Fig 1.
Outline of VR behavior test setup.
A) Overview of the equipment setup for the VR experiment (A1), the harness used to hold the mouse (A2), a mouse in a harness on the spherical treadmill (A3), and an image of a mouse performing in the VR environment (A4). B) Logic diagram of the VR program. C) Illustration of the conditions required for a mouse to receive a reward at the designated target (green). The shaded area represents where the mouse must be when facing the target to receive a reward. D) A view of both the colored/luminance targets (top), and the moving bars (bottom). Arrows are added to the moving bars to indicate the direction of movement.
Fig 2.
Training outline and representative tracing results.
A) An outline of the training schedule for the color/luminance recognition. B) Representative tracings (red dots and lines) of the movement of the mouse in the VR arena (black circle). The color-coded rectangles around the circle indicate the visual targets. B1 shows a representative tracing of a mouse that has just started training. B2 shows a representative tracing of a mouse that has been trained for 21 days.
Fig 3.
The time course of the recognition of color/luminance targets.
The average number of rewards per session, the average distance traveled to reach a reward, and the average time required to reach a reward are analyzed as a function of training time. A) The average number of rewards per session of untrained mice (Acclimation, D1-D3.5), mice after Baseline training (Baseline, D6.5-D9), and mice at the end of training (Data Acq, D19-D21.5). An ANOVA test shows that the differences among these three groups are statistically significant (F(2,75) = 62.18, P<0.001). Fisher’s PLSD post-hoc analysis shows that the differences between all three pair-wise comparisons are statistically significant (Acclimation versus Baseline, P<0.001; Acclimation versus Data Acq., P<0.001; Baseline versus Data Acq., P<0.001). B) The average number of rewards and failures during the Baseline and Data Acquisition phases. A Chi-squared test shows that the difference between these two groups is statistically significant (F = 114, P<0.001). C) The average number of rewards per session as a function of training time. D) The total distance traveled per session during the Acclimation, Baseline, and Data Acquisition phases. An ANOVA test shows that the differences among these three groups are statistically significant (F(2,75) = 63.64, P<0.001). Fisher’s PLSD post-hoc analysis shows that the differences between all three pair-wise comparisons are statistically significant (Acclimation versus Baseline, P<0.001; Acclimation versus Data Acq., P<0.001; Baseline versus Data Acq., P<0.001). E) The average distance traveled per reward during the Baseline and Data Acquisition phases. A paired Student t-test shows that the difference between these two groups is not statistically significant (df = 18, t = 0.02, P>0.05). F) The percent of the total distance traveled that resulted in rewards during the Baseline and Data Acquisition phases. A paired Student t-test shows that the difference between these two groups is statistically significant (df = 22, t = 4.69, P<0.001). G) The average time required for reaching a reward per session during the Baseline and Data Acquisition phases. A paired Student t-test shows that the difference between these two groups is statistically significant (df = 18, t = 2.24, P<0.05). In all panels except panel C, the number in each column indicates the number of test sessions of 4 mice. Data are presented in the bar graphs as mean ± SEM in this figure and all of the following figures.
Fig 4.
The time course of the recognition of moving bars.
A) The outline of the training schedule for the recognition of moving bars. B) Representative tracing results of a mouse that has just started training (Acclimation, B1), and a mouse that has been trained to recognize the downward moving target for 21 total days (Data Acquisition, B2). C) The average number of rewards per session during the Baseline phase (day 6.5–9), and the Data Acquisition phase (day 19–21.5). A paired Student t-test shows that the difference between these two groups is statistically significant (df = 41, t = 3.76, P<0.001). D) The average number of rewards and failures during the Baseline and Data Acquisition phases. A Chi-squared test shows that the difference between these two groups is statistically significant (F = 63.9, P<0.001). E) The average distance traveled for each reward during the same two phases as in Panel C. A paired Student t-test shows that the difference between these two groups is not statistically significant (df = 37, t = 1.28, P>0.05). F) The average time required for each reward during the same two phases as Panel C. A paired Student t-test shows that the difference between these two groups is statistically significant (df = 37, t = 3.13, P<0.01). The number in each column indicates the number of test sessions from 7 mice.
Fig 5.
Mice retain memory of their VR perception.
A) The outline of the training/testing schedule of the memory test of VR recognition. B) The average number of rewards per session as a function of training time during stage 1 (left, teal) and stage 3 (right, black). A linear regression analysis shows a linear fitting for both phases with a R2 = 0.61 and 0.45, respectively. C) The number of rewards per session of the Acclamation phase (Stage 1), Baseline phase (Stage 1), Data Acquisition phase (Stage 1), the first six sessions of Stage 3 of Additional Training (Retrain), and the final six sessions of Stage 3 Final Acquisition (Final). An ANOVA test shows that the differences among these five groups are statistically significant (F(4,145) = 29.66, P<0.001). Fisher’s PLSD post-hoc analysis shows that the number of rewards of the final six sessions (Final) is significantly higher than any other training phase. In addition, the number of rewards of the initial six sessions of the third stage (Retrain phase) is statistically higher than that of the Acclimation phase of the first stage (Fisher’s PLSD, P<0.0001) but is not statistically different from that of the Baseline (Fisher’s PLSD, p>0.05). However, it is statistically different from that of the Data Acquisition phases of the first stage (Fisher’s PLSD, P<0.05). The number in each column indicates the number of test sessions for 7 mice during stage 1, and 3 mice during stage 3. D) Cumulative distributions of the frequency of rewards/session in stage 1 and stage 3. A Kolmogorov-Smirnov (K-S) test confirmed that the difference of the distributions of the rewards/session of these two stages is statistically significant (P<0.001).