Repeatability of objective measurements of fixational eye motion measured by Tracking Scanning Laser Ophthalmoscopy (TSLO) in healthy controls and comparisons with the Readalyzer

Abstract

Purpose : The purpose of this study was to compare TSLO measurements of fixational eye motion on two days separated by 1 week. 10-second recordings were acquired and microsaccade number, amplitude, and peak velocity were extracted for healthy control subjects and compared with outputs from the Readalyzer.

Methods : Twenty-one young, healthy control subjects (11 male, 10 female) aged 23-31 participated in the study. At least three 10-second videos of the subjects fixating the upper right hand corner of the 5° 840nm imaging raster were acquired for each eye on two separate days. Offline, custom Matlab software was used to stabilize the images, extract eye motion at 960Hz, and compute microsaccade number, amplitude, and peak velocity. All subjects completed the Readalyzer on the first day. Test-retest reliability was evaluated between days using the Intraclass Correlation Coefficient (ICC_2,2) as an indicator of the consistency of measurements. A linear regression model with age and sex weighted by the number of microsaccades compared TSLO and Readalyzer measurements.

Results : Each subject had a minimum of 2 videos from the left and/or right eye for each day, with a mean (SD) of 6.1 (2.2) and 7.1 (2.5) on days 1 and 2. Videos were analyzed if at least 7700 samples were present in the extracted 10-second eye trace. There were an average (SD) of 15 (7) microsaccades/video on day 1 and 16 (8) on day 2. Average peak velocity and saccade amplitudes for each video averaged (SD) 31.7deg/s (14.9) and 0.32° (0.19) on day 1 and 31.4deg/s (12.7) and 0.31° (0.17) on day 2. Test-retest reliability statistics for the TSLO measurements were excellent (peak velocity 0.965; amplitude 0.947). TSLO microsaccade average peak velocity and amplitude were both positively linearly related to Readalyzer outputs of fixations/100 words (p<0.0001; R²=0.341;0.396), regressions/100 words (p<0.0001; R²=0.378;0.444), Regression/Fixation Ratio (p<0.0001; R²=0.391;0.440), and negatively linearly related to Reading Rate (p<0.0001; R²=0.230;0.299).

Conclusions : Precise and direct microscopic measurements of fixational eye motion of healthy control subjects were reliable and consistent across days as measured with TSLO. Pre-clinical measurements of eye motion with TSLO may provide valuable insight into clinical test of oculomotor function.

This is a 2020 ARVO Annual Meeting abstract.