A statistical criterion to establish normal ranges for age in a contrast sensitivity function test

In a previous work (Opt. Appl. 39(2), 2009, pp. 415–428) we established the characteristics that a computer-based contrast sensitivity function (CSF) measurement system has to be used in the opthalmological clinic. In order to obtain a generalized use of CSF in clinics and as a screening tool, the necessity to incorporate a normality range by age was also suggested. It will also be important to establish how many reference curves are necessary, because in the last decades, different ranges have been presented in the literature. In the present work, our purpose was to show how to distribute the observers in terms of the statistical variations of CSF as a function of age in a normal population of healthy eyes. We then evaluated the utility of these curves in the detection of vision problems and, finally, the possibility of using them as a screening tool considering a reduced number of spatial frequencies. We used a computer-based CSF measurement system to present sinusoidal gratings whose values range from 1 to 24 cycles per degree. Three different groups (control, clinical and non-clinical) of subjects were considered. From the statistical analysis we obtained two ranges of normality, based on significant differences that appear around the age of 50. As we were interested in evaluating if this separation could increase the sensitivity of the test, we also performed a series of measurements in a clinical environment. As an interesting possibility of usage of a vision test is screening, we also measured people in conditions relatively different to those found in laboratories or clinics. We observed that this division into two ranges allows a better discrimination, especially for young adults. Measurements show an improvement of 22% in the detection of vision anomalies.