Purchase this article with an account.
Fang Hou, Luis A Lesmes, Woojae Kim, Hairong Gu, Mark Pitt, Jay Myung, Zhong-Lin Lu; Evaluating the sensitivity for detecting contrast sensitivity changes using the quick CSF method. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):3900.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The contrast sensitivity function (CSF) has shown promise for monitoring the progression of vision loss in eye disease or its remediation with treatment (Barnes, et al., 2004). The quick CSF is a novel Bayesian adaptive method developed to enable the wide, easy application of precise CSF testing. Here, we 1) induced predictable visual changes in a large sample of observers, 2) determined the quick CSF’s sensitivity for detecting CSF changes, and 3) used the large dataset to empirically determine the minimum sample sizes and testing times needed to detect CSF changes.
CSFs of 112 college students with normal vision were assessed by the quick CSF with a 10-letter identification task in low, medium, and high mean luminance conditions: 2.65 (L), 20.2 (M) and 95.4 (H) cd/m2. Visual acuity (VA) was measured using the Snellen eye chart. Data were analyzed using Bayesian and conventional statistics. For each observer, CSF metrics across different luminance conditions were calculated in each trial, via bootstrap statistics for the area under the log CSF (AULCSF), computed by resampling from the Bayesian posterior distribution of the CSF.
1) Average AULCSF values obtained in the L and M conditions were only 38.7% and 72.1% of those obtained in the H condition, while average relative VA values were 95% and 99%. Although both AULCSF (F(2,335)=293, p<0.001) and VA (F(2,332)=5.89, p<0.001) reductions were significant, the magnitudes of AULCSF changes were much greater than those of VA reduction (F(1,443)=461, p<0.001).<br /> 2) After 10 trials, we needed only 1, 3, and 17 subjects to reliably (with 95% posterior probability) detect AULCSF changes in comparisons of L-H, L-M, and M-H conditions, respectively; after 25 trials, we needed only 1 subject to detect all the changes. In contrast, we needed 16, 20, and 82 subjects to reliably detect VA changes in those comparisons<br /> 3) With 10 subjects, we needed 3, 4, and 11 quick CSF trials to reliably detect AULCSF changes between the L-H, L-M and M-H luminance conditions, respectively. With 25 subjects, we needed only 2, 3, and 7 trials.
This assay calibration study demonstrates that the quick CSF method is very sensitive in detecting CSF changes. Whereas the current study uses the quick CSF to detect visual changes modeled by different luminance conditions, future studies will examine the performance of the method in clinical settings.
This PDF is available to Subscribers Only