Abstract
Purpose: To test the accuracy, tolerance and repeatability of the COAS aberrometer (Wavefront Sciences, Inc.) with model eyes and normal human eyes. Methods: Model eyes were constructed from six PMMA, single-surface lenses with known characteristics. Accuracy was evaluated by comparing theoretical predictions with measured spherical aberration, coma, defocus, and astigmatism. Tolerance to axial and lateral misalignment was measured by controlled displacements of the model eye relative to the aberrometer. Repeatability was tested with measurements taken within seconds or across several days on the same model eye. The same tolerance and repeatability experiments were then repeated on human eyes. Results: Accuracy of spherical aberration and coma agreed closely with theoretical predictions (e.g. for all six aspheric models, the mean difference between predicted and measured Z4\0 was 0.006 microns). Defocus and astigmatism were accurately measured (e.g. measured defocus was within +/− 0.25 diopters over a +/− 3 D range of refractive error). Axial displacements over the range +/− 2.5 cm had little effect on measurements for myopic and emmetropic model eyes. Also, lateral displacements over the range +/−1.5 mm did not produce significant coma. The standard deviations of repeated measurements of higher order RMS were less than 1% of the mean over seconds and less than 8% over days for model eyes. Tolerance to small lateral displacements was also observed for human eyes. Rotational (fixation) errors over a 3° range increased measurement variability, but not enough to fully account for long-term variations in human eyes. Conclusion: The COAS aberrometer can measure 2nd, 3rd, and 4th order aberrations accurately and repeatably on model eyes. Variability in human eye measurements is mostly due to fixational error and other physiological factors, rather than instrument fluctuations or errors of alignment.