Abstract
Purpose :
Retinoscopy and eccentric photorefraction (EPR) are similar and very useful methods for obtaining objective estimates of refractive errors. In 1998, Roorda employed the EPR and a Badal optometer to achieve higher sensitivity and accuracy for spherical refraction [1]. However, astigmatism measurements have not been successful. Because a single measurement in EPR and retinoscopy provides only information along one specifically selected meridian, a sphero-cylindrical refraction would require no less than three measurements along independent meridians. In this study, we reconsider the feasibility and accuracy of achieving astigmatism in EPR.
Methods :
We present three possible mathematical models to describe astigmatism of the human eyes: the elliptical radius, the elliptical power, and the pure 2nd-order wavefront aberration models. Both elliptical models are corrected to the ocular axial length; i.e., their base radius and base power are spherical-equivalent dependent. The discrepancies between the models and the potential errors caused by the 3-meridian analysis in EPR are investigated mathematically.
Results :
For the same degree of sphero-cylinder prescription, we examine the refractive power beyond the two principle meridians. The elliptical radius model is always more hyperopic than the wavefront model and the elliptical-power model is always slightly more myopic. The maximum discrepancy occurs at meridian around midway of the two principle meridians, and it increases with roughly the square power of the cylinder. Applying the wavefront model to analyze the 3-meridian data of an elliptical-radius-like eye would result in errors on both sphere and cylinder. These errors could be significant at a high astigmatism, and the error is more significant for hyperopia than myopia (for example, for a 10D cylinder, the sphere/cylinder error are 0.52D/-1.04D for hyperopia vs 0.25D/-0.49D for myopia).
Conclusions :
The elliptical radius model may portray human eyes with the most proper degree of spherical aberration. The wavefront model may represent the spectacle and contact lens correction of refractive error the best. From our study, the three models are all capable to provide proper prediction of astigmatism for EPR measurement with error less than 0.25D If human eyes are not presented with significant amount of high-order aberration and the astigmatism is less than 5D.
[1] A. Roorda, W. R. Bobier, M. C. W. Cambell, Vis. Res. 38, 1934 (1998)
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.