TREYESCAN: Development of an Eye Tracking Test for the Measurement of Compensatory Eye Movements for Driving with Visual Field Defects

Abstract

Purpose : A reliable screening method for persons with visual field defects is vital for safe mobility. Presently, static visual field tests are used, but these fixate the patient’s gaze, and do not measure compensatory viewing needed for dynamic traffic situations. Research on this subject is important, as the prevalence of age-related field disorders and the number of older drivers is rising due to a prolonged life expectancy. The aim was to select relevant traffic scenes for TREYESCAN (Traffic eye scanning & compensation analyzer), which measures compensatory viewing with head and eye movements.

Methods : Traffic scenes were recorded with a wide field lens (~120°), whilst driving 2 routes. Interesting scenes were chosen and included in 6 videos of ~8 minutes. Ten raters with extensive driving experience defined relevant areas of interest (AOI). Participants with no ophthalmic diagnoses and at least 5 years of driving experience were included. They watched the 6 videos on a wide screen of 100° field of view, the Pupil Core eye tracker measured the gaze and head movements were permitted in all directions (Fig. 1). They were instructed to view the scenes as if they were driving the car and to press a button when an action was required. Based on their experience and the AOI rating, the best scenes were chosen for inclusion.

Results : Ten raters, with 38.3±9.0 (mean±SD) years of driving experience, identified a total of 349.8±32.8 AOIs per video, with 117.8±4.8 relevant AOIs identified by 5 or more raters. Twenty participants aged 39.2±15.2 with 20.4±15.1 years of driving experience were included. Some participants felt dizzy during scenes with fast turns or speed bumps, hence these situations were removed (12.0 min). In addition, we removed scenes in which an AOI suddenly came into view from behind the vehicle, as this may give a startled sense to participants and information from the side mirrors is not provided (3.5 min). Then, scenes with only central screen AOIs (12.5 min) and a large number of overlapping AOIs (3.5 min), were also removed.

Conclusions : The TREYESCAN traffic scenes (2 videos of ~8 min) will be used in a case-control study, in which eye movements of glaucoma patients will be compared to normal-sighted persons. We expect to discover if glaucoma patients show sufficient scanning and compensatory eye movements for mobility demands.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

 

Fig 1. TREYESCAN setup