Non-aversive Animal Restraint Enabling Recording of Optomotor Reflex in Ground Squirrels

Abstract

The optomotor reflex (OMR) provides a behavioral assessment of an animal’s contrast sensitivity and visual acuity. Mice or rats are typically placed directly onto a small circular platform by hand; however, handling animals like this can stimulate stress and anxiety, which introduce confounding factors when interpreting data. It has been shown that non-aversive handling methods, such as picking up mice or rats in a familiar tunnel/tube, can reduce anxiety. This is of particular interest in studies where animals display heightened stress, overactivity, or motor dysfunction, resulting in an inability to stay on the platform. A team led by Drs. Kiyoharu J. Miyagishima and Francisco M. Nadal-Nicolás have redesigned the conventional OMR platform to provide semi-closed containment. This makes it possible for the first time to record the optomotor reflex in the 13-lined ground squirrel, which is one of the few mammals that can see color. It has a visual streak with a high density of cones similar to the human macula providing an attractive model for studying effects on the cone visual system.

Introduction

Ground squirrels offer a unique and valuable model for studying visual function due to their diurnal lifestyle and cone-dominated retinas, which closely mimic human vision¹. This, coupled with the practical advantages of their use (smaller size, lower husbandry cost) compared to primates, make them an ideal model for translational research into retinal injury, disease, or ocular side effects of drugs on color vision changes^2,3,4,5,6. Preclinical research on retinal diseases relies on the ability to assess whether an animal can see. The optomotor reflex (OMR) offers a clinically relevant method for evaluating visual function in animals without the need for anesthesia, enabling an accurate measurement of visual acuity^7,8. In this test, the animal follows a rotating stripe pattern with its gaze. The rotating stripes induce the perception of global motion in the environment, triggering an involuntary head movement - the OMR. This reflex serves to visually stabilize the environment and can be used to quantify how well an animal can see by incrementally adjusting the spatial frequency or contrast of the stimulus. As the stimulus becomes more difficult to perceive, the OMR is eventually not triggered, allowing the determination of visual acuity or contrast threshold. The advantage of the OMR is that it is a reflex. Thus, no training is required, and head fixation or surgery is unnecessary. The reflex is present even in very young animals (upon eye-opening)^9,10, making it possible to study early-onset diseases. Additionally, since it is noninvasive, longitudinal measurements can be made enabling tracking of disease progression.

Most OMR testing is performed using custom or commercial systems that consist of a raised platform where the animal is placed, surrounded by four screens displaying rotating black and white stripe patterns. The stripes move in one direction (clockwise or counterclockwise), creating a motion stimulus that elicits a visual tracking response from the animal^{10,11,12,13,14,15,16}. A camera is mounted above the testing arena to capture the animal’s visually evoked head movements in response to the rotating stimulus. These head movements are tracked and analyzed to determine the animal’s ability to perceive and respond to motion, thereby assessing its visual acuity.

Although the OMR is routinely used in mice and rats, measurement of the optomotor response in 13-lined ground squirrels has been challenging due to their hyperactivity and inability to stay on elevated platforms, even with prior conditioning. To circumvent this, we developed a platform that provides semi-containment, facilitating the transfer of the animals to the arena. The goal of this method is to reduce the time for the animals to adapt to the experimental setting, enabling the recording of animals that display uncooperative behavior or have difficulty staying on the platform and improving the inclusion of animals in studies. We have also defined a protocol to measure visual acuity in squirrels using an experimentally defined contrast setting and have adjusted the camera settings for improved imaging results, enabling automated analysis and tracking.

Protocol

The procedures involving animals were conducted in adherence to the laws of the United States and the regulations of the Department of Agriculture. All experiments with animals were conducted according to protocols approved by the Animal and Care and Use Committee of the National Eye Institute at the National Institutes of Health. Male and female 13-lined ground squirrels (ages 6–8 months) used in this study were purchased from the University of Wisconsin Oshkosh Squirrel Colony. Mice (rd10: 1 month, CX3CR1^GFP/+: 6.5 months) used in this study were bred in our animal facility. Animals and their cages were randomly selected and assigned to experimental groups.

1. Preparation of the prototype platform and modified handling tunnel

NOTE: Round flat platforms are commonly employed in systems designed to measure the optomotor response¹³. However, hyperactive rodent models, such as squirrels, often struggle to maintain stability on these platforms. This design provides a semi-enclosed space, which helps the animals habituate quickly and reduces anxiety-like behaviors. This improvement enhances the animal’s ability to focus and minimizes instances of imbalance or falls during recordings.

1. In CAD drawing software (e.g., OpenSCAD or FreeCAD), design the platform (gray) and the enclosure (red) as shown in Figure 1.

   1. Ensure that the key dimensions for the squirrel platform are maintained as the length and width (L × W): 129 mm × 81 mm, the thickness: 7 mm, and the height of the wall that extends around the perimeter: 8 mm. Use the provided free CAD standard document file for the prototype platform (squirrel) used in this proof-of-concept study (Supplementary Coding File 1).

   2. Ensure that the key dimensions for the removable top enclosure (looking top down) are maintained as the length (119.5 cm) and width (76.2 cm), the height of the sides (50.8 mm), and the part’s thickness (3 mm). Use the provided free CAD standard document file to replicate the removable top enclosure (squirrel) employed in this proof-of-concept study (Supplementary Coding File 2).

2. Adjust the diameter of the through hole in the base of the platform as needed to fit the pedestal used to elevate the platform. The diameter pictured is 40 mm.

3. Print the platform using a white opaque filament. This provides maximal contrast for the camera, enabling the tracking software to correctly detect the animal’s contour on the platform.

4. Print the removable top enclosure with opaque dark filament (e.g., red). This provides a sense of shelter for the rodents, which facilitates guiding them into the tunnel before they are lifted and transferred to the arena.

5. Have a machinist cut the acrylic extruded clear square tube using a table saw so that the long end is ~ 101.6 mm in length. Be sure to use protective eye wear and a push stick to guide the acrylic through the blade while keeping your hands safe from potential injury.

6. Using the table saw or a handheld rotary tool equipped with an abrasive cutting wheel, remove one of the long rectangular faces. This creates an open top through which the camera mounted above will have an unobstructed view of the animal’s position and head movements.

7. Clean all edges with a hand file or sandpaper. Assemble the modified handling tunnel using the Acrylic Extruded Clear Square Tube cut to length with the 3D printed removable top enclosure by placing the enclosure on top of the open face.

Figure 1: Images of the computer-aided designs and photographs of the non-aversive platform and top enclosure used to transfer squirrels to the arena.

(A) Orthographic view of the squirrel platform designed to rest on top of the pedestal. (B) Bottom view of the squirrel platform. (C) Top view of the squirrel platform, rotated along the z-axis to highlight the depth of the edges. (D) Orthographic view of the 3D-printed enclosure serving as a removable lid for the acrylic tunnel. (E) Side view of the 3D-printed enclosure. (F) Top view of the 3D-printed enclosure. (G) Side view photograph of the modified rectangular acrylic tunnel placed on the squirrel platform. (H) Top view photograph of the modified rectangular acrylic tunnel on the squirrel platform. (I) Photograph of the modified rectangular tunnel with the assembled 3D-printed enclosure placed on the squirrel platform. (J) Photograph of a squirrel transferred to the arena using the modified rectangular tunnel with the 3D-printed enclosure assembly. The use of images is approved by the Head of the Veterinary Research and Resources Section (NEI)

2. Setting up the OMR arena

1. Disinfect the platform and the square tube enclosure assembly with a cleaner/disinfectant before and in between animals to remove unwanted olfactory cues.

   NOTE: While leather gloves are used to handle the animals and come into direct contact with them, they are not worn during the recordings and do not cause any lasting or observable confounding effects.

2. Tear a small piece of paper towel measuring approximately 50 mm × 50 mm and fold it to cover the top surface of the pedestal. Place the 3D-printed platform onto the pedestal (14 inch high) with the piece of paper towel in place.

   NOTE: The paper adjusts the press fit of the platform, prevents unintended movement, provides a uniform white background for imaging (if the pedestal’s surface is not white), and acts as an absorbent layer in case the animal urinates during the recording, preventing the platform from becoming slippery.

3. Wearing leather work gloves, gently guide the animal into the square tube enclosure assembly.

4. Transfer the animal to the arena by placing the enclosure assembly onto the platform (Figure 1J).

5. Carefully detach the removable top enclosure leaving the animal in the acrylic extruded clear square tube on the platform. Close the door to the OMR system

3. Setting up the OMR software parameters

NOTE: The pictured OMR system uses a compact industrial camera featuring an IR-sensitive 1/3” CMOS sensor with a global shutter at 25 frames per second (fps). The camera is equipped with an F1.6 wide-angle lens, enabling a complete top-down view of the arena. Preliminary visual acuity tests conducted on squirrels using the standard 100% contrast setting revealed that the squirrels’ visual acuity exceeded the resolution limits of the hardware. The 23.8” full HD In-Plane Switching screens used in the setup have a resolution of only 1920 pixels in width, and the software is therefore hard-coded with a maximum spatial frequency of 2 cyc/°. For comparison, the visual acuity of C57Bl/6 mice is approximately 0.3–0.5 cyc/°. At 2 cyc/°, each stripe averages 5.3 pixels in width. With a rotation speed of 12°/s and a frame rate of 60 frames per second, the pattern shifts by 0.2° per frame, equivalent to 4.2 pixels per frame, assuming uniform pixel speed. At spatial frequencies near 2 cyc/°, the stripe pattern approaches the resolution limit where aliasing may occur, causing the pattern to appear as if it is moving backward. To prevent this artifact, the software is hard-coded to cap the spatial frequency at 2 cyc/°.

Thus, for squirrels we devised a protocol where we perform the acuity test but at a much lower contrast. We selected a contrast setting (12.78%) where the squirrels were unable to see the striped pattern at 2 cycles/°.

1. In the live video panel, ensure that the animal is adapting well to the new environment and is beginning to move less and calming down.

   1. At 2 cyc/°, each stripe averages 5.3 pixels in width. Set the rotation speed to 12°/s and a frame rate of 60 frames per second; the pattern shifts by 0.2° per frame, equivalent to 4.2 pixels per frame, assuming uniform pixel speed. Set the spatial frequencies near 2 cyc/° so the stripe pattern approaches the resolution limit where aliasing may occur, causing the pattern to appear as if it is moving backward. To prevent this artifact, hard code the software to cap the spatial frequency at 2 cyc/°.

      NOTE: For squirrels, we devised a protocol where we perform the acuity test but at a much lower contrast. We selected a contrast setting (12.78%) at which the squirrels were unable to see the striped pattern at 2 cycles/°.

2. Download Supplementary Coding File 3. Open the OptoDrum software.

3. Under the Settings tab, click on Import Settings. Load the Supplementary Coding File 3.

4. Under the Settings tab, set the Background threshold offset to 17, the minimum animal size [px] to 30, and the maximum tail width [% of size] to 1.

5. Under camera settings, leave Invert Video and Manual Camera control unchecked and IR Light Off.

6. Widen the position and size of the region of interest for tracking to accommodate the squirrel if it leans outside of the opening of the tube. The suggested values are Position: X: 359, Y: 237, Size: X: 570, Y: 455. Do not extend the ROI over the Stimuli screen, as it can lead to false positives.

7. Select Session Configuration from the top panel (Figure 2A).

   1. For the staircase, set estimated spatial acuity as follows. Set expected acuity to 1.800 cyc/° (648 cycles); set optimal stimulus resolution to 0.5 cyc/° (180 cycles); set measurement resolution to 0.100 cyc/° (36 cycles).

   2. For the staircase, set the number of required confirmations as follows. Set for failed trials to 3; set for successful trials to 2.

      NOTE: The number of successful trials can be increased to further reduce the occurrence of false positive confirmations. However, with the use of this platform and modified handling tunnel, the error rate remains acceptably low. This is primarily due to the squirrels’ relative stillness and the infrequency of rotational head movements aside from those elicited by visually evoked stimuli.

8. Under the Staircase tab (Figure 2B), set antialiasing to width to 3 px and leave sinusoidal unchecked.

9. For the testing criteria, lock the Contrast at 12.78% and lock the rotation speed at 12 °/s.

   NOTE: At the 12.78% contrast setting, the squirrels’ spatial acuity is limited to approximately 1.5–1.8 cyc/°. No longer exceeding the maximum hardware settings, this configuration enables the assessment of spatial acuity in response to disease or injury.

10. Once two of the three parameters are locked, the Auto Set the parameters option will become active (no longer grayed out). Check the box next to Auto Set the Parameters.

11. Set the rotation direction depending on whether visual impairment is expected to be in one or both eyes. Note that for each eye, only motion in the temporal-to-nasal direction elicits tracking. Consequently, assess injury to the left eye using clockwise (CW) stimulation, as counterclockwise (CCW) stimulation is expected to elicit normal responses. For degeneration or injury expected in both eyes, rotation of the stimulus in both directions may be used.

12. Click on Start Trial to begin the presentation of the stimulus. The camera above will begin recording and tracking the animals’ head movements. While the stimulus is being presented, click on the Analysis Tab for real-time viewing of the Angular Velocities (head rotation), Track Quality, and Score parameters. At the end of each stimulus presentation, the software will automatically determine whether to advance (indicated by the green checkmark) to the next Cyc/°, repeat the trial, or reduce the Cyc/° (indicated by the red X).

13. During the recording, if the tracking software incorrectly flips the nose (red) and tail (green) markers, press Ctrl on the keyboard to allow manual override to correct the head-tail orientation.

14. If the animal falls from the platform, pause the trial by pressing Pause Evaluation or clicking on the Space Bar. Remove the handling tunnel from the platform and use it with the lid for safe handling of the animal to place it back onto the platform. Once the animal is ready, click on Resume Evaluation to continue the trial.

15. Once all the trials have been completed, click on the Summary Tab to show the successful trials in green and the unsuccessful trials in red. The trial circled in green indicates the visual acuity threshold for the squirrel using the contrast of 12.78%. Return the animal carefully to its home cage.

Figure 2: Software settings for measuring squirrel spatial acuity.

(A) Parameters adjusted to accommodate the squirrels’ high spatial acuity compared to other rod-dominant rodents. The expected acuity, or maximum acuity the animal can achieve, is set to 1.8 cyc/°. Since squirrels can perceive beyond 2 cyc/° for most contrasts, the contrast is reduced to 12.78%. The optimal stimulus resolution, defined as the cyc/° used to initiate the test, is set to 0.5 cyc/°. This value is positioned well below the visual acuity threshold to ensure the detection of any potential visual deficits. The measurement resolution is configured to 0.1 cyc/° to enable fine adjustments as the software refines the stimulus to determine the animal’s visual acuity threshold. (B) Graphical representation of the parameters outlined in (A).

Representative Results

OMR is considered successful once the software identifies the visual threshold determined by the animal’s performance in spatial frequency discrimination trials. Specifically, the threshold is defined as follows: two successful trials at a given spatial frequency (measured in cyc/°), followed by three failed trials at the next higher spatial frequency. OMR was performed on 13-lined ground squirrels (6–8 months of age) using the standard round flat platform (Figure 3A) to compare with the non-aversive platform and handling tunnel (Figure 3B). When animals were placed on the standard round flat platform, they invariably fell off and were unable to complete the automated visual threshold assessment within the allotted time (10 min). In one instance, an animal leaped from the platform and grabbed the camera above. While it is technically possible to measure OMR under these conditions, it requires numerous attempts and several hours per animal. In contrast, the non-aversive platform and handling tunnel enable successful OMR measurements in 100% of animals without the need for prior acclimation or positive reward conditioning.

Figure 3: Comparison of the standard round flat platform with the non-aversive platform.

(A) Successive top-view images of a thirteen-lined ground squirrel (TLGS) on the standard platform demonstrate how easily the animals leave the platform, interrupting trials and preventing experiment completion. (B) TLGS on the non-aversive platform successfully completed a trial to detect the visually evoked optomotor response (OMR). The animal remains on the platform for the entire duration of the trial and until the conclusion of the experiment. (C) Boxplots showing latency to fall for animals (n=4) placed on either the standard or non-aversive platform. (Standard Platform: TLGS-1 (4 falls), TLGS-2 (7 falls), TLGS-3 (4 falls), TLGS-4 (12 falls) within 10 minutes without visual stimulation). (D) Boxplots showing the duration of individual trials (including both successful and unsuccessful trials) in response to clockwise (CW), counterclockwise (CCW), or stimulation in both directions. Note for both (D) and (E), data for the standard platform is unavailable, as no animals remained on the platform long enough to complete the experiment. (E) Boxplots showing the duration of the OMR test resulting in the successful detection of the animals’ visual acuity threshold. Using the non-aversive platform, visual acuity thresholds are consistent regardless of the type of stimulus applied.

The latency to fall for each platform is presented in Figure 3C, with a maximum observation time of 10 min. Animals were divided into two groups and initially assigned to either the standard or non-aversive platform in the absence of external stimuli. To account for potential preconditioning effects, the animals were subsequently swapped between platforms following the completion of the initial tests. Animals placed on the standard platform frequently fell within seconds, whereas the same animals placed on the non-aversive platform consistently remained on the platform without falling. A side-by-side comparison of the duration of individual trials (Figure 3D) and the time required to complete a successful OMR test, culminating in the detection of spatial acuity thresholds (Figure 3E), is presented for the non-aversive platforms under different stimuli. Notably, data for the standard platform are unavailable, as no animals remained on the platform long enough to complete a test.

Figure 4A shows a photo of a 13-lined ground squirrel inside the modified tunnel after being transferred to the platform in the OMR device. The tunnel’s transparent acrylic allows the animal to view the four monitors independently of the direction it is facing. The tunnel provides sufficient space for the animal to reorient itself and turn around. Figure 4B shows a top-view image captured using the camera mounted above. Using the correct camera parameters (see step 3), the tracking software outlines the squirrel and tracks its head movements. Figure 4C shows a screen capture of an animal viewing the (CW) stimulus presentation. Figure 4D provides a zoomed-in image overlaying two screen captures taken approximately 2 s apart, illustrating the reflexive change in head position as the animal tracks the stimulus. Figure 4E shows real-time analysis of the squirrel’s head movements which the software uses to automatically determine whether the head motions are visually evoked or random. Note that during the period between 11 s and 13 s, the tracking quality (second panel from the top) is successful. The orange backgrounds indicate the time periods during which clockwise stimulation is presented, while the blue indicates counterclockwise stimulation. For the following data, four naïve, uninjured animals were measured, including two males and two females.

Figure 4: Successful tracking of TLGS using the non-aversive platform.

(A) Photograph of a TLGS on the non-aversive platform (with the 3D-printed enclosure removed). (B) Top-view image from a camera mounted above the arena shows that the non-aversive platform does not obstruct tracking of the animal’s movements. Animal orientation is defined by the nose (red) and tail (green). The region of interest (ROI) is sufficiently large to capture the animal’s head movements, which may extend beyond the platform’s confines. (C) Top-view image of a squirrel during a trial, with a blue arrow indicating the direction of the stimulus presentation and a black arrow indicating the direction of head rotation. (D) Zoomed-in overlay image of the squirrel’s head, showing head rotation through two images taken approximately two seconds apart. (E) Automated analysis of a trial for the squirrel is shown in panel (A-C). The visually evoked OMR, like the one shown in (C), is represented by positive peaks (circled in yellow) in the plot of tracking quality. Note: The stimulus is presented when the center of the animal (yellow dot, shown in (B)) is stationary. The duration of the stimulus is 5 s, followed by a 1s interval. The tracking score increases upon successive tracking of the stimuli, as indicated by the positive tracking of the CCW stimulus starting at t=7s (first green arrow) and ending at t=8s (first magenta arrow), followed by the positive tracking of the CW stimulus starting at t=11s (second green arrow) and ending at t=13s (second magenta arrow).

Figure 5 shows the acquisition summary following the successful completion of an OMR experiment. Using the Staircase parameters (see step 3.7.1), the software automatically adjusts the stimulus pattern by ± 1–3 cyc/° (Figure 5A). Note the duration (Figure 5A, far right column) required for the software to determine whether the animal was tracking the stimulus. Typically, for stimuli that resulted in successful tracking, the duration of presentation was under 30 s. On average, the visual acuity threshold is determined within 7.7 ± 1.6 min (n=4) for stimuli in both directions, 7.3 ± 2.2 minutes for CCW, and 7.1 ± 1.2 minutes for CW. Figure 5B provides a graphical summary of the data. The green circled data point indicates the threshold of 576 cycles (1.6 cyc/°). Figure 5C demonstrates that in naïve squirrels subjected to both clockwise and counterclockwise stimulation, the visual acuity could not be determined within 20 min using the standard round flat platform, as the squirrels were unable to remain on the platform. However, the modified platform and tunnel successfully enabled measurement of the visual acuity threshold in four out of four animals (1.6 ± 0.1 cyc/°). Since the optomotor response is strongest to temporal to nasal stimulation, we used clockwise stimulation to elicit the optomotor response in the left eye, and counterclockwise was used for the right eye (Figure 5D-E). We gently sutured the animal’s left eye closed under isoflurane anesthesia and allowed the animal to recover. We then repeated the test using either CCW (Figure 5D) or CW (Figure 5E) only stimulation. Under CCW stimulation (Figure 5D), there is no statistical difference (p = 0.06) between the naïve condition (1.6 ± 0.1 cyc/°) versus left eyelid closed (1.8 ± 0.1 cyc/°). In contrast, when the animals were presented with CW stimulation (Figure 5E), visual acuity was greatly impaired in the group with the left eyelid closed (0.011 cyc/°, the lowest setting possible) compared to the naïve group (1.5 ± 0.2 cyc/°; p = 0.001). This straightforward test validates the protocol and highlights the effectiveness of the non-aversive platform in rapidly and reliably measuring the optomotor response in hyperactive animals, such as the thirteen-lined ground squirrel. We further evaluated the non-aversive platform using an optic nerve crush (ONC) model, a well-established experimental paradigm³ for investigating traumatic optic neuropathy resulting from blast wave or head trauma, as well as optic nerve degeneration associated with glaucoma. A schematic illustration of the optic nerve crush in squirrels is depicted in Figure 5F. Optic nerve crush (ONC) was performed as previously described³ under general anesthesia, induced with 3–5% isoflurane and maintained at 1–2% via nosecone. Topical 0.5% proparacaine was applied to the surgical eye for local anesthesia. To prevent infection, topical antibiotics were administered, followed by lubricating gel to both eyes to prevent corneal dehydration during anesthesia. Ketoprofen (5 mg/kg, IP) was administered for postoperative analgesia. Body temperature was carefully maintained throughout the procedure. For the following experiment, the optic nerve of the right eye was carefully crushed in awake (non-hibernating) squirrels using curved-tip forceps (Figure 5G). The ONC procedure induces retinal ganglion cell death and axonal degeneration, thereby disrupting visual signal transmission and functionally rendering the right eye blind. To assess visual acuity 4 days post-ONC we delivered either CW or CCW stimulation and determined the acuity threshold at which the animal could still perceive the stimulus at 12.78% contrast. Under CW stimulation, the left (unaffected) eye exhibited a visual acuity of 1.5 ± 0.2 cyc/°, comparable to the naïve condition (Figure 5E). In contrast, the ONC-affected right eye failed to detect the stimulus, with a visual acuity of 0.011 cyc/°; p = 0.0001). Thus, this approach of using a non-aversive platform broadens the applicability of OMR as a clinically relevant screening tool for assessing visual function in disease progression or drug discovery impacting cone function.

Figure 5: Validation of measuring squirrel OMR using the non-aversive platform.

(A) Table summarizing the results of a complete OMR experiment conducted with the non-aversive platform and protocol. (B) Graphical summary of the results in (A), showing a visual acuity threshold of 1.6 cyc/° at 12.78% contrast in response to stimuli in both directions. (C) Comparison of the visual acuity detection performance between the standard platform and the non-aversive platform n=4 animals. (D-E) Validation of the non-aversive platform for measuring the animal’s visual acuity by selectively evaluating the contributions of the left and right eyes to the optomotor response based on stimulus direction. (D) With the left eyelid held closed, there is no observed difference in response to CCW stimulation, indicating that the right eye predominantly mediates the OMR to CCW stimulation. (E) With the left eyelid held closed, animals fail to detect CW stimuli, confirming the necessity of the left eye for CW stimulus detection. (F) Schematic illustration of the 13-lined ground squirrel optic nerve crush procedure (Image was generated in Biorender). The TLGS has a linear optic nerve head that runs horizontally across the retina and forms a Y shape as it exits the back of the eye. (G) Screenshot of the optic nerve crush during a live surgical procedure. The crush is performed approximately 2mm from the center of the optic disc. (H) Boxplot of visual acuity threshold in response to CW or CCW stimulation. **** p= 0.0001, Student’s T-test.

Additionally, we investigated whether the non-aversive platform could provide similar benefits for more commonly used laboratory rodents, such as mice and rats. A key challenge in these species is the tendency of animals to fall or leap from the platform, which we aimed to mitigate through platform modification. The base of the platform is scaled in size to accommodate mice, ensuring compatibility with the experimental setup. The FCStd file for the prototype platform (mouse) used in this proof-of-concept study is provided (Supplementary Coding File 4). This file contains the detailed dimensions necessary for fabrication and replication of the mouse platform. For mice, the platform is raised to 18 inches from the floor of the arena. While falling from this height does not result in injury, frequent falls can be a reason for removing an animal from a study, thereby reducing the sample size (N; Figure 6A). Additionally, certain mouse models of disease, such as neurofibromatosis 1 (NF1 mice), exhibit poor balance or locomotor abilities, making them more prone to falling¹⁷. We modified the platform’s size (Figure 6B) for tests of CX3CR1^GFP/+ mice, a commonly used microglia reporter line bred on a C57BL6 background. Similar to squirrels, the mice have the freedom to orient themselves on the platform and can face either direction (Figure 6C). The tracking software (settings provided in Supplementary Coding File 5) is able to monitor the animal’s position and head movements without interference from the sides of the acrylic tunnel (Figure 6D). Unlike squirrels and humans, which possess cone-rich regions (cone-rod ratios of 50:1)^18,19, mice have relatively poor visual acuity, necessitating adjustments to the stimulus parameters (Figure 6E), including increasing the contrast to the maximal value of 99.72%. A graphical representation of a representative OMR experiment using the non-aversive platform is shown in Figure 6F. The measured visual acuity threshold is circled in green. A table showing the results of the individual trials that culminated in the identification of the visual acuity threshold for this mouse is shown in Figure 6G. The duration of each OMR test (Figure 6H) and the lengths of each individual trial (Figure 6I) are similar for both the standard and non-aversive platforms. Figure 6J, we manually counted the number of fecal pellets inside the behavioral arena at the conclusion of the trials as an index of stress or anxiety²⁰. Although there was a trend toward more fecal pellets in mice using the standard platform (2.9 ± 1.5) compared to the non-aversive platform (2.1 ± 1.5), the difference did not reach statistical significance. This was likely due to three mice that climbed onto the side wall of the acrylic tunnel and were unable to descend, requiring assistance. The ordeal appeared to cause notable stress, significantly increasing their fecal output compared to those who stayed in the tunnel. This suggests that slight modifications to the tunnel height may be necessary to discourage such behavior. Importantly, the non-aversive tunnel (0.42 ± 0.1 cyc/°) yielded similar visual acuity measurements as the standard platform (0.44 ± 0.1 cyc/°; Figure 6K) and consistent with previously reported values for WT mice (0.43 ± 0.05 cyc/°)²¹. The major advantage of the non-aversive platform was its significant reduction in the number of falls from the platform (Figure 6L) compared to the standard platform. This improvement makes it possible to perform behavioral assessment of visual acuity in more lines of transgenic mice, which may not have been feasible with the standard platform.

Figure 6: Validation of the non-aversive platform for use with mice.

(A) Sequential images showing that mice on the standard round flat platform can leap down or fall, disrupting experiments. (B) The non-aversive platform design can be adapted to accommodate a mouse-sized rectangular tunnel (Length: 95mm, Width: 50.8mm). (C) Photograph of a mouse positioned on the platform within the testing arena. Note the transparent acrylic does not distort the light passing through it. (D) Top view image captured with the camera mounted above the platform, showing a mouse on a scaled-down version of the non-aversive platform. The image demonstrates successful tracking of the animal’s outline. (E) Stimulus parameters for recording the OMR in mice with expected visual acuity similar to wild type (WT). (F) Graphical representation of an OMR experiment confirming successful visual acuity measurement. (G) Table summarizing the results from a complete OMR experiment conducted using the non-aversive platform. (H) Boxplot comparing the duration of successful OMR tests performed using the standard platform versus the non-aversive platform. (I) Boxplot illustrating trial lengths conducted with the standard or non-aversive platform. (J) Fecal pellet output, used as a measure of stress, comparing the standard and non-aversive platforms. (K) Comparison of measured visual acuity using the standard versus the non-aversive platform, highlighting comparable values. (L) Boxplot quantifying the number of falls from each platform during the OMR tests. The non-aversive platform shows a significantly lower number of falls compared to the standard platform (* p<0.05, Student’s T-test).

To further evaluate the utility of the non-aversive platform, we compared its performance with the standard round platform in 1-month-old rd10 mice, a model of inherited retinal degeneration (retinitis pigmentosa)²². These mice harbor a mutation in the β subunit of PDE6, a gene encoding a key photoreceptor protein, leading to progressive photoreceptor degeneration²³. Given the expected decline in visual acuity in rd10 mice, stimulus parameters were adjusted accordingly. Specifically, we set the expected acuity at 0.35 cyc/°, the optimal stimulus resolution at 0.031 cyc/°, and the measurement resolution at 0.031 cyc/° (Figure 7A-B). For the following tests, mice were divided into two groups and initially placed on either the standard or non-aversive platform (Figure 7C). After completing the first set of experiments to assess visual acuity (Round 1), the animals were retested using the alternate platform (Round 2). While not statistically significant, OMR experiments conducted using the non-aversive platform tended to have shorter overall durations (Figure 7D), whereas the length of individual trials remained comparable between the two platforms (Figure 7E). Notably, mice tested on the non-aversive platform tended to produce fewer fecal pellets, suggesting reduced stress (Figure 7F). Visual acuity measurements were comparable between platforms, although the non-aversive platform exhibited reduced variability (Figure 7G). One mouse developed cataracts before testing, and its visual acuity measurement is marked with an x to denote it as an outlier. Importantly, the non-aversive platform significantly reduced the number of falls per animal during testing compared to the standard round platform (Figure 7H), further supporting its advantages in maintaining stable behavioral engagement and minimizing animal distress.

Figure 7: Software settings for measuring spatial acuity in rd10 mice at P30.

(A) Parameters adjusted to accommodate the mouse’s declining spatial acuity compared to WT mice. The expected acuity, or maximum acuity the animal can achieve, is set to 0.35 cyc/°. While the contrast is set to a maximal value of 99.72%. The optimal stimulus resolution, defined as the cyc/° used to initiate the test, is set to 0.031 cyc/°. The measurement resolution is configured to 0.031 cyc/° to enable fine adjustments as the software refines the stimulus to determine the animal’s visual acuity threshold. (B) Graphical representation of the parameters outlined in (A). (C) Representative images of rd10 mice on either the standard (left) or non-aversive (right) platforms. (D) Boxplot comparing the duration of successful OMR tests performed using the standard platform and the non-aversive platform using rd10 mice. (E) Boxplot illustrating the length of trials conducted with the standard or non-aversive platform using rd10 mice. (F) Fecal pellet output is used as a measure of stress to compare the standard versus non-aversive platform. (G) Comparison of measured visual acuity using the standard versus the non-aversive platform, highlighting comparable values achieved with each. (H) Boxplot quantifying the number of falls from each platform during the OMR tests. The non-aversive platform shows a significantly lower number of falls compared to the standard platform (* p<0.05, Student’s T-test).

In summary, the findings demonstrate that the non-aversive platform provides a significant improvement over the standard round platform for assessing visual function using the optomotor response (OMR) across multiple species, including ground squirrels and mice. By minimizing stress and reducing the frequency of falls, the non-aversive platform enhances animal engagement and ensures more reliable data collection without the need for extensive acclimation or positive reinforcement. This was particularly evident in hyperactive species such as the 13-lined ground squirrel, where the standard platform failed to yield any usable OMR measurements, whereas the non-aversive platform enabled 100% successful threshold determinations. Furthermore, the platform’s efficacy was validated in a model of traumatic optic neuropathy, confirming its utility for functional assessments of vision loss. The ability to extend this approach to commonly used rodent models, including rd10 mice with inherited retinal degeneration, underscores its broad applicability for preclinical vision research. By improving experimental efficiency and reducing variability, the non-aversive platform has the potential to facilitate high-throughput screening of therapeutic interventions for retinal and optic nerve diseases while promoting animal welfare through stress reduction. Future studies may refine the tunnel design to further optimize animal behavior and ensure its compatibility with a wider range of species and experimental conditions.

Discussion

Critical steps

Facilitating Tunnel Entry - One important point to consider is that the squirrels require the 3D-printed enclosure assembly to serve as a temporary lid for the acrylic rectangular tunnel; without it, they will not readily enter. In our facility, the 13-lined ground squirrels are regularly housed with a red rat-sized cylinder (Length: 15.24 cm, Diameter: 7.62 cm). The purpose of the 3D-printed enclosure is to provide familiarity with their enrichment tunnels, making handling them easier. It is important to wear heavy leather gloves when handling the squirrels, as, despite their seemingly harmless appearance, they can bite or scratch and are as unpredictable as any other rodent. Animals can be encouraged to enter the tunnel and can be transferred by gently placing a hand over each end of the tunnel, blocking both ends. This method causes the least amount of stress.

Surface Cleaning Protocol-Another important step is to clean all surfaces that come into contact with the squirrels using a non-toxic, non-corrosive disinfectant that also deodorizes in a single step. Proper cleaning reduces the risk of pathogen transmission and enhances data integrity by eliminating residual scents that could distract animals during OMR testing.

Platform Design and Printing - The 3D-printed platforms should be fabricated using white or near-white filament to provide optimal contrast between the squirrel and the platform. This is particularly important as the squirrels’ distinctive coat pattern-characterized by alternating brown and light-colored stripes and spots-can make it challenging for tracking software to clearly define their outline.

Modifications and troubleshooting

Platform Design Enhancements - The initial engineering prototype of the platform can be further improved by applying Design for Manufacturing and Assembly (DFMA) principles, which focus on simplifying fabrication and reducing assembly costs. In the final platform design (see Supplementary Coding File 6), several modifications have been implemented. The platform is now secured to the pedestal by a cylinder on the bottom of the platform, providing a more stable and reliable attachment mechanism. The rectangular transparent acrylic side walls can now be easily cut with a handheld rotary tool, eliminating the need for specialized machining. These walls are assembled directly into the base using newly integrated printed slots for a secure fit. To further enhance stability, the acrylic side walls can be bonded using a biocompatible, medical-grade adhesive epoxy selected for its strong plastic adhesion and sterilization resistance. These modifications offer two key benefits: 1) Improved manufacturability and cost reduction - The revised design allows the use of flat acrylic sheets instead of prefabricated rectangular cylinders, providing greater flexibility in thickness selection and maximizing optical clarity. To prevent mice from climbing the side walls, which were previously 2 inches tall, the walls can be extended to a height of 3–3.5 inches. 2) Enhanced platform stability - It removes the need for the through-hole cutout in the center of the platform which had allowed the initial 3D-printed design to fit over the pedestal even when excess material from the printing process narrowed the diameter. The revised design removes this dependency, securely attaching the platform to the pedestal and eliminating the need for paper spacers as press-fit solutions. These enhancements optimize the platform for scalability, reliability, and reproducibility, making it more suitable for widespread use in experimental applications.

Protocol Adjustments for Impaired Visual Acuity - For studies involving animals with expected visual impairments, the protocol can be customized by adjusting the optimal stimulus resolution to suit their visual capabilities. Additionally, increasing the required confirmations of successful trials from the default of 2 to 3 can help minimize false positives, ensuring greater accuracy in the results.

Managing Animal Behavior During Testing - The use of an acrylic tunnel proved effective in reducing stress among test subjects; however, some squirrels displayed behaviors indicative of excessive comfort, such as grooming or even falling asleep. To maintain engagement with the visual stimulus, a light tap on the table or testing apparatus can serve as a gentle cue, redirecting their attention without introducing undue stress.

Furthermore, the modular design of the platform suggests that incorporating the tunnel into the animals’ housing environment as part of their enrichment could be advantageous. Familiarization with the tunnel prior to testing may enhance the ease of transfer to the experimental arena, further mitigating stress. Although the number of fecal pellets (a common measure of stress²⁰) were reduced in mice, it did not rise to the level of significance. Nonetheless, the trend suggests potential stress alleviation through environmental familiarity.

In the case of squirrels, fecal pellet counts were not recorded due to logistical constraints. The animals required transportation across campus to access the optomotor device, exposing them to both environmental and motion-related stressors which could confound stress measurements based on the platform. Additionally, the natural seasonal reduction in food intake during winter, which facilitates hibernation, leads to decreased fecal output. As a result, fecal pellet counts would not have served as a reliable stress measure in these animals. These findings highlight the importance of species-specific considerations when assessing stress indicators, especially in non-traditional laboratory species. Therefore, as an alternative methodology, we evaluated the number of falls from the platform.

Limitations of the method

This study assessed visual acuity at a single contrast level (12%) due to the squirrel’s exceptionally high visual acuity, which exceeds the hardware capabilities of the testing system. Lowering the contrast increases the difficulty of tracking the visual stimuli, enabling more precise determination of the visual thresholds. One limitation of the design is that when using stimulation in only one direction (CW or CCW), the animals tend to track the stimulus and then rest their head against the acrylic side wall, which prevents further rotation in the same direction. To address this, a light tap on the table or OMR device is needed to recenter the animal’s attention (head position), allowing it to respond to the next stimulus. In general, the tracking of squirrels is more robust than that of mice, and when using stimuli in both directions, it leads to a more efficient determination of visual acuity.

Another limitation is the potential influence of external factors such as time of day (circadian rhythms)²⁴ on the optomotor reflex. Squirrels may additionally be influenced by the time of the season, as they typically hibernate in the fall and winter^25,26. While some research animals fail to hibernate and remain active during colder months, seasonal influences on visual testing cannot be entirely excluded.

Environmental conditions within the testing apparatus can also pose challenges. The compact design of most OMR systems, with four monitors operating in a confined space, can lead to elevated temperatures inside. Some systems, like the one used in this study, feature USB-powered fans positioned at the bottom of the arena to improve air circulation and have adjustable speeds that can be controlled to maintain the appropriate temperature throughout the experiment. To further mitigate heat buildup and its potential effects on the OMR, it is recommended to leave the door open to the OMR system between animals to allow for improved air circulation. While the OMR test is highly sensitive and capable of detecting subtle changes in visual function²⁷, it is limited by its reliance on reflexive behavior. Unlike tests assessing cortical vision, OMR does not measure conscious visual processing. Additionally, the OMR may fail to detect focal retinal damage, such as laser-induced injuries, if sufficient functional retina remains intact.

Lastly, the OMR is a color-blind response, as seen in humans and other species like zebrafish, where color and motion perception are distinct processes. However, using the 13-lined ground squirrel model and testing at low rotation velocities where color can be perceived²⁸, may allow for investigation into ocular side effects of drugs on color vision changes. It remains unclear which photoreceptors are involved in large-field motion detection in squirrels. Exploring this aspect may open new avenues for understanding retinal function and developing innovative models of retinal disease. The ability to modify the color of the bar stimulus using RGB values (0–255) will facilitate future research in this promising area.

Significance of the method

The non-aversive platform enables robust behavioral assessment of visual function in squirrels and enhances the OMR test across other rodent species as well. Rodents exhibiting excessive locomotor activity, with squirrels being an extreme example, may display diminished visual performance due to their hyperactivity interfering with their awareness of the visual stimuli¹⁵. By mitigating such behavioral interference, the non-aversive platform has the potential to help normalize results across strains, thereby reducing variability and improving the reliability of visual performance assessments. Furthermore, this platform could be applied to emerging animal models, such as the tree shrew, a diurnal species with a cone-dominant retina and high visual acuity similar to squirrels²⁹. Consequently, the non-aversive platform broadens the applicability of the optomotor response test for investigating visual function. It also serves as a complementary tool to established methodologies, including electroretinography (ERG)¹ and other behavioral assessments like the water maze³⁰, enhancing the toolkit available to vision research.

Importance and potential applications of the method in specific research areas

The non-aversive platform enhances the acquisition of optomotor response (OMR) data across diverse rodent models, expanding the scope of the optomotor response test to emerging species with complex visual systems and offering increased flexibility for large-scale studies aimed at accurately assessing visual function. By incorporating modifications to the standard OMR protocol, we demonstrate that the 13-lined ground squirrel can be reliably recorded using this method. This capability highlights the platform’s potential for drug screening applications, including evaluating compounds that may induce or alleviate visual impairments.

Supplementary Material

Supplementary Coding File 3

Supplementary Coding File 3: Tracking settings file for squirrel.

Supplementary Coding File 2

Supplementary Coding File 2: CAD document file for the removable top enclosure for squirrels.

Supplementary Coding File 1

Supplementary Coding File 1: CAD standard document file for the prototype platform for squirrels.

Supplementary Coding File 5

Supplementary Coding File 5: Tracking settings file for mouse.

Supplementary Coding File 4

Supplementary Coding File 4: FCStd file for the prototype platform for mouse.

Supplementary Coding File 6

Supplementary Coding File 6: CAD for the final prototype platform for squirrels used here.

Materials.

Name	Company	Catalog Number	Comments
12V Max Lithium-ion Cordless Rotary Tool Kit	Dremel 8220	F0138220JA	To cut and polish the Acrylic Square Tube. https://www.dremel.com/us/en/p/8220-1-28-f0138220aa
6–0 COATED VICRYL (polyglactin 910) Suture	Ethicon	J670G	For suturing the eye lid. https://www.ethicon.com/na/epc/code/j670g?lang=en-default
Clear Colorless Acrylic Square Tube (2" x 12")	Canal Plastics Center	RT-300161	Clear rectangular tunnel (for mouse). https://www.canalplastic.com/products/clear-colorless-acrylic-square-tube
Clear Colorless Acrylic Square Tube (3" x 12")	Canal Plastics Center	RT-300168	Clear rectangular tunnel (for squirrel). https://www.canalplastic.com/products/clear-colorless-acrylic-square-tube
DREMEL EZ SpeedClic: Plastic Cutting Wheels	Dremel	SC476	For cutting the Acrylic Square Tube
F270 3D printer	Stratasys		To print engineering grade 3D models of the platforms. https://support.stratasys.com/en/printers/fdm-legacy/uprint
Fluriso (Isoflurane)	VetOne	502017	General Anesthesia. Liquid for Inhalation. http://vetone.net/Default/GetFile?id=7dda0c60-45fa-e711-bf2e-0024e8785118
FreeCAD 1.0			Open-source parametric 3D modeler
Igor Pro software Version 6.3.7.2	WaveMetrics, Inc		For plotting the graphs. https://www.wavemetrics.com/
Low-Profile Anesthesia Masks for Traditional Vaporizers for SomnoFlo. Extra Large (animals over 300 g)	Kent Scientific corporation	SOMNO-0804	For providing general anesthesia. https://www.kentscientific.com/products/low-profile-anesthesia-masks-for-somnosuite/
MakerBot Print for Windows version 4.10.1.2056	MakerBot Industries, LLC		Software for orienting the parts for printing
MakerBot Replicator Desktop 3D Printer (5th Gen)	MakerBot Industries, LLC	MP05825	For 3D printing prototype versions of the platforms. https://www.rnd-tech.com/product/makerbot-replicator-desktop-3d-printer-5th-gen/
Microsoft Excel	Microsoft		Spreadsheet software to analyze the data. microsoft.com/en-us/microsoft-365/excel
Multipurpose leather work gloves	Steiner Industries	SPC02	To handle squirrels. https://www.steinerindustries.com/leather-palm/product/steiner-spc02-leather-palm-work-glove
OptoDrum Plus	STRIA.TECH		Commercial device for measuring the optomotor reflex (visual acuity and contrast sensitivity) https://stria.tech/products/optodrum/
OptoDrum software Version 1.7.3	STRIA.TECH		Software for performing automated measures of the optomotor reflex.
Peroxigard Ready to Use (cleaner/disinfectant)	Peroxigard	PRTU242101	To clean/disinfect the platform and the acrylid square tube between experiments. https://peroxigard.com/product-information/rtu/
Pureline M6000 Anesthesia Machine with O2 Concentrator	Penn Veterinary Supply, Inc.	SUPM6000	For animal anesthesia. https://www.pennvet.com/customer/portal/catalog/home?urile=wcm:path%3APennVet+Catalog/Product+Catalog/SUPM6000/Pureline+M6000+Anesthesia+Mach+w_O2+Conc
uPrint SE Plus 3D printer	Stratasys		To print engineering grade 3D models of the platforms. https://www.makerbot.com/makerbot-print/
White opaque MakerBot PLA Filament, Large Spool for Replicator+ (0.9kg, 2lb)	MakerBot Industries, LLC	MP05780	Filament to 3D print the platforms. https://store.ultimaker.com/3d-printer-materials/replicator-series-materials/makerbot-pla-material-large-spool-for-replicator
Pharmacological treatment
Neomycin and Polymyxin B Sulfates and Bacitracin Zinc Ophthalmic Ointment, USP	Bausch & Lomb	RX-0069	Triple ophthalmic antibiotic for preventing ocular infections
Proparacaine Hydrochloride 0.5%	Akorn	17478–263–12	Local anesthetic for ophthalmic instillation