4.
 
Mean optomotor responses and their standard deviations of C57BL/6 wild-type mice at different spatial frequencies. (A) Angular running speed and (B) angular orientation speed. The horizontal error bar illustrates the uncertainty of the spatial frequency variable, originating from the fact that the mice could change the viewing angle by moving closer to the stripe pattern. Data from seven animals are shown, with three or more animals tested at each data point. Angular running speed had more statistical power (P < 0.005, variance ratio test). The responses of the mice were significantly different from the condition in which no visual stimuli were present, for spatial frequencies up to 0.1 cyc/deg at 30 cd/m2 (P < 0.05, Dunnett test). If compared with the null hypothesis, responses were significantly different from zero up to 0.3 cyc/deg (P < 0.05, variance ratio test). The responses at 0.1 cd/m2 were decreased, but not significantly different from the responses at 30 cd/m2 (P > 0.05, Dunnett test). Responses at 0.005 cd/m2 were significantly reduced compared to the response at 30 cd/m2 (P < 0.05, Dunnett test). Responses in complete darkness were neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the response when no visual stimuli were present (P > 0.005, Dunnett test).
Figure 4.
 
Mean optomotor responses and their standard deviations of C57BL/6 wild-type mice at different spatial frequencies. (A) Angular running speed and (B) angular orientation speed. The horizontal error bar illustrates the uncertainty of the spatial frequency variable, originating from the fact that the mice could change the viewing angle by moving closer to the stripe pattern. Data from seven animals are shown, with three or more animals tested at each data point. Angular running speed had more statistical power (P < 0.005, variance ratio test). The responses of the mice were significantly different from the condition in which no visual stimuli were present, for spatial frequencies up to 0.1 cyc/deg at 30 cd/m2 (P < 0.05, Dunnett test). If compared with the null hypothesis, responses were significantly different from zero up to 0.3 cyc/deg (P < 0.05, variance ratio test). The responses at 0.1 cd/m2 were decreased, but not significantly different from the responses at 30 cd/m2 (P > 0.05, Dunnett test). Responses at 0.005 cd/m2 were significantly reduced compared to the response at 30 cd/m2 (P < 0.05, Dunnett test). Responses in complete darkness were neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the response when no visual stimuli were present (P > 0.005, Dunnett test).
Figure 5.
 
Optomotor responses of C57BL/6 wild-type mice at different spatial frequencies, for angular running speed (A) and angular orientation speed (B), as measured in the small drum. Error bars are as in Figure 4 . Data from 15 animals contributed to the curve, with 5 or more animals for each data point. Mice showed significant responses for spatial frequencies up to 0.5 cyc/deg. Eight of 10 responses were significantly different from zero (P < 0.05, variance ratio test) and significantly different from the responses when no visual stimuli were present (P < 0.05, Dunnett test).
Figure 5.
 
Optomotor responses of C57BL/6 wild-type mice at different spatial frequencies, for angular running speed (A) and angular orientation speed (B), as measured in the small drum. Error bars are as in Figure 4 . Data from 15 animals contributed to the curve, with 5 or more animals for each data point. Mice showed significant responses for spatial frequencies up to 0.5 cyc/deg. Eight of 10 responses were significantly different from zero (P < 0.05, variance ratio test) and significantly different from the responses when no visual stimuli were present (P < 0.05, Dunnett test).
Figure 6.
 
Optomotor responses of Rhodopsin knockout (RHO−/−) mice are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At 30 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test). At 0.1 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test) nor significantly different from the responses when no visual stimuli were present (P > 0.05, Dunnett test).
Figure 6.
 
Optomotor responses of Rhodopsin knockout (RHO−/−) mice are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At 30 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test). At 0.1 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test) nor significantly different from the responses when no visual stimuli were present (P > 0.05, Dunnett test).
Figure 7.
 
Optomotor responses of mice lacking rod function (CNGB1−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars as in Figure 4 . At 30 cd/m2, the mice showed significant responses at 0.03 and 0.05 cyc/deg (P < 0.05, Dunnett test). At both 0.1 and 0.005 cd/m2, responses were random (P > 0.05, Dunnett test).
Figure 7.
 
Optomotor responses of mice lacking rod function (CNGB1−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars as in Figure 4 . At 30 cd/m2, the mice showed significant responses at 0.03 and 0.05 cyc/deg (P < 0.05, Dunnett test). At both 0.1 and 0.005 cd/m2, responses were random (P > 0.05, Dunnett test).
Figure 8.
 
Optomotor responses of mice lacking cone function (CNGA3−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . In this mutant, the angular orientation speed provided higher significances (P < 0.0007, variance ratio test). At 30 cd/m2, the mice showed significant responses up to 0.2 cyc/deg (P < 0.05, Dunnett test). At 0.005 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test).
Figure 8.
 
Optomotor responses of mice lacking cone function (CNGA3−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . In this mutant, the angular orientation speed provided higher significances (P < 0.0007, variance ratio test). At 30 cd/m2, the mice showed significant responses up to 0.2 cyc/deg (P < 0.05, Dunnett test). At 0.005 cd/m2, the mice showed significant responses up to 0.1 cyc/deg (P < 0.05, Dunnett test).
Figure 9.
 
Optomotor responses of mice lacking both rod and cone function (CNGA3−/−RHO−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At both 30 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the responses without visual stimulation (P > 0.05, Dunnett test).
Figure 9.
 
Optomotor responses of mice lacking both rod and cone function (CNGA3−/−RHO−/−) are plotted against spatial frequency for angular running speed (A) and angular orientation speed (B). Error bars are as in Figure 4 . At both 30 and 0.005 cd/m2, the responses were randomly distributed and neither significantly different from zero (P > 0.05, variance ratio test), nor significantly different from the responses without visual stimulation (P > 0.05, Dunnett test).
Table 1.
 
Relative Responses in Terms of Angular Running Speed in Wild-type and Knockout Mice
Table 1.
 
Relative Responses in Terms of Angular Running Speed in Wild-type and Knockout Mice
Response (%) % Response When No Visual Stimulus Was Present (Rotating Drum) % Response When No Visual Stimulus Was Present (Stationary Drum)
30 cd/m2 0.1 cd/m2 0.005 cd/m2 0 cd/m2
Wild-type 100 76 46 −9 18 3
RHO−/− 24 −1 −14
CNGB1−/− 41 −11 −4
CNGA3−/− 70 8
CNGA3−/−RHO−/− 15 15
Table 2.
 
Relative Responses in Terms of Angular Orientation Speed in Wild-type and Knockout Mice
Table 2.
 
Relative Responses in Terms of Angular Orientation Speed in Wild-type and Knockout Mice
Response (%) % Response When No Visual Stimulus Was Present (Rotating Drum) % Response When No Visual Stimulus Was Present (Stationary Drum)
30 cd/m2 0.1 cd/m2 0.005 cd/m2 0 cd/m2
Wild-type 100 48 41 12 −29 33
RHO−/− 74 7 −53
CNGB1−/− 60 −7 −3
CNGA3−/− 139 82
CNGA3−/−RHO−/− 24 24
×
×

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.

×