ABSTRACT
BACKGROUND: In low vision patients, the assessment of contrast sensitivity is an essential tool to determine the stage of visual impairment. However, traditional contrast sensitivity tests rely on verbal feedback, and the expertise of the examiner. OBJECTIVE: In the current study, a fast, OKN-based virtual diagnosis tool was developed estimating contrast sensitivity automatically without active cooperation of the patient as well as the practitioner within 3.5 minutes. METHODS: In a HTC Vive headset with an SMI-eye tracker, a virtual rotating drum was implemented, and an algorithm was developed, evaluating the occurrence of an OKN. The tool was evaluated in healthy subjects as well as under low vision simulation for two spatial frequencies and four contrasts. It was then compared to two contrast sensitivity estimates based on manual report on the orientation of static gratings as well as the movement direction of translating gratings. RESULTS: An algorithm was developed, which matched ground truth ratings of occurrence of OKN with an accuracy of 88 %. Furthermore, known differences in contrast sensitivity between healthy and low vision conditions as well as a decrease in contrast sensitivity for lower spatial frequencies was successfully reproduced in the developed tool. CONCLUSIONS: The developed OKN-based sensitivity test represents a reliable proof of concept for technology readiness of virtual reality-based screening tools of visual function in practice, specifically in patients with difficulties to report perception verbally, or under conditions, where no experienced examiner is present.
