Abstract
Purpose :
To develop optoretinography as a biomarker, we quantify the repeatability of the single cone optoretinogram (ORG) in healthy subjects within and across sessions.
Methods :
Healthy subjects (N=4, 28-41 years) were imaged with an adaptive optics scanning laser ophthalmoscope (AOSLO, 785 nm, 1° x 1° field at 1.5° temporal retina). In each of two sessions, 100 AOSLO videos of the cone mosaic were acquired with a visible stimulus (545Δ10 nm, 2.75µW/deg2, 0.056s exposure duration) presented 1s after the start of recording. Twenty control (non-stimulus) videos were also acquired. The confocal intensities of single cones were normalized and standardized (Cooper et al., 2020). Videos were randomly sorted into two groups of 50 and an inclusion criteria for each cone was set as: cone signals were successfully measured for at least 10 out of the first 18 frames following the stimulus onset in at least 13 of the 50 videos in each group and session. The single-cone ORG time course was calculated using a moving root mean square (RMS) over a 5-frame window, with subtraction of the corresponding quantity from the control videos. ORG amplitude was determined as the 95th percentile of the ORG within one second of stimulus onset.
Results :
Across subjects, 1007-2274 cones were included in the analysis. To investigate within and across session variation, ORG amplitudes were averaged over all cones. A one-way ANOVA showed no significant difference in the average ORG amplitude between groups in either session 1 [F (1,7) = 0.005, p = 0.943] or session 2 [F (1,7) = 0.604, p = .466]. A two-way ANOVA (Subject and Session) showed no significant difference across sessions [F (1,8) = 0.15, p = 0.709] nor an interaction [F (3,8) = 0.01, p = 0.998]. There was, however, a significant difference across subjects [F (3,8) = 5.27, p = 0.027]. Bland-Altman plots of the repeated ORG measures showed no systematic patterns, either within or across sessions. On average ~2% of cones remained in the lower 5th percentile of ORG amplitude while ~92% of cones remained in the upper 95th percentile, both within and across sessions.
Conclusions :
The single cone ORG exhibited reasonable repeatability within and across sessions. It was possible to identify cones that had consistently low responses and to produce a reliable population measure by averaging single cone ORGs. Measuring single cone response has the potential to localize visual function assessment across health and disease.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.