Abstract
Purpose :
A large variation in the cone directionality, or ρ, value of the Stiles-Crawford Effect of the First Kind (SCE-I) has been reported depending on the experimental parameters, measurement approaches (e.g. psychophysical vs. optical imaging), retinal location and inter-subject differences. We hypothesize that the difference is, in part, due to the wavelength and exit pupil sizes used when capturing images. Here, empirical optical measurements are compared to numerical models to investigate this hypothesis and narrow the uncertainty in the interpretation.
Methods :
Measurements were performed at 4o temporal retina (TR) using an adaptive optics flood illuminated fundus camera with a moving entrance pupil across a distance of ±1.25 mm while capturing high-resolution retinal images through a dilated 6 mm pupil. Cone directionality was determined at 4o TR on a normal human subject using 650 and 700 nm imaging light. For comparison, numerical analysis was performed on idealized 5 µm diameter cones aligned towards the pupil center with negligible disarray. The experimental situation was simulated at wavelengths 550, 600, 650 and 700 nm allowing determination of the effective ρ values.
Results :
The ρ values from the AO imaging did not vary with wavelength and corresponded well with the numerical analysis. Overall, the results were in good agreement with most ρ values reported in the literature of approximately 0.05 mm-2. The model showed a weak wavelength dependence with smallest directionality at long wavelengths, while being also very sensitive to the exact diameter of the pupil used to capture the retinal images – Figure 1. For small pupils, the effective directionality was large, whereas for large pupils, the directionality was reduced.
Conclusions :
Our analysis shows that both wavelength and pupil size are critical parameters when analyzing the SCE-I with objective methods and that the directionality parameter found in this way is not strictly equal to the directionality determined with psychophysical methods.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.