Clinical testing of contrast thresholds using a commercial television monitor system

Abstract

Background: The Medmont AT‐20 has incorporated a contrast threshold test using a predetermined letter size that can be applied in clinical settings. This paper describes a pilot study that evaluates this technology and the effects of certain parameters on test outcomes.Methods: A photometric calibration of the test was performed to define the relationship between the AT‐20 scale and Weber contrast (W%). We determined the effects of repeated measures (precision), target size (6/6 to 6/96), viewing duration (50 to 1,000 msec), defocus (+0.50 to +1.50 DS) and a macula scotoma on thresholds. The accuracy of the staircase (PEST) procedure was evaluated with and without false‐negative responses.Results: The AT‐20 scale has an almost linear relationship to a logarithmic transformation of W% and provides a suitable measure of contrast threshold. In the absence of monitor calibration, threshold uncertainty could be as great as 0.22 log units (W%) compared with published norms. We found that threshold variability averaged ± 7.1 AT‐20 scale units (95 per cent limits of agreement) and was proportional to threshold magnitude. One dioptre of defocus decreased thresholds by about one log unit (W%) for a 6/24 target. We propose that a 6/24 letter shown for 500 msec should provide a useful target for most clinical settings. The PEST procedure can yield endpoints in 47 (±12) seconds, is robust to false negative (FN) responses and gives abnormal thresholds in the presence of a macula scotoma.