Purchase this article with an account.
JS McLellan, SA Burns; The Nonrandom Nature of Wave Aberrations is not a Sampling Artifact . Invest. Ophthalmol. Vis. Sci. 2002;43(13):1897.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: We have previously shown that the eye's measured wave aberrations (expressed as Zernike coefficients) produce consistently better MTFs than simulated sets of aberrations having the same RMS error, but randomized signs, for both monochromatic (OSA, '00) and polychromatic (ARVO '01) conditions. The current work investigates whether this effect may be an artifact of the coefficient-fitting process, due to smoothing of the measured wavefront. If this were the case, randomizing the coefficient signs could produce more turbulent wavefronts even though the RMS is maintained. Methods: Wave aberrations (530 nm) were measured for 40 subjects using a spatially resolved refractometer. Data at each wavelength were fit with Zernike polynomials up to the seventh order. Simulated aberration sets were produced by randomizing the signs of each subject's actual aberrations, while maintaining the RMS error for each term. MTFs were calculated for these simulations and compared to the actual MTFs. The wavefronts for all data sets and simulations were then computationally resampled to produce a new estimate of the aberrations for each, and MTFs were calculated as before. Results: Results are summarized by the MTF Ratio: real MTF / mean simulated MTF, as a function of spatial frequency. For the measured coefficients, the mean MTF Ratio is ≷1.0 up to 50 cpd. For both measured data and simulations, the RMS errors for the resampled wavefronts are lower than those of the originals, but the RMS of the resampled data differs from that of the resampled simulations by only ∼1%. The MTF ratio for the resampling shows the same relation as that of the original analysis. For both the original data set and the resampled data set, there is a peak elevation in MTF ratio between 5 and 15 cpd. Conclusion: The current results support the hypothesis that the eye's aberrations are not random. The eye's optical quality is better than would be expected for random sign aberrations of the same magnitude. This effect can not be ascribed to an artifact of the data fitting process.
This PDF is available to Subscribers Only