Rapid Quantification of Adult and Developing Mouse Spatial Vision Using a Virtual Optomotor System

Abstract

purpose. To develop a simple, rapid method of quantifying the spatial vision of mice.

methods. A rotating cylinder covered with a vertical sine wave grating was calculated and drawn in virtual three-dimensional (3-D) space on four computer monitors facing to form a square. C57BL/6 mice standing unrestrained on a platform in the center of the square tracked the grating with reflexive head and neck movements. The spatial frequency of the grating was clamped at the viewing position by repeatedly recentering the cylinder on the head. Acuity was quantified by increasing the spatial frequency of the grating until an optomotor response could not be elicited. Contrast sensitivity was measured at spatial frequencies between 0.03 and 0.35 cyc/deg.

results. Grating acuity was measurable on the day of eye opening (postnatal day [P]15: mean acuity, 0.031 cyc/deg) and reached a maximum (∼0.4 cyc/deg) by P24. A peak in the contrast sensitivity function emerged on P16 (4.7, or 21% contrast at 0.064 cyc/deg). The peak remained at 0.064 cyc/deg and climbed to a maximum sensitivity of 24.5, or 4% contrast, by P29. Acuity was obtained in each mouse in <10 minutes, and a detailed contrast sensitivity curve was generated in approximately 30 minutes.

conclusions. The virtual optomotor system provides a simple and precise method for rapidly quantifying mouse vision. Behavioral measures of vision in mice are essential for interpreting the results of experiments designed to reveal the cellular and molecular mechanisms of vision and visual development and for evaluating potential treatments for visual diseases.

Numerous studies have made use of genetically altered mice in investigations of the function of the visual system ^{1 2 3 4} and the mechanisms of developmental visual plasticity, ^{5 6} and many mouse models of visual system diseases have been produced and studied. ^{7 8 9} Despite the growing use of mice in vision research, little basic information is available on the spatial vision of mice or on how experimental manipulations of the visual system affect the mouse’s vision. A major reason for this deficit is the lack of fast and simple behavioral methods to quantify vision in the mouse. ¹⁰  

Reinforcement-based visual discrimination tasks have been developed for quantifying the spatial vision of mice. ^{11 12} Although these tasks have been used to measure the effects of experimental manipulations on the mouse’s vision, ^{12 13 14} they require a substantial investment of time (≥1 week) to generate valid psychophysical thresholds, and their use is effectively limited to juvenile-adult mice because younger animals apparently lack the cognitive capacity to learn the tasks quickly. 

Tests of optomotor responses in mice hold the promise of overcoming some of the limitations of visual discrimination tasks, because they do no not require reinforcement training to measure vision. Optomotor responses in rodents have been studied for some time, with a mechanical apparatus ¹⁵ that consists of a drum with printed stimuli on the inside wall and that rotates around the animal. The difficulty of controlling the speed of the drum and its position in relation to the animal combined with the problem of printing precise visual stimuli and exchanging them rapidly, however, have precluded using this device for the measurement of visual thresholds in the mouse. 

As a result of the limitations of existing methods, no behavioral measure of vision have been used at all in most of the experimental studies of the mouse’s vision. There is also little basic information available on normal visual behavior in the mouse, and mice with genetic modifications affecting the visual system are being produced but not screened for vision. At a minimum, behavioral measures of spatial vision (acuity and contrast sensitivity) are needed. In addition, many of the mutations and manipulations in mice that target the visual system have developmental consequences, and interpretations of the experiments would benefit from knowledge about the early development of vision. 

We addressed the need for a simple and rapid method of quantifying the mouse’s spatial vision by developing a virtual-reality optomotor system. We use this methodology to measure the grating acuity and contrast sensitivity of adult and developing mice from the day of eye opening. 

Materials and Methods

Animals

C57BL/6 mice, originally obtained from the Jackson Laboratory (Bar Harbor, ME), were used in the study. The C57BL/6 strain was chosen because of its wide use in laboratory studies and because it is a popular background strain for evaluating the behavioral consequences of transgenic mutations. Breeding pairs and their pups were housed in Plexiglas cages (46 × 26 × 16 cm [L × W × H]) in a room with an ambient temperature of 21°C, 35% relative humidity, a 12-hr light–dark cycle, and where food and water were available ad libitum. Soon after the pups were born, the male parent was removed from the cage. Litters were weaned at 25 days of age, at which time the males and females were separated and group housed under identical circumstances. All experimental procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were authorized by the University of Lethbridge Animal Care Committee, which approves only procedures that are conducted in accordance with the standards of the Canadian Council on Animal Care. 

Virtual Optomotor System

A virtual cylinder comprising a vertical sine wave grating was projected in three-dimensional (3-D) coordinate space on computer monitors arranged in a quadrangle (square) around a testing arena. The testing arena consisted of a Plexiglas box (39 × 39 × 32.5 cm [L × W × H]) with rectangular openings (33.5 × 26.5 cm [W × H]) on each wall painted flat white on the inside. A mirror with a small hole in its center was placed tangentially to the bottom of the openings. A platform was positioned 13 cm above the floor by securing a white Plexiglas disc (diameter, 5.3 cm) to a threaded bolt that passed through the hole in the mirror to the bottom of the apparatus, where it was attached with a nut. A mirror with a large central access hole (diameter, 25.3 cm) was also situated tangentially to the top of the openings. A vented and hinged lid (30.5 × 30.5 cm) enclosed the top of the apparatus. A camera (FireWire iSight; Apple Computer Corp., Mountain View, CA) was positioned directly above the platform by attaching it to the lid with a holding sleeve. One of four 17-in. LCD computer monitors (model 1703FP; Dell, Phoenix, AZ) was attached to each outside wall of the apparatus, so that all monitors projected through the rectangular openings into the arena. Whisper fans were used to cool the monitors and vent the testing arena. Figure 1 shows a schematic representation of the apparatus, which was located on a table in a dimly lit, quiet room. 

A computer program (OptoMotry; CerebralMechanics, Lethbride, Alberta, Canada), running on a dual processor (G4 or G5 Power Macintosh; Apple Computer Corp.), was used to drive video cards (OpenGL-compatible 7000 Mac Edition; ATI Radeon, Markham, Ontario, Canada) and project on the monitors a virtual cylinder in 3-D coordinate space (Fig. 2A) . The gamma response of the monitors was used to linearize the output to the screens, and the screen luminance was adjusted to equalize the screen intensity (black mean, 0.22 cd/m²; white mean, 152.13 cd/m²). Visual stimuli were drawn on the walls of the cylinder, the image of which was extended by the floor and ceiling mirrors. Therefore, from the perspective of the platform, each monitor appeared as a window on a surrounding 3-D world. The software also controlled the speed of rotation and geometry of the cylinder and the spatial frequency and contrast of the stimuli and enabled live video feedback of the testing arena. Movie 1 (see) shows examples of mice behaving in the apparatus. The mice occasionally repositioned themselves, but between these large movements, the head and upper body slowly rotated, following the grating pattern on the screens. A red crosshair in the video frame indicated the center of the cylinder rotation. 

Behavioral Testing

Mice were placed one at a time on the platform, the lid of the box was closed, and the animals were allowed to move freely. As the mouse moved about the platform, the experimenter followed the mouse’s head with a crosshair superimposed on the video image. The x-y positional coordinates of the crosshair were used to center the rotation of the cylinder at the mouse’s viewing position, thereby maintaining the virtual walls of the cylinder at a constant distance from the animal and effectively “clamping” the spatial frequency of the grating (Fig. 1C) . When a grating perceptible to the mouse was projected on the cylinder wall and the cylinder was rotated (12 deg/sec), the mouse normally stopped moving its body and would begin to track the grating with reflexive head movements in concert with the rotation (Figs. 2B 2C) . An experimenter assessed whether the animals tracked the cylinder by monitoring in the video window the image of the cylinder, the animal, and the crosshair simultaneously. If the mouse’s head tracked the cylinder rotation, which was evident as movement against the stationary arms of the crosshair, it was judged that the animal could see the grating. 

If, during the course of testing an animal slipped or jumped off the platform, it was simply returned to the platform and testing was resumed. Whenever possible, experimenters were blind to the treatment and age of the animals, as well as to the animal’s previously recorded thresholds. All animals were habituated before the outset of testing with handling and by placing them on the platform for a few minutes at a time. The mice were generally tested during the first few hours of their daylight cycle (12-hour light-dark; light on at 7 AM), normally for 5 to 30 minutes at a time. 

Determination of Visual Thresholds

When measuring grating acuity, we projected a homogeneous gray stimulus on the cylinder at the beginning of each testing session. After placing the animal on the platform and closing the lid, the experimenter waited until the animal stopped moving, at which time the gray was replaced with a low-spatial-frequency (∼0.1 cyc/deg) sine wave grating (100% contrast) of the same mean luminance and moving in one direction. The animal was assessed for tracking behavior for a few seconds, and then the gray stimulus was restored. This procedure was repeated until unambiguous tracking was observed. The short testing epochs reduced the possibility of the mouse’s adapting to the stimulus and established that each animal was capable of tracking when a salient stimulus was present, and initiating the testing with a low-spatial-frequency grating enabled each mouse’s optomotor response to be typified. Using either a staircase or method-of-limits procedure, we then systematically increased the spatial frequency of the grating until the animal no longer responded. Occasionally, during testing, sudden reversals of grating drift direction, sudden changes in luminance (e.g., jumps to black or white), squeaking noises, or taps on the lid were interspersed with the grating presentations to induce the animal to stop moving, which facilitated more rapid testing. The process of incrementally changing the spatial frequency of the test grating was repeated a few times until the highest spatial frequency that the mouse could track was identified as the threshold. A threshold for each direction of rotation was assessed this way, and the highest spatial frequency tracked in either direction was recorded as the threshold. 

A contrast-sensitivity function was assessed by using the general procedures just described. The differences included that testing at a spatial frequency began with a grating of 100% contrast, which was then systematically reduced until the contrast threshold was identified. In addition, a contrast threshold was identified at six spatial frequencies between 0.03 and 3.5 cyc/deg (0.031, 0.064, 0.092, 0.103, 0.192, 0.272 cyc/deg). The threshold at a spatial frequency was calculated as a Michelson contrast from the screen’s luminances (maximum –minimum)/(maximum + minimum). The contrast sensitivity (the reciprocal of the threshold) was then plotted against spatial frequency on a log–log graph. 

Timeline of Experiments

The acuity and contrast sensitivity of mice from two separate litters was measured daily from the day of eye opening (postnatal day [P]15) to P35 and regularly thereafter into adulthood (P90–P125). 

Results

Qualitative Observations

In the course of developing the mouse virtual optomotor system, we investigated the utility of several different design features of the apparatus, software, and testing protocols. In doing so, we found that one of the most important features of the methodology was that mice were allowed to move freely on the platform. In our initial designs, we used a Plexiglas cylinder to create vertical walls around the testing platform to confine the mice. This was more of a hindrance than it was a benefit, however, because animals would spend a large amount of time in a testing session investigating the cylinder walls with their paws and vibrissa, which reduced the number of tracking episodes. We were also concerned that the Plexiglas cylinder might distort the image of the gratings for the animal and increase the variability of our threshold measurements. By adopting an exposed platform with no barriers between the animal and the monitors, we reduced the number of distractions for the animals and the possible distortion of the visual stimulus. 

The size of the testing platform was also an important variable in the procedures. If, for example, we used a large platform, the mice would persistently walk around the perimeter parallel to the cylinder. Because tracking could only be judged when the mice were not moving, the large platform reduced the number of occasions in a testing session when the animals tracked the moving grating. Conversely, if a very small platform was used, animals were preoccupied with balancing in position, and as a result, would often jump off. The 5.3-cm platform we adopted was a compromise that seemed to work best; it was small enough to allow the mice to stand securely and move freely, but it restricted movements to pivots and forced the animals to face the screens with the head extended beyond the edge of the platform, all of which facilitated the measurement of optomotor responses. The relatively small platform also decreased the potential for the vibrissal contact with the surface that might provide contradictory somatosensory feedback to the animal that the substrate was not moving when the visual world was rotated. 

We also tested the need for using mirrors to extend the screen images. Mice would readily pursue a rotating cylinder displayed on the monitors alone; however, the mirrors noticeably increased the overall frequency of tracking episodes, most likely because the reflected stimulus then occupied a greater proportion of their visual field. In addition, the visual cliff provided by the reflection of the cylinder reduced the number of times the mice would jump down to the floor of the testing arena. 

Centering the rotation of the cylinder at the animal’s viewing position was also shown to be an important feature of the system. Its effectiveness could easily be seen, when, after determining the grating threshold with the animal’s head a few centimeters from the edge of the platform, the center of the rotation was quickly moved to the center of the platform. Under these circumstances, a stimulus that would not elicit tracking when the rotation was centered on the animal’s head would cause the animal to track, undoubtedly because the spatial frequency of the grating was decreased by “apparently” moving the cylinder closer to the animal. Consequently, systems that do not have a tracking feature would produce more variable results, because the cylinder would not be maintained at a constant “virtual” distance from the animal’s eyes. 

Although testing could proceed rapidly in the virtual optomotor system, the duration of the assessment period was limited by the tolerance of the animal for staying relatively calm on the platform and not jumping down to the floor. In most cases, animals could be tested for 10 to 15 minutes at a time without difficulty; more than enough time to measure acuity. However, animals from time to time became restless and endeavored to get off the platform. If this occurred, we found that it could be remedied by returning the animal to its holding cage for 10 to 15 minutes. In general then, a flexible approach to testing the animals with close attention paid to their behavior is the best procedure for generating reliable data. 

Finally, we found that the most efficient procedure for generating thresholds was to have the observer judge whether the animal tracked or not, but in principle, this could lead to experimenter bias in the results. To check this, we had two experienced observers test the same animals. In addition, we tested some animals in a blind fashion, where the observer could not see the direction of motion and simply recorded the direction of tracking. In all cases, the visual thresholds were not different. 

Quantification of Visual Thresholds

All animals in this study proficiently tracked a rotating grating from the day of eye opening onward, and grating and contrast thresholds were readily generated at all ages tested. Figure 3 displays graphically the development of grating acuity in 17 mice that were measured from P15 to adulthood. On P15, the average acuity was near 0.04 cyc/deg. The threshold increased rapidly over the next 2 weeks to near 0.4 cyc/deg by P24 and did not change appreciably thereafter. There was remarkably little variability in the measures between animals on any given day, and once acuity reached a maximum, there was great measurement consistency and stability for each animal, notwithstanding that the experimenter was blind to the values generated previously. At low spatial frequencies, robust (1–2 cm; 0.5–2 seconds), sweeping head and neck movements were generated, but as the spatial frequency of the grating was increased, the extent and duration of the movements decreased until at threshold, no movements were generated. Therefore, detecting movements near threshold consisted of identifying small (<0.5 cm) and brief (<0.5 second) tracking episodes. 

Contrast sensitivity also developed rapidly after eye opening. On P15, at only 0.031 could a contrast threshold (1.51) be recorded. By P16, however, thresholds could be measured at spatial frequencies up to 0.103 cyc/deg and an inverted U-shaped curve emerged with a peak (4.66) at 0.064 cyc/deg, which remained as the peak of the contrast sensitivity curve throughout development, but the sensitivity at each spatial frequency developed with an idiosyncratic pattern. Overall, contrast sensitivity increased until P30, after which no changes were distinguished in the curve. In a similar fashion to that observed when measuring acuity, there was little variability between the thresholds of individual mice, and there was great reproducibility in the measures. In addition, the magnitude of the optomotor response decreased as the contrast of the grating decreased until the contrast threshold was reached, at which point no movements were generated. Figure 4A shows graphically a sample of contrast sensitivity curves generated by 11 mice between P16 and adulthood. Figure 4B displays a detailed plot of the developmental changes in contrast sensitivity at each spatial frequency tested. 

Discussion

Comparison with Existing Methodologies

We are aware of only one study measuring the visual tracking capabilities of mice by using a mechanical optomotor device. ¹⁶ Because that study did not measure visual thresholds, just the amount of turning, it is difficult to compare the results directly with those of our study. 

Other studies, including our own, have obtained threshold measurements in mice by using simple discrimination tasks. There are several important differences, however, between the virtual optomotor task and reinforcement-based tasks in the measurement of mouse visual thresholds. One difference is that highly reliable grating thresholds can be generated in the virtual optomotor task in a matter of a few minutes, both in adults and young animals from the day of eye opening. By comparison, more than 1 week is needed to generate a valid grating threshold in mature animals in the visual water task, ^{13 14} and it is difficult to train and test mice younger than 1 month of age. The grating thresholds obtained from reinforcement-based tasks, however, are consistently higher than those generated in the virtual optomotor task. For example, we report herein that maximum grating acuity is near 0.4 cyc/deg, but Gianfranceschi et al., ¹² using a terrestrial visual discrimination task, and ourselves, using the visual water task, ^{11 13 14} have reported maximum values between 0.5 and 0.6 cyc/deg. 

We believe there are at least two possible explanations for the discrepancy in the values generated with the virtual optomotor task and visual discrimination tasks. The first is that the visual pathways subserving behavioral responses in optomotor and reinforcement-based tasks process different features of the retinal output, and therefore, reflect different measures of vision. There are clear parallels between the measurement of optokinetic eye movements and head tracking in the virtual optomotor task, and it is known that optokinetic eye movements are largely driven by subcortical, low-frequency visual pathways. Mouse grating acuity generated in the visual water task, however, is significantly reduced by lesions of V1 (manuscript submitted). Consequently, it is possible that our head-tracking thresholds measure the function of subcortical pathways whereas, the visual water task measures cortical vision. The second possibility is that differences in grating thresholds between optomotor and reinforcement-based tasks reflect differences in behavioral responses. Because the optomotor response is graded with less vigorous tracking near threshold, it is possible that a weak sensory signal is present without any apparent behavioral response. In either case, the differences in the measures show that discrimination-based tasks and optomotor tasks are not interchangeable, and prudence should be exercised when comparing threshold values generated with the different methodologies. 

Mechanisms Underlying the Development of Thresholds

The mechanisms underlying the development of grating acuity and contrast sensitivity were not specifically investigated in the present study, but any interpretations of the data are complicated by the fact that the development of visual thresholds could be influenced by changes in the visual system, the motor system, or both. In terms of the visual system, however, several inferences can be drawn from the developmental patterns of grating acuity and contrast sensitivity. First, head-tracking is present on the day of eye opening, indicating that the visuomotor circuitry is already in place when high-quality visual information normally becomes available. Second, both grating acuity and contrast sensitivity develop rapidly over the 2 weeks after eye opening, suggesting that visuomotor circuitry requires experience to develop fully. Third, the time course for the maturation of contrast sensitivity is different at different spatial frequencies. For example, 0.064 cyc/deg, the peak of the contrast sensitivity curve at all ages tested, reaches its maximum at 28 days. In contrast, 0.031 cyc/deg reaches a maximum contrast sensitivity much earlier, and all the other spatial frequencies tested reach mature values after P30 (Fig. 4B) . This suggests that there are multiple spatial frequency channels for visual contrast sensitivity. 

Conclusions

The virtual optomotor task enables spatial visual thresholds to be measured rapidly and without specific reinforcement training. This makes it possible to measure vision in mice from the day of eye opening and thereby enables lifespan measures of vision. The developmental profiles of acuity and contrast sensitivity present a dynamic picture of emerging function in the mouse visual system, suggesting that a wide range of visual phenotypes in mice may be identifiable with the task. In addition, the electronic generation of stimuli, which made the rapid switching of frequency, direction, and contrast possible and was essential for the success of the present study, opens the door to many other visual tests ^{17 18} in mice. 

 Supported by a grant from the Natural Sciences and Engineering Research Council of Canada.

 Submitted for publication May 14, 2004; accepted July 4, 2004.

 Disclosure: G.T. Prusky, CerebralMechanics (I); N.M. Alam, None; S. Beekman, None; R.M. Douglas, CerebralMechanics (I)

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

 

Figure 1.

 

Schematic representation of the optomotor testing apparatus. (A) Side view. A mouse is placed on a platform positioned in the middle of an arena created by a quad-square of computer monitors. Sine wave gratings drawn on the screens are extended vertically with floor and ceiling mirrors. A video camera is used to monitor the animal’s behavior from above. (B) Top view. The mouse is surrounded by 360° of gratings and is allowed to move freely on the platform.

Figure 2.

 

Virtual geometry and optomotor response. (A) A virtual cylinder is projected in 3-D coordinate space on the monitors. The head of the mouse centers the rotation of the cylinder. (B) When the cylinder is rotated, the mouse tracks the drifting grating with head and neck movements. (C) A single-frame video camera image of a mouse tracking the cylinder grating. The four-line crosshair is positioned between the eyes of the mouse, and the coordinates are used to center the rotation of the cylinder.

Figure 3.

 

Development of grating acuity. On the day of eye-opening (P15) the average grating threshold of 17 mice was near 0.04 cyc/deg. The acuity increased rapidly over the following 10 days and maximized near 0.4 cyc/deg at P25. No changes in acuity were detected thereafter. Vertical bars represent ± SEM and are often not visible, as they are smaller than the symbols.

Figure 4.

 

Development of contrast sensitivity. (A) A selection of contrast sensitivity curves generated from mice aged 16 days to adult. Contrast sensitivity increased as the mice aged until reaching its maximum around 1 month of age. At 16 days, 0.064 cyc/deg emerged as the spatial frequency with the highest contrast sensitivity and remained the peak at all subsequent ages. Each point represents the average of 11 mice. Vertical bars, ±SEM. (B) Contrast sensitivity values at all ages measured, plotted separately for each of the six spatial frequencies tested. Each spatial frequency had a distinct developmental profile.

Supplementary Materials

Movie 1 - 7.39 MB 

The authors thank Virnell Frandsen and Leslie Nelson for their valuable assistance.