The black eye C57BL/6N male mice (body weight 20–25 g, 8 weeks old, OrientBio, Seoul, Korea) were employed in the experiment. Mice were equipped for behavioral tests as mentioned previously [8, 10, 12, 13]. Briefly, under general anesthesia with isoflurane, a minimal incision in the scalp was made after application of lidocaine cream to reduce pain or discomfort. The head fixation pedestal was made with tow nuts and four screws. Nuts were put on bregma and lambda of the cranium, and screws were embedded between the nuts. Then, mice were permitted to recuperate for at least 3 days after surgery. A drop of physostigmine salicylic solution (Eserine, Sigma-Aldrich Korea, Seoul, Korea) was given to the eyes in preparation for eye movement recording to keep the pupil size stable during the recording. Mice were held in a factory animal holder in the middle of a turntable, and they were adapted to constraint in an animal carrier for 15 min in the dark and then 15 min in the light without any excitation. After two days of acclimation, calibration was performed, which is the process of converting 2-demensional linear visual stimuli on the screen into angular eye movement. The radius of pupil, which is necessary for computing eye movement gain, could be measured through the calibration process. The calibration equations and processes were based on those used in Stahl et al. [14], and mice and the container were put at the same location so that calibration could be done at recordings after calibration.

CCD camera (IPX – VGA210, IMPERX, Boca Raton, FL, USA) with infrared (IR) filter (LP830) was used to capture the image of the eye, which was then transferred to a desktop PC using a camera link grabber board (PCI – 1426, National Instruments, TX, USA). A single IR-LED was put around the camera to establish a reference corneal reflex for calibration as well as IR illumination was produced using IR-LED. A motor-mounted drum with alternating black and white vertical stripes was used to provide optokinetic stimulation (AKM22E – VBBNR – 00, Kollomorgen, Radford, VA, USA). For the output and input between the PC and the monitor, a data collection (DAQ) PCI board (PCI – 6230, National Instruments, Austin, TX, USA) was utilized. Several virtual instruments built in LabView (National Instruments, Austin, TX, USA) were used to process the obtained images.

To assess the baseline performance of the mice utilized in the research, three baseline ocular-motor responses were investigated such as OKR, VOR in the light as well as VOR in the darkness. The horizontal OKR gain was determined by applying with 0.5 Hz frequency and 50 (peak-to-peak) amplitude of rotation in the horizontal position in the light to the drum. The hOKR gain was calculated as the ratio of peak-to-peak eye velocity to peak-to-peak drum oscillation velocity. Twelve cycles of elicited eye movements were chosen for average from sixty cycles, free of motion parallax and eye blinking, and the given stimulation and response were suited to sine curves. As previously described, we used a custum-made program in LabView to analyze the research data [8, 10, 12, 13]. The standard deviation (SD) of the group averages is represented in the text. The changes in OKR gain following hOKR training were compared between groups with different training methodologies using Mann-Whitney U test or Kruskal–Wallis test (SPSS v. 17.0, IBM Corp., Armonk, NY, USA), and significant value is considered with p value 0.05. All procedures were approved by the Institutional Animal Care and Use Committee of Seoul National University College of Medicine.

Before induction of OKR adaptation, we examined the baseline performance of OKR in mice. The OKR gain was measured in the light condition. The OKR drum stimulation was given by 5° amplitude-sinusoidal rotation with different frequencies including 0.1, 0.25, 0.5 and 1.0 Hz. The OKR gain was calculated as the ratio of maximal eye velocity to maximal drum velocity. These baseline performance characteristics were consistent with those of our previous observations [8, 12, 13]. Then, we induced OKR adaptation by continuous oscillation of optokinetic drum while the mice are remained stationary with eyes opened. Considering that OKR gain was measured with drum oscillation of 0.5 Hz frequency and 5° amplitude, we used 3 different categories of learning paradigm as follows: (1) Optokinetic drum oscillation for 60 min with same amplitude and different frequency (0.25 Hz and 5°, 0.5 Hz and 5°, and 1 Hz and 5°; see Fig. 1). (2) Optokinetic drum oscillation for 60 min with same frequency and different amplitude (0.5 Hz and 2.5°, 0.5 Hz and 5°, and 0.5 Hz and 10°; see Fig. 2). (3) Training with serial combination of different frequency or amplitude (see Figs. 3,4).

Fig. 1.

The OKR adaptation after training protocols with different frequency of drum rotation. (A) After 60 min OKR training with 0.25 Hz and 5° drum oscillation, OKR gain was enhanced from 0.190 $\pm$ 0.046 to 0.754 $\pm$ 0.074. Then, the OKR gain after 24 hr in the dark was 0.771 $\pm$ 0.049 (n = 5). (B) After 60 min OKR training with 0.5 Hz and 5° drum oscillation, OKR gain increased from 0.248 $\pm$ 0.078 to 0.776 $\pm$ 0.060. Then, the OKR gain after 24 hr in the dark was 0.753 $\pm$ 0.052 (n = 5). (C) After 60 min OKR training with 1 Hz and 5° drum oscillation, OKR gain increased from 0.233 $\pm$ 0.023 to 0.329 $\pm$ 0.015. Then, the OKR gain after 24 hr in the dark was 0.447 $\pm$ 0.018 (n = 5). (D) Comparison of OKR gain increase immediately after the 60 min training among three protocols indicated that the average ratio of post-training hOKR gain to pre-training hOKR gain was significantly different among three protocols (p = 0.008, Kruskal–Wallis test).

Fig. 2.

The OKR adaptation after training protocols with different amplitude of drum rotation. (A) After 60 min OKR training with 0.5 Hz and 2.5° drum oscillation, OKR gain increased from 0.182 $\pm$ 0.030 to 0.740 $\pm$ 0.063. Then, the OKR gain after 24 hr in the dark was 0.760 $\pm$ 0.039 (n = 5). (B) After 60 min OKR training with 0.5 Hz and 5° drum oscillation, OKR gain increased from 0.248 $\pm$ 0.078 to 0.776 $\pm$ 0.060. Then, the OKR gain after 24 hr in the dark was 0.753 $\pm$ 0.052 (n = 5). (C) After 60 min OKR training with 0.5 Hz and 10° drum oscillation, OKR gain increased from 0.196 $\pm$ 0.044 to 0.516 $\pm$ 0.053. Then, the OKR gain after 24 hr in the dark was 0.609 $\pm$ 0.050 (n = 5). (D) Comparison of OKR gain increase immediately after the 60 min training among three protocols indicated that the average ratio of post-training hOKR gain to pre-training hOKR gain was significantly different among three protocols (p = 0.031, Kruskal–Wallis test).

Fig. 3.

The OKR adaptation after training protocols with combination of different frequency of drum rotations. (A) The OKR training protocol was composed of serial combination of 20 min OKR training with 0.25 Hz and 5° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 1 Hz and 5° drum oscillation. The OKR gain was enhanced from 0.176 $\pm$ 0.046 to 0.672 $\pm$ 0.109 following 20 min OKR training with 0.25 Hz and 5° drum oscillation, slightly increased to 0.755 $\pm$ 0.109 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and decreased to 0.612 $\pm$ 0.117 following 20 min OKR training with 1 Hz and 5° drum oscillation. Then, the OKR gain after 24 hr in the dark was 0.750 $\pm$ 0.060 (n = 5). (B) The OKR training protocol was composed of serial combination of 20 min OKR training with 1 Hz and 5° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 0.25 Hz and 5° drum oscillation. The OKR gain increased from 0.192 $\pm$ 0.043 to 0.249 $\pm$ 0.069 following 20 min OKR training with 1 Hz and 5° drum oscillation, increased to 0.547 $\pm$ 0.133 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and further increased to 0.753 $\pm$ 0.042 following 20 min OKR training with 0.25 Hz and 5° drum oscillation. Then, the OKR gain after 24 hr in the dark was 0.739 $\pm$ 0.028 (n = 5). (C) Comparison of OKR gain increase immediately after the 60 min training between two protocols showed that the average ratio of post-training hOKR gain to pre-training hOKR gain was not significantly different between two protocols (p = 0.841, Mann-Whitney U test).

Fig. 4.

The OKR adaptation after training protocols with combination of different amplitude of drum rotations. (A) The OKR training protocol was composed of serial combination of 20 min OKR training with 0.5 Hz and 2.5° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 0.5 Hz and 10° drum oscillation. The increase in OKR gain was obtained from 0.203 $\pm$ 0.050 to 0.598 $\pm$ 0.033 following 20 min OKR training with 0.5 Hz and 2.5° drum oscillation, slightly increased to 0.707 $\pm$ 0.038 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and decreased to 0.542 $\pm$ 0.151 following 20 min OKR training with 0.5 Hz and 10° drum oscillation. Then, the OKR gain after 24 hr in the dark was 0.664 $\pm$ 0.051 (n = 5). (B) The OKR training protocol was composed of serial combination of 20 min OKR training with 0.5 Hz and 10° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 0.5 Hz and 2.5° drum oscillation. The increase in OKR gain was obtained from 0.229 $\pm$ 0.036 to 0.406 $\pm$ 0.088 following 20 min OKR training with 0.5 Hz and 10° drum oscillation, further increased to 0.674 $\pm$ 0.038 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and slightly increased to 0.735 $\pm$ 0.045 following 20 min OKR training with 0.5 Hz and 2.5° drum oscillation. Then, the OKR gain after 24 hr in the dark was 0.731 $\pm$ 0.048 (n = 5). (C) Comparison of OKR gain increase immediately after the 60 min training between two protocols showed that the average ratio of post-training hOKR gain to pre-training hOKR gain did not exhibit significant difference between two protocols (p = 0.095, Mann-Whitney U test).

The OKR adaptation was investigated by measuring OKR gain change after 60 min training of which the protocol consisted of optokinetic drum oscillation with same amplitude and different frequency. After 60 min OKR training with 0.25 Hz and 5° drum oscillation, OKR gain increased from 0.190 $\pm$ 0.046 to 0.754 $\pm$ 0.074 (Fig. 1A, n = 5). Then, mice were kept for 24 hr in the dark, and OKR gain at the end of 24 hr in the dark was 0.771 $\pm$ 0.049 (Fig. 1A). After 60 min OKR training with 0.5 Hz and 5° drum oscillation, OKR gain increased from 0.248 $\pm$ 0.078 to 0.776 $\pm$ 0.060. Then, the OKR gain after 24 hr in the dark was 0.753 $\pm$ 0.052 (Fig. 1B, n = 5). After 60 min OKR training with 1 Hz and 5° drum oscillation, OKR gain increased from 0.233 $\pm$ 0.023 to 0.329 $\pm$ 0.015. Then, the OKR gain after 24 hr in the dark was 0.447 $\pm$ 0.018 (Fig. 1C, n = 5). The quantity of OKR gain increase immediately after the 60 min OKR training was evaluated among three training protocols. The average ratio of post-training OKR gain to pre-training OKR gain was 413.7 $\pm$ 91.9% in 0.25 Hz and 5° group, 342.0 $\pm$ 126.9% in 0.5 Hz and 5° group, and 142.4 $\pm$ 15.7% in 1 Hz and 5° group, which was significantly different among groups (Fig. 1D, p = 0.008, Kruskal–Wallis test).

Then, we investigated the amount of OKR adaptation after training protocols with different amplitude of drum oscillation. After 60 min OKR training with 0.5 Hz and 2.5° drum oscillation, OKR gain increased from 0.182 $\pm$ 0.030 to 0.740 $\pm$ 0.063 (Fig. 2A, n = 5). Then, mice were maintained for 24 hr in the dark, and OKR gain at the end of 24 hr in the dark was 0.760 $\pm$ 0.039 (Fig. 2A). After 60 min OKR training with 0.5 Hz and 5° drum oscillation, as shown in Fig. 1B, OKR gain was enhanced from 0.248 $\pm$ 0.078 to 0.776 $\pm$ 0.060. Then, the OKR gain after 24 hr in the dark was 0.753 $\pm$ 0.052 (Fig. 2B, n = 5). After 60 min OKR training with 0.5 Hz and 10° drum oscillation, the increase of OKR gain was obtained from 0.196 $\pm$ 0.044 to 0.516 $\pm$ 0.053. Then, the OKR gain after 24 hr in the dark was 0.609 $\pm$ 0.050 (Fig. 2C, n = 5). The amount of OKR gain increase immediately after the 60 min OKR training was evaluated among three training protocols. The average ratio of post-training OKR gain to pre-training OKR gain was 413.3 $\pm$ 65.2% in 0.5 Hz and 2.5° group, 342.0 $\pm$ 126.9% in 0.5 Hz and 5° group, and 276.0 $\pm$ 72.6% in 0.5 Hz and 10° group, which was significantly different among groups (Fig. 2D, p = 0.031, Kruskal–Wallis test).

The amount of OKR adaptation after training with serial combination of different frequency was tested. Training protocol consisted of three consecutive 20 min drum oscillations with different frequency while keeping the amplitude of drum rotation constant. First, the OKR training was conducted by serial combination of 20 min OKR training with 0.25 Hz and 5° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 1 Hz and 5° drum oscillation (Fig. 3A). The increase of OKR gain was obtained from 0.176 $\pm$ 0.046 to 0.672 $\pm$ 0.109 following 20 min OKR training with 0.25 Hz and 5° drum oscillation, slightly increased to 0.755 $\pm$ 0.109 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and decreased to 0.612 $\pm$ 0.117 following 20 min OKR training with 1 Hz and 5° drum oscillation. Mice were maintained in the absolute dark for 24 hr after training, and the OKR gain at the termination of 24 hr dark rearing was measured as 0.750 $\pm$ 0.060 (n = 5). Then, the OKR training was conducted by serial combination of 20 min OKR training with 1 Hz and 5° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 0.25 Hz and 5° drum oscillation (Fig. 3B). The increase of OKR gain was measured from 0.192 $\pm$ 0.043 to 0.249 $\pm$ 0.069 following 20 min OKR training with 1 Hz and 5° drum oscillation, increased to 0.547 $\pm$ 0.133 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and further expanded to 0.753 $\pm$ 0.042 following 20 min OKR training with 0.25 Hz and 5° drum oscillation. Mice were maintained in the absolute dark for 24 hr after training, and the OKR gain was measured as 0.739 $\pm$ 0.028 (n = 5) at the termination of 24 hr dark rearing. The immediate increase of OKR gain following the 60 min training was compared between the protocol with increasing frequency (Fig. 3A) and that with decreasing frequency (Fig. 3B). The average ratio of post-training OKR gain to pre-training OKR gain was lower in the protocol with increasing frequency (374.3 $\pm$ 134.8, left bar in Fig. 3C) than that with decreasing frequency (402.9 $\pm$ 59.7, right bar in Fig. 3C), which, however, was not significantly different between two protocols (p = 0.841, Mann-Whitney U test).

The amount of OKR adaptation after training with serial combination of different amplitude was tested. Training protocol consisted of three consecutive 20 min drum oscillations with different amplitude while keeping the frequency of drum rotation constant. First, the OKR training was conducted by serial combination of 20 min OKR training with 0.5 Hz and 2.5° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 0.5 Hz and 10° drum oscillation (Fig. 4A). The increase in OKR gain was obtained from 0.203 $\pm$ 0.050 to 0.598 $\pm$ 0.033 following 20 min OKR training with 0.5 Hz and 2.5° drum oscillation, slightly increased to 0.707 $\pm$ 0.038 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and decreased to 0.542 $\pm$ 0.151 following 20 min OKR training with 0.5 Hz and 10° drum oscillation. Mice were maintained in the absolute dark for 24 hr after training, and the OKR gain at the termination of 24 hr dark rearing was measured as 0.664 $\pm$ 0.051 (n = 5). Then, the OKR training was conducted by serial combination of serial combination of 20 min OKR training with 0.5 Hz and 10° drum oscillation, 20 min OKR training with 0.5 Hz and 5° drum oscillation, and 20 min OKR training with 0.5 Hz and 2.5° drum oscillation (Fig. 4B). The increase in OKR gain was obtained from 0.229 $\pm$ 0.036 to 0.406 $\pm$ 0.088 following 20 min OKR training with 0.5 Hz and 10° drum oscillation, further increased to 0.674 $\pm$ 0.038 following 20 min OKR training with 0.5 Hz and 5° drum oscillation, and slightly increased to 0.735 $\pm$ 0.045 following 20 min OKR training with 0.5 Hz and 2.5° drum oscillation. Mice were maintained in the absolute dark for 24 hr after training, and the OKR gain at the termination of 24 hr dark rearing was measured as 0.731 $\pm$ 0.048 (n = 5). The OKR gain increase immediately after the 60 min training was compared between the protocol with increasing amplitude (Fig. 4A) and that with decreasing amplitude (Fig. 4B). The average ratio of post-training OKR gain to pre-training OKR gain was lower in the protocol with increasing amplitude (266.3 $\pm$ 40.6, left bar in Fig. 4C) than that with decreasing amplitude (329.0 $\pm$ 59.1, right bar in Fig. 4C), which, however, was not significantly different between two protocols (p = 0.095, Mann-Whitney U test).