Comparison of computerized system and human observers for vision assessment based on optomotor response in Rhodopsin knockout mice

Abstract

Purpose : Detection of optomotor response (OMR) is a convenient approach to evaluate visual functions in rodent. We have developed a computer vision system to detect real time OMR to circumvent the manpower cost issues with systems that need human observers to visually detect OMR. The present study validate the automated system against human observers using in a genetically modified mouse strain with photoreceptor degeneration, Rhodopsin knockout (Rho^-/-) mice.

Methods : Rho^-/- mouse was placed on a pedestal that was placed in the center of a standard OMR system that contained an enclosure formed by four 15-inch LCD monitors for displaying black and white stripes. The OMR of Rho^-/- mice was followed starting from 6 weeks old because it is thought to cover before functional vision loss was detectable to blindness (as previously reported for most Rho^-/- mice). By varying the size, direction and width of stripes, contrast threshold at a fixed spatial frequency (0.20 cyc/deg) and visual acuity based on stripe width at 100% contrast of Rho^-/- mice, were measured by observers (12 eyes; 6 – 12 weeks old) and the automated system (11 eyes; 6-16 weeks old). Linear regression analysis was used to assess the vision changes over time, for both methods.

Results : Automated system and human observers both detected similar progressive decline of visual acuity over time. According to linear regression, the declining rate was 0.061 cpd/week (auto system, R2=0.787) and 0.059cpd/week (human, R2=0.641), respectively. Similarly, automated system and human experimenter both found similar increased contrast threshold over time. According to linear regression, minimal visible contrast increase 9.0% per week (auto, R2=0.71) and 8.4% per week (human, R2=0.71), respectively.

Conclusions : For detection of vision changes in Rhodopsin knockout mice, the automated system performed consistently with human observers. The results suggest that the auto system can be a reliable approach to vision assessment in mice with photoreceptor degeneration. It could be used as a platform for drug screening against photoreceptor degeneration in genetically modified mice.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.