A new source of variance in visual acuity

In an ocularly healthy population, decimal visual acuities range from 1 to 3. We wondered how this spread can be understood. Using a maximum likelihood adaptive procedure, "visual acuity" was measured in a healthy population with four stimulus sets: unblurred Landolt Cs and Cs blurred with a Gaussian of width sigma = 2.12', 4.24' and 8.48'. A simple model based on scale invariance of the visual system was applied. This model was tested by predicting the outcomes of the 2.12' measurements based on the other measurements. The minimum angle of resolution (MAR) values found are closely proportional to "equivalent blur" of the stimulus defined as the convolution of a value for intrinsic blur of the eye and added blur. The proportionality factor is different between individuals and is an important source for the spread in acuities found in a healthy population. The differences between the proportionality factors are, interpreted as differences in the (neuronal) judgment capability between individuals. The total standard deviation of log(MAR) found in our study was 0.11. This value can be subdivided in 0.06 for the (neuronal) judgment capability, 0.08 for the intrinsic retinal blur and a measurement accuracy of 0.04. (C) 2003 Elsevier Ltd. All rights reserved.